Accord Study

Protocol
(May 11, 2005 Version)
ACCORD Protocol – May 11, 2005 Version

Action to Control Cardiovascular Risk in Diabetes
(ACCORD) Trial
Protocol
Page Number
Abstract .............................................................................................................................. 1
Chapter 1 – Introduction and Background.......................................................................... 3
Chapter 2 – Participant Selection and Follow-up ............................................................. 28
Chapter 3 – Interventions.................................................................................................. 53
Chapter 4 – Participant Safety and Confidentiality .......................................................... 82
Chapter 5 – Clinical Outcome Measures .......................................................................... 87
Chapter 6 – Health-Related Quality of Life/Cost-Effectiveness Outcome Measures ...... 92
Chapter 7 – Statistical Considerations............................................................................ 101
Chapter 8 – Data Management and Training.................................................................. 112
Chapter 9 – Adherence ................................................................................................... 115
Chapter 10 – Ancillary Studies ...................................................................................... 118
Chapter 11 – Publication Policy ..................................................................................... 120
Chapter 12 – Quality Control ......................................................................................... 123
Chapter 13 – Study Organization ................................................................................... 127
Chapter 14 – References.................................................................................................. 133
Appendices:
I. Model Informed Consent Document…………………………………….152
II. Charges to the Subcommittees of the ACCORD Steering Committee….163
III. ACCORD Conflict of Interest Policy – General Principles……………..166

1
ABSTRACT
Patients with type 2 diabetes mellitus die of cardiovascular disease (CVD) at rates two to
four times higher than non-diabetic populations of similar demographic characteristics. They also
experience increased rates of nonfatal myocardial infarction and stroke. With the growing
prevalence of obesity in the United States, CVD associated with type 2 diabetes is expected to
become an even greater public health challenge in the coming decades than it is now. Expected
increases in event rates will be associated with a concomitant rise in suffering and resource
utilization. Despite the importance of this health problem in the North American population,
there is a lack of definitive data on the effects of intensive control of glycemia and other CVD
risk factors on CVD event rates in diabetic patients.
The overall goal of the Action to Control Cardiovascular Risk in Diabetes (ACCORD)
trial is to address this challenge by testing three complementary medical treatment strategies for
type 2 diabetes to enhance the options for reducing the still very high rate of major CVD
morbidity and mortality in this disease.
The design is a randomized, multicenter, double 2 X 2 factorial design in 10,000 patients

with type 2 diabetes mellitus. The trial is designed to test the effects on major CVD events of
intensive glycemia control, of treatment to increase HDL-cholesterol and lower triglycerides (in
the context of good LDL-C and glycemia control), and of intensive blood pressure control (in the
context of good glycemia control). All 10,000 participants will be in the overarching glycemia
trial. In addition, one 2 X 2 trial will also address the lipid question in 5,800 of the participants
and the other 2 X 2 trial will address the blood pressure question in 4,200 of the participants.
The three specific primary ACCORD hypotheses are as follow. In middle-aged or older
people with type 2 diabetes who are at high risk for having a cardiovascular disease (CVD) event
because of existing clinical or subclinical CVD or CVD risk factors:
(1) does a therapeutic strategy that targets a HbA1c of < 6.0% reduce the rate of CVD
events more than a strategy that targets a HbA1c of 7.0% to 7.9% (with the
expectation of achieving a median level of 7.5%) ?
(2) in the context of good glycemic control, does a therapeutic strategy that uses a fibrate
to raise HDL-C/lower triglyceride levels and uses a statin for treatment of LDL-C
reduce the rate of CVD events compared to a strategy that only uses a statin for
treatment of LDL-C?
(3) In the context of good glycemic control, does a therapeutic strategy that targets a
systolic blood pressure (SBP) of < 120 mm Hg reduce the rate of CVD events
compared to a strategy that targets a SBP of < 140 mm Hg?
The primary outcome measure for the trial is the first occurrence of a major
cardiovascular disease event, specifically nonfatal myocardial infarction, nonfatal stroke, or
cardiovascular death.

2
The ACCORD study is designed to have:

• 89% power to detect a 15% treatment effect of intensive glycemic control compared
with standard glycemic control,
• 87% power to detect a 20% treatment effect of lipid control through LDL-C treatment
and fibrates compared with lipid control using LDL-C treatment alone,
• 94% power to detect a 20% treatment effect of intensive blood pressure control
compared with standard blood pressure control.
Secondary hypotheses include treatment differences in other cardiovascular outcomes,

total mortality, microvascular outcomes, health-related quality of life, and cost-effectiveness.
The 10,000 participants will be treated and followed for about 4 to 8 years (approximate
mean of 5.6 years) at approximately 60 Clinical Sites administratively located within 7 Clinical
Center Networks in the United States and Canada. Recruitment will occur in two non-
contiguous periods: an initial period that began in January 2001 for the Vanguard Phase of the
trial (during which 1184 participants were randomized) and then a subsequent period beginning
in January 2003 and ending in September 2005. Follow-up is scheduled to end in June 2009,
with the primary results announced in early 2010.

3
Chapter 1
Introduction and Background
Patients with type 2 diabetes mellitus die of cardiovascular disease (CVD) at rates two to
four times higher than nondiabetic populations of similar demographic characteristics. They also
experience increased rates of nonfatal myocardial infarction and stroke. Diabetes is a complex
metabolic disorder with abnormalities in carbohydrate, lipid, and protein metabolism, often
accompanied by other CVD risk factor abnormalities, such as elevated blood pressure. The
combination of diabetes with hypertension and/or dyslipidemia confers a much higher risk than
each one alone. Diabetes increases the risk of cardiovascular events two-to-three-fold at every
level of blood pressure (BP) and total serum cholesterol, and in diabetic patients there is a graded
increase in risk across the ranges of BP and total serum cholesterol. In addition, patients with
type 2 diabetes often have low plasma HDL-cholesterol levels, putting them at increased risk for
CVD, and there are data supporting a role for lowering triglycerides and raising HDL-cholesterol
levels for primary and secondary prevention of CVD in diabetic patients.
With the growing prevalence of obesity in the United States, CVD associated with type 2
diabetes is expected to become an even greater public health challenge in the coming decades
than it is now. Expected increases in event rates will be associated with a concomitant rise in
suffering and resource utilization. Despite the importance of this health problem in the North
American population, there is a lack of definitive data on the effects of intensive control of
glycemia and other CVD risk factors on CVD event rates in diabetic patients. Scientists on three
panels convened or sponsored by the National Institutes of Health since 1997 concluded a trial
was needed to determine the effects on macrovascular disease of aggressive glycemic, lipid,
and/or blood pressure control in type 2 diabetic patients.
The overall goal of the Action to Control Cardiovascular Risk in Diabetes (ACCORD)
trial is to address this problem by testing three complementary medical treatment strategies for
type 2 diabetes to enhance the options for reducing the still very high rate of major CVD
morbidity and mortality in this disease. The design is a randomized, multicenter, double 2 X 2
factorial design in 10,000 patients with type 2 diabetes mellitus. The trial is designed to test the
effects on major CVD events of intensive glycemia control, of treatment to increase HDL-
cholesterol and lower triglycerides (in the context of good LDL-C and glycemia control), and of
intensive blood pressure control (in the context of good glycemia control). All 10,000
participants will be in the overarching glycemia trial. In addition, one 2 X 2 trial will also
address the lipid question in 5,800 of the participants and the other 2 X 2 trial will address the
blood pressure question in 4,200 of the participants. Thus each participant will be in a 2 X 2 trial
testing 2 treatment strategies of 2 interventions, one of which is always glycemic control and the
other is either lipid or blood pressure control.
The primary outcome measure for the trial is the first occurrence of a major
cardiovascular event, specifically nonfatal myocardial infarction, nonfatal stroke, or
cardiovascular death. Participants will be recruited over two non-contiguous periods (described
in Sections 1.6 and 7.1) and followed for about 4 to 8 years (approximate mean of 5.6 years).

4
The three primary ACCORD hypotheses are to determine if the rate of major
cardiovascular events in type 2 diabetic patients at increased risk for CVD can be reduced by:
(1) Intensive glycemic control compared with standard glycemic control
(2) Lipid control through drug treatment to raise HDL-C and lower triglyceride levels in
the context of LDL-C treatment compared to LDL-C treatment alone.
(3) Intensive blood pressure control compared with standard blood pressure control

If more than one of the more intensive treatment groups experience significantly lower
major CVD event rates than the respective control groups, clinicians' choices may be further
guided by 1) effects on secondary clinical outcomes, including microvascular disease, adverse
effects, and quality of life; 2) subgroup analyses of effects in combined versus single factor
approaches; 3) resource requirements, including medical care costs; and 4) patient acceptance
and tolerance of various classes of medications.
1.1 Diabetes and Glycemia
1.1.a Diabetes and Cardiovascular Disease
Diabetes mellitus is a common disorder that is frequently misunderstood and poorly

treated. Although usually thought of in terms of the acute symptoms and long-term
consequences associated with elevated glucose levels, diabetes is a complex metabolic disorder
with abnormalities in carbohydrate, lipid, and protein metabolism. It is not a single disease and
although common forms of diabetes are associated with an increased risk of CVD, the type of
diabetes may have implications for the approach to preventing its cardiovascular complications.
Current recommendations for CVD prevention generally apply to both type 1 and type 2
diabetes. In the United States, approximately 10 percent of diabetic patients have type 1 diabetes
(previously called insulin-dependent diabetes, or IDDM) while approximately 90 percent have
type 2 diabetes (previously called non-insulin-dependent diabetes, or NIDDM). The optimal
treatment for the young, type 1 diabetic patient with severe, labile hyperglycemia may not be the
best treatment for the older, type 2 diabetic with mild, stable glucose elevations. Type 1 diabetes
is characterized by severe insulin deficiency, and restoration of normal glucose levels by
intensive insulin therapy may be more successful in reducing risk of all chronic complications of
this disorder. Type 2 diabetes, a more complex disease with generally elevated levels of insulin
resistance and variable levels of circulating insulin, is often accompanied by multiple other CVD
risk factor abnormalities, such as elevated blood pressures and lipids. While glucose control also
appears important for type 2 patients, it is critical not to overlook treatment of these other CVD
risk factors, which may have a greater or lesser effect than glucose control on prevention of CVD
complications. Nevertheless, several recent studies indicate that in clinical practice neither
hyperglycemia nor other CVD risk factors are adequately controlled in patients with diabetes
(Savage 1998).

5
Declines in CVD mortality in the United States in the past 30 years have been smaller
among diabetic patients than among non-diabetic patients. Compared to their non-diabetic
counterparts, the relative risk of CVD for men with diabetes is 2 to 3, and for women with
diabetes is 3 to 4 (Stamler 1993, Kannel 1979, Fuller 1983, Barrett-Connor 1991, Goldbourt
1993, Manson 1991). Population-based studies suggest that approximately 45% of white adults
with diabetes have coronary heart disease compared to 25% in non-diabetic individuals (Wingard
1995). The annual risk of fatal and nonfatal CVD in middle-aged diabetic individuals is 2 to 5%
(Stamler 1993, Morrish 1991, ETDRS Investigators 1992, Damsgaard 1992, Neil 1993). This
risk is independent of the risk associated with other risk factors such as hypercholesterolemia,
smoking, and hypertension (Stamler 1993). Diabetic patients with other CVD risk factors are at
greater risk than non-diabetic individuals. Data collected in the recent Heart Outcomes
Prevention Evaluation Study (HOPE 2000) confirm these high risks and show that they apply
even in 1999, despite the use of therapies proven to reduce CVD risk. (At baseline, 56% of
placebo patients were on aspirin, 20% on diuretics, 29% on beta-blockers and 22% on lipid
lowering agents.) In this large multicenter trial, 1769 high-risk people with diabetes but without
clinical CVD who were randomized to placebo experienced a 4.5-year rate of myocardial
infarction, stroke or cardiovascular death of 19.8% (4.4%/year).
Patients with diabetes also have an even worse prognosis following a cardiovascular
event. Prospective studies report that the relative risk of mortality following a myocardial
infarction is 2 to 3 times higher in diabetic compared to non-diabetic individuals (Behar 1997,
Mak 1997). This higher risk also applies to diabetic patients with unstable angina. In a recent
unpublished analysis of data from the international OASIS registry (Yusuf 1998) of hospitalized
unstable angina patients (21% with diabetes), the relative risk for MI, stroke, or CVD mortality
within 2 years of admission was two-fold higher (RR=1.8; 95%CI 1.6-2.2) in diabetic patients
compared to non-diabetic patients; the absolute rate in diabetic patients was 16.9% (versus 9.7%
in non-diabetic patients).
1.1.b Glucose as a Continuous Risk Factor for Cardiovascular Disease
Diabetes is a metabolic disease characterized by hyperglycemia, in which the defining

glucose cutoffs are those that predict a high subsequent risk of eye and kidney disease (Expert
Committee on the Diagnosis and Classification of Diabetes Mellitus 1997). As noted above,
people with diabetic-range hyperglycemia are also at high risk for CVD. This suggests that
hyperglycemia is also a risk factor for CVD. Indeed, large prospective epidemiologic studies
(summarized in Table 1.1) have consistently shown that in patients with diabetes, the higher the
glucose, the higher the incidence of CVD (Moss 1994, Kuusisto 1994, Andersson 1995, Gall
1995, Agewall 1997, Turner 1998, Wei 1998, Hadden 1997, Fu 1993). Taken together, these
studies suggest that the risk of a CVD event rises approximately 10-30% for every 1% increase
in HbA1c. This estimate is supported by the UKPDS observational results showing that the
incidence of MI rose 14% per 1% rise in HbA1c (Stratton 2000).
As the diagnostic thresholds for diabetes were not chosen on the basis of predicting a
high CVD risk, there is no a priori reason that the glucose-CVD risk relationship should not
extend below these microvascular risk cutoffs that characterize diabetes. Indeed, recent
epidemiologic studies have clearly shown that this relationship extends well below diabetic
glucose thresholds and may extend down to normal fasting and postprandial levels (Coutinho
6
1999, Gerstein 1999, Gerstein 1996, Gerstein 1997, Balkau 1998, Bjornholt 1999, Haffner
1998). These observations strongly support the hypothesis that lowering glucose to levels within
the normal, non-diabetic range may prevent CVD. The mean normal fasting and 2-hour post-load
plasma glucose levels are 92 mg/dl (5.11 mmol/l) and 97 mg/dl (5.39 mmol/l), respectively
(Cowie 1995).
Table 1.1: Relationship Between Glycemia and Risk of CVD in type 2 Diabetes
Study N Age F/U, Glycemia Outcome Rate (%) Relative RR / 1%
yrs Risk HbA1c
Increase
Andersson 411 66 7.4 SMBG> 7.8 mmol/l Death 44 vs 32 1.4 N/A

(1995) (140.4 mg/dl) vs <7.8
mmol/l
Kuusisto 229 68 3.5 HbA1c>7 vs <7 CHD 12 vs 3 4.3 N/A

(1994) HbA1c>7 vs <7 Death 20 vs 13 1.6
All CHD
Gall (1995) 328 56 5.3 HbA1c>7.8 vs <7.8 CV Death 10.4 vs 4.6 2.2 1.3
Agewall 94 67 6.3 N/A CV Death N/A N/A 1.54

(1997)
Lehto 1059 58 7.2 HbA1c>10.7 vs<10.7 CHD N/A 1.4 N/A

(1997) Death
Wei (1998) 4875 52 7.5 FPG 8-11.5 mmol/l CV Death 6.3 vs 2.8 2.9 N/A
(144 –207 mg/dl) vs
<8 mmol/l
Turner 3055 52 7.9 HbA1c >7.5 vs <6.2 Fatal MI N/A 1.72 N/A
(1998) Any 1.52 1.11
MI/angina
Moss (1994) 1780 66.6 8.3 N/A IHD N/A N/A 1.1
Death 1.17
Stroke
Death
Fu (1993) 479 61.2 4 HbA1c>8.4 vs <6.3 ECG 30.8 vs 1.5 1.17

MI/angina 20.3
N: sample size; FPG: fasting plasma glucose; IHD: ischemic heart disease; F/U: follow-up; RR: relative
risk; SMBG = self monitoring of blood glucose.
1.1.c Glucose Reduction to Lower the Risk of Cardiovascular Disease
The possibility that blood glucose level may be a modifiable CVD risk factor is supported
by the above epidemiologic data. It is also supported by a growing body of data from clinical
trials (Table 1.2). The UKPDS is the first trial to show that a policy of intensive glycemic control
using oral agents or insulin can reduce clinical outcomes in patients with type 2 diabetes. In the
main study of 3867 individuals with newly diagnosed type 2 diabetes, a fasting plasma glucose
<6 mmol/L (108mg/dl) was targeted by initial therapy with either a sulfonylurea (SU) or insulin.
7
Other agents were added when needed. Using this approach, the intensive group achieved a
median HbA1c of 7.0% (interquartile range 6.2-8.2%) over a 10-year period and experienced a
25% relative risk reduction (RRR) in microvascular outcomes and a 12% RRR in all diabetes-
related endpoints compared to a policy that achieved a median HbA1c of 7.9% during this period
(UKPDS 1998a). There was a strong trend towards a reduced risk of MI with an observed RRR
of 16% (95%CI 0%-29%;P=0.052). This 16% RRR for MI per 0.9% decrease in median HbA1c
over 10 years is consistent with the 18% RRR in MI per 1% decrease in HbA1c observed in the
UKPDS and other epidemiologic analysis (Section 1.1 and Table 1.1). As noted in Table 1.2, the
results for the group initially randomized to insulin were similar to the results for the intensive
group as a whole.
Unfortunately, stable degrees of glucose control in either of the randomized groups could
not be maintained. Therefore, the median HbA1c in the intervention group was 6.6% (IQR 5.9-
7.5%), 7.5% (IQR 6.6-8.8%) and 8.1% (IQR 7.0-9.4%) in the first, second and third 5-year
intervals respectively. The median HbA1c in the conventional group during these 3 periods was
7.4%, 8.4% and 8.7% respectively. Expressed differently, 50% of newly diagnosed UKPDS
participants in the intensive group had HbA1c values >7.0% during the first 10 years of follow-
up, and 25% had values >8.2% during this period. In essence, the UKPDS showed that delaying
the rise in HbA1c by 5 years and maintaining good control for at least the first 5 years led to
clinically important differences in CVD events.
A separate randomization of 1704 obese participants in 15 UKPDS centers allocated 342

participants to intensive control with metformin, 951 to sulfonylureas or insulin and 411 to
conventional control (UKPDS 1998). The median HbA1c was 7.4% in the metformin/other
intensive group and 8.0% in the conventional group during the first 10 years of follow-up.
Despite a more modest separation in HbA1c, the metformin group had a 32% risk reduction in
any diabetes-related endpoint, a 42% risk reduction in diabetes-related death, a 36% risk
reduction in all-cause mortality, and a 39% risk reduction in MI. There was a nonsignificant 29%
reduction in microvascular outcomes. Conversely, intensive therapy with sulfonylureas or insulin
did not significantly reduce outcomes. Indeed, the metformin group had statistically better
outcomes than the other intensive groups for any diabetes-related endpoint, all-cause mortality,
and stroke. However, if the results of the metformin arm are combined with the
sulfonylurea/insulin arm, the results support the cardiovascular benefit of glucose lowering. In
this analysis, the risk of myocardial infarction or stroke in the metformin intensive group would
be 19.3% and the risk in the conventional group would be 23.4% (relative risk reduction = 18%).
Similarly, the relative risk reduction for the combined outcome of myocardial infarction, stroke,
and cardiovascular death (assuming that cardiovascular death accounted for 80%
of all deaths) was 21.5%. As noted above, the observed median difference in HbA1c was 0.6%
(UKPDS, 1998b).
Despite the impressive results with the obese patients in the metformin study, another
randomization of obese and non-obese intensive group participants in which metformin was
added to a sulfonylurea if the fasting plasma glucose was 6.1-15 mmol/l (109.8–270 mg/dl) led
to 96% increase in diabetes-related deaths, and a 60% increase in all-cause mortality. This
surprising observation was not apparent after a combined analysis with the treatment group
starting with metformin and with epidemiologic analysis of the data, and remains unexplained.

8
Nevertheless, it increases uncertainty regarding the best treatment approach for patients with
type 2 diabetes.
Taken together these UKPDS reports show that a policy of improving glycemia in
patients with type 2 diabetes reduces clinically important outcomes. The benefit is especially
clear for microvascular disease, although there is a trend towards reduced macrovascular disease.
In light of strong epidemiologic evidence that the risk of CVD rises as the glucose level rises,
and the results of the UKPDS, it is likely that the CVD outcomes would have been reduced to a
greater degree had stable tight glycemic control been achieved in the intervention group. This
hypothesis clearly requires testing in prospective trials of high-risk patients followed for several
years, and is the primary basis for the ACCORD Trial.
In addition to the UKPDS, other trials of tight glycemic control in patients with diabetes
further support the hypothesis that tight glycemic control is cardio-protective (Table 1.2). The
Kumamoto study of insulin-mediated intensive control in thin patients with type 2 diabetes
reported a CVD event rate of 0.6/100 patient-years in the intensive group and 1.3 in the
conventional group (Ohkubo 1995) (i.e. a nonsignificant RRR of 46%). In the DIGAMI study of
insulin-mediated glycemic control after a myocardial infarction, a HbA1c of 7.1% vs 7.9% after
1 year of therapy was associated with a 29% lower mortality rate (Malmberg 1995). In the
variable insulin dose arm of the UGDP study, there was also a nonsignificant trend in favor of
reduced CV deaths (Genuth 1996). This controversial study reported an increased CVD mortality
in a tolbutamide arm after 6 years, which was therefore discontinued. Finally, a recent meta-
analysis of all intervention studies in patients with type 1 diabetes showed that intensive therapy
with insulin reduced macrovascular events by 45% (95%CI 22%-65%) (Lawson 1999) and the
development of a first event by 28% (P=NS). Although these studies were not powered to detect
an effect of tight control on CVD outcomes, the results of this meta-analysis also support the
hypothesis that glucose-lowering may prevent CVD outcomes.
In contrast to the evidence cited above, the possibility that intensive glycemic control
may worsen CVD outcomes was raised by the feasibility phase of the VACS-DM trial, in which
the intensively treated group had a nonsignificant increase in the risk of CVD events (Abraira
1997). This observation remains unexplained, but may have been related to the short duration of
the trial, the use of a sulfonylurea-class drug in the intensive group but not in the conventional
group (or of the sulfonylurea used, glipizide), or the relatively few events. Nevertheless, the
results highlight residual uncertainty regarding the potential CVD benefits of glycemic control,
and the importance of testing if glycemic control with various strategies prevents CVD events.

9
Table 1.2: Glucose Lowering Trials and CVD in People with Diabetes
Study Yrs HbA1c HbA1c Therapy Outcome Relative Risk RRR

(Intense) (Control) Reduction (micro)
(RRR) (CVD)
UKPDS (1998) 10 7.0% (113%) 7.9%(127%) Insulin/SU MI 16% (0,29) 25%
UKPDS (1998) 10.7 7.4%(119%) 8.0% (129%) Metformin MI 39% (11,59) 29%(NS)
Kumamoto 6 7.1% (111%) 9.4%(147%) Insulin CV Events 46% (NS) 65%

(Ohkubo 1995)
VACSDM 2.3 7.1% (116%) 9.3% (152%) Insulin/SU CV Events -40% (5,-108) N/A
(Abraira 1997)
DIGAMI 1 7.1% (range) 7.9%(range) Insulin Mortality 29% (4,51) N/A

(Malmberg 1995)
UGDP(IVAR)* 12.5 FPG 130-146 FPG 170-186 Insulin CV Deaths 9% (NS) 9%(NS)
(Genuth 1996) (7.2-8.1 mmol/l) (9.4-10.3 mmol/l)
Type 1 DM** 2-7 7.6% 8.7% Insulin Any Event 45% (22,65) not
(Lawson 1999)
calculated
Type 1 DM** 2-7 7.6% 8.7% Insulin First Event 28% (-17,56)
(Lawson 1999)
SU: sulfonylurea; micro: microvascular disease; NS: not significant in the report; DM: diabetes mellitus
*From the variable insulin dose arm of the UGDP in which a fasting plasma glucose of 130-146 mg/dl (7.2-8.1 mM)
was achieved. Results are expressed as the reported value and the % above the upper limit of normal for the assay
used (different assays were used in different sites in DIGAMI). Results for surrogate markers are shown (eye exam,
poor visual acuity or severe retinal changes).
**From a meta-analysis of all studies of tight control in type 1 diabetes
1.1.d Glucose Reduction and Adverse Events
The risks of glucose reduction are mainly those of hypoglycemia and weight gain, and in
randomized trials of people with type 2 diabetes these risks are highest in insulin-treated
individuals. Table 1.3 lists the actual risks in major trials that suggest that between 2-3% of
patients with type 2 diabetes who achieve close to optimal glycemic control with intensified
insulin therapy will have a severe hypoglycemic reaction annually. This rate may change with
newer approaches to therapy and with increased self-management education.
In addition to the risks of glucose-lowering per se, adverse effects due to the agents used
to lower glucose may also occur. These effects include the possibility that SUs may increase the
risk of arrhythmias, especially in an ischemic myocardium (Smits 1995), that metformin
increases the risk of lactic acidosis and gastrointestinal symptoms, and that thiazolidinediones
increase the risk of liver toxicity and are associated with mild anemia and edema (DeFronzo
1999).

10
Table 1.3 : Risks of Tight Control with Insulin in Patients with type 2 Diabetes
Study HbA1c Hypoglycemia Mean Weight Gain
Severe Any
UKPDS (1998) 7.1% (115%)* 1.8%/yr 28%/yr 4 kg > control (10 yrs)
Kumamoto 7.1% (111%) 0% 1.9%/yr BMI incr 20.5 - 21.2 (6 yrs)

(Ohkubo 1995)
VACSDM (Abraira 1997) 7.3% (120%) 3%/yr 41%/yr Same as control
DIGAMI (Malmberg 1995) 7.0% Control N/A 1 kg/yr
UGDP(IVAR) FPG 6.7 mmol/l (120.6 3.2% N/A 0.2%/yr

(Genuth 1996) mg/dl)
NB: severe hypoglycemia is defined as an episode requiring third party assistance; *HbA1c in the group randomized
to initial therapy with insulin; not different from control rate (Malmberg et al, personal communication); variable
insulin group in the University Group Diabetes Program; fasting plasma glucose at the end of the study (mmol/l);
BMI: body mass index
1.1.e Rationale for a Trial of Glucose Lowering to Prevent Cardiovascular Disease
Epidemiologic and clinical trial evidence strongly support the hypothesis that glucose is a
modifiable risk factor for CVD in people with diabetes, and that achieving near-normal glycemia
will prevent CVD events. Unfortunately, the clinical trial data are insufficient to prove the
hypothesis and definitive conclusions regarding the results of a therapy cannot be made from
epidemiologic analyses alone because they do not correct for the possibility that outcome and
glycemic control may be confounded with other unmeasured variables. Possible reasons for a
failure to demonstrate a statistically significant benefit of glucose control on CVD risk in the
UKPDS include low MI event rates. For example, the rates of MI in the control and intervention
group in the UKPDS were 17.4 and 14.7/1000 patient-years respectively (UKPDS 1998a); the
3867 patients would have been sufficient to detect a 20% risk reduction (but not a lower
reduction) with approximately 80% power. The fact that normal glucose levels were not
maintained throughout the study in the intervention group is also a limitation.
Support for the benefits of glucose lowering are balanced by several concerns: a)
aggressive glycemic lowering has clear risks (see Section 1.1.d), b) there is no definitive clinical
trial evidence for CVD benefits of glucose lowering, c) there is no definitive clinical trial
evidence of microvascular benefits for HbA1c levels below 7-7.5%, d) the largest clinical trial
(the UKPDS) was done in relatively low-risk newly diagnosed individuals, and e) few data are
available regarding the CVD effect of glycemic control on high-risk older individuals with well-
established diabetes. These considerations strongly support the need to determine definitively the
CVD efficacy and risks of intensive glycemic control in people with type 2 diabetes.
ACCORD participants will be randomized to two targeted levels of glucose control.

Participants randomized to the intensive group will have a HbA1c target of < 6.0%. Patients

11
randomized to the conventional group will have a HbA1c target of 7.0% to 7.9% (with the
expectation of achieving a median level of 7.5%).
1.1.f Key Methodologic Questions
1.1.f.1 Can We Achieve a HbA1c Target of 6% in the ACCORD Intensive Group?
Prior to the ACCORD Vanguard, the answer to this question for middle-aged patients
with established diabetes was unknown, and attempts to achieve this target have not been
reported. Reasons for this include the following: a) the continuous relationship between
hyperglycemia and CVD in people with type 2 DM was not clearly described until after 1995; b)
until the UKPDS was published, there were concerns (from the UGDP, VACSDM and biologic
data) that tight glycemic control in people with type 2 diabetes, with insulin, SU or metformin
could increase the risk of CVD, hypoglycemia and weight gain, and there was considerable
debate over whether or not it would decrease the risk of microvascular disease; c) until recently,
the value of combining oral agents and insulin was unknown and discouraged, and there were
few data in support of such a strategy - for example the UKPDS and VACSDM were both
designed as monotherapy trials in which a second agent was added after the first one failed -
indeed in the UKPDS, a second agent was added only when the fasting plasma glucose exceeded
270 mg/dl (15 mmol/l); d) only SUs were available in the United States until metformin was
introduced - several other oral agents have been introduced since then; e) there was no point-of-
care HbA1c testing available in earlier trials; f) the intensity of follow-up in large trials may have
been inadequate to achieve tight metabolic control - for example, participants in the UKPDS
intensive group only attended clinic visits every 3-4 months and had HbA1c measured only
every 6 months (UKPDS 1998a), and in the Kumamoto study participants were seen only every
6 months (Ohkubo 1995); g) postprandial glucose levels were not explicitly targeted - for
example in the VACSDM study intensive therapy participants were asked to do twice daily
capillary blood sugar testing before meals and once weekly 3 am testing (Abraira 1995); h)
normal HbA1c values were not always targeted - the goal of therapy in the Kumamoto study was
a HbA1c < 7% (Ohkubo 1995).
The above shows that most large studies have been able to achieve good mean HbA1c
values of 7% using relatively simple monotherapy-based approaches with modest follow-up
protocols. Whether better levels can be achieved by the comprehensive intensive protocol
discussed above remains unknown and required testing in the Vanguard Phase of ACCORD. At
least 2 small studies have shown that normal HbA1c levels can be achieved in people with type 2
DM using insulin alone. In one study of 14 obese individuals (mean age 59; BMI 31 kg/m2) with
a mean HbA1c of 7.7%, twice daily insulin injections (without oral agents) reduced this value to
5.1% within 4 months (Henry 1993, Henry 1996). During this time, no severe (and only
minimally mild) hypoglycemic episodes occurred. However, mean weight increased by 9% (8.7
kg). In another small study of 14 individuals (mean age 50; mean glycated Hb 13%), continuous
subcutaneous insulin for a 3 week period achieved a normal glycated Hb of 8.1% (normal range
was 6.3%-8.2% in this assay) (Garvey 1985). In ACCORD, eligibility criteria have been
selected to enhance the likelihood of being able to achieve this target (see Section 2.1.a). The
ACCORD Vanguard Phase established the feasibility of achieving HbA1c levels less than 6.0%

12
in a substantial portion of patients when this level is the goal and there is the ability to use
multiple medications.
1.1.f.2 Is it Ethical to Target a HbA1c of 7.0 to 7.9% in the ACCORD Standard Group and
What is the Risk-Benefit Relationship?
The UKPDS reported that for newly diagnosed obese and non-obese people, a policy of
tight glycemic control over a median of 10 years reduced clinically important diabetes-related
endpoints and microvascular events by 12% and 25% respectively (UKPDS 1998a). The
absolute risk reductions for these outcomes were approximately 5% and 3% respectively over the
10-year period.
The HbA1c in that trial rose over the duration of follow-up; the median HbA1c during
the 10-year period was 7%, and 1 out of 4 patients had a value >8.2%. The median and 75th
percentiles of HbA1c during the 1st, 2nd and 3rd 5-year follow-up period are noted in Table 1.4.
The UKPDS also showed that for obese individuals, a policy of tight control starting with
metformin (which achieved a median HbA1c of 7.4% during a median follow-up period of 10.7
years) reduced the risk of diabetes-related endpoints, diabetes-related death, and MI by 32%,
42% and 39% respectively (compared to a HbA1c of 8%). The approximate absolute risk
reductions were 13%, 5%, and 7% respectively for these outcomes (UKPDS 1998b).
ACCORD will utilize a treatment protocol that will introduce metformin early and will
target a HbA1c of 7.0 to 7.9% (inclusive) in the standard group. This value is consistent with the
intervention group's values in the UKPDS analysis of the effect of metformin in obese people
(which showed a large absolute benefit on clinical endpoints). It is also consistent with a
substantial portion of the intervention group’s values in the UKPDS analysis of the effects of
sulfonylurea and insulin in obese and non-obese people (Table 1.4). Moreover, it is lower than
the HbA1c usually achieved in most people with type 2 diabetes and lower than the median
HbA1c noted at baseline in the ACCORD Vanguard Phase. Therefore, people in the ACCORD
standard group will have a drug and HbA1c treatment policy that is consistent with what was
proven effective in the UKPDS.
With this background, it is expected that participants in both the intensive and standard
group will be experiencing reductions in HbA1c levels in ACCORD and would thus be expected
to derive microvascular benefits from participating.

13
Table 1.4: UKPDS HbA1c Values
Population (Drug)/Period Median HbA1c (%) 25th - 75th %-ile % of participants

HbA1c cut-points above 7.5%
Obese/Non-obese (SU & Insulin)/ 7.0 6.2-8.2 >25%
1st 5 yrs 6.6 5.9-7.5 25%
2nd 5 yrs 7.5 6.6-8.8 50%
3rd 5 yrs 8.1 7.0-9.4 >50%
Obese (Metformin)/ 7.4 N/A N/A
1st 5 yrs 6.7 N/A N/A
2nd 5 yrs 7.9 N/A N/A
3rd 5 yrs 8.3 N/A N/A
For the standard group, the challenge is to minimize the risks of severe hypoglycemia,
while at the same time lowering glucose sufficiently to reduce the risk of microvascular events
from the risk which that group would have otherwise incurred if they had continued on their pre-
ACCORD glycemic therapy. Therefore, if there is no CV benefit to intensive glycemic control,
the risks of being treated intensively will likely outweigh the benefits; if there is a CV benefit of
intensive glycemic control, the risks of being treated conventionally may outweigh any benefits.
Table 1.5 lists estimated relative and absolute risks and benefits for various degrees of
glycemic control (i.e. HbA1c levels) that will be achieved in the standard group and is based on
several assumptions:
a) The baseline (pre-randomization) HbA1c will be 8.5% (which it was in the Vanguard).
b) The annual absolute risk of severe hypoglycemia will be greater than 2.5% per 1% fall in
HbA1c from baseline. Thus this estimate represents the minimum risk.
c) The microvascular benefits for ACCORD standard participants will be as noted in the
UKPDS epidemiologic analysis (Stratton 2000): a 37% relative risk reduction for every
1% fall in HbA1c.
d) There is a linear relationship between HbA1c and both risk and benefit within the HbA1c
range of 7.0% to 8.5%.
These estimates facilitate a comparison of the impact of 2 different final conventional

group HbA1c levels and illustrate the risk-benefit trade-off within the HbA1c range of 7.1% to
7.9%. As noted in Figure 1.1 that was derived from these estimates, the risk of severe
hypoglycemia (even when minimum estimates are used) clearly rises more steeply than the fall
in the risk of microvascular events. Thus, a final standard group median HbA1c of 7.7% versus
7.5% would lead to a 5-year absolute risk of microvascular events (i.e. mainly laser therapy) of
8% versus 7%. This 1% absolute risk reduction over 5 years would be countered by a 2.5%
absolute risk increase of severe hypoglycemia (i.e. 10% versus 12.5%) over 5 years.
14
Table 1.5: Effect of the Degree of Fall in HbA1c in Standard ACCORD Participants
Microvascular Events Severe
Absolute
Final Hypoglycemia
HbA1c fall
HbA1c RRR Annual ARR Absolute Risk Absolute Risk
from 8.5%
Annual 5 yr Annual 5 yr
None 8.5% 0 0 ~ 2.3%* 11.5% N/A N/A
0.8% 7.7% 30% 0.7%# 1.6% 8% >2% >10%

1% 7.5% 37%** 0.9%** 1.4%** 7% >2.5% >12.5%
1.2% 7.3% 44% 1% 1.3% 6.5% >3% >15%
1.4% 7.1% 52% 1.2% 1.1% 5.5% >3.5% >17.5%
RRR – relative risk reduction; ARR – absolute risk reduction; *The actual risk may be
greater for the older, high cardiovascular risk people in ACCORD than in the UKPDS. #
calculated as (2.3-1.6); ** from Stratton 2000.
Figure 1.1:
Projected Microvascular versus Severe Hypoglycemia Risks
20 Microvascular Events
15 Severe Hypoglycemia
5 Year Rate
(%)
10
5
0
7.1 7.3 7.5 7.7 7.9
Final HbA1c

15
1.1.f.3 Can ACCORD Achieve and Maintain an Absolute HbA1c Difference Approaching a
Target 1.5% Between the Intensive and Standard Groups?
Given the currently available data, a HbA1c difference of approximately 1.5% is

estimated to be a conservative target that would lead to a 15% or greater reduction in CVD
events. Although review of the UKPDS data of obese participants suggests that a lower
differential of 1% may be adequate, sufficient uncertainty regarding these estimates exists to
justify the 1.5% differential. Other trials such as the UKPDS (UKPDS 1998a, UKPDS 1998b),
Kumamoto study (Ohkubo 1995), VACSDM (Abraira 1997), UGDP (Genuth 1996) and
DIGAMI (Malmberg 1995) study have successfully achieved and maintained separation of the
HbA1c using different protocols with less flexibility and choice of therapy.
Targeting a between-group difference in HbA1c that is lower than 1.5% may jeopardize
ACCORD’s chance of achieving an adequate HbA1c difference. In light of the high importance
of achieving and maintaining a HbA1c difference that is sufficient to test the research question,
ACCORD has adopted a delta of 1.3% as an alert level. Greater separations are, however,
expected in response to the novel approaches to glycemic control that will be employed for both
treatment groups (see Section 3.2). Nevertheless, if the HbA1c separation falls below 1.3% in
participants with at least 2 years of follow-up, the progress of the trial will be carefully
scrutinized by the investigators, and actions will be taken to increase the separation. The Data
Safety and Monitoring Board will also monitor the HbA1c levels and separation.
1.2 Diabetes and Dyslipidemia
1.2.a Type 2 Diabetes and Dyslipidemia
One of the goals of ACCORD is to determine if more aggressive control of diabetic

dyslipidemia, specifically raising HDL-cholesterol and lowering triglycerides (TG), in the
context of desirable levels of LDL-cholesterol, will provide greater benefit than only having
desirable levels of LDL-cholesterol. The reason for choosing to address this question is that a
dyslipidemia characterized by low HDL-cholesterol and high TG levels, with average LDL-
cholesterol levels, is typical of type 2 diabetes mellitus. Albrink and coworkers first reported
links between hypertriglyceridemia and insulin resistance (Davidson 1965), but it was the work
of Reaven and Farquhar and their colleagues that clearly defined this link (Reaven 1967, Olefsky
1974). Since then, numerous investigators conducting either small, detailed physiologic studies
or larger epidemiologic studies have confirmed the relationship between type 2 diabetes, insulin
resistance and increased blood levels of very low density lipoprotein (VLDL) TG (Olefsky 1974,
Albrink 1980, Laws 1997, Howard 1998, Bonora 1998). In vivo studies of lipoprotein
metabolism have indicated that insulin resistant states are associated with increased assembly
and secretion of apoprotein B100 (apoB)-containing lipoproteins. Thus, increased secretion of
both VLDL TG and apoB (Sigurdsson 1976, Kissebah 1982, Ginsberg 1982, Howard 1983) is a
central abnormality in individuals with insulin-resistance/type 2 diabetes. It is believed that
increased free fatty acid flux to the liver in insulin-resistant individuals drives TG synthesis and
assembly of VLDL. Reduced activity of the key enzyme in triglyceride removal from plasma,
lipoprotein lipase (LpL), is also important. LpL is an insulin regulated enzyme in muscle and
fat, and has been shown to be modestly reduced in many patients with type 2 diabetes (Taskinen
1987). In patients with type 2 diabetes, hyperglycemia may contribute to increased VLDL
16
secretion as well, although correction of blood glucose levels seems to only partly reverse the
dyslipidemia (Ginsberg 1991).
Patients with type 2 diabetes have low plasma HDL-cholesterol levels, and this does not
seem to be related to either glycemic control or mode of treatment (Hollenbeck 1986, Gordon
1977). A consistent finding has been the inverse relationship between plasma insulin (or C-
peptide) and HDL- cholesterol levels (Uusitupa 1986). The degree of insulin resistance also
appears to be related inversely to HDL-C concentrations (Laakso 1990). Increased secretion of
apo B-containing lipoproteins could result in increased cholesterol ester transfer protein (CETP)-
mediated transfer of HDL cholesteryl esters to those lipoproteins (Tall 1986, Bagdade 1993), and
this would explain the reduced levels of plasma HDL-cholesterol in patients with type 2 diabetes.
The finding of triglyceride-enriched HDL-C particles in patients with this disorder supports this
scheme. Increased hepatic lipase (HL) activity may also contribute to the development of low
HDL-cholesterol (Horowitz 1993, Lamarche 1999). ApoAI and AII levels are reduced
consistently as well, and fractional catabolism of apoAI is increased in type 2 patients with low
HDL-C (Golay 1987), as it is in nondiabetic patients with similar lipoprotein profiles (Le 1988,
Nicoll 1980, Brinton 1991).
Small dense LDL-C are commonly present in patients with type 2 diabetes and is most
likely an integral part of the dyslipidemia of insulin resistance (Feingold 1992, Reaven 1993).
Thus, increased plasma levels of VLDL TG can stimulate CETP-mediated transfer of LDL
cholesteryl esters to VLDL in exchange for TG. The TG-enriched LDL-C is then modified by
LPL and/or HL, producing small dense LDL-C.
1.2.b Evidence that LDL-Cholesterol Lowering Reduces Cardiovascular Events
Several primary and secondary prevention trials have demonstrated remarkable

reductions in CHD events and mortality in high-risk patients, and the issue for ACCORD is how
to apply this evidence to LDL-C treatment goals in the lipid portion of the trial. Beginning with
the landmark Coronary Primary Prevention Trial in which cholestyramine was used (Lipid
Research Clinics Program 1984), and continuing through the secondary prevention trials of
HMG CoA-reductase inhibitors (or “statins”) such as 4S (Scandinavian Simvastatin Survival
Study Group 1994), CARE (Sacks 1998), LIPID (LIPID 1998) and the primary prevention statin
trials such as WOSCOPS (Shepherd 1995) and AFCAPS/TEXCAPS (Downs 1998), treatment to
lower LDL-C has resulted in consistent reductions in cardiovascular events. The data from the
trials validate the algorithm for cholesterol treatment suggested by the National Cholesterol
Education Program (NCEP) Adult Treatment Panel III that sets initiation and treatment targets
for cholesterol lowering in patients with various levels of risk. The rationale of the NCEP panel
was that those at the highest risk would benefit the most and warrant the most aggressive
treatment. Characteristics of the landmark statin trials published through 1998 are summarized in
Table 1.6.

17
Table 1.6: Landmark Statin Trials

LDL-Cholesterol (mg/dl)
Population Trial N Eligibility Mean On Placebo Mean On Statin Agent
Primary Prevention
HIGH LDL-C WOSCOPS 6595 >155 (4.03 mmol/l) 191(4.97 mmol/l) 140 (3.64 mmol/l) Pravastatin
LOW LDL-C AFCAPS* 6605 > 125 (3.25 mmol/l) 156 (4.05 mmol/l) 115 (2.99 mmol/l) Lovastatin
Secondary Prevention
HIGH LDL-C 4S 4444 >155 (4.03 mmol/l) 186 (4.84 mmol/l) 121 (3.15 mmol/l) Simvastatin
LOW LDL-C CARE 4159 >115 (2.99 mmol/l) 135 (3.51 mmol/l) 98 (2.54 mmol/l) Pravastatin
LOW LDL-C LIPID 9014 >115 (2.99 mmol/l) 150 (3.9 mmol/l) 113 (2.94 mmol/l) Pravastatin
N: sample size *AFCAPS participants were particularly healthy: no unstable hypertension; diabetic patients with
HbA1c > 20% upper limit of normal excluded.
Table 1.7 summarizes events in the major primary and secondary prevention statin trials.
Table 1.7: Event Rates in Statin Trials.

Event Rate
Population Trial N Placebo Statin Event Definition
Primary Prevention
HIGH LDL-C WOSCOPS 6595 248/3293=7.9%=1.6%/yr 174/3302=5.5%=1.1%/yr Fatal CHD + NFMI
LOW LDL-C AFCAPS 6605 183/3301=5.5%=1.1%/yr 116/3304=3.5%=0.7%yr Fatal + NFMI, USA,

sudden death
HIGH LDL-C 4S 4444 622/2223=28%=5.2%/yr 431/2221=19%=3.5%yr * Fatal CHD + NFMI
LOW LDL-C CARE 4159 274/2078=13%=2.6%/yr 212/2081=10%=2%/yr Fatal CHD + NFMI
LOW LDL-C LIPID 9014 715/4502=15%=2.6%/yr 557/4512=12%=2%/yr Fatal CHD + NFMI
N: sample size; MFMI: nonfatal MI; USA: unstable angina
* = 2%/yr for the 40% of 4S participants whose LDL-C was <95 mg/dl (2.47 mmol/l) on treatment
1.2.c Evidence that LDL-Cholesterol Lowering Reduces Cardiovascular Events in People

with type 2 Diabetes
In several of the secondary prevention studies there were small numbers of individuals
with type 2 diabetes. Subgroup analyses show very high rates of CHD events and mortality in
patients with type 2 diabetes, and demonstrate substantial reductions in outcomes in the treated
groups consistent with overall results of these trials. (LIPID 1998, Goldberg 1998, Haffner
1999). The results of those subgroup analyses are presented in Table 1.8.

18
Table 1.8: Event Rates in Diabetic Patients in Statin Trials

# Event Rate
Diabetic
Trial Placebo Statin Event Definition
Patients
Primary Prevention
HIGH LDL-C WOSCOPS 70 --- ---
LOW LDL-C AFCAPS 155 6/71 = 8.4% = 1.6%/yr 4/84 = 4.8% = 0.9%/yr Fatal + NFMI, USA,
sudden death
HIGH LDL-C 4S 202 44/97=45% = 8.4%/yr 24/105=22%= 4.2%/yr Fatal CHD + NFMI
LOW LDL-C CARE 602 112/304=37%= 7.4%/yr 81/282=29%= 5.8%/yr Fatal CHD + NFMI
+revasc
LOW LDL-C CARE* 3.9%/yr 3.1%/yr Fatal CHD + NFMI*
LOW LDL-C LIPID 782 88/386=23% = 3.8%/yr 76/396=19%= 3.1%/yr Fatal CHD + NFMI
Abbreviations are in Table 1.7. *Data from CARE are corrected assuming that the proportion of events attributable
to revascularizations in diabetic patients equals that in non-diabetic patients. CARE and 4S data for diabetic patients
have been published separately (Goldberg 1998, Haffner 1999). Rates in WOSCOPS too small to report.
1.2.d Evidence that Going Beyond LDL-Cholesterol Lowering to Raise HDL-Cholesterol

and Lower Triglycerides May Lead to Further Reductions in Cardiovascular
Events in People with Type 2 Diabetes
The evidence presented above indicates clearly that lowering LDL-cholesterol levels is
beneficial for non-diabetic patients and people with diabetes. However, the event rates in the
treated diabetic subgroups are similar to the rates observed in the non-diabetic placebo groups.
That is, the risk among diabetic patients is not “normalized”. This raises the question about the
potential benefit of going beyond simple LDL-cholesterol-lowering. One option would be to
treat the lipid abnormalities characteristic of diabetes patients. In fact, there are data supporting
a role for lowering triglycerides and raising HDL-cholesterol levels in primary and secondary
prevention trials. In the Helsinki Heart Study, gemfibrozil lowered LDL-C modestly but also
lowered triglycerides and raised HDL-C, and the reduction in cardiac events in that primary
prevention trial was linked by multiple regression analysis to the rise in HDL-cholesterol. There
were too few diabetic patients in that study to observe a significant benefit in that group,
although a trend toward benefit was seen (Frick 1987). The recent VA-HIT trial (Rubins 1999)
indicated that in men with CHD and LDL-cholesterol levels of about 110 mg/dl (2.86 mmol/l),
treatment with gemfibrozil reduced new events by 22% over a five-year period. Gemfibrozil
treatment was associated with a 25% lowering of triglycerides, a 7% increase in HDL-
cholesterol, and no change in LDL-cholesterol. About 25% of the 2500 men in the trial had
diabetes, and this group appeared to have both a much higher event rate in the placebo group
(37% over five years) and a similar 22% reduction in events in the gemfibrozil treated group.
The results of VA-HIT are summarized in Table 1.9.

19
Table 1.9: Event Rates in VA-HIT

Population N Event Rate Event Definition
Placebo Fibrate
Overall 2531 275/1267=22%=4.3%/yr 219/1264=17%=3.3%/yr CHD Death, NFMI, stroke
Diabetic Participants 627 116/318=36%=7%/yr 88/309=28%=5.5%/yr CHD Death, NFMI, stroke
A key fact to note regarding VA-HIT was that the initial LDL-C level was in a range
considered to be near target. However, even with a near target LDL-C, the placebo group rates
for the diabetic patients were almost twice as high as the rates for the nondiabetic patients. More
importantly, VA-HIT has demonstrated that increasing HDL-C and lowering TG can provide
significant additional benefits for patients with a near target LDL-C. To extrapolate from the
VA-HIT study, it could be hypothesized that even after statin therapy (with lowering of LDL-
cholesterol from an average level of about 140 mg/dl (3.64 mmol/l) [the expected level of LDL-
C in a diabetic population] to a target LDL-C of about 115 mg/dl (2.99 mmol/l)), the addition of
a fibrate could further reduce event rates significantly. This is the basis for the ACCORD lipid
intervention hypothesis.
1.2.e Rationale for Trial of Fibrate + Statin vs. Placebo + Statin.
As noted above, the very high five-year event rates in VA-HIT participants with diabetes
in the placebo group (36%) and in the fibrate treated group (28%) indicate a need to answer the
question as to whether combined therapy with statin and fibrate would provide greater benefit
than therapy with statin alone.
Regarding trials in diabetic patients, there is only one completed fibrate-only trial and
only one fibrate-only trial underway. The Diabetes Atherosclerosis Intervention Study (DAIS)
was an angiographic trial in which 418 patients with diabetes (mean HbA1c=7.5%) and coronary
artery disease were randomized to fenofibrate or placebo and followed for a mean of about 40
months. The fenofibrate group had a statistically significant smaller increase in percent diameter
stenosis and a statistically significant smaller decrease in minimum lumen diameter. Although
not powered for clinical events, there were fewer cardiac events in the fenofibrate group (38
versus 50) (DAIS 2001). The Fenofibrate Intervention and Event Lowering in Diabetes (FIELD)
study is a clinical event trial of 8,000 diabetic patients largely without coronary artery disease
that will not be completed for several years.
The only other trial attempting to address the issue of therapy with statin plus a fibrate in
patients with diabetes was the Lipid Diabetes Study (LDS), which was using fixed doses of
cerivastatin and fenofibrate (vs placebos) in a 2 X 2 factorial design in a primary prevention
setting. However, this trial was recently terminated after cerivastatin was removed from the
market. Therefore, ACCORD is now the only trial addressing this important issue. Also, the
ACCORD protocol differs from the original LDS protocol in that ACCORD includes both
primary and secondary prevention groups and evaluates lipid treatment in the context of
protocol-specified glucose control.

20
Other strategies that could be used to obtain better outcomes in diabetic patients beyond
average or near target LDL-cholesterol levels are improved risk stratification and achievement of
lower LDL-cholesterol levels. Improved risk stratification implies the ability to better identify
those individuals at highest risk than is now feasible. Approaches to this include measurement of
serum markers (e.g. C reactive protein) or direct non-invasive quantification of vascular disease
(e.g. with coronary calcium screening). Data supporting this approach, while intriguing, are
tentative at the present time.
A strategy of achieving lower LDL-cholesterol than currently recommended is an

attractive one, but there are 2 large clinical trials ongoing that will address this question,
although not exclusively in diabetic patients. The Treating to New Targets (TNT) trial will
randomize 8600 patients with coronary heart disease to high or low-dose atorvastatin and follow
them for clinical events. The Study of the Effectiveness of Additional Reductions in Cholesterol
and Homocysteine (SEARCH) trial is a similar trial with high vs low dose simvastatin in 12,000
patients with coronary heart disease.
An additional rationale for studying the efficacy of combined therapy with a statin and
fibrate is that this is an increasingly used combination that must be proved safe as well as
effective. Data from a number of small clinical trials suggests that the incidence of myositis,
defined as muscle pain and plasma CPK level greater than 10 times upper limit of normal, is
about 1%. In those small, tightly controlled trials, there were no cases of rhabomyolysis. This
purported safety profile must be confirmed in a large trial and placed in the context of the
hypothesized additional benefit achieved by the combined treatments.
1.2.f Justification for Use of Simvastatin in Lipid Trial Participants
ACCORD is a trial of a high-risk diabetic population, one with prevalent vascular

disease, evidence of subclinical disease, or the presence of multiple CVD risk factors. The data
presented above, as well as national guidelines, support the use of statins for such patients. Thus,
in each of the trials noted above, the diabetic subset had much higher event rates than the overall
group. Even after treatment with a statin, the diabetic patients had event rates in the range of
3.1-4.2% per year. In AFCAPS/TEXCAPS, a primary prevention trial of relatively low risk
individuals, the diabetic subgroup on lovastatin had an event rate that was 50% greater than the
overall trial population (Downs 1998). In the Heart Protection Study (Heart Protection Study
[HPS] Collaborative Group 2002), patients with diabetes without prior CHD events who were on
simvastatin had a 2.8% yearly event rate.
Although the NCEP guidelines define diabetes as a CHD-equivalent (and has LDL-C >
130 mg/dl as the initiation for pharmacologic treatment and the goal at 100 mg/dl), and an update
of the guidelines states that an LDL-C goal of <70 mg/dL is a therapeutic option for patients with
existing cardiovascular disease (Grundy 2004), the guidelines also note that drug treatment is
optional for CHD patients with an LDL-C between 100 and 129 mg/dl, inclusive, and that
clinical judgment may call for deferring drug therapy in this subcategory because of limited data
identifying the exact levels for either initiation or goals. As noted above, placebo treatment in the
three major statin trials was associated with the following event rates and corresponding on-
treatment LDL-C: 4S 5.2%/year, 186 mg/dl; LIPID 2.6%/year, 150 mg/dl; CARE 2.6%/year,
135 mg/dl. Thus a 20% lower LDL-C in LIPID compared to 4S was associated with an event
21
rate that was half a great. However, a 10% lower LDL-C in CARE compared to LIPID was
associated with no fewer events. Similarly, pooled analysis of CARE and LIPID data showed no
benefit of treatment at the lowest quintile of baseline LDL-C (median of 117 mg/dl [3.04
mmol/l], Sacks 1999). On the other hand, the results from the Heart Protection Study (HPS)
demonstrate benefit of lowering LDL-C in a high risk group even when baseline LDL-C levels
are low. Specifically, five-year CHD event rates were reduced from 22.2% to 17.6% in subjects
who had a baseline LDL-C less than 116 mg/dl. Further, five-year CHD event rates were
reduced from 21.0% to 16.4% in subjects with baseline LDL-C levels less than 100 mg/dl. More
recently, the Treating to New Targets (TNT) trial of persons with existing CHD found that
lowering LDL-C levels to an average of 77 mg/dL using atorvastatin 80 mg/day, compared with
an average of 101 mg/dL using atorvastatin 10 mg/day, resulted in a significant 22% reduction in
major cardiovascular events over a median of 4.9 years of follow-up [LaRosa 2005]. Thus, based
on the HPS results and other evidence such as TNT, ACCORD will treat all primary prevention
participants in the lipid portion of the trial with 20 mg/day simvastatin and all secondary
prevention participants with 40 mg/day. In addition, the dose of simvastatin will be increased
from 20 mg/day to 40 mg/day in participants who begin the trial as primary prevention who then
have a cardiovascular event during the course of the trial or whose LDL-C is consistently greater
than 100 mg/dl (2.59 mmol/l). The estimated in-trial mean LDL-cholesterol level of participants
in the lipid component of ACCORD is estimated to be approximately 82 mg/dl (2.12 mmol/l)
(see Section 3.3.c).
(It is to be noted that under the Vanguard Protocol [dated September 13, 2001],
participants in the lipid trial were titrated from 0 to 20 mg of simvastatin for the purpose of
achieving an LDL-C of approximately 100 mg/dl [2.6 mmol/l]. Under this main trial protocol, all
lipid trial participants, including those randomized during the Vanguard who provide consent,
will be assigned 20 or 40 mg simvastatin, depending on their CVD status.)
Available data from the major secondary prevention trials published prior to HPS indicate
that when individuals have baseline LDL levels greater than 120 mg/dl, lowering LDL-C to
below that level is associated with benefit that is similar regardless of the exact LDL-C
concentration that is achieved. Thus, recurrent event rates related to the average on-treatment
LDL-C in the active treatment groups were: LIPID 2%/year, average 113 mg/dl; POSCH
2%/year, 111 mg/d; CARE 2%/year, 97 mg/dl; and for the 40% subset of participants in 4S
who lowered their LDL-C to < 100 mg/dl, 2%/year, 95 mg/dl. These studies had similar event
rates for on-treatment LDL-C that ranged from 95 mg/dl to 113 mg/dl. On the other hand, in
HPS, benefit of simvastatin treatment was observed even when baseline LDL-C levels were less
than 100 mg/dl. We believe that the ACCORD lipid trial protocol, in which we will treat all
primary prevention participants with 20 mg simvastatin and all secondary prevention participants
with 40 mg, regardless of baseline LDL-cholesterol levels, is consistent with all of the published
trial results. Because 40 mg of simvastatin may increase the risk for adverse events, particularly
in the patients receiving fenofibrate, participants will be followed closely and CPK regularly
measured. LDL-cholesterol levels will be monitored by the Coordinating Center and any
participant who is on 40 mg of simvastatin and whose LDL-cholesterol is consistently greater
than 120 mg/dl will be unmasked and treated appropriately.

22
1.3 Diabetes and Hypertension
1.3.a Diabetes and Cardiovascular Disease
Diabetes mellitus increases the risk of cardiovascular events two-to-three-fold at every

level of SBP or diastolic BP (DBP) and in diabetic patients there is a graded increase in risk
across the entire range of blood pressure levels (Stamler 1993). Therefore, diabetes and
hypertension combined confer a much higher risk than either one alone. In part because of this
higher risk, even at high normal levels of BP, JNC VI recommended beginning drug treatment in
diabetic patients if the SBP is >130 mm Hg or the DBP is >85 mm Hg, and BP goals are <130/85
mm Hg (JNC VI 1997). However, at the time these recommendations were made, there were no
completed clinical trials supporting the recommendations.
1.3.b Trials of Reducing Blood Pressure in Diabetic Patients
Table 1.10 describes the clinical trials of blood pressure lowering in diabetic patients. In
the 583 participants with type 2 diabetes mellitus in SHEP, major cardiovascular disease events
were reduced by 34% (Curb 1996). Although this was the same risk reduction as in nondiabetic
participants, the absolute risk reduction was twice as great for diabetic participants. The SHEP
BP entry criterion was a SBP 160-219 mm Hg; the treatment goal was <160 mm Hg and at least
20 mm Hg reduction from baseline. Systolic BP was reduced from 170 to 143 mm Hg.
Subsequent to JNC VI, the Hypertension Optimal Treatment (HOT) study reported that in
the diabetic subgroup (n=1,501) major cardiovascular events were reduced by 51% (P=0.005) in
those randomized to a DBP goal of <80 mm Hg compared to a goal of <90 mm Hg: 12 versus 24
events/1000 patient-years (Hansson 1998). However, this was a post hoc analysis and the
number of events was relatively small. The achieved BP for the more intensive group in the
diabetic patients has not been reported, but for all hypertensive patients it was 140/81 mm Hg.
There were no differences in cardiovascular events between randomized groups in the entire
18,790 hypertensive patients in HOT. In the United Kingdom Prospective Diabetes Study
(UKPDS 1998), 1,148 hypertensive type 2 diabetic patients were randomized to either tight BP
control (<150/85 mm Hg) or less tight BP control (<180/105 mm Hg). In that trial, diabetes
related endpoints were reduced by 24% (P=0.005), deaths related to diabetes by 32% (p=.019),
strokes by 44% (p=.013), and microvascular endpoints by 37% (p=.009) after a median follow-
up of 8.4 years (UKPDS 1998a). Although not statistically significant, all-cause mortality was
reduced by 18% and MI by 21%. Average BP over 9 years was 144/82 mm Hg and 154/87 mm
Hg in the tight and less tight BP control groups, respectively, for a BP difference of 10/5 mm Hg.
In a placebo-controlled trial of treatment of isolated systolic hypertension, the Systolic
Hypertension in Europe (Syst-Eur) Trial, the 492 patients with diabetes were reported in a post
hoc analysis to have significant reductions in CVD mortality, all CVD events, and stroke with
the mean SBP reduced from 175 to 153 mm Hg (Tuomilehto 1999). Entry criteria were similar
to SHEP (SBP 160-219 mm Hg), and the goal was to reduce SBP at least 20 mm Hg to <150 mm
Hg. The Appropriate Blood Pressure Control in Diabetes (ABCD) Trial, a prospective,
randomized, masked trial in 470 hypertensive diabetic patients (type 2), compared the effects of
moderate control of BP (target DBP 80-89 mm Hg) with those of intensive control of BP (DBP

23
75 mm Hg) on the incidence and progression of diabetic nephropathy, retinopathy,

cardiovascular disease and neuropathy (Schrier 1996, Estacio 1998). The results of the
microvascular outcomes of the ABCD trial BP comparison have recently been reported (Estacio
2000). The mean blood pressure achieved in the intensive group was 132/78 mm Hg and was
138/86 mm Hg in the moderate control group. There were no differences in any microvascular
endpoints for the 2 BP goals (and no microvascular differences between nisoldipine vs enalapril
in the more intense group). The BP delta was 6/8 mm Hg, although the goals were just for DBP.
The intensive therapy group had a lower mortality rate, 5.5% vs 10.7% (p=0.037), but there were
no statistically significant differences in MI, cerebrovascular events, or CHF to account for the
mortality difference (Estacio 2000).
Table 1.10: Clinical Trials of Blood Pressure Lowering in Diabetic Patients
Mean BP, Mean BP,

N Risk
Trial Duration less more Initial Therapy Outcome
Reduction
intense intense
SHEP Stroke 22% (ns)
(Curb 583 5 years 155/72* 143/68* Chlorthalidone CVD events 34%
1996) CHD 56%
Syst-Eur Stroke 69%
(Tuomilehto 492 2 years 162/82 153/78 Nitrendipine
1999)
CV events 62%
CV events 51%
HOT
MI 50%
(Hansson 1,501 3 years 144/85* 140/81* Felodipine
Stroke 30% (ns)
1998)
CV mortality 67%
Diabetes-related
UKPDS endpoints: 34%
Captopril or
(UKPDS 1,148 8.4 years 154/87 144/82 deaths: 32%
atenolol
1999a) Strokes 44%
Microvascular 37%
CCr nc
Albuminuria nc
ABCD
Nisoldipine or Retinopathy nc
(Estacio 470 5.3 years 138/86 132/78
enalapril Neuropathy nc
2000)
Mortality 49%
MI, CVA, CHF ns
BP = blood pressure, ns = not significant, nc = no change
* BP in diabetic + non-diabetic population, since BP not reported for diabetic patients alone
Therefore, the HOT and UKPDS studies provide the most definitive clinical trial evidence
to date and support BP goals in diabetic hypertensive patients of <150/85 mm Hg (UKPDS) and
DBP <80 mm Hg (HOT). Based on these goals, as well as achieved BP levels in other trials,
including SHEP, all of the trials are consistent with SBP goals of 140 mm Hg in diabetic patients
and none, including ABCD, have confirmed benefit to lower goals than this.
The ALLHAT study was begun in 1994 and includes more than 15,000 diabetic patients
(Davis 1996). It is primarily designed to compare 4 different classes of antihypertensive drugs.
The BP goal of therapy is at least <140/90 mm Hg. Neither this trial, nor others in progress, will
provide data on the added effect on CVD morbidity and mortality of BP-lowering on top of
glycemic control in diabetic patients. ACCORD will address this issue and should also provide
24
the first clinical trial data on the possible benefit of treating to more aggressive BP goals
(compared with the UKPDS, for example) in preventing CVD in diabetic patients.
1.3.c Trials Regarding Choice of Antihypertensive Drug
Ongoing trials, such as ALLHAT, will clarify whether there are important differences in
CVD outcomes among various classes of antihypertensive agents in patients with type 2 diabetes
mellitus and hypertension (Davis 1996, Cutler 1998). Results from the 15,000+ diabetic
hypertensive participants within ALLHAT (randomized to receive chlorthalidone, amlodipine,
lisinopril, or doxazosin in a double-masked design) should give more definitive direction for
antihypertensive drug therapy for ACCORD, although the projected end of follow-up for
ALLHAT is not until 2002. In early 2000, however, the doxazosin arm of ALLHAT was
stopped because of a significantly higher incidence of cardiovascular events in the doxazosin
group versus the chlorthalidone group (ALLHAT 2000). In the diabetic subgroup of ALLHAT,
the rates of CVD and CHF were significantly higher in participants randomized to doxazosin
(relative risk = 1.24 [P<0.0001] and = 2.14 [P<0.0001], respectively). Otherwise, existing data
do not clearly mandate one antihypertensive drug class for this population.
Major CVD events were reduced in the diabetic subgroups in SHEP (Curb 1996) and
HDFP with therapy initiated with a diuretic. In the 758 patients in the tight control group of
UKPDS, the ACE inhibitor captopril and the beta-blocker atenolol were equally effective in
reducing the incidence of diabetic macrovascular and microvascular complications (UKPDS
1998b). In the Captopril Prevention Project (CAPPP), there were no significant differences in
CVD mortality or MI for captopril versus conventional treatment with diuretics and/or beta-
blockers in the nearly 11,000 hypertensive patients (although strokes were 25% more frequent
with captopril). However, in a post hoc subgroup analysis in the 572 patients with diabetes, the
risk reduction for the primary CVD endpoint was 41% (P=0.019) with captopril vs conventional
treatment (Hansson 1999). In the second Swedish Trial in Old Patients with Hypertension
(STOP-2), there was no difference for the primary outcome (cardiovascular mortality) between
patients randomized to diuretics and/or beta-blockers versus ACE inhibitors versus calcium
antagonists, both overall and in the 719 patients with diabetes (Hansson 1999). In a post hoc
analysis of the diabetic subgroup (n=492) of the Syst-Eur Trial, an antihypertensive regimen
initiated with the dihydropyridine calcium channel blocker nitrendipine reduced CVD mortality
and events compared to placebo (Tuomilehto 1999).
Several relatively small controlled trials in diabetic hypertensive patients have reported
lower cardiovascular event rates with an ACE inhibitor compared with a calcium channel
blocker. The 470 diabetic hypertensive participants in the ABCD trial had a 7-fold higher
incidence of fatal and nonfatal MIs with the dihydropyridine calcium channel blocker nisoldipine
than with the ACE inhibitor enalapril through five years of follow-up (Estacio 1998), although
microvascular outcomes were not different between the two drugs (Estacio 2000). In the
Fosinopril Amlodipine Cardiovascular Events Trial (FACET), 380 diabetic hypertensive patients
experienced a 51% lower incidence of the combination of acute MI, hospitalized angina, and
stroke with fosinopril compared with amlodipine (P=0.03) over 2.8 years of follow-up (Tatti
1998).

25
Therefore, diuretics, ACE inhibitors, beta-blockers, and calcium channel blockers have
been associated with reduced major macrovascular or microvascular events in diabetic
hypertensive patients compared with placebo or a less intensively treated control group in
randomized controlled trials. Comparisons between drugs are less clear, except for the higher
risk with an alpha blocker seen in ALLHAT (ALLHAT 2000). It would also appear reasonable
to avoid treating hypertension in diabetic patients with single-drug therapy with a calcium
channel blocker until more data are available.
1.4 Conclusions of Recent Expert Panels Convened to Discuss Diabetes and

Cardiovascular Disease
The report of the Macrovascular Disease Subcommittee of the NIH-sponsored Diabetes

Conference in September 1997 “enthusiastically recommended a large scale clinical trial to
determine whether the level of glucose control … will decrease the incidence of CHD in a
diabetic population." The report also noted that "about 50 percent of excess heart disease in
diabetic patients can be attributed to associated abnormalities in other known CVD risk factors"
and that "the same risk factors that predict large vessel disease (i.e. stroke, heart attack and
peripheral arterial disease) in the general population also affect the diabetic." A trial comparing
the cost-effectiveness of different therapeutic approaches (such as contrasting optimal glucose
control with aggressive lipid lowering) was advocated by some members of the panel. The panel
also emphasized the importance of selecting diabetic patients at particularly high risk for
developing CVD for inclusion in any trial.
Similar conclusions were reached by the NHLBI Special Emphasis Panel on Prevention
and Treatment of Cardiovascular Disease in Diabetes Mellitus. Notably, a number of panel
members recommended testing not only the relative benefit of different diabetic regimens, but
also different target levels or intensities of treatment for lipids or blood pressure, using a factorial
design. The rationale for a factorial design is that although a number of studies in progress are
collectively addressing treatment of lipids, blood pressure or glycemic control in diabetic
patients, none of them would shed light on the comparative benefit of treating hyperglycemia
and aggressively treating blood pressure and lipids.
Additional support for a large clinical trial testing the benefit of tight glycemic, lipid, and
blood pressure control was given by an ad hoc advisory group convened by NHLBI in May
1998.
1.5 Specific ACCORD Hypotheses
ACCORD is designed as a double 2x2 factorial design with factors consisting of:
intensive versus standard glycemic control, intensive versus standard blood pressure control, and
in the presence of desirable LDL-C levels, fibrate use versus placebo. As shown in Figure 1.2
below, all 10,000 participants will be randomized to the glycemic interventions; 5,800
participants meeting the lipid entry criteria will be randomized to the lipid interventions in one
2x2 trial; 4,200 participants who meet the blood pressure entry criteria will be randomized to the
blood pressure interventions in the second 2x2 trial.

26
Figure 1.2:
Projected Allocation of Participants in ACCORD
Lipid Trial SBP Trial

Fibrate Placebo Intensive Standard
Intensive 1450 1450 1050 1050 5000

10,000
Participants
in Glycemia
Trial
Standard 1450 1450 1050 1050 5000
2900 2900 2100 2100
Participants not recruited for the lipid trial will be referred to their usual source of care
for treatment of any lipid abnormalities. Similarly, participants not recruited for the blood
pressure trial will be referred for treatment of any blood pressure abnormality.
Recommendations for goals of these treatments will be provided (see Section 3.5). High risk
participants with and without a history or evidence of vascular disease will be recruited at
approximately 60 Clinical Sites administratively located within 7 Clinical Center Networks in
the United States and Canada. Recruitment will occur over two non-contiguous periods
(described below and in Section 7.1) and participants followed for about 4 to 8 years
(approximate mean of 5.6 years).
The three specific primary ACCORD hypotheses are:
In middle-aged or older people with type 2 diabetes who are at high risk for having a
cardiovascular disease (CVD) event:
(2) in the context of good glycemic control, does a therapeutic strategy that uses a fibrate
to raise HDL-C/lower triglyceride levels and uses a statin for treatment of LDL-C
reduce the rate of CVD events compared to a strategy that only uses a statin for
treatment of LDL-C?
systolic blood pressure (SBP) of < 120 mm Hg reduce the rate of CVD events more
than a strategy that targets a SBP of < 140 mm Hg?

27

The intervention-specific secondary hypotheses are specified in Section 7.1.b. The intervention-
specific subgroup hypotheses are specified in Section 7.1.c.
1.6 Timetable/The Vanguard Phase
ACCORD will be conducted over a 101/2 year period, from October 1, 1999 to
March 31, 2010. There are eight operational phases for the trial:
Phase # of Months in Phase Calendar Time Trial Activities

I 10 10/1/99 to 7/30/00 Initial Protocol Development
II 2 8/1/00 to 9/30/00 Procedure Finalization /Training
III 3 10/1/00 to 12/31/00 Vanguard Startup and Screening
IV 24 1/1/01 to 12/31/02 Vanguard Recruitment/Follow-up/
Review/Protocol Revision
V 33 1/1/03 to 9/30/05 Main Recruitment and Follow-up
VI 41 10/1/05 to 2/28/09 Follow-up Only
VII 4 3/1/09 to 6/30/09 Participant Close-out
VIII 9 7/1/09 to 3/31/10 Analysis and Reporting
As noted above, recruitment will occur in two non-contiguous periods: an initial period
that began in January 2001 in the Vanguard Phase (Phase IV) of the trial (during which
approximately 1200 participants were recruited), and then a subsequent period beginning in
January 2003 (after review of the vanguard data) and ending in September 2005 (during which
the remainder of the 10,000 participants will be recruited).
During Phase IV, the ACCORD investigators, the Data and Safety Monitoring Board,
and the National Heart, Lung, and Blood Institute monitored the feasibility of the Vanguard
protocol. The specific goals of the Vanguard, which were used to judge its success, are described
in Section 7.5. After extensive review of the data, the ACCORD Protocol was revised to increase
the likelihood of achieving all of the trial objectives.

28
Chapter 2
Participant Selection and Follow-up
2.1 Eligibility Criteria
The objective of setting inclusion/exclusion criteria is to identify a trial population that

will ensure adequate event rates for statistical power, provide maximum generalizability, and
maximize safety. Inclusion/exclusion criteria were made as simple as possible.
In addition to fulfilling the overarching glycemia trial entry criteria, to be eligible

for ACCORD a screenee also needs to fulfill the entry criteria for either the lipid and/or
blood pressure components of the trial. To reduce the possibility of bias by having clinic
staff decide whether a screenee should be in the lipid or blood pressure component,
eligibility for both components needs to be assessed.
2.1.a Inclusion Criteria
1. Type 2 diabetes mellitus defined according to the 1997 ADA criteria:

• Fasting plasma glucose >126 mg/dl (>7.0 mmol/l), or
• Symptoms of hyperglycemia with casual plasma glucose > 200 mg/dl (>11.1
mmol/l), or
• 2 hour plasma glucose > 200 mg/dl (>11.1 mmol/l) after a 75 gram oral
glucose load
2. HbA1c (obtained within 3 months prior to anticipated date of randomization):

• 7.5 to 11%
a) if on insulin, < 1 u/kg plus on 0 or 1 oral agent, or
b) if not on insulin, on 0, 1, or 2 oral agents
• 7.5 to 9%
a) if on insulin < 1 u/kg plus on 2 oral agents, or
b) if not on insulin plus on 3 oral agents, or
c) if on insulin > 1 u/kg plus 0 oral agents
Oral agents include: a) insulin secretagogues (sulfonylurea, meglitinides),

b) biguanides, c) insulin enhancers (thiazolidinediones)
The upper limits for HbA1c were selected to increase the likelihood of reaching
the study’s HbA1c targets. The lower limit was selected to allow for further
reduction should the participant be assigned to the intensive glycemic group.
3. Known diabetes duration > 3 months
4. Stable diabetes therapy for > 3 months (dose of any 1 antihyperglycemic drug has
not changed by more than two-fold and new agents have not been added within
the previous 3 months)

29
5. Age at Randomization:
• 40 to 79 years (inclusive) for anyone with a history of clinical cardiovascular
disease (defined below in Item #6A), or
• 55 to 79 years (inclusive) for anyone without a history of clinical
cardiovascular disease (defined below in Item #6A)
6. At high risk of CVD events, defined as:
A. Presence of clinical cardiovascular disease.

• previous myocardial infarction (MI)
• previous stroke
• History of coronary revascularization (e.g., coronary artery bypass graft
surgery, stent placement, percutaneous transluminal coronary angioplasty,
or laser atherectomy)
• History of carotid or peripheral revascularization (e.g., carotid
endarterectomy, lower extremity atherosclerotic disease atherectomy,
repair of abdominal aorta aneurysm, femoral or popliteal bypass)
• angina with ischemic changes (resting ECG), ECG changes on a graded
exercise test (GXT), or positive cardiac imaging study
or
B. If no clinical cardiovascular disease, evidence in the last 2 years suggesting a
high likelihood of cardiovascular disease. Specifically, the presence of one
of the following:
• Microalbuminuria
• Ankle brachial index < 0.9 (by simple palpation)
• LVH by ECG or ECHO
• > 50% stenosis of a coronary, carotid, or lower extremity artery
or
C. The presence of at least 2 of the following factors that increase CVD risk:
• On lipid lowering medication or untreated LDL-C >130 mg/dl (3.38
mmol/l)
• Low HDL-C (< 40 mg/dl (1.04 mmol/l) for men and < 50 mg/dl (1.29
mmol/l) for women)
• On BP lowering medication or untreated SBP >140 mm Hg or DBP > 95
mm Hg.
• Current cigarette smoking
• Body mass index > 32 kg/m2
Note: Category A represents secondary prevention participants. Categories B

and C together represent primary prevention participants.

30
2.1.b Exclusion Criteria
Exclusion criteria were selected to enhance safety and adherence.
1. History of hypoglycemic coma/seizure within last 12 months

2. Hypoglycemia requiring 3rd party assistance in last 3 months with concomitant
glucose < 60 mg/dl (3.3 mmol/l)
3. History consistent with type 1 diabetes
4. Unwilling to do frequent capillary blood glucose self-monitoring or unwilling to
inject insulin several times a day
5. BMI > 45 kg/m2
6. Serum Creatinine > 1.5 mg/dl (132.6 umol/l) obtained within the previous 2 months
7. Transaminase >2 times upper limit of normal or active liver disease
8. Any ongoing medical therapy with known adverse interactions with the glycemic
interventions (e.g., corticosteroids, protease inhibitors)
9. Cardiovascular event or procedure (as defined for study entry) or hospitalization for
unstable angina within last 3 months
10. Current symptomatic heart failure, history of NYHA Class III or IV congestive heart
failure at any time, or ejection fraction (by any method) < 25%
11. A medical condition likely to limit survival to less than 3 years or a malignancy other
than non-melanoma skin cancer within the last 2 years
12. Any factors likely to limit adherence to interventions. For example,
• dementia
• alcohol or substance abuse
• plans to move in the next 2 years.
• history of unreliability in medication taking or appointment keeping
• significant concerns about participation in the study from spouse, significant
other, or family members
• lack of support from primary health care provider
13. Failure to obtain informed consent from participant
14. Currently participating in another clinical trial. Note: Patient must wait until the
completion of his/her activities or the completion of the other trial before being
screened for ACCORD
15. Living in the same household as an already randomized ACCORD participant.
16. Any organ transplant
17. Weight loss > 10% in last 6 months
18. Pregnancy, currently trying to become pregnant, or of child-bearing potential and not
practicing birth control

31
19. Participants with recurrent requirements for phlebotomy or transfusion of red blood
cells.
2.1.c Additional Eligibility Criteria for Participants in the Lipid Component of

ACCORD
Participants eligible for the glycemic component of the trial will also be eligible
for the lipid component if the following criteria are met. Screening lipids may either be
measured at a local laboratory or obtained from medical records. If obtained from
medical records, use the most recent values recorded within the previous 12 months. If
there are no lipid values recorded in the medical records within the previous 12 months, a
blood test must be performed by the local laboratory.
• 60 mg/dl < LDL-C < 180 mg/dl (1.55 to 4.65 mmol/l) if not on a lipid-
lowering agent during screening, or, if on a lipid-lowering agent, the LDL-C
needs to be between the drug/dose-specific cut points inclusive found in Table
2.1.
and
• HDL-C less than 55 mg/dl (1.42 mmol/l) for women or Blacks/African-
Americans, or HDL-C less than 50 mg/dl (1.29 mmol/l) for all other gender-
race groups
and
• Triglycerides <750 mg/dl (8.47 mmol/l) on no therapy or < 400 mg/dl (4.52
mmol/l) on treatment with lipid lowering drugs
The rationale for the lower LDL-C limit is to exclude people with already low
LDL-C levels because they would be exposed to a statin, which would likely reduce their
LDL-C levels to very low, possibly harmful levels. The rationale for the upper LDL-C
limit is that patients with higher LDL-C often would require a higher dose of a statin than
ACCORD would provide, which would place them at higher risk for adverse events if
randomized to a fibrate. The rationale for the HDL-C limit is that increasing HDL-C may
have little effect among participants in whom HDL-C is already high. The triglyceride
limits were selected for participant safety.
The additional exclusion criteria for the lipid intervention are:

• known hypersensitivity to statins or fibrates
• requirements for use of erythromycin, clarithromycin, cyclosporine,
systemic azole antifungals, or nefazodone or trazodone
• refusal to stop current lipid-lowering drugs .
• history of pancreatitis
• untreated or inadequately treated thyroid disease
• women who are breast feeding
• documented previous occurrence of myositis/myopathy
• pre-existing gallbladder disease (eg., history of gallstones)

32
The recruitment goal for the lipid 2 X 2 trial is 5,800 participants.
Table 2.1: LDL-C Eligibility Ranges for Screenees on a Lipid-Lowering

Agent (By Agent and Dose) (08/03/04 Revision)
In mg/dl, In mmol/L ,
Estimated % the LDL-C Must the LDL-C Must
Lipid Lowering Agent Dose LDL-C Reduction Be Between (inclusive): Be Between (inclusive):
Atorvastatin (Lipitor) 2.5 mg 25 45 - 135 1.16 - 3.49

Atorvastatin (Lipitor) 5 mg 29 43 - 128 1.10 - 3.30
Simvastatin (Zocor) 5 mg 26 44 - 133 1.15 - 3.44

Lovastatin (Mevacor) 10 mg 18 49 - 148 1.27 - 3.82

Pravastatin (Pravachol) 10 mg 22 47 - 140 1.21 - 3.63

Fluvastatin (Lescol) 20 mg 22 47 - 140 1.21 - 3.63

Fluvastatin (Lescol) 40 mg 24 46 - 137 1.18 - 3.54
Rosuvastatin (Crestor) 5 mg 40 36 - 108 0.93 - 2.79

Ezetimbe (Zetia) 10 mg 17 50 - 149 1.29 - 3.86
Fenofibrate any 5 57 - 171 1.47 - 4.42

Niacin any 10 54 - 162 1.40 - 4.19
Resin any 10 54 - 162 1.40 - 4.19
All Others any 0 60 - 180 1.55 - 4.65
2.1.d Additional Eligibility Criteria for Participants in the Blood Pressure

Component of ACCORD
Participants eligible for the glycemic component of the trial will also be eligible
for the blood pressure component:
• If the systolic blood pressure is between 130 and 160 mm Hg, inclusive, and
the patient is on 0, 1, 2, or 3 antihypertensive medications, or
• If the systolic blood pressure is between 161 to 170 mm Hg, inclusive, and
the patient is on 0, 1, or 2 antihypertensive medications, or

33
• If the systolic blood pressure is between 171 to 180 mm Hg, inclusive, and the
patient is on 0 or 1 antihypertensive medication.
and
• If:
dipstick protein in a spot urine is < 2+, or
the protein-to-creatinine ratio in a spot urine is <700 mg/gm creatinine, or
24-hour protein excretion is <1.0 gm/24 hours
For screenees who are not currently on blood pressure (BP)-lowering medication,
there must be documentation of SBP > 130 mm Hg on at least 2 occasions.
The recruitment goal for the blood pressure 2 X 2 trial is 4,200 participants.
2.2 Recruitment: Informed Consent, Screening, Baseline
The ACCORD recruitment goal is a minimum sample size of 10,000 participants:

50% females, 33% racial and ethnic minorities, and 50% primary prevention (no history of
clinical CVD as defined in 2.1.a.6). Specific community resources will be used to target high
risk and minority/under-served populations to ensure adequate representation of these groups in
ACCORD. Recruitment strategies that worked well in other trials related to diabetes will be
used. Centralized training for CCN and Clinical Site staffs regarding recruitment issues will be
provided before recruitment begins. Several recruitment strategies were used successfully
during the Vanguard Phase, including chart review and review of patients within investigator
practices. During the main trial, additional strategies will be employed, including advertising.
2.2.a Informed Consent
To participate in ACCORD, participants must provide written, informed consent using

procedures reviewed and approved by each Clinical Site’s local Institutional Review Board.
Even though consent to participate in ACCORD must be obtained for all stages of the study,
the process and timing of consent may vary by clinic. Descriptions of each Clinical Site’s
consent procedures are included as part of the Manual of Procedures, and copies of each
Clinical Site’s consent documents are kept at the Coordinating Center. The consent forms must
include all procedures done as part of screening, a possible run-in, and follow-up. The
elements of consent are presented in Section 4.5 and a model informed consent document is in
Appendix I.
Of special concern regarding informed consent is the collection of blood samples for
genetic analysis. The consent forms will clearly indicate that a sample may be drawn for this
purpose, but that the participant has the right to refuse this procedure. The portion of the
informed consent document describing the genetics component of ACCORD uses the multi-
level approach recommended by the NHLBI Panel on “Opportunity and Obstacles to Genetics
Research in NHLBI Clinical Studies.” Also, the confidentiality of the data will be maintained.

34
2.2.b Screening and Possible Run-In (Self-Monitoring of Blood Glucose)
Potential participants can be recruited for ACCORD through either of two sequences of
screening or pre-randomization visits. One sequence would be used for those patients who are
currently in the practices of the Clinical Sites within the ACCORD network. A second
sequence would be used for those patients who come from outside the ACCORD Clinical Sites
and are, therefore, less well known to the ACCORD clinical center staff.
Prior to randomization, potential participants will be asked to provide evidence

that they can routinely monitor their capillary blood sugars. This evidence may be from a
diary, self-monitoring blood glucose (SMBG) device that they bring to the clinic, or, if
such retrospective data cannot be presented, then the screenee must prospectively go
through a 2 to 4 week pre-randomization run-in period. If the data are obtained from a
diary or from a SMBG device, then at least 2 weeks of data must be available and the
screening visit cannot occur on the same day as the randomization visit.
2.2.b.1 Existing Populations in the Clinical Site Practices-Medical record searches or

reviews of existing databases can be done initially by setting up the searches using the
characteristics that match the inclusion/exclusion criteria. Additional “hand searches” may be
necessary using the remaining inclusion/exclusion criteria not already part of the existing
database but part of the patient’s existing clinical record. It is likely that all or most all of the
inclusion/exclusion criteria will be available in most medical records.
2.2.b.2 Individuals Recruited Outside Existing Clinical Site Practices-Individuals

identified by any media strategy or who are otherwise identified outside of the practice of the
ACCORD clinical center will have to be appropriately screened. While general screening of
the population for abnormal fasting glucose levels is not permitted, referrals from health fairs or
community screenings conducted by others may be a useful source of participants.
2.2.c Screening Visits/Baseline Visit
The following are key elements of the screening and baseline visits.
A. Center notified of individual’s interest in study (for individuals outside practice)

1. Response to media
2. Phone number of individual available
B. Phone Contact
1. Age determined (if unknown)
2. Administer phone screen to determine initial potential eligibility
C. Screening Visit 1
1. Screening consent, if required by the sites’s IRB
2. Obtain HIPAA authorization
3. Patient to sign a release of information to collect documentation
4. Begin collection of baseline information, including additional eligibility
information.

35
D. Screening Visit 2
1. Perform required labs
2. Continue collection of baseline/eligibility information
3. Possible run-in period begins
E. Baseline visit (Randomization Visit)
1. Criteria for self-monitoring of blood glucose is satisfied
2. Confirmation that all inclusion/exclusion criteria satisfied
3. Perform physical examination
4. Randomization Consent
5. Patient randomized
6. Trial intervention begins
2.3 Schedule of Follow-up Visits
As described in Tables 2.2A through 2.2F, post-randomization follow-up visit

schedules differ by treatment group assignment. For participants in the intensive
glycemia group (regardless of BP/Lipid trial assignment) and for participants in the
standard glycemia group + intensive blood pressure group, post-randomization visits will
occur at Months 1, 2, 3, 4, and every 2 months thereafter. For participants in the standard
glycemia group + either the standard BP or lipid trial, post-randomization visits will
occur at Months 1, 4, and every 4 months thereafter. Additional visits can be scheduled
as needed to monitor and assure appropriate implementation of the study interventions.
For the purpose of event ascertainment, all participants in all treatment groups will be
queried regarding the occurrence of a possible event on the same schedule, specifically
every 4 months.
2.4 Procedures by Visit
Clinical center staff will be treating and following six different types of
participants in ACCORD. These are participants who are randomized to the:
• Intensive Glycemia and Intensive Blood Pressure Groups
• Intensive Glycemia and Standard Blood Pressure Groups
• Intensive Glycemia Group and in the Lipid Trial
• Standard Glycemia and Intensive Blood Pressure Groups
• Standard Glycemia and Standard Blood Pressure Groups
• Standard Glycemia Group and in the Lipid Trial.
Note that for the lipid trial, participants in the masked fibrate and placebo groups
will be treated identically by clinic personnel.
Scheduled examination components are shown by treatment group assignment

and by visit in Tables 2.2A through 2.2F. Because during follow-up the components of
the visits differ according to the portion of the trial the participant is in, these tables are
specific to the glycemia/blood pressure/lipid trial treatment group assignment.

36
Assessments performed at the various visits include questionnaires, physical

examinations, other clinical studies, laboratory tests, and performance of study-related
procedures as described below. Baseline characteristics to define the patient population
include sociodemographics, anthropometrics, blood pressure, pulse, current and past medical
history, basic physical examination, concomitant medications, laboratory, and quality of life
measurements.
2.4.a Questionnaires
2.4.a.1 Sociodemographics
Information is collected during screening and baseline regarding age, ethnicity,

gender, level of education, persons living with participants and United States ZIP
code/Canadian postal code. These data will be used to identify eligible participants and to
characterize the final study population. Social Security Number/Medicare
Number/Canadian Social Insurance Number/Provincial Health Insurance Number will be
collected for tracking purposes.
2.4.a.2 Medical History
Medical history data are collected at baseline in the form of a detailed initial
medical history and collected at specified follow-up visits in the form of an abbreviated
interval history. Important aspects of the medical history include eligibility criteria,
allergies, cardiovascular disease, smoking status, and diabetes. The presence of CVD
prior to entry into the study serves as an eligibility and stratification factor. Data
regarding the duration of diabetes and the presence of complications of diabetes are
important for descriptive purposes, subgroup analyses, and prognostic analyses.
2.4.a.3 Concomitant Medications
Information regarding the participants’ concomitant medication therapy is

collected and documented at baseline and then reviewed and revised at annual follow-up
visits. Appropriate sources for obtaining this information include participant (significant
other) report, current pharmacy action profiles, and verification of medications
documented in the medical record. Although data are collected on all standing therapies,
emphasis is placed on concurrent antihypertensive, glycemic and lipid-lowering therapy
as well as background risk reduction (eg., aspirin) therapy.
2.4.a.4 Diet, Physical Activity, Health-Related Quality of Life Substudy, Cost

Effectiveness Substudy, Eye Substudy, Memory in Diabetes (MIND)
Substudy
Diet and physical activity data are collected from a random sample of 2000
participants at Baseline, Month 12, Month 36, and Month 48. This random sample will
also participate in the Health-Related Quality of Life Substudy (see Section 6.2), which is
itself a random sample nested within the 4288 participants participating in the Cost

37
Effectiveness Substudy (see Section 6.3). As with the diet and physical activity data,
HRQL data will be collected at Baseline, Month 12, Month 36, and Month 48. Cost data
will be collected at baseline and every four months for the duration of the trial.
For the ACCORD Eye Substudy, conducted in a subset of 4065 participants, a full
ophthalmologic examination and fundus photography will be performed at Baseline and
at Month 48.
For the Memory in Diabetes (MIND) Substudy, conducted in a subset of 2,800

participants, a battery of cognitive neuropsychological tests will be conducted at Month
1, Month 20, and Month 40. (The Month 1 Visit will serve as the Baseline Visit.) In
addition to the neuropsychological tests, a subsample of 640 MIND participants will have
a Baseline and Month 40 MRI examination.
2.4.b Physical Examination Measures
2.4.b.1 Anthropometric Measurements
Body fat is a significant predictor for the onset of diabetes, as well as for
subclinical and clinically manifested cardiovascular disease. Excessive body and
abdominal obesity also hinders diabetes control and increases the likelihood of the
development of cardiovascular disease in this patient population. Successful
management of type 2 diabetes includes exercise and dietary modification with the goal
of reducing total body fat, particularly abdominal fat. It is the intent of this study to
gather data that will elucidate the impact of body fat and body composition on the course
of cardiovascular disease among patients with diabetes without extreme burden to study
participants and clinical investigators.
Anthropometric measures gathered for ACCORD include (1) standing height, (2)
weight, and (3) waist circumference. Body mass index (BMI, calculated as kg/m2) is
commonly used in clinical trials and population-based epidemiologic studies as an
estimate of overweight/obesity. Guidelines are currently available for the assessment of
overweight and obesity based on BMI values. BMI correlates well with adipose tissue
composition measured by more burdensome procedures such as cardiothoracic scan,
underwater weighing, and bioelectrical impedance. Similarly, abdominal obesity, as
assessed by a measurement of waist circumference, is an easily measured indicator that
has been shown to be predictive of both diabetes and cardiovascular disease risk.
2.4.b.2 Blood Pressure and Pulse
Using an automated device (the Omron 907), blood pressure (BP) and pulse are
measured three times at each clinic visit. The seated BP and pulse readings for
ACCORD are the averages of the first, second and third systolic and diastolic BP's and
pulses.

38
2.4.b.3 Other Physical Examination Components
The physical examination includes the items noted above (anthropometric

measurements, ascertainment blood pressure and heart rate) and a system-oriented
approach for the remainder of the examination. Participants will undergo both full
physical examinations and abbreviated aspects of the examination during the course of
their participation in the trial. Elements of the examination to be completed will vary
depending upon the time and type of visit (initial, interval, annual, final) and will comply
with recommended standards of diabetes care.
The systems physical examination includes: general survey, skin, head, ears, eyes,
nose, throat (including funduscopic) neck, chest, heart, abdomen, musculoskeletal/
extremities, pulse assessment, and neurological (including lower extremity).

39
Table 2.2A: Scheduled Examination Components by Visit: For Participants Randomized to the Intensive Glycemia + Intensive Blood Pressure Groups
Schedule in Months
τ
Evaluations Scrn BL 0.5 1 2 3 4 5 6 7 8 9 10 11 12 14 16 18 20 22 24 Q2 Q4 Q12 Q24 36 40 48 prn1 Exit
Clinic Visit X X X X X X prn X prn X prn X prn X X X X X X X X X X
BP/Pulse X X X X X X X X X X X X X X X X X X
† † † † † † † † †
Weight X X X X X X X X X X X X X X X X X X X
γ
BP Mileposts Μ Μ Μ Μ Μ Μ Μ
∆
HbA1c (POC) X X X X X prn X prn X prn X prn X X X X X X X X X
HbA1c C C C C C C C C C
FPG C C C C C C C
Potassium C C C C C C C C
Creatinine L C C C C C C C C
Lipid Profile L C C C C C
ALT L C C C C C C C C
CPK C C
Urinalysis L C C C C C
ECG L C C C C
Events X X X X X X X X X
Diet,Phys Actv* X X X X
HRQL* X X X X
Costs* X X X X X X X X
Eye Substudy Cφ C
Visual Acuity X X X X
MIND:Cognitiveξ X X X
MIND: MRI Cλ C λλ
Serum Storage C C C C C
EDTAPlasma Storage C C
Urine Storage C C
Phone f/u# X X X X X Intensive Glyc Group: Phone calls must be made between all regularly scheduled clinic visits

40
Notes for Table 2.2A:
X: This evaluation/procedure applies at this visit
τ: A second screening visit is required to document hypertension for potentially eligible screenees (see Figure 2.2):
(a) not currently on antihypertensive therapy
(b) but who had a SBP > 130 mm Hg on the first clinic visit
(c) and for whom there is no notation in the medical record of another SBP > 130 mm Hg within 3 months prior to randomization.
1: prn (‘as needed’) includes:

(a) Monitoring K+ and Creatinine after starting and/or significantly changing ACEI, ARB and thiazides
(b) Patients initiated on thiazolidinediones will be monitored, as recommended by the manufacturer, with ALT levels every two months for the first 12
months after the initiation of this therapy, and annually thereafter.
γ: Milepost blood pressure visits (marked as M) are only for participants who are assigned to the intensive BP group. After 2 years of follow-up, these visits
will occur annually.
∆: Each participant in the Intensive Glycemic Group will have a point-of-care (POC) HbA1c measurement at each clinic visit.
* These evaluations will be done in a subset of participants (4288 participants in the Cost Study and, within this subset, 2000 will complete HRQL, diet and
physical activity assessments [i.e., all 2000 participants are in HRQL/diet/physical activity])
φ
For the Eye Substudy (in a subset of 4065 participants), the baseline eye exam/fundus photography can be performed up to 2 months post-randomization.
ξ
For the clinics participating in the MIND Cognitive Substudy (conducted in a subset of 2,800 participants), a battery of cognitive neuropsychological tests
will be obtained at 1, 20 and 40 months post-randomization. (The 1 month visit will serve as the baseline visit.)
λ In addition to the neuropsych tests, a subsample of 640 MIND participants will have a baseline MRI within 45 days after the baseline neuropsych test date.
λλ Participants in the MRI portion of MIND will have a follow-up MRI +/- 45 days around the 40 month neuropsych test date.
# In addition to the phone contacts noted in the table, calls must also be made between all other regularly scheduled clinic visits
† Measurement documented in source notes only.
Scrn=Screening Visits; BL=Baseline Visit; C=Central reading center or lab; POC=Point of Care; L=Local lab; BP=blood pressure; CPK=Creatine
phosphokinase; FPG=fasting plasma glucose

41
Table 2.2B: Scheduled Examination Components by Visit: For Participants Randomized to the Intensive Glycemia + Standard Blood Pressure Groups
Schedule in Months
τ
BP/Pulse X X X X X X X X X X X
† † † † † † † † †
γ
BP Mileposts (none)
∆
FPG C C C C C C C
CPK C C
ECG L C C C C
HRQL* X X X X
φ
Eye Substudy C C
ξ
MIND:Cognitive X X X
λ
MIND: MRI C C λλ
Urine Storage C C
Phone f/u# X X X X X X Intensive Glyc Group: Phone calls must be made between all regularly scheduled clinic visits

42
Notes for Table 2.2B:

γ: Milepost blood pressure visits are only for participants in the Intensive BP group.
φ
ξ
# In addition to the phone contacts noted in the table, calls must also be made between all other regularly scheduled clinic visits.
† Weight measurement documented in source notes only.

43
Table 2.2C: Scheduled Examination Components by Visit: For Participants Randomized to the Intensive Glycemia Group + Lipid Trial
Schedule in Months
τ
BP/Pulse X X X X X X X X X X
† † † † † † † † †
γ
BP Mileposts (none)
∆
FPG C C C C C C C
Potassium C C C C C C
Creatinine L C C C C C C C C C C
σ
Lipid Profile L C C C C C C C C
ALT L C C C C C C C C C
CPK C C C C C C C C C
ECG L C C C C
HRQL* X X X X
φ
Eye Substudy C C
ξ
λ
MIND: MRI C C λλ
Urine Storage C C
Phone f/u# X X X X X Intensive Glyc Group: Phone calls must be made between all regularly scheduled clinic visits

44
Notes for Table 2.2C:

φ
ξ
# In addition to the phone contacts noted in the table, calls must also be made between all other regularly scheduled clinic visits.
σ An additional lipid profile would be required at the next 4 month visit (after dietary/adherence counseling) if notified by the Coordinating Center that the
LDL-C has exceeded 130 mg/dl (3.36 mmol/L) and/or that the triglyceride level has exceeded 750 mg/dl (8.47 mmol/l) (see Section 3.3.c for details)

45
Table 2.2D: Scheduled Examination Components by Visit : For Participants Randomized to the Standard Glycemia + Intensive Blood Pressure Groups
Schedule in Months
τ
Clinic Visit X X X X X X prn X X X X X X X X X X X X X
BP/Pulse X X X X X X X X X X X X X X X X X X
† † † † † † † † †
γ
BP Mileposts Μ Μ Μ Μ Μ Μ Μ
∆
HbA1c (POC) (as needed)
FPG C C C C C C C
CPK C C
ECG L C C C C
HRQL* X X X X
Eye Substudy Cφ C
MIND:Cognitiveξ X X X
MIND: MRI Cλ C λλ
Urine Storage C C
Phone f/u (as needed)

46
Notes for Table 2.2D:

γ: Milepost blood pressure visits (marked as M) are only for participants who are assigned to the intensive BP group. After 2 years of follow-up, these visits
will occur annually.
∆: Only participants in the Intensive Glycemic Group need to have a point-of-care (POC) HbA1c measurement at each clinic visit.
φ
ξ

47
Table 2.2E: Scheduled Examination Components by Visit:For Participants Randomized to the Standard Glycemia + Standard Blood Pressure Groups
Schedule in Months
τ
Clinic Visit X X X X X X X X X X X X
BP/Pulse X X X X X X X X X X X
†
Weight X X X X X X X X X X X
γ
BP Mileposts (none)
∆
HbA1c (POC) (as needed)
FPG C C C C C C C
CPK C C
ECG L C C C C
HRQL* X X X X
φ
Eye Substudy C C
ξ
λ
MIND: MRI C C λλ
Urine Storage C C

48
Notes for Table 2.2E:

φ
ξ

49
Table 2.2F: Scheduled Examination Components by Visit: For Participants Randomized to the Standard Glycemia Group + Lipid Trial
Schedule in Months
τ
Clinic Visit X X X X X X X X X X X X
BP/Pulse X X X X X X X X X X
†
Weight X X X X X X X X X X X
γ
BP Mileposts (none)
∆ (as needed)
HbA1c (POC)
FPG C C C C C C C
Potassium C C C C C C
Creatinine L C C C C C C C C C C
σ
Lipid Profile L C C C C C C C C
ALT L C C C C C C C C C
CPK C C C C C C C C C
ECG L C C C C
HRQL* X X X X
φ
Eye Substudy C C
ξ
λ
MIND: MRI C C λλ
Urine Storage C C

50
Notes for Table 2.2F:

φ
ξ
† Measurement documented in source notes only
σ An additional lipid profile would be required at the next 4 month visit (after dietary/adherence counseling) if notified by the Coordinating Center that the
LDL-C has exceeded 130 mg/dl (3.36 mmol/L) and/or that the triglyceride level has exceeded 750 mg/dl (8.47 mmol/l) (see Section 3.3.c for details)

51
2.4.c Other Clinical Measures
2.4.c.1 Ankle Brachial Index
The ankle brachial index (ABI) is a hemodynamic measure that identifies and
quantifies severe arterial obstructive disease in the lower extremities. The ABI is be a
measure of subclinical cardiovascular disease, and persons with low ABI may at
increased risk of clinical cardiovascular disease. In ACCORD, the ABI may be measured
during the screening process to assist in the identification of a high risk subgroup of
persons with diabetes but no clinical cardiovascular disease. Measurement of ABI is not
required.
2.4.c.2 Electrocardiography
A 12-lead ECG is obtained at baseline in order to assess eligibility, and at the

biennial follow-up visits (i.e., every 2 years) and close-out visit to ascertain the
occurrence of silent (unrecognized) MI. The baseline ECG is used to identify previous
(including silent) MIs, and to identify evidence of left ventricular hypertrophy.
2.4.d Laboratory Procedures
The schedule for laboratory procedures is shown in Tables 2.2A through 2.2F.
Data regarding glycemic control (fasting plasma glucose and HbA1c) are important for
determining eligibility status (see Section 2.1). During follow-up, HbA1c levels are used
to enable the titration of hypoglycemic therapy to goals. Level of control also serves as
an important variable in analyses exploring the mechanism of effect of hypoglycemic
therapy on outcomes.
Blood and urine samples will be stored for future measurements of other less
traditional risk factors. White blood cells will also be stored for future DNA extraction
for genetic studies. It may prove possible to identify subgroups, defined by specific
genes or genetic markers, which respond differentially to the various treatment strategies.
For safety purposes, potassium, ALT, CPK and creatinine measurements will be
performed periodically (see Tables 2.2A through 2.2F).
2.4.e Drug Dispensing, Ordering, Storage, and Disposal
Drug Dispensing
The complexity created by the large number of medications and multiple

treatment strategies requires substantial attention to the process of medication dispensing.
All study medications dispensed to the participants will be labeled and identified with the
study name, participant’s name, medication name, strength and quantity, directions for
use, and authorized prescriber’s name. An emergency study-related phone number for
study drug information will also appear on the label. All participants are instructed orally

52
on medication administration. Written instructions will also be provided. (See also

Chapter 9: Adherence.)
Participants receive medication supplies at scheduled visits in sufficient quantity

to last until the next scheduled visit. Medication dispensing may occur in the intervening
periods between visits in case of emergency, loss, or schedule changes. A tracking
mechanism is maintained for all dispensing actions. It is recommended that authorized
dispensing personnel be limited in number to assure proper adherence with established
accountability and dispensing procedures.
Drug Supply Ordering
Each Clinical Site, upon completion of procedures for study initiation, will
receive a standard initial shipment (determined by the Coordinating Center and prepared
by the Drug Distribution Center) of study drug supplies for each portion of the trial. It is
expected that this initial shipment will suffice for a specified number of visits for a given
number of randomized participants. Subsequent ordering for these and additional
participants will then become the responsibility of each Clinical Site.
The Drug Distribution Center (DDC) in consultation with each Clinical Site sets
inventory levels for each item. When an item reaches the reorder point, additional stock is
automatically shipped from the DDC.
Drug Receipt and Storage
Drug shipments are sent to the Clinical Site in care of a designated staff member.
The shipment is inspected for damage and its contents reconciled with the accompanying
ACCORD Shipping Notice. The inventory is logged using the established tracking
mechanism. Packing slips are filed in a secure location. Any damage or discrepancies in
the shipment are to be reported promptly to the Drug Distribution Center for corrective
action. Each Clinical Site is responsible for storing the study drug supplies in a locked,
secure area with limited access. Manufacturer recommendations and local policies for
drug storage are followed.
Drug Disposal
Clinical Sites are authorized to destroy ACCORD stock locally, complying with
any local policies and procedures. Destruction will be documented on the ACCORD
Local Destruction Form, with a copy sent to the DDC. All study drugs are labeled with an
expiration date. Prior to expiration, the DDC will automatically ship replacement stock.
Notification of these shipments will be made via the Coordinating Center. Once
replacement stock is received the clinical site will destroy expired stock and document
destruction as described above.

53
Chapter 3
Interventions
3.1 Introduction
ACCORD is designed to test the effects on CVD events of (1) intensive glycemia
control compared with the current standard of care for glycemia, (2) raising HDL-
cholesterol and lowering triglycerides with fibric acid therapy in the context of desirable
LDL-C, and (3) intensive blood pressure control compared with standard blood pressure
control.
This chapter presents descriptions of the three trial interventions. The chapter also
presents the lifestyle/background recommendations provided for all ACCORD
participants.
All interventions and lifestyle recommendations will begin at randomization.
3.2 Glycemic Control Intervention
3.2.a Glycemia Research Question
cardiovascular disease (CVD) event, does a therapeutic strategy that targets a HbA1c of < 6.0%
reduce the rate of CVD events more than a strategy that targets a HbA1c of 7.0% to 7.9% (with
the expectation of achieving a median level of 7.5%) ?
3.2.b Research Design
Ten thousand (10,000) individuals with type 2 diabetes who meet the ACCORD
eligibility criteria (see Section 2.1) will be randomized to one of two different glycemic targets:
an HbA1c of < 6% or an HbA1c of 7.0% to 7.9%. Several approaches will be used to achieve
and maintain near normal glycemia in the intensive group, including a minimum of bimonthly
visits, telephone contacts, point-of-care HbA1c testing, targeting postprandial and preprandial
glucose levels, aggressive early use and titration of several different oral agents, self-titration
strategies, early use of insulin, and emphasis on combinations of agents.
3.2.c Glycemic Targets
In this trial, participants will be randomized to one of two treatment groups based on the
targeted level of glycemic control. Both the intensive and standard therapy groups will utilize
all currently available glucose-lowering therapies. The two treatment groups will have
different glycemic targets and will have different thresholds of glycemic control at which
therapeutic changes will be considered (Table 3.1).

54
Table 3.1: Glycemic Targets and Thresholds for Action for ACCORD
“Action Required” Threshold
Group HbA1c Targets HbA1c > 50% of SMBG Results/4 days
> 7.9%* or < 6.5%# (anytime)
or
Standard Therapy 7 – 7.9% fasting/ac < 90 mg/dl (5.0 mmol/l)#
6.6%-6.9% # (twice
consecutively)
fasting/ac > 100 mg/dl (5.6 mmol/l)
Intensive Therapy < 6.0% ≥ 6.0%* or
2 hrs pc > 140 mg/dl (7.8 mmol/l)*
pc: postcibal; ac: antecibal; SMBG: self monitoring of blood glucose; *antihyperglycemic therapy will be
advanced if either the HbA1c or the SMBG “action required” criteria are met at any participant encounter
#
therapy with drugs that increase the risk of hypoglycemia (e.g. insulin, sulfonylureas, meglitinides) will
be reduced to avoid hypoglycemia if these criteria are met
To achieve these glycemic targets, participants will require self-management education

and dietary and lifestyle interventions, as well as pharmacologic therapy. They will also require
different drug choices and treatment intensities. For example, within 6 months of
randomization, most intensive group participants will likely be on 3 or more injections of
insulin per day in addition to 2 or 3 oral agents. Conversely, standard therapy participants are
less likely to be on insulin, will be on < 2 injections per day if insulin is used, and will be
taking fewer oral agents. Moreover, the frequency with which self-management behavior is
applied and participants are contacted will vary between the two levels of glycemic control.
3.2.d Self-Management Education
The goal of self-management education is to empower the participant to take

responsibility for making the day-to-day changes in therapy required to maintain the targeted
level of glycemic control. Proficiency in self-monitoring of blood glucose (SMBG) is a key
component of self-management, as is knowledge of how to use SMBG data to alter therapy to
achieve target glycemia. The importance of self-management and SMBG will be stressed.
Indeed, SMBG will be expected of all ACCORD participants, and unwillingness or inability to
do SMBG is an exclusion criterion (Chapter 2). Instructions and information on SMBG will be
made available to all participants.
3.2.e Dietary and Lifestyle Interventions
All participants will be provided with the same dietary and lifestyle recommendations to
optimize their glucose control. These will include: a) advice that blood glucose control may be
more critical than weight control in reducing the risk of complications of diabetes; b) teaching
dietary principles including carbohydrate counting; c) advice to engage in regular aerobic
exercise (if medically fit to do so according to the physician who provides their medical care);
d) teaching the technical and interpretative skills of blood glucose monitoring; and e) education
of participants’ families regarding the management of hypoglycemia.

55
Specific dietary and exercise recommendations will be tailored to each participant.

Because group consultation is as effective as individual consultation for achieving glycemic
improvement, sites may utilize either approach.
3.2.f Approach to Targeting and Achieving Different Levels of Glycemic Control
Targeting and achieving two different levels of glycemic control (Table 3.1) without
causing clinically significant hypoglycemia is critically important to the success of the trial.
Differences in visit frequency, the intensity and frequency of inter-visit contacts, the prompt
response to HbA1c results, the frequency of SMBG, and different approaches to self-
adjustment of glycemic therapy based on SMBG results and carbohydrate intake (if on insulin)
will be used to achieve these two levels of glycemic control. Table 3.2 summarizes the
different approaches that will be implemented in the standard and intensive groups to target the
levels described in Table 3.1. As noted above, self management and SMBG are part of every
participant’s care in ACCORD. The standard and intensive groups will differ in the intensity
of these activities as noted in Table 3.2.
Table 3.2 Achieving Glycemic Goals

Standard Group Intensive Group
Visits (1st 4 months) Monthly – Q 4 mo* Monthly
Visits (> 4 months) Q 2 – 4 mo* Q 2 mo
Phone contact Participant initiated (prn) Research staff initiated (≥1 inter-visit)
Supplemental contact Severe hypoglycemia/hyperglycemia Severe hypoglycemia OR
HbA1c in action required range HbA1c in action required range OR
Frequent (>50%/4 days) premeal SMBG SMBG in action required range (based on
levels <90 mg/dl (5.0 mmol/l) review of logbooks)
Point of Care HbA1c Optional Mandatory
Routine use of postprandial No Yes
SMBG values to guide therapy
SMBG freq.a (not on insulin) ≤7/wk (daily at different times or >1/day on ≥ 2/day and 4/day if glucose is > target (2
certain days) ac/day and 2 pc/day)
SMBG freq. a (on insulin) ≤3/day 4-8/d (at least 2 ac/day and 2 pc/day;
occasional 3 am test prn)
Self titration principles Avoid severe hypoglycemia and premeal Avoid severe hypoglycemiab AND
SMBG levels < 90 mg/dl (5.0 mmol/l) Adjust Rx q4d AND
Use CHO/patterns (if on insulin Rx)
Initial Minimum Rx Diet/lifestyle Diet/lifestyle AND 2 oral agents
Insulin Use (when needed) Generally ≤ 2 injections/day Flexible
* depending on the blood pressure group to which the person has been assigned; aless frequent if goals are
achieved; b including avoiding SMBG levels < 70 mg/dl (3.9 mmol/l) on > 1/4 of the readings.
3.2.g Visit Frequency and Inter-Visit Contacts
Tables 2.2A through 2.2F describe the activities to be performed at each ACCORD
follow-up visit. Participants will have different scheduled visit frequencies based on their
allocated glycemic therapy group, their most recent HbA1c (if a central measurement and a
point-of-care measurement differ, the higher of the two will be used), and other clinically
important considerations, such as hypoglycemic episodes. The importance of contacting the
research staff if any of the following occurs will be reinforced: any major illness or

56
hospitalization, any new diagnosis or drug prescription, any episode of hypoglycemia requiring
assistance, or any other concerns regarding their therapy. Supplemental visits will be arranged
whenever required.
Both the standard and intensive therapy groups will have a visit 1 month after
randomization. Subsequently, individuals in the standard therapy group who are also allocated
to intensive blood pressure therapy will have monthly visits until month 4, and then bimonthly
visits for the rest of the trial; the remaining standard therapy participants will have a visit at 4
months and then every 4 months thereafter. Conversely, all intensive group participants will
have monthly visits for the first 4 months and bimonthly visits thereafter. In addition, the
research staff will contact all intensive group participants on at least one occasion between
these visits (by telephone, FAX or email) to reinforce adherence, answer any questions, check
for serious adverse events (including severe hypoglycemic episodes requiring third party
assistance), review self-monitoring of blood glucose (SMBG) records, and determine whether a
supplemental visit is required. Supplemental contacts will occur for any participant who has
experienced an episode of severe hypoglycemia or whose last HbA1c is in the “action
required” range, and for any intensive therapy participant whose SMBG values are above the
targets noted in Table 3.1. Finally, all intensive therapy participants will be asked to mail,
email, FAX or telephone biweekly logs of their capillary glucose values so that the research
assistant can respond to them in a proactive fashion.
3.2.h Response to HbA1c Results
Local immediate measurement of HbA1c (using a point-of-care testing system at each

Clinical Site) will be used to guide prompt changes in therapy at each visit. Such an approach
provides immediate feedback regarding glycemic control to both patients and clinical staff, and
has been shown to lead to better glycemic control than more conventional laboratory-based
approaches. Every participant in the Intensive Group will have a point-of-care HbA1c
measurement at each clinic visit; this measurement may also be made in the standard group at
the discretion of the research staff. The results will be recorded, along with the action taken in
response to the result. Such action must be taken and documented whenever a participant’s
HbA1c is within the “action required” ranges specified in Table 3.1. The central lab and point-
of-care HbA1c results will be compared regularly to ensure that they are similar. If there is a
systematic difference between these results, the “action required” HbA1c thresholds in Table
3.1 (that were chosen based on “gold standard” HbA1c results measured in a central lab) may
be translated into “point-of-care” HbA1c thresholds for action to ensure that these systematic
differences are taken into account. For example, if the point-of-care HbA1c result consistently
reads 0.2% lower than the central lab result, the “action required” threshold for a point-of-care
measurement in the intensive group would be 5.8%.
HbA1c will also be measured centrally every 4 months. This measure will be used as
the HbA1c value for reporting the study results, and provides a quality control check for the
individual point-of-care samples. Sites will be notified by the Coordinating Center (CC)
whenever a participant’s centrally measured HbA1c is in the “action required” range; such
notification will be linked to a note reminding the Clinical Site of the participant’s treatment
group assignment and the glycemic goals for that group. A response from the Clinical Site

57
regarding the changes in therapy made to achieve or maintain target levels will be required on
case report forms after any such notification.
3.2.i Frequency of Self-Monitoring of Blood Glucose (SMBG)
All participants will also be asked to do self-monitoring of blood glucose (SMBG)

according to the frequency noted in Table 3.3. Less frequent testing may be acceptable if
participants have safely achieved the glycemic targets specific for their group. The SMBG
results will be used to ensure that individuals in both groups are not having frequent
hypoglycemic episodes (defined in Table 3.2) and to guide adjustments in therapy to prevent
hypoglycemia.
In addition, these levels will be used for the intensive therapy participants to intensify
therapy. The next dose or drug will be introduced for individuals in whom >50% of the fasting
SMBG values exceed 100 mg/dl (5.6 mmol/l) and in whom > 50% of the 2 hour postprandial
values exceed 140 mg/dl (7.8 mmol/l). Thus, therapy in the intensive group will be intensified
on the basis of either these SMBG values, or any HbA1c >6% (provided that intensification is
not contraindicated in the judgment of the investigator because of frequent severe
hypoglycemic episodes or other serious adverse effects).
Table 3.3: Self-Monitoring Blood Glucose (SMBG) Targets and Frequency

SMBG Standard Group Intensive Group
≤7 tests/week
≥ 2/day; QID if > target
Frequency (diet/oral tx only) (daily at varying times or more
(2 ac/day and 2 pc/day)
frequently on selected days)
4-8 tests/day
≤3 tests/day
Frequency (if on insulin) (at least 2 ac/day and 2 pc/day; and the
occasional 3 am test prn)
*the target is for at least 50% of the SMBG values to be in this range
3.2.j Self-titration of Anti-hyperglycemic Therapy
Standard therapy participants will be provided with simple algorithms to allow them to
self-titrate their oral therapy or insulin to avoid hypoglycemia. They will also be instructed to
call the clinic if they are recording frequent low SMBG values (see Table 3.2); if they have any
episode of severe hypoglycemia; if they are experiencing frequent episodes of symptomatic
hypoglycemia (>1/week); or if they have any symptoms of hyperglycemia. In these instances,
therapy can be adjusted.
Intensive therapy participants will be provided with algorithms to allow them to self-
titrate their oral therapy or insulin (i.e., make changes every 4 days) according to the pattern of
their SMBG results and to avoid hypoglycemia or hyperglycemia. Moreover, participants
requiring insulin will also be taught how to vary their dose according to the carbohydrate
content of meals, with supplemental adjustments for ambient glucose levels and variations in
exercise.

58
3.2.k Adjustment of Glycemic Therapy
The target and “Action Required” HbA1c and SMBG values for both groups are noted
in Table 3.1. These targets will be achieved by using the same combination of dietary, lifestyle
and pharmacologic approaches in both groups. As outlined above, however, the groups will
differ in the intensity of follow-up, frequency of changes to glycemic therapy, and self-titration
interventions. Whenever antihyperglycemic therapy needs to be increased (to reduce the
HbA1c), participants will either move to a higher dose of their current therapy or, if already on
the highest dose, will move to the next agent. For example, if action is required for a
participant on maximum dose of metformin, sulfonylurea and a thiazolidinedione, evening
insulin will be added.
The suggested algorithm for pharmacologic interventions is shown in Figures 3.1 and
3.2. For participants on intensive therapy whose HbA1c values are in the “Action Required”
range (i.e., > 6%), it calls for immediate institution of combination therapy with 2 classes of oral
agents. It also calls for self-titration of therapy between visits for the intensive group as
described in Section 3.2.j and in Table 3.2, and for titration of therapy at the visits based on the
HbA1c or the SMBG results.
The exact time at which insulin will be started in individuals not taking insulin at the
time of randomization is not explicitly defined. Nevertheless, evening basal insulin will be
added for intensive group participants on maximal oral therapy whenever their glucose values
are in the “Action Required” range as noted in Table 3.1. Moreover, sites will be prompted to
add rapid acting insulin to intensive group individuals whose HbA1c is in the “action required”
range with postprandial SMBG levels > 140 mg/dl (7.8 mmol/l). Figure 3.3 describes the
algorithm for the use of insulin.
Antihyperglycemic therapy will not be reduced for participants in either group whose
HbA1c is within or above the target range (noted in Table 3.1) unless required because of
severe hypoglycemia or adverse effects.
Antihyperglycemic therapy will be reduced for participants in the standard group for
the following reasons (Figure 3.2):
1. any severe hypoglycemia
2. more than 1 episode of symptomatic hypoglycemia per week
3. > 50% of SMBG levels < 90 mg/dl (5 mmol/l)
4. adverse effects of antihyperglycemic drugs
5. HbA1c < 6.5% on one occasion or 6.6-6.99% on 2 consecutive occasions and either
on insulin or a secretagogue, a history of 1 or more episodes of symptomatic
hypoglycemia since the previous visit, or 1 or more SMBG levels below 90 mg/dl (5
mmol/l) since the previous visit.

59
3.2.l Glycemia Medications Available Within ACCORD
The following classes of antihyperglycemic drugs are available within ACCORD:

a) biguanides (e.g., metformin)
b) secretagogues (e.g., sulfonylureas such as glimepiride and meglitinides such as repaglinide)
c) thiazolidinediones (e.g., rosiglitazone)
d) alpha-glucosidase inhibitors (e.g., acarbose)
e) insulins (e.g., NPH, ultralente, glargine, aspart, regular).
3.2.m Alternatives and Contraindications for Glucose-Lowering Drugs
Acarbose may be used at the investigator’s discretion to deal with postprandial spikes
that may be difficult to control with other medications. Whether or not acarbose is used does
not influence the algorithm in Figure 3.1.
Repaglinide, an insulin secretagogue in the meglitinide (benzoic acid derivative) class,

may be substituted for sulfonylurea therapy in those individuals with erratic meal schedules or
with hypoglycemia or sustained postprandial hyperglycemia. Repaglinide and sulfonylureas
should not be combined because they are both insulin secretagogues.
Metformin may have gastrointestinal side effects especially if high initial doses are
used. Therapy will therefore be initiated at a dose of 500 mg with dinner, increasing the dose
by 500 mg every week until the patient meets target goals or reaches the clinically effective
maximum dose of 1000 mg twice/day or is unable to tolerate higher doses. Contraindications to
the use of metformin include a) serum creatinine >1.4 mg/dl for women or > 1.5 mg/dl for
men, b) drug-treated congestive heart failure, c) severe obstructive pulmonary disease, d)
evidence of significant impairment of hepatic function (AST or ALT > 2.5 times the upper
limit of normal), e) ongoing metabolic or respiratory acidosis, or f) other high risk condition
for the development of acidosis or cardiovascular collapse.
Thiazolidinediones may cause fluid retention (including edema, anemia and CHF),
liver toxicity, ovulation, and weight gain. Contraindications to their use include: a) ALT >2.5X
upper limit of normal at start of therapy, or b) NYHA Class III or IV CHF. They should be
used with caution in patients with prior edema. Rosiglitazone will be the thiazolidinedione
provided by the study. Patients with mildly elevated liver enzymes (ALT levels <2.5X upper
limit of normal) at baseline or during therapy with rosiglitazone should be evaluated to
determine the cause of the liver enzyme elevation. Initiation of, or continuation of, therapy
with rosiglitazone in patients with mild liver enzyme elevations should proceed with caution
and include close clinical follow-up, including more frequent liver enzyme monitoring, to
determine if the liver enzyme elevations resolve or worsen. If at any time ALT levels increase
to >3X the upper limit of normal in patients on therapy with rosiglitazone, liver enzyme levels
should be rechecked as soon as possible. If ALT levels remain >3X the upper limit of normal,
therapy with rosiglitazone should be discontinued.

60
Sulfonylurea contraindications include a) the use of repaglinide, b) severe allergic

reaction to sulfa containing compound (anaphylaxis, Stevens-Johnson).

61
Figure 3.1
Treatment Algorithm for Intensive Glycemic Therapy Group (Goal: HbA1c<6%)
Enter on 1 oral agent* Enter on 2 oral agents*
Enter on no oral agent* On SU On TZD On SU and On Metformin and On SU and TZD

Convert to On Met Convert to Metformin TZD Convert to Protocol
Protocol SU Protocol TZD Convert to Protocol SU Convert to Protocol TZD Agents
Initiate SU and Initiate and titrate Initiate SU Initiate - titrate SU Titrate Metformin Titrate SU
MET Metformin Titrate Metformin Consider Max Titrate Metformin Consider Max Consider Max
Titrate Metformin
Action Action Action Action Action Action Action

No required? No required? No required? No required? No required? No required? No required?
Yes Yes Yes Yes Yes Yes Yes
At maximal At maximal At maximal At maximal At maximal At maximal At maximal

Met dose? Met dose? Met dose? Met dose? Met dose? Met dose? Met dose?

Titrate SU Titrate SU Titrate SU Titrate TZD Titrate SU Titrate TZD Titrate TZD
if not max if not max if not max
Action Action Action Action Action Action Action

No required? No required? No required? No required? No required? No required? No required?
At maximal At maximal At maximal At maximal At maximal At maximal At maximal

SU dose? SU dose? SU dose? TZD dose? SU dose? TZD dose? TZD dose?
Yes
Add and Titrate 3rd agent
Titrate SU to max clinically effective dose
Titrate Metformin to 1000 mg BID
Titrate TZD to max dose
Intensive Therapy Group Or
Action required? If clinically indicated may go straight to Insulin Therapy
HbA1c>6% or SMBG data (>50%) For Insulin initiation and/or adjustment, go to Figure 3-3.
Pre meal > 100 mg/dl (5.6 mmol/l)
Post-meal > 140 mg/dl (7.8 mmol/l)
Repaglinide may be substituted for SU if clinically appropriate

62
Figure 3.2:
Treatment Group Algorithm for Standard Glycemia Therapy Group (Goal: HbA1c 7% to 7.9%)
Start Here:
Check HbA1c
Optimize MNT and

No Increase 1 Agent by 1 Dose
HbA1c Increment OR
< 9% Optimize MNT and
Add an Agent if indicated
Yes
Optimize MNT OR
HbA1c Yes Increase 1 Agent by 1 Dose
> 8% Increment OR
Optimize MNT X 1-2 Months Add an Agent if Indicated
and Repeat HbA1c
No
HbA1c Yes No Action

7 - 7.99% Required
No
STEP A
HbA1c Repeat No Return to
Any Severe Hypoglycemia Yes HbA1c
6.6-6.99% HbA1c STEP A
OR < 6.99%
Symptomatic Hypoglycemic Episodes > 1/wk No Next Visit
Yes
OR
No
>50% of CBG Levels <90mg% (5 mM)
OR
Adverse Effects of Drugs HbA1c
< 6.5%
On Insulin or Secretagogue
OR No
No Action
Any Symptomatic Hypoglycemia
Yes Required
OR
Any CBG <90mg% (5 mM)
Reduce Dose of a Yes

Glucose Lowering Drug

63
Figure 3.3:
Use of Insulin for Participants On Maximal Oral Therapy
FOR INTENSIVE GROUP FOR STANDARD GROUP

PARTICIPANTS PARTICIPANTS
Add Evening Glargine; Add Evening NPH;

Start @ 10 U or 0.2U/kg/day
Start @ 10 U or 0.2U/kg/day
Titrate Dose Titrate Dose
Adjust Adjust
Doing No Doing No pills or
pills or Well?*
Well?* insulin as insulin as
needed needed
Yes
Yes
HbA1c No HbA1c No HbA1c No

Too High? No Change Too High? Too Low?
Yes Yes
Yes
No On Insulin No On Insulin
> 0.5U/kg/ > 0.5 U/kg/
day? day?
Reduce/
Yes Yes stop
Stop Secretagogue; Insulin
Stop Secretagogue or one
Add AM NPH oral med
Add/titrate Aspart/
Lispro (or R) Titrate Insulin
before 1,2, or 3 Dose(s) as No
meals, as needed Needed Change
Adjust
Adjust Doing
No pills or
Doing No pills or insulin
Well?*
Well?* insulin as as
needed needed
Yes
Yes Yes
HbA1c No HbA1c No
HbA1c No Too High? Too Low?
Yes
Too High? No Change
Yes
Add/titrate R
at Bkfst/Dinner
*Doing well: no severe hypoglycemic or adverse event or no reason to reduce therapy (as described in Figure 3.2)

64
3.3 Lipid Intervention
3.3.a Lipid Research Question
In middle-aged or older people with type 2 diabetes who are at high risk of having
a CVD event and in the context of good glycemic control, does a therapeutic strategy that
uses a fibrate to raise HDL-C/lower triglyceride levels and uses a statin for treatment of
LDL-C reduce the rate of CVD events compared to a strategy that only uses a statin for
treatment of LDL-C? The specific fibrate to be used in ACCORD is fenofibrate and the
specific statin is simvastatin.
The lipid component of ACCORD is a fully masked, randomized trial of 5,800

participants. Eligible participants will be randomized to fenofibrate or placebo; all
participants will be treated with simvastatin. To be eligible for the lipid trial, the
observed or estimated LDL-C at screening (in the absence of treatment) must be between
60 and 180 mg/dl (1.55 and 4.65 mmol/l), inclusive. HDL-C must be less than 55 mg/dl
(1.42 mmol/l) for women or Blacks/African-Americans, or less than 50 mg/dl (1.29
mmol/l) for all other gender-race groups. Other eligibility criteria are noted in Section
2.1.
The upper limit for triglyceride (TG) eligibility for screenees not on a lipid
lowering agent is 750 mg/dl (8.47 mmol/l) and 400 mg/dl (4.52 mmol/l) for screenees on
a lipid-lowering agent. It is expected that initial diet and glucose control will rapidly
reduce TG levels in the very few participants near these limits. If an untreated participant
has a TG level between 400 (4.52 mmol/l) and 750 mg/dl (8.47 mmol/l), he/she will have
a beta-quantification performed by ultracentrifugation by the Central Chemistry
Laboratory to allow direct determination of LDL-cholesterol level. Ten percent of
participants are expected to be in this range.
The 4,200 participants who are not enrolled in the lipid portion of ACCORD (i.e.,
the 4,200 participants in the blood pressure portion of the trial) will be treated by their
usual physicians (who may also be study investigators). The recommended LDL-C goals
for these 4,200 participants will be based on the National Cholesterol Education Program
(NCEP) guidelines (National Cholesterol Education Program 2001). Based on published
data on the percent of participants reaching goals, it is expected that this group will have
a mean LDL-cholesterol of about 110 mg/dl (2.84 mmol/l). As noted in Section 1.2.f, the
2001 NCEP guidelines define diabetes as a CHD-equivalent.
Participants who were on a lipid-lowering agent at screening must agree to stop

that treatment and be changed to simvastatin.
The starting dose of masked fenofibrate/placebo medication will be determined by

the calculated glomerular filtration rate (GFR) using the baseline serum creatinine level
and the abbreviated MDRD equation (Levey 2003). Those participants with a baseline
GFR >50 ml/min/1.73m2 will begin at a starting dose of 160 mg of fenofibrate or

65
identical placebo tablet. Those with a calculated GFR between 30 and <50 will start at
the reduced dose of 54 mg/day fenofibrate or placebo (or will be placed on 160 mg tablet
every other day if the 54 mg dose is unavailable). The masked medication should be
administered with the morning meal.
Participants in the lipid trial will have serum creatinine measured every four
months during follow-up. If the participant had started on the 160 mg dose of the masked
medication, this dose will be down-titrated if the participant’s estimated GFR falls
between 30 and <50 mL/min/1.73m2 on two consecutive measurements taken four
months apart. Participants with GFRs in this range will receive either 54 mg/day (or 160
mg every other day) of fenofibrate or matching placebo.
If the estimated GFR falls below 30 mL/min/1.73m2 at any time, the Coordinating
Center will notify the clinic site that a confirmatory blood draw for repeat estimated GFR
will be required within 2 weeks. If the confirmatory estimated GFR is below
30mL/min/1.73m2, the masked study medication will be permanently discontinued,
regardless of fenofibrate or placebo assignment.
The starting dose of open-labeled simvastatin will be determined by presence of

cardiovascular disease at randomization. Primary prevention participants (those
participants without clinical cardiovascular disease) will start at a simvastatin dose of 20
mg/day, administered once daily after the evening meal or at bedtime. Secondary
prevention participants (those with a history of clinical cardiovascular disease as defined
in Chapter 2, Section 2.1.a.6.A.) will start at a simvastatin dose of 40 mg/day.
For participants starting at 20 mg/day of simvastatin, if the LDL-C is greater than

100 mg/dl (2.59 mmol/l) on two consecutive follow-up visits, the daily dose of
simvastatin will be increased to 40 mg. Additionally, if a cardiovascular event occurs (as
defined in Chapter 2, Section 2.1.a.6.A) during follow-up, the participant’s simvastatin
dose will be increased to 40 mg/day. If, during follow-up, the LDL-C is > 120 mg/dl (>
3.10 mmol/l) on two consecutive measurements following titration of simvastatin to 40
mg/day, the participant will be referred to their own physician for individualized
treatment. This is described below in Section 3.3.c.
The order of therapy will be simvastatin first (at randomization), with the
fenofibrate/placebo started at the next monthly visit. Participants and physicians will be
masked to fibrate/placebo assignment, and to LDL-cholesterol, triglyceride, and HDL-
cholesterol levels throughout the trial. This will be the only fully masked part of the
ACCORD study.
During the trial, a fasting plasma lipid profile is scheduled to be obtained and
centrally analyzed at four months, eight months, twelve months and yearly thereafter (see
Tables 2.2C and 2.2F). Participants who have triglyceride levels greater than 400 mg/dl
(4.4 mmol/l) at any time will have a beta-quantification performed to allow for
determination of LDL-cholesterol levels at all time points. Safety profiles, including liver
function tests and CPK levels, will be determined at one month, four months, eight

66
months, and twelve months for the first year and annually thereafter. To monitor renal
function during follow-up, all lipid participants will be required to have an additional
tube of blood drawn for creatinine at the routine blood draw every 4 months, which will
be analyzed centrally (as noted in Tables 2.2C and 2.2F). If at any time the participant
has relevant symptoms or signs suggestive of drug-induced toxicity, liver function tests
and/or CPK levels will be obtained through the Central Laboratory.
3.3.c. Lipid Goals/Safety Issues
The goal of statin therapy is to achieve LDL-C values consistent with current
NCEP and ADA guidelines. Under this lipid trial protocol, primary prevention
participants will be on 20 mg simvastatin (which could conservatively lower LDL-C by
30%) and secondary prevention participants will be on 40 mg simvastatin (which could
lower LDL-C by 40%). In addition, any participant on 20 mg simvastatin whose follow-
up LDL-C values are greater than 100 mg/dl (2.59 mmol/l) on two consecutive occasions
and any primary prevention participant who experiences a cardiovascular event (Section
2.1.a.6.A) will be placed on 40 mg/day simvastatin Using these assumptions/expectations
as guides, the following conservative estimates are made:
Estimated Expected Mean

Mean LDL-C On-treatment
Baseline LDL-C* in Strata LDL-C in Strata
> 60 to < 80 mg/dl 70 mg/dl 46 mg/dl
(> 1.55 to < 2.07 mmol/L) (1.81 mmol/L) (1.19 mmol/L)
> 80 to < 100 mg/dl 90 mg/dl 59 mg/dl
> 100 to < 120 mg/dl 110 mg/dl 73 mg/dl
> 120 to < 140 mg/dl 130 mg/dl 86 mg/dl
> 140 to < 160 mg/dl 150 mg/dl 90 mg/dl
> 160 to < 180 mg/dl 170 mg/dl 102 mg/dl
*This would be the observed LDL for participants not on a lipid-lowering agent at baseline,
but an estimated LDL for participants on a lipid-lowering agent.
Estimation based on the expected LDL effects of the drug/dose participant is taking.

67
It is further estimated that 5% of the participants would be in the first stratum at

baseline, 15% in the second, 20% in the third, 25% in the fourth, 20% in the fifth, and
15% in the sixth. Thus, the expected overall mean on-treatment LDL-C would be
approximately 82 mg/dl mg/dl (2.12 mmol/L).
Also, because the upper limit for entry LDL-C is 180 mg/dl, and because 40 mg
simvastatin should provide about an average 40% percent reduction in LDL-cholesterol,
it is expected that few participants will have an on-treatment LDL-C of more than 120
mg/dl. However, if a participant has an LDL-cholesterol level that is persistently greater
than 120 mg/dl (3.10 mmol/l) even with treatment of 40 mg/day simvastatin, ACCORD
will, consistent with NCEP guidelines, take the participant off the masked study
medication and continue treatment with simvastatin until placed on a non-study statin by
his/her primary caregiver.
Specifically, if the measured LDL-C goes above 120 mg/dl (3.10 mmol/l) the
Coordinating Center will notify the clinic staff who ought to confirm compliance with the
study statin, refer the participant to a nutritionist for dietary instruction/reinforcement (if
appropriate), and schedule a blood draw for the visit four months from the visit at which
the LDL-C was above 120 (3.10). This blood specimen needs to be sent to the ACCORD
Central Chemistry Laboratory for lipid analysis.
If the participant has an LDL-C above 120 mg/dl (3.10 mmol/l) on two
consecutive visits after titrating simvastatin to 40 mg/day (even after compliance review
and dietary counseling), the following will occur:
• The investigator will be notified by the Coordinating Center to take the
participant off the fibrate/placebo pills.
• The participant will remain on simvastatin 40 mg/day until placed on non-
study statin by his/her primary caregiver.
• The site staff will make an appointment with the participant's doctor for
follow-up.
• The site staff will also provide a letter for the participant to take to his/her
physician for the follow-up visit. This letter will include the blood lipid
values and describes the medication regimen the participant was on when
the blood was drawn.
• The site staff will confirm that the participant had visited their physician.
• From that point on, the participant would be treated for lipids by his/her
personal physician and given results of any ACCORD lipid determinations
to share with this physician.
If the centrally measured LDL-C is ever less than 40 mg/dl (1.03 mmol/l) during
follow-up, the Coordinating Center will advise the clinic site. Clinic personnel should
then determine compliance with study statin and fibrate/placebo (to make sure that the
participant is not taking more than the prescribed number of pills daily), refer participant
to nutritionist for dietary counseling to ensure that the participant is eating a balanced,
adequate diet, and schedule a blood draw for the visit four months from the visit at which

68
the LDL-C was less than 40 (1.03). If the centrally measured LDL-C is ever less than 40
mg/dl (1.03 mmol/l) on two consecutive measurements, the following will occur:
participant off simvastatin.
• The participant will remain on the masked study medication.
As a minimal goal, all participants will have LDL-C lower than 120 mg/dl (3.10
mmol/l) and triglycerides less than 750 mg/dl (8.47 mmol/l) during the study.
Triglyceride values will be maintained at a level that does not pose a risk of pancreatitis.
If the centrally measured triglyceride ever exceeds 750 mg/dl (8.47 mmol/l) during
follow-up, the Coordinating Center will advise the clinic site. Clinic personnel should
then determine compliance with study statin and fibrate/placebo, refer participant to
nutritionist for dietary instruction/reinforcement (if deemed appropriate) and determine
and modify potential exacerbating disorders i.e. alcohol or simple sugar intake,
hypothyroidism, hyperglycemia. Also, the clinic needs to schedule a blood draw for the
visit four months from the visit at which the triglyceride exceeded 750 (8.47).
If the triglyceride exceeds 750 mg/dl (8.47 mmol/l) on two consecutive measurements,
even after the above measures have been conducted, the following will occur:
participant off simvastatin and the masked fibrate/placebo medication.
• The participant will be dispensed 160 mg/day tablet of fenofibrate or 600
mg BID of gemfibrozil until placed on nonstudy fibrate by his/her primary
caregiver.
• The site staff will make the appointment for follow-up by the participant’s
physician and will confirm that the appointment was kept.
• The site staff will also provide a letter for the participant to take to their
physician for the follow-up visit. This letter will include the blood lipid
values and describes the medication regimen the participant was on when
the blood was drawn.
• From that point on, the participant will be treated for their lipids by his/her
personal physician and given results of any ACCORD lipid determinations
to share with this physician.
If the estimated GFR falls below 30 mL/min/1.73m2 at any time, a confirmatory

blood draw will be required within 2 weeks. If the estimated GFR remains below 30
mL/min/1.73m2, the masked study medication will be permanently discontinued,
regardless of fenofibrate or placebo assignment.
If the masked fibrate/placebo study medication is stopped for any reason, neither
the participant nor the clinic staff need to be unmasked regarding the study medication’s
true identity, unless there are other circumstances dictating unmasking.

69
3.4 Blood Pressure Control Intervention
3.4.a Blood Pressure Research Question
In middle aged or older people with type 2 diabetes who are at high risk of having
a CVD event and in the context of good glycemic control, does a therapeutic strategy that
targets a systolic blood pressure (SBP) of < 120 mm Hg reduce the rate of CVD events
The blood pressure component of ACCORD is an unmasked, open labeled,

randomized trial of 4,200 participants. Eligibility criteria are described in Sections 2.1.a,
2.1.b, and 2.1.d.
If the investigator believes the participant is likely to be eligible for the BP

intervention, medications may be adjusted prior to the randomization visit to determine
whether the participant’s SBP will rise or fall to the BP criteria. No more than 2 visits
after any adjustment of any antihypertensive therapy will be permitted for a participant to
meet the BP eligibility criteria before entry into ACCORD.
If previously untreated for hypertension, a participant should have documentation

of SBP >130 mm Hg on 2 visits within 3 months prior to the randomization visit in order
to be included in the BP intervention.
There are no diastolic blood pressure (DBP) inclusion criteria.
3.4.c Blood Pressure Goals
Participants eligible for the BP component will be randomized to one of two

goals: SBP <120 mm Hg for the more intense goal and SBP <140 mm Hg for the less
intense goal. Figures 3.4 and 3.5 describe the treatment algorithms for the two blood
pressure treatment groups.
For ACCORD participants not participating in the blood pressure portion of

ACCORD (i.e., the 5,800 participants in the lipid portion of the trial), recommendations
for BP treatment will be made to their usual source of care. (See Section 3.5.e).
3.4.d Antihypertensive Classes (Agents)
Use of once daily preparations of the study antihypertensive agents will be

encouraged unless alternate dosing frequency (e.g., BID) is indicated.

70
The following classes of agents may be used and are provided by the study.
Angiotensin converting enzyme (ACE)-inhibitors

Diuretics
Beta-blockers
Dihydropyridine and Non-dihydropyridine CCBs
Alpha-blockers
Angiotensis II receptor blockers (ARBs)
Sympatholytics
Alpha-beta blockers
Combinations:
Thiazide diuretic/potassium sparing diuretic
Beta-blocker/diuretic
ACE-inhibitor/diuretic
ARB/diuretic
Dihydropyridine CCB/ACE-inhibitor
Non-dihydropyridine CCB/ACE-inhibitor
The investigator may select among the available ACCORD antihypertensive

medications for initiation of therapy. Other drugs not supplied by the trial may be used as
the investigator determines appropriate. However, all antihypertensive regimens should
include a drug class associated with reduced cardiovascular events in diabetic
participants: diuretic, beta-blocker, calcium channel blocker, or ACE inhibitor. Based on
currently completed trials, some experts believe that monotherapy with a calcium channel
blocker may be less desirable in a diabetic patient. If an alpha blocker is used, it should
be used in combination with at least one agent proven to reduce cardiovascular events in
diabetic hypertensive patients. For participants in the intensive BP group (Figure 3.4), a
combination of a diuretic and either an ACE inhibitor or a beta-blocker should be
initiated at randomization. Drug doses should be increased and/or additional
antihypertensive medications should be added at each visit in the intensive group until the
participant's goal has been reached.
For participants in the intensive blood pressure group, “Milepost Visits” will
occur at 4 month intervals for the first 2 years of follow-up and annually thereafter. If at
a Milepost Visit the SBP is not less than 120 mm Hg, then an antihypertensive drug from
a different class than what is being taken must be added, unless there are compelling
reasons to wait. Milepost Visits do not apply to the standard blood pressure group.
Medication doses may be decreased or medications changed whenever the

ACCORD therapist considers it clinically appropriate, such as when adverse effects occur
that are believed to be secondary to an antihypertensive medication. Rechallenge is
encouraged if a trial off the medication is not associated with resolution of the adverse
effect, or if the adverse experience was not serious and the agent is strongly indicated for
another condition (e.g., an ACE inhibitor in heart failure).

71
Most multi-drug regimens are more effective if a diuretic is included as one of the
agents. Regimens are more effective if the drugs combined have very different
mechanisms of action. For example, an ACE inhibitor or ARB will usually be more
effective when combined with a diuretic or calcium antagonist than with a beta-blocker.
Only a few specific combinations are to be avoided, such as a beta-blocker with the
calcium antagonists verapamil or diltiazem. However, a beta-blocker combines very
effectively and safely with a dihydropyridine calcium antagonist. It is expected that most
ACCORD participants in the intensive BP intervention group will require at least 2 and
up to 5 antihypertensive medications to achieve their BP goals. If a participant is not at
goal on 4 drugs, consultation with the Clinical Center Network is recommended.
3.4.e Initiation of Blood Pressure Therapy
It is recommended that the BP intervention begin at the first visit at the same time
glycemia treatment is initiated. Intensive group participants should be seen at least
monthly until at BP goal (< 120 mm Hg). Once a participant’s BP goal has been
achieved, the antihypertensive medication regimen may still be altered subsequently to
maintain BP near goal, to avoid excessive hypotension, or to alleviate or minimize
adverse effects.
3.4.f Achieving and Monitoring Blood Pressure Control
Figures 3.4 and 3.5 describe the treatment algorithms for the two blood pressure
treatment groups. The BP treatment protocol of ACCORD is designed to be flexible in
terms of choice and dose of drugs. For the intensive BP group, the algorithm is
structured for adding additional medications for those participants who are above their
BP goals at the Milepost Visits.
At the point of randomization, all participants in the intensive group of the

hypertension study will automatically be assigned a series of milepost dates. Milepost
dates will be assigned for the entire duration of the study. Between these designated
visits, the ACCORD therapist may adjust dose of medications within the recommended
dose range or add medications. However once a milepost date has been reached and the
participant remains above goal BP the therapist is required to add an additional class of
drug to the existing regimen.
Each clinic and individual participant will have their BP and drug status monitored
closely by the Coordinating Center and CCN. The clinical center will be notified before
the Milepost Visit that adding drug is required if BP is above goal at that visit. In
situations where adding a drug could in the opinion of the therapist be potentially harmful
to the participant then adding a drug at this visit can be waived, however the therapist
must justify this decision on a “Milepost Exception Form.” The number of Milepost
exception forms will be closely monitored in each ACCORD clinic and regular feedback
provided to the clinic for the degree of adherence to the drug protocol.

72
For intensive group participants, once the ACCORD participant has been
prescribed 5 drugs, if the BP remains above goal at subsequent milepost visits it will be
permitted to substitute a different class into the regimen instead of adding another drug,
or increasing the dose of drug.
Therefore, action is required at each milepost visit throughout the duration of the
study for those intensive group participants who remain above their initial goal pressure
of < 120 mm Hg.
Also, if the SBP is > 120 mm Hg at any regular clinic visit for a participant in the
intensive BP group, blood pressure medications must be added or titrated and the
participant seen monthly until the SBP goes below 120 mm Hg or until a clinical decision
is made that therapy should not be increased further. (See Figure 3.4). Again: if the visit
is a Milepost Visit, BP the therapist must add an additional class of drug to the existing
regimen.
For standard BP group participants, medication dose titration or addition of

another drug is indicated if SBP >160 mm Hg at a single visit or >140 mm Hg at two
successive visits. Down-titration (‘step-down’) of therapy in the standard group is
allowed at the discretion of the ACCORD therapist, after consultation with the
participant, if the SBP < 130 mm Hg at a single visit or < 135 mm Hg at two consecutive
visits. (See Figure 3.5).
The blood pressure treatment protocol in ACCORD is designed to treat SBP

intensively to a goal less than 120 mm Hg in the intensive arm and to <140 mm Hg in the
standard arm of the blood pressure study. Evidence from clinical trials, such as UKPDS
and HOT, supports a SBP goal of 140 mm Hg in persons with hypertension and diabetes.
It is not known if lowering SBP to the more intensive ACCORD goal of 120 mm Hg is
beneficial or even harmful in patients like those entered into the ACCORD trial
compared with the standard goal of 140 mm Hg. ACCORD is designed to provide
definitive evidence for the intensive control of SBP in Type 2 Diabetics. Based on this
rationale, step-down (a reduction of dose or number of antihypertensive drugs) can be
done with participants in the standard goal group. Step-down is allowed at the discretion
of the ACCORD therapist, after consultation with the participant, when the SBP has been
<135 mm Hg on two successive clinic visits or is < 130 mmHg at any single visit.
3.4.g Monitoring Potassium and Creatinine
If an ACE-inhibitor, AII receptor blocker (ARB), or a thiazide is started or if the

dose is significantly increased, the potassium and creatinine levels should be monitored.

73
Figure 3.4: Treatment Algorithm for Intensive Blood Pressure Group

(Goal: SBP < 120 mm Hg)
Start Here
Monitor as Designated
Through Follow-up
Is Yes
SBP<120 mm Hg Continue Therapy**
at this visit?
No
You must :
A) Add Therapy Not Already
in Use*** (ACEI*,
Is This a Yes thiazide,ß-blocker, CCB,
Milepost reserpine, or α-blocker)
Visit? AND
B) See participant monthly
until SBP<120 mm Hg†
No
You must:
A) Titrate or Add Therapy Not
Already in Use*** (ACEI*, thiazide,
ß-blocker, CCB, reserpine, or
α-blocker)
AND
B) See participant monthly until SBP<120
mm Hg†
* ARB can be considered as a substitute for participants who do not tolerate ACEI therapy
** Unless side effects warrant change in therapy
*** Consult with the Clinical Center Network before adding a fifth antihypertensive medication
† or until a clinical decision is made that therapy should not be increased further

74
Figure 3.5: Treatment Algorithm for Standard Blood Pressure Group

(Goal: SBP < 140 mm Hg)
Start Here
Monitor as Designated
Through Follow-up
No Is Yes
SBP<130 mm Hg
at this visit?
Is
SBP < 135 mm Hg Yes
Step-down
on 2 consecutive visits?
No
Titrate or Add Therapy

Is SBP > 160 mm Hg Not Already in Use***
at this visit or Yes
> 140 mm Hg on (ACEI*, thiazide,
2 consecutive visits? ß-blocker, CCB,
reserpine, or α-blocker)
No
Continue Therapy**
* ARB can be considered as a substitute for participants who do not tolerate ACEI therapy
** Unless side effects warrant change in therapy
*** Consult with the Clinical Center Network before adding a fifth antihypertensive medication

75
3.5 Lifestyle Recommendations and Background Therapy
The purpose of including lifestyle recommendations and background therapy in

ACCORD is two fold. First, it fosters high quality general diabetes care in all ACCORD
participants in accordance with current practice guidelines. Second, it minimizes bias by
increasing the likelihood that background therapies that alter the risk of cardiovascular
events and that are not being studied in ACCORD are utilized equally across all study
arms. The background therapy recommendations will be provided to the participants and
their physicians. Background therapy is considered part of usual recommended care for
diabetes and, as such, is not covered by research study costs. The delivery of these
background therapies will be left up to the participants’ own clinicians.
Lifestyle therapy, which includes medical nutrition therapy (MNT) and physical
activity, is somewhat different. The primary glycemia research question is not a question
exclusively of drug effects, but of the effects of the higher degree of glycemic control.
To achieve a difference in HbA1c between the intensive and standard arms requires both
lifestyle and medication interventions. Counseling in both MNT and physical activity is
expected to be delivered by ACCORD clinicians. Participants randomized to the
intensive glycemic control arm of the study will receive more intensive reinforcement of
lifestyle therapy by way of more frequent assessments and instruction in order to achieve
their treatment goals.
The Lifestyle and Background Therapy Working Group will coordinate the
provision of relevant participant educational materials to be made available for study-
wide use. These will include the topics of general diabetes care, medical nutrition
therapy, physical activity, smoking cessation, and anti-thrombotic therapy and will
complement educational materials related to the glycemia, blood pressure, and lipid
interventions that are part of the trial. Unlike most educational materials for diabetes, the
ACCORD materials will not include specific goals for glucose, HbA1c, blood pressure,
and lipids, as these will depend on randomized treatment assignment.
General diabetes education should be carried out by the ACCORD clinics in

accordance with national standards for diabetes self-management education programs
(ADA 1999). Achieving the lifestyle and background therapy goals will require a
coordinated team effort along with education and/or supplemental materials.
Periodic reviews for updating recommendations will be conducted.
3.5.a Medical Nutrition Therapy
Medical Nutrition Therapy (MNT) consists of weight control and dietary

modification. The American Diabetes Association (ADA) position statement on
“Nutrition Recommendations and Principles for People with Diabetes Mellitus” reports
that “medical nutrition therapy is integral to total diabetes care and an essential
component of successful diabetes management” (ADA 2000a). Physical activity is a

76
closely related component of diabetes care. MNT is considered integral to the current
study for achieving optimal diabetes management.
The overall goal of MNT is to assist individuals with diabetes in making changes
in nutrition habits and body weight leading to improved metabolic control. There is no
one proven strategy or method that can be universally implemented, but
recommendations regarding weight control and nutritionally adequate meal plans can be
made to foster progress toward the goal.
The following recommendations and general principles will apply to all

participants in ACCORD, regardless of randomized group assignment. However, as
noted above, participants randomized to the intensive glycemic control arm of the study
will receive more intensive reinforcement of MNT.
3.5.a.1 Weight Control
Participants who are considered overweight (BMI > 25 kg/m2 according to the
NHLBI Obesity Education Initiative) are advised and encouraged to lose 10% of their
current weight, or 5-9 kg, whichever is less, over a 6-month period. A moderate caloric
restriction (250-500 calories less than the average daily intake calculated from a food
history) and a nutritionally adequate meal plan should be encouraged. Non-overweight
participants are encouraged to maintain weight. These guidelines are based on the
NHLBI Obesity Education Initiative (Obesity Education Initiative Expert Panel 1998).
3.5.a.2 Dietary Modification
Dietary modifications are recommended based both on glycemic control and also
on control and prevention of CVD risk factors common in people with diabetes. For
individuals using insulin therapy, it is recommended that meals be eaten at consistent
times synchronized with the time-action of the insulin preparation used (ADA 2000a).
The National Cholesterol Education Program (NCEP) Step I diet is recommended

(National Cholesterol Education Program 1993). This includes limiting fat intake to
<30% of total calories with saturated fat restricted to <10% of total calories.
Polyunsaturated fat intake should be < 10% of calories with monounsaturated fat
consumption 10-15% of calories. Cholesterol intake should be limited to < 300 mg/day.
If LDL-cholesterol is persistently elevated after a sufficient trial and compliance with the
Step I diet, the Step II diet should be prescribed. This diet calls for further reduction in
saturated fat and cholesterol (NCEP 1993). Some authorities recommend minimizing
transfatty acid intake as well because they increase LDL-cholesterol, decrease HDL-
cholesterol and are associated in cohort studies with increased CVD incidence (Ascherio
1997; Ascherio 1999; Mann 1997).
The percent of calories from carbohydrate intake will vary and is individualized
based on the participants’ eating habits, glucose, and lipid goals. First priority should be

77
given to the total amount of carbohydrate consumed rather than the source of the
carbohydrate (ADA 2000).
A reduced sodium intake (2.4-3.0 g/day) is recommended for all people with
diabetes (ADA 2000a). For people with hypertension and nephropathy, less than 2
gms/day is recommended by the ADA (ADA 2000a). The National High Blood Pressure
Education Program recommends less than 2.4 g/day of sodium for those with mild to
moderate hypertension (Joint National Committee on Prevention, Detection, Evaluation,
and Treatment of High Blood Pressure 1997) and to prevent hypertension (National High
Blood Pressure Education Program 1993). Therefore, ACCORD will recommend that
dietary sodium intake be reduced to less than 2.4 gms/day for the ACCORD study
population and less than 2 g/day for those with nephropathy.
It is recommended that trial participants limit daily alcohol intake to no more than
1 ounce (30ml) of ethanol for men and 0.5 ounces (15 ml) for women (ADA 2000a).
One ounce of ethanol is equivalent to 24 ounces (720 ml) of beer, 10 ounces (300 ml) of
wine or 2 ounces (60ml) of 100-proof whiskey. These recommendations are consistent
with JNC VI (Joint National Committee on Prevention, Detection, Evaluation, and
Treatment of High Blood Pressure 1997) and the U.S. Dietary Guidelines (USDA &
USDHHS 1995).
3.5.b Physical Activity
Physical activity recommendations for ACCORD participants are made in

accordance with current national recommendations for physical activity in the general
population (Pate 1995; USDHHS 1996) and for physical activity in diabetic patients
(ACSM 1994; ACSM & ADA 1997; ADA 2000). Regular moderate-to-vigorous aerobic
physical activity can improve metabolic control in people with diabetes and help with
weight loss and weight control.
Participants will be encouraged to accumulate 30 minutes or more of moderate-

intensity aerobic physical activity on 5 or more days of the week. Moderate-intensity
aerobic activity is defined as repetitive motion using large muscle groups that increases
the heart rate to 50-70% of maximal, is perceived as fairly light to somewhat hard, or is
equivalent in perceived intensity to brisk walking (3-4 miles per hour for most people, or
walking as if you are in a hurry). Maximal heart rate can be estimated by subtracting age
from 220. Persons on beta-blockers cannot use the heart rate criterion, as beta-blockers
attenuate the increase in heart rate, so perceived exertion or comparable intensity to brisk
walking should be recommended. Thirty minutes of physical activity may be
accumulated in bouts of 8-10 minutes. Warm-up and cool-down activities will be
encouraged, as will foot protection and inspection and maintenance of hydration (ADA,
2000b).
Sedentary ACCORD participants will be encouraged to increase their physical

activity levels gradually, starting with lower-intensity, shorter-duration, and less frequent
activities (eg., moderately paced walking for 5 minutes twice a week) and increasing

78
gradually over weeks or months to moderate-intensity and longer-duration activities (eg.,

brisk walking for 20-30 minutes, 5 days a week). The CCNs and Clinical Sites are
encouraged to develop and maintain lists of low cost or free local resources for safe
physical activity to provide to patients.
To prevent adverse events, patients with proliferative retinopathy, nephropathy, or

peripheral neuropathy with loss of protective sensation should be advised to avoid
vigorous or strenuous exercise, high-impact exercise (eg., jogging, high-impact aerobics,
racquet sports, competitive sports), weight training, and for those with peripheral
neuropathy, prolonged walking (ACSM & ADA 1997). Recommended exercises include
brisk walking (3-4 miles per hour), swimming, stationary cycling, and rowing.
In accordance with national recommendations for people with diabetes and at high
risk for underlying cardiovascular disease (ACSM 1994; Mahler 1995; Pate 1995; ACSM
& ADA 1997; ADA 2000b), screening should be considered for ACCORD participants
beginning an unsupervised exercise program or increasing their intensity of physical
activity. The recommended screening is an exercise stress test, or documentation of an
exercise stress test within the previous 3 months, that is negative for ischemia and
significant arrhythmias at a workload of 4-5 METS (i.e., moderate intensity, equivalent to
brisk walking). Persons continuing their current regular physical activity or increasing
duration of activity at the same intensity do not need this screening. Persons
experiencing symptoms of ischemia during physical activity should undergo diagnostic
evaluation.
3.5.c Smoking Cessation
There are consistent results from both cross-sectional and prospective studies
showing enhanced risk for micro- and macrovascular disease, as well as premature
mortality, from the combination of smoking and diabetes. The smoking cessation
literature is extensive, generally well-designed, and encouraging regarding the impact of
cost-effective practical office-based interventions. System-based approaches that make
smoking cessation intervention a routine part of office contacts and provide multiple
prompts, advice, assistance, and follow-up support are particularly effective. Although
there is minimal information on the effectiveness of cessation interventions specifically
for people with diabetes, there is no reason to assume that cessation intervention would
be less effective in this population. The following recommendations are based on
guidelines of the American Diabetes Association (ADA 2000c; Haire-Joshu 1999) and
Agency for Health Care Policy and Research (Fiore 1996).
All participants who are tobacco users will be strongly encouraged to stop. The
widely accepted AHCPR guidelines and model include the following steps:
1. ASK – each participant about tobacco use

2. ADVISE – current smokers to quit using a brief, unambiguous, strong and
personalized message
3. ASSESS – current smokers’ willingness to quit

79
4. ASSIST – current smokers who express willingness to make a quit attempt by

developing a quit plan, encouraging adjunctive pharmacotherapy, and providing
supplementary materials
5. ARRANGE follow-up, either by a health care provider or in a specialized smoking
cessation program
3.5.d Antithrombotic Therapy
Large-scale collaborative trials and meta-analyses of trials support the view that
low-dose aspirin lowers the rate of recurrent cardiovascular events in men and women
with diabetes and cardiovascular disease (Antiplatelet Trialists’ Collaboration 1994;
ETDRS Investigators 1992; Johnson 1999). Substantial evidence suggests that low-dose
aspirin therapy should also be used as a primary prevention strategy in men and women
with diabetes who are at high risk for cardiovascular events (Steering Committee of the
Physicians’ Health Study Research Group 1989; ETDRS Investigators 1992). Based on
these studies, the American Diabetes Association also recommends low-dose aspirin as
secondary prevention, and as primary prevention in high-risk men and women over the
age of 30 with diabetes (ADA 2000d; Colwell 1997). All of the ACCORD participants
will fall into one of these two categories. Aspirin is safe and effective across a dosage
range from 75-325 mg daily (Antiplatelet Trialists’ Collaboration 1994; Johnson 1999).
Therefore, aspirin 75-325 mg daily is recommended for all ACCORD participants unless
contraindicated by allergy, bleeding disorder, recent gastrointestinal bleeding or need for
anticoagulant therapy.
3.5.e Treatment of Hypertension and Dyslipidemia
Because ACCORD includes a full factorial trial for both the blood pressure and
lipid components, there will be participants enrolled in the lipid component who may
have hypertension and participants in the blood pressure component who have
dyslipidemia. These participants will receive care for their conditions, if present, from
their usual source of medical care.
Blood pressure and lipid goals for treatment are recommended in these patients
based on the investigators' synthesis of clinical trial evidence. In some cases, these may
differ from national recommendations from consensus panels. It is suggested that the
patient’s physician take all these guidelines under consideration when individualizing
treatment for each patient.
A blood pressure goal of <140/85 mm Hg is recommended for participants not in

the BP intervention. This goal is supported by evidence for cardiovascular disease
prevention. National guidelines that recommend a lower goal in people with diabetes of
<130/85 mm Hg, including JNC VI (Joint National Committee on Prevention, Detection,
Evaluation, and Treatment of High Blood Pressure 1997) and the American Diabetes
Association (ADA 2000j) will also be provided.

80
For participants not in the lipid intervention, an LDL-cholesterol goal of <100

mg/dl is recommended. Information on national guidelines will also be provided that
recommend behavioral and/or pharmacologic treatment for LDL-cholesterol levels >100
mg/dl in people with diabetes whether or not CHD is present (ADA 2000e, National
Cholesterol Education Program 2002).
As current guidelines are revised, the most current recommendations will be

conveyed to the participants’ physicians.
3.5.f Angiotensin-Converting-Enzyme Inhibitors (ACE-Inhibitors)
Evidence for the effectiveness of angiotensin-converting-enzyme (ACE)

inhibition in reducing adverse outcomes in patients with type 2 diabetes has been
increasing. Meta-analyses and randomized clinical trials have shown that ACE inhibitors
reduce mortality in acute MI, with greater effect sizes in diabetic patients (ACE Inhibitor
Myocardial Infarction Collaborative Group 1998); (Zuanetti 1997), and reduce sudden
cardiac death following acute MI (Domanski, et al., 1999). Meta-analysis has shown that
mortality and subsequent hospitalizations for CHF are reduced in patients with
congestive heart failure treated with ACE-inhibitors, including patients in all four NYHA
CHF classes if their ejection fractions is < 35% (Garg 1995).
Some randomized controlled trials have suggested that ACE-inhibitors are

superior to other treatments for hypertension in hypertensive type 2 diabetic patients
(Estacio 1998; Tatti 1998; Hansson 1999a) Other studies have found ACE-inhibitors to
be equivalent to other treatments (Hansson 1999b; UKPDS 1998). Given these
inconsistent results, there is no clear consensus at the present time to use ACE-inhibitors
over other antihypertensive treatments. However, the Heart Outcomes Prevention
Evaluation Study (HOPE) compared ACE-inhibitors with placebo in participants at high
risk for cardiovascular disease and found a 25% reduction in the combined outcome of
MI, stroke, and cardiovascular death, which did not appear to be explained by the degree
of blood pressure reduction (Yusuf 2000). The effect was significant in participants with
and without diabetes, participants with and without a past history of coronary artery
disease, and participants with and without microalbuminuria (Yusuf 2000). In the
subgroup of patients with diabetes, there was a reduction in relative risk in patients with
and without microalbuminuria and in patients with and without hypertension, although
effects in some of the subgroups did not reach significance (HOPE 2000).
There is strong evidence that ACE-inhibitors improve renal outcomes

(nephropathy and albumin excretion) in type 2 diabetes when compared with placebo in
hypertensive participants as well as non-hypertensive participants with and without
microalbuminuria (Ravid 1996; Ravid 1998; Yusuf 2000).
Based on this evidence, the ACCORD study will recommend the use of ACE-
inhibitors for reducing cardiovascular morbidity and mortality in patients who have
experienced acute MI, congestive heart failure, nephropathy, and in patients with type 2
diabetes with at least one additional risk factor for cardiovascular disease.

81
3.5.g Diabetes Related General Medical Care
ACCORD participants will also receive diabetes related general medical care.
The following recommendations are based on guidelines of the American Diabetes
Association (ADA 2000f ).
1. All participants should receive an annual dilated eye and visual exam by an
ophthalmologist or optometrist (ADA 2000g).
2. All participants should receive a foot examination at least annually to assess skin
integrity, foot structure and biomechanics, vascular status, and protective
sensation (ADA 2000h, Mayfield 1998). A standardized examination will be
described in the Manual of Procedures.
3. Patients with diabetes (in particular those with end organ complications of cardiac
and renal disease) are at high risk for cardiopulmonary complications,
hospitalization, and death from influenza and pneumococcal disease. Although
there are few clinical trials of influenza and pneumococcal vaccine efficacy
specifically in patients with diabetes, subgroup analyses of patients with diabetes
reported in clinical narrative and case-control studies support the fact that
vaccination against influenza has been effective in reducing hospital admissions
during influenza epidemics (Smith 2000; Nichol 1994). Therefore, all
participants should receive annual influenza vaccine and, if previously
unvaccinated, one dose of pneumococcal vaccine (ADA 2000i).
3.5.h Monitoring Lifestyle Recommendations and Background Therapy
Use of aspirin therapy, ACE-inhibitors, and smoking status will be documented

on regular visit forms with the same frequency for all study participants. Weight will be
measured at all clinic visits. A foot examination will be conducted as part of the baseline
and annual study physical examinations. Visual acuity will be measured at baseline,
every other year thereafter, and at the end of the study. Diet and physical activity will be
measured at baseline, and at years 1, 3, and 4.

82
Chapter 4
Participant Safety and Confidentiality
4.1 Introduction
Assuring participant safety and the confidentiality of participant data are essential
components of ACCORD. Each participating investigator has primary responsibility for the
safety of the individual participants under his/her care, while the Data and Safety Monitoring
Board will have primary responsibility for monitoring the accumulating study data for signs of
adverse trends in morbidity/mortality and drug toxicity.
4.2 Exclusions
Persons with contraindications to the study statin or fibrate therapy drug will not be
eligible to be enrolled in the lipid component. Exclusions are detailed in Sections 2.1.b (General
Exclusions), 2.1.c (Lipid Component Exclusions), and 2.1.d (Blood Pressure Component
Exclusions).
4.3 Adverse Reactions and Discontinuation of Study Drug
Given the number of drugs employed in ACCORD, adverse reactions could be caused by
single drugs or by drug-drug interactions. Recently, among the thiazolidinedione class of drugs
for type 2 diabetes mellitus, troglitazone was removed from the U.S. market based upon FDA's
review of liver toxicity data, which suggested that it is more toxic than two newer agents from
the same class, rosiglitazone (Avandia) and pioglitazone (Actos) (FDA 2000). Myopathy and
rhabdomyolysis have been reported after combination therapy with statins and fibrates (Ellen
1998, Pierce 1990).
Possible adverse effects of the study drugs will be assessed at each follow-up visit by
patient history, including hypoglycemia episodes and, in the lipid component, muscle pain.
Chemistry tests will be performed periodically to monitor safety issues, as indicated in Tables
2.2A through 2.2F. Patients initiated on thiazolidinediones will be monitored, as recommended
by the manufacturer, with ALT levels every two months for the first 12 months, and annually
thereafter. Patients treated with ACE inhibitors, AII receptor blockers, or diuretics will be
monitored for hypo- or hyperkalemia and for renal dysfunction. Patients in the lipid portion of
the trial will be monitored periodically with CPK enzymes and creatinine levels. Serious and
unexpected reactions to study drugs will be reported to the FDA. Hypoglycemia episodes
requiring assistance by medical/paramedical personnel will be reported to the Coordinating
Center as Serious Adverse Experiences (SAEs) using study-specific forms.
ACCORD will monitor the frequency and severity of muscle symptoms at every clinic
visit. For participants in the lipid trial, CPK concentrations will be measured at baseline, 1
month, 4 months, 8 months and 12 months and as needed for moderate to severe unexplained

83
muscle symptoms. In addition to the FDA definition of myopathy, we will monitor by treatment
group the frequency and severity of reported muscle symptoms and the frequency of CPK
elevations with and without symptoms. These procedures will provide valid and important data
regarding the occurrence and severity of myopathy. To insure participant safety, patients will be
withdrawn from lipid-lowering therapy for CPK in excess of 10 times the upper limit of normal
in the absence of symptoms or for CPK in excess of 5 times the upper limit of normal in the
presence of symptoms. In addition, reductase inhibitor therapy will be avoided in participants
who developed myositis while taking the fibrate placebo. This procedure should ensure that no
one who was assigned to the fibrate placebo is rechallenged with a reductase inhibitor (the only
active agent for the participant).
As described in Section 3.3.b, participants in the lipid trial will have serum creatinine
measured at baseline and at least every four months thereafter. The starting dose of masked
fenofibrate/placebo medication will be determined by the calculated glomerular filtration rate
(GFR) using the baseline serum creatinine level and the abbreviated MDRD equation (Levey
2003). Those participants with a baseline GFR >50 ml/min/1.73m2 will begin at a starting dose
of 160 mg of fenofibrate or identical placebo tablet. Those with a calculated GFR between 30
and <50 will start at the reduced dose of 54mg/day fenofibrate or placebo. If the participant had
started on the 160 mg dose of the masked medication, this dose will be down-titrated if the
participant’s estimated follow-up GFR falls between 30 and <50 mL/min/1.73m2 on two
consecutive measurements taken four months apart. Participants with GFRs in this range will
receive either 54 mg/day of fenofibrate or matching placebo. If the estimated GFR falls below
30 mL/min/1.73m2 at any time, the Coordinating Center will notify the clinic site that a
confirmatory blood draw for repeat estimated GFR will be required within 2 weeks. If the
confirmatory estimated GFR is below 30 mL/min/1.73m2, the masked study medication will be
permanently discontinued, regardless of fenofibrate or placebo assignment.
If necessary, the study physician may at his/her discretion reduce or stop administration
of any study drug. Depending on the situation, the change may be temporary or permanent.
Situations that may require temporary reduction or elimination of a study medication include:
worsening congestive heart failure, acute myocardial infarction, severe hypoglycemia episodes,
and other illnesses. Events that may require permanent cessation of a study drug include:
jaundice, myopathy, other adverse drug reaction, need for active therapy with closely related
compounds, cardiac transplantation, repeated severe hypoglycemia episodes, other conditions,
and participant request.
4.4 Unmasking Procedure for the Fibrate/Placebo Intervention
In some special circumstances (e.g., a medical emergency), a patient’s assigned

fibrate/placebo treatment group may need to be revealed. If the request to unmask comes from a
physician unfamiliar with the patient (e.g., emergency medicine department), the caller will be
referred to that participant’s ACCORD physician or site coordinator to discuss the relevant
medical issues. If appropriate, an effort will be made to maintain the blinding of the patient and
the ACCORD staff.

84
The ACCORD Drug Distribution Center (DDC) will be contacted for the purpose of
unmasking a participant’s therapy. The telephone number is (505) 248-3203 and it is available
24 hours a day (answering service after hours). The caller should state that he/she is calling in
reference to the ACCORD trial. The appropriate DDC personnel will be contacted to respond to
the call. They will record the participant’s ACCORD I.D. number, name of physician requesting
code break and reason the unmasking is necessary.
The DDC must have either the participant’s I.D. number or fibrate/placebo bottle number
in order to access the patient’s therapy (active drug or placebo). Any unmasking done by the
DDC will be reported immediately to the Clinical Site investigator or coordinator for inclusion in
the patient’s ACCORD medical record. The CCN PI, Project Office and Coordinating Center
will also be notified of any unmasking that occurred, although the clinic staff should remain
masked if possible.
4.5 The Elements of Informed Consent
Consent to participate in a research study includes the elements listed below. The
ACCORD consent process will include all elements. (A model informed consent document is in
Appendix I).
• Participants must be advised that the study involves research. Staff must explain the
purposes of the research, the expected duration of participation, and a description of
the procedures to be followed, including identification of experimental procedures.
• Anticipated benefits of the trial must be explained to the participant.
• Attendant discomforts and risks “reasonably to be expected” must be described.
• Appropriate alternative procedures that might be advantageous for the participant

must be disclosed.
• The extent, if any, to which confidentiality of records identifying the participant will
be maintained must be described.
• Prospective participants must be advised of the availability or non-availability of

medical treatment or compensation for physical injuries incurred as a result of
participation in the study, and if available, what they consist of, or where further
information can be obtained.
• Persons responsible for the study must explain whom a participant can contact for
answers to pertinent questions about the research and his or her rights, and whom to
contact in the event of a research-related injury.
• Participants must be told that participation is voluntary, refusal to participate will

involve no penalty or loss of benefits to which the participant is otherwise entitled,

85
and the participant may discontinue participation at any time without penalty or loss
of benefits to which he or she is otherwise entitled.
4.6 Confidentiality
The confidentiality of all participant information (including but not limited to any genetic
analysis) must be protected at the Clinical Sites, the Clinical Center Networks, and the
Coordinating Center. Paper records and computer files must be appropriately safeguarded from
unauthorized access.
Paper records for study participants will be stored at the Clinical Sites. Copies of signed
informed consents and records pertaining to SAEs and study-defined clinical events, including
necessary medical records, will be stored at the Coordinating Center. These records will receive
the same care as would ordinary medical records. They will be stored in locked filing cabinets
and/or filing rooms within secure office space. Only study personnel who have completed
ACCORD training in data handling will have access to study forms.
Similar care will be used in the handling of the computer records of study data stored at
each Clinical Site. Access to the data in the local ACCORD database will be controlled by a
system of user identification names and passwords. Each Clinical Site staff member must
complete the ACCORD data handling training program before being given an ID and password
to use the data system. The privileges allowed to each ID can be individually specified by the
local Clinic Coordinator. All passwords stored within the system will be encrypted using Secure
Socket Layer (SSL) encryption.
Confidentiality of information within the Coordinating Center will be protected through a

variety of procedures and facilities:
1. The confidential nature of the data collected, processed, and stored at the Coordinating
Center is explained to all new personnel, who must sign a confidentiality certification
after discussion with their supervisor.
2. All access to Coordinating Center office space containing data is controlled through a
single door, which is locked with a keypunch lock. This door remains locked at all
times.
3. All participant data sent to the Coordinating Center is encrypted as described above.
4. All participant data stored on the Wake Forest University’s mainframe computers are
likewise encrypted. In addition, all such databases are protected by passwords that
must be supplied before the data can be accessed. Passwords are released only to
Coordinating Center staff with a need to use the particular file, and are changed on a
regular schedule.

86
5. All printouts, plots, and reports containing individually identifiable data are produced
on printers and plotters within the Coordinating Center’s secure office space. All such
reports are kept in locked storage cabinets within the Coordinating Center.
6. No participant identifiers will be present on any data for files transmitted to the Data
and Safety Monitoring Board or the sponsor.
4.7 Vanguard Participants
The Vanguard Phase participants will be asked to provide consent to be treated and
followed according to this revised main trial protocol.

87
Chapter 5
Clinical Outcome Measures
5.0 Outcomes
This chapter describes the components of the ACCORD primary and secondary clinical
outcomes. The cardiovascular disease (CVD) events occurring during follow-up will be
classified by a Working Group of the Morbidity and Mortality subcommittee.
5.1 Primary (Macrovascular) Outcome
The primary endpoint for ACCORD is the composite outcome of death from
cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. Cardiovascular deaths
are defined in Section 5.1.a, myocardial infarctions are defined in Section 5.1.b, and strokes are
defined in Section 5.1.c
5.1.a Cardiovascular Death
5.1.a.1 Unexpected death: Unexpected death presumed to be due to ischemic cardiovascular

disease, occurring within 24 hours of the onset of symptoms without confirmation of
cardiovascular disease, and without clinical or post mortem evidence of other etiology.
5.1.a.2 Fatal Myocardial infarction (MI): death within 7 days of the onset of documented MI (see
5.1.b).
5.1.a.3 Congestive heart failure (CHF): death due to clinical, radiological or postmortem
evidence of CHF without clinical or postmortem evidence of an acute ischemic event
(cardiogenic shock to be included).
5.1.a.4 Death after invasive cardiovascular interventions: death associated with the intervention,
i.e., within 30 days of cardiovascular surgery, or within 7 days of cardiac catheterization,
arrhythmia ablation, angioplasty, atherectomy, stent deployment, or other invasive
coronary or peripheral vascular intervention.
5.1.a.5 Documented arrhythmia: death due to bradyarrhythmias or tachyarrhythmias not

associated with an acute cardiac ischemic event.
5.1.a.6 Death following non-cardiovascular surgery: death due to cardiovascular causes as

defined in 5.1.a.1-5.1.a.5, 5.1.a.7-5.1.a.8 within 30 days of surgery.
5.1.a.7 Stroke: death due to stroke occurring within 7 days of the signs and symptoms of a stroke
(see 5.1.c).

88
5.1.a.8 Other cardiovascular diseases: death due to other vascular diseases including pulmonary
emboli and abdominal aortic aneurysm rupture.
5.1.a.9 Presumed cardiovascular death: Suspicion of cardiovascular death with supporting

clinical evidence that may not fulfill criteria otherwise stated. Example: Patient admitted
with typical chest pain of 3 hours duration and treated as an MI, but without ECG and
enzymatic documentation to meet usual criteria.
5.1.b Myocardial Infarction
The definitions for MI are presented below. If necessary for a definition, prolonged
ischemic symptoms must last 20 minutes, and the cardiac enzymes of interest are Troponin T or
I and/or serum CK-MB mass. Silent MIs will be identified by the ACCORD ECG Reading
Center.
5.1.b.1 Q-wave MI: Diagnosis based on the occurrence of a compatible clinical syndrome with
prolonged ischemic symptoms, associated with the development of new significant Q
waves (defined in the ECG Reading Center Manual of Procedures). Diagnostic
elevation of cardiac enzymes will include: increase in CK-MB mass to a level > twice
the upper limit of normal, and/or and increase in Troponin T or I to a level that
indicates myonecrosis in the laboratory performing the study.
5.1.b.2 Non Q-wave MI: Diagnosis based on the occurrence of a compatible clinical
syndrome with prolonged ischemic symptoms, associated with elevation of serum
enzymes, as for Q-wave MI. Only in the case that both Troponin and CK-MB mass
measurements are not available, would the elevation of total CK to > twice the upper
limit of normal qualify for diagnosis.
5.1.b.3 Silent (unrecognized) MI: development of new significant Q waves without other
evidence of myocardial infarction (the date of event will be assigned halfway between
the date of discovery and last normal ECG).
5.1.b.4 Probable non Q-wave MI: Diagnosis based on the occurrence of a compatible clinical
syndrome with prolonged ischemic symptoms, without documentation of cardiac
enzyme elevation, but associated with the development of new and persistent
significant ST-T changes (>24 hr in duration). (Changes are defined in the ECG
Reading Center Manual of Procedures).
5.1.b.5 MI after cardiovascular invasive interventions Diagnosis based upon the occurrence of
CK-MB (or Troponin) elevations to a level increased 3-5 times normal for the
laboratory performing the studies, occurring within 7 days of cardiac catheterization,
arrhythmia ablation, angioplasty, atherectomy, stent deployment or other invasive
coronary, carotid or peripheral vascular intervention.

89
5.1.b.6 MI after coronary bypass graft surgery: Diagnosis based upon the occurrence of CK-
MB (or Troponin) elevations to a level increased > 5-10 times normal for the laboratory
performing the studies, occurring within 30 days of cardiac surgery.
5.1.b.7 MI after non-cardiovascular surgery: MI (as defined above, occurring within 30 days of
non-cardiovascular surgery.
5.1.c Stroke
5.1.c.1 Definite ischemic stroke: CT or MRI scan within 14 days of onset of a focal neurological
deficit lasting more than 24 hours with evidence of brain infarction (mottled cerebral
pattern or decreased density in a compatible location), no intraparenchymal hemorrhage
by CT/MRI, no significant blood in the subarachnoid space by CT/MRI or by lumbar
puncture, or autopsy confirmation. A nonvascular etiology must be absent.
5.1.c.2 Definite primary intracerebral hemorrhage: Focal neurological deficit lasting more than
24 hours. Confirmation of intraparenchymal hemorrhage in a compatible location with
CT/MRI scan within 14 days of the deficit onset, or at autopsy, or by lumbar puncture.
5.1.c.3 Subarachnoid hemorrhage: Sudden onset of a headache, neck stiffness, loss of

consciousness. There may be a focal neurological deficit, but neck stiffness is more
prominent. Blood in the subarachnoid space by CT/MRI or lumbar puncture or
intraventricular by CT/MRI.
5.1.c.4 Stroke of unknown type etiology: Definite stroke of unknown etiology when CT, MRI, or
autopsy are not done. Information is inadequate to diagnose ischemic (infarction),
intracerebral hemorrhage, or subarachnoid hemorrhage.
5.1.c.5 Non-fatal stroke after cardiovascular invasive interventions: stroke (as defined in 5.1.c.1-
5.1.c.4) associated to the intervention within 30 days of cardiovascular surgery, or within
7 days of cardiac catheterization, arrhythmia ablation, angioplasty, atherectomy, stent
deployment or other invasive coronary or peripheral vascular interventions.
5.1.c.6 Non-fatal stroke post non-cardiovascular surgery: stroke (as defined in 5.1.c.1-5.1.c.4)
occurring within 30 days of non-cardiovascular surgery.
5.2 Secondary Outcomes
The secondary endpoints for ACCORD are as follows. (See Section 7.1.b for the
intervention-specific secondary hypotheses.)
• An expanded macrovascular outcome, specifically the combination of the primary

endpoint plus any revascularization (defined in Section 5.3.a) plus hospitalization for
congestive heart failure (defined in Section 5.3.b)
• Total mortality

90
• Cardiovascular mortality
• Major coronary heart disease event, specifically fatal events (defined in Section
5.1.a.1 through 5.1.a.6 and 5.1.a.8 through 5.1.a.9), nonfatal myocardial infarction
(defined in Section 5.1.b), and unstable angina (defined in Section 5.3.c).
• Total stroke, specifically fatal strokes (defined in Section 5.1.a.7) and nonfatal strokes
(defined in Section 5.1.c).
• Congestive Heart Failure Death (defined in Section 5.1.a.3 ) or Hospitalization for
Congestive Heart Failure (with documented clinical and radiological evidence)
• Health-related quality of life (see Chapter 6)
• Cost-effectiveness (see Chapter 6)
• The main microvascular outcome of the ACCORD study is the primary outcome of
the ACCORD Eye Substudy, namely: “the combined outcome of progression of
diabetic retinopathy of at least 3 stages on the ETDRS scale, photocoagulation, or
vitrectomy for diabetic retinopathy”. This substudy will only take place in
approximately half of the ACCORD study population.
• A second composite microvascular endpoint will be examined in the entire ACCORD
population: fatal or non-fatal renal failure (as defined in 5.4.a.3) or retinal
photocoagulation or vitrectomy for diabetic retinopathy. This endpoint essentially
replicates the composite microvascular endpoint in the UKPDS.
5.3 Other ACCORD Outcomes
5.3.a All cardiovascular revascularization procedures, including:

• PTCA (balloon)
• PTCA with stent
• CABG
• Carotid angioplasty with stent
• Carotid endarterectomy
• Peripheral angioplasty with or without stent
• Peripheral vascular surgery (including aortic aneurysm repair)
• Limb amputation: including partial or digit amputation due to vascular disease.
5.3.b Unstable angina: new onset exertional angina, accelerated or rest angina, or both, and at
least 1 of the following (Downs 1998):
a) at least 1-mm ST segment deviation and reversible defect on stress perfusion study, or
b) angiographic findings of at least 90% epicardial coronary artery or at least 50%
stenosis in the left main coronary artery, or
c) at least 1-mm ST segment deviation with pain on ECG stress testing and/or rest ECG
and evidence of at least 50% stenosis in a major epicardial coronary artery.
5.3.c Total cancer mortality, including:

• Primary site of cancer is gastrointestinal
• Primary site of cancer is lung.
• Primary site of cancer is breast
• Primary site of cancer is prostate.

91
• Primary site of cancer is brain.

• Primary site of cancer is ‘other’
• Primary site of cancer is multi site.
• Primary site of cancer is genito-urinary.
5.4 Microvascular Outcomes
5.4.a Development of nephropathy

5.4.a.1 Doubling of serum creatinine or a 20 ml/min/1.73m2 decrease in estimated
glomerular filtration (GFR) as estimated by the MDRD equation (GFR = 186 x
(serum creatinine in mg/dl)-1.154 x (age in years)-0.203 x 0.742 in females x 1.210 in
African-Americans
5.4.a.2 Development of macroalbuminuria (albumin/creatinine ratio > 300 mg albumin
per gram creatinine in random urine sample)
5.4.a.3 Development of renal failure as defined by renal transplantation or initiation of
dialysis or a rise in serum creatinine > 3.3 mg/dl in the absence of an acute
reversible cause
5.4.a.4 Composite nephropathy outcome (any of the three above-named outcomes)
5.4.a.5 Development of microalbuminuria (albumin/creatinine ratio > 30 mg albumin per
gram creatinine in a random urine sample)
5.4.b Development of diabetic eye complication

5.4.b.1 Use of retinal photocoagulation or vitrectomy for diabetic retinopathy
5.4.b.2 Cataract extraction
5.4.b.3 Three-line change in visual acuity using Log MAR visual acuity chart
5.4.b.4 Severe vision loss due to diabetes (<5/200)
5.4.c Development of diabetic neuropathy

5.4.c.1 Scoring > 2.5 on the Michigan Neuropathy Screening Instrument (this is a
composite neuropathy endpoint employed in the DCCT study)
5.4.c.2 Loss of vibratory sensation (128 Hz tuning fork)
5.4.c.3 Loss of ankle jerk during Jendrassic maneuver
5.4.c.4 Loss of pressure sensation (monofilament with 10 gram force)

92
Chapter 6
Health-Related Quality of Life/
Cost-Effectiveness Outcome Measures
6.1 Hypotheses
6.1.a. Health-Related Quality of Life (HRQL) Hypotheses
6.1.a.1 Primary HRQL Hypotheses (Main Effects Comparisons)
(1) Intensive control of blood glucose compared with less intensive control in patients with
diabetes will:
a) decrease symptoms and side effects as assessed by the Symptoms Distress In Diabetes
Questionnaire.
b) decrease symptoms and disability associated with cardiovascular events as assessed by
the SF-36v2 and the Health Utilities Index III.
c) improve self-reported treatment satisfaction (and consequently treatment adherence
and drop-outs) as assessed by the Diabetes Treatment Satisfaction Questionnaire
(DTSQ).
(2) Treatment of lipids with statins and fibrates (to lower LDL-C and triglycerides and raise
HDL-C) compared with treatment of lipids with statins alone (to lower LDL-C) will:
Questionnaire.
(DTSQ).
(3) Intensive control of blood pressure compared with less intensive control in patients with
diabetes will:
Questionnaire.
(DTSQ).
6.1.a.2 Secondary HRQL Hypotheses (Pair-wise “Reference Case” Comparisons)
(1) Intensive control of glucose and blood pressure, when compared to standard treatment for
glucose and blood pressure, will:

93
a) decrease symptoms and side effects as assessed by the Symptoms Distress In

Diabetes Questionnaire.
b) decrease symptoms and disability associated with cardiovascular events as
assessed by the SF-36v2 and the Health Utilities Index III.
(DTSQ).
(2) Intensive control of glucose and lipids (with fibrate) when compared to standard treatment
for glucose and placebo will
a) decrease symptoms and side effects as assessed by the Symptoms Distress In
Diabetes Questionnaire.
b) decrease symptoms and disability associated with cardiovascular events as
assessed by the SF-36v2 and the Health Utilities Index III.
(DTSQ).
6.1.a.3 Event-Related HRQL [Feeling Thermometer (FT)] Hypotheses
(1) Participants experiencing cardiovascular events in the preceding 4 months will show
greater declines in feeling thermometer (pre-post) ratings than those who do not experience
events.
(2) Participants experiencing >3 hypoglycemic episodes in the preceding 4 months will show
greater declines in feeling thermometer ratings than those without hypoglycemic episodes
(3) Participants in the intensive glucose control group will show more rapid recovery of feeling
thermometer ratings after cardiovascular events than those participants in the standard
glucose control group
6.1.b. Cost-Effectiveness Hypotheses
6.1.b.1 Primary Cost-Effectiveness Hypotheses (Main Effects Comparisons)
(1) The incremental cost effectiveness ratio of intensive control of blood glucose, in the
presence of lipid therapy or blood pressure control, will not be larger than the maximum
acceptable "ceiling" level of the cost per cardiovascular disease free year gained and cost
per quality adjusted life-year (QALY) gained when compared to less intensive control.
(2) The incremental cost effectiveness ratio of intensive lipid treatment (adding a fibrate to a
statin), in the presence of control of glucose, will not be larger than the maximum
acceptable "ceiling" level of the cost per cardiovascular disease free year gained and cost
per quality adjusted life-year (QALY) gained when compared to less intensive control.

94
(3) The incremental cost effectiveness ratio of intensive blood pressure control, in the presence
of control of glucose, will not be larger than the maximum acceptable "ceiling" level of the
cost per cardiovascular disease free year gained and cost per quality adjusted life-year
(QALY) gained when compared to less intensive control.
6.1.b.2 Secondary Cost-Effectiveness Hypotheses (Pair-wise “Reference Case” Comparisons)
(1) The incremental cost effectiveness ratio of intensive control of blood glucose and blood
pressure will not be larger than the maximum acceptable "ceiling" level of the cost per
cardiovascular disease free year gained and cost per quality adjusted life-year (QALY)
gained when compared to standard treatment for glucose and blood pressure.
(2) The incremental cost effectiveness ratio of intensive control of blood glucose and lipids
(with fibrate) will not be larger than the maximum acceptable "ceiling" level of the cost per
cardiovascular disease free year gained and cost per quality adjusted life-year (QALY)
gained when compared to standard treatment for glucose and placebo.
6.2 Health-Related Quality of Life (HRQL)
6.2.a Rationale
HRQL measurement is a key means of determining the value of ACCORD interventions

and outcomes from the patient’s point of view. While mortality reduction is a secondary outcome in
ACCORD, all other outcome measures concern nonfatal health outcomes. HRQL assessment
enables an assessment of the relative importance of these various outcomes (from amputation to
cognitive impairment) to the patients themselves. HRQL assessment provides an understanding of
the balance between the burdens and benefits of intensive glucose and lipid control, and intensive
glucose and blood pressure control from the patient’s point of view. HRQL information should also
provide valuable insight into adherence results and into the practicality of clinically implementing
these interventions after the trial is over. Assessment of HRQL in ACCORD is designed to study
treatment effects from the patient’s point of view concerning: 1) short term symptoms and 2)
longer-term rates of macrovascular and microvascular events.
The ACCORD HRQL instruments were selected based upon the following criteria: 1)
brevity, 2) inclusion of the major dimensions shown in the literature to be affected by diabetes and
its treatment, 3) proven responsiveness to treatment-related changes, and 4) appropriateness for the
age range in ACCORD and ethnic diversity of type 2 diabetic patients. For these purposes, specific
symptoms distress measures, generic health status, and depression measures were chosen.
6.2.b Health Related Quality of Life (HRQL)
The HRQL measures will be administered to a random sample of 250 participants in each
cell of the 8 ACCORD treatment groups (2000 participants in total) at 0, 12, 36 and 48 months.
This sample of 2000 participants will be nested within the larger random sample of 4288

95
participants participating within the cost effectiveness assessment substudy (described in Section
6.3).
General health status will be measured by the self-administered SF-36v2. This is the most
widely used general health status measure with extensive validation and population norms available.
It allows comparison of the ACCORD population with those of other studies and other chronic
diseases. Eight scale scores will be generated in the following domains: general health, physical
function, role-physical, role-emotional, vitality, social function, mental health, and pain. It can also
be scored in terms of physical health and mental health component scores. The SF-36v2 offers
expanded response options on the role function items, offering greater sensitivity in this area.
The Symptom Distress In Diabetes Questionnaire developed by Testa and

colleagues (1993,1994 Phase V Technologies) will be used. This self-administered instrument
has been shown to be responsive to improved glycemic control in a randomized, double-masked
placebo-controlled clinical trial of diet and either 5 or 20 mg of glipizide (Testa 1998). Based
upon existing data and item pools, a 60-item version of this instrument has been developed for
ACCORD by eliminating questions while maintaining internal consistency. Furthermore, a 7-
item diabetes-specific treatment satisfaction measure (validated in previous trials) will be used to
assess the impact of intensive treatment.
Due to its documented relation with cardiovascular events and glycemic control, clinical
depression will also be measured. The 9-item depression measure from the Patient Health
Questionnaire (PHQ) will be used. The PHQ is the self-report version of the PRIME-MD, a well-
validated psychiatric diagnostic interview for use in primary care settings. The PHQ depression
measure offers the briefest measure that provides diagnostic information, severity information,
and responsiveness to depression treatment.
Effect sizes on the Symptom Distress Measure ranged from 0.6 to 0.2 SD units in the
glipizide trial. There is evidence from previous treatment trials that hypoglycemic, lipid-
lowering, and anti-hypertensive drug effects and their reflection in patient-rated HRQL will
show non-additive properties. Therefore, it is of interest to compare individual cells receiving
different combinations of interventions, rather than just marginal effects of each intervention. To
allow us to address possible non-additive treatment effects, a sample of participants randomized
to each 2x2 trial will be assessed for HRQL. Sample size calculations used an ANOVA model
and assumed similar treatment effects within each trial (glycemic plus lipid or glycemic plus
blood pressure). To detect a 0.3 SD difference between the group that receives intensive
therapies and the group that receives no intensive therapies, an estimated 250 participants per
group would provide approximately 90% power. The total sample size would thus be 2000 of
the 10,000 randomized participants (=250 X 8 treatment groups).
6.2.c Health State Utility Measure
If cost-effectiveness analyses are going to include the patient’s perspective they must
assess the value or utility of the patient’s health state. Interviews such as the standard gamble or
time trade-off generate utilities from the study patients themselves. The use of either approach

96
will be burdensome in ACCORD. Therefore the most valid alternative is to use a measure with
previously derived population-based utility values.
The Health Utilities Index (HUI), is one such measure, and will be used to assess health
state utility. The HUI is a general HRQL measure which includes a health-status classification
system and a preference-based scoring formula. The HUI Mark 3 (HUI3) has eight attributes
(vision, hearing, speech, ambulation, dexterity, emotion, cognition, and pain) with five to six
levels per attribute. With the recent release of a multiplicative multi-attribute utility function for
the HUI3 system, users are now able to generate utility scores for HUI3 health states. The HUI3
scoring function is based on preference measurements obtained from a random sample of the
general population (>16 years of age) in Hamilton, Ontario, Canada. There is a high level of
agreement between directly measured utility scores for HUI3 health states and scores obtained
using the multiplicative function.
The HUI is a self-administered 15-item instrument that takes approximately 8 minutes to

complete. It will be administered to the full ACCORD sample at the same intervals as the HRQL
instrument is administered to the sub-sample. The measurement intervals are at 0, 12, 36, 48
months and study exit. This is important for the following reasons. First, it allows the sub-
samples assessed for HRQL and the sub-samples assessed for cost effectiveness to be linked in
analysis. Second, it allows for detection of the HRQL effects of diabetes complications occurring
in a minority of ACCORD participants, which requires a large sample size. Third, it allows the
calculation of valid incremental cost-utility scores for the various intensities of treatments to be
tested in ACCORD.
6.2.d Event-Related HRQL
Patients eventually accommodate to most decrements in health status. In the HRQL

literature, this has been referred to as “response shift.” In order to capture the effect of events
(such as MI, hospitalization or side-effects prompting medication discontinuation or drop-out)
close to the time when they occur, the single-item “feeling thermometer” from the EUROQOL
instrument will be used as part of the Interval History Form (every four months) for the full
sample. This asks patients to rate their health state from 0 (worst imaginable health state) to 100
(best imaginable health state). Through selection of various events already recorded in the
ACCORD database and an appropriate control group, the HRQL change related to many
different kinds of events can be assessed using this single item.
6.3 Cost Effectiveness Assessment
6.3.a Rationale
For this trial, the economic research questions are: 1) is the intensive glycemic therapy
more cost-effective than the standard glycemic therapy? 2) is the intensive lipid-lowering therapy
more cost-effective than standard lipid-lowering therapy? 3) is the intensive hypertension
therapy more cost-effective than standard hypertension therapy? and 4) is the intensive glycemic
therapy in combination with intensive blood pressure therapy or intensive glycemic therapy in

97
combination with intensive lipid therapy more cost-effective than the standard therapy? These
questions will be addressed by conducting incremental cost-effective analyses in which the net
costs and net effectiveness of intensive therapy defined by the main trial to standard therapy will
be calculated and expressed as a series of ratios. The perspective of this economic evaluation
will be a single payer of national health care system.
6.3.b Effectiveness
The primary endpoints defined by the main trial are considered as primary outcome
measures for this economic evaluation. Three primary effectiveness measures are identified and
include: 1) CVD free-year gained, 2) life-year gained and 3) quality-adjusted life-year (QALY)
gained. CVD free-year is defined as time until first occurrence of CVD endpoints. The measure
of life-year gained is determined by the difference in number of life–years between intensive
therapy and standard therapy. QALY’s will be calculated using utility values derived from the
HUI-3.
The cost effectiveness sample will comprise a total of 4,288 randomly selected
participants from the ACCORD study population.
6.3.c Resources
Therapies are conceptualized as having three stages: 1) initiation of the therapy, 2)

monitoring and maintenance of the ongoing therapy, and 3) treatment of side effects and
complications.
Resources consumed will be classified into the following categories: 1) initiation of the
therapy; 2) maintenance of the ongoing therapy, ambulatory services, and diabetes supplies; 3)
inpatient services.
6.3.d Costs
Under the perspective of a national health care system, all direct medical costs associated
with treatment of Type 2 diabetes and its complications and costs for treating adverse effects of
the therapy will be considered. These costs will include costs of inpatient care, outpatient care,
medications, medical equipment, supplies, laboratory tests, overhead, labor, and fringe benefits.
The participant's costs such as waiting time, transportation, lodging, and informal care arising
from the disease will not be included. Likewise, opportunity costs of premature death,
productivity loss, and long-term disability will not be considered in this study.
6.3.d.1 Cost Data Collection
To reduce the burden on data collection for economic analysis, data being collected from
the main trial will be used to the extent possible. Much of the data such as primary endpoints
and resources for the initial and the ongoing therapy will be routinely collected according to the
design of the main trial. The following sections describe methods for use in the collection of
medical resource consumption data, which are not collected by the main trial. In general, two

98
approaches will be used to collect these data: the case report form and use of administrative data
systems, including HCFA, VA, HealthPartners, and the Canadian Health System.
6.3.d.2 Intensive and standard therapy
Labor and fringe benefits of providers, overhead and resource used for patient
management, including telephone calls, letters, team meetings, and adherence activities, will be
collected at clinic site level. The estimated allocation of these resources to each therapy per
patient will be recorded on the Clinic Resource and Cost Questionnaire. Data on medications,
tests, and medical supplies for the therapies will be derived from the main trial.
6.3.d.3 Inpatient Care
Data on hospitalization will be collected at the patient-level. Cost for hospital care
represents a disproportionate share of direct medical costs (~70 to ~80%) and shows extreme
variability. Therefore, it is important to collect hospitalization data from all CEA study patients
in order to derive cost estimates with reasonably narrow confidence intervals. We will collect
hospital admission data primarily using the hospitalization form and through available
administrative data systems. Research staff at each clinic site will obtain a copy of the discharge
summary for each hospital admission in patients who participate in the CEA and then send a
copy of the discharge summary to the coordinating center. All study patients will be requested to
consent to the release of their medical records when they participate in the study. The
administrative claims data will be used to examine data quality since we are not able to obtain
hospitalization data for all study patients through the administrative data systems. The time
period for collection of hospitalization data by the hospitalization form is at each follow-up visit
and by the administrative data system is every two years. A specially trained medical coder at
the coordinating center will map diagnoses and procedures into DRGs.
6.3.d.4 Ambulatory Care
Use of ambulatory care services will be collected for all patients through self-reporting at
each follow-up visit and recorded into the clinic follow-up questionnaire. The data include the
number of clinic and/or physician office visits, the number of emergency room visits, and
number and type of outpatient diagnostic tests and procedures. Data on medication use will be
collected from the main trial.
6.3.d.5 Unit Cost
Primary and secondary data sources will be used to calculate unit costs of resources used
to reflect the cost for consuming an itemized service. Unit cost of hospital stay will be based on
Diagnosis-Related Group (DRG) of the Medicare. Unit costs for outpatient services, outpatient
procedures, laboratory tests, and consultations will be estimated using HCFA Medicare data.
The unit cost for physician services will be calculated using the Medicare Fee Schedule. The
unit cost of labor and fringe benefits and equipment and supplies consumed in case management
services will be assessed from secondary sources. The unit cost of medications will derive from
average wholesale prices using the Medical Economics Data Red Book. The total cost of each

99
treatment is calculated by multiplying the quantity consumed of each type of resource by the unit
cost. A discount rate of 5% will be used to adjust inflation over the study years.
6.3.e. Sample Size for Cost-Effectiveness
The sample size estimates are based on the following assumptions:
• Patients are randomized in a double 2 X 2 layout and each cell has the same expected
sample size.
• The decision rule in any health care system is that the intensive treatment should be
implemented instead of the standard treatment if the incremental cost-effectiveness ratio
(ICER) of the intensive treatment is less than maximal willingness to pay for additional
health effect (Rc).
• There is no correlation between cost and health effect.
• The test of interest is that the observed ICER derived from the trial is significantly less
than the ceiling cost-effectiveness ratio, Rc.
We defined the observed ICER as a ratio of extra cost to additional CVD free-year gained
comparing intensive treatment and standard treatment within the trial period. The following table
contains sample size estimates for the main effect of each of the intervention using Briggs’s
method [Brigg and Gray, 1998; Brigg and Tambour, 1998]. These estimates are based on the
results of UKPDS [UKPDS, 1998]. In the UKPDS study, the median duration of follow up was
8.4 years. The mean difference in time free from diabetes related end points including coronary
heart disease, cerebrovascular disease, amputations, laser treatment for retinopathy, cataract
extraction, renal failure and death, was 0.55 years. The median patient follow-up year in
ACCORD is expected to be 5.6 years, which is 33% shorter than UKPDS. We assume mean
difference in years free from CVDs would be 20% longer, that is, 0.66 (0.55+0.11) years in
ACCORD because of more intensive therapy in at least two major CVD risk factors. In addition,
we assume cost of treatment for type 2 diabetes in ACCORD is similar to UKPDS. The range of
Rc values is chosen based on O’Brien’s study [O’Brien, 1998]. The $30,000 represents the upper
limit of ceiling cost-effectiveness ratio (Rc). The total sample size for the CEA component is
4288, that is, 536 in each cell. It has 80% power to test the null hypothesis: the observed ICER >
Rc (=$15,000) at α=0.05 level. All eligible patients will be randomly allocated to each cell
through a multi-stage randomization process.
Table 6.1 Sample Size and Power of ACCORD CEA

Maximal willingness
to pay for additional Power
health effect (Rc). 70% 80% 90%
$5,000 607 772 1033
$10,000 459 583 781
$15,000 422 536 718
$20,000 405 515 689
$25,000 396 503 673
$30,000 389 495 663

100
6.3.f Data Analysis for Cost-Effectiveness
Two methods of cost-effectiveness analysis (CEA) and cost-utility analysis (CUA) will
be used in the economic evaluation. The ratios of cost to outcome derived from CEA/CUA are
used to compare cost-effectiveness among treatment strategies. An incremental cost-
effectiveness ratio (ICER) will be calculated, which provides a summary of the cost-
effectiveness of one intervention relative to the other.
The basic formula to calculate incremental CEA ratio and CUA ratio of a specific
treatment A relative to the reference treatment B is presented as following:
ICERCEA = (Mean Cost treatment A - Mean Cost treatment B )

(Mean Effecttreatment A - Mean Effect treatment B)
ICERCUA = (Mean Cost treatment A - Mean Cost treatment B )

(Mean QALY treatment A – Mean QALY treatment B)
The ratio of incremental cost to increment effectiveness represents cost-effectiveness of

the specific treatment. This ratio is a point estimate. Bootstrap methods will be used to calculate
confidence intervals for cost-effectiveness ratios. In addition, sensitivity analyses will be
performed in order to examine effects of key parameters on cost-effectiveness ratios.

101
Chapter 7
Statistical Considerations
7.1. Design
ACCORD is designed with factors consisting of: intensive versus standard glycemic
control, intensive versus standard blood pressure control, and in the presence of LDL-C
lowering, fibrate versus placebo. All 10,000 participants will be randomized to the glycemic
groups, and to either the blood pressure or lipid groups in two non-overlapping 2 X 2 layouts
(Figure 1.2): 5800 participants will be randomized to the lipid groups and 4200 to the blood
pressure groups. Throughout this chapter, a “+” will refer to the more intensive level of an
intervention whereas a “-” will refer to the less intensive level.
The first year of recruitment, treatment, and follow-up was designated the Vanguard Phase
of the trial. The specific goals of the Vanguard, which were used to judge its success, are
described in Section 7.5. In this phase, 1184 participants were randomized over a 5-month period
and then recruitment was suspended. Vanguard participants will have approximately 18 months of
follow-up, on average, when randomization for the main trial begins. Approximately 8800
participants will be randomized during the main component of ACCORD study. Randomization for
this phase will take place from January 2003 through June 2005. All participants (including the
1184 recruited in the Vanguard Period) will be treated and followed through June 2009. Thus, the
length of follow-up in ACCORD will range from 4.0 to 8.4 years (approximate mean of 5.6 years).
Participants eligible for the lipid intervention but not the blood pressure intervention will
be randomized to one of the four cells in the lipid and glycemic control 2 X 2 layout, and
similarly for those eligible for the blood pressure intervention but not the lipid intervention.
Participants eligible for both the lipid and the blood pressure interventions will be randomly
assigned to one or the other trial. This random assignment will be weighted to ensure that the
4200 BP/5800 Lipid split is efficiently achieved. Data from the Vanguard portion of ACCORD
indicates that 30% of participants will be eligible for both blood pressure and lipid interventions,
30% will be eligible for the lipid but not the blood pressure intervention, and 40% will be
eligible for the blood pressure, but not the lipid intervention.
During the Vanguard period, all randomizations were stratified by clinical center network
and baseline CVD status (either primary or secondary prevention) using permuted blocks. All
randomizations for the main trial will be stratified by clinical site. The reason for the change to
use clinical site rather than clinical center network and CVD status as stratification factors is to
provide more balance in the types of participants seen within clinical sites. Improved balance
should result in clinic personnel having more experience in following the protocol for ACCORD
participants randomized to each of the eight possible conditions. With randomization of 10,000
participants, it is anticipated that there will be good balance related to CVD status without
including this as a randomization factor.

102
7.1.a Primary Hypotheses
cardiovascular disease (CVD) event:
(2) in the context of good glycemic control, does a therapeutic strategy that uses a
fibrate to raise HDL-C/lower triglyceride levels and uses a statin for treatment of
LDL-C reduce the rate of CVD events compared to a strategy that only uses a
statin for treatment of LDL-C?
systolic blood pressure (SBP) of < 120 mm Hg reduce the rate of CVD events
Analyses of each of the primary hypotheses will be conducted within separate models as
comparisons of the marginal effects and not as comparisons among the individual cells. All
participants will be included in the analysis for the glycemia hypothesis. Each hypothesis will be
tested using a 2-sided probability of Type 1 error = 0.05.
7.1.b Secondary Hypotheses
Several secondary hypotheses will be tested for each of the glycemia, lipid and blood
pressure hypotheses. The hypotheses are to determine whether more intensive treatment
compared to standard treatment reduces the occurrence of:
1) an expanded macrovascular disease outcome, consisting of the primary outcome plus

revascularization plus hospitalization for heart failure (defined in Section 5.2)
2) total mortality
3) each of the separate components of the primary outcome
4) a composite microvascular disease outcome, including kidney and eye disease (defined in
Section 5.2) (with neuropathy added for the glycemia trial)
These outcomes will be analyzed as comparisons of marginal effects. HRQL and cost-
effectiveness are also examined as secondary outcomes and are described in Chapter 6.

103
7.1.c Subgroup Hypotheses
The two subgroup hypotheses for the glycemia intervention are to determine if:
(1) Effects of glycemic control on the primary outcome are the same across baseline
levels of HbA1c, and
(2) Effects of glycemic control on the primary outcome are independent of effects due to
the blood pressure and lipid interventions.
The three subgroup hypotheses for the lipid intervention are to determine if the benefits
of fibrate ( in the context of desirable levels of LDL- C and good glycemic control) are:
(1) Equal across levels of LDL-C measured prior to initiation of fibrate therapy,
(2) Equal across HDL-C levels measured prior to initiation of fibrate therapy, and
(3) Equal across triglyceride levels measured prior to initiation of fibrate therapy.
Consistency of the effects for the glycemia, lipids, and blood pressure interventions will
also be examined in subgroups defined by gender, age, race/ethnicity, and presence of clinical
CVD at baseline (i.e., primary and secondary prevention participants), and the presence/absence
of the other interventions.
7.2 Analysis Plan
7.2.a Plan for Primary Hypotheses
Intention to Treat -- Primary comparisons of intervention groups will be performed

according to the intention-to-treat principle. All randomized participants in these analyses will be
grouped according to their intervention assignment at randomization, regardless of adherence.
Analytical Techniques -- Separate models will be used to test the primary hypothesis
associated with each intervention. The main comparisons of the intervention groups with respect
to the distribution of time until first identification of a CVD endpoint (described in detail in
Chapter 5) will be based on survival analysis methods. Failure time will be measured from the
time of randomization.
To test each primary hypothesis, a proportional hazards model will be used (Cox 1972)
incorporating adjustment for important factors specified below. This will be the primary
analysis.
Glycemic Hypothesis: The glycemic hypothesis will be tested in all 10,000 randomized
participants. The model to be fit will contain separate indicator variables that identify
participants: (a) in the BP trial, (b) in the BP trial AND randomized to the BP(+) intervention, (c)
in the lipid trial, (d) in the lipid trial AND randomized to fibrate(+), and (e) randomized to
intense glycemic control. In addition to these variables, indicator variables will be included that
identify: (f) secondary prevention participants, and (g) Clinical Center Networks. Our reasoning

104
for including term (f) is that secondary prevention participants should have higher event rates
than primary prevention participants. Likewise, term (g) will be included because the clinical
networks contain very different types of participants that may have different event rates. For
example, the VA clinics will primarily consist of men. The main comparison in this model will
be based on the chi-square statistic from a likelihood ratio test obtained from proportional
hazards models with/without term (e).
Lipid Hypothesis: The lipid hypothesis will be tested in approximately 5800

participants. The model to be fit will contain terms (d), (e), (f) and (g). This hypothesis will be
tested using a likelihood ratio test for models with/without term (d).
Blood Pressure Hypothesis: The blood pressure hypothesis will be tested in

approximately 4200 participants. The model to be fit will contain terms (b), (e), (f) and (g). This
hypothesis will be tested using a likelihood ratio test for models with/without term (b).
Kaplan-Meier (Kaplan & Meier 1958) estimates of survival will be obtained for the
intervention and control groups for each hypothesis. Estimates for the proportion of participants
who remain event free at pre-specified time points, and the associated confidence intervals, will
be constructed (Peto 1977). The hazard functions will be assessed for proportionality using
log/log plots of survival and Schoenfeld residuals. An unadjusted analysis (i.e., a log-rank test)
will also be performed.
7.2.b Secondary Hypotheses
Testing each of the secondary hypotheses for glycemic, blood pressure and lipid control
involves analysis of the time until the occurrence of a secondary outcome. The planned analyses
for each of these outcomes will parallel the analyses performed for the primary outcome.
Proportional hazards models containing the terms specified in the models presented in Section
7.2.a will be specified to test each of the secondary hypotheses. The one exception to this
analysis plan will be exclusion of the term controlling for Clinical Center Networks.
7.2.c Subgroup Hypotheses
Testing each of the subgroup hypotheses will be carried out using survival analysis
methods, as all subgroup hypotheses involve time until the occurrence of the primary outcome.
For each subgroup hypothesis, the proportional hazards model used to address the primary
hypotheses will serve as the base model to which additional terms will be added to test each
subgroup hypothesis.
To address the glycemia subgroup hypothesis to determine if relative risks for the
primary outcome are the same across levels of HbA1c, a term representing HbA1c levels will be
entered into the proportional hazards model. A test of the interaction between this term and the
term representing the glycemic intervention effect will address this initial subgroup hypothesis.
To address the second glycemia subgroup hypothesis, whether the effects of glycemic control on
the primary outcome are independent of effects due to the blood pressure and lipid interventions,

105
the significance of interactions between the terms representing each of the interventions will be
investigated.
Each of the three subgroup hypotheses for the lipid intervention will also be investigated
through the use of interaction terms in the proportional hazards model. In particular, these
hypotheses will be investigated in three separate models by testing the significance of the
interactions between the variable representing the fibrate intervention and variables
characterizing: (1) baseline LDL-C levels, (2) baseline HDL-C levels, and (3) baseline
triglyceride levels.
Finally, consistency of effect in demographic and primary/secondary prevention

participants, and in the separate 2 X 2 trials, will be tested by stratified analyses and by
investigating the significance of the interaction between the variable representing the
intervention and variables characterizing subgroup membership.
7.3 Power Considerations
7.3.a Summary
Given the assumptions presented below, the ACCORD study is designed to have:
• 89% power to detect a 15% treatment effect of intensive glycemic control compared with
standard glycemic control,
• 87% power to detect a 20% treatment effect of lipid control through LDL-C lowering and
fibrates compared with lipid control using LDL-C lowering alone,
• 94% power to detect a 20% treatment effect of intensive blood pressure control compared
with standard blood pressure control.
7.3.b Computational Details and Sensitivity Analyses
As described below in Section 7.3.c, population event rates have been estimated from two
major observational cohort studies conducted in the United States, (ARIC and CHS). The
strategy for power calculations is to estimate cumulative event rates for each of the 8 cells of the
design and then average the event rates for the “+” and “-” intervention cells appropriate for
comparing the more and less intensive levels of the intervention (i.e., 4 cells with “+” versus 4
cells with“-” for glycemia, 2 cells with “+” versus 2 cells with “-” for fibrate, and 2 cells with
“+” versus 2 cells with “-” for blood pressure control). Finally, a simple binomial power
calculation was performed based on the two averaged event rates.
The difficult aspect of this calculation is estimation of the event rates in the two “control”
cells where only less intensive interventions “-” are applied (e.g., the less intense BP and less
intense glycemic interventions for participants in the BP and glycemic interventions). The
strategy for estimating these control event rates is to use the ARIC and CHS data to select
patients similar to those who meet the various eligibility criteria, calculate an event rate for the

106
selected patients, and then adjust this event rate to reflect differences between ARIC/CHS and
ACCORD.
Assuming that 40% of participants will be secondary prevention and 60% will be primary
prevention, then using the ARIC and CHS data, the annual population event rates are taken to be
6.56% among those who satisfy both the BP and lipid criteria and 5.60% among those eligible
for the lipid intervention. Details of these calculations are deferred to the next section. As an
approximation, we use the rate 6.56% for those randomized to the BP and glycemia 2 X 2, and
the rate 5.60% for those randomized to the lipid and glycemia 2 X 2.
For the power computations, these rates are further adjusted by applying:
• 25% reduction in event rate due to healthy volunteer/background therapy effect in ACCORD,
• 18% reduction in event rate due to lowering HbA1c from the mean of 9.3% observed in
ARIC/CHS to 7.5% in ACCORD (assuming a 1.5% reduction in HbA1c is associated with a
15% effect),
• 5% increase in event rate when silent MI is added because silent MI was not included in the
ARIC and CHS rates,
• 1% reduction in event rate per 1 mg/dl drop in LDL-C from the mean of 120 mg/dl observed
in ARIC and CHS.
As a result of placing all participants in the lipid and glycemia 2x2 on 20 mg of a statin,
the mean LDL-C is assumed to be approximately 95 mg/dl for these ACCORD participants, and
125 mg/dl for participants enrolled in the BP and glycemia 2x2.
Based on these assumptions, event rates for the two cells involving less intensive
interventions were calculated.
• Participants in the lipid trial and randomized to fibrate (-) and glycemia (-) have an estimated
event rate of
5.6% x (1-0.25) x (1-0.18) x (1.05) x (1-0.25) = 2.71%.
• Participants in the blood pressure trial and randomized to blood pressure (-) and glycemia
(-) have an estimated event rate of
6.56% x (1-0.25) x (1-0.18) x (1.05) x (1+0.05) = 4.45%.
For each of these rates, the rate expected after K years of follow-up was calculated using
the formula (1-(1-rate)K). Vanguard participants will have a mean follow-up of approximately
8.2 years, whereas participants randomized in the main trial will have a mean follow-up of
approximately 5.2 years. For each of the 8 cells in the design (Figure 1.2), the expected
intervention effect was then applied to the above K-year rates to obtain the expected rate within
the cell separately for Vanguard and main trial participants. Event rates contained in cells used to
address each hypothesis were averaged by applying appropriate weights reflecting the number of
participants randomized to each cell.

107
Power computations are based on having 10,000 participants in the glycemia

intervention, 5800 in the lipid intervention, and 4200 participants in the blood pressure
intervention, and assume that the main trial will enroll 40% secondary prevention participants.
The Vanguard period had approximately 35% secondary prevention participants. The power is
computed based on effect sizes of 14.7% (12% for Vanguard participants and 15% for main trial
participants) for the glycemia intervention and 20% for the lipid and BP interventions, with
adjustments of the main effects for attenuation attributable to the factorial design (i.e., assume
that each intervention has the anticipated treatment effect). Type I error is set at 0.05 for each of
the three primary hypothesis.
Power for each intervention in ACCORD is presented in Table 7.1.
Table 7.1 Power for ACCORD Trial

Glycemia Lipid Blood Pressure
Power 89 87 94
Sensitivity analyses were also performed to determine how the mixture of

primary/secondary prevention participants recruited during the ACCORD main trial would affect
the power for each intervention. In Table 7.2, underlying assumed annual event rates and
associated power are provided for assumptions of 65%, 60%, 55%, or 50% main trial primary
prevention participants. Note that the 60% primary prevention assumption corresponds to the
power calculations presented in Table 7.1.
Table 7.2. Sensitivity of Power to Percentage Mix of Secondary Prevention Participants
(Assuming mean LDL-C=95 mg/dl in participants in lipid intervention

and 125 mg/dl in participants not in lipid intervention)
% Primary in Annual Event Rates in Cells Defined By Glycemia (-)* Power
Main Trial BP (-) Fibrate (-) Glycemia Lipid BP
0.65 0.0428 0.0259 0.88 0.85 0.93
0.60 0.0443 0.0270 0.89 0.87 0.94
0.55 0.0458 0.0281 0.90 0.88 0.94
0.50 0.0472 0.0291 0.91 0.89 0.95
Two additional sensitivity analyses were undertaken. The first investigated the power for
the glycemia intervention if the lipid and BP interventions did not work. In this situation,
ACCORD has greater than 92% power for the glycemia hypothesis under each of the
primary/secondary mixtures presented in Table 7.2. In addition, the power was estimated if the
use of 20 mg of a statin resulted in a 20% reduction in the ARIC/CHS rates rather than the 25%
assumed in Table 7.2. The effect of this changed assumption was to increase the glycemia power
by 1% and the lipid power by 2% in each row of Table 7.2.

108
7.3.c Details of Estimation of Event Rates In ARIC and CHS
The Atherosclerosis Risk in Communities (ARIC) study is a population-based study of

cardiovascular disease and atherosclerosis in 15,792 middle-aged men and women from 4 U.S.
communities. Participants were 45-65 years old at the time of the baseline examination, which
was conducted between 1987 and 1989. Participants have been contacted annually by phone
since their baseline exam to obtain information regarding hospitalizations and cardiovascular
events, which were subsequently validated by medical record review. Three follow-up
examinations were conducted at 3-yr intervals between 1990 and 1998.
The Cardiovascular Health Study (CHS) is also a population based study of risk factors
for cardiovascular and cerebrovascular disease in elderly men and women from four U.S.
communities. The cohort consisted of 5,201 community-dwelling adults aged 65 years or older
who had a baseline clinic visit in 1989-1990 (original cohort) and an additional 687 African-
American adults aged 65 years or older who had a baseline clinic visit in 1992-1993 (new
cohort). Incident CVD events, including myocardial infarction (MI), stroke, transient ischemic
attack (TIA), and death, were identified by 6-month telephone calls and annual clinic visits.
Further verification was done by the CHS Events Subcommittee, using Medicare Part A hospital
discharge lists, hospital records, outpatient records, and physician reports.
ARIC and CHS data were used to determine the event rates that would be found among
age-eligible diabetic participants. Secondary prevention participants were defined as those who
had a history of MI, stroke or revascularization. Primary + RF participants were defined as non-
secondary participants who were either current smokers, or met the ACCORD definitions for
HDL-C and ABI risk factor criteria. Further exclusion criteria that were applied to define the
sample were: age=>55 years, 7.5%<=HbA1c<=11%, LDL-C<=170 mg/dl, Trig<150mg/dl. For
these computations, the ACCORD lipid and blood pressure selection criteria were applied to the
primary and secondary prevention participants. HbA1c data were not available. For the HbA1c
criterion, Fasting plasma glucose (FPG) was required to be < 216 and > 140 mg/dl. The upper
limit was obtained through use of the regression equation from Avignon (1997). That paper gave
a formula correlating HbA1c and FPG from linear regression: y = 17.0x + 29.1, r = 0.62 where
y=FPG and x=HbA1c, so HbA1c = (FPG - 29.1)/17. Table 7.3 contains a summary of yearly
event rates estimated from CHS and ARIC after applying the exclusion criteria. An event was
defined as the occurrence of fatal CHD or fatal/non-fatal stroke or myocardial infarction.
Averaging results from the two studies with equal weight, the event rate is estimated to be
approximately 3.6% per year in primary + RF participants and 8.6% per year in secondary
prevention participants.
Table 7.4 presents similar statistics to Table 7.3, except in this table systolic blood
pressure was restricted to be between 130 and 170. Overall, the event rate is estimated to be
approximately 4.6% per year in primary + RF participants and 9.5% per year in secondary
prevention participants.

109
Table 7.3 Event Rates From CHS and ARIC

After Application of Lipid and HbA1c Selection Criteria
Rate per Mean Mean Mean Mean HbAlc

year Age (years) LDL-C (mg/dl) SBP (mm Hg) Based on FPG (%)
CHS
Secondary 12.1% 73 113 137 8.3
Primary + RF 5.0% 72 116 139 9.1
ARIC
Secondary 5.0% 61 129 130 9.8
Primary + RF 2.1% 60 127 129 9.7
Total
Secondary 8.6%* 67 121 134 9.1
Primary + RF 3.6%* 66 121 134 9.4
* 0.4 x 8.6% + 0.6 x 3.6% = 5.6% (see Section 7.3.b)
Table 7.4 Event Rates From CHS and ARIC

After Application of Lipid, BP and HbA1c Selection Criteria
Rate per Mean Mean Mean Mean HbAlc

year Age (years) LDL-C (mg/dl) SBP (mm Hg) Based on FPG (%)
CHS
Secondary 13.6% 73 115 146 8.4
Primary + RF 6.3% 74 118 146 9.7
ARIC
Secondary 5.4% 61 135 144 9.2
Primary + RF 2.8% 61 126 143 10.1
Total
Secondary 9.5%** 67 125 145 8.8
Primary + RF 4.6%** 68 122 145 9.9
** 0.4 x 9.5% + 0.4 x 4.6% = 6.56% (see Section 7.3.b)

110
7.4 Statistical Reports
7.4.a Steering Committee Reports
Periodic reports will be generated for the Steering Committee, Clinical Center Networks,
and Clinical Sites. These reports will include information on recruitment, loss to follow-up,
adherence, baseline covariate information on the comparability of treatment groups, and adverse
events. Information will be stratified by Clinical Center Networks and Clinical Sites. Other
reports will include information on quality control for central facilities and data entry.
7.4.b Data and Safety Monitoring Board (DSMB) Reports
The role and composition of the Data and Safety Monitoring Board are described
in Section 13.8. Meetings of the DSMB will be held at least annually. Material for these
meetings will be distributed two weeks in advance of the meetings. Up-to-date statistical
analyses will be provided to the DSMB in preparation for their meetings. The analyses
will include data on recruitment, outcome measures, any side- or safety effects,
adherence, and quality control, and will be designed in cooperation with the DSMB.
Interim analyses of the intervention effectiveness will be performed at times coinciding
with the meetings of the DSMB, and will be controlled to protect the overall Type I error
of the trial. These results will be for the use of the DSMB and will not be revealed to the
investigators. The purpose of these analyses will be for the DSMB to assess the trial
progress with respect to intervention efficacy and safety, for possible recommendations
regarding early termination of the entire trial or any individual intervention.
Interim analyses will be performed periodically for the Data and Safety
Monitoring Board (DSMB). Monitored parameters will include the following:
1. HbA1c separation between groups
2. HbA1c distribution within groups
3. Primary outcome results
4. Conditional power
5. Adverse events
6. Recruitment progress
7. Other Event Rates
7.5 Vanguard Phase Criteria
Recruitment and a minimum of one-year treatment/follow-up of 1184 participants

was designated the Vanguard phase of the trial. The purpose of the Vanguard phase was
to determine the feasibility and success of implementing the protocol treatments and of
achieving the treatment goals throughout the ACCORD clinical sites. The following
were the specific goals of the Vanguard phase, which were used to judge its success:

111
Recruitment. The goal was to randomize 1,000 Vanguard participants within four
months after the first randomization.
Glycemia control. There were two glycemia control goals, which were analyzed
in participants with at least 8 and 12 months of post-randomization follow-up:
a) the median achieved HbA1c in the intensive group would be less than 6.5%,
and
b) the difference in median HbA1c between the standard and intensive groups
would be at least 1.3%.
Blood Pressure. The study target is 20 mm Hg difference between the two BP

arms. Evidence suggests that a 10-12 mm Hg difference may produce a 20% effect on
CVD events. The target for the end of the Vanguard period was 10 mm Hg, which
allowed time to titrate medications. An achieved mean SBP < 130 mm Hg in the
intensive BP group was considered adequate for the Vanguard phase, if it also was 10
mm Hg lower than the less intensive group.
Lipid. Targets for assessment of the fibrate intervention were based on adherence.
The target adherence rate to fibrate and to statin was at least 80% as measured by simple
self-report.
The Vanguard Phase was successful in achieving almost all of these goals, and the
trial protocol was modified to increase the likelihood of achieving all of the goals. These
modifications are incorporated into this main trial protocol.

112
Chapter 8
Data Management and Training
8.1 Overview: Use of the World Wide Web
All Clinical Center Networks and Clinical Sites will use the World Wide Web (WWW) to
enter ACCORD data collected on forms from participants seen within the Clinical Sites. In
situations where an internet connection is not available at a Clinical Site, the Clinical Center
Network will be responsible for entering data from participants’ forms. Each Clinical Site will have
a password protected area on the ACCORD home page through which data will be entered.
Documentation of the data entry system will be maintained at the Coordinating Center (CC). In
addition, training materials for measurement and data entry personnel will be available in
downloadable format on the ACCORD web site. Site-specific reports relating to patient
demographics, recruitment goals, etc., among other reports, will be available on the web site.
Data security in the web-based data system uses 128-bit encryption and Secure Socket
Layer (SSL). Recovery from disasters such as natural phenomenon (water, fire, or electrical) is
possible through the ability to reconstruct both the database management system and the data up to
the last back-up through the use of nightly backups. This will ensure optimal recovery of data
systems in the event of a disaster. Back-up tapes are kept in a locked, fire and waterproof storage
cabinet away from the computer room. Additional back-up tapes will be stored at another location
on the Wake Forest University School of Medicine campus.
8.2 Flow of Data from Trial Units to Analytical Databases
8.2.a Data from Clinical Sites and Clinical Center Networks
Patient Randomization: An internet-based, web browser randomization procedure will be

employed in ACCORD. Clinical Sites access the study web site and initiate the interactive
randomization page. Entry into this area is password protected and encrypted. Once security
requirements are satisfied, a series of questions follows, interrogating the user for identifying and
eligibility information. If these questions have been appropriately answered, a patient identification
number is issued. When the session is complete, an e-mail process is spawned and a record of the
transaction is sent to the Clinic Coordinator, the Regional Coordinator at the appropriate Clinical
Network Center, and the Project Manager at the CC indicating that the patient has been properly
randomized and appended to the database.
Patient Tracking: The Patient Tracking System (PTS) is a fully integrated tracking and
notification system that advises clinic staff about patient follow-up windows, as well as projected
clinic and laboratory workload for a week at a time (longer if necessary). Tracking a patient begins
at screening and continues automatically throughout the project by integrating patient follow-up
data with predetermined follow-up "windows". When a participant is enrolled into the study, a
schedule of target dates for each of the visits is automatically generated by the CC. The report
details the recommended "windows" that each visit should fall into and a case file is created for the
participant. In addition, the clinic will generate personalized form letters to be sent at the prescribed
intervals reminding patients (and clinic staff) of an impending visit. At the end of each week, a

113
listing of patients that are due for follow-up detailing each patient's required tests is transmitted to
the clinic to assist in staff and laboratory resource allocation for that week. Reports detailing
deviations from the protocol are automatically generated and transmitted to the clinic via e-mail
attachments. These data will also be available in the study web site.
Data Entry: Data entry screens are developed in HTML, with a Cold Fusion to Oracle (a
relational database management system) backend. The images on the screens mirror the data
collection forms for ease and accuracy of entry.
As patient visits are completed, and hard copy forms are filled out, the clinic coordinator
reviews each form for accuracy and completeness, including laboratory reports and any supporting
documentation (pertinent hospital records, etc.). Once any data problems have been resolved, data
are entered by clinic staff into the computer via the web-based browser application. During data
entry, key variables are checked for accuracy with the assignment of ranges. Where data are
entered outside of preset ranges, entry is denied pending the review for accuracy. Override
capabilities exist; however, these responses are flagged for review upon receipt by the CC. The
Project Manager will reconcile any responses that continue to be questionable. Also, a sample of
key forms is required to be double-keyed for entry verification and identification of problem
fields/forms.
8.2.b Data from Central Chemistry Laboratory and ECG Reading Center
Laboratory specimens and electrocardiographic data are sent to the Central Chemistry
Laboratory and ECG Reading Center from the Clinical Sites on a fixed schedule. The Clinical Sites
log each shipment specifying patient identification and visit sequence in a computerized format.
This information includes dates for specimen/test acquisition as well as shipping. The Central
Chemistry Laboratory and ECG Reading Center provide results to the CC on live internet feed,
which will include all log information as well as the date of analysis. Depending on clinic needs,
reports will be sent to assist in the clinical functions (eg., providing timely feedback to the clinic on
any measurement that exceeds a predefined alert level).
8.2.c Analytical Database Closure
At regular intervals, data queries will be carried out using SAS/Connect to perform
consistency checks on key variables and forms. Although much of this will be done at the data
entry level in the clinic, this additional pass through the data will serve as a quality control check
before the data files are merged with the permanent SAS datasets on the CC Sun computer.
Upon study completion, after all clinic and laboratory data have been collected and filtered
through the quality control routines, the Oracle database will be converted to SAS datasets and
certified. The database will be taken off-line and archived on magnetic tape and/or CD-ROM. The
final SAS datasets will be certified and issued version numbers to synchronize analytic efforts and
distributed in accordance with ACCORD Steering Committee and NHLBI policy. The choice of
media on which to copy and distribute copies of the database to the investigators will depend upon
the systems and media available at that point in time. Final data tapes and documentation will be
sent to NHLBI.

114
8.3 Feedback to Clinical Sites and Clinical Center Networks
As described in Chapter 12, a routine system of data edit reports will be generated to help
ensure that all data are entered in timely and complete manner. These reports will include both
the assessment for each Clinical Site of the time between data collection and entry, and the
generation of reports by the CC of missing items. These reports will be provided to the Clinical
Center Networks, Clinical Sites, and Measurement Procedure/Quality Control (MP/QC)
Subcommittee on a regular basis so that data collection items that are troublesome can be
identified and Clinical Sites not meeting the standards set by the MP/QC Subcommittee can be
notified. Network Coordinators will be copied on all data reports for Clinical Sites within their
network and asked to follow-up on any action that needs to be taken.
8.4 Training
Each Clinical Center Network will appoint a Training and Quality Control Liaison. This
person will be responsible for the maintenance of measurement and training standards at the
Clinical Sites including training of new personnel and re-certification for existing staff. The
Training and Quality Control Liaison will be familiar with all measurement requirements for
ACCORD and provide input into the scheduling of clinic activities so that there is adequate time
for clinic staff to carry out their responsibilities while meeting quality standards.
Key clinic staff from each Clinical Center Network and Clinical Site will be trained at an
initial central training session. These key staff will be responsible for training and re-training
other staff members at the Clinical Site. Specific procedures for training clinic staff to obtain
ACCORD data is provided in the Manual of Procedures (MOP). A copy of the MOP is on the
ACCORD web site. This searchable electronic copy allows clinic personnel to quickly find an
answer to any question about procedures.

115
Chapter 9
Adherence
9.1. Background and Rationale
The overall adherence approach for this large and lengthy trial should be based on two
essential principals (Probstfield 1986, Probstfield 1990). First, keys to good adherence in a
clinical trial are anticipation and a prevention oriented approach. Second, effective adherence
plans for a clinical trial are implemented during the protocol development and recruitment
periods, and revised during follow-up as needed.
As part of the central pretrial training sessions, all investigators and clinical coordinators
will receive instruction on adherence issues. Additionally, they will periodically have refresher
instruction in the overall adherence program throughout the follow-up period. The key
components of an overall adherence program are described in section 9.5.
9.2 Determining Adherence Potential
A run-in period, during which potential participants will be asked to monitor capillary
blood sugars, is employed to screen out individuals who may have difficulty adhering to the
requirements of the overarching glycemia component of the trial.
9.3 Monitoring Adherence
Adherence will be monitored by self-report of each of the prescribed medications for glycemic
control and for the lipid lowering and blood pressure control at each visit. This approach has been used
in many other trials. Also, pill counts of selected medications will be conducted during the Vanguard
Phase. Details for this monitoring are provided in the Manual of Procedures (MOP).
9.4 Procedures for Maintaining/Improving Adherence
The details for an overall adherence program are provided in the MOP. Briefly, an
overall program for adherence must be implemented at the time that recruitment is started. This
will include a centralized training and regular refresher courses on adherence issues for the
Clinical Center Network (CCN) and Clinical Site staffs throughout the trial.
The Recruitment and Retention Subcommittee will meet by conference call on a monthly
basis during the follow-up portion of the trial for the purpose of monitoring adherence
performance. They will review data provided by the Coordinating Center that will be directed
primarily at assessing adherence at the CCN level. The variables reviewed will include the level
of adherence for each of the three interventions. Also, the number of participants off an assigned
intervention, the number of dropouts, and the number of participants lost-to-follow-up will also
be reviewed. Adherence at the individual Clinical Site level will be reviewed by the respective
CCNs. Guidance to the individual CCNs will be provided as needed.

116
9.5 Components of an Overall Adherence Program
The key components of an overall adherence program include the following.
• Develop a bottom line for adherence that cannot be transgressed by any trial participant.
Although quite simple in statement, the key issue in a trial with the intent to alter the natural history
of a disease is the status of every participant with respect to the primary outcome measure at
termination of follow-up. This allows an intention-to-treat analysis to be performed without
reservation.
• Set goals for adherence before the trial starts. Implement strategies that make those goals possible.
Evaluate them periodically. Power calculations are routinely made taking into account numbers of
participants who will go off medication and those who may crossover treatment assignment during
the trial. These estimates should be used regularly as minimum goals for adherence in the study.
• Do not randomize all “number eligible” screenees. The eligibility criteria in this trial are very
wide to allow maximum numbers to be eligible for enrollment. While the ability to discern
objectively those who will not perform well in the study is limited, any exhibition of hesitancy or
inability to proceed through the pre-randomization period should be regarded as a caution against
randomization of that screenee.
• Pay attention to signs and symptoms of potential poor adherence. Codification of red flags for
potential poor adherence has been used previously in trials. These may help Clinical Site staff
identify potential non-adherers at any time during trial conduct.
• Use an adherence team approach, if possible. More than one individual sees screenees and
participants in a clinic. All interactive information can be useful in the maintenance of good
adherence.
• Use a constant caretaker model, if possible. Participant interaction with the same staff person
consistently is thought to be useful. Use when possible. Transitions to other staff may be
necessary. Make them as smooth as possible.
• If necessary, modify dosing regimen. The protocol-prescribed dose of the medications may be
altered in order to accommodate the participant. Every attempt should be made to keep these
intervals short, but long-term accommodations may be necessary.
• Use adherence techniques. There are a host of adherence techniques that have been used
previously in trials, eg., occasion cards, appointment reminders, intervening phone calls, etc. These
will be systematically reviewed for staff use.
• Use the behavioral counseling approach. Interviewing and counseling techniques have been
shown to aid staff in sustaining long term adherence performance. These include identification of
barriers to adherence and individualized problem solving.

117
• Have an intervention program for poor adherers. Poor adherers and drop-outs are recoverable to
productive trial participation, as shown in other studies. Instruction to staff will be provided for the
approach to these challenging participants.
• Have a maintenance plan for all participants. Sustaining long-term adherence in trials is a
challenging task to begin with, and will be a key issue in this trial with its lengthy intervention and
follow-up.
• Adherence Techniques. The specific techniques will be included in an Adherence Binder that
each site will have at their clinics, along with samples of letters and ongoing additional tools
for the clinical sites and CCN’s to use. Centralized training of techniques will occur for
ACCORD clinic staff.

118
Chapter 10
Ancillary Studies
10.1 Introduction
In addition to the main ACCORD protocol, investigators may wish to perform ancillary
studies using the ACCORD population, blood or urine samples, or collected data.
An ancillary study is an investigation with objectives that are not related to the ACCORD
protocol but uses ACCORD participants, samples, or data collected by ACCORD. In most cases,
an ancillary study will involve acquisition of additional data that are not compiled as part of the
standard ACCORD data set. An ancillary study may or may not use all randomized participants.
Investigators are encouraged to propose and conduct ancillary studies. Such studies
enhance the value and productivity of ACCORD, and help ensure the continued interest of the
diverse group of investigators who are critical to the success of the trial as a whole. These studies
provide an exceptional opportunity for investigators, either within or outside of ACCORD, to
conduct additional projects at minimal cost. In general, ancillary studies will require additional
funding from the NIH or other sources.
10.2 Application Review Process
To protect the integrity of ACCORD, all ancillary studies must be reviewed and approved
before access to ACCORD data, samples, or participants is permitted. Investigators will not be
allowed access to the ACCORD participants, samples, or database without approval. New
ancillary study proposals should be sent to the ACCORD Laboratory & Ancillary Studies (LAS)
Subcommittee. Ancillary study forms can be obtained by calling the Coordinating Center or
accessing the LAS Subcommittee website. When the application is complete, the study proposal
will be sent to the LAS Subcommittee for review. The LAS Subcommittee will have two weeks
to review the proposal and make a recommendation to the Steering Committee. Preliminary
approval/disapproval will be made by the Steering Committee, with a final recommendation for
approval/disapproval made by the Data and Safety Monitoring Board to the NHLBI Director.
An ACCORD principal investigator or co-investigator must be included as the principal

investigator and/or co-investigator in every ancillary study proposal. If Coordinating Center
resources are to be used, arrangements must be made with the Coordinating Center Principal
Investigator. In general, costs associated with ancillary study data management at the
Coordinating Center must be budgeted into each ancillary study.
All proposed ancillary studies must be submitted to the LAS Subcommittee in time for
review, circulation to appropriate committees, and to obtain clearance prior to submission to a
funding agency. As a beneficiary of a collaborative study, each investigator must realize that
other investigators must be given an opportunity to participate in proposed studies and to offer a
critique of the proposal. Such collaboration will often strengthen the ancillary study. Studies

119
submitted for approval less than 60 days prior to a funding application deadline may not receive
timely approval.
During the review process, highest priority will be given to studies which:
• do not interfere with the main ACCORD objectives,
• have the highest scientific merit,
• produce the least burden on ACCORD participants,
• have objectives closest to those of ACCORD,
• require the unique characteristics of the ACCORD cohort, and
• provide opportunities for more junior investigators to serve as the PI of a project.
Investigators with approved ancillary studies will report to the Chair of the LAS Subcommittee
every year regarding the status of study funding, initiation and terminations dates, success of data
collection, and any presentations and publications derived from the ancillary study. A written
progress report on ancillary studies will be made once a year to the LAS Subcommittee and to
the Steering Committee.

120
Chapter 11
Publication Policy
11.1 Data Analysis and Release of Results
The Coordinating Center will be responsible for the collection, storage, analysis, and
release of all ACCORD collaborative study data. Analyses from the main database will be
performed by the Coordinating Center. In the case of ancillary studies, the Coordinating Center
will review the data analyses of manuscripts using the ACCORD database. Distributed data
analysis may be necessary if proposed analyses require special expertise that does not exist at the
Coordinating Center, or if a particular analysis cannot be completed by the Coordinating Center
within a reasonable time period. In both of these situations, verification of final distributed
analyses will be performed at the Coordinating Center.
11.2 Manuscript Proposal Process
The review process of an ACCORD manuscript begins with the submission of a

manuscript proposal. Instructions on the format of publications can be obtained by accessing the
ACCORD website. The completed manuscript proposal will be submitted to the ACCORD
Publications and Presentations (P&P) Subcommittee for review and will be forwarded to the
Steering Committee for approval and Writing Group nominations. Nomination of additional
Writing Group members to the list of submitted names present on a proposal is the responsibility
of the Steering Committee. Each center that collects or processes the data used in a paper, and
each unit represented on the Steering Committee, may nominate a Writing Group representative.
Based on the nominations received from the centers, the P&P Subcommittee will select Writing
Group members and a Chairperson for each manuscript. Generally, a Writing Group will consist
of approximately five members with investigators from every CCN having the opportunity to
participate on papers. The P&P Subcommittee may change the composition of a Writing Group
that has failed to produce the required manuscript according to the schedule originally agreed
upon by the Group and the P&P Subcommittee.
A limited number of ACCORD manuscripts, such as the major design paper and the main
papers describing treatment effects on the major endpoints, will be authored by the ACCORD
Study Group with reference to all investigators in an appendix. For some major papers, named
authors may be suggested by the P&P Subcommittee with final determination by voting
members of the Steering Committee. Named authorship for other papers can be suggested by the
proposal’s originator and will include a limited number of named authors who comprise the
Writing Group. As allowed by the editorial policy of individual journals, an appendix describing
the structure of the ACCORD organization and containing the names of all the ACCORD
investigators will be included with publications.
The ACCORD Steering Committee will determine priorities for scheduling a start date
for a manuscript. A priority rating of 1 indicates highest priority and a 5 indicates the lowest

121
priority. The Writing Group Chairperson is responsible for all phases of manuscript preparation,
from conception through publication. Members of the Writing Group are responsible for
performance of tasks assigned by the Chairperson within the allotted time period. Each member
is expected to actively participate in the preparation of the manuscript. Selection of the journal
for initial submission of the manuscript is delegated to the Writing Group, with input from the
P&P Subcommittee and the Steering Committee.
11.3 Approval of Manuscripts, Abstracts, and Presentations
All manuscripts must be reviewed by the ACCORD P&P Subcommittee and the NHLBI
Project Office. Prior to submission for publication, manuscripts must be approved by the P&P
Subcommittee and Steering Committee. Manuscripts with NHLBI co-authors must also be
approved by NHLBI. The Committees will have 30 days to review and provide comments
regarding necessary revisions. After revision, a final copy of the manuscript with the cover letter
to the journal should be sent to the chair of the P&P Subcommittee and to the Coordinating
Center, in addition to all co-authors.
The ACCORD P&P Subcommittee will maintain a current list of all relevant meetings
and their deadlines for submission of abstracts. No abstract shall be submitted to any national or
international organization for consideration prior to review by the P&P Subcommittee and the
NHLBI Project Office. Additionally, prior to submission, abstracts must be approved by the P&P
Subcommittee and the Steering Committee. Abstracts of papers for presentations are expected to
be based on active manuscripts. Abstracts should be submitted to the P&P Subcommittee at least
two weeks prior to the abstract deadline and will be reviewed within 48 hours. If the P&P
Subcommittee review is favorable, the abstract will be sent simultaneously to the Steering
Committee and NHLBI Project Office for review. The Steering Committee and NHLBI will have
48 hours to review and to respond to the abstract. If an abstract is approved by the Steering
Committee, the Writing Group Chair will be notified by the P&P Subcommittee Chair and will
be given clearance to submit the abstract.
Presentations given at national or international scientific meetings or events sponsored by

industry need prior clearance by the P&P Subcommittee. Any ACCORD investigator who
receives a personal invitation to make a presentation should immediately notify the P&P
Subcommittee of the sponsor, date and topic of the presentation. The approval process for these
presentations will follow the same guidelines as specified for ACCORD abstracts. If information
is to be presented that is not based on previously approved reports, prior approval must be
granted by the P&P Subcommittee.
Presentations at local meetings of any previous published or presented ACCORD data do

not need prior clearance by the P&P Subcommittee. However, as with all presentations, the P&P
Subcommittee should be notified of these presentations.
A standard set of PowerPoint slides presenting the design and the rationale of trial will be
developed by the P&P Subcommittee and placed on the ACCORD website for downloading.
Presenters are encouraged to use these slides as part of any presentation.

122
11.4 Recruitment Materials
For purposes of participant recruitment, presentations will use approved materials that
will be maintained at the Coordinating Center. Any material to be distributed to the public for
recruitment, in addition to approved ACCORD recruitment material, must be approved by the
P&P Subcommittee and reviewed by the Project Office. A standard set of PowerPoint slides,
developed in collaboration with the ACCORD Recruitment and Retention Subcommittee, will be
downloadable from the ACCORD website.

123
Chapter 12
Quality Control
12.1 Introduction
An important feature of the ACCORD Trial is a concern for the integrity and high quality
of the study data. This concern is reflected in the detail provided in the protocol regarding initial
screening and recruitment of participants, data acquisition at baseline and follow-up visits,
reading and/or interpretation of the results, and their analysis and publication. There are two
primary purposes for quality control: to historically document the level of quality and to provide
feedback to the clinical and laboratory centers in order to maintain and improve the quality of the
study data over the course of the examination. The Measurement Procedures and Quality
Control Committee will establish guidelines for quality assurance and quality control while much
of the monitoring and analysis will be carried out by the Coordinating Center and the Network
Monitoring Centers. The Measurement Procedures and Quality Control Committee will report
any areas of concern to the Steering Committee for consideration.
This chapter outlines the type of quality assurance activities that will be conducted in the
ACCORD Trial. Two phrases are used. The first, quality assurance, is the collection of manuals
and procedures that will be in place to assure the integrity of the data. A subset of these
procedures is referred to as quality control, which describes the monitoring and analytic activities
that assess performance during data collection and its processing.
12.2 Manual of Procedures
As with any multicenter study, standardization of study procedures is very important in

the ACCORD Trial. A Manual of Procedures (MOP) will be developed that will include the
detailed descriptions of all trial procedures. This MOP will be used for training purposes and as
a reference for all study investigators and staff. The MOP will be an important aspect of efforts
to standardize important study procedures across clinical sites in the ACCORD Trial.
Attention has been placed on the standardization of important study procedures, including
the use of a central lab and ECG reading center, standard equipment in the clinics for HgbA1c
measurement and BP measurement, standard equipment for use by the participants for home
glucose monitoring, and standard examinations for visual acuity and foot involvement.
Furthermore, attention has been placed on the use of standard event definitions and event
validation procedures. Event tracking will include redundancy, such as, questions about multiple
types of events and searches of databases (e.g., HCFA, VA, NDI).
12.3 Study Forms and Data Entry Procedures
Attention will be placed on quality assurance concepts during the development of forms.
Forms will be printed with question-by-question instructions on the reverse of the preceding
page for easy reference. The forms will be translated into web-based data entry screens. These

124
screens will enable the incorporation of range and logical checks at the time of data entry. These
features will contribute to quality assurance.
12.4 Training
As with any multicenter study, training of staff and pilot testing of procedures will be
crucial to efforts to standardize procedures and assure data quality. We plan to have two different
training models available, central training and the train-the-trainer approach. In the central
training aspects of the ACCORD training effort, all relevant staff members from all clinical sites
will be convened in a single, centrally administered training session. This approach is cost-
efficient and contributes to uniformity of the training experience and thereby to uniformity of
data quality across sites. In the train-the-trainer aspect of the ACCORD training effort, clinical
center network (CCN) staff will provide training sessions to persons who were unable to attend
the central training session and to newly hired staff as turnover occurs. The Vanguard period will
serve as a test of feasibility of study procedures; participants recruited during the Vanguard will
be retained in the trial. The Steering Committee will organize a central training session prior to
the Vanguard and Main trial periods; furthermore, the CCNs will organize CCN training and
refresher training sessions, as needed.
12.5 Data Queries
The Coordinating Center will be responsible for data editing, which will include checks
for missing data and crosschecks for inconsistencies. The Coordinating Center will produce data
query requests that will be distributed directly to the appropriate clinical center. Clinical center
staff will be responsible for responding to the data queries in a timely manner. The Coordinating
Center will generate a summary indicating the number and types of queries generated by clinic
and network. This report will include the number of queries unresolved for more than 30 days.
This report will be shared with the Measurement Procedures and Quality Control Subcommittee
and Clinical Center Network Monitoring Center investigators and staff for quality control
purposes.
12.6 Quality Control Reports
The Measurement Procedures and Quality Control Subcommittee will develop quality
indicators that will be tracked in routine quality control reports in the ACCORD Trial. These
reports will serve both purposes described above, that is, for historical documentation of the
quality of the data collection process and for providing feedback to individual clinic sites. These
reports will be generated by the Coordinating Center and distributed to the CCN monitoring
centers. Investigators and staff at the CCN monitoring centers will be responsible for
disseminating this information to the appropriate investigators and staff at the clinics in their
networks. These reports will be used to inform discussions that will take place during regularly
scheduled telephone contacts and site visits.
Quality Control reports will focus on measures of process, impact and outcomes.
Examples of process measures that will be tracked for quality control purposes include:
1. late submission of data forms
2. the number of participants with missed or late visits

125
3. proportion of study participants off study medications

4. the number and dose of hypoglycemic and antihypertensive medications
5. adherence to the lipid-lowering assignment.
Examples of impact measures that will be tracked for quality control purposes include:
1. control of glycemia with HgbA1c results (central lab)
2. control of lipid and lipoprotein concentrations (central lab)
3. control of blood pressure with automated BP measurements
Examples of outcome measures that will be tracked for quality control purposes include:
1. Documentation for reported study events
2. Proportion of participants with ECG submitted to central ECG Reading Center
3. Proportion of participants with urine samples submitted for micro/macroalbuminuria
assessment
12.6.1 Deviations from protocol
Adherence to the study protocol is crucial to collection of high quality data and to the
internal validity of the trial. Thus, the Medical Intervention Subcommittee will define important
deviations from the intervention protocol for tracking purposes. A clinic-site-specific report
describing important protocol deviations will be disseminated to the CCN monitoring centers for
quality control purposes. Copies of these reports and a summary report describing important
protocol deviations on a study-wide basis will be shared with the Measurement Procedures and
Quality Control Subcommittee and the Steering Committee.
12.6.2 Monitoring the Clinical Centers in the Networks
Within each clinical center network (CCN), the physicians and nurse/study coordinators at the
CCN monitoring center will be responsible for monitoring activities and performance at the clinical
centers in the network. This team will coordinate the research activities of the study and maintain
effective communications with the clinical centers, the coordinating center, the project office, the
support centers (lab, ECG, and drug distribution center) as well as the other clinical center networks.
The primary role is to support the clinical centers in recruitment and maintenance of the respective
clinical center participant cohorts in the trial. The CCN team will assist in solving problems related to
quality control, protocol adherence, recruitment and retention for the clinical sites.
12.7 Site Initiation
The initiation of a clinical site to enroll and randomize participants into the ACCORD
trial will occur after a series of preliminary tasks are completed. These will include approval
from the local or network ethics board, completion of all required FDA forms and letters of
agreement, training of staff, development of clinical site recruitment and retention plan, and
receipt of study supplies and medication. The individual networks will provide appropriate
assistance to the clinical sites as needed with regard to site visits prior, at time of, or following

126
the clinical site’s initial enrollment and randomization of participants to ensure the sites are
comfortable with the process.
12.8 Site Visits
During recruitment and follow-up, personnel from the CCN monitoring centers will site
visits each clinical center in their network on a regular basis to promote communication, answer
questions, and ensure that study procedures are understood and conducted correctly. The site
visit program will provide a mechanism to encourage the effective and standardized delivery of
recruitment efforts, intervention programs, and the collection of appropriate and valid data within
each ACCORD clinic site. Site visits may also be performed if consistent departures from the
Protocol and MOP are detected. Refresher training and training of new staff may be done as
needed during these visits depending on the availability of staff. Site visits will provide the
opportunity for CCN monitoring center investigators and staff to review the operations of clinics
in their networks. As needed, representatives from the Coordinating Center, Project Office,
other CCNs and the Steering Committee may attend specific site visits. Usual activities at sites
visits may include reviews of clinic staffing levels and duties, discussions of clinic flow,
inspections of clinic space and records, review of study drug accountability, reviews of the
quality control reports described above, reviews of maintenance logs for important study
equipment, confirmation of participant’s consents, inclusion and exclusion criteria, and source
documentation; presence of regulatory documents; review of recruitment and adherence strategies and
trouble-shooting of problems. After each site visit, two types of reports will be produced. The
first will be a frank discussion at the end of the visit between the site visit team, the clinic PI and
key clinic staff. The site visit team will prepare written reports, detailing problems and offering
potential solutions regarding the activities of the site. A detailed report of the team’s observations
and recommendations will be sent to the PI of the site being reviewed, the PI of the Coordinating
Center and the Chair of the Measurement Procedures and Quality Control Subcommittee. The
Quality Control Committee will regularly review 10% of the site visit reports submitted and
recommend reporting changes as needed.
12.9 Laboratory and ECG Center Quality Control
The ACCORD Measurement Procedures and Quality Control Subcommittee will work
with the Central Laboratory and the ECG Reading Center to develop quality control procedures
to ensure high quality data. The results of quality control procedures performed at the Central
Laboratory and the ECG Reading Center will be reported on a regular basis to the Measurement
Procedures and Quality Control Subcommittee and the Steering Committee.

127
Chapter 13
Study Organization
13.1 Overview
The ACCORD organizational structures and responsibilities are similar to those of other,
large multicenter clinical trials sponsored by government or industry. Seven Clinical Center
Networks and a Coordinating Center are contracted by the National Heart, Lung and Blood
Institute to work together through the Steering Committee to successfully design and conduct
the trial (see Figure 13.1). In addition, there is a Central Chemistry Laboratory, an ECG
Reading Center, and a Drug Distribution Center. Scientific leadership is provided by the
Steering Committee (see Figure 13.2).
13.2 Clinical Center Networks and Clinical Sites
The 10,000 ACCORD participants will be recruited, randomized, treated, and followed
through a system of seven Clinical Center Networks (CCNs). Each CCN consists of a network
of collaborating Clinical Sites, which consist of medical facilities and/or individual practices that
will be involved in the evaluation, enrollment and treatment of patients in the trial. CCN
investigators will work with their Clinical Sites during the trial on issues related to recruitment,
compliance with protocol, and quality control. While these Clinical Sites will interact principally
through their CCNs, for matters such as data collection these sites will transmit their data directly
to the Coordinating Center and the other central units. Similarly, data queries will be sent
directly to the Clinical Sites, with copies to the appropriate CCN.
13.3 The Coordinating Center
The Coordinating Center (CC), with input from the ACCORD Steering Committee, will
be responsible for coordinating the protocol writing activities; developing and distributing the
Manual of Procedures; training trial personnel in the standardized protocol implementation and
data collection; providing rapid feedback to the CCNs and Core Laboratories on the quality of
data submitted and proposing corrections; collecting all trial data, including clinical outcomes;
and analyzing all data; and preparing reports for the Data and Safety Monitoring Board. The CC
will conduct yearly visits to each CCN to monitor and assure high performance throughout the
trial. CC investigators and staff are also active members of each of the Steering Committee
subcommittees.
During the recruitment phases of the trial, the CC will be responsible for monitoring
patient recruitment and will provide weekly reports to the CCNs, the Chair and Vice-chair of the
Steering Committee, the NHLBI Project Office, Core Laboratories, and the Drug Distribution
Center. Included in the reports will be measures of progress in recruiting women and minorities.
These weekly reports will assist the Recruitment and Retention Subcommittee with evaluating

128
and correcting recruitment problems. The CC will also develop (with the assistance from the
Steering Committee) criteria for the certification of new Clinical Sites.
13.4 ECG Reading Center, Central Chemistry Laboratory,

and the Drug Distribution Center
The ECG Reading Center and the Central Chemistry Laboratory provide central
interpretation of resting ECG, HbA1c and other blood measurements on trial participants. Each
core unit is responsible for development and distribution of specific measurement procedures,
timely data gathering, and analysis. The Drug Distribution Center (DDC) is responsible for the
development and implementation of plans for drug acquisition, packaging, labeling, and
dispensing according to the study protocol. This Center also assists with monitoring compliance
and provides data to the Coordinating Center for further analyses.
13.5 NHLBI Project Office and Other Government Representatives
ACCORD is sponsored by the National Heart, Lung, and Blood Institute (NHLBI). The
NHLBI Project Office is responsible for the administration and monitoring of the trial.
Representatives from this Office participate in the scientific, general organizational, and fiscal
management of the trial. Statistical consulting is provided by NHLBI statisticians. The National
Institute of Diabetes and Digestive and Kidney Disease (NIDDK) and the Centers for Disease
Control and Prevention (CDC) are co-sponsors of ACCORD. In addition to NHLBI personnel,
representatives from these agencies actively participate as scientists in the Steering Committee.
13.6 The ACCORD Steering Committee, Committee of Investigators,

Subcommittees of the Steering Committee, and Executive Committee
The ACCORD Steering Committee provides the overall leadership for the study and
establishes scientific and administrative policy. It is composed of the Principal Investigators
from the seven Clinical Center Networks, the Principal Investigator from the Coordinating
Center, the NHLBI Project Officer, the Chairs of the three major intervention working groups
(glycemia, lipid, and blood pressure), the Steering Committee Chair, and the Steering Committee
Vice-Chair. This committee oversees the overall conduct of the trial throughout all Phases. This
committee, along with the Committee of Investigators, developed the trial design, prepared the
final protocol, and approved the study forms and manual of operations. During the data
collection phases of the trial, this committee oversees data collection practices and procedures to
identify and correct deficiencies. The committee will also consider and adopt changes in the
study protocol or procedures as necessary during the course of the trial.
The Steering Committee is chaired by the Steering Committee Chair, who will serve as
the senior executive officer of the investigative group. A Vice-Chair will assist the Chair with his
responsibilities. Voting Steering Committee members are the Principal Investigators from the
seven Clinical Center Networks, the Principal Investigator from the Coordinating Center, and the
NHLBI Project Officer. If a CCN or the CC PI cannot make a meeting at which a vote is taken,
then the Co-Principal Investigator may vote (with the understanding that the PI is assured that
the Co-PI is fully informed about the issue). The Steering Committee Chair, or Vice-Chair in his

129
absence, votes only to make or break a tie. Co-investigators, CC staff, NIH staff, consultants,
and opinion leaders may also be invited to attend meetings.
All other ACCORD investigators, co-investigators, and senior staff represent the
ACCORD Committee of Investigators.
There are nine standing subcommittees of the Steering Committee (Figure 13.2). The
charges to these subcommittees are presented in Appendix II. These are the Design and Analysis
Subcommittee, the Medical Interventions Subcommittee, the Recruitment and Retention
Subcommittee, the Measurement Procedures and Quality Control Subcommittee, the Morbidity
and Mortality Subcommittee, the Publications and Presentations Subcommittee, the Health-
related Quality of Life/Cost-effectiveness Subcommittee, the Laboratory and Ancillary Studies
Subcommittee, and the Operations Subcommittee.
An Executive Committee acts as the operational arm of the Steering Committee and
makes decisions on behalf of the Steering Committee on day-to-day operational issues requiring
immediate action. It makes recommendations to the Steering Committee for consideration. It
will meet at least biweekly by conference call to review trial progress and any study issues that
may arise. This committee will also develop time lines for the accomplishment of tasks and will
develop Steering Committee Meeting agendas.
The members of the Executive Committee include the Steering Committee Chair,
Steering Committee Vice-chair, Coordinating Center personnel, Project Office personnel, one
CCN PI, and the Chair of the Operations Committee. The CCN Representative PI is appointed
annually so that each PI has the opportunity to serve. The Chair of the Operations Committee,
who is a CCN Coordinator, is also appointed annually.
13.7 The Protocol Review Committee
An independent Protocol Review Committee (PRC) reviewed the original Vanguard

Protocol. Members of the Committee, appointed by the Director of NHLBI, were senior experts
in the areas of cardiovascular medicine, diabetes, biostatistics, and bioethics. The Study Chair,
the Vice-Chair, the senior staff of the Coordinating Center, the CCN PI’s, and representatives
from the NHLBI and other sponsoring Federal agencies and Institutes participated in PRC
meetings as non-voting members. The Protocol Review Committee met at the end of the
protocol development phase of the trial and reported to the NHLBI regarding the scientific merit
and feasibility of the trial, and made a recommendation to the NHLBI that the protocol be
approved.
13.8 The Data and Safety Monitoring Board
During the recruitment and follow-up phases of the trial, an independent Data and Safety
Monitoring Board (DSMB) will monitor data and oversee patient safety. Members of the Board,
appointed by the Director of NHLBI, are senior experts in the areas of cardiovascular medicine,
diabetes, biostatistics, and bioethics. The Study Chair, the Vice-Chair, the Principal Investigator
and senior staff of the Coordinating Center, and representatives from the NHLBI and other

130
sponsoring Federal agencies and Institutes will participate in DSMB meetings as non-voting
members. The DSMB will meet at least once a year to monitor safety, to advise the NHLBI
about study progress, including contractor performance, and to make recommendations to the
NHLBI regarding study continuation and protocol changes. In addition, the Coordinating Center
will provide data to the DSMB Chair at his/her request at regular intervals to ensure early
identification of any major adverse outcomes of therapy.
13.9 The Hypoglycemia Monitoring Committee
A Hypoglycemia Monitoring Committee (HMC) will provide external oversight and

monitoring for cases of serious hypoglycemia events in ACCORD and provide feedback to
enhance participant safety. All members of the HMC are endocrinologists with experience in the
care of diabetes. The HMC is considered a subcommittee of the ACCORD Data and Safety
Monitoring Board (DSMB) and, as such, is appointed by the NHLBI and reports to the DSMB
and to the NHLBI. The committee complements DSMB monitoring by providing more frequent
monitoring than provided by the DSMB and focusing on individual cases. The committee can
recommend collection of additional information to ensure that data collected are appropriate for
the monitoring needs and can recommend establishment of additional processes within the study
to ensure that procedures are in place to enhance participant safety. The committee can
recommend to NHLBI that the ACCORD DSMB review specific issues, or convene by
conference call or in person, at times other than regularly scheduled DSMB meetings if the need
arises.
13.10 Role of Industry
Industry may contribute resources to the study and will be acknowledged appropriately.
However, the scientific decisions and governance of the trial will be determined by the Steering
Committee, as per NHLBI Policy.
13.11 Conflict of Interest Policy
The ACCORD investigators have established a policy regarding Conflict of Interest.

This policy is presented in Appendix III. This policy was developed to meet two goals. First,
the investigators wished to maintain the confidence that advice was being given, and decisions
made, in as unbiased and fully informed manner as possible. Second, the investigators wished
that the processes and results of the trial would meet public standards of conduct.

131
Figure 13.1: ACCORD Organizational Chart

Protocol Review Data and Safety
Committee NHLBI Monitoring Board
Clinical Center Networks
Northeast Canada Southeast Ohio-Michigan

Coordinating Drug
Center Distr Cntr
VA Network Minnesota Western
Central Central
Clinical Clinical Clinical Clinical Clinical Clinical Clinical ECG Lab Chem Lab
Sites Sites Sites Sites Sites Sites Sites

132
Figure 13.2: ACCORD Committees

Executive Steering
Committee Committee
Design & Recruitment & Measurement Morbidity & Lab &

Analysis Retention Procedures & Mortality Ancillary
Subcommittee Subcommittee Quality Control Subcommittee Studies
Subcommittee Subcommittee
Health Related Medical Publications &

Quality of Life/ Operations
Subcommittee Interventions Presentations
Cost-Effectiveness Subcommittee Subcommittee
Subcommittee
Glycemia Blood Lipid Lifestyle/

Working Pressure Working Background
Group Working Group Working
Group Group

133
Chapter 14
References
Abraira C, Colwell JA, Nuttall F, Sawin CT, Henderson W, Comstock JP, Emanuele NV, Levin
SR, Pacold I, Lee HS. Cardiovascular events and correlates in the veterans affairs diabetes
feasibility trial. Arch Intern Med 1997;157:181-188.
Abraira C, Colwell JA, Nuttall FQ, Sawin CT, Johnson-Nagel N, Comstock JP, Emanuele NV,
Levin SR, Henderson W, Lee HS. Veterans Affairs Cooperative Study on glycemic control and
complications in type II diabetes (VACSDM). Results of the feasibility trial. Diabetes Care
1995;18:1113-1123.
ACE Inhibitor Myocardial Infarction Collaborative Group. Indications for ACE inhibitors in the
early treatment of acute myocardial infarction: Systematic overview of individual data from
100,000 patients in randomized trials. Circulation; 1998;97:2202-2212.
ACSM (American College of Sports Medicine). Exercise for Patients with Coronary Artery
Disease. Med Sci Sports Exerc 1994;26:i -v.
ACSM & ADA (American College of Sports Medicine and American Diabetes Association)
Joint Position Statement. Diabetes mellitus and exercise. Med Sci Sports Exerc 1997;29(12):i-
vi.
ADA (American Diabetes Association). Nutrition recommendations and principles for people
with diabetes mellitus (position statement). Diabetes Care 2000;23(Suppl. 1): S43-S46. (2000a)
ADA (American Diabetes Association). Diabetes mellitus and exercise (position statement).
Diabetes Care 2000;23(Suppl. 1): S50-S54. (2000b).
ADA (American Diabetes Association). Smoking and diabetes (position statement). Diabetes
Care 2000;23(Suppl. 1): S63-S64. (2000c).
ADA (American Diabetes Association). Aspirin therapy in diabetes (position statement).

Diabetes Care 2000;23(Suppl. 1):S61-S62. (2000d)
ADA (American Diabetes Association). Management of dyslipidemia in adults with diabetes

(position statement). Diabetes Care 2000;23(Suppl 1):57-60. (2000e).
ADA (American Diabetes Association). Standards of medical care for patients with diabetes
mellitus (position statement). Diabetes Care 2000;23(Suppl. 1):S32-S42. (2000f).
ADA (American Diabetes Association). Diabetic retinopathy (position statement). Diabetes

Care 2000;23(Suppl. 1):S73-S76. (2000g).

134
ADA (American Diabetes Association). Preventive foot care in people with diabetes (position
statement). Diabetes Care 2000;23(Suppl. 1):S55-S56. (2000h)
ADA (American Diabetes Association). Immunization and the prevention of influenza and
pneumococcal disease in people with diabetes (position statement). Diabetes Care
2000;23(Suppl. 1):S91-S93. (2000i).
ADA (American Diabetes Association). Management of dyslipidemia in adults with diabetes.

Diabetes Care 23:Suppl 1: S57-S60, 2000. (2000j)
ADA (American Diabetes Association). National standards for diabetes self-management

education programs and American Diabetes Association review criteria. Diabetes Care
1999;22(Suppl 1):S111-114.
Advisory Committee on Immunization Practices (ACIP): Prevention and control of influenza:

recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR
1997;46 (No. RR-9):1–25.
Agewall S, Wikstrand J, Ljungman S, Fagerberg B, Risk Factor Intervention Study Group.

Usefulness of microalbuminuria in predicting cardiovascular mortality in treated hypertensive
men with and without diabetes mellitus. Am J Cardiol 1997;80:164-169.
Albert PS. Longitudinal data analysis (repeated measures) in clinical trials. Stat Med
1999:18:1707-1732.
Albrink MJ, Krauss RM, Lindgren FT, Von der Groeben VD, Wood PDS. Intercorrelations
among high density lipoproteins, obesity, and triglycerides in a normal population. Lipids
1980;15:668-678.
ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group. Major
cardiovascular events in hypertensive patients randomized to doxazosin vs chlorthalidone: the
Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).
JAMA 2000;283:1967-1975.
Andersson DKG, Svardsudd K. Long-term glycemic control relates to mortality in type II

diabetes. Diabetes Care 1995;18:534-543.
Antiplatelet Trialists’ Collaboration. Collaborative overview of randomized trials of antiplatelet

therapy I: prevention of death, myocardial infarction, and stroke by prolonged antiplatelet
therapy in various categories of patients. BMJ 1994;308:1-72, 81-106.
Antman EM, Tanasijevic MJ, Thompson B, Schactman M, McCabe CH, Cannon CP, Gischer
GA, Fung AY, Thompson C, Wybenga D, Braunwald E. Cardiac-specific troponin I levels to
predict the risk of mortality in patients with acute coronary syndromes. N Engl J Med.
1996;335:1342-1349.

135
Armitage,P., McPherson, C.K. and Rowe, B.C. Repeated significance tests on accumulating
data. J Roy Stat Soc, Series A. 1969;132:235-244.
Ascherio A, Katan MB, Zock PL, Stampfer MJ, Willett WC. Trans fatty acids and coronary heart
disease. N Engl J Med 1999;340(25):1994-8.
Ascherio A, Willett WC. Health effects of trans fatty acids. Am J Clin Nutr 1997;66(4
Suppl):1006S-1010S.
Bagdade JD, Lane JT, Subbaiah PV, Otto ME , Ritter MC. Accelerated cholesteryl ester transfer
in noninsulin-dependent diabetes mellitus. Atherosclerosis 1993;104:69-77.
Balkau B, Shipley M, Jarrett RJ, Pyorala K, Pyorala M, Forhan A, Eschwege E. High blood
glucose concenttration is a risk factor for mortality in middle-aged nondiabetic men. 20-year
follow-up in the Whitehall Study, the Paris Prospective Study, and the Helsinki Policemen
Study. Diabetes Care 1998;21:360-367.
Barrett-Connor E, Cohn BA, Wingard DL, Edelstein SL. Why is diabetes mellitus a stronger
risk factor for fatal ischemic heart disease in women than in men? The Rancho Bernardo Study.
JAMA 1991;265:627-631.
Bauer, P. A sequential elimination procedure for choosing the best population(s) based on
multiple testing. J Stat Plan and Infer 1989;21:245-252.
Behar S, Boyko EJ, Reicher-Reiss H, Goldbourt U. Ten year survival after acute myocardial
infarction: comparison of patients with and without diabetes. Am Heart J 1997;133:290-296.
Berger M, Muhlhauser I. Diabetes care and patient-oriented outcomes. JAMA 1999;281:1676-

1678.
Betensky, R.A. Sequential analysis of censored survival data from three treatment groups.
Biometrics 1997;53:794-806.
Bjornholt JV, Erikssen G, Aaser E, Sandvik L, Nitter-Hauge S, Jervell J, Erikssen J, Thaulow E.

Fasting blood glucose: An underestimated risk factor for cardiovascular death. Results from a
22-year follow-up of healthy nondiabetic men. Diabetes Care 1999;22:45-49.
Bonora E, Kiechl S, Willeit J, Oberhollenzer F, Effer G, Targher G, Alberiche M, Bonadonna

RC, Muggeo M. Prevalence of insulin resistance in metabolic disorders. The Bruneck Study.
Diabetes 1998;47:1643-1649.
Boyle MH, Furlong W, Feeny D, Torrance GW, Hatcher J. Reliability of the Health Utilities
Index--Mark III used in the 1991 Cycle 6 Canadian General Social Survey Health Questionnaire.
Quality of Life Research 1995;4:249-257.

136
Briggs AH, Gray AM. Power and sample size calculation for stochastic cost-effectiveness
analysis. Medical Decision Making 1998;18 (supplement): S81-S92.
Briggs AH, Tambour M. The design and analysis of stochastic cost-effectiveness studies for the
evaluation of health care interventions. Stockholm School of Economics, Working Paper Series
in Economics and Finance 1998;234: 1-22.
Brinton EA, Eisenberg S, Breslow JL. Increased apo A-I and apo A-II fractional catabolic rate in
patients with low high density lipoprotein-cholesterol levels with or without
hypertriglyceridemia. J Clin Invest 1991;87:536-544.
Campeau L, Knatterud GL, Domanski M, Hunninghake DB, White CW, Geller NL, Rosenberg
Y, for The Post Coronary Artery Bypass Graft Trial Investigators. The effect of aggressive
lowering of low-density lipoprotein cholesterol levels and low-dose anticoagulation on
obstructive changes in saphenous-vein coronary-artery bypass grafts. N Engl J Med.
1997;336:153-162.
Carew BD, Ahn SA, Boichot HD, Dierenfeldt BJ, Dolan NA, Edens TR, Weiner DH, Probstfield
JL. Recruitment strategies in the studies of left ventricular dysfunction (SOLVD): strategies for
screening and enrollment in two concurrent but separate trials. The SOLVD Investigators.
Control Clin Trials 1992;13:325-338.
Carney RM, Rich MW, Freedland KE, Saini J, te Velde A, Simoene C, Clark K. Major
depressive disorder predicts cardiac events in patients with coronary artery disease. Psychosom
Med 1988;50:627-633.
Colwell JA: Aspirin therapy in diabetes (Technical Review). Diabetes Care 1997;20:1767–1771.
Cosgrove N, Borhani NO, Bailey G, Borhani P, Levin J, Hoffmeier M, Krieger S, Lovato LC,
Petrovich H, Vogt T, Wilson AC, Breeson V, Probstfield JL. Mass mailing and staff experience
in a total recruitment program for a clinical trial: the SHEP experience. Systolic Hypertension in
the Elderly Program. Cooperative Research Group. Control Clin Trials 1999;20:133-148.
Coutinho M, Gerstein HC, Wang Y, Yusuf S. The relationship between glucose and incident
cardiovascular events. A metaregression analysis of published data from 20 studies of 95,783
individuals followed for 12.4 years. Diabetes Care 1999;22:233-240.
Cowie CC, Harris MI. Physical and metabolic characteristics of persons with diabetes. In: Harris
MI, Cowie CC, Stern MS, Boyko EJ, Reiber GE, Bennett PH, editors. Diabetes in America. NIH
Publication No. 95-1468. 2nd ed. National Institutes of Health, 1995:117-164.
Cox DR, Oakes D. Analysis of survival data. Chapman and Hall: New York. 1984.
Cox DR. Regression models and life tables. J Royal Stat Soc 1972;34B:187-197.

137
Curb JD, Pressel SL, Cutler JA, Savage PJ, Applegate WB, Black H, Camel G, Davis BR, Frost
PH, Gonzalez N, Guthrie G, Oberman A, Rutan GH, Stamler J. Effect of diuretic-based
antihypertensive treatment on cardiovascular disease risk in older diabetic patients with isolated
systolic hypertension. JAMA 1996;276:1886-1892.
Cutler JA. Calcium-Channel Blockers for Hypertension -- Uncertainty Continues. N Engl J

Med 1998;338:679-681.
DAIS (Diabetes Atherosclerosis Intervention Study) Investigators. Effect of fenofibrate on

progression of coronary-artery disease in type 2 diabetes: the Diabetes Atherosclerosis
Intervention Study, a randomised study. Lancet 2001;357:905-910.
Damsgaard EM, Froland A, Jorgensen OD, Mogensen CE. Eight to nine year mortality in
known non-insulin dependent diabetes and controls. Kid Int 1992;42:731-735.
Davidson PC, Albrink MJ. Insulin resistance in hyperglyceridemia. Metabolism 1965;14:1059-

1070.
Davis BR, Cutler JA, Gordon DJ, Furberg CD, Wright JT, Cushman WC, Grimm RH, LaRosa, J,
Whelton PK, Perry HM, Alderman MH, Ford CE, Oparil S, Francis C, Proschan M, Pressel S,
Black HR, Hawkins CM for the ALLHAT Research Group. Rationale and Design for the
Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).
Amer J Hypertens 1996;9:342-360.
DeFronzo RA. Pharmacologic therapy for type 2 diabetes mellitus. Ann Intern Med
1999;131:281-303.
DeMets DL, Gail MH. Use of logrank tests and group sequential methods at fixed calendar
times. Biometrics 1985;41:1039-1044.
Domanski MJ, Exner DV, Borkowf CB, Geller NL, Rosenberg Y, Pfeffer MA. Effect of
angiotensin converting enzyme inhibition on sudden cardiac death in patients following acute
myocardial infarction. A meta-analysis of randomized clinical trials. JACC 1999;33:598-604.
Downs JR, Clearfield M, Weis S, Whitney E, Shapiro DR, Beere PA, Langendorfer A, Stein EA,
Kruyer W, Gotto AM. Primary prevention of acute coronary events with lovastatin in men and
women with average cholesterol levels. Results of AFCAPS/TexCAPS. JAMA 1998;279:1615-
1622.
Drummond, Michael F., Bernie O'Brien, Greg Stoddart, and George W. Torrance. Methods for
the Economic Evaluation of Health Care Programmes. Second Edition. Oxford: Oxford
University Press, 1997.
Ellen RL, McPherson R. Long-term efficiency and safety of fenofibrate and a statin in the
treatment of combined hyperlipidemia. Am J Cardiol 1998;81(4A):60B-65B.

138
Estacio RO, Jeffers BW, Gifford N, Schrier RW: The effect of blood pressure control on
diabetic microvascular complications in patients with hypertension and type 2 diabetes. Diabetes
Care 2000;23 (Suppl. 2):B54–B64.
Estacio RO, Jeffers BW, Hiatt WR, Biggerstaff SL, Gifford N, Schrier RW. The effect of
nisoldipine as compared with enalapril on cardiovascular outcomes in patients with
non-insulin-dependent diabetes and hypertension. N Engl J Med 1998;338:645-652.
ETDRS Investigators. Aspirin effects on mortality and morbidity in patients with diabetes
mellitus. Early treatment of diabetic retinopathy study report 14. JAMA 1992;266:1292-1300.
Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the
Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care
1997;20:1183-1197.
Expert Panel on Detection Evaluation and Treatment of high blood cholesterol in adults.
Summary of the second report of the National Cholesterol Education Program Expert Panel
(NCEP) on detection, evaluation, and treatment of high blood cholesterol in adults (Adult
Treatment Panel II). JAMA 1993;269:3015-3023.
Expert Panel on Detection, Evaluation, and Treatment of high blood cholesterol in adults.
Executive summary of the third report of the National Cholesterol Education Program (NCEP)
Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult
Treatment Panel III). JAMA 2001;285:2486-2497.
FDA 2000. Rezulin to be withdrawn from the market. March 21, 2000. Accessed at
http://www.fda.gov/bbs/topics/NEWS/NEW00721.html).
Feingold KR, Grunfeld C, Pang M, Doerrler W, Krauss RM. LDL subclass phenotypes and
triglyceride metabolism in non-insulin-dependent diabetes. Arterio Thrombo 1992;2:1496-1502.
Follman DA, Prochan MA, Geller NL. Monitoring pairwise comparisons in multiarmed clinical
trials. Biometrics 1994;50:325-336.
Frasure-Smith N, Lesperance F, Talajic M. Depression following myocardial infarction, JAMA

1993;270:1819-1825.
Frasure-Smith N, Lesperance F, Talajic M. The impact of negative emotions on prognosis

following myocardial infarction: is it more than depression? Health Psychology 1995;14:388-
398.
Frick MH, Elo O, Haapa K, Heinonen OP, Heinsalmi P, Helo P, Huttunen JK, Kaitaniemi P,
Koskinen P, Manninen V, et al. Helsinki Heart Study: primary-prevention trial with gemfibrozil
in middle-aged men with dyslipidemia. Safety of treatment, changes in risk factors, and
incidence of coronary heart disease. N Engl J Med 1987;317:1237-1245.

139
Fu CC, Chang CJ, Tseng CH, Chen MS, Kao CS, Wu TJ, Wu HP, Chiang LN, Chen CJ, Tai TY.
Development of macrovascular diseases in NIDDM patients in northern Taiwan. A 4-yr follow-
up study. Diabetes Care 1993;16:137-143.
Fuller JH, Shipley MJ, Rose G, Jarrett RJ, Keen H. Mortality from coronary heart disease and
stroke in relation to degree of glycemia: the Whitehall study. Br Med J Clin Research Ed
1983;287:867-870.
Gall M-A, Borch-Johnsen K, Hougaard P, Nielsen FS, Parving H-H. Albuminuria and poor
glycemic control predict mortality in NIDDM. Diabetes 1995;44:1303-1309.
Garg R, Yusuf S. Overview of randomized trials of angiotensin-converting enzyme inhibitors on

mortality and morbidity in patients with heart failure. Collaborative Group on ACE Inhibitor
Trials JAMA 1995;273:1450-1456.
Garvey WT, Olefsky JM, Griffin J, Hamman RF, Kolterman OG. The effect of insulin treatment
on insulin secretion and insulin action in type II diabetes mellitus. Diabetes 1985;34:222-234.
Gavard JA, Lustman PJ, Clouse RE. Prevalence of depression in adults with diabetes. An
epidemiological evaluation. Diabetes Care 1993;16:1167-1178.
Genuth S. Exogenous insulin admininstration and cardiovascular risk in NIDDM and IDDM.
Ann Intern Med 1996;124:104-109.
Gerstein HC, Pais P, Pogue J, Yusuf S. Relationship of glucose and insulin levels to the risk of
myocardial infarction: a case-control study. J Am Coll Cardiol 1999;33:612-619.
Gerstein HC, Yusuf S. Dysglycaemia and risk of cardiovascular disease. Lancet 1996;347:949-
950.
Gerstein HC. Glucose: a continuous risk factor for cardiovascular disease. Diabetic Med
1997;14 (Suppl 3):S25-S31.
Ginsberg H, Grundy SM. Effect of caloric restriction on very low density lipoprotein triglyceride
metabolism in subjects with diabetes mellitus. Diabetologia 1982;23:421-425.
Ginsberg HN. Lipoprotein physiology in nondiabetic and diabetes states. Diabetes Care
1991;14:839-855.
Glick HA, Polsky D, Willke RJ, and Schulman KA. A comparison of preference assessment
instruments used in a clinical trial: Responses to the visual analog scale from the EuroQol EQ-
5D and the Health Utilities Index. Medical Decision Making 1999;19:265-275.
Golay A, Zech L, Shi M-Z. High density lipoprotein (HDL) metabolism in noninsulin-
dependent diabetes mellitus: measurement of HDL turnover using tritiated HDL. J Clin
Endocrinol Metab 1987;65:512-518.

140
Goldberg RB, Mellies MJ, Sacks FM, Moye LA, Howard BV, Howard WJ, Davis BR, Cole TG,
Pfeffer MA, Braunwald E. Cardiovascular events and their reduction with pravastatin in diabetic
and glucose-intolerant myocardial infarction survivors with average cholesterol levels: subgroup
analyses in the cholesterol and recurrent events (CARE) trial. The Care Investigators. Circulation
1998;98:2513-2519.
Goldbourt U, Yaari S, Medalie JH. Factors predictive of long-term coronary heart disease
mortality among 10059 male Israeli civil servants and municipal employees. A 23 year mortality
follow-up in the Israeli Ischemic Heart Disease Study. Cardiology 1993;82:100-121.
Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR. Diabetes blood lipids and the
role of obesity in coronary heart disease risk for women. The Framingham study. Ann Int Med
1977;87:393-397.
Grundy SM, Benjamin IJ, Burke GL, Chait A, Eckel RH, Howard BV, Mitch W, Smith SC,
Sowers JR. Diabetes and cardiovascular disease. A statement for healthcare professionals from
the American Heart Association. Circulation 1999;100:1134-1146.
Hadden DR, Patterson CC, Atkinson AB, Kennedy L, Bell PM, McCance DR, Weaver JA.
Macrovascular disease and hyperglycaemia: 10-year survival analysis in type 2 diabetes mellitus:
the Belfast Diet Study. Diabet Med 1997;14:663-672.
Haffner SM, Alexander CM, Cook TJ, Boccuzzi SJ, Musliner TA, Pedersen TR, Kjekshus J,
Pyorala K. Reduced coronary events in simvastatin-treated patients with coronary heart disease
and diabetes or impaired fasting glucose levels: subgroup analyses in the Scandinavian
Simvastatin Survival Study. Arch Intern Med 1999;159:2661-2667.
Haffner SM. The importance of hyperglycemia in the nonfasting state to the development of
cardiovascular disease. Endocrine Reviews 1998;19:583-592.
Haffner SM. Lehto S. Ronnemaa T. Pyorala K. Laakso M. Mortality from coronary heart disease
in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial
infarction. N Engl J Med 1998;339:229-234.
Haffner SM. Diabetes, hyperlidemia, and coronary artery disease. Am J Cardiol 1999;83:17F-
21F.
Haire-Joshu D, Glasgow RE, Tibbs TL: Smoking and diabetes (Technical Review). Diabetes
Care 1999;22:1887-1898.
Hansson L, Zanchetti A, Carruthers SG, Dahlof B, Elmfeldt D, Julius S, Menard J, Rahn KH,
Wedel H, Westerling S. Effects of intensive blood-pressure lowering and low-dose aspirin in
patients with hypertension: principal results of the Hypertension Optimal Treatment (HOT)
randomised trial. HOT Study Group. Lancet 1998;351:1755-62.

141
Hansson L, Lindholm LH, Niskanen L, Lanke J, Hedner T, Niklason A, Luomanmaki K, Dahlof

B, deFaire U, Morlin C, Karlberg BE, Wester PO, Bjorck JE. Effect of angiotensin-converting-
enzyme inhibition compared with conventional therapy on cardiovascular morbidity and
mortality in hypertension: the Captopril Prevention Project (CAPPP) randomized trial. Lancet
1999;353:611-616 (1999a).
Hansson L, Lindholm LH, Ekbom T, Dahlof B, Lanke J, Schersten B, Wester PO, Hedner T,
deFaire U, for the STOP-Hypertension-2 study group. Randomized trial of old and new
antihypertensive drugs in elderly patients: cardiovascular mortality and morbidity in the Swedish
Trial in Old Patients with Hypertension-2 (STOP-2) study. Lancet 1999;354:1751-1756 (1999b).
Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol
lowering with simvastatin in 20,536 high risk individuals: a randomised placebo-controlled trial.
Lancet 2002;360:7-22.
Henry RR, Gumbiner B, Ditzler T, Wallace P, Lyon R, Glauber HS. Intensive conventional
insulin therapy for type II diabetes. Metabolic effects during a 6-mo outpatient trial. Diabetes
Care 1993;16:21-31.
Henry RR. Glucose control and insulin resistance in non-insulin-dependent diabetes mellitus.
Ann Intern Med 1996;124:97-103.
Hollenbeck CB, Chen Y-D, Greenfield MS, Lardinois CK, Reaven GM. Reduced plasma high
density lipoprotein-cholesterol concentrations need not increase when hyperglycemia is
controlled with insulin in noninsulin-dependent diabetes mellitus. J Clin Endocrinol Metab
1986;62:605-608.
HOPE (Heart Outcomes Prevention Evaluation Study) Investigators. Effects of ramipril on

cardiovascular and microvascular outcomes in people with diabetes mellitus: results of the
HOPE study and Micro-HOPE substudy. Lancet 2000;355:253-259.
Horowitz BS, Goldberg IJ, Merab J, Vanni T, Ramakrishnan R, Ginsberg HN. Increased
plasma and renal clearance of an exchangeable pool of apolipoprotein A-I in subjects with low
levels of high density lipoprotein cholesterol. J Clin Invest 1993;91:1743-1752.
Howard BV, Mayer-Davis EJ, Goff D, Zaccaro DJ, Laws A, Robbins DC, Saad MF, Selby J,
Hamman RF, Krauss RM, Haffner SM. Relationships between insulin resistance and lipoproteins
in nondiabetic African Americans, Hispanics, and Non-Hispanic Whites: The Insulin Resistance
Atherosclerosis Study. Metabolism 1998;47:1174-1179.
Howard BV, Reitman JS, Vasquez B, Zech L. Very-low density lipoprotein triglyceride
metabolism in non- insulin-dependent diabetes mellitus: relationship to plasma insulin and free
fatty acids. Diabetes 1983;32:271-276.
Hsu, J.C. and Edwards, D.G. Sequential multiple comparisons with the best. JASA
1983;78:958-964.

142
Hughes, M.D. Stopping guidelines for clinical trials with multiple treatments. Stat Med
1993;12:901-915.
Jennison, C. and Turnbull, B.W. Exact calculations for sequential t, chi-sq and F tests.
Biometrika 1991;78:133-141.
Johnson ES, Lanes SF, Wentworth CE, Satterfield MH, Abebe BL, Dicker LW. A
meteregression analyis of the dose-response effect of aspirin on stroke. Arch Intern Med
1999;159:1248-1253.
Joint National Committee on Prevention, Detection, Evaluation and Treatment of High Blood
Pressure. The sixth report of the Joint National Committee on Prevention, Detection, Evaluation,
and Treatment of High Blood Pressure. Arch Intern Med 1997;157:2413-2446.
Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood
Pressure. The Sixth Report of the Joint National Committee on Prevention, Detection,
Evaluation, and Treatment of High Blood Pressure (JNC VI). Arch Intern Med 1997;157:2413-
2446.
Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. 1980. New York:
Wiley.
Kannel WB, McGee DL. Diabetes and cardiovascular disease. The Framingham study. JAMA
1979;241:2035-2038.
Kaplan E, Meier P. Nonparametric estimation from incomplete observations. JASA 1958;8:699-

711.
Kissebah AH, Alfarsi S, Evans DJ, Adams PW. Integrated regulation of very-low-density
lipoprotein triglyceride and apolipoprotein-B kinetics in non-insulin- dependent diabetes
mellitus. Diabetes 1982;31:217-225.
Kuusisto J, Mykkanen L, Pyorala K, Laakso M. NIDDM and its metabolic control predict
coronary heart disease in elderly subjects. Diabetes 1994;43:960-967.
Laakso M, Sarlund H, Mykkanen L. Insulin resistance is associated with lipid and lipoprotein
abnormalities in subjects with varying degrees of glucose tolerance. Arteriosclerosis
1990;10:223-231.
Ladwig KH, Roll G, Breithardt G, Budde T, Borggrefe M. Post-infarction depression and

incomplete recovery 6 months after acute myocardial infarction. Lancet 1994;343:2-3.
Lamarche B, Uffelman KD, Carpentier AC, Lewis GF. Triglyceride enrichment of HDL
enhances in vivo metabolic clearance of HDL apo A-1 in healthy men. J Clin Invest
1999;103:1191-1199.

143
Lan, K.K.G. and DeMets, D.L. Discrete sequential boundaries for clinical trials. Biometrika
1983;70:659-663.
Lan, K.K.G. and DeMets, D.L. Group sequential procedures: calendar versus information time.
Stat Med 1989;8:1191-1198.
LaRosa JC, Grundy SM, Waters DD, Shear C, Barter P, Fruchart J-C, Gotto AM, Greten H,
Kastelein JJP, Shepherd J, Wenger NK, for the Treating to New Targets Investigators. Intensive
lipid lowering with atorvastatin in patients with stable coronary disease. New Engl J Med
2005;352:1425-1435.
Laws A, Hoen HM, Selby JV, Saad MF, Haffner SM, Howard BV. Differences in insulin
suppression of free fatty acid levels by gender and glucose tolerance status. Relation to plasma
triglyceride and apolipoprotein B concentrations. Insulin Resistance Atherosclerosis Study
(IRAS) Investigators. Artheroscler Thromb Vasc Biol 1997;17:64-71.
Lawson M, Gerstein HC, Tsui E, Zinman B. Effect of intensive therapy on early macrovascular
disease in young individuals with type 1 diabetes. A systematic review and meta-analysis.
Diabetes Care 1999;22:(Suppl.2)B35-B39
Le N-A, Gibson JC, Ginsberg HN. Independent regulation of plasma apolipoprotein CII and CIII
concentrations in very low density and high density lipoproteins: implications for the regulation
of the catabolism of these lipoproteins. J Lipid Res 1988;29:669-677.
Lee, J.W. and DeMets, D.L. Sequential comparison of changes with repeated measurements
data. JASA 1991;86:757-762.
Lehto S, Ronnemaa T, Haffner SM, Pyorala K, Kallio V, Laakso M. Dyslipidemia and

hyperglycemia predict coronary heart disease events in middle-aged patients with NIDDM.
Diabetes 1997;46:1354-1359.
Levey AS, Coresh J, Balk E, Kausz AT, Levin A, Steffes MW, Hogg RJ, Perrone RD, Lau J,
Eknoyan G. National Kidney Foundation practice guidelines for chronic kidney disease:
Evaluation, classification, and stratification. Ann Intern Med 2003;139:137-147.
LIPID (Long-Term Intervention with Pravastatin in Ischemic Disease) study group. Prevention
of cardiovascular events and death with pravastatin in patients with coronary heart disease and a
broad range of initial cholesterol levels. N Engl J Med 1998;339:1349-1357.
LRC-CPPT Investigators. The Lipid Research Clinics Coronary Primary Prevention Trial
resulds. JAMA 1984;251;351-364.
LRC-CPPT Investigators. Recruitment for clinical trials: the Lipid Research Clinics Coronary
Primary Prevention Trial experience. Its implications for future trials. Circulation. 1982;66:IV1-
78.

144
Lovato LC, Hill K, Hertert S, Hunninghake DB, Probstfield JL. Recruitment for controlled
clinical trials: literature summary and annotated bibliography. Control Clin Trials 1997;18:328-
352.
Lui, K.L. A simple generalization of the O’Brien and Fleming group sequential test procedure to
more than two treatment groups. Biometrics 1993;49:1216-1219.
Lustman PJ, Griffith LS, Clouse RE, Freedland KE, Eisen SA, Rubin EH, Carney RM, McGill
JB. Effects of nortriptyline on depression and glucose regulation in diabetes: results of a double-
blind, placebo-controlled trial. Psychosom Med 1997;59:241-250.
Lustman PJ, Griffith LS, Freedland KE, Kissel SS, Clouse RE. Cognitive behavior therapy for
depression in Type 2 diabetes mellitus. Ann Intern Med 1998;129:613-621.
Lustman PJ, Harper GW. Nonpsychiatric physicians' identification and treatment of depression
in patients with diabetes. Compr Psychiatry 1987;28:22-27.
Mahler DA; Froelicher VF; Miller NH, et al. Kenney WL, Humphrey RH, and Bryant CX,
editors. ACSM Guidelines for Exercise Testing and Prescription. 5 ed. Baltimore: Williams &
Wilkins;1995. 373p.
Mak KH, Moliterno DJ, Granger CB, Miller DP, White HD, Wilcox RG, Califf RM, Topol EJ.
Influence of diabetes mellitus on clinical outcome in the thrombolytic era of acute myocarial
infarction. GUSTO-1 investigators. Global Utilization of Streptokinase and Tissue Plasminogen
Activator for Occluded Coronary Arteries. J Am Coll Cardiol 1997;30:171-179.
Malmberg K, Ryden L, Efendic S, Herlitz J, Nicol P, Waldenstrom A, Wedel H, Welin I.

Randomized trial of insulin-glucose infusion followed by subcutaneous insulin treatment in
diabetic patients with acute myocardial infarction (DIGAMI study): Effects on mortality at 1
year. J Am Coll Cardiol 1995;26:57-65.
Mann JI. The role of nutritional modifications in the prevention of macrovascular complications
of diabetes. Diabetes 1997;46 Suppl 2:S125-30.
Manson JE, Coldlitz GA, Stampfer MJ, Willett WC, Krolewski AS, Rosner B, Arky RA, Speizer
FE, Hennekens CH. A prospective study of maturity-onset diabetes mellitus and risk of
coronary heart disease and stroke in women. Arch Int Med 1991;151:1141-1147.
Mayfield JA, Reiber GE, Sanders LJ, Janisse D, Pogach LM: Preventive foot care in people with
diabetes (Technical Review). Diabetes Care 1998;21:2161–2177.
McGill JB, Lustman PJ, Griffith LS, Freedland KE, Gavard JA, Clouse RE. Relationship of
depression to compliance with self-monitoring of blood glucose (Abstract). Diabetes
1992;41:A84.

145
Mittmann N, Trakas K, Risebrough N, Liu BA. Utility scores for chronic conditions in a
community-dwelling population. Pharmacoeconomics, 1999;15:369-376.
Morrish NJ, Stevens LK, Fuller JH, Keen H, Jarrett RJ. Incidence of macrovascular disease in
diabetes mellitus: the London follow-up to the WHO multinational study of vascular disease in
diabetics. Diabetologia 1991;34:584-589.
Mosca L, Grundy SM, Judelson D, King K, Limacher M, Oparil S, Pasternak R, Pearson TA,
Redberg RF, Smith SC, Winston M, Zinberg S. Guide to preventive cardiology for women.
Circulation 1999;99:2480-2484.
Moss SE, Klein R, Klein BEK, Meuer SM. The association of glycemia and cause-specific
mortality in a diabetic population. Arch Intern Med 1994;154:2473-2479.
National Cholesterol Education Program. Third report of the expert panel on detection,
evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III): full
report. Available at: http://nhlbi.nih.gov/guidelines/cholesterol/atp3_rpt.htm (accessed
September 25, 2002).
National High Blood Pressure Education Program. Working Group Report on Primary
Prevention of Hypertension. NIH Publ. No. 93-2669, 1993.
National Institutes of Health Obesity Education Initiative Expert Panel on the Identification,
Evaluation, and Treatment of Overwieght and Obesity in Adults. Clinical guidelines on the
identification, evaluation and treatment of overweight and obesity in adults. Executive Summary.
1-20. 1999.
Neil A, Hawkins M, Potok M, Thorogood M, Cohen DA, Mann J. A prospective population-

based study of microalbuminuria as a predictor of mortality in NIDDM. Diabetes Care
1993;16:996-1003.
Nichol KL, Margolis KL, Wuorenma J, Von Sternberg T. The efficacy and cost effectiveness of
vaccination against influenza among elderly persons living in the community. N Engl J Med
1994;331:778–784.
Nicoll A, Miller NE, Lewis B. High density lipoportein metabolism. In: Advances in Lipid
Research. AnonymousAcademic Press, New York, 1980;54-54.
Obesity Education Initiative Expert Panel. Clinical guidelines on the identification, evaluation,
and treatment of overweight and obesity in adults -- the evidence report. National Institutes of
Health. Obes Res 1998;6(Suppl 2):51S-209S.
O'Brien PC, Fleming TR. A multiple testing procedure for clinical trials. Biometrics
1979;35:549-556.

146
O’Brien JA, Shomphe LA, Kavanagh PL, Raggio G, Caro JJ. Direct medical costs of
complications resulting from type 2 diabetes in the U.S. Diabetes Care 1998;21:1122-1128.
Ohkubo Y, Kishikawa H, Araki E, Miyata T, Isami S, Motoyoshi S, Kojima Y, Furuyoshi N,

Shichiri M. Intensive insulin therapy prevents the progression of diabetic microvascular
complications in Japanese patients with non-insulin-dependent diabetes mellitus: a randomized
prospective 6-year study. Diab Res Clin Pract 1995;28:103-117.
Olefsky J, Reaven GM, Farquhar JW. Effects of weight reduction on obesity. Studies of lipid
and carbohydrate metabolism in normal and hyperlipoproteinemic subjects. J Clin Invest
1974;53:64-76.
Olefsky JM, Farquhar JW, Reaven GM. Reappraisal of the role of insulin in
hypertriglyceridemia. Am J Med 1974;57:551-560.
Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, Bouchard C, Buchner D, Ettinger W,
Heath GW, King AC, et al. Physical activity and public health: A recommendation from the
Centers for Disease Control and Prevention and the American College of Sports Medicine.
JAMA 1995;273;402-407.
Patrick, Donald L., and Pennifer Erickson. Health Status and Health Policy: Quality of Life in
Health Care Evaluation and Resource Allocation. New York: Oxford University Press, 1993.
Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto
J, Smith PG. Design and analysis of randomized clinical trials requiring prolonged observations
of each patient. II. Analysis and examples. Br J Cancer 1977;35:1-25.
Pierce LR, Wysowski DK, Gross TP. Myopathy and rhabdomyolysis associated with lovastatin-
gemfibrozil therapy. JAMA 1990;264:71-75.
Probstfield JL, Russell ML, Henske JC, Reardon R, Insull W, Jr. Dropouts from a clinical trial,
their recovery and characterization: A basis for dropout management and prevention. In:
Shumaker S, O'Kene J, Shron E, eds. Handbook of Health Behavior. New York, NY: Springer
Publishing Co. 1990:376-400.
Probstfield JL, Russell ML, Henske JC, Reardon RJ, Insull W, Jr. Successful program for
recovery of dropouts to a clinical trial. Am J Med 1986;80:777-784.
Probstfield JL. The clinical trial prerandomization compliance (adherence) screen. In: Cramer
JA, Spilkereds. Patient Compliance in Medical Practice and Clinical Trials. New York, NY:
Raven Press; 1991:323-334.
Proceedings of the Workshop on the Recruitment Experience in NHLBI- sponsored Clinical

Trials. April 14 and 15, 1986. Controlled Clin Trials 1987;8:1S-150S.

147
Prochan, M.A., Follman, D.A., and Geller, N.L. Monitoring multi-armed trials. Stat Med
1994;13:1441-1452.
Ravid M, Brosh D, Levi Z, Bar-Dayan Y, Ravid D, Rachmani R. Use of enalapril to attenuate

decline in renal function in normotensive, normoalbuminuric patients with type 2 diabetes
mellitus. A randomized, controlled trial. Ann Intern Med 1998;128:982-988.
Ravid M, Lang R, Rachmani R, Lishner M. Long-term renoproprotective effect of angiotensin-

converting enzyme inhibition in non-insulin-dependent diabetes mellitus. A 7-year follow-up
study. Arch Intern Med 1996;156:286-289.
Reaven GM, Chen Y-D, Jeppesen J, Maheux P , Krauss RM. Insulin resistance and
hyperinsulinemia in individuals with small, dense, low density lipoprotein particles. J Clin
Invest 1993;92:141-146.
Reaven GM, Lerner RL, Stern MP, Farquhar JW. Role of insulin in endogenous
hypertriglyceridemia. J Clin Invest 1967;46:1756-1767.
Reboussin, D.M. “Flexible monitoring of multi-arm clinical trials with repeated measurements”,
ASA 1994 Proceedings of the Biopharmaceutical Section, 1995.
Rizzo JA, Sindelar, JL. Linking health-related quality-of-life indicators to large national data
sets. PharmacoEconomics 1999;16:473-482.
Rubin RR, Peyrot M. Quality of Life in Diabetes. Diab Metab Res Rev 1999;15:205-208.
Rubins HB, Robins SJ, Collins D, Fye CL, Anderson JW, Elam MB, Faas FH, Linares E,
Schaefer EJ, Schectman G, Wilt TJ, Wittes J. Gemfibrozil for the secondary prevention of
coronary heart disease in men with low levels of high-density lipoprotein cholesterol. Veterans
Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N Engl J Med
1999;341:410-418.
Sacks FM, Braunwald E, Simes RJ, Packard C, Shepherd J, Furberg CD. Pretreatment lipid and
coronary event reduction with Pravastatin – The Prava-3 Pooling Project. Circulation 1999:100
(Supp I);I-739.
Sacks FM, Moye LA, Davis BR, Cole TG, Rouleau JL, Nash DT, Pfeffer MA, Braunwald E.
Relationship between plasma LDL concentrations during treatment with pravastatin and
recurrent coronary events in the Cholesterol and Recurrent Events trial. Circulation
1998;97:1446-1452.
Sacks FM, Pfeffer MA, Moye LA, Rouleau JL, Rutherford JD, Cole TG, Brown L, Warnica JW,
Arnold JMO, Wun C, Davis BR, Braunwald, E, for the Cholesterol and Recurrent Events Trial
Investigators. The effect of pravastatin on coronary events after myocardial infarction in patients
with average cholesterol levels. N Engl J Med 1996;335:1001-1009.

148
Samsa G, Edelman D, Rothman M, Williams GR, Lipscomb J, Matchar D. Determining

clinically important differences in health status measures: a general approach with illustration
using the Health Utilities Index Mark II. PharmacoEconomics 1999;15:141-155.
Savage PJ. Cardiovascular complications of diabetes mellitus. What we know and what we
need to know about their prevention. Ann Intern Med 1996;124(1 pt 2):123-126.
Scandinavian Simvastatin Survival Study Group. Randomised trial of cholesterol lowering in

4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S).
Lancet 1994;344:1383-1389.
Schrier RW, Estacio RO, Jeffers B. Appropriate Blood Pressure Control in NIDDM (ABCD)
Trial. Diabetologia 1996;39:1646-1654.
Shepherd J, Cobbe SM, Ford I, Isles CG, Lorimer AR, MacFarlane PW, McKillop JH, Packard
CJ. Prevention of coronary heart disease with pravastatin in men with hypercholesterolemia. N
Engl J Med 1995;333:1301-1307.
Sigurdsson G, Nicoll A, Lewis B. Metabolism of very low density lipoproteins in

hyperlipidemia: studies of apolipoprotein B kinetics in man. Eur J Clin Invest 1976;6:167-177.
Slud, E.V. and Wei, L.J. Two-sample repeated significance tests based on the modified
Wilcoxon statistic. JASA 1982;77:862-868.
Smith SA, Poland GA. The use of influenza and pneumococcal vaccines in people with diabetes
(Technical Review). Diabetes Care 2000;23:95-108.
Smits P, Thien T. Cardiovascular effects of sulphonylurea derivatives. Implications for the

treatment of NIDDM? Diabetologia 1995;38:116-121.
Spitzer RL, Kroenke K, Williams JBW, and the Patient Health Questionnaire Primary Care
Study Group. Validation and utility of a self-report version of PRIME-MD. JAMA
1999;282:1737-1744.
Stamler J, Vaccaro O, Neaton JD, Wentworth D. Diabetes, other risk factors, and 12-yr
cardiovascular mortality for men screened in the Multiple Risk Factor Intervention Trial.
Diabetes Care 1993;16:434-444.
Steering Committee of the Physicians’ Health Study Research Group. Final report of the aspirin
component of the ongoing Physicians’ Health Study. N Engl J Med 1989;321:129-135.
Stratton IM, Adler AI, Neil AW, Matthews DR, Manley SE, Cull CA, Hadden D, Turner RC,
Holman RR, on behalf of the UK Prospective Diabetes Study Group. Association of glycaemia
with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35):
prospective observational study. BMJ 2000;321:405-411

149
Tall AR. Plasma lipid transfer proteins. J Lipid Res 1986;27:361-367.
Taskinen M-R. Lipoprotein lipase in diabetes. Diabetes Metab Rev 1987;3:551-570.
Tatti P, Pahor M, Byington RP, Di Mauro P, Guarisco R, Strollo G, Strollo F. Outcome results
of the Fosinopril Versus Amlodipine Cardiovascular Events Randomized Trial (FACET) in
patients with hypertension and NIDDM. Diabetes Care 1998;21:597-603.
Testa MA, Simonson DC, Turner RR. Valuing quality of life and improvements in glycemic
control in people with type 2 diabetes mellitus. Diabetes Care 1998;21:C44-C52.
Testa MA, Simonson DC. Health economic benefits and quality of life during improved
glycemic control in patients with type 2 diabetes mellitus. JAMA 1998;280:1490-1496.
Tsiatis AA. Repeated significance testing for a general class of statistics used in censored
survival analysis. JASA 1982;77:855-861.
Tuomilehto J, Rastenyte D, Birkenhager WH, Thijs L, Autikainen R, Bulpitt CJ, Fletcher AE,
Forette F, Goldhaber A, Palatini P, Sarti C, Fagard R. Effects of calcium-channel blockade in
older patients with diabetes and systolic hypertension. Systolic Hypertension in Europe Trial.
N Engl J Med 1999;340:677-684.
U.S. Department of Agriculture and U.S. Department of Health and Human Services (USDA &
USDHHS). Nutrition and Your Health: Dietary Guidelines for Americans. USDA Home and
Garden Bulletin No. 232, 1995.
U.S. Department of Health and Human Services (USDHHS). Physical Activity and Health: A
Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services,
Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and
Health Promotion, 1996.
UKPDS (United Kingdom Prospective Diabetes Study Group). Quality of life in type 2 diabetic
patients is affected by complications but not by intensive policies to improve blood glucose or
blood pressure control (UKPDS 37). Diabetes Care 1999;22:1125-1136.
UKPDS (United Kingdom Prospective Diabetes Study). Effect of intensive blood glucose
control with metformin on complications in overweight patients with type 2 diabetes (UKPDS
34). Lancet 1998;352:854-865.
UKPDS (United Kingdom Prospective Diabetes Study). Intensive blood-glucose control with
sulphonylureas or insulin compared with conventional treatment and risk of complications in
patients with type 2 diabetes (UKPDS 33). Lancet 1998;352:837-853.
UKPDS (United Kingdom Prospective Diabetes Study Group). Cost effectiveness analysis of
improved blood pressure control in hypertensive patients with type 2 diabetes (UKPDS 40). BMJ
1998;317: 720-726.

150
UKPDS (United Kingdom Prospective Diabetes Study Group). Efficacy of atenolol and captopril
in reducing risk of macrovascular and microvascular complications in type 2 diabetes. (UKPDS
39) BMJ 1998;317:713-720.
UKPDS (United Kingdom Prospective Diabetes Study Group). Risk factors for coronary artery
disease in on-insulin dependent diabetes. (UKPDS 23) BMJ 1998;316:823-828.
UKPDS (United Kingdom Prospective Diabetes Study Group). Tight blood pressure control and
risk of macrovascular and microvascular complications in type 2 diabetes. (UKPDS 38) BMJ
1998;317:703-713.
Uusitupa M, Siitonen O, Voutilainen E, Aro A, Hersio K, Pyorala K, Penttila I, Ehnholm C.

Serum lipids and lipoproteins in newly diagnosed non-insulin- dependent (type II) diabetic
patients with special reference to factors influencing HDL-cholesterol and triglyceride levels.
Diabetes Care 1986;9:17-22.
Wei M, Gaskill SP, Haffner SM, Stern MP. Effects of diabetes and level of glycemia on all-
cause and cardiovascular mortality. The San Antonio Heart Study. Diabetes Care 1998;21:1167-
1172.
Wei, L.J., Su, J.Q. and Lachin, J.M. Interim analyses with repeated measurements in a
sequential clinical trial. Biometrika 1990;77:359-364.
Wingard DL, Barrett-Connor E. Heart Disease and Diabetes. In: Harris MI, Cowie CC, Stern
MS, Boyko EJ, Reiber GE, Bennett PH, editors. Diabetes in America. 2 ed. Washington:
National Institutes of Health, 1995:429-448.
WOSCOPS Study Group. Influence of pravastatin and plasma lipids on clinical events in the
West of Scotland Coronary Prevention Study. Circulation 1998;97:1440-1445.
Wu, M.C. and Lan, K.K.G. Sequential monitoring for comparison of changes in a response
variable in clinical studies. Biometrics 1992;48:765-779.
Yusuf S, Flather M, Pogue J, Hunt D, Varigos J, Piegas L, Avezum A, Anderson J, Keltai M,

Budaj A, Fox K, Ceremuzynski L. Variations between countries in invasive cardiac procedures
and outcomes in patients with suspected unstable angina or myocardial infarction without initial
ST elevation. OASIS (Organisation to Assess Strategies for Ischaemic Syndromes) Registry
Investigators [see comments]. Lancet 1998;352:507-514.
Yusuf S, Sleight P, Pogue J, Bosch J, Davies R, Dagenais G. Effects of an angiotensin-

converting-enzyme inhibitor, ramipril, on cardiovascular events in high-risk patients. The Heart
Outcomes Prevention Evaluation Study Investigators. N Engl J Med 2000;342:145-153.

151
Zuanetti G, Latini R, Maggioni AP, Franzosi M, Santoro L, Tognoni G, for the GISSI-3
Investigators. Effect of the ACE inhibitor lisinopril on mortality in diabetic patients with acute
myocardial infarction: Data from the GISSI-3 study. Circulation 1997;96:4239-4245.

152
Appendix I:
Model Informed Consent Document
(Consent Version Date: May 11, 2005)
ACTION TO CONTROL CARDIOVASCULAR RISK IN DIABETES (ACCORD)
Principal Investigator(s)__________________________________
You are invited to join in a research study called Action to Control Cardiovascular Risk in Diabetes
(ACCORD), which is sponsored by the National Heart, Lung and Blood Institute (part of the U.S. federal
government). The investigators listed above are in charge of the study. Other professional persons may
help them or act for them.
What are some general things you should know about research studies? Research studies are
designed to gain scientific knowledge that may help other people in the future. You may or may not
receive any direct benefit from participating. There may also be risks associated with participating in
research studies.
Your participation is voluntary. You may refuse to participate, or may withdraw your consent to
participate in any study at any time, and for any reason, without jeopardizing your future care at this
institution or your relationship with your doctor. You do not have to participate in research in order to
receive treatment.
Details about this study are discussed below. It is important that you understand this information so that
you can decide in a free and informed manner whether you want to participate. You will be given a copy
of this consent form. You are urged to ask the investigators named above, or staff members who may
assist them, any questions you have about this study at any time.
STUDY PURPOSE
What is the purpose of the study and how long will it last? Type 2 diabetes is very common in
North America. People with Type 2 diabetes have a higher chance of getting heart disease or stroke than
people without diabetes. The purpose of the ACCORD study is to determine the best approaches to lower
the risk of heart disease and stroke in people with Type 2 diabetes.
ACCORD will answer three research questions. In diabetes, the level of sugar in the blood is too high.
So the first question is to determine the effects of lowering blood sugar to a level below that normally
targeted in current clinical practice, compared with a level that is usually targeted. Many diabetic patients
have high blood pressure. So the second question is to determine the effects of lowering blood pressure
to a level below that normally targeted in current clinical practice, compared with a level usually targeted.
Many diabetic patients also have problems with their blood lipids (like cholesterol, fat-like materials in the
blood). So the third question is to determine the effects of treating several components of blood lipids
compared with treating only one component. Each of these questions is described in more detail below.
You are being invited to participate in ACCORD because you have Type 2 diabetes along with other
factors that increase your chance of having future heart disease and stroke, or you may already have had
heart disease or stroke. Your participation in the study will last until 2009. However, study results will be
reviewed regularly to see if the trial should be stopped earlier than this. Most participants will be in the
ACCORD study between 5 ½ and 8 ½ years.
The total number of participants will be about 10,000 from 77 clinics throughout the United States and
Canada. The study will involve approximately ___ patients at the ______ clinical site. ACCORD recruited
about 1,200 participants during the Vanguard (pilot) portion of the trial in 2001 and these participants
are still being treated and followed.

153
STUDY SUMMARY
What will happen if you take part in this study? Initial visits will be conducted to determine
whether you qualify for the study. These are called "screening" visits. Your medical history, blood
pressure, and past blood sugar and cholesterol measurements will be reviewed to determine whether you
qualify for the study. You will have a short physical exam, and one tube (about 2 teaspoonfuls) of blood
may be collected and tested for creatinine (a measure of kidney function), lipids and liver function. Some
urine will also be collected and tested for protein.
If you qualify for the study and volunteer to participate, your study doctor will treat your blood sugar and
either your blood pressure or your blood lipids according to the ACCORD study protocol. You and your
personal physician are still responsible for other parts of diabetes care, including general preventive
measures, foot care, and eye care. If you are not in the blood pressure part of ACCORD, your personal
physician will still be responsible for treating your blood pressure. If you are not in the blood lipids part
of ACCORD, your personal physician will still be responsible for treating your blood lipids (such as blood
cholesterol). In addition, you will still need to see your personal physician(s) for all other medical care.
Blood sugar treatment groups. If you qualify and consent, you will be randomly assigned (like the
flip of a coin) to one of the two blood sugar goals. The "intensive" goal is a blood sugar level lower than
the current recommended value. The "standard" goal is a blood sugar level similar to the current
recommended value. Your current treatment for diabetes (if any) will be changed to study treatment
based on the goal to which you are assigned. Your study treatment will use available and approved
diabetes treatments (oral medications and/or insulin as may be required).
If you are randomized to the intensive blood sugar goal, it is very likely that you may need one or more
of the following: a) at least 2 oral medications; b) 3 or more insulin injections per day; c) frequent self-
adjustment of insulin; and d) frequent home glucose monitoring. This means you will probably have to
take several pills, give yourself insulin injections with a small needle, and do finger sticks to test your
blood sugar up to eight times a day.
The degree of control of blood sugar is best measured by a test called hemoglobin A1c. This test gives
an average of your sugar values during the past 2 to 3 months. If you are in the intensive blood sugar
treatment group, the goal will be to keep your hemoglobin A1c at less than 6.0% (which is about an
average blood sugar of 115 mg/dl (6.4 mmol/L)). This level is much lower than usually achieved in
clinical practice. If you are in the standard blood sugar treatment group, the goal will be to keep your
hemoglobin A1c value between 7.0% and 7.9% with the average around 7.5% (average blood sugar of
160 mg/dl (8.9 mmol/L)). This level is also lower than that usually achieved in clinical practice. Lowering
hemoglobin A1c to this level from higher levels has been shown to reduce complications of diabetes like
eye and kidney diseases. Your diabetes medications may be adjusted upwards or downwards, as your
study doctors try to reach these blood goals safely.
Compared to the intensive target of a hemoglobin A1c of less than 6.0%, the standard hemoglobin A1c
target of 7.5% has a somewhat higher risk for some diabetes complications. These include eye disease
(retinopathy), kidney disease (nephropathy), and abnormal nerve function (neuropathy). On the other
hand, a hemoglobin A1c of less than 6.0% will increase somewhat the risk for developing serious low
blood sugar reactions (hypoglycemia) and weight gain. Whether the lower hemoglobin A1c target gives
more or less protection against cardiovascular disease (such as heart attack or stroke) is not known. This
is what ACCORD is trying to find out.
In the standard group, ACCORD will take action and recommend treatment to lower your blood sugar if
your hemoglobin A1c value becomes greater than 7.9%. If your hemoglobin A1c drops below 7.0% and
you are taking insulin or a secretagogue (like glimepiride or repaglinide) , we may reduce your diabetes
treatment to try to bring your value above 7.0%. In the intensive group, if your hemoglobin A1c value
becomes even slightly greater than 6.0%, we will increase your treatment.
Depending on your initial blood pressure and blood cholesterol results, you will also be asked to
participate in either the blood pressure or cholesterol parts of the study. You must participate in one or

154
the other (based on your qualifications) to participate fully in ACCORD.

Blood pressure treatment groups. Blood pressure lowering can prevent heart disease, stroke, and
kidney disease. There is some evidence that lowering blood pressure further than current practice might
help prevent heart disease and stroke in people with diabetes. This possibility needs careful testing in a
study such as this one.
If you qualify for the blood pressure portion of the study, you will be randomly assigned (like the flip of a
coin) to one of two blood pressure goals. The "intensive" goal is a blood pressure level lower than that
already proven to reduce heart disease and stroke. The "standard" goal is a blood pressure level similar
to that already proven to reduce disease. Your study doctor will choose the medications he/she feels will
be best for treating your blood pressure. Therefore, your current blood pressure medication (if any)
could be changed or continued. If you do not reach your blood pressure goal, your study doctor will
change your treatment until you do.
Blood lipid treatment groups. Lowering blood cholesterol can prevent heart disease and stroke.
There is also some evidence that changing other blood lipids by lowering triglycerides (a type of fat in the
blood) and raising HDL-cholesterol (the good cholesterol) may prevent heart disease in people with
diabetes. This possibility needs careful testing in a study such as this one.
If you are eligible to participate in the blood lipid study, your current cholesterol medication treatment (if
any) will be stopped and changed to the study medication. You will be treated with cholesterol-lowering
medication commonly known as a "statin". The statin used in ACCORD is called simvastatin.
The dose of simvastatin you are started on will depend on your medical history. If you have had a heart
attack, stroke, heart surgery, surgery on your arteries (blood vessels) or angina (chest pain) with
changes in an electrocardiogram (ECG or EKG), you will be started on 40 mg of simvastatin. If you have
not had any of those, you will receive 20 mg a day of simvastatin.
Regardless of your assigned dose of simvastatin, you will be randomly assigned (like the flip of a coin) to
a medication known as a fibrate to lower your triglycerides and raise your HDL-cholesterol, or to a
placebo (a pill that does not contain any medicine). The fibrate used in ACCORD is called fenofibrate.
Neither you nor your doctor will know which study treatment (placebo or fibrate) you are receiving. If it
becomes necessary to know for medical reasons, the information will be made available.
If you begin ACCORD at the 20 mg dose of simvastatin and your cholesterol levels remain higher than
the currently recommended level, or if you have a heart attack, stroke, heart surgery, surgery on your
arteries (blood vessels) or angina (chest pain) with changes in an electrocardiogram (ECG or EKG) during
the study, your dose of simvastatin will be increased to 40 mg per day. If your cholesterol level remains
too high despite treatment with the increased dose of simvastatin, you will be taken off the lipid study
medications and sent to your personal doctor to get appropriate treatment to reduce your cholesterol
level.
Genetic component. Genetic research will be done as part of this study. You may, if you wish,
volunteer for the genetic portion of the study. If you volunteer to participate in the genetic portion of
ACCORD, your blood will be stored for genetic (DNA) analysis. The genetic portion of ACCORD is
described in more detail below. You do not need to agree to participate in the genetic studies to
participate in the main ACCORD study.
Visit schedule and measurements. If you qualify for ACCORD and are assigned to the standard
blood glucose group and either the lipid trial or the standard blood pressure group, you will be asked to
visit the clinic at one month, four months, and every four months thereafter for the duration of the trial.
If you are assigned to any of the other groups, you will be asked to come every month for the first four
months of the study and then at least every two months thereafter until the end of the study.
At each clinic visit, your health will be reviewed, and any symptoms you may have will be discussed with
the study doctor or nurse or other study staff. Your weight, blood pressure, and heart rate will be
measured, and your study medications will be reviewed to make sure you are taking them correctly. You

155
will receive nutrition and physical activity recommendations and will be taught how to follow them. In
addition, a member of your ACCORD study care team may contact you by phone between your clinic
visits to determine how you are feeling and whether or not further action is required to control your blood
sugar or blood pressure levels.
You will have blood specimens (up to five tablespoons) drawn every four months for the first year and
once a year thereafter. These tests will measure blood sugar, potassium, kidney function, and liver
function. You will also be asked to allow blood and urine specimens to be taken and stored for future
non-genetic studies. Also, additional blood samples may be taken occasionally to monitor your
treatments for safety, which may require you to come in for additional visits.
Some urine will be collected at the baseline visit and every two years thereafter so that it can be
examined for urine protein and creatinine (a measure of kidney function). You will also have an
electrocardiogram (a recording of the electrical activity of the heart, also called an ECG or an EKG) at
baseline and every two years thereafter. A limited eye exam will be done every other year.
If you are in the cholesterol study, your blood cholesterol will be measured every four months during the
first year and every year thereafter until the end of the study. You will also have blood drawn every four
months throughout the study to check your kidney function. If you are not in the cholesterol study, you
will have your cholesterol measured every year.
As part of diabetes management, you will be expected to check your own blood sugar, as discussed later.
If you are assigned to the "intensive" blood sugar goal you will have more frequent blood sugar testing
by the clinic. This testing will range from once per month during the first 4 months of treatment to every
two months thereafter.
You also have about a 1-in-5 chance of being chosen to complete questionnaires about your quality and
activities of life, and your diet and physical activity levels. These questionnaires will be given at the
beginning of the study, your 1 year visit, 3 year visit, and 4 year visit. The questionnaires will take
about one hour of your time. In addition, you may be chosen to participate in a group where health care
costs will be monitored (and you would be asked to give permission to obtain records from any
hospitalizations).
Certain medical procedures are recommended for people with diabetes that are not part of the research
study. These include annual eye exams by an ophthalmologist, annual foot exams, annual flu and
pneumococcal vaccinations, and electrocardiograms (ECGs or EKGs). The study eye examination does
not replace the recommended annual eye exams by an experienced eye care professional, such as an
ophthalmologist (a doctor who specializes in the diagnosis and treatment of eye diseases).
During the course of the trial, our central Coordinating Center at Wake Forest University School of
Medicine, or its representatives may contact you, about your participation in the trial. For example, you
may be asked if you are having any trouble taking any of your medications. You may also be asked how
you are feeling and whether you have been in the hospital for any reason, why and where you were
hospitalized.
POTENTIAL RISKS OF PARTICIPATING IN THE ACCORD STUDY

What are the possible risks and discomforts? Due to unknown risks and potential harm to the
unborn fetus, sexually active women of childbearing potential must use a reliable method of birth control
while participating in this study. Reliable methods of birth control are considered to be abstinence (not
having sex), oral contraceptives (the pill), intrauterine device (IUD), DepoProvera, Norplant, tubal ligation
(tubes tied), or vasectomy of the partner (with confirmed negative sperm counts) in a monogamous
relationship (same partner). An acceptable, although less reliable method, involves the careful use of
condoms and/or a spermicidal foam or gel along with a diaphragm, cervical cap, or sponge. We
encourage you to discuss this issue further with your doctor if you have any questions.
If you are a pregnant woman, you cannot participate in this study. Because some methods of birth

156
control are not 100% reliable, a negative pregnancy test is required at least 10 days after your last
normal menstrual period if you are a sexually active woman of childbearing potential.
This study requires that blood be drawn from a vein in your arm several times during the study. Drawing
blood may result in pain at the point of puncture, a feeling of faintness, irritation of the vein, and bruising
or bleeding at the site of the needle stick. There is also a very slight possibility of an infection at the
needle puncture site. The study visits, procedures, and lab work might be more often than your medical
conditions usually require, but they are very important for the study.
This study requires daily finger-stick measurements of your blood sugar level. You probably have
experience testing your blood sugar by finger-stick before coming into the study. You need to test your
blood sugar daily because it is very important for the study that you keep your blood sugar values at the
assigned goal. If you are assigned to the intensive blood sugar goal, there is a good chance that at some
point you will be asked to do up to eight finger sticks a day to properly correct your blood sugar. Your
blood sugar checks will be reviewed by clinic personnel and will be used to figure out your treatment
plan. Clinic personnel, or others working for ACCORD, may contact you to discuss your blood sugar
results.
Treating blood sugar in persons with diabetes can sometimes cause blood sugar to be too low. This
condition, called "hypoglycemia", can result from changing diet, exercise, or medication. Symptoms are
usually mild but sometimes can be more serious.
Mild symptoms of hypoglycemia include hunger, anxiety, dizziness, or light-headedness. Sometimes
there is sweating, fatigue or mild confusion, tremors (shaking) or palpitations (feeling your heart beating
in your chest). Hypoglycemia may cause loss of consciousness. If this occurs while operating machinery
such as driving a car, it can result in injury or even be life threatening.
In rare cases, hypoglycemia can be very severe and require emergency treatment or hospitalization.
Severe hypoglycemia may cause brain damage, coma, or death. Severe hypoglycemia can occur in any
patient taking medication to lower blood sugar. It is more likely to occur in those treated with insulin to
achieve lower glucose targets, as in the intensive treatment group of this study.
A sugar-containing drink such as fruit juice usually quickly relieves the milder symptoms. You may be
given sugar pills to raise your blood sugar if you have symptoms. Medications are sometimes needed to
treat severe hypoglycemia. These may include intravenous (I.V.) fluids or injections of glucagon, a
medication that rapidly increases blood sugar.
Regardless of which blood sugar treatment group you are assigned to, safety will always be of first
importance when changes in the management of your blood sugar are made. Based on data from
previous studies it is estimated that, in the intensive group, about six out of 100 participants will have a
serious complication (such as hospitalization or emergency room visit for hypoglycemia) every year. In
the standard group about 2 participants may have such a complication every year. In either group,
ACCORD doctors and nurses will take action to lessen the risk of hypoglycemia should it occur too often
or in a severe form. On the other hand participants in the standard group may have a somewhat higher
risk of complications related to diabetes (like eye, kidney disease or abnormal nerve function). It is
estimated that, in the intensive group, about one out of 100 participants will have such a complication
every year. In the standard group about 1.5 participants may present such a complication every year.
If you are assigned to the intensive blood pressure group, you may experience blood pressure that is too
low. Symptoms of low blood pressure may be mild, such as feeling a little lightheaded, or less often may
be more severe, such as dizziness, fatigue, or fainting. Sitting or lying down often relieves these
symptoms. You should notify your clinic doctor or nurse if you have these symptoms. Clinic staff will
follow you closely to lower your chances of having too-low blood pressure.
What are the side effects of the medicines used in the study? All drugs have a potential risk of an
allergic reaction, which if not treated quickly, could become life threatening.
You may have side effects from the specific medications chosen as treatments. Medications that may be

157
used at this time in ACCORD are listed below. Additional medications may be chosen in the future. The
ACCORD staff will tell you about any new medicines that they may give you.
Possible side effects for the classes of medications include the following. Your doctors have ways to
manage these effects.
Blood sugar treatments
Sulfonylureas [glimepiride]: The most common side effects associated with this family of medicines
include hypoglycemia (low blood sugar), weight gain, and allergies. Very rarely, blood cell abnormalities
may occur. Your doctor has ways of managing the blood cell abnormalities.
Biguanides [metformin]: Common side effects associated with this drug class include nausea, vomiting,
diarrhea, bloating, loss of appetite, or metallic taste in the mouth. These usually get better after the first
few weeks of treatment. If these treatments are stopped, the side effects will go away over a day or two.
Very rarely, people can have a severe reaction known as lactic acidosis (a condition that occurs when
your body fluids and tissues have too much acid in them). Lactic acidosis almost always occurs in people
with advanced kidney disease, liver disease or heart failure, and in people who drink alcohol heavily.
Every effort will be made to avoid using this drug in people with those conditions.
Thiazolidinediones (TZDs) [rosiglitazone, pioglitazone]: The most common side effects related to this
group of medicines include fluid retention (a condition that occurs when your body holds in too much
water) and weight gain. Although the 4 mg/day dose of rosiglitazone (the TZD to be used in ACCORD) is
the only dose of rosiglitazone that has been approved by the U.S. FDA for use with insulin, higher doses
of rosiglitazone, which you may be placed on, have been combined with insulin in medical practice. The
use of drugs like rosiglitazone together with insulin may cause fluid retention, which could lead to or
worsen heart failure. Heart failure is a decreased ability to pump enough blood throughout the body.
Symptoms of heart failure include shortness of breath, cough, fatigue, tiredness, ankle swelling, or
weight gain. If your doctor prescribes insulin together with rosiglitazone, you will be monitored closely
for these symptoms, so that the medications can be adjusted or, if necessary, stopped.
Although there has been no report of liver difficulties with rosiglitazone, a related medication was
removed from the market due to rare, severe liver reactions. Thus, if you require this medication, you
will need to have blood tests looking for liver problems every two months for the first year after you
begin the medication and once a year thereafter.
Insulin [various short-, intermediate-, or long-acting forms, including aspart and glargine]: Potential side
effects related to insulin use include: low blood sugar, low potassium in the blood, allergies or skin
changes.
Meglitinides [repaglinide]: Common side effects include headache, upper respiratory infections, nausea,
vomiting, constipation, and diarrhea. The most serious side effect is hypoglycemia.
Alpha-Glucosidase Inhibitors [acarbose]: Side effects include flatulence (gas), diarrhea and
abdominal discomfort These are generally mild to moderate in severity and usually diminish in frequency
and intensity with time. Very rarely, this medication may cause skin reactions, hepatitis, and/or jaundice
(yellowing of the skin or whites of the eyes, indicating possible liver problems).
Blood pressure treatments
Angiotensin Converting Enzyme Inhibitors (ACE-I) [benazepril, lisinopril, ramipril]: Potential side
effects associated with this type of medicine include: dizziness, headache, fatigue, nausea, diarrhea,
cough, rash, high potassium in the blood, low blood pressure upon standing, harm to kidney function and
rarely angioedema (swelling of the face, lips and tongue that can result in difficulty breathing or in rare
cases, death).
Diuretics [chlorthalidone, hydrochlorothiazide]: Potential side effects associated with this class of
medication also known as "water pills" include: muscle cramps, nausea, vomiting, diarrhea, dizziness,
rash, weakness, low blood pressure, low potassium, high blood sugar, partial or total lack of ability to

158
perform sexual function, and gout (a painful joint condition that occurs when too much acid and salt build
up in the blood stream and joints).
Beta Blockers [metoprolol]: The most common side effects associated with this group of medicines
include: dizziness, fatigue, stomach upset, depression, cold hands and feet, low blood pressure, changes
in heart rhythm and heart rate, and decrease in sexual function. Beta-blockers may also hide some of
the symptoms but not the hazards of low blood sugar. If you begin taking these medications, you should
not stop taking them without talking to your study doctor first.
Calcium Channel Blockers [isradipine, diltiazem, amlodipine, nifedipine]: The most frequent side
effects associated with these medications are: ankle or foot swelling, dizziness, flushing, palpitations
(awareness of your heartbeat), headache, fatigue, nausea and abdominal discomfort. Occasionally,
severe hypotension (abnormally low blood pressure) may occur when starting these medications or
adjusting their dose. Rarely, increased angina (chest pain) and myocardial infarctions (heart attacks)
may occur in people with severe coronary artery disease. When combined with a Beta Blocker, the
medication nifedipine may cause congestive heart failure (a decreased ability to pump enough blood
through the body), which can be serious but is very rare.
Alpha Blockers [terazosin]: Potential side effects associated with this category include: fainting,
dizziness, fatigue, swelling, low blood pressure, partial or total lack of ability to perform sexual function,
changes in heart rhythm and certain blood cell abnormalities.
A-II Receptor Blockers [candesartan, valsartan]: The most common side effects are dizziness,
headache, fatigue, diarrhea, muscular-skeletal pain. More serious side effects are angioedema (swelling
of the face, lips and tongue that can result in difficulty breathing or in rare cases, death) and severe
hypotension. This family of drugs may also affect your kidney function. Your doctor may do blood tests
to see if your kidneys are performing properly.
Loop Diuretic [furosemide]: rare side effects include thrombocytopenia (low platelet count), rash,
pancreatitis (inflammation of the pancreas), and jaundice (yellowing of the skin or whites of the eyes,
indicating possible liver problems). Serious side effects include abnormalities in blood cells.
Sympatholytics [reserpine]: The most common side effects include dizziness, dry mouth, nausea,
vomiting, nasal congestion, peripheral edema (too much fluid in the body's tissues), stomach cramps,
headache, impotence, depression, nervousness, shortness of breath, nightmares, difficulty with urination,
shaky hands, and anorexia (poor appetite). More serious side effects include dysrhythmias (heart rhythm
abnormalities), black tarry stools, hematemesis (vomiting blood), bradycardia (slow heart rate), chest
pain, and thrombocytopenia (low platelet count).
Vasodilators [hydralazine]: Side effects include headache, tachycardia (fast heart rate), angina (chest
pain), and palpitations. Rare but more serious side effects include abnormalities in blood cells and lupus-
like syndrome.
Potassium Sparing Diuretics [triamterene]: The most common side effects include diarrhea, nausea,
vomiting, gastrointestinal distress, dizziness, dry mouth, pruritis (itching), rash, sensitivity to light,
weakness, hypotension, muscle cramps, blood chemical imbalances (such as too much potassium),
impaired kidney function, elevated uric acid, blood cell abnormalities and reduced folic acid stores. More
serious possible side effects include increased acid in the blood and shock due to an allergic reaction to
the medication.
Alpha-beta blockers [carvedilol]: The most common side effects are dizziness and fatigue. The more
serious side effects include AV block (a heart rhythm disturbance), bradycardia (slow heart rate),
thrombocytopenia (low platelet count), and bronchospasm (tightening of breathing airways). Alpha-beta-
blockers may also hide some of the symptoms but not the hazards of low blood sugar.
Lipid treatments
HMG-CoA Reductase Inhibitors (statins) [simvastatin]: Common side effects associated with this
class of cholesterol-lowering medications include: headache, dizziness, stomach upset. Rare, but more

159
serious side-effects are muscle aches, rash and elevated liver enzymes (indicating possible liver
problems) in the blood. (Also, see 'Drug Interactions' discussed below.)
Fibrates [fenofibrate]: Potential side effects associated with these medications include: abdominal pain,
stones in the gall bladder, jaundice (yellowing of the skin and/or whites of the eyes, indicating possible
liver problems), headache, change in taste, elevated liver and kidney function tests, and certain
abnormalities in blood cells. Your study doctor has ways to manage these blood cell abnormalities.
Fenofibrate could possibly harm the kidney. Blood tests will be done regularly to look at your kidney
functioning. If your results are not normal your dose of fenofibrate or placebo (whichever you are on)
will be reduced. If your values do not improve, the medication will be stopped entirely. After your dose
is reduced or stopped, your study doctor will continue to monitor your kidney function. (Also, see 'Drug
Interactions' discussed below.)
Drug Interactions
What are some of the ways the study drugs can interact? The Food and Drug Administration
(FDA) has approved all drugs that will be used in ACCORD. Most have been used for many years.
Therefore, we know much about the way these drugs work and how they interact with other drugs -
especially other treatments that will be used in this study.
Researchers know that using a sulfonylurea (a type of drug that lowers blood sugar) with certain other
drugs should be avoided. Your study doctor will make sure that you do not take these kinds of medicines
together.
Researchers also know that using statins and fibrates together may increase the chance for certain side
effects such as liver problems and muscle pain and inflammation. These side effects are rare, but are
more likely at higher statin doses. If your dose of simvastatin is increased to 40 mg per day, your chance
of side effects may be increased. Many doctors use simvastatin and fenofibrate together, and the
ACCORD trial will use caution whenever you are given this combination. Additionally, the ACCORD clinic
will be checking your blood to make sure that the study medications are not harming your liver or
muscles. These tests will be done at 1, 4, 8, and 12 months after you begin the medications, and every
year after that. If your study doctor thinks that the statin and fibrate medicines are causing problems for
you, then he/she may take you off one or both these medicines.
If you are eligible to be in the lipid portion of ACCORD and if you are on warfarin (also called Coumadin),
your personal doctor will be informed both by phone and in writing that you may be on fenofibrate.
Because the use of fenofibrate generally means that your dose of warfarin should be reduced to avoid
excessive risk of bleeding, you will be tested to see how fast your blood clots. This blood test can be
done by either the ACCORD clinic or by your private doctor. You will not be randomized until the
ACCORD clinic staff speaks with your private doctor about monitoring the appropriate dose of warfarin for
you. If you are placed on warfarin during the study, you will need to make sure that your private doctor
is reminded that you may be on fenofibrate.
POTENTIAL BENEFITS
What are the possible benefits? The ACCORD treatment may or may not be of personal benefit to
you. The information gathered from the study will be very important for the treatment of diabetes in the
future. There will be no charge to you for any of the required tests and procedures performed during
your participation in this study. Clinic visits, physical exams, laboratory tests, electrocardiograms and any
other procedures associated with the research aspects of this study are paid for by the study. In addition,
your medications for the blood sugar control as well as for the blood pressure control portion or blood
lipid control portion of ACCORD (whichever part you are in) will be provided to you free of charge. You
will not be paid for your participation in this study.

160
ALTERNATIVE TREATMENTS
If you chose not to participate, what other options do you have? You do not have to participate
in this research study in order to receive treatment. A number of treatments are available for diabetes,
high blood pressure, or high cholesterol. These treatments include drugs, diet, exercise, and weight loss.
If you decide to stop participating in this study, your personal doctor should manage your medical care.
NEW INFORMATION
What if we learn about new risks during the study? You will be given any new information gained
during the course of the study that might affect your health, welfare, or willingness to stay in the
ACCORD study. Results of your laboratory tests and clinical measurements will be provided to you to
share with your personal physician.
PRIVACY
How will your privacy be protected? Any information obtained about you during this study will be
treated as strictly confidential to the full extent permitted by applicable law. To ensure confidentiality, a
code number will be assigned to you. Your name and any other potentially identifying information will
not be used on any data or samples you provide. However, your name and Social Security and Medicare
numbers will be recorded and stored centrally to help the study keep track of any illnesses you may
experience. Also, in order to receive supplies (glucose strips) to measure your own blood glucose during
the trial, you will need to provide the information that will permit billing for Medicare (if you are covered)
and/or other insurance you may have (if you have it.) You will not be identified in any report or
publication about this study.
Your records for this study may be reviewed by authorized representatives from the National Heart, Lung,
and Blood Institute, the Food and Drug Administration (FDA) and monitoring personnel from the
________ Clinical Center Network Office for the study at __________and by the committee in charge of
protecting research participants at________________.
At the end of the study, all forms with your name or other identifying information will be kept in a locked
room for a period of five years. Only your study doctor or co-workers assisting the doctor will have
access to these forms. After five years, the forms will be destroyed.
Also at the end of the study, the Coordinating Center will provide the National Heart, Lung, and Blood
Institute (NHLBI) data from the study, without personal identifying information such as your name,
address, Social Security number, or Medicare number. Blood, urine, and/or tissue samples or other
materials taken from you during the study will be considered donated by you to medical research. These
materials may also be provided to the NHLBI at the end of the study, again without personal identifying
information. The data and/or materials may be shared with other scientists who meet NHLBI
requirements including treating the data or materials as medically confidential, obtaining approval from
their Human Subjects review boards, and agreeing not to share the data or materials with other parties.
Drug companies that have contributed drugs, and in some cases money, to the ACCORD study also will
be provided study data without any personal identifying information.
U.S. Federal Certificate of Confidentiality. It is particularly important to you to know that ACCORD
has been granted a Certificate of Confidentiality from the United States Federal Government to make sure
we can best protect your privacy. This certificate means that the ACCORD researchers cannot be forced
to tell anyone not connected with the study about your participation. This includes courts and police.
The researchers will only release information if you request it.
There are some limits to the researcher's ability to maintain your confidentiality. If we learn that keeping
information private would immediately put you in danger, or put someone else we know about in danger,
then we will have to tell the appropriate agencies to protect you or the other person.

161
INJURY
What will happen if you become ill during the study or suffer a complication related to the
treatment that you are receiving as part of the study? While it is not likely that you will suffer
major health problems as a result of your participation in this study, the medical treatment that is a part
of this study carries a small risk of serious health problems. Of course, should a problem occur, or should
you need emergency medical help, necessary emergency care would be provided and the investigator
working with you would help you find a doctor to continue your care if needed. Any cost of medical
care that results from such a health problem will be your responsibility and will not be paid for by the
National Heart, Lung, and Blood Institute, the study investigators, or the hospital or clinic conducting this
study.
QUESTIONS ABOUT THE STUDY AND YOUR RIGHTS

What if you have questions about this study? For questions about the study or in the event of a
research-related injury, contact the study investigator, __________, at ________________________
[INCLUDE AFTER-HOURS NUMBER].
What if you have questions about your rights as a participant? For questions about your rights as
a research participant, you may contact the Chairman of the Institutional Review Board, which is a group
of people who review the research to protect your rights as a research participant, at
________________. You will be given a copy of this consent form.
What if you want to stop before your part in the study is complete? Taking part in this study is
voluntary. You may choose not to take part or you may leave the study at any time. Refusing to
participate or leaving the study will not result in any penalty or loss of benefits to which you are entitled.
Your study doctor also has the right to stop your participation in this study at any time. This could be
because you have had an unexpected reaction, or have failed to follow instructions, or because the entire
study has been stopped.
GENETIC STUDIES
What is the goal of the genetic studies? One goal of ACCORD is to examine your genetic material
(DNA) and its relationship to the effects of the treatments. If you volunteer to participate in the genetic
studies you will be asked for a sample of blood (about 1 teaspoon) to obtain DNA from your blood cells.
Information gained from research on your DNA may be used to develop new ways to detect or treat
major diseases.
Will the DNA samples be shared with other institutions? If you agree to participate in the genetics
portion of the study, the ACCORD Central Laboratory may share DNA samples with researchers
participating in ACCORD. If you give permission, samples may also be shared with other research
laboratories studying the genetics of type 2 diabetes and the development of heart and blood vessel
diseases, other major diseases, health conditions, or risk factors. The scientists from these laboratories
would be given the DNA without any information to identify you.
How will genetic information be kept private? Only the ACCORD Central Laboratory will have
access to the samples. No other individual, including your spouse, parents, children, physician or
employer will have access to the stored sample or information gained from your stored sample. At the
end of the study, your samples may be provided to other investigators under certain conditions, without
any personal identifying information (See Privacy section above).
How long will the DNA samples be kept? Your sample may be kept until it is no longer of scientific
value. If, at any time during the study, you decide that you do not wish to have your DNA sample stored
any longer, you may notify your ACCORD study coordinator and the sample will be destroyed.
Who owns the samples? By checking "yes" at the end of this document, you volunteer to provide

162
genetic samples for medical research purposes. Your DNA will not be sold to anyone or to institutions or
companies for financial gain or commercial profit without your consent. Also, neither you nor your heirs
will receive money from any discoveries or inventions made using the information and/or specimens you
provide. There is no cost to you or your insurance company for the storage and use of the samples.
Will you receive study results of research involving your samples? You will not be informed of
the results of the research performed on your genetic blood sample, although genetic tests may be
developed after a study of samples in the ACCORD study. If there is any new information about genetic
testing for type 2 diabetes and its relationship to heart and blood vessel diseases or other health
conditions, you will be informed by your study doctor if this information may be important to you or your
family.
PARTICIPANT'S AGREEMENT FOR THE GENETIC PORTION OF ACCORD
Please check one of the following choices:

____ Yes, I agree to participate in the genetic portion of ACCORD
____ No, I do not agree to participate in the genetic portion of ACCORD
If you agreed to participate in the genetic portion of ACCORD, please check one of the following
regarding diseases to be studied:
____ I agree to allow my genetic sample to be studied for genes related to any major disease or
health condition or risk factors.
____ I agree to allow my genetic sample to be studied ONLY for genes related to diabetes,
blood pressure, blood cholesterol abnormalities, heart disease, other cardiovascular diseases,
kidney diseases, or other risk factors for heart disease or for diabetes.
If you agreed to participate in the genetic portion of ACCORD, please check one of the following
regarding investigators who will have access to the genetic samples:
____ I agree to allow my genetic samples to be used for research by ACCORD investigators as
well as by other researchers who meet NHLBI standards and procedures.
____ I agree to allow my genetic samples to be used ONLY for research by ACCORD
investigators.
PARTICIPANT'S AGREEMENT FOR ACCORD STUDY

I have read the information provided above. I voluntarily consent to participate in the
ACCORD study.
Participant's signature Date
Printed name of participant
___________________
Signature of person obtaining consent Date
______________________________________
Printed name of person obtaining consent

163
Appendix II:
Charges to the Subcommittees
of the ACCORD Steering Committee
During the protocol development phase of ACCORD, the Subcommittees of the Steering
Committee were responsible for developing specific portions of the protocol and for making
recommendations to the ACCORD Steering Committee for approval. During the data collection
phases of the trial, the Subcommittees will be responsible for monitoring specific portions of the
conduct of the trial and will provide periodic status reports to the Steering Committee.
Medical Interventions Subcommittee
This subcommittee developed the medical intervention plans for the trial, including the
glycemic, lipid, and blood pressure controls. This work was accomplished through three
intervention-specific working groups within the subcommittee. A fourth working group, the
Lifestyle/Background Therapy Working Group, developed the plans for smoking cessation,
weight control, exercise improvement, dietary modifications, and background pharmacologies.
During the data collection phases of the trial, the Medical Interventions Subcommittee will
monitor the progress of protocol-specified medical management strategies, as well as adherence
to the study medications and lifestyle changes. The subcommittee will develop strategies to
maximize adherence to medications and lifestyle modifications. An additional charge to this
subcommittee is to monitor the safety of the interventions and to make recommendations
regarding possible changes to the protocol/MOP because of safety concerns.
Recruitment and Retention Subcommittee
This subcommittee developed the trial eligibility criteria, as well as the screening and
recruitment strategies for patient accrual. During the recruitment phases of the trial, this
subcommittee will monitor recruitment and screening, and will identify/assist the Clinical Center
Networks (and their component clinics) experiencing recruitment difficulties. Adjustments to
eligibility criteria, if necessary to improve overall participant accrual, will be considered by this
group. During the follow-up phases of the trial, this subcommittee will monitor all aspects of
participant retention, including visit and procedure adherence.
Measurement and Quality Control Subcommittee
This subcommittee developed the general data collection forms for use in the trial (in
conjunction with other ACCORD subcommittee recommendations) and identified (with input
from other subcommittees) clinical laboratory data to be collected. This subcommittee will also
establish criteria under which the clinics, the Coordinating Center, and Core Units are expected
to perform. This subcommittee will review all aspects of quality control monitoring and will act
on these reports. Deviations from performance levels will be brought to the attention of this

164
subcommittee by the Coordinating Center. The monitoring activities will include, but not be
limited to, monitoring data quality, timeliness, completeness; monitoring alert levels; monitoring
data entry, and, with respect to the core labs, reviewing the processing of samples. Reports from
the Site Visitors to the Clinical Center Networks and to the Core Labs will be reviewed by this
Subcommittee to determine whether action should be taken.
Design and Analysis Subcommittee
This subcommittee reviewed alternative designs for the trial, including the impact of
various designs on sample size, statistical power, and patient recruitment. This subcommittee
will of necessity work closely with the Medical Interventions Subcommittee and the Recruitment
and Retention Subcommittee on the development of analysis plans for recruitment and adherence
monitoring.
Health-related Quality of Life/Cost-Effectiveness Subcommittee
This subcommittee established which measures of health-related quality of life and

cost-effectiveness are best studied in this trial, and developed plans for analyses of these data.
During follow-up, this subcommittee will monitor and assess the progress of this portion of the
trial and will prepare reports to the Steering Committee.
Operations Subcommittee
Selected staff of the CCNs will meet as an Operations Subcommittee to discuss and
review the progress of the trial. The purpose of this group is to assure communication among the
study sites with respect to overall study coordination. Also attending these meetings will be
representatives from the Coordinating Center and the Project Office, who will inform and train
the Project Coordinators on trial procedures. The CCN Project Coordinators, who are most
aware of the day-to-day issues at the sites, will be an invaluable resource to the trial and will be
invited to make recommendations regarding the conduct of the trial to the Steering Committee
for review and consideration.
Morbidity and Mortality Subcommittee
This subcommittee recommended the definitions to be used for the classification of study
events that comprise the primary and secondary ACCORD outcome measures. This
Subcommittee will develop the procedures for collecting the relevant information from the
clinical centers, develop the procedures to classify each applicable event, and develop quality
control procedures for the ascertainment and classification of these clinical events. During the
data collection phases of the trial, this subcommittee will oversee the work of the Event
Classification Working Group (made up of ACCORD investigators, who may or may not be on
the Morbidity and Mortality Subcommittee), who will meet on a regular basis, and who will use
the procedures and criteria adopted by the trial to classify the occurrence of clinical events in a
masked fashion and to monitor event ascertainment/classification quality control.

165
Publications and Presentations Subcommittee
This subcommittee developed the policies and procedures by which ACCORD

investigators will conduct analyses, write papers, and make presentations. Included in the
responsibilities of this subcommittee are approving analyses/papers/presentations, soliciting
writing group members, and monitoring the progress of all proposed papers to ensure their
prompt completion and publication.
Laboratory and Ancillary Studies Subcommittee
This subcommittee will review procedures regarding the collection and storage of body
fluids and specimens, provide appropriate recommendations to the Steering Committee regarding
the collection and storage, and develop policies regarding access to the stored fluids and
specimens. The subcommittee will also be responsible for reviewing ancillary study proposals,
providing feedback to the Principal Investigator of the proposal, and making recommendations to
the Steering Committee regarding the proposals.

166
Appendix III:
ACCORD Conflict of Interest Policy
(Revised 08/24/01)
General Principles
1. This full policy is to be made public on our Website and in publications when possible.
2. The primary concerns are twofold. First, that the ACCORD investigators maintain the
internal integrity of the study by which we mean the confidence among ourselves
(investigators and staff) as we develop and modify the detailed protocol, that advice is being
given and decisions are being made in as unbiased and fully informed manner as possible.
Second, that we maintain the external integrity of the study by which we mean the
acceptance of our process and results as having met public standards of conduct.
3. To meet these goals we will obtain full disclosure by all of the key members of the study
(defined below) of their, and their immediate family’s, financial relationships with all
pharmaceutical and biomedical companies judged to have an active or potential interest
in the conduct and outcome of the study. These are to be reported on a standard form,
each of which will be reviewed on at least an annual basis, or more frequently if there is a
significant change from the last report, by a small subset of the Executive Committee
(termed the Oversight Committee). The Oversight Committee will be comprised of the
Chair of the Steering Committee, the PI of the Coordinating Center, and the NHLBI
Project Officer. The information to be reported will be detailed, but will not include
specific dollar amounts, although the definitions below require that certain relationships
be segregated by those above and below certain dollar thresholds.
4. All of the study PIs, Co-PIs, and the Steering Committee and its various subcommittees’
members are covered by this policy.
5. A conflict of interest will not necessarily exclude any member of the study from
participating in study discussions, unless required in individual cases by the Oversight
Committee. However, full disclosure of all potential conflicts of interest will be made at
each meeting to all attendees in an effective, but non-cumbersome manner. This includes
the full Steering Committee as well as each of its subcommittees.
6. A significant financial conflict of interest, defined below, will cause a person to recuse
himself or herself from voting on all issues related to the conflict. All such actions will be
recorded and kept as part of the study record in the Coordinating Center. This policy
applies most especially to the subcommittees making recommendations to the full
Steering Committee during the protocol design phase, as well as to the Steering
Committee itself.
7. All financially relevant relationships are to be reported. Only those relationships that are
between the individual and the specific company (rather than between the individual’s

167
parent institution and the specific company, for example) present the potential for a
significant financial conflict of interest, defined under paragraphs 9a and 9b below.
Specifically, research funding for contracts or grants to the parent institution which
provide support to the individual, his/her laboratory or his/her close scientific
collaborators is not ordinarily judged to present the potential for a financial conflict of
interest, although such awards are to be fully disclosed as a part of this policy.
8. Those financially relevant relationships that are to be reported include employment,

consultancies, board memberships, honoraria, stock ownership or options, grants,
contracts, patents received or pending, and royalties. The Oversight Committee will
decide, with #9 below as a guideline, whether any of these and other relationships in
each individual case are significant enough to warrant recusal from voting or
discussions.
9. A significant financial relationship is defined to exist:
a. when the dollar amount awarded on an annual basis, or expected-to-be awarded on

an annual basis, with regard to each related corporate relationship exceeds $25,000.
The Oversight Committee may also judge lower dollar amounts as significant in
specific/individual circumstances.
b. or when there is any equity holding in a related company (excluding mutual funds
and blind trusts). Again the Oversight Committee may decide in individual
circumstances that the equity holdings are relatively minor enough to not present a
real conflict of interest.
10. Significant financial relationships in existence since October 1, 1999 between ACCORD
investigators and all pharmaceutical and biomedical companies judged to have an active
or potential interest in the conduct and outcome of the study will be described in all study
reports and publications. Similar relationships, but which are not significant, as well as
actions taken early in the design phase of ACCORD that end significant financial
relationships (e.g., stock divestment) will all be described on the ACCORD web site, but
will not ordinarily be listed in study reports or publications. In addition we will
obviously meet or exceed the reporting standards of the journals publishing our
manuscripts.

Accord Study

Uploaded by

Copyright:

Available Formats

Accord Study

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Accord Study

Uploaded by

Copyright:

Available Formats

Protocol

(May 11, 2005 Version)

ACCORD Protocol – May 11, 2005 Version

Chapter 1 – Introduction and Background.......................................................................... 3

Chapter 2 – Participant Selection and Follow-up ............................................................. 28

Chapter 4 – Participant Safety and Confidentiality .......................................................... 82

Chapter 5 – Clinical Outcome Measures .......................................................................... 87

Chapter 6 – Health-Related Quality of Life/Cost-Effectiveness Outcome Measures ...... 92

Chapter 7 – Statistical Considerations............................................................................ 101

Chapter 8 – Data Management and Training.................................................................. 112

Chapter 9 – Adherence ................................................................................................... 115

Chapter 10 – Ancillary Studies ...................................................................................... 118

Chapter 11 – Publication Policy ..................................................................................... 120

Chapter 12 – Quality Control ......................................................................................... 123

Chapter 13 – Study Organization ................................................................................... 127

Chapter 14 – References.................................................................................................. 133

I. Model Informed Consent Document…………………………………….152

II. Charges to the Subcommittees of the ACCORD Steering Committee….163

III. ACCORD Conflict of Interest Policy – General Principles……………..166

ACCORD Protocol – May 11, 2005 Version

The design is a randomized, multicenter, double 2 X 2 factorial design in 10,000 patients

ACCORD Protocol – May 11, 2005 Version

The ACCORD study is designed to have:

Secondary hypotheses include treatment differences in other cardiovascular outcomes,

ACCORD Protocol – May 11, 2005 Version

ACCORD Protocol – May 11, 2005 Version

Secondary hypotheses include treatment differences in other cardiovascular outcomes,

1.1 Diabetes and Glycemia

1.1.a Diabetes and Cardiovascular Disease

Diabetes mellitus is a common disorder that is frequently misunderstood and poorly

ACCORD Protocol – May 11, 2005 Version

1.1.b Glucose as a Continuous Risk Factor for Cardiovascular Disease

Diabetes is a metabolic disease characterized by hyperglycemia, in which the defining

Andersson 411 66 7.4 SMBG> 7.8 mmol/l Death 44 vs 32 1.4 N/A

Kuusisto 229 68 3.5 HbA1c>7 vs <7 CHD 12 vs 3 4.3 N/A

Agewall 94 67 6.3 N/A CV Death N/A N/A 1.54

Lehto 1059 58 7.2 HbA1c>10.7 vs<10.7 CHD N/A 1.4 N/A

Fu (1993) 479 61.2 4 HbA1c>8.4 vs <6.3 ECG 30.8 vs 1.5 1.17

1.1.c Glucose Reduction to Lower the Risk of Cardiovascular Disease

A separate randomization of 1704 obese participants in 15 UKPDS centers allocated 342

ACCORD Protocol – May 11, 2005 Version

ACCORD Protocol – May 11, 2005 Version

Study Yrs HbA1c HbA1c Therapy Outcome Relative Risk RRR

UKPDS (1998) 10 7.0% (113%) 7.9%(127%) Insulin/SU MI 16% (0,29) 25%

Kumamoto 6 7.1% (111%) 9.4%(147%) Insulin CV Events 46% (NS) 65%

DIGAMI 1 7.1% (range) 7.9%(range) Insulin Mortality 29% (4,51) N/A

1.1.d Glucose Reduction and Adverse Events

ACCORD Protocol – May 11, 2005 Version

Study HbA1c Hypoglycemia Mean Weight Gain

Kumamoto 7.1% (111%) 0% 1.9%/yr BMI incr 20.5 - 21.2 (6 yrs)

VACSDM (Abraira 1997) 7.3% (120%) 3%/yr 41%/yr Same as control

DIGAMI (Malmberg 1995) 7.0% Control N/A 1 kg/yr

UGDP(IVAR) FPG 6.7 mmol/l (120.6 3.2% N/A 0.2%/yr

1.1.e Rationale for a Trial of Glucose Lowering to Prevent Cardiovascular Disease

ACCORD participants will be randomized to two targeted levels of glucose control.

ACCORD Protocol – May 11, 2005 Version

1.1.f Key Methodologic Questions

1.1.f.1 Can We Achieve a HbA1c Target of 6% in the ACCORD Intensive Group?

ACCORD Protocol – May 11, 2005 Version

ACCORD Protocol – May 11, 2005 Version