Comparison of Pioglitazone vs Glimepiride on

Progression of Coronary Atherosclerosis in Patients

With Type 2 Diabetes: The PERISCOPE Randomized
Online article and related content Controlled Trial
current as of May 5, 2008.
Steven E. Nissen; Stephen J. Nicholls; Kathy Wolski; et al.
JAMA. 2008;299(13):1561-1573 (doi:10.1001/jama.299.13.1561)

http://jama.ama-assn.org/cgi/content/full/299/13/1561

Correction Contact me if this article is corrected.

Citations This article has been cited 4 times.

Contact me when this article is cited.

Topic collections Drug Therapy, Other; Endocrine Diseases; Diabetes Mellitus; Randomized
Controlled Trial; Prognosis/ Outcomes; Cardiovascular System; Cardiovascular
Disease/ Myocardial Infarction; Drug Therapy
Contact me when new articles are published in these topic areas.
Related Articles published in Does PERISCOPE Provide a New Perspective on Diabetic Treatment?
the same issue Philippe Gabriel Steg et al. JAMA. 2008;299(13):1603.

Subscribe Email Alerts

http://jama.com/subscribe http://jamaarchives.com/alerts

Permissions Reprints/E-prints
permissions@ama-assn.org reprints@ama-assn.org
http://pubs.ama-assn.org/misc/permissions.dtl

Downloaded from www.jama.com by guest on May 5, 2008

 ORIGINAL CONTRIBUTION JAMA-EXPRESS

Comparison of Pioglitazone vs Glimepiride

on Progression of Coronary Atherosclerosis
in Patients With Type 2 Diabetes
The PERISCOPE Randomized Controlled Trial
Steven E. Nissen, MD Context No antidiabetic regimen has demonstrated the ability to reduce progression of
Stephen J. Nicholls, MBBS, PhD coronary atherosclerosis. Commonly used oral glucose-lowering agents include sulfonyl-
Kathy Wolski, MPH ureas, which are insulin secretagogues, and thiazolidinediones, which are insulin sensitizers.

Richard Nesto, MD Objective To compare the effects of an insulin sensitizer, pioglitazone, with an in-
sulin secretagogue, glimepiride, on the progression of coronary atherosclerosis in pa-
Stuart Kupfer, MD tients with type 2 diabetes.
Alfonso Perez, MD Design, Setting, and Participants Double-blind, randomized, multicenter trial at
Horacio Jure, MD 97 academic and community hospitals in North and South America (enrollment Au-
gust 2003-March 2006) in 543 patients with coronary disease and type 2 diabetes.
Robert De Larochellière, MD
Interventions A total of 543 patients underwent coronary intravascular ultraso-
Cezar S. Staniloae, MD nography and were randomized to receive glimepiride, 1 to 4 mg, or pioglitazone, 15
Kreton Mavromatis, MD to 45 mg, for 18 months with titration to maximum dosage, if tolerated. Atheroscle-
rosis progression was measured by repeat intravascular ultrasonography examination
Jacqueline Saw, MD
in 360 patients at study completion.
Bo Hu, PhD
Main Outcome Measure Change in percent atheroma volume (PAV) from base-
A. Michael Lincoff, MD line to study completion.
E. Murat Tuzcu, MD Results LeastsquaresmeanPAVincreased0.73%(95%CI,0.33%to1.12%)withglimep-
for the PERISCOPE Investigators iride and decreased 0.16% (95% CI, −0.57% to 0.25%) with pioglitazone(P=.002). An
alternative analysis imputing values for noncompleters based on baseline characteristics

A
LTHOUGH MANAGEMENT OF showed an increase in PAV of 0.64% (95% CI, 0.23% to 1.05%) for glimepiride and a
glucose levels represents one decrease of 0.06% (−0.47% to 0.35%) for pioglitazone (between-group P=.02). Mean
of the principal treatment (SD) baseline HbA1c levels were 7.4% (1.0%) in both groups and declined during treat-
ment an average 0.55% (95% CI, −0.68% to −0.42%) with pioglitazone and 0.36% (95%
goals of diabetes therapy, it CI, −0.48% to −0.24%) with glimepiride (between-group P=.03). In the pioglitazone group,
has been difficult to demonstrate a fa- compared with glimepiride, high-density lipoprotein levels increased 5.7 mg/dL (95% CI,
vorable effect of improved glycemic 4.4 to 7.0 mg/dL; 16.0%) vs 0.9 mg/dL (95% CI, −0.3 to 2.1 mg/dL; 4.1%), and median
control on the macrovascular compli- triglyceride levels decreased 16.3 mg/dL (95% CI, −27.7 to −11.0 mg/dL; 15.3%) vs an
cations of this disease.1,2 No antidia- increase of 3.3 mg/dL (95% CI, −10.7 to 11.7 mg/dL; 0.6%) (P⬍.001 for both comparisons).
betic regimen has demonstrated the Median fasting insulin levels decreased with pioglitazone and increased with glimepiride
ability to reduce the progression of (P⬍.001). Hypoglycemia was more common in the glimepiride group and edema, frac-
coronary atherosclerosis. Accord- tures, and decreased hemoglobin levels occurred more frequently in the pioglitazone group.
ingly, there is little evidence to sup- Conclusion In patients with type 2 diabetes and coronary artery disease, treatment with
port a preference of one class of glucose- pioglitazone resulted in a significantly lower rate of progression of coronary atherosclero-
lowering medication over any other as sis compared with glimepiride.
a means to reduce atherosclerotic dis- Trial Registration clinicaltrials.gov Identifier: NCT00225277
ease burden.3 Sulfonylureas have been JAMA. 2008;299(13):1561-1573 www.jama.com
available for decades, lower blood glu-
cose by acting as insulin secreta- Author Affiliations and a List of the PERISCOPE
gogues, and represent one of the most commonly-used classes of antidia- Investigators appear at the end of this article.
betic therapy. Thiazolidinediones Corresponding Author: Steven E. Nissen, MD,
Department of Cardiovascular Medicine, Cleveland
(TZDs) are a relatively new class of an- Clinic Foundation, 9500 Euclid Ave, Cleveland, OH
For editorial comment see p 1603.
tidiabetic agents that reduce glucose pri- 44195 (nissens@ccf.org).

©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, April 2, 2008—Vol 299, No. 13 1561

Downloaded from www.jama.com by guest on May 5, 2008

PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

sent. Participants aged 35 to 85 years

Figure 1. Flow of Patients Through the Trial
were eligible if they had a baseline gly-
1636 Patients screened
cated hemoglobin (HbA1c) level of 6.0%
to 9.0% (if taking a glucose-lowering
1089 Excluded
medication) or 6.5% to 10% (if not cur-
890 Did not meet inclusion rently receiving drug therapy). Pa-
or met exclusion criteria
89 Withdrew consent tients were required to have coronary
110 Other angiography performed for clinical in-
dications that demonstrated at least 1
547 Randomized angiographic stenosis with at least 20%
narrowing. A “target vessel” for IVUS
273 Randomized to receive glimepiride 274 Randomized to receive pioglitazone examination was required to have less
273 Received study drug 270 Received study drug than 50% obstruction throughout a
as assigned as assigned
40-mm or longer segment.
92 Did not complete end point assessment 91 Did not complete end point assessment
Patients were excluded for type 1
2 Died before final intravascular 1 Died before final intravascular diabetes, if they were taking 3 or more
ultrasound obtained ultrasound
87 Final intravascular ultrasound not 85 Final intravascular ultrasound not antidiabetic medications, or if they
obtained
3 Final intravascular ultrasound not
obtained
5 Final intravascular ultrasound not
had received any TZD within the past
analyzable analyzable 12 weeks. Other major exclusion
criteria were a serum creatinine level
181 Included in primary analysis 179 Included in primary analysis of ⬎2.0 mg/dL, triglyceride level of
273 Included in safety analysis 270 Included in safety analysis more than 500 mg/dL, uncontrolled
4 Excluded (did not receive
study drug) hypertension (blood pressure ⬎160/
100 mm Hg despite therapy), active
liver disease, or a left main coronary
marily by increasing insulin sensitiv- curate and reliable approach to evalu- artery stenosis of more than 50%.
ity in peripheral tissues.4 ating the effect of therapies on progres- Each of these exclusion criteria was
These 2 approaches to management sion of coronary atherosclerosis. 9 selected to exclude patients who
of hyperglycemia have important dif- Pioglitazone was selected to represent might not complete a full 18 months
ferences in their metabolic and physi- the TZD class and glimepiride chosen of treatment after randomization.
ological effects.3 Two TZDs are cur- as representative of the sulfonylurea All diabetic medications, including
rently marketed: pioglitazone and class. The goal of this study was to di- insulin, were permitted during the
rosiglitazone. Both agents reduce in- rectly compare the effectiveness of these study with the exception of a TZD,
flammatory biomarkers, while piogli- 2 alternative approaches, an insulin- sulfonylurea, or other insulin secreta-
tazone also produces greater elevations providing vs an insulin-sensitizing gogue. An independent committee
of high-density lipoprotein choles- strategy, in reducing progression of ath- blinded to treatment assignment cen-
terol (HDL-C) and reduces triglycer- erosclerosis in patients with type 2 dia- trally adjudicated adverse cardiovas-
ide levels.5 In separate randomized clini- betes and coexisting coronary artery cular events.
cal trials, pioglitazone decreased the rate disease.
of progression of carotid intimal me- Titration of Study Drug
dial thickness and showed some evi- METHODS If the patient was taking a sulfonyl-
dence for a reduction in adverse car- Study Design urea or other insulin secretagogue prior
diovascular outcomes.6,7 The effects of The PERISCOPE Trial (Pioglitazone to randomization, the drug was discon-
sulfonylureas on cardiovascular dis- Effect on Regression of Intravascular tinued. Participants naive to glucose-
ease have been controversial.8 How- Sonographic Coronary Obstruction Pro- lowering therapy at screening or tak-
ever, the majority of recent analyses spective Evaluation) was a prospec- ing less than glimepiride, 2 mg/d (or the
suggest a neutral effect for this class of tive, randomized, multicenter, double- equivalent dosage of another sulfonyl-
agents on cardiovascular outcomes.1,3 blind clinical trial (FIGURE 1). The study urea), took 15 mg of pioglitazone or 1
We compared the effects of these 2 was designed by the Cleveland Clinic mg of glimepiride. For those taking 2
widely used, but distinctly different, Coordinating Center for Clinical Re- mg/d or more of glimepiride (or the
classes of oral hypoglycemic agents on search (C5 Research) in collaboration equivalent) or metformin mono-
the rate of progression of coronary ath- with the sponsor. The institutional re- therapy, the starting dosage was 30 mg
erosclerosis. We used intravascular ul- view boards of participating centers of pioglitazone or 2 mg of glimepiride.
trasonography (IVUS) because this approved the protocol and all partici- If fasting capillary blood glucose mea-
imaging modality is considered an ac- pants provided written informed con- surements exceeded 140 mg/dL at sub-
1562 JAMA, April 2, 2008—Vol 299, No. 13 (Reprinted) ©2008 American Medical Association. All rights reserved.

Downloaded from www.jama.com by guest on May 5, 2008

 PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

sequent visits, the study drug was in- A follow up IVUS examination was plied by the median number of matched
creased to the next level (2 mg or 4 mg performed after 18 months of treatment cross-sections in pullbacks for all
for glimepiride or 30 mg or 45 mg for in all patients who provided informed patients completing the trial. This pro-
pioglitazone). Study drug was titrated consent, regardless of whether they con- cedure adjusts for pullbacks with dif-
to the maximum dose by 16 weeks, if tinued to take the study drug (modified fering numbers of measured cross-
tolerated. Metformin, insulin, or both intent-to-treat population). If a patient sections, resulting in an equal weighting
could be added or increased in dosage required cardiac catheterization for a of each individual patient in comput-
at the discretion of the investigator to clinical indication between 12 and 18 ing efficacy results. An additional sec-
achieve the target HbA1c level of less months, a follow-up IVUS examination ondary efficacy parameter, change in
than 7.0%. If a patient experienced hy- wasperformedtoavoidexposingpatients atheroma volume in the most dis-
poglycemic symptoms, other glucose- to the risk of an additional catheteriza- eased 10-mm subsegment, was calcu-
lowering therapies were reduced to tion at study completion. lated by first determining the 10 con-
maintain maximal dosages of study The follow-up IVUS examination was tiguous cross-sections with the greatest
medication. performed in the same coronary seg- atheroma volume at baseline, then com-
ment imaged during baseline exami- paring atheroma volume at follow-up
Randomization and Allocation nation. The motorized pullback was ini- for these same cross-sections.
Concealment tiated at the same side branch originally PAV was selected as the primary effi-
The patients and all study personnel selected during the baseline examina- cacyparameterbecausethisendpointhas
were blinded to treatment assign- tion to ensure that the same segment exhibitedthelowestvariabilityinmultiple
ment. The randomization was per- was imaged during follow-up. Using previous IVUS trials for a diverse range
formed using an interactive voice– digitized images, personnel blinded to of therapeutic interventions.10-12,14-16
response system that used a block size clinical characteristics and treatment as-
of 4. The study specified a balanced signments performed measurements for Statistical Methods
(1:1) treatment allocation stratified ac- cross-sections spaced at 1.0-mm inter- For continuous variables with a nor-
cording to diabetes treatment status at vals. For each analyzed cross-section, mal distribution, the mean and 95%
the time of randomization (treatment- the operator measured the area of the confidence intervals (CIs) are re-
naive or ⬍2 mg/d of glimepiride or external elastic membrane (EEM) and ported. For variables not normally dis-
equivalent vs ⱖ2 mg/d of glimepiride the lumen. The accuracy and repro- tributed, median and interquartile
or equivalent, metformin mono- ducibility of this method has been re- ranges are reported and 95% CIs around
therapy, or sulfonylurea-metformin ported previously.17 median changes were computed using
combination therapy). bootstrap resampling. Intravascular ul-
Calculation of Efficacy Parameters trasonography efficacy parameters were
Intravascular Ultrasound The primary efficacy parameter, change analyzed using analysis of covariance
Examination and Measurement in percent atheroma volume (PAV), was and reported as least square means and
Patients underwent coronary angiog- calculated as previously described.10-16 95% CIs using a linear model that in-
raphy and the operator selected a coro- cluded treatment group, pooled cen-
nary vessel for intravascular ultra- ter, and baseline values as covariates.
(EEMCSA − LUMENCSA)
sound (IVUS) examination. The PAV = × 100 All P values are 2-sided and not ad-
methods for IVUS image acquisition EEMCSA justed for multiple testing. P values of
and measurement have been de- ⱕ.05 were considered significant. A
scribed in prior publications.10-16 Briefly, where EEMCSA is the external elastic sensitivity analysis was performed using
the operator was instructed to select the membrane cross-sectional area and a multiple imputation procedure
longest and least angulated epicardial LUMEN C S A is the luminal cross- (PROC MI and ANALYSE in SAS ver-
coronary artery for IVUS examina- sectional area. sion 9.1; SAS Institute Inc, Cary, North
tion. The ultrasound probe was passed Several secondary IVUS efficacy para- Carolina) to impute the IVUS end
over a guidewire into the distal vessel meters were prespecified, including the points for patients who did not have a
just beyond a side branch, which served change in average maximum ath- follow-up IVUS performed.
as a reference point. Subsequently, a eroma thickness, computed by mea- For the primary efficacy parameter,
motor drive was engaged that with- suring the maximum thickness of the change in PAV, the sample size was se-
drew the ultrasonic transducer at 0.5 atheroma in each cross section, then av- lected to provide 90% power at a 2-sided
mm/s. The examination was screened eraging these values for the entire ar- ␣ of .05 to detect a treatment differ-
for image quality in the core labora- tery. The secondary efficacy param- ence of 1.8% assuming a 5.0% stan-
tory and only patients meeting pre- eter “normalized total atheroma dard deviation. A treatment difference
specified image quality requirements volume” was calculated as the average of 1.8% and standard deviation of 5.0%
were eligible for randomization. atheroma area per cross section multi- was based on the differences between
©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, April 2, 2008—Vol 299, No. 13 1563

Downloaded from www.jama.com by guest on May 5, 2008

PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

ever, despite randomization, small,

Table 1. Baseline Demographic Characteristics and Medications at the Time of
Randomization but statistically significant imbalances
Glimepiride Pioglitazone were observed for 2 characteristics. Pa-
(n = 273) (n = 270) P Value a tients randomized to receive glime-
Age, mean (SD), y 59.7 (9.1) 60.0 (9.4) .64 piride were more likely to report a
Men, No. (%) 180 (65.9) 186 (68.9) .39 history of hypertension, although base-
Race, No. (%) b line blood pressures were similar
White 220 (80.6) 225 (83.3)
between treatment groups (TABLE 2). Pa-
Black 27 (9.9) 30 (11.1)
.17 tients randomized to receive piogli-
Asian 16 (5.9) 12 (4.4)
tazone were more likely to be former
Native American 10 (3.7) 3 (1.1)
smokers, whereas patients randomized
Ethnicity, No. (%)
Hispanic 71 (26) 63 (23.3) .21 to receive glimepiride were more likely
Weight, mean (SD), kg 92.8 (18.5) 94.2 (19.7) .28 to be current smokers. Race and ethnic-
BMI, mean (SD) 32.0 (5.2) 32.1 (5.3) .91 ity were self-reported by selecting op-
Smoking status, No. (%) tions from a list and were collected to
Current 53 (19.4) 31 (11.5) .01 determine whether these characteris-
Past 119 (43.6) 147 (54.4) .44 tics influenced the response of patients
Never 99 (36.4) 89 (33.0) .42 to the study treatments. Characteris-
Duration of disease, median (IQR), mo tics were similar in the 360 patients who
Diabetes, mo 71.0 (30.0-131.0) 70.0 (27.0-129.0) .74
completed the trial and the 183 pa-
Coronary disease 8.0 (1.0-56.0) 9.0 (1.0-50.0) .90
tients who did not (data not shown).
Hypertension, No. (%) 250 (91.6) 225 (83.3) .002
A high percentage of patients re-
Prior myocardial infarction, No. (%) 70 (25.6) 83 (30.7) .08
Medication use, No. (%)
ceived concomitant medications of es-
Aspirin 251 (91.9) 242 (89.6) .29 tablished value in prevention of cardio-
␤-Blocker 211 (77.3) 205 (75.9) .84 vascular complications of diabetes.
ACE inhibitor or ARB 229 (83.9) 217 (80.4) .30 Greater than 80% of patients were re-
Statin 224 (82.0) 220 (81.5) .87 ceiving an angiotensin converting en-
Other lipid-lowering agent 17 (6.2) 13 (4.8) .50 zyme inhibitor or angiotensin receptor
Metformin 174 (63.7) 176 (65.2) .69 blocker, more than 75% were receiving
Insulin 63 (23.1) 49 (18.1) .09 a ␤-blocker, approximately 90% were
Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; BMI, body mass index, cal- taking aspirin, and nearly 90% were re-
culated as weight in kilograms divided by height in meters squared; IQR, interquartile range.
a Comparisons were made after controlling for pooled center. ceiving lipid-lowering therapy. The mean
b Self-report by participant by selecting options from a list.
titrated daily dosage of glimepiride was
2.9 mg, and the mean titrated dosage of
the 2 active therapies in patients density lipoprotein cholesterol (LDL-C) pioglitazone was 37.4mg.
with diabetes in a previous trial com- to mmol/L, multiply by 0.0259; and in-
paring moderate and intensive lipid- sulin to pmol/L, multiply by 6.945. Laboratory Results
lowering therapies.11 Approximately Table 2 summarizes laboratory val-
330 patients with evaluable IVUS ex- RESULTS ues, body weight, and blood pressures
aminations at baseline and follow-up Patient Population for the 360 patients who had evalu-
were required. Assuming a dropout rate Between August 5, 2003, and March 31, able IVUS examinations at both base-
of 25%, randomization of approxi- 2006, 547 patients were randomized line and follow-up. FIGURE 2 shows
mately 440 randomized individuals was and 543 received study drug at 97 cen- changes during the course of the trial
anticipated. Due to an observed drop- ters in North America and South for HbA1c levels, fasting plasma glu-
out rate of approximately 35% during America, 273 in the glimepiride group cose, HDL-C, and high-sensitivity C-
study conduct, the enrollment target and 270 in the pioglitazone group. reactive protein. At baseline, in both
was increased to 540 patients. Analy- Evaluable baseline and follow-up IVUS treatment groups, mean HbA1c concen-
ses were performed using SAS version examinations were available in 360 pa- tration was 7.4%, LDL-C averaged ap-
8.2 (SAS Institute Inc). tients (66%), 181 in the glimepiride proximately 95 mg/dL, HDL-C was
To convert creatinine to µmol/L, group and 179 in the pioglitazone slightly higher than 40 mg/dL, and fast-
multiply values by 88.4; C-reactive pro- group. The disposition of patients in the ing triglycerides was about 145 mg/
tein to nmol/L, multiply by 9.524; tri- trial is shown in Figure 1. dL. Blood pressure averaged approxi-
glycerides to mmol/L, multiply by Most demographic characteristics and mately 128/75 mm Hg. There were
0.0113; blood glucose to mmol/L, mul- baseline medications were similar in no major differences in laboratory val-
tiply by 0.0555; HDL-C and low- both treatment groups (TABLE 1). How- ues for patients randomized to the 2
1564 JAMA, April 2, 2008—Vol 299, No. 13 (Reprinted) ©2008 American Medical Association. All rights reserved.

Downloaded from www.jama.com by guest on May 5, 2008

 PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

Table 2. Baseline and On-Treatment Vital Signs and Laboratory Values in Participants Completing the Trial (n = 360) a
Glimepiride Pioglitazone
Parameters (n = 181) (n = 179) P Value b
Baseline Values
HbA1c, mean (SD), % 7.4 (1.0) 7.4 (1.0) .76
Fasting blood glucose, mean (SD), mg/dL 148.0 (43.4) 147.2 (41.0) .77
Fasting insulin levels, median (IQR), µU/mL 22.5 (15.5 to 32.0) 21.0 (16.0 to 33.0) .52
Cholesterol, mean (SD), mg/dL
Total 163.8 (38.0) 160.8 (35.7) .47
LDL 94.4 (32.9) 93.5 (30.7) .89
HDL 43.4 (13.9) 40.8 (11.5) .05
Fasting triglycerides, median (IQR), mg/dL 145 (104 to 216) 139 (104 to 198) .28
C-reactive protein, median (IQR), mg/L 3.0 (1.3 to 5.2) 2.6 (1.2 to 6.5) .57
Brain-type natriuretic peptide, median (IQR), pg/mL 24.0 (11 to 42) 22.0 (11 to 52) .60
Blood pressure, mean (SD), mm Hg
Systolic 128.6 (17.1) 127.8 (16.6) .95
Diastolic 75.2 (9.4) 75.7 (10.7) .45
Weight, mean (SD), kg 93.3 (18.5) 94.3 (19.5) .65
Average Values During Treatment c
HbA1c, mean (SD), % 7.0 (1.0) 6.9 (0.9) .04
Fasting blood glucose, mean (SD), mg/dL 147.9 (33.8) 139.3 (29.1) .01
Fasting insulin levels, median (IQR), µU/mL d 26.5 (16.0 to 36.0) 16.0 (12.0 to 23.0) ⬍.001
Cholesterol, mean (SD), mg/dL
Total 165.9 (35.7) 164.5 (35.9) .95
LDL 96.1 (30.4) 95.6 (28.9) .82
HDL 43.7 (12.4) 46.8 (11.7) .03
Fasting triglycerides, median (IQR), mg/dL 152 (108 to 220) 119 (87 to 166) ⬍.001
C-reactive protein, median (IQR), mg/L d 2.2 (1.1 to 4.1) 1.2 (0.7 to 3.1) ⬍.001
Brain-type natriuretic peptide, median (IQR), pg/mL 22.6 (11.5 to 41.2) 32.4 (17.0 to 67.3) ⬍.001
Blood pressure, mean (SD), mm Hg
Systolic 130.5 (11.9) 128.1 (11.1) .07
Diastolic 75.9 (6.7) 74.5 (7.4) .05
Weight, mean (SD), kg d 94.9 (19.1) 97.8 (21.1) .16
Change From Baseline e,f,g
HbA1c level, LS mean (95% CI), % −0.36 (−0.48 to −0.24) −0.55 (−0.68 to −0.42) .03
Fasting blood glucose, LS mean (95% CI), mg/dL 0.41 (−3.66 to 4.48) −8.5 (−12.73 to −4.27) .003
Fasting insulin levels, median (95% CI), µU/mL, [% change] d 1.33 (−0.5 to 5.0) [8.5] −5.0 (−7.0 to −4.0) [−28.3] ⬍.001
Cholesterol, LS mean (95% CI), mg/dL [% change]
Total 1.16 (−2.9 to 5.3) [3.1] 2.5 (−3.3 to 8.3) [4.4] .39
LDL 1.1 (−2.4 to 4.6) [6.9] 2.1 (−1.5 to 5.8) [6.6] .69
HDL 0.9 (−0.3 to 2.1) [4.1] 5.7 (4.4 to 7.0) [16.0] ⬍.001
Fasting triglycerides, median (95% CI), mg/dL [% change] 3.3 (−10.7 to 11.7) [0.6] −16.3 (−27.7 to −11.0) [−15.3] ⬍.001
C-reactive protein, median (95% CI), mg/L [% change] d −0.4 (−0.9 to −0.2) [−18.0] −1.0 (−1.5 to −0.8) [−44.9] ⬍.001
Brain-type natriuretic peptide, median (95% CI), pg/mL 0.58 (−2.0 to 4.5) 8.0 (5.0 to 10.5) .001
Blood pressure, median (95% CI), mm Hg
Systolic 2.3 (0.9 to 3.7) 0.1 (−1.4 to 1.5) .03
Diastolic 0.9 (0.1 to 1.7) −0.9 (−1.7 to −0.01) .003
Weight, median (95% CI), kg d 1.6 (0.8 to 2.4) 3.6 (2.8 to 4.4) ⬍.001
Abbreviations: ACE, angiotensin-converting enzyme inhibitor; ARB, angiotensin-receptor blocker; CI, confidence interval; HDL, high-density lipoprotein; IQR, interquartile range;
LDL, low-density lipoprotein; LS, least squares.
SI conversion factors: To convert C-reactive protein to nmol/L; multiply by 9.524; fasting blood glucose to mmol/L, multiply by 0.0555; fasting insulin to pmol/L, multiply by 6.945;
HDL, LDL, and total cholesterol to mmol/L, multiply by 0.0259; and triglycerides to mmol/L, multiply by 0.0113.
a All participants with a baseline and follow-up intravascular ultrasound examination who contributed to the primary efficacy parameter were included in this analysis.
b P values for average follow-up values were generated from a 2-way analysis of variance model with terms for baseline, treatment, and pooled center.
c Mean (SD) or median (IQR) average postrandomization values.
d Last observation carried forward.
e Least-square means for absolute and percentage change (%) were estimated using a 2-way analysis of variance with terms for treatment, pooled center and baseline value.
f For laboratory parameters with a nonnormal distribution (fasting insulin, triglycerides, and C-reactive protein), percentage change was estimated using the logarithm of the ratio of
follow-up value to baseline value as the dependent variable.
g 95% CIs around medians were calculated using bootstrap resampling.

©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, April 2, 2008—Vol 299, No. 13 1565

Downloaded from www.jama.com by guest on May 5, 2008

PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

treatment groups, although patients vs 3.3 mg/dL (95% CI, −10.7 to 11.7 mg/ associated with greater mean weight
randomized to glimepiride had a dL; 0.6%) and high-sensitivity C-reactive gain, 3.6 kg (95% CI, 2.8 to 4.4 kg) vs
slightly higher mean HDL-C level, 43.4 protein, −1.0 mg/L (95% CI, −1.5 to −0.8 1.6 kg (0.8 to 2.4 kg; P ⬍.001), and an
mg/dL compared with 40.8 mg/dL for mg/L; 44.9%) vs −0.4 mg/L (95% CI, −0.9 increase in brain-type natriuretic pep-
pioglitazone (P =.05). to −0.2 mg/L; 18%) showed more favor- tide levels, 8.0 pg/mL (95% CI, 5.0 to
During the first 24 weeks of the trial, ableeffectsinthepioglitazonegroupcom- 10.5 pg/mL) vs 0.58 pg/mL (95% CI,
HbA1c values were reduced comparably paredwiththeglimepiridegroup(P⬍.001 −2.0 to 4.5 pg/mL; P =.001; Table 2).
in the 2 treatment groups, but subse- for both comparisons). Glimepiride in-
quently diverged, with average postran- creased median fasting insulin levels by IVUS Results
domization values slightly lower in the 1.33 µU/mL (95% CI, −0.5 to 5.0 µU/mL; TABLE 3 summarizes the results for the
pioglitazone group, 6.9% compared with 8.5%), whereas pioglitazone reduced me- primary and secondary IVUS efficacy
7.0% in the glimepiride group (P=.04; dian fasting insulin levels by 5.0 µU/mL parameters. The primary efficacy mea-
Table 2 and Figure 2). The mean HDL-C (95% CI, −7.0 to −4.0 µU/mL; 28.3%; sure, least square mean change in per-
levelsincreasedfrombaselinesignificantly P⬍.001). cent atheroma volume, increased 0.73%
moreinthepioglitazonetreatmentgroup, Systolic blood pressure increased in (95% CI, 0.33% to 1.12%) in the glimep-
5.7 mg/dL (95% CI, 4.4 to 7.0 mg/dL; both treatment groups, but the me- iride group (P⬍.001 for change from
16.0%) compared with 0.9 mg/dL (95% dian increase was smaller in the piogl- baseline) and decreased 0.16% (95% CI,
CI, −0.3 to 2.1 mg/dL; 4.1%) for the itazone group, 0.1 (95% CI, −1.4 to 1.5 −0.57% to 0.25%) in the pioglitazone
glimepiride group, (P⬍.001; Table 2 and mm Hg) compared with the glimep- group (P=.44 for change from baseline;
Figure 2). Changes from baseline in me- iride group, 2.3 (95% CI, 0.9 to 3.7 between-treatment groups, P = .002).
dianlevelsofbothtriglycerides,−16.3mg/ mm Hg; P = .03). Compared with Analysis of the primary end point ad-
dL(95%CI,−27.7to−11.0mg/dL;15.3%) glimepiride, pioglitazone treatment was justing for baseline differences in smok-

Figure 2. Changes in Mean Hemoglobin A1c, Fasting Plasma Glucose Levels, High-Density Lipoprotein Cholesterol, and C-Reactive Protein
During the Trial (n = 360)

Hemoglobin A1c
Fasting plasma glucose
Mean Fasting Plasma Glucose, mg/dL

8.5 220

200
8.0
180
Mean HbA1c, %

7.5
160

7.0 140

120
6.5
100
6.0
80

5.5 60
0 8 16 24 32 40 48 56 64 72 0 8 16 24 32 40 48 56 64 72
Weeks Following Randomization Weeks Following Randomization

High-density liproprotein cholesterol

C-reactive protein
65
Median C-Reactive Protein, mg/L

6.0 Glimepiride
60 Pioglitazone
Mean HDL-C, mg/dL

55 5.0

50 4.0
45
3.0
40
2.0
35

30 1.0

25 0
0 8 16 24 32 40 48 56 64 72 0 8 16 24 32 40 48 56 64 72
Weeks Following Randomization Weeks Following Randomization

SI conversion factors: To convert C-reactive protein to nmol/L; multiply by 9.524; fasting plasma glucose to mmol/L, multiply by 0.0555; high-density lipoprotein
cholesterol to mmol/L, multiply by 0.0259; and triglycerides to mmol/L, multiply by 0.0113. Error bars indicate standard deviations for the hemoglobin A1c, glucose,
and high-density lipoprotein cholesterol panels and interquartile range for the C-reactive protein panel.

1566 JAMA, April 2, 2008—Vol 299, No. 13 (Reprinted) ©2008 American Medical Association. All rights reserved.

Downloaded from www.jama.com by guest on May 5, 2008

 PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

ing did not significantly alter the re- IVUS Results Imputing Clinical Outcomes
sults. A secondary efficacy measure, Noncompleters and Adverse Events
change in maximum atheroma thick- Because of the relatively high rate (34%) TABLE 5 shows centrally adjudicated
ness increased in the glimepiride group, of noncompletion of patients random- clinical outcomes, investigator-
0.011 mm (95% CI, −0.0002 to 0.022 ized in the trial (ie, no analyzable IVUS), reported adverse events, laboratory ab-
mm) and decreased in the pioglitazone a sensitivity analysis was performed in normalities, and reasons for discon-
group, −0.011 mm (95% CI, −0.022 to which values for the primary and sec- tinuing drug therapy during the trial.
0.0004 mm; between groups, P=.006). ondary end points were imputed for each An independent committee blinded to
Another secondary IVUS parameter, nor- randomized patient not completing the treatment assignment centrally adju-
malized total atheroma volume, showed trial. The imputation technique assigned dicated major adverse cardiovascular
a greater reduction for pioglitazone com- changes in these efficacy end points events. Adjudicated end points in-
pared with glimepiride: −5.5 mm3 (95% for each noncompleter based on the pa- cluded cardiovascular and noncardio-
CI, −8.67 to −2.38 mm3) vs −1.5 mm3 tient’s baseline characteristics, includ- vascular death, nonfatal myocardial in-
(95% CI, −4.50 to 1.54 mm3) that did not ing demographics, laboratory values, and farction and stroke, hospitalization for
reach statistical significance (P = .06), baseline atheroma volumes. The result- unstable angina or congestive heart fail-
with a significant within-group de- ing analyses are shown in TABLE 4. ure, and coronary revascularization.
crease from baseline for the piogli- For the primary efficacy parameter, PAV, The investigators reported other listed
tazone group (P⬍.001) and no change the glimepiride group showed a mean adverse events, including hypoglyce-
for the glimepiride group (P=.34). A sec- increase of 0.64% (95% CI, 0.23% to mia, angina pectoris, peripheral edema,
ondary efficacy measure, change in the 1.05%) vs −0.062% (95% CI, −0.47% hypertension, and bone fractures.
10 mm of the most diseased segment to 0.35%) for the pioglitazone group The frequency of major adverse car-
showed no difference between the 2 (P=.02). Although all P values for IVUS diovascular events was similar in both
treatment groups. FIGURE 3 shows a con- efficacy parameters were reduced in sta- treatment groups, although the trial was
sistent treatment effect across mul- tistical significance using the imputed not powered to detect differences in
tiple prespecified and exploratory sub- results, the overall interpretation of the morbidity and mortality. Among inves-
groups with no statistical heterogeneity. study was not altered. tigator-reported adverse events, hypo-

Table 3. Baseline, Follow-up, and Change From Baseline in Intravascular Ultrasound End Points
Glimepiride Pioglitazone
(n = 181) (n = 179)
P
Mean (SD) Median (IQR) Mean (SD) Median (IQR) Value a
Baseline Examination
Percent atheroma volume, % b 40.3 (8.9) 40.3 (34.7 to 45.9) 40.6 (8.4) 40.3 (34.1 to 46.0) .54
Maximum atheroma thickness, mm c 0.82 (0.26) 0.80 (0.64 to 0.98) 0.81 (0.25) 0.79 (0.61 to 1.00) .94
Normalized total atheroma volume, c mm3 219.8 (95.2) 197.8 (148.1 to 277.7) 207.5 (83.8) 190.9 (147.6 to 254.5) .27
Atheroma volume in 10-mm most diseased 64.7 (31.5) 62.1 (40.9 to 86.6) 62.7 (28.1) 59.4 (43.6 to 78.7) .59
segment, c mm3
Follow-up Examination
Percent atheroma volume, % b 41.0 (9.0) 40.5 (35.2 to 46.9) 40.5 (8.5) 40.5 (33.6 to 46.3) .73
Maximum atheroma thickness, mm c 0.83 (0.26) 0.81 (0.64 to 0.99) 0.80 (0.24) 0.76 (0.62 to 0.97) .39
Normalized total atheroma volume, c mm3 217.7 (95.3) 192.6 (150.9 to 278.3) 200.8 (81.6) 184.5 (144.6 to 248.4) .13
Atheroma volume in 10-mm most diseased 62.4 (31.2) 57.8 (39.5 to 83.1) 60.0 (27.5) 57.9 (39.7 to 77.8) .62
segment, c mm3
Nominal Change From Baseline
LS Mean P Value Change LS Mean P Value Change P
(95% CI) From Baseline (95%CI) From Baseline Value d
Percent atheroma volume, % b 0.73 (0.33 to 1.12) ⬍.001 −0.16 (−0.57 to 0.25) .44 .002
Maximum atheroma thickness, mm c 0.011 (−0.0002 to 0.022) .054 −0.011 (−0.022 to 0.0004) .06 .006
Normalized total atheroma volume, c mm3 −1.5 (−4.50 to 1.54) .34 −5.5 (−8.67 to −2.38) ⬍.001 .06
Atheroma volume in 10-mm most diseased −2.1 (−3.33 to −0.84) .001 −2.0 (−3.33 to −0.67) .003 .93
segment, c mm3
Abbreviation: CI, confidence interval; IQR, interquartile range; LS, least squares.
a P values for baseline and average follow-up values were generated from a 2-way analysis of variance model with terms for treatment and pooled center.
b Primary efficacy parameter.
c Secondary efficacy parameter.
d P values from 2-way analysis of variance with terms for treatment and pooled center and baseline value as covariates.

©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, April 2, 2008—Vol 299, No. 13 1567

Downloaded from www.jama.com by guest on May 5, 2008

PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

glycemia was more common in the (P = .004). Decreases in hemoglobin COMMENT

glimepiride group 37.0% vs 15.2% were more common in the piogli- Atherosclerosis in patients with diabe-
(P⬍.001) and edema more common in tazone group (P =.01). An increase in tes is particularly aggressive, character-
the pioglitazone group, 17.8% vs 11% fractures and reductions in hemoglo- ized by higher cardiovascular event rates
(P = .02). Angina pectoris was re- bin level are both effects previously ob- and a greater severity of coronary ob-
ported in 12.1% of glimepiride- served with TZD therapy. More pa- structive disease.18-21 Cardiovascular dis-
treated patients and 7.0% of piogli- tients randomized to pioglitazone had ease represents the ultimate cause of
tazone-treated patients (P=.05). In the an increase in blood urea nitrogen lev- death in approximately 75% of pa-
pioglitazone group, fractures oc- els (P = .009), but there were no differ- tients with diabetes.22 Accordingly, de-
curred in 3.0% of patients compared ences in incidence of elevations in cre- fining the optimal strategy for manage-
with none in the glimepiride group atinine (Table 5). ment of coronary disease in the diabetic

Figure 3. Primary Efficacy Parameter (Percent Atheroma Volume) in Prespecified Subgroups

Percent Atheroma Volume,

No. of Patients LS Mean Change From Baseline, % (95% CI) Treatment
Difference, % Favors Favors
Baseline Characteristicsa Glimepiride Pioglitazone Glimepiride Pioglitazone (95% CI) P Value Pioglitazone Glimepiride

Prespecified subgroups
Age, y
≤60 93 93 0.59 (–0.01 to 1.19) –0.20 (–0.81 to 0.41) –0.79 (–1.61 to 0.04)
>60 .92
88 86 0.72 (0.12 to 1.33) –0.11 (–0.75 to 0.54) –0.83 (–1.67 to 0.01)
Sex
Men 119 131 0.51 (0.02 to 0.100) –0.18 (–0.66 to 0.30) –0.69 (–1.36 to –0.01)
Women .30
62 48 1.04 (0.28 to 1.80) –0.08 (–1.00 to 0.84) –1.12 (–2.28 to 0.04)
BMI
≤Median (31.7) 85 97 0.71 (0.08 to 1.34) –0.11 (–0.64 to 0.62) –0.72 (–1.58 to 0.14)
>Median .74
95 81 0.70 (0.13 to 1.28) –0.09 (–0.99 to 0.29) –1.05 (–1.91 to –0.20)
Systolic blood pressure,
mm Hg
<130 100 91 0.67 (0.06 to 1.27) –0.11 (–0.76 to 0.53) –0.78 (–1.64 to 0.07)
≥130 .85
81 88 0.72 (0.15 to 1.30) –0.09 (–0.67 to 0.48) –0.82 (–1.61 to –0.03)
Statin therapy
Yes 152 154 0.71 (0.27 to 1.15) –0.38 (–0.84 to 0.07) –1.09 (–1.72 to –0.47)
No .07
29 25 0.17 (–1.08 to 1.43) 0.92 (–0.45 to 2.29) 0.75 (–1.13 to 2.63)
Diabetes duration, mo
≤Median (71) 93 94 0.84 (0.24 to 1.43) –0.71 (–1.30 to –0.12) –1.54 (–2.38 to –0.71)
>Median .14
88 85 0.82 (0.17-1.48) 0.35 (–0.32 to 1.01) –0.48 (–1.35 to 0.40)
HbAIC, %
≤Median (7.3) 97 98 1.17 (0.61 to 1.73) 0.20 (–0.37 to 0.76) –0.97 (–1.75 to –0.19)
>Median .63
84 81 0.39 (–0.25 to 1.04) –0.39 (–1.07-0.28) –0.79 (–1.70 to 0.12)

Exploratory nonprespecified subgroups

HDL-C, mg/dL
≤Median (40) 98 79 0.91 (0.39 to 1.44) 0.08 (–0.53 to 0.68) –0.84 (–1.63 to –0.05)
>Median .72
79 96 0.35 (–0.31 to 1.02) –0.43 (–1.07 to 0.22) –0.78 (–1.68 to 0.12)
LDL-C, mg/dL
≤Median (90) 87 92 1.02 (0.43 to 1.61) 0.54 (–0.09 to 1.17) –0.48 (–1.30 to 0.35)
>Median .53
94 87 0.43 (–0.21 to 1.07) –0.75 (–1.38 to –0.12) –1.18 (–2.06 to –0.30)
Triglycerides, mg/dL
≤Median (143) 94 87 1.24 (0.66 to 1.82) –0.18 (–0.80 to 0.44) –1.42 (–2.24 to –0.59)
>Median .22
87 92 0.32 (–0.32 to 0.95) –0.21 (–0.85 to 0.43) –0.52 (–1.40 to 0.36)
Metabolic syndrome
Yes 148 147 0.64 (0.20 to 1.08) –0.27 (–0.73 to 0.18) –0.91 (–1.52 to –0.30)
No .94
33 32 0.75 (–0.51 to 2.02) –0.36 (–1.59-0.86) –1.12 (–3.17 to 0.93)
High-sensitivity CRP, mg/L
≤Median (2.8) 93 86 0.47 (–0.13 to 1.06) –0.37 (–1.01 to 0.28) –0.83 (–1.70 to 0.04)
>Median .75
84 89 0.86 (0.23 to 1.49) 0.07 (–0.56 to 0.70) –0.79 (–1.64 to 0.07)
Atheroma volume, %
≤Median (40.4) 90 88 0.51 (–0.13 to 1.15) –0.27 (–0.93 to 0.39) –0.77 (–1.68 to 0.13)
>Median .64
91 91 0.95 (0.40 to 1.50) –0.005 (–0.58 to 0.57) –0.96 (–1.74 to –0.17)

–3 –2 –1 0 1 2 3
Difference in LS Mean Change
From Baseline, % (95% CI)

P values for subgroup interaction were generated from a 2-way analysis of variance model with terms for treatment and pooled center and baseline value. BMI indi-
cates body mass index, calculated as weight in kilograms divided by height in meters squared; CI, confidence interval; CRP, C-reactive protein; HbA1c, hemoglobin A1c;
HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; LS, least squares.
aMedian values are shown in parentheses.

1568 JAMA, April 2, 2008—Vol 299, No. 13 (Reprinted) ©2008 American Medical Association. All rights reserved.

Downloaded from www.jama.com by guest on May 5, 2008

 PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

Table 4. Change From Baseline in Intravascular Ultrasound End Points Using Imputed Values for Noncompleters
Least Squares Mean Change From Baseline

Glimepiride Pioglitazone
(n = 273) (n = 270)
Between-
P Value (Change P Value (Change Group
LS Mean (95% CI) From Baseline) a LS Mean (95% CI) From Baseline) b P Value a
Percent atheroma volume, % b 0.64 (0.23 to 1.05) .003 −0.062 (−0.47 to 0.35) .77 .02
Maximum atheroma thickness, mm c 0.008 (−0.002 to 0.019) .12 −0.011 (−0.023 to 0.001) .07 .01
Normalized total atheroma volume, c mm3 −1.37 (−4.37 to 1.62) .37 −5.35 (−9.47 to −1.22) .01 .14
Atheroma volume in 10-mm most diseased −2.11 (−3.37 −0.85) .002 −2.52 (−3.73 to −1.31) ⬍.001 .66
segment, c mm3
Abbreviations CI, confidence interval, LS, least squares.
a P values from 2-way analysis of variance with terms for treatment, pooled center, and baseline value as covariates.
b Primary efficacy parameter.
c Secondary efficacy parameter.

population has important public health sistent with good diabetes management tions in the incidence of macrovascu-
implications. Although cardiovascular (ⱕ7.0%). Thus, pioglitazone showed the lar disease from much larger decreases
disease represents a critical source of ability to further reduce disease progres- in HbA 1 c levels than observed in
morbidity and mortality, there exist few sion on a background of contemporary PERISCOPE.24,25 In the United King-
data to support the preference of any spe- medical therapy. dom Prospective Diabetes Study (UK-
cific glucose-lowering regimen to pre- There are several potential explana- PDS), intensive therapy achieved a HbA1c
vent these complications.3 Since the in- tions for these findings. The sulfonyl- level of 7.0% vs 7.9% for standard
troduction of sulfonylureas nearly 50 urea, glimepiride, could have in- therapy but showed no reduction in
years ago, several additional classes of creased the rate of disease progression macrovascular complications.25 Re-
oral hypoglycemic agents have been in- or alternatively, pioglitazone could have cently, the intensive treatment arm of the
troduced. We sought to clarify the ef- exerted an antiatherosclerotic effect, or Action to Control Cardiovascular Risk
fects on coronary disease progression for both. We think that a proatherogenic in Diabetes (ACCORD) study was
2 of these classes, a sulfonylurea, an in- effect of glimepiride is unlikely. Low- stopped by the data monitoring com-
sulin secretagogue, and a TZD, an in- density lipoprotein cholesterol levels are mittee because the intensive treatment
sulin sensitizer, 2 diametrically oppo- highly predictive of coronary progres- group (median HbA 1c level 6.4%)
site strategies for management of sion rates in IVUS trials. The rate of pro- showed increased mortality compared
hyperglycemia. gression of atherosclerosis observed in with the standard treatment group (me-
The observation of a significant ben- the glimepiride treatment group is con- dian HbA1c levels 7.5%).26
efit for pioglitazone treatment represents, sistent with results in other IVUS trials The HbA1c levels observed in the
toourknowledge,thefirstdemonstration for patients treated to comparable PERISCOPE trial were approximately
of the ability of any hypoglycemic agent LDL-C levels.11-16 However, the rate of midway between the intensive and the
to slow the progression of coronary ath- progression in the pioglitazone group standard treatment groups in ACCORD
erosclerosisinpatientswithdiabetes.Evi- was substantially lower than might be and were similar to the intensive treat-
dence for a slowing of disease progression expected for a population with ob- ment group in UKPDS. Moreover, a
has proven a very challenging end point served LDL-C levels, particularly con- post hoc exploratory analysis examin-
in recent years with the prominent fail- sidering that prior IVUS studies have ing outcomes based on whether pa-
ure of several promising approaches.15,16 demonstrated a higher rate of coro- tients reached an HbA1c of 7% or lower
Many observers have attributed the dif- nary disease progression in patients showed that the achieved HbA1c level
ficulty in further reducing atherosclero- with diabetes compared with those who was not an important mediator of the
sis progression using novel therapies to do not have diabetes.23 atherosclerosis progression rate (P value
the effectiveness of contemporary medi- The precise mechanism for the ob- for interaction = .74). Accordingly, it
calmanagement,includingstatinsandan- served antiatherosclerotic effect of pio- seems unlikely that a 0.19% differ-
tihypertensiveagents.Inthecurrentstudy, glitazone remains uncertain. It seems un- ence in HbA1c levels could explain the
a high percentage of patients received op- likely that the small differences in HbA1c observed benefits of pioglitazone on dis-
timal treatment with nearly 90% of pa- levels between treatment groups (0.19%) ease progression.
tientsreceivinglipid-loweringtherapy.Av- could explain the observed differences Alternatively, differences between
erage LDL-C levels and blood pressures in the rate of disease progression. Stud- these 2 treatments in nonglycemic ef-
during treatment were below the current ies of glycemic management in type 2 fects seem more likely to explain the ob-
guideline targets. HbA1c levels were con- diabetes have failed to show reduc- served outcomes. Diabetes promotes
©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, April 2, 2008—Vol 299, No. 13 1569

Downloaded from www.jama.com by guest on May 5, 2008

PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

atherosclerosis in the setting of pro- a 16% increase in HDL-C, and substan- mains theoretically possible that an in-
found metabolic and physiologic ef- tial reductions in both triglyceride lev- sulin-sensitizing therapy confers
fects. Adverse metabolic effects in- els (15%) and C-reactive protein lev- independent antiatherosclerotic ben-
clude abnormalities in lipid metabolism, els (45%). A more favorable effect on efits over an insulin-providing treat-
an increased incidence of hyperten- blood pressure was also observed. Pio- ment strategy. These hypotheses will be
sion, promotion of coagulation, and glitazone also significantly reduced fast- explored in subsequent post hoc analy-
increased systemic inflammation.27 Sev- ing-serum insulin levels. Insulin resis- ses of the current trial.
eral biomarkers associated with ath- tance, resulting in hyperinsulinemia, is It cannot be assumed that the ob-
erosclerosis progression were favor- associated with adverse cardiovascu- served benefits of pioglitazone repre-
ably affected by pioglitazone, including lar outcomes. 28 Accordingly, it re- sent a “class effect” of TZDs. The only

Table 5. Clinical End Points, Adverse Events, Laboratory Abnormalities, and Reasons for Discontinuing Participation (Safety Population, n = 543)
No. (%) of Patients

Glimepiride Pioglitazone
(n = 273) (n = 270) P Value
Centrally Adjudicated Cardiovascular Events a
Composite of cardiovascular death, nonfatal MI, or nonfatal stroke 6 (2.2) 5 (1.9) .78
Composite of cardiovascular death, nonfatal MI, nonfatal stroke, 13 (4.8) 11 (4.1) .70
hospitalization for unstable angina, or congestive heart failure
Composite of cardiovascular death, nonfatal MI, nonfatal stroke, 41 (15.0) 40 (14.8) .95
coronary or carotid revascularization, hospitalization for
unstable angina, or congestive heart failure
Cardiovascular death 1 (0.36) 3 (1.1) .37
Noncardiovascular death 1 (0.36) 0 (0.0) ⬎.99
Nonfatal MI 4 (1.5) 2 (0.7) .69
Nonfatal stroke 1 (0.36) 0 (0.0) ⬎.99
Hospitalization for unstable angina 2 (0.7) 4 (1.5) .45
Coronary revascularization 30 (11.0) 29 (10.7) .93
Hospitalization for congestive heart failure 5 (1.8) 4 (1.5) ⬎.99
Investigator-Reported Adverse Events
Hypoglycemia 101 (37.0) 41 (15.2) ⬍.001
Angina pectoris 33 (12.1) 19 (7.0) .05
Peripheral edema 30 (11.0) 48 (17.8) .02
Hypertension 24 (8.8) 13 (4.8) .07
Bone fracture 0 (0) 8 (3.0) c .004
Central Laboratory Abnormalities
ALT ⬎3 ⫻ ULN 3 (1.1) 2 (0.7) ⬎.99
Serum urea nitrogen ⬎30 mg/dL 13 (4.8) 29 (10.7) .009
Creatinine ⬎2.0 mg/dL 2 (0.7) 3 (1.1) .69
Hematocrit decrease ⬎20% 3 (1.1) 8 (3.0) .12
Hemoglobin decrease ⬎3 g/dL 2 (0.7) 11 (4.1) .01
Reasons for Discontinuation of Drug Therapy (n = 192) b
Adverse event 34 (12.5) 30 (11.1) .63
Lack of efficacy 1 (0.4) 4 (1.5) .21
Lost to follow-up 6 (2.2) 4 (1.5) .75
Study termination at site 9 (3.3) 7 (2.6) .63
Protocol violation 3 (1.1) 6 (2.2) .34
Voluntary withdrawal by participant 34 (12.5) 40 (14.8) .42
Investigator’s discretion 8 (2.9) 6 (2.2) .60
Total Not Completing the Trial 95 (34.8) 97 (35.9) .78
Abbreviations: ALT, alanine aminotransferase; AST, aspartate aminotransferase; MI, myocardial infarction ULN, upper limit of normal.
SI conversation factors: To convert ALT to µkat/L, multiply by 0.0167; creatinine to µmol/L, multiply by 76.25; urea nitrogen to mmol/L, multiply by 0.357;
a Adverse events were collected for 30 days following each participant’s last dose.
b This represents patients who discontinued before the final visit (week 172) regardless of whether a final intravascular ultrasound was obtained.
c Six women, 2 men; location: foot, 2; humerus, 2; lower limb, 2; malleolar, 1; metacarpal, 1; facial, 1; rib, 1; upper limb, 1; wrist, 1 (total number ⬎8, because of multiple fractures
in the same patient).

1570 JAMA, April 2, 2008—Vol 299, No. 13 (Reprinted) ©2008 American Medical Association. All rights reserved.

Downloaded from www.jama.com by guest on May 5, 2008

 PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

other agent in the class, rosiglitazone, We recognize that the current study array of prespecified and exploratory
increases LDL-C approximately 18%, has limitations. The withdrawal rate of subgroups. These finding may have im-
and exhibits less favorable effects on patients in this trial (35%) was higher portant implications for defining the op-
other biomarkers, including HDL-C and than for most recent IVUS stud- timal strategy for management of pa-
triglycerides.5,29 In addition, rosiglita- ies.11,12,14-16 We attribute the higher drop- tients with type 2 diabetes and coronary
zone has been associated with an in- out rate to the difficulty of maintain- atherosclerosis.
creased risk of myocardial infarction ing patients on complex antidiabetic
Published Online: March 31, 2008 (doi:10.1001/
or other ischemic events in several regimens for the course of the trial. This jama.299.13.1561).
meta-analyses, whereas a similar meta- observation is consistent with the ap- Author Affiliations: Department of Cardiovascular
analysis of pioglitazone showed a re- proximate 30% withdrawal rate in a Medicine (Drs Nissen, Nicholls, Lincoff, and Tuzcu and
Ms Wolski) and Department of Quantitative Health
duction in the risk of adverse cardio- contemporary, noninvasive diabetes Sciences (Dr Hu), Cleveland Clinic, Cleveland, Ohio;
vascular outcomes.30,31 Furthermore, a trial of the same duration and using the Department of Cardiovascular Medicine, Lahey Clinic,
Burlington, Massachusetts (Dr Nesto); Department of
well-controlled randomized trial in an- same comparators.7 The demographic Clinical Science, Takeda Global Research and Devel-
other vascular bed, the carotid artery, and laboratory characteristics of com- opment, Deerfield, Illinois (Drs Kupfer and Perez); De-
partment of Cardiology, Clinica Chutro, Colon, Ar-
demonstrated an absence of progres- pleters and noncompleters were simi- gentina (Dr Jure); Department of Cardiology, Quebec
sion of intimal medial thickness for pio- lar and a post hoc sensitivity analysis Heart Institute, Hôpital Laval, Quebec City, Quebec
glitazone-treated patients and progres- imputing efficacy values for noncom- (Dr De Larochellière); Cardiovascular Division, St Vin-
cent’s Hospital Manhattan, New York, New York (Dr
sion in glimepiride-treated patients.7 A pleters did not alter interpretation of the Staniloae); Department of Medicine, Division of Car-
placebo-controlled clinical outcomes trial. Intravascular ultrasonography diology, Atlanta VA Medical Center, Atlanta, Geor-
gia (Dr Mavromatis), and Division of Cardiology, Van-
trial investigating pioglitazone re- evaluates the effect of therapies on the couver General Hospital, Vancouver, British Columbia,
ported a reduction in major adverse car- change in atherosclerotic disease bur- Canada (Dr Saw).
Author Contributions: Dr Nissen had full access to all
diovascular events that did not reach den, not morbidity and mortality. The of the data in the study and takes responsibility for
conventional levels of statistical sig- magnitude of the differences in pro- the integrity of the data and the accuracy of the data
analysis.
nificance (hazard ratio, 0.90; 95% CI, gression rates observed in the current Study concept and design: Nissen, Nesto, Kupfer, Perez,
0.80-1.02; P=.095), but showed a sig- trial are similar to the results reported Tuzcu.
nificant reduction in the composite of for other beneficial therapies, such as Acquisition of data: Nissen, Kupfer, Perez, Jure,
De Larochellière, Staniloae, Mavromatis, Saw, Tuzcu.
death, myocardial infarction, and stroke statins or blood pressure–lowering Analysis and interpretation of data: Nissen, Nicholls,
(hazard ratio, 0.84; 0.72-0.98, P=.027).6 agents.11,12,14 However, clinical out- Wolski, Nesto, Kupfer, Perez, Hu, Lincoff, Tuzcu.
Drafting of the manuscript: Nissen.
The benefits of any therapy must al- comes trials, not surrogate end point Critical revision of the manuscript for important in-
ways be evaluated in the context of treat- studies, are the preferred approach to tellectual content: Nissen, Nicholls, Wolski, Nesto,
Kupfer, Perez, De Larochellière, Jure, Staniloae,
ment-emergent adverse effects. In the the determination of the benefits of any Mavromatis, Saw, Hu, Lincoff, Tuzcu.
current study, both regimens were well therapeutic intervention, particularly in Statistical analysis: Nissen, Wolski, Nesto, Hu.
tolerated but exhibited a different pat- the context of the current controversy Obtained funding: Nissen, Perez, Tuzcu.
Administrative, technical, or material support: Nissen,
tern of adverse effects. Hypoglycemia regarding the cardiovascular effects of Nicholls, Kupfer, Perez, De Larochellière, Saw, Lincoff,
was more prevalent in the glimepiride TZDs. Unfortunately, few current trials Tuzcu.
Study supervision: Nissen, Kupfer, Perez, Mavromatis.
group and edema more common in the are comparing clinical outcomes for al- Financial Disclosures: Dr Nissen reports that the Cleve-
pioglitazone group, but neither case ternative antidiabetic therapies. A re- land Clinic Coordinating Center for Clinical Research
has received research support to perform clinical trials
resulted in a difference in rates of dis- cently announced large comparative from Pfizer, Astra Zeneca, Sankyo, Takeda, sanofi-
continuation of therapy (Table 5). Frac- trial of several diabetes therapies, in- aventis, Lilly, Roche, Daiichi-Sankyo, and Novartis. Dr
tures occurred in 3.0% of pioglitazone- cluding rosiglitazone, will not report Nissen consults for many pharmaceutical companies,
but requires them to donate all honoraria or consult-
treated patients and none of the findings until 2014.32 ing fees directly to charity so that he receives neither
glimepiride-treated patients. Patients Despite these limitations, the cur- income nor a tax deduction. Dr Nicholls reports re-
ceiving honoraria from Pfizer, AstraZeneca, Takeda,
gained weight on both regimens, but the rent study provides potentially useful and Merck Schering-Plough, consultancy fees from
weight gain was approximately 2 kg insights into the cardiovascular ef- AstraZeneca, Pfizer, Roche, Novo-Nordisk, Liposcience,
and Anthera Pharmaceuticals and research support
greater in the pioglitazone group. There fects of these 2 diabetes treatment strat- from AstraZeneca and Lipid Sciences. Dr Lincoff
were no observed differences in major egies. The findings of the PERISCOPE reports receiving research support to perform clinical
trials through the Cleveland Clinic Coordinating
cardiovascular morbidity or mortality, study support the conclusion that treat- Center for Clinical Research from Takeda, sanofi-
although the trial was not powered to ment with the insulin-sensitizing TZD, aventis, Eli Lilly, Pfizer, Centocor, Dr Reddy’s
assess clinical outcomes. We therefore pioglitazone compared with glimepi- Laboratory, Medicines Company, Roche, Daiichi-
Sankyo, Schering-Plough, Scios, and Astra-Zeneca. Dr
conclude reduced atherosclerosis pro- ride can prevent the progression of ath- Lincoff reports that he consults for a number of phar-
gression was associated with a reason- erosclerosis in patients with type 2 dia- maceutical companies, but requires them to donate
all honoraria or consulting fees directly to charity so
able safety profile, generally compa- betes during 18 months of treatment. that he receives neither income nor tax deductions.
rable with glimepiride, although there Patients randomized to pioglitazone ex- Drs Kupfer and Perez are employees of Takeda. Dr
Nesto reports receiving honoraria from GlaxoSmith-
were distinct differences in the types of hibited a lower rate of progression of Kline, Takeda, Merck, Pfizer, and sanofi-aventis. Dr
adverse effects for the 2 drugs. coronary atherosclerosis across a wide Saw reports receiving honoraria from sanofi-aventis,

©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, April 2, 2008—Vol 299, No. 13 1571

Downloaded from www.jama.com by guest on May 5, 2008

PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

GlaxoSmithKline, and Cordis. Ms Wolski reports no versity of California San Diego Medical Center, San lottesville, M. Ragosta; Sentara Norfolk General Hos-
relationships. Dr Tuzcu reports receiving consultancy Diego; Theo Meyer, MD, University of Massachu- pital, R. Stine; Hunter Holmes McGuire VA Medical
fees from Pfizer and honoraria from Pfizer and Merck. setts Medical Center, Worcester. Center, Richmond, A. Minisi; Washington: The Ever-
Funding/Support: This study was financially sup- In addition to the authors, the following investiga- ett Clinic, Everett, P. Huang, M. Mulumudi; and Wis-
ported by Takeda Pharmaceuticals North America Inc. tors participated in this study: consin: Cardiology Associates, Green Bay, D. Jenny.
Role of the Sponsor: The sponsor participated in dis- United States: Alabama: University of Alabama at Canada: Alberta: University of Alberta Hospital, Ed-
cussions regarding study design and protocol devel- Birmingham Hospital, V. Misra; California: Diabetes monton, J. Burton; British Columbia: Vancouver Hos-
opment and provided logistical support during the trial. and Lipid Management Center, Huntington Beach, P. pital and Health Sciences Center, C. Buller, J. Saw; Vic-
Monitoring of the study was performed by the spon- Rosenblit; LAC-USC Medical Center, Los Angeles, R. toria Heart Institute, J. D. Hilton; Manitoba: Health
sor, who also maintained the trial database. The IVUS Sarma; UCLA Interventional Cardiology Research, Los Sciences Centre, Winnipeg, J. Ducas; New Bruns-
end points were determined by the Cleveland Clinic Angeles, J. Tobis; Mercy Heart Institute, Sacra- wick: St John Regional Hospital, R. Teskey; Ontario:
Coordinating Center for Clinical Research. Statistical mento, K. Singh; Naval Medical Center, San Diego, London Health Sciences Centre, W. Kostuk; Univer-
analyses were independently performed by both the J. Cavendish, C. Ledford; San Diego VA Medical Cen- sity of Ottawa Heart Institute, M. LeMay; and Que-
sponsor and the coordinating center and the results ter, W. Penny; University of California San Diego bec: Montreal: Institut de Cardiologie de Montreal,
cross-checked by statisticians with both organiza- Medical Center, E. Mahmud; St Joseph’s Medical Cen- R. Bonan; CHUM–Hopital Saint-Luc, A. Kokis; Chum
tions. The manuscript was prepared by the corre- ter, Stockton, H. Madyoon; Los Angeles Biomedical – Hopital Notre-Dame, F. Reeves; Hopital du Sacre-
sponding author and modified after consultation with Research, Torrance, M. Budoff; Connecticut: Cardi- Coeur de Montreal, M. Doucet; Hopital Maisonneuve-
the other authors. The sponsor was permitted to re- ology Associates of Southern Connecticut, Bridge- Rosemont, C. Constance; and Sainte Foy: Hopital La-
view the manuscript and suggest changes, but the fi- port, C. Landau; University of Connecticut Health Cen- val, R. De Larochellière.
nal decision on content was exclusively retained by ter, Farmington, M. Azrin; District of Columbia: George South America: Argentina: Buenos Aires: Instituto Car-
the academic authors. The study contract specified Washington University Hospital, J. Reiner; Florida: Bay diovascular de Buenos Aires, J. Belardi; Sanatorio Mi-
transfer of a copy of the study database to the coor- Pines VA Medical Center, T. Theodoropoulos, M. Van- tre, E. Blumberg; Sanatorio Guemes, G. Bortman; Hos-
dinating center for independent analysis and granted dormael; Jacksonville Center for Clinical Research, M. pital Italiano, A. Fernandez; Instituto Dupuytren, P.
the academic authors the unrestricted rights to pub- Koren; University of Florida Health Sciences Center, Kantor; Hospital Britanico de Buenos Aires, C. Majul;
lish the results. Jacksonville, M. Costa; Georgia: Atlanta VA Medi- Fundacion Favaloro, O. Mendiz; Instituto Alexander
cal Center, Decatur, K. Mavromatis; Indiana: Cardio- Fleming, A. Pocovi; Cordoba: Hospital Privado, M.
Independent Statistical Analysis: An independent sta-
Amuchastegui; Sanatorio Allende, L. Guzman; Sana-
tistical analysis was conducted by Kathy Wolski, MPH, vascular Clinics, PC, Merrillville, T. Nguyen; Ken-
torio Parque, H. Luquez; Clinica Chutro, H. Jure.
from the Department of Cardiovascular Medicine at tucky: St Elizabeth Medical Center, Edgewood, D.
Chile: Santiago: Hospital Clinico de la Pontificia Uni-
the Cleveland Clinic Foundation. Ms Wolski received Courtade; University of Kentucky Medical Center, Lex-
versidad Catolica de Chile, R. Corbalan, E. Guarda;
the trial database from the sponsor, which included ington, K. Ziada; University of Louisville, S. Ikram, M.
Hospital Clinico de la Universidad de Chile, G. Dus-
all the raw data, not just derived datasets; and inde- Leesar; Louisiana: Ochsner Clinic Foundation, New
saillant; and Hospital San Juan de Dios, A. Bellet.
pendently computed the IVUS efficacy parameters, Orleans, S. Ramee, J. Reilly; Maine: Northeast Car-
Additional Contributions: We thank the following in-
safety measures, laboratory parameters, and demo- diology Associates, Bangor, M. Rowe; Cardiovascu-
dividuals in conducting this study: Takeda Global Re-
graphic variables. Ms Wolski is employed by the Cleve- lar Consultants of Maine, Scarborough, W. Dietz;
search and Development: Brigit Isaacson, MT (ASCP),
land Clinic Cardiovascular Coordinating Center, which Michigan: McLaren Hospital, Flint, A. DeFranco, W.
MBA, Pat Naughton, RN, MS, Yinzhong Chen, PhD,
received compensation from the sponsor for conduct- Rivera; Borgess Research Institute, Kalamazoo, T. Fisch-
and Andrew Roberts, MS. Cleveland Clinic Coordi-
ing the trial, including reimbursement for statistical ser- ell; Cardiology Consultants, PC, Petoskey, H. Colfer; nating Center for Clinical Research: Craig Balog, BS,
vices. St Joseph Mercy Medical Center – Oakland, Pon- and Amy Hsu, MS. Cleveland Clinic Intravascular Ul-
An additional independent statistical analysis was per- tiac, K. Patel; William Beaumont Hospital, Royal Oak, trasound-Angiography Core Laboratory: Sorin Brener,
formed by Bo Hu, PhD, from the Department of Quan- C. Grines; Michigan Cardiovascular Institute, Sagi- MD, Tim Crowe, BS, William Magyar, BS, Jordan An-
titative Health Sciences at Cleveland Clinic. Dr Hu is naw, R. Gudipati; Michigan Heart, PC, Ypsilanti, B. drews, BS, Eva Balazs, BS, Tammy Churchill, BS, Anne
also an adjunct assistant professor in the Department McCallister; Minnesota: Abbott Northwestern Hos- Colagiovanni, BS, Kelly Emerick, BS, Jessica Fox, BA,
of Statistics at Case Western Reserve University. Dr pital, Minneapolis, I. Chavez; St Mary’s Duluth Clinic Teresa Fonk, BS, Tom Ivanc, MS, Karilane King, MS,
Hu received the entire trial database and indepen- Health System, A. Deibele; Central Minnesota Heart Aaron Loyd. BS, Erin Mayock, BS, Kara McInturff, BS,
dently confirmed Ms Wolski’s findings. Dr Hu re- Center, St Cloud, B. Erickson; Missouri: University of Roman Poliszczuk, BS, Rhiannon Regal, BS, Andrea
ceived compensation from the coordinating center, but Missouri Hospitals and Clinics, Columbia, R. Webel; Winkhart, BS, and Jay Zhitnik, BS. The Cleveland Clinic
not from the sponsor, for statistical services. The re- New Jersey: Cooper Health System, Camden, S. Werns; employees did not receive any specific compensation
sults reported in this article are based on the indepen- New York: Capital Cardiology Associates, Albany, A. for participation in this clinical trial.
dent analyses. DeLago; Montefiore Medical Center, Bronx, V.
Investigators and Committees of the Piogl- Srinivas; StVincent’s Hospital-Manhattan, C. Sta-
itazone Effect on Regression of Intravascular Sono- niloae; Winthrop University Hospital, Mineola, T. Ker- REFERENCES
graphic Coronary Obstruction Prospective Evalu- win; Mt Sinai Medical Center, New York, B. Baliga,
ation (PERISCOPE) Trial R. Yanagisawa; Southbay Cardiovascular Associates, 1. UK Prospective Diabetes Study (UKPDS) Group. In-
Executive Committee: Steven E. Nissen, MD (Chair), West Islip, P. Lee; Buffalo Cardiology and Pulmo- tensive blood-glucose control with sulphonylureas or
Cardiovascular Medicine, Cleveland Clinic, Ohio; nary Associates, Williamsville, J. Corbelli; Ohio: Uni- insulin compared with conventional treatment and risk
Richard Nesto, MD, Lahey Clinic, Burlington, Massa- versity Hospitals of Cleveland, R. Nair; The Ohio State of complications in patients with type 2 diabetes (UK-
chusetts; Ralph D’Agostino Sr, PhD, Boston Univer- University Medical Center, Columbus, R. Magorien; PDS 33) [published correction appears in Lancet.
sity, Boston, Massachusetts; Alfonso Perez, MD, Tak- University of Cincinnati, T. Helmy, S. A. Najeed, J. 1999;354(9178):602]. Lancet. 1998;352(9131):837-
eda Global Research and Development Center, Schmidt; North Ohio Heart Center, Westlake, I. 853.
Deerfield, Illinois (nonvoting). Ahmed; Oklahoma: Southwest Cardiology Associ- 2. Stettler C, Allemann S, Jüni P, et al. Glycemic con-
Steven M. Haffner, MD, University of Texas Health ates, Oklahoma City, M. Yasin; Oklahoma VA Medi- trol and macrovascular disease in types 1 and 2 dia-
Science Center, San Antonio, was a member of the cal Center, Oklahoma City, E. Schechter; University betes mellitus: meta-analysis of randomized trials. Am
Executive Committee until November 2007, when he of Oklahoma Health Sciences Center, Oklahoma City, Heart J. 2006;152(1):27-38.
was asked to resign. J. Saucedo; Oklahoma Heart Institute, Tulsa, W. Leim- 3. Bolen S, Feldman L, Vassy J, et al. Systematic re-
Data and Safety Monitoring Board: Edward Horton, bach; Pennsylvania: Bryn Mawr Hospital, J. Martin; view: comparative effectiveness and safety of oral
MD, Boston, Massachusetts; George Bakris, MD, Uni- Milton S. Hershey Medical Center, Hershey, M. Kozak; medications for type 2 diabetes mellitus. Ann Intern
versity of Chicago, Chicago, Illinois; Sidney Gold- University of Pittsburgh Medical Center, O. Marro- Med. 2007;147(6):386-399.
stein, MD, Henry Ford Hospital, Detroit, Michigan; quin; Rhode Island: Rhode Island Hospital, Provi- 4. Staels B, Fruchart JC. Therapeutic roles of peroxi-
Barry Davis, MD, PhD, University of Texas School of dence, B. Sharaf. South Carolina: Upstate Cardiol- some proliferator-activated receptor agonists. Diabetes.
Public Health, Houston; Spencer King, MD, Fuqua ogy PA, Greenville, S. Johnson; Texas: Amarillo Heart 2005;54(8):2460-2470.
Heart Center, Atlanta, Georgia, James Harper, PhD, Group, Amarillo, E. Rivera; Houston Northwest Medi- 5. Goldberg RB, Kendall DM, Deeg MA, et al. A com-
Omnicare Clinical Research, Covington, Kentucky. cal Center, Houston, H. S. Bhatia, I. Lieber; Texas Car- parison of lipid and glycemic effects of pioglitazone and
Clinical End Point Committee: William Gaasch, MD diac Center, Lubbock, F. Shoukfeh; University of Texas rosiglitazone in patients with type 2 diabetes and
(Chairman) Director, Cardiovascular Research, La- Health Science Center, San Antonio, M. Feldman; Ten- dyslipidemia. Diabetes Care. 2005;28(7):1547-1554.
hey Clinic, Burlington, Massachusetts; I. K. Jang, MD, nessee: Knoxville Cardiovascular Group, Knoxville, D. 6. Dormandy JA, Charbonnel B, Eckland DJA, et al.
Mass General Hospital, Boston, Massachusetts; Ajit B. Wortham; VA Memphis, K. B. Ramanathan, Park- Secondary prevention of macrovascular events in pa-
Raisinghani, MD, University of California San Diego way Cardiology Associates, Oak Ridge, M. Sharma; tients with type 2 diabetes in the PROactive Study
Medical Center, San Diego; Ori Ben-Yehuda, MD, Uni- Virginia: University of Virginia Health System, Char- (PROspective pioglitAzone Clinical Trial In macroVas-

1572 JAMA, April 2, 2008—Vol 299, No. 13 (Reprinted) ©2008 American Medical Association. All rights reserved.

Downloaded from www.jama.com by guest on May 5, 2008

 PIOGLITAZONE VS GLIMEPIRIDE AND CORONARY ATHEROSCLEROSIS PROGRESSION

cular Events): a randomised controlled trial. Lancet. pears in N Engl J Med. 2006;355(6):638. N Engl J Med. bolic control and prevalent cardiovascular disease in
2005;366(9493):1279-1289. 2006;354(12):1253-1263. non-insulin-dependent diabetes mellitus (NIDDM): The
7. Mazzone T, Meyer PM, Feinsein SB, et al. Effect 16. Nissen SE, Tardif JC, Nicholls SJ, et al; ILLUS- NIDDM Patient Outcome Research Team. Am J Med.
of pioglitazone compared with glimepiride on ca- TRATE Investigators. Effect of torcetrapib on the pro- 1997;102(1):38-47.
rotid intima-media thickness in type 2 diabetes; a ran- gression of coronary atherosclerosis [published cor- 25. UK Prospective Diabetes Study (UKPDS) Group.
domized trial. JAMA. 2006;296(21):2572-2581. rection appears in N Engl J Med. 2007;357(8):835]. Intensive blood-glucose control with sulphonylureas
8. Eurich DT, McAlister FA, Blackburn DF, et al. Ben- N Engl J Med. 2007;356(13):1304-1316. or insulin compared with conventional treatment and
efits and harms of antidiabetic agents in patients with 17. Schoenhagen P, Sapp SK, Tuzcu EM, et al. Vari- risk of complications in patients with type 2 diabetes
diabetes and heart failure: systematic review. BMJ. ability of area measurements obtained with different (UKPDS 33). Lancet. 1998;352(9131):837-853.
2007;335(7618):497. intravascular ultrasound catheter systems: impact on 26. For Safety, NHLBI Changes Intensive Blood Sugar
9. Nicholls SJ, Sipahi I, Schoenhagen P, Crowe T, Tuzcu clinical trials and a method for accurate calibration. Treatment Strategy in Clinical Trial of Diabetes and
EM, Nissen SE. Application of intravascular ultra- J Am Soc Echocardiogr. 2003;16(3):277-284. Cardiovascular Disease [news release]. National Heart,
sound in anti-atherosclerotic drug development. Nat 18. Haffner SM, Lehto S, Ronnemaa T, Pyorala K, Lung, and Blood Institute Communications Office; Feb-
Rev Drug Discov. 2006;5(6):485-492. Laakso M. Mortality from coronary heart disease in ruary 3, 2006. http://public.nhlbi.nih.gov/newsroom
10. Nissen SE, Tsunoda T, Tuzcu EM, et al. Effect of subjects with type 2 diabetes and in nondiabetic sub- /home/GetPressRelease.aspx?id=2551 Accessed
recombinant ApoA-I Milano on coronary atheroscle- jects with and without prior myocardial infarction. March 7, 2008.
rosis in patients with acute coronary syndromes: a ran- N Engl J Med. 1998;339(4):229-234. 27. Beckman JA, Creager MA, Libby P. Diabetes and
domized controlled trial. JAMA. 2003;290(17):2292- 19. Brener SJ, Lytle BW, Casserly IP, Topol EJ, Lauer atherosclerosis: epidemiology, pathophysiology, and
2300. MS. Propensity analysis of long-term survival after sur- management. JAMA. 2002;287(19):2570-2582.
11. Nissen SE, Tuzcu EM, Brown BG, et al. Effect of gical or percutaneous revascularization in patients with 28. Fontbonne A, Charles MA, Thibult N, et al. Hy-
intensive compared with moderate lipid-lowering multivessel coronary disease and high-risk features. perinsulinaemia as a predictor of coronary heart dis-
therapy on progression of coronary atherosclerosis: a Circulation. 2004;109(19):2290-2295. ease mortality in a healthy population: the Paris Pro-
randomized controlled trial JAMA. 2004;291(9): 20. Scholte AJ, Schuijf JD, Kharagjitsingh AV, Jukema spective Study, 15-year follow-up. Diabetologia. 1991;
1071-1080. W, Pundziute G, van der Wall EE, Bax JJ. Prevalence 34(5):356-361.
12. Nissen SE, Tuzcu EM, Libby P, et al; CAMELOT of coronary artery disease and plaque morphology as- 29. Avandia (rosiglitazone maleate) tablets: prescrib-
Investigators. Effect of antihypertensive agents on car- sessed by multi-slice computed tomography coro- ing information [package insert]. Research Triangle
diovascular events in patients with coronary disease nary angiography and calcium scoring in asymptom- Park, NC: GlaxoSmithKline; 2007. http://www.fda
and normal blood pressure: the CAMELOT study: a atic patients with type 2 diabetes. Heart. 2008;94 .gov/cder/foi/label/2007/021071s023lbl.pdf. Ac-
randomized controlled trial. JAMA. 2004;292(18): (3):290-295. cessed March 6, 2008.
2217-2225. 21. Waller BF, Palumbo PJ, Lie JT, Roberts WC. Sta- 30. Nissen SE, Wolski K. Effect of rosiglitazone on the
13. Nissen SE, Tuzcu EM, Schoenhagen P, et al; Re- tus of the coronary arteries at necropsy in diabetes risk of myocardial infarction and death from cardio-
versal of Atherosclerosis with Aggressive Lipid Low- mellitus with onset after age 30 years: analysis of 229 vascular causes [published correction appears in N Engl
ering (REVERSAL) Investigators. Statin therapy, LDL diabetic patients with and without clinical evidence of J Med. 2007;357(1):100]. N Engl J Med. 2007;356
cholesterol, C-reactive protein, and coronary artery coronary heart disease and comparison to 183 con- (24):2457-2471.
disease. N Engl J Med. 2005;352(1):29-38. trol subjects. Am J Med. 1980;69(4):498-506. 31. Lincoff AM, Wolski K, Nicholls SJ, Nissen SE. Pio-
14. Nissen SE, Nicholls SJ, Sipahi I, et al; ASTEROID 22. Gu K, Cowie CC, Harris MI. Diabetes and de- glitazone and risk of cardiovascular events in patients
Investigators. Effect of very high-intensity statin cline in heart disease mortality in US adults. JAMA. with type 2 diabetes mellitus: a meta-analysis of ran-
therapy on regression of coronary atherosclerosis: the 1999;281(14):1291-1297. domized trials. JAMA. 2007;298(10):1180-1188.
ASTEROID trial. JAMA. 2006;295(13):1556-1565. 23. Nicholls SJ, Tuzcu EM, Kalidindi S, et al. Effect of 32. Rubin R. Heart attack warning added to Avandia
15. Nissen SE, Tuzcu EM, Brewer HB, et al; ACAT Intra- diabetes mellitus on progression of coronary athero- label: FDA also asks drugmaker to conduct study. USA
vascularAtherosclerosisTreatmentEvaluation(ACTIVATE) sclerosis and arterial remodeling. J Am Coll Cardiol. TODAY. http://www.usatoday.com/printedition/life
Investigators. Effect of ACAT inhibition on the progres- In press. /20071115/d_avandia15.art.htm. Accessibility veri-
sion of coronary atherosclerosis [published correction ap- 24. Meigs JB, Singer DE, Sullivan LM, et al. Meta- fied March 9, 2008.

©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, April 2, 2008—Vol 299, No. 13 1573

Downloaded from www.jama.com by guest on May 5, 2008