Lower Risk of Hypoglycemia with Insulin Detemir than with

Neutral Protamine Hagedorn Insulin in Older Persons with Type 2

Diabetes: A Pooled Analysis of Phase III Trials
Alan J. Garber, MD, PhD, Per Clauson, MD, PhD,w Claus B. Pedersen, MSc,w and
Klaus Klendorf, MD, DMScz

OBJECTIVES: To compare the safety and efficacy of insu- and younger persons with DM at similar levels of HbA1c.
lin detemir with that of neutral protamine Hagedorn (NPH) J Am Geriatr Soc 55:1735–1740, 2007.
insulin in older (aged
65) and younger (aged 18–64) per- Key words: insulin; hypoglycemia; aged; type 2 diabetes
sons with type 2 diabetes mellitus (DM). mellitus
DESIGN: Pooled, post hoc analysis of data from three
open-label, randomized studies.
SETTING: Three multinational Phase III trials.
PARTICIPANTS: Four hundred sixteen older and 880
younger persons with DM, treated for 22 to 26 weeks with
basal insulin plus mealtime insulin or oral agents.
MEASUREMENTS: Hemoglobin A1c (HbA1c), fasting
plasma glucose, glucose variability, hypoglycemic episodes. T ight glycemic control in persons with type 2 diabetes
mellitus (DM) has been shown to decrease the inci-
dence and progression of late diabetic complications.1 As a
RESULTS: Mean treatment difference for HbA1c (insulin result, national diabetes organizations have recommended
detemir–NPH insulin) indicated that insulin detemir was hemoglobin A1c (HbA1c) targets of 6.5% or less (American
not inferior to NPH insulin for both age groups (0.035%,
Association of Clinical Endocrinologists)2 and less than
95% confidence interval (CI) 5  0.114–0.183 and
7.0% (American Diabetes Association; ADA).3 Currently,
0.100%, 95% CI 5  0.017–0.217, for older and younger
only approximately half of U.S. adults reach the more-con-
persons, respectively). Relative risk of all hypoglycemic
servative ADA target (o7.0%).4 Hypoglycemia is an im-
episodes (insulin detemir/NPH insulin) was 0.59 (95% CI-
portant, although by no means the only, barrier to achieving
0.42–0.83) for older persons and 0.75 (95% CI-0.59–0.96)
for younger persons. Adverse events were similar between this goal. Unfortunately, the incidence of hypoglycemia in-
treatments. Fasting plasma glucose was similar between creases with more-aggressive treatment regimens.1 Also, as
treatments (mean treatment difference 0.97 mg/dL, 95% glycemic control improves, the plasma glucose threshold for
CI 5  8.01–9.95, and 4.69 mg/dL, 95% CI 5  2.30– counterregulatory responses and reaction time decreases,
11.67, for older and younger persons, respectively). Mean which narrows the window for appropriate intervention to
treatment difference for weight was  1.02 kg (95% CI occur.5 This also potentially increases the risk of severe hy-
 1.61 to  0.42) and  1.13 (95% CI  1.58 to  0.69) poglycemia.
for older and younger persons, respectively. The reported incidence of severe or major hypoglyce-
CONCLUSION: Previously reported benefits of insulin mia (e.g., variably defined as requiring the assistance of a
detemir, particularly less hypoglycemia and less weight third party or hospitalization; requiring intravenous glu-
gain, compared with NPH insulin, were the same for older cose or glucagon; blood glucose o50 mg/dL) for patients
with DM has been variable, but most studies suggest that it
occurs much less frequently in patients with type 2 DM,
than in those with type 1.6–8 It is likely that the true fre-
quency is underestimated, particularly in elderly people,
From the Division of Diabetes, Endocrinology and Metabolism, Baylor because of inclusion in most surveys of people in the early
College of Medicine, Houston, Texas; wGlobal Development, Novo Nordisk,
Bagsvaerd, Denmark; zDiabetes Department, Kge Hospital, stages of the disease, who have treatment regimens posing
Kge, Denmark. little risk of hypoglycemia. However, in one study, con-
Address correspondence to Alan J. Garber, MD, PhD, Baylor College of ducted in community diabetes clinics, when sequential
Medicine, Faculty Center, Suite 1000, 1709 Dryden Rd, BCM-620, Houston, patients were matched for duration of insulin therapy,
TX 77030. E-mail: agarber@bcm.tmc.edu there was no difference in the incidence of severe hypogly-
DOI: 10.1111/j.1532-5415.2007.01414.x cemia between people with type 1 and type 2 DM.9 This is

JAGS 55:1735–1740, 2007

r 2007, Copyright the Authors
Journal compilation r 2007, The American Geriatrics Society 0002-8614/07/$15.00
1736 GARBER ET AL. NOVEMBER 2007–VOL. 55, NO. 11 JAGS

consistent with the reduced counterregulatory responses sulin analogs, or OADs for at least 2 months and HbA1c of
reported for patients in the later stages of DM.10,11 12.0% or less. In Study 3,19 only subjects with HbA1c of
Hypoglycemia is of particular concern for older per- 7.5% to 10.0% were enrolled. Persons using high insulin
sons, who may be more vulnerable because of reduced doses (41.4 U/kg per day), a basal insulin dose less than
insulin clearance, associated comorbidities, and possibly 30% of the total daily insulin dose, or having major diabetic
a greater risk of hypoglycemia-associated autonomic complications or other significant diseases, such as liver and
failure.8,10–12 A laboratory study of two groups of men, kidney diseases, were not included in the studies.
seven aged 60 to 70 and seven aged 22 to 26, showed that Insulin detemir or NPH insulin was administered once
older persons with DM were more prone to severe cognitive or twice daily in combination with bolus insulin (insulin
impairment and to have poorer subjective awareness that aspart or human soluble insulin) or with OAD treatment for
their blood glucose was falling into the hypoglycemic range 22 to 26 weeks. The morning dose of basal insulin was
than younger persons.13 Another consideration is that administered before or at breakfast. The evening dose of
older persons often live alone, and consequently, fear of basal insulin was given at bedtime in Study 117 and Study
hypoglycemia may be a significant barrier to adherence in 2,18 whereas in Study 3,19 insulin detemir was administered
this age group. 1 hour before dinner or 1 hour after dinner, and NPH
One advance in the treatment of DM is the devel- insulin was administered within 1 hour before dinner
opment of insulin analogs, which have improved pharma- or at bedtime. The insulin dose was adjusted during the
cokinetic and pharmacodynamic profiles over those of studies to achieve the following targets: fasting blood
conventional human insulin products.14 Insulin detemir glucose (FBG) 72 to 126 mg/dL (4–7 mmol/L),17 FBG less
is a long-acting basal insulin analog that has a prolonged than 108 mg/dL (o6 mmol/L),18 fasting plasma glucose
duration of action without a pronounced peak.15,16 Several (FPG) less than 108 mg/dL,19 postprandial blood glu-
individual clinical studies have reported that it has a similar cose less than 180 mg/dL (o10 mmol/L) 17 or less than
effect on HbA1c as neutral protamine Hagedorn (NPH) in- 162 mg/dL (o9 mmol/L).18
sulin but with a lower risk of hypoglycemia and less weight
gain than NPH insulin in general populations of patients Efficacy and Safety Measures
with DM.17–20 To evaluate whether similar benefits were
HbA1c at the end of treatment was the primary endpoint in
obtained for older (aged
65) as well as for younger (aged
all three studies. Secondary endpoints and their methods of
18–64) persons, data were pooled from three published
assessment differed somewhat between the studies. Data on
Phase III studies involving persons with DM.17–19 In addi-
glucose variability were pooled from the two basal-bolus
tion, the effect of treatment with insulin detemir on the
studies (Studies 1 and 2).17,18 In Study 1, variability was
general health profile of older persons was assessed.
assessed as within-person variation in self-measured FBG
(SMFBG) during the final 7 days of treatment,17 whereas in
METHODS Study 2, it was assessed during the final 4 days of treat-
This was a pooled analysis of three multicenter, open-label, ment.18 Glucose variability data from Study 3 were not
parallel, randomized clinical Phase III studies (described included, because that trial was not a basal-bolus study.
below) of 22 to 26 weeks’ duration to compare the effect of Laboratory values for FPG were reported in Studies 117 and
insulin detemir and NPH insulin on glycemic control and 3, and these results were pooled in the present analysis.19
hypoglycemia in older and younger persons with DM. The Additional secondary endpoints included basal insulin
intention-to-treat (ITT) analysis set included all persons doses after study treatment (including dose ratios) and
exposed to at least one dose of trial product. The completer body weight.
population was defined as participants who finished Safety endpoints summarized here included all treat-
the trial, and the efficacy analysis set included completers ment-emergent hypoglycemic episodes (all, major, minor,
having an HbA1c measurement at the end of the trial. and symptoms only) for all trials. Nocturnal hypoglycemic
episodes (11:00 p.m. to 6:00 a.m.) were described sepa-
 Study 1: 26-week basal-bolus insulin regimen; once or rately. All adverse events (AEs) (including severity and re-
twice daily treatment with insulin detemir or NPH in- lationship to trial product), all serious AEs (SAEs), and all
sulin; insulin aspart as bolus insulin; 505 persons in ITT AEs leading to withdrawal were also summarized. SAEs
population; mean age  standard deviation, 60.4  8.6; were defined as death; a life-threatening experience; in-
randomized 2:1 detemir:NPH.17 subject hospitalization or prolongation of existing hospi-
 Study 2: 22-week basal-bolus insulin regimen; once talization; persistent or significant disability or incapacity;
or twice daily treatment with insulin detemir1insulin important medical events that, based upon appropriate
aspart or NPH insulin1human soluble insulin; 394 medical judgement, might jeopardize the health of the
persons in ITT population; mean age 58.2  9.3; ran- subject or might require medical or surgical intervention to
domized 1:1 detemir:NPH.18 prevent one of the outcomes listed in this definition.
 Study 3: 26-week basal-only insulin regimen; twice daily Laboratory results were examined descriptively.
treatment with insulin detemir or NPH insulin was
added to prior therapy with oral antidiabetic drugs
(OADs); 475 persons in ITT population; mean age Hypoglycemic Episodes
60.8  9.2; randomized 1:1 detemir:NPH.19 Subjects recorded hypoglycemic episodes in their diaries
throughout the trial. They were asked to record date, time
Enrollment criteria were persons of either sex aged 18 and of hypoglycemic episode, time and type of last antidiabetic
older with DM for at least 1 year treated with insulin, in- treatment before episode, time of last main meal before
JAGS NOVEMBER 2007–VOL. 55, NO. 11 DETEMIR LOWER HYPOS THAN NPH IN OLDER PERSONS 1737

episode, whether the episode was symptomatic, whether

they were able to treat themselves, and the blood or plasma
glucose level before treating the episode (if available). The
answer to the question ‘‘Subject able to treat the episode by
him-/herself’’ was to be reported as ‘‘No’’ if another person
had to administer food, glucagon, or intravenous glucose to
the subject because of severe central nervous system dys-
function associated with the hypoglycemic episode. Hypo-
glycemic episodes were defined as major if requiring
assistance by another person because of central nervous
system symptoms and minor if they were self-treated and
documented by a blood glucose value less than 50 mg/dL
(2.8 mmol/L) or a plasma glucose value less than 56 mg/dL
(3.1 mmol/L). Symptom-only episodes were defined as
those that were self-managed but without a glucose mea-
surement or with a plasma glucose of 50 mg/dL or greater. Figure 1. Trial diagram.
Hypoglycemic episodes were only reported as AEs if they
fulfilled the criteria of SAEs.
treated with insulin detemir, compared with 4.8% of older
Statistical Methods persons and 4.3% of younger persons treated with NPH
The endpoints HbA1c (noninferiority), FPG, and body (Figure 1). The main reasons for discontinuation were
weight at the end of treatment were analyzed for each age reported as nonadherence to treatment or other protocol
group using analysis of variance models including treatment deviations, personal reasons, ineffective therapy, and AEs,
and trial as fixed effects and HbA1c, FPG, and weight at mostly unrelated to the trial medications. The efficacy anal-
baseline as covariates for the respective endpoints. Within- ysis set included 1,296 completers from the ITT population
person variation in SMFBG was analyzed using a mixed (724 for insulin detemir and 572 for NPH; 416 older and
model with subject as a random effect. Mean basal insulin 880 younger persons). Five completers (4 younger and 1
doses at the end of treatment (including dose ratios) were older) were excluded from the efficacy analysis set, because
calculated. Analysis of hypoglycemic episodes was per- they were missing HbA1c values at end of trial. Overall, of
formed using a negative binomial distribution with differ- the four age- and treatment-group combinations, 93.6%
ent means according to treatment. Serious adverse events to 95.2% of the ITT population was in the efficacy anal-
related to hypoglycemia and withdrawals due to hypogly- ysis set.
cemia were summarized descriptively.
Efficacy
RESULTS After 22 to 26 weeks, HbA1c with insulin detemir was not
inferior to NPH insulin (mean treatment difference, insulin
Subjects detemir–NPH insulin, 0.035%, 95% CI 5  0.114  0.183
Baseline characteristics were balanced between treatments for older persons and 0.100%, 95% CI 5  0.017  0.217,
for both age groups. The disease duration was longer for for younger persons. For both older and younger persons,
older persons (Table 1). The ITT population (N 5 1,374) mean FPG with insulin detemir was similar to that with
was 773 for insulin detemir and 601 for NPH. Withdrawals NPH insulin after 24 or 26 weeks of treatment (mean
were 5.9% of older persons and 6.0% of younger persons treatment difference, insulin detemir  NPH insulin,

Table 1. Baseline Characteristics of Persons with Type 2 Diabetes Mellitus Pooled from Three Randomized Phase III
Trials Comparing Once- or Twice-Daily Treatment with Insulin Detemir with Treatment with Neutral Protamine
Hagedorn (NPH) Insulin: Data from Studies 1, 2, and 3
Older (Aged
65) Younger (Aged 18–64)

Characteristics Detemir (n 5 239) NPH (n 5 177) Detemir (n 5 485) NPH (n 5 395)

Men, n (%) 117 (50) 88 (50) 214 (40) 212 (50)

Women, n (%) 122 (50) 89 (50) 271 (60) 183 (50)
Age, mean (range) 70 (65–91) 70 (65–79) 55 (29–64) 55 (29–64)
Weight, kg, mean (range) 81 (43–125) 79 (52–126) 85 (44–128) 85 (42–150)
Body mass index, kg/m2, mean (range) 29 (19–48) 29 (19–40) 30 (18–49) 30 (18–52)
Duration of diabetes mellitus, years, mean (range) 14 (1–40) 14 (1–38) 11 (1–38) 12 (1–39)
Hemoglobin A1c, %, mean (range) 8.1 (5.4–11.9) 8.1 (5.3–11.6) 8.2 (5.1–12.3) 8.2 (5.0–12.6)
FPG, mmol/L, mean (range) 10.4 (2.9–20.1) 10.5 (5.3–20.4) 10.6 (2.7–21.9) 10.7 (3.2–24.3)
FPG, mg/dL, mean (range) 187.0 (52.9–361.6) 188.5 (94.9–367.6) 191.0 (49.0–393.7) 192.8 (56.9–437.6)

FPG 5 fasting plasma glucose.

1738 GARBER ET AL. NOVEMBER 2007–VOL. 55, NO. 11 JAGS

0.97 mg/dL, 95% CI 5  8.01  9.95 for older persons and quent in older than younger persons (5.1% vs 1.8% for
4.69 mg/dL, 95% CI 5  2.30  11.67 for younger per- NPH insulin vs 2.1% and 1.0% for insulin detemir, for
sons). Within-person variation (expressed as SD (CV%)) in older and younger persons, respectively). Older persons
SMFBG was significantly lower after 22 to 26 weeks of treated with insulin detemir had significantly lower risk of
treatment with insulin detemir than with NPH insulin for minor (P 5.005) and ‘‘symptoms only’’ (P 5.011) episodes
older persons (24.3 mg/dL (18.6%) for insulin detemir vs than with NPH insulin, whereas younger persons had sig-
27.2 mg/dL (20.8%) for NPH insulin, Po.05) and for nificantly lower risk of minor (P 5.031) episodes and sim-
younger persons (22.6 mg/dL (16.9%) for insulin detemir vs ilar risk of ‘‘symptoms only’’ episodes than with NPH
25.8 mg/dL (19.0%) for NPH insulin, Po.001). insulin. No hypoglycemic episodes led to withdrawal in any
of the three studies.
Insulin Dose The proportion of persons who experienced AEs and
the total number of treatment-emergent AEs (i.e., occurring
For younger people, the mean daily basal insulin dose was
from the first day of dosing to 7 days after the final dose,
0.63  0.45 U/kg for insulin detemir and 0.48  0.28 IU/kg
inclusive), SAEs, and AEs leading to withdrawal for both
for NPH insulin (insulin detemir:NPH insulin ratio 1:31).
age groups are shown in Table 3. The majority of events
Older people had similar doses (0.59  0.44 U/kg for insu-
were mild or moderate, with only 4.5% (82/1,827 events;
lin detemir and 0.46  0.26 U/kg for NPH insulin, insulin
44 in older persons, 38 in younger persons) listed as severe.
detemir:NPH insulin ratio 1:29). Two of the trials (Studies
Most SAEs were not considered to be related to the study
117 and 218) were basal-bolus studies, whereas Study 319
drugs, with most of those that were involving hypoglycemic
used twice-daily basal insulin in conjunction with OADs.
episodes or accidental overdoses. Similar proportions of
Thus, although the mean basal doses are useful for com-
people in both treatment arms experienced AEs for both age
paring doses in older versus younger persons in this pooled
groups, without any consistent pattern. The frequency of
analysis, comparison with other studies has little meaning,
AEs of moderate severity appeared to be slightly higher for
because these values reflect the composite effect of different
insulin detemir for both age groups (Table 3), although the
trial designs and use of different bolus insulins.
frequency of SAEs was higher with NPH insulin than with
insulin detemir in older persons (10.2% vs 4.2%). AEs
Safety probably or possibly related to trial product varied between
The frequency of hypoglycemic episodes is described in Ta- the three studies, but overall, similar proportions of AEs in
ble 2. Relative risk (RR) of overall hypoglycemia was sta- each treatment group were deemed to be related to trial
tistically significantly lower with insulin detemir than with product (72/1,085 events (6.6%) for insulin detemir vs 43/
NPH insulin (RR 5 0.59, P 5.002 for older persons; 742 events (5.8%) for NPH), with the only such AE that
RR 5 0.75, P 5.022, for younger persons; Table 2). RR of appeared to be more frequent with insulin detemir than
all nocturnal episodes was significantly lower with insulin NPH insulin being injection site reactions (16 vs 7 events
detemir (Po.001) in younger persons but not significantly reported). There was no indication that this varied accord-
different for the two treatments in older persons. Major ing to age group.
hypoglycemic episodes (overall or nocturnal) were uncom- Standard laboratory parameters (including lipids)
mon overall, so formal statistical comparisons were not showed no clinically relevant changes for either treatment
performed. Numerically, they were somewhat more fre- or age group at the end of the study.

Table 2. Frequency of Hypoglycemia in Older and Younger Persons with Diabetes Mellitus During 22 to 26 Weeks
of Treatment with Once- or Twice-Daily Insulin Detemir or Neutral Protamine Hagedorn (NPH) Insulin, Pooled from
Three Randomized Phase III Trials
Older (Aged
65) (n 5 416) Younger (Aged 18–64) (n 5 880)

Detemir (n 5 239) NPH (n 5 177) Detemir (n 5 485) NPH (n 5 395)

w w
Episodes n (%) Hypoglycemic Episodes RR 95% CI n (%) Hypoglycemic Episodes RR 95% CI

Overall hypoglycemic episodes

All 131 (55) 1,139 128 (72) 1,368 0.59 0.42–0.83 267 (55) 1,675 237 (60) 1,772 0.75 0.59–0.96
Major 5 (2) 6 9 (5) 13 ND ND 5 (1) 5 7 (2) 13 ND ND
Minor 89 (37) 446 99 (56) 576 0.56 0.37–0.84 172 (35) 732 181 (46) 798 0.72 0.54–0.97
Symptoms 111 (46) 687 112 (63) 779 0.61 0.42–0.89 205 (42) 937 187 (47) 959 0.78 0.59–1.03
Nocturnal episodes
All 59 (25) 208 69 (39) 238 0.71 0.43–1.17 102 (21) 215 129 (33) 349 0.50 0.36–0.69
Major 3 (1) 3 3 (2) 6 ND ND 0 (0) 0 2 (1) 4 ND ND
Minor 35 (15) 89 47 (27) 112 0.64 0.36–1.14 59 (12) 103 94 (24) 202 0.40 0.27–0.60
Symptoms 40 (17) 116 47 (27) 120 0.77 0.42–1.42 63 (13) 112 64 (16) 143 0.65 0.42–1.00

Number of persons in intention-to-treat population.
w
Number and percentage of persons with hypoglycemia.
ND 5 not done; RR 5 relative risk of insulin detemir vs NPH; CI 5 confidence interval.
JAGS NOVEMBER 2007–VOL. 55, NO. 11 DETEMIR LOWER HYPOS THAN NPH IN OLDER PERSONS 1739

Table 3. Adverse Events in Older and Younger Persons with Diabetes Mellitus During 22–26 Weeks of Treatment with
Once- or Twice-Daily Insulin Detemir or Neutral Protamine Hagedorn (NPH) Insulin, Pooled from Three Randomized
Phase III Trials
Older (
65) (n 5 416) Younger (18–64) (n 5 880)

Detemir (n 5 239) NPH (n 5 177) Detemir (n 5 485) NPH (n 5 395)

Adverse Events n (%)w AEs

Total 149 (62.3) 340 95 (53.7) 238 288 (59.4) 745 224 (56.7) 504
By severity
Mild 109 (45.6) 207 79 (44.6) 157 230 (47.4) 509 186 (47.1) 380
Moderate 75 (31.4) 114 35 (19.8) 56 136 (28.0) 216 75 (19.0) 106
Severe 10 (4.2) 19 18 (10.2) 25 19 (3.9) 20 16 (4.1) 18
Related to insulin 15 (6.3) 20 15 (8.5) 19 43 (8.9) 52 21 (5.3) 24
Leading to withdrawal 6 (2.5) 8 3 (1.7) 3 10 (2.1) 11 2 (0.5) 3
 Number of persons in intention-to-treat population.
w
Number and percentage of persons with adverse events (AEs).

Weight 0.46–0.48 IU/kg for NPH), although insulin doses across

Older and younger persons gained statistically significantly treatments should not be compared with other studies,
less body weight with insulin detemir than with NPH because the mean values obtained in this pooled analysis
insulin (0.69 vs 1.62 kg for older persons, Po.001; 0.80 vs reflect the effect of combining twice-daily basal-only studies
1.93 kg for younger persons, Po.001). The treatment (requiring a higher basal insulin dose) with basal-bolus
difference was  1.02 (95% CI 5  1.61 to  0.42) for studies (requiring a lower basal insulin dose). Also, in the
older persons and  1.13 (95% CI 5  1.58 to  0.69) for latter, there is an effect from using different bolus insulins
younger persons. (aspart vs regular human insulin).25
Consistent with what was previously reported for all
age groups combined in two of three individual studies,
DISCUSSION there was less within-person variation in SMFBG for people
In this pooled post hoc analysis of three Phase III trials using insulin detemir than with NPH insulin (Po.05 for
involving 1,296 people with DM treated with insulin older persons and Po.001 for younger persons). There was
detemir or NPH for 22 to 26 weeks, lower risk of hypo- also less weight gain (Po.001) with insulin detemir than
glycemia was observed with insulin detemir than with NPH NPH insulin for both age groups. The decreased variability
insulin, for older and younger persons. This was accom- may explain in part the differential frequency of hypogly-
plished while reducing weight gain by 1.01 to 1.13 kg and cemia observed for insulin detemir and NPH insulin.
providing similar levels of glycemic control (end of trial Given the propensity for hypoglycemia-associated au-
HbA1c 7.2–7.3% and FPG 143–148 mg/dL) with both tonomic failure in advanced DM,10 the contribution of
insulins. The overall safety profile was similar between older age and impaired awareness to risk of severe hypo-
treatments for both age groups. glycemia,8,12 and more-severe cognitive impairment during
A risk reduction of 41% for all hypoglycemic episodes hypoglycemia in older persons,13 an insulin formula-
was observed with insulin detemir, which was clinically as tion having a significant reduction in risk of hypoglycemia
well as statistically significant (RR 5 0.59, P 5.002 for is likely to be particularly welcome for this patient popu-
older persons; RR 5 0.75, P 5.02 for younger persons). lation.
Major (nocturnal and overall) episodes were uncommon for It is possible, but not proven, that the lower risk of
both treatment groups and both age groups. That the risk hypoglycemia contributes in turn to the relative weight
reduction was greater for the older persons is particularly benefit of insulin detemir. In one exploratory analysis of
encouraging, because this population may be especially data from a 26-week randomized trial, there was no cor-
reluctant to intensify regimens. There are no comparable relation between weight gain and hypoglycemia for people
data comparing hypoglycemic events with NPH insulin using insulin detemir (P 5.20), in contrast to a significant
with those for the other basal insulin analog, insulin glar- correlation between greater number of hypoglycemic events
gine, separately for older and younger persons with DM, and mean weight gain for NPH insulin (P 5.003).26
although with respect to all age groups combined, insulin Although the weight-sparing effect for insulin detemir
glargine has generally been associated with less hypoglyce- versus NPH insulin has been previously reported in Phase
mia, and nocturnal hypoglycemia in particular, than NPH III trials in type 1 and type 2 DM,27 this pooled analysis
insulin.21–24 demonstrates that the weight-sparing action is retained for
In the current pooled analysis, basal insulin require- older and younger persons with DM. This adds to the un-
ments were similar within treatment group for both derstanding of the clinical performance of insulin detemir,
age groups but were somewhat higher for insulin detemir which remains the only basal insulin analog to have a con-
than for NPH insulin (0.59–0.63 U/kg for detemir vs sistent weight-sparing effect. Studies using the other basal
1740 GARBER ET AL. NOVEMBER 2007–VOL. 55, NO. 11 JAGS

insulin analog, insulin glargine, have not consistently 7. Donnelly LA, Morris AD, Frier BM et al. DARTS/MEMO collaboration. Fre-
reported this weight-sparing effect, with one study report- quency and predictors of hypoglycaemia in Type 1 and insulin-treated Type 2
diabetes: A population-based study. Diabet Med 2005;22:749–755.
ing less weight gain than with NPH insulin28 and others 8. Henderson JN, Allen KV, Deary IJ et al. Hypoglycaemia in insulin-treated Type
reporting similar weight gain with glargine and NPH insu- 2 diabetes: Frequency, symptoms, and impaired awareness. Diabet Med
lin29,30 in DM. Older persons and younger persons may be 2003;20:1016–1021.
concerned about weight gain, so the results of this study 9. Hepburn DA, MacLeod KM, Pell AC et al. Frequency and symptoms of
hypoglycaemia experienced by patients with type 2 diabetes treated with
should reassure both subgroups that weight gain is not in- insulin. Diabet Med 1993;10:231–237.
evitable with intensive insulin therapy. 10. Segel SA, Paramore DS, Cryer PE. Hypoglycemia-associated autonomic failure
The results of this post hoc analysis of data pooled from in advanced type 2 diabetes. Diabetes 2002;51:724–733.
three Phase III trials comparing insulin detemir and NPH 11. Cryer PE. Hypoglycaemia: The limiting factor in the glycaemic management of
type I and type II diabetes. Diabetologia 2002;45:937–948.
insulin suggest that the global benefits of insulin detemir 12. Holstein A, Plaschke A, Egberts EH. Clinical characterisation of severe hy-
Fsimilar HbA1c with less hypoglycemia, less variation in poglycaemiaFa prospective population-based study. Exp Clin Endocrinol
SMFBG, and less weight gainFapply in older and younger Diabetes 2003;111:364–369.
persons with DM. This may allow older persons desiring 13. Matyka K, Evans M, Lomas J et al. Altered hierarchy of protective responses
against severe hypoglycemia in normal aging in healthy men. Diabetes Care
better control to titrate insulin with greater confidence or, 1997;20:135–141.
for patients already at or near target, to use insulin with 14. Hirsch IB. Insulin analogues. N Engl J Med 2005;352:174–183.
greater safety and less worry, without compromising cur- 15. Brunner GA, Sendhofer G, Wutte A et al. Pharmacokinetic and pharmaco-
rent levels of control. dynamic properties of long-acting insulin analogue NN304 in comparison to
NPH insulin in humans. Exp Clin Endocrinol Diabetes 2000;108:100–105.
16. Heinemann L, Sinha K, Weyer C et al. Time-action profile of the soluble, fatty
acid acylated, long-acting insulin analogue NN304. Diabet Med 1999;16:
ACKNOWLEDGMENTS 332–338.
17. Haak T, Tiengo A, Draeger E et al. Lower within-subject variability of fasting
Conflict of Interest: This study was supported by Novo- blood glucose and reduced weight gain with insulin detemir compared to NPH
Nordisk. The original three randomized trials used in this insulin in patients with type 2 diabetes. Diab Obes Metab 2005;7:56–64.
post hoc analysis were also supported by Novo Nordisk. 18. Raslova K, Bogoev M, Raz I et al. Insulin detemir and insulin aspart: A
Alan Garber has served on several advisory boards for promising basal-bolus regimen for type 2 diabetes. Diabetes Res Clin Pract
2004;66:193–201. Erratum in: Diab Res Clin Pract 2006;72:112.
Novo Nordisk. Klaus Kolendorf has acted as an Investiga- 19. Hermansen K, Davies M, Derezinski T et al. A 26-week, randomized, parallel,
tor in Novo Nordisk’s Phase III clinical trials. treat-to-target trial comparing insulin detemir with NPH insulin as add-on
Author Contributions: Alan Garber designed the meta- therapy to oral glucose-lowering drugs in insulin-naive people with type 2
analysis; interpreted the results; was involved in data anal- diabetes. Diabetes Care 2006;29:1269–1274.
20. Philis-Tsimikas A, Charpentier G, Clauson P et al. Comparison of once-daily
ysis, drafting, and critical revision of the manuscript; and is insulin detemir with NPH insulin added to a regimen of oral antidiabetic drugs
primary author of the manuscript. Per Clauson and Claus in poorly controlled type 2 diabetes. Clin Ther 2006;28:1569–1581.
Pedersen were involved in study concept and design, data 21. Wang F, Carabino JM, Vergara CM. Insulin glargine: A systematic review of a
gathering, data analysis, drafting, and critical revision of long-acting insulin analogue. Clin Ther 2003;25:1541–1577.
22. Rosenstock J, Dailey G, Massi-Benedetti M et al. Reduced hypoglycemia risk
the manuscript. Klaus Kolendorf was involved in data with insulin glargine: A meta-analysis comparing insulin glargine with human
analysis, drafting, and critical revision of the manuscript. NPH insulin in type 2 diabetes. Diabetes Care 2005;28:950–955.
Sponsor’s Role: Two authors are employees of Novo 23. Bullano MF, Al-Zakwani IS, Fisher MD et al. Differences in hypoglycemia
Nordisk, which also provided support for medical writing event rates and associated cost-consequence in patients initiated on long-acting
and intermediate-acting insulin products. Curr Med Res Opin 2005;21:
assistance in the form of copyediting and redrawing 291–298.
artwork to meet the Journal’s requirements. 24. Eliaschewitz FG, Calvo C, Valbuena H et al. HOE 901/4013 LA Study Group.
Therapy in type 2 diabetes: Insulin glargine vs. NPH insulin both in combi-
nation with glimepiride. Arch Med Res 2006;37:495–501.
25. Home PD, Lindholm A, Riis A. European Insulin Aspart Study Group. Insulin
REFERENCES aspart vs. human insulin in the management of long-term blood glucose
1. UK Prospective Diabetes Study (UKPDS) Group. Intensive blood-glucose con- control in type 1 diabetes mellitus: A randomized controlled trial. Diabet Med
trol with sulphonylureas or insulin compared with conventional treatment and 2000;17:762–770.
risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 26. Davies M, Derezinski T, Kim H et al. No correlation between weight gain and
1998;352:837–853. number of hypoglycemic events in patients with type 2 diabetes treated with
2. Lebovitz HE, Austin MM, Blonde L et al. ACE/AACE Diabetes Recommen- insulin detemir as compared to NPH insulin. Diabetes 2006;55(Suppl 1):
dations Implementation Writing Committee. ACE/AACE consensus confer- A466.
ence on the implementation of outpatient management of diabetes mellitus: 27. Fritsche A, Haring H. At last, a weight-neutral insulin? Int J Obes 2004;
Consensus conference recommendations. Endocr Pract 2006;12(Suppl 1): 28(Suppl 2):S41–S46.
6–12. 28. Rosenstock J, Schwartz SL, Clark CM Jr. et al. Basal insulin therapy in type 2
3. American Diabetes Association. Clinical practice recommendations. Diabetes diabetes: 28-week comparison of insulin glargine (HOE 901) and NPH insulin.
Care 2005;28(Suppl 1):S1–S3. Diabetes Care 2001;24:631–636.
4. Resnick HE, Foster GL, Bardsley J et al. Achievement of American Diabetes 29. Yki-Jarvinen H, Dressler A, Ziemen M. HOE 901/300s Study Group. Less
Association clinical practice recommendations among U.S. adults with diabe- nocturnal hypoglycemia and better post-dinner glucose control with bedtime
tes, 1999-2002: The National Health and Nutrition Examination Survey. Di- insulin glargine compared with bedtime NPH insulin during insulin combi-
abetes Care 2006;29:531–537. nation therapy in type 2 diabetes. HOE 901/3002 Study Group. Diabetes Care
5. Korzon-Burakowska A, Hopkins D, Matyka K et al. Effects of glycemic con- 2000;23:1130–1136.
trol on protective responses against hypoglycemia in type 2 diabetes. Diabetes 30. Riddle MC, Rosenstock J, Gerich J. Insulin Glargine 4002 Study Investigators.
Care 1998;21:283–290. The Treat-to-Target Trial: Randomized addition of glargine or human NPH
6. Cryer PE, Davis SN, Shamoon H. Hypoglycemia in diabetes. Diabetes Care insulin to oral therapy of type 2 diabetic patients. Diabetes Care 2003;26:
2003;26:1902–1912. 3080–3086.