n engl j med 362;2 nejm.org april 22, 2010
1463
The new england
journal
of medicine
established in 1812
april 22, 2010
vol. 362 no. 16
Effect of Nateglinide on the Incidence of Diabetes
and Cardiovascular Events
The NAVIGATOR Study Group*
ABSTR ACT
The authors are listed in the Appendix.
Address reprint requests to Dr. Rury R.
Holman at the Diabetes Trials Unit, Ox-
ford Centre for Diabetes, Endocrinology,
and Metabolism, Churchill Hospital, Ox-
ford OX3 7LJ, United Kingdom, or at rury
.holman@dtu.ox.ac.uk.
*The names of the investigators and
members of the committees in the Na-
teglinide and Valsartan in Impaired Glu-
cose Tolerance Outcomes Research
(NAVIGATOR) Study Group are listed in
Supplementary Appendix 1, available with
the full text of this article at NEJM.org.
This article (10.1056/NEJMoa1001122) was
published on March 14, 2010, and was up-
dated on March 29, 2010, at NEJM.org.
N Engl J Med 2010;362:1463-76.
Copyright © 2010 Massachusetts Medical Society.
Background
The ability of short-acting insulin secretagogues to reduce the risk of diabetes or
cardiovascular events in people with impaired glucose tolerance is unknown.
Methods
In a double-blind, randomized clinical trial, we assigned 9306 participants with
impaired glucose tolerance and either cardiovascular disease or cardiovascular risk
factors to receive nateglinide (up to 60 mg three times daily) or placebo, in a 2-by-2
factorial design with valsartan or placebo, in addition to participation in a lifestyle
modification program. We followed the participants for a median of 5.0 years for
incident diabetes (and a median of 6.5 years for vital status). We evaluated the effect
of nateglinide on the occurrence of three coprimary outcomes: the development of
diabetes; a core cardiovascular outcome that was a composite of death from cardio-
vascular causes, nonfatal myocardial infarction, nonfatal stroke, or hospitalization
for heart failure; and an extended cardiovascular outcome that was a composite of
the individual components of the core composite cardiovascular outcome, hospital-
ization for unstable angina, or arterial revascularization.
Result s
After adjustment for multiple testing, nateglinide, as compared with placebo, did
not significantly reduce the cumulative incidence of diabetes (36% and 34%, respec-
tively; hazard ratio, 1.07; 95% confidence interval [CI], 1.00 to 1.15; P = 0.05), the core
composite cardiovascular outcome (7.9% and 8.3%, respectively; hazard ratio, 0.94,
95% CI, 0.82 to 1.09; P = 0.43), or the extended composite cardiovascular outcome
(14.2% and 15.2%, respectively; hazard ratio, 0.93, 95% CI, 0.83 to 1.03; P = 0.16).
Nateglinide did, however, increase the risk of hypoglycemia.
Conclusions
Among persons with impaired glucose tolerance and established cardiovascular dis-
ease or cardiovascular risk factors, assignment to nateglinide for 5 years did not re-
duce the incidence of diabetes or the coprimary composite cardiovascular outcomes.
(ClinicalTrials.gov number, NCT00097786.)
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org by JEAN-CHRISTOPHE PHILIPS on May 3, 2010 .
The
new england journal
of
medicine
n engl j med 362;16 nejm.org april 22, 2010
1464
P
ersons with impaired glucose tol-
erance are at increased risk for type 2 dia-
betes mellitus and cardiovascular disease
1-3
;
therefore, treatments that might reduce the inci-
dence of diabetes and associated cardiovascular
disease and deat h are potentially important.
3
The
risk of diabetes is reduced with lifestyle interven-
tions that involve increasing physical activity and
r e d u c i n g w e i g h t
4-6
and with metformin,
6
acar-
bose,
7
or rosiglitazone
8
therapy, but no trials to
date have been powered to consider cardiovascu-
lar outcomes.
Among persons with type 2 diabetes, reducing
glycemia results in a small reduction in the risk
of major macrovascular events.
9
Glucose levels af-
ter a glucose challenge, however, are more closely
associated with cardiovascular risk than are fast-
ing glucose levels,
2
suggesting that postprandi-
al glycemia may be a distinct therapeutic target.
Lowering postprandial glucose levels with the
alpha-glucosidase inhibitor acarbose has been re-
ported to decrease the risk of myocardial infarc-
tion among persons with impaired glucose tol-
erance.
10
In a large, prospective study involving persons
with impaired glucose tolerance and cardiovascu-
lar disease or cardiovascular risk factors, we evalu-
ated an alternative postprandial glucose-lower-
ing approach that used the short-acting insulin
secretagogue nateglinide, in addition to a lifestyle
modification program. The aim of the Nateglin-
ide and Valsartan in Impaired Glucose Tolerance
Outcomes Research (NAVIGATOR) trial
11
was to
determine whether the risk of diabetes and car-
diovascular events could be reduced in this popu-
lation.
Methods
Study Oversight
The trial was approved by the ethics committee
at each participating center, and all participants
provided written informed consent. The study was
sponsored by Novartis Pharma, was designed by
the sponsor in collaboration with an academic
executive committee, and was monitored by an
independent data and safety monitoring commit-
tee (see Supplementary Appendix 1, available
with the full text of this article at NEJM.org).
11
Data were collected, managed, and analyzed by
the sponsor, with oversight by the executive com-
mittee, and analyses were replicated by an inde-
pendent academic statistician from the London
School of Hygiene and Tropical Medicine. The
manuscript was prepared by the writing group
(see Section 1 in Supplementary Appendix 1),
whose members had unrestricted access to the
data, and was revised subsequently by all the au-
thors. All the authors made the decision to sub-
mit the manuscript for publication and assume
responsibility for the accuracy and completeness
of the data. The NAVIGATOR trial protocol is avail-
able in Supplementary Appendix 2.
Study Participants
Subjects were eligible for inclusion in the study if
they had impaired glucose tolerance,
3
a fasting
plasma glucose concentration of at least 95 mg per
deciliter (5.3 mmol per liter) but less than 126 mg
per deciliter (7.0 mmol per liter), and one or more
cardiovascular risk factors (in the case of sub-
jects who were 55 years of age or older) or known
cardiovascular disease (in the case of subjects who
were 50 years of age or older) (
Section 2
in Sup-
plementary Appendix 1). Subjects were excluded
if they had abnormal laboratory test results or
concomitant conditions that could interfere with
the assessment of the safety or efficacy of the
study drug or if they had taken antidiabetic med-
ication within the previous 5 years.
11
Study Medication
Participants were randomly assigned, with the use
of an interactive voice-response telephone system,
to nateglinide, at a dose of 60 mg taken before
meals three t imes daily, or placebo and, in a 2-by-
2 factorial design, to valsartan or placebo. Both
the participants and the investigators were un-
aware of the treatment assignments. Nateglinide
was initially dispensed at a dose of 30 mg, with
an increase to the full dose of 60 mg after 2 weeks.
Reductions in the dose were permitted if there
were side effects. The comparison of valsartan
with placebo is reported elsewhere in this issue
of the Journal.
12
Lifest yle Modification
All subjects were required to participate in a st udy-
specific lifestyle modification program that was
designed to reduce the risk of diabetes (see Sec-
tion 3 in Supplementary Appendix 1). The aim of
the program was to help participants achieve and
maintain a 5% weight loss, reduce the amount of
saturated and total fats in their diets, and in-
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org by JEAN-CHRISTOPHE PHILIPS on May 3, 2010 .
Nateglinide, Diabetes, and Cardiovascular Events
n engl j med 362;16 nejm.org april 22, 2010
1465
crease their physical activity to 150 minutes week-
ly.
5,6
To facilitate adherence, site personnel who
were trained to administer the program provided
materials to participants at each clinic visit and re-
inforcement through interim telephone contacts.
Study Procedures
After the dose-adjustment phase, participants re-
turned for study visits every 6 months. The fast-
ing plasma glucose level was measured every 6
months for 3 years and annually thereafter. Oral
glucose-tolerance tests were performed yearly. On
the mornings of study visits, the administration
of the study medication was delayed until after the
glucose tests had been performed so that the glu-
cose measurements would not be affected by the
study drugs.
Study Outcomes
Initially, there were two coprimary outcomes: in-
cident diabetes mellitus and an extended compos-
ite cardiovascular outcome (death from a cardio-
vascular cause, nonfatal myocardial infarction,
nonfatal stroke, hospitalization for heart failure,
arterial revascularization, or hospitalization for
unstable angina). A third coprimary core cardio-
vascular outcome (death from a cardiovascular
cause, nonfatal myocardial infarction, nonfatal
stroke, or hospitalizat ion for heart failure), which
had initially been designated as a secondary out-
come, was added, as described previously.
11,13
Diabetes was considered to be present
14,15
if the
participant had a fasting plasma glucose level of
126 mg per deciliter or more or a glucose level of
200 mg per deciliter (11.1 mmol per liter) or more
2 hours after a glucose challenge — confirmed by
an oral glucose-tolerance test within 12 weeks af-
ter the elevated glucose value was recorded. The
date of the onset of diabetes was considered to be
the date of the first elevated glucose value. An in-
dependent committee, whose members were un-
aware of the treatment assignments, adjudicated
cases in which diabetes was diagnosed by other
means. Participants in whom diabetes developed
could have open-label antidiabetic medication —
with the exception of prandial glucose regulators
— added to their regimen.
Deaths, hospitalizations, and potential cardio-
vascular events that did not result in hospitaliza-
tion were adjudicated by an independent committee
whose members were unaware of the treatment
assignments. (Definitions of these outcomes are
provided in Section 4 in Supplementary Appen-
dix 1.)
Statistical Analysis
The targeted sample size of 9152 participants was
determined on the basis of the estimated number
of participants that would be needed to provide
the power required for the valsartan-versus-pla-
cebo component of the study.
11,13
We estimated
that diabetes would develop in at least 3000 of
these participants, with the result that the study
would have more than 99% power to detect a 25%
reduction in the risk of incident diabetes.
Since the effects of the two drugs (nateglinide
and valsartan) on the three coprimary outcomes
were examined in a factorial manner, we adjusted
for the three tests that were performed for each
drug, but not across drugs. We report two-sided
P values throughout, as well as protocol-specified
one-sided values for the coprimary outcomes and
their components. The 2.5% one-sided family-wise
type I error rate for each drug was controlled with
the use of a closed-testing procedure that initially
assigned one fifth of the alpha to the diabetes
outcome and four fifths to the two cardiovascular
outcomes (since we anticipated that there would be
more cases of diabetes than of cardiovascular out-
comes). This allowed for the testing of each copri-
mary outcome even if the others did not differ
significantly between the study groups but would
usually require a two-sided P value of less than 0.01
for the difference in the diabetes outcome to be
considered significant (Section 5 in Supplementary
Appendix 1). An O’Brien–Fleming type of alpha
spending approach accounted for the interim
efficacy analysis that was performed in Novem-
ber 2005, when 30% of the target number of par-
ticipants had had an extended cardiovascular out-
come (one-sided alpha spent: 0.00004).
16
Log-rank tests, stratified according to the pres-
ence or absence of a history of cardiovascular dis-
ease and to randomized assignment to valsartan
or placebo, were used to compare the nateglinide
and placebo groups with respect to the time to
the first event in the extended or core cardiovas-
cular outcome. A Cox discrete-time proportion-
al-odds model
17
was used to assess incident dia-
betes, given the fixed-time schedule for glucose
measurements. We also performed predefined
analyses of the components of the composite car-
diovascular outcome, exploratory outcomes (the
time to death from all causes and the time to
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org by JEAN-CHRISTOPHE PHILIPS on May 3, 2010 .
The
new england journal
of
medicine
n engl j med 362;16 nejm.org april 22, 2010
1466
cardiovascular-related hospitalization), indexes of
hyperglycemia, and body weight. We tested for a
possible interaction between valsartan and nat-
eglinide for each reported outcome. The effects
of the study treatment were evaluated in pre-
specified subgroups (Section 6 in Supplementary
Appendix 1).
11
Baseline characteristics of the par-
ticipants, safety assessments, and additional mea-
surements during the trial were compared with
the use of summary statistics.
Results
Study Participants
Of 43,502 participants who were screened at 806
centers in 40 countries, 9518 were randomly as-
signed to a study drug between January 4, 2002,
and Januar y 29, 2004. The most common reasons
for exclusion were ineligibility (96.6%) and refusal
to participate (2.5%). After randomization, 212
participants at 10 sites were excluded when the
sites were closed owing to deficiencies in the ad-
herence to Good Clinical Practice guidelines, leav-
ing 9306 participants whose data were included
in the final analyses (Fig. 1).
Among participants who were taking the study
drug at 6 months, fewer people in the nateglinide
group than in the placebo group were taking the
full dose of the study drug (92.0% vs. 94.2%,
P<0.001). At 1 year, 79.8% of the surviving partici-
pants in the nateglinide group and 80.8% of those
in the placebo group were still receiving the study
medication (P = 0.26); the respective percentages
were 74.7% and 75.9% at 3 years (P = 0.20) and
69.9% and 71.0% at 5 years (P = 0.28).
The baseline characteristics of the nateglinide
and placebo groups did not differ significantly
(
Table 1
). Of the 9306 participants, 3547 (38.1%)
reported that a primary relative had diabetes and
2266 (24.3%) had known cardiovascular disease.
There were no significant differences between
the groups in concomitant treatments at base-
line or at study end. At study end, less than 3%
of the participants in whom diabetes had not
developed were taking open-label glucose-lower-
ing therapies.
Follow-up
The fasting plasma glucose level or the plasma
glucose level 2 hours after a glucose challenge was
measured at the closeout visit or during the final
6 months of the study in 80% of the surviving
participants who had not withdrawn consent and
in whom diabetes had not developed. A total of
609 participants in the nateglinide group (13.1%)
and 602 in the placebo group (12.9%) were lost to
follow-up or withdrew from the study; however,
because many of these participants were lost to
follow-up or withdrew consent late in the study,
informat ion on vit al st atus was available for 95.7%
of the possible follow-up time in both groups.
The median follow-up time for data on vital status
was 6.5 years, and the median follow-up times
for data on the diabetes, extended cardiovascu-
lar, and core cardiovascular outcomes were 5.0,
6.3, and 6.4 years, respectively.
Study Outcomes
Coprimary Diabetes Outcome
Diabetes developed in 1674 participants in the
nateglinide group (36.0%) and in 1580 participants
in the placebo group (33.9%) (hazard ratio with
nateglinide, 1.07; 95% confidence interval [CI],
1.00 to 1.15; P = 0.05) (
Table 2
and Fig. 2A). The
effect of nateglinide on the progression of im-
paired glucose tolerance to diabetes was consis-
tent across all prespecified subgroups except for
subgroups specified according to sex and fasting
plasma glucose level (Section 6A in Supplemen-
tary Appendix 1).
Coprimary Cardiovascular Outcomes
The extended composite cardiovascular outcome
occurred in 658 participants in the nateglinide
group (14.2%) and in 707 participants in the pla-
cebo group (15.2%) (hazard ratio with nategli-
nide, 0.93; 95% CI, 0.83 to 1.03; P = 0.16 (
Table 2
and Fig. 2B). The core composite cardiovascular
outcome occurred in 365 participants in the na-
teglinide group (7.9%) and in 387 participants in
the placebo group (8.3%) (hazard ratio with na-
teglinide, 0.94; 95% CI, 0.82 to 1.09; P = 0.43) (
Ta-
ble 2
and Fig. 2C). Nateglinide treatment had no
significant effect on the coprimary cardiovascu-
lar outcomes in any of the prespecified subgroups
(Sections 6B and 6C in Supplementary Appendix 1).
Components of Primary Outcomes and Exploratory
Outcomes
As shown in Table 2 and Figure 2D, there were
no significant differences between the nateglin-
ide and placebo groups with respect to the com-
ponents of the extended cardiovascular compos-
ite outcome, including death from cardiovascular
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org by JEAN-CHRISTOPHE PHILIPS on May 3, 2010 .
Nateglinide, Diabetes, and Cardiovascular Events
n engl j med 362;16 nejm.org april 22, 2010
1467
causes, or with respect to the prespecified explor-
atory outcomes of hospitalization for a cardiovas-
cular cause and death from any cause.
No interactions between nateglinide and val-
sartan were seen for any of the outcomes reported
(Section 7 in Supplementary Appendix 1).
Glycemia
During the course of the study, participants in the
nateglinide group had lower mean fasting plas-
ma glucose levels than did those in the placebo
group; the mean difference was 0.47 mg per deci-
liter (95% CI, 0.05 to 0.90) (0.03 mmol per liter
[95% CI, 0.003 to 0.05]) (P = 0.03) (Fig. 3A). How-
ever, glucose levels 2 hours after a glucose chal-
lenge were higher in the nateglinide group than
in the placebo group; the mean difference was
4.37 mg per deciliter (95% CI, 2.80 to 5.93) (0.24
mmol per liter [95% CI, 0.16 to 0.33]) (P<0.001)
(Fig. 3B).
In the two exploratory analyses we conducted,
incident diabetes defined on the basis of fasting
6 col
33p9
9518 Were randomly assigned to receive either
nateglinide or placebo
43,502 Patients with cardiovascular disease or
cardiovascular risk factors were screened
39,282 (90.3%) Were screened once
4133 (9.5%) Were screened twice
87 (0.2%) Were screened three
or four times
32,984 Were excluded
24,023 Did not meet inclusion criteria for
impaired glucose tolerance or
cardiovascular risk factors
6719 Had an abnormal laboratory value
1652 Had an abnormal test result
853 Withdrew consent
293 Had other reasons
215 Were ineligible owing to medical
history
171 Were taking excluded medication
58 Had an intercurrent medical event
9306 Were included in the final analysis
212 Were excluded because
of protocol deficiencies at site
4645 Were assigned to receive nateglinide
(with or without valsartan)
4661 Were assigned to receive placebo
(with or without valsartan)
919 Discontinued
446 Were lost to follow-up
310 Died
163 Withdrew participation
914 Discontinued
459 Were lost to follow-up
312 Died
143 Withdrew participation
3726 (80%) Completed the trial 3747 (80%) Completed the trial
AUTHOR:
FIGURE:
RETAKE:
SIZE
4-C H/TLine Combo
Revised
AUTHOR, PLEASE NOTE:
Figure has been redrawn and type has been reset.
Please check carefully.
1st
2nd
3rd
Holman
1 of 3
ARTIST:
TYPE:
ts
04-22-10JOB: 36216 ISSUE:
Figure 1. Enrollment and Follow-up of Study Participants.
Copyright © 2010 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org by JEAN-CHRISTOPHE PHILIPS on May 3, 2010 .
