Diuretics, Beta-blockers and Statins Increase
the Risk of Diabetes in Patients with Impaired
Glucose Tolerance:
Insights from the NAVIGATOR study
Lan Shen, MD; Bimal R. Shah, MD, MBA; Eric M. Reyes, Ph.D.; Laine Thomas,
Ph.D.; Peter Diem, MD; Lawrence A. Leiter, MD; Bernard Charbonnel, MD;
Viacheslav Mareev, MD; Edward Horton, MD; Steven M. Haffner, MD; Vladimir
Soska, MD; Rury Holman, MD; Angelyn Bethel, MD; Frank Schaper, MD; Jie Lena
Sun, MS; John McMurray, MD; Robert Califf, MD; Henry Krum, MBBS, Ph.D.
Disclosures
L. Shen: None. B. Shah: None. E. Reyes: None. L. Thomas: None. P. Diem:
Research Grant; Significant; Novartis. Consultant/Advisory Board; Modest;
Novartis.L. Leiter: Research Grant; Significant; Novartis.
Consultant/Advisory Board; Modest; Novartis. B. Charbonnel:
Consultant/Advisory Board; Modest; Novartis. V. Mareev: Speakers Bureau;
Modest; Novartis. Consultant/Advisory Board; Modest; Novartis. E. Horton:
Consultant/Advisory Board; Modest; Novartis. S. Haffner: None. V. Soska:
None. R. Holman: Research Grant; Significant; Novartis.
Consultant/Advisory Board; Modest; Novartis. M. Bethel: Research Grant;
Modest; Novartis, Bayer, Merck, Amylin, Lilly. F. Schaper: Research Grant;
Significant; Novartis. J. Sun: None. J. McMurray: Other; Modest; Employer
(Glasgow University) paid for time as member of executive committee of
NAVIGATOR and travel and accommodations for NAVIGATOR
executive/steering committee meetings. R. Califf: Research Grant;
Significant; Novartis. Consultant/Advisory Board; Modest; Novartis. H.
Krum: Research Grant; Significant; Novartis.
Background
• Diuretics, β-blockers, and statins are the
cornerstone of therapy in patients with CVD
• Even with their robust evidence of benefit, more
recent data suggests that the long-term use may
increase fasting glucose levels
• These concerns have led the US FDA to mandate a
label change for statins in 2012
• However, there are limited large scale studies using
serial glucose measures that have linked these
medications to new onset diabetes (NOD)
Objective
To examine the association of new onset
diabetes (NOD) with the initiation of βblockers, thiazide diuretics, or statins in
high risk, treatment naïve patients.
NAVIGATOR Study
• A multinational, randomized design study testing the
efficacy and safety of long term administration of
nateglinide and valsartan in the prevention of
diabetes and cardiovascular outcomes
• Study population is 9518 patients with impaired
glucose tolerance (IGT) and at least one other CV
risk factor enrolled between January 4, 2002 to
January 29, 2004.
• NAVIGATOR excluded patients with diabetes at
baseline
Methods
• Therefore, we evaluated treatment initiation of the
following medications on naïve patients.
• β-blockers
• Diuretics
• Statins
• CCBs – (metabolically negative control)
• Median follow-up time: 5.0 years
Methods
• Endpoint: New onset diabetes (NOD)
• NOD defined as:
• Fasting plasma glucose >126 mg/dl (7.0mmol/L) OR
Glucose level >200 mg/dl (11.1 mmol/L) 2 hours after
OGTT
AND
• Confirmed by OGTT 12 weeks after the elevated
glucose value was recorded
• Glucose testing frequency:
• Fasting plasma glucose level was measured every 6
months for the first 3 years and then annually
• Oral glucose tolerance tests were performed annually
Methods
• We examined the timing of the starting of each
medication during study period
• We estimated the effect of receiving treatment on
progression to NOD using Marginal Structural Models
– In short, this is a Cox Proportional Hazard model adjusting
cofounders that change over time
– We present a single hazard ratio for treatment that
represents the difference in rate of progression to NOD
between someone who receives these medications and
someone who does not
Results
Medication Initiation
N=5637
N=6343
N=6143
N=6290
Results
Baseline Characteristics
β –blockers
Yes
n=993
Diuretics
Statins
No
Yes
No
Yes
No
n=4644 n=1425 n=4918 n=1474 n=4669
Age (Median)
64
62
64
62
63
63
Hypertension (%)
80
69
84
65
77
84
Congestive Heart
Failure (%)
4
2
3
2
3
4
Hx CV events (%)
30
19
34
32
31
29
Hx stroke (%)
5
3
6
3
4
3
Hx Heart Failure (%)
4
2
3
2
3
4
Fasting Glucose
6
6
6
6
6
6
5.8
5.8
5.8
5.8
5.8
5.8
HbA1c
Results
Incidence of New Onset Diabetes
Unadjusted HR
(95% CI)
Adjusted HR
(95% CI)
Diuretics
1.48
(1.30,1.68)
1.35
(1.12,1.61)
Statins
1.38
(1.21, 1.57)
1.30
(1.09, 1.55)
β-blockers
1.35
(1.16, 1.57)
1.19
(0.97,1.46)
CCBs
1.31
(0.97, 1.31)
1.14
(0.91, 1.42)
Limitations
• Post-hoc analysis of a clinical trial so residual
confounders cannot be excluded
• No information on dose-response for medication
• No information on the adherence to specified
medication
• Small size of population in β-blocker subgroup
contributes to the limited power of the association in
the β-blocker subgroup
Conclusion
• In high-risk patients with IGT, the use of diuretic and
statin therapy is associated with NOD
• β-blocker therapy has a borderline effect towards the
development of NOD, but the effect was not significant
after adjustment
• Surveillance of glucose levels should be considered in
high risk patients taking these medications
• Future studies are warranted to examine net benefit/risk
profile of those medications in high risk CVD patients
Thank you