Author + information
- Received February 6, 2004
- Revision received March 19, 2004
- Accepted March 23, 2004
- Published online August 4, 2004.
- Carl J. Pepine, MD, MACC⁎ ( and )
- Rhonda M. Cooper-DeHoff, PharmD
- ↵⁎Reprint requests and correspondence:
Dr. Carl J. Pepine, Division of Cardiovascular Medicine, University of Florida, 1600 S.W. Archer Road, Gainesville, Florida 32610-0277, USA.
The prevalence of diabetes is increasing, and patients with diabetes are at increased risk of adverse cardiovascular outcomes. Recently, the results from 11 large randomized clinical trials have suggested a difference in the emergence of new diabetes according to cardiovascular medication use. Treatment with angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, and calcium antagonists yielded a lower incidence of diabetes development than beta-blockers and diuretics. Physicians should consider this possible diabetes consequence when prescribing long-term beta-blockers and diuretics, particularly in patients at high risk of developing diabetes.
The epidemic of diabetes in the U.S. and elsewhere has been well publicized and has devastating implications on future cardiovascular (CV) disease and its adverse outcomes. Patients with diabetes have a two- to four-fold increase in CV mortality (1), and diabetes is the leading cause of end-stage renal disease (2). Recently, major emphasis has been placed on educating the public regarding the hazards of inactivity and obesity, which require long-term patient behavior modification to alter and which are major risk factors for diabetes development. However, much less attention has been given to physician practice patterns relative to drug prescribing, particularly in patients at high risk of developing diabetes.
Because CV disease, particularly coronary artery disease (CAD), and its risk conditions are life long, patients will require use of CV medications for many years. This prolonged medication exposure has led to questions regarding the potential of certain drugs to protect against or hasten diabetes development, although they are useful for the management of CAD and associated co-morbidities like hypertension (3–5). To this end, pharmacologic and epidemiologic studies have documented that certain drugs, primarily thiazide diuretics and beta-blockers (BBs), frequently used in patients with CAD or at high risk of CAD, are associated with metabolic disturbances that may result in increased insulin resistance (6–9). However, other angiotensin-active agents, primarily angiotensin-converting enzyme (ACE) inhibitors and angiotensin receptor blockers (ARBs), appear to improve insulin sensitivity and glucose metabolism (6). Calcium antagonists (CAs), on the other hand, appear to be neutral with regard to metabolic effects (10–12).
Since 1999, there have been 11 prospective, randomized clinical trials comprising over 106,000 patients with or at high risk of CV disease treated with standard CV therapies for prevention management (13–23). Included among these are a total of 14,590 black and 13,391 Hispanic patients, and 94% of all patients had a mean age of ≥60 years at the time of enrollment. Data from these trials add to the growing evidence that supports the notion that certain CV drugs have important effects on diabetes development, and the strength of this evidence is now too strong for physicians to ignore. These trials include over 88,000 patients who did not have diabetes at enrollment. After follow-up, ranging from one to eight years, there are consistent findings related to classes of drug treatment and the development of new diabetes. The purpose of this paper is to examine this evidence base, which is summarized in Figure 1.
The initial report of a difference in new diabetes emergence in a large randomized CV disease trial came from the Captopril Prevention Project (CAPPP) (13), where 10,413 hypertensive, nondiabetic patients were randomized to either a captopril or BB with or without thiazide diuretic treatment group. After a mean follow-up of 6.1 years, 6.5% of those assigned to captopril versus 7.3% of those assigned to BB/thiazide diuretic developed diabetes. A short time later, the Swedish Trial in Old Patients with Hypertension-2 (STOP-2) study investigators (14) reported in 5,895 nondiabetic, hypertensive elderly patients, a non-significant trend in the development of new diabetes by comparing treatment groups containing either BBs and/or diuretics (atenolol, metoprolol, pindolol, or hydrochlorothiazide [HCTZ] plus amiloride) (4.9%), an ACE inhibitor (either enalapril or lisinopril) (4.7%), or a CA (either felodipine or isradipine) (4.8%).
The Heart Outcomes Prevention Evaluation (HOPE) then reported on 5,720 nondiabetic patients with or at high risk of CAD, randomized to ramipril or placebo plus usual care medications (15). Usual care included CAs, BBs, and/or diuretics in most cases. After 4.5 years, 3.6% of those assigned to ramipril and 5.4% of those assigned to placebo developed new diabetes. Recently, the HOPE investigators reported results from HOPE-TOO, which extends follow-up to 7.1 years in a subset of 6,786 patients. Despite the majority (67%) of HOPE-TOO patients being continued on or switched to ramipril during the additional 2.6 years of follow-up, of those who were nondiabetic at HOPE entry and continued to be followed, 7.2% and 10.2% of those originally assigned to ramipril and placebo, respectively, developed diabetes.
Next, the Intervention as a Goal in Hypertension Treatment (INSIGHT) investigators reported on 5,019 nondiabetic, hypertensive patients randomized to either the nifedipine gastrointestinal transport system (GITS) or thiazide diuretic co-amilozide with or without a BB (16). After almost five years, 5.4% of those assigned to nifedipine GITS and 7.0% of those assigned the thiazide diuretic developed new diabetes. The Losartan Intervention Evaluation (LIFE) group then reported that among 7,998 nondiabetic, hypertensive patients with left ventricular hypertrophy, new diabetes occurred in 6% of those assigned to losartan and 8% in those assigned to atenolol with or without HCTZ (17). Shortly thereafter, the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) investigators reported that among 21,294 nondiabetic patients with hypertension, 9.8% of those assigned to amlodipine, 8.1% assigned to lisinopril, and 11.6% assigned to chlorthalidone developed diabetes (18).
The Second Australian National Blood Pressure Study (ANBP2) revealed that among 5,626 nondiabetics, 4.54% of those treated with enalapril compared with 6.58% of those treated with HCTZ developed diabetes (19). Soon thereafter, the Study on Cognition and Prognosis in the Elderly (SCOPE) investigators reported that of the 4,342 nondiabetics enrolled, 4.3% of those treated with candesartan compared with 5.3% of the those treated with placebo (and background antihypertensives) developed diabetes (20). Then, the Antihypertensive Treatment and Lipid Profile in a North of Sweden Efficacy Evaluation (ALPINE) trialists reported on 392 nondiabetic, low-risk Scandinavian hypertensive patients (21). They found new diabetes in only 0.5% of those assigned to candesartan with or without felodipine and 4.0% of those assigned to HCTZ with or without atenolol.
The Candesartan in Heart Failure Assessment of Reduction in Mortality and Morbidity (CHARM) investigators subsequently reported that among 5,439 nondiabetic chronic heart failure patients, 6.0% of those assigned to candesartan and 7.4% of those assigned to placebo added to a background of BBs and diuretics in most cases, developed diabetes (22). Finally, the International Verapamil-Trandolapril Study (INVEST) trialists reported that in 16,176 nondiabetic, hypertensive patients with CAD, the incidence of new diabetes was significantly lower in the verapamil sustained release/trandolapril strategy (7%), compared with the atenolol/HCTZ strategy (8.2%) (23). Treatment with HCTZ was associated with new diabetes in both strategies, whereas increased exposure to the ACE inhibitor trandolapril in the verapamil sustained release strategy appeared to be associated with more protection from new diabetes than the atenolol/HCTZ strategy.
These data from 11 trials appear consistent. In each of these trials, the point estimate for risk of new diabetes suggested fewer cases in one randomly assigned treatment group than the other, and in nine of these 11 trials, the difference was statistically significant. The percent reduction in new diabetes according to the type of treatment medication is displayed in Figure 2.There were many common findings comparing the studies. All were prospective, randomized, and included patients with or at high risk of CAD, and the majority had less than optimal blood pressure. Most compelling is that the treatment groups containing either agents blocking the effects of angiotensin II (ACE inhibitors or ARBs) and/or CAs had fewer patients who developed diabetes than treatment groups containing diuretics and/or BBs.
As expected, there are some limitations to these trials. Important among these are that only 7 of the 11 trials were double-blind, whereas the other 4 (13,14,19,23) utilized the prospective randomized open blinded end point (PROBE) design (24). In some of these trials, thiazide diuretics and/or BBs were used in some patients assigned to agents blocking the effects of angiotensin II (ACE inhibitors or ARBs) (13,15–20,22,23) or CAs. This design probably minimized the differences observed in emergence of new diabetes between the treatment groups in these trials. The definition of “diabetes” differed among the trials. The same or similar blood pressure reduction was not achieved in each of the treatment groups in all of the studies, which may have contributed to the disproportionate development of diabetes between treatment groups in some of the studies (17,18,20). Lastly, STOP-2 and SCOPE (14,20), which demonstrated a non-significant trend, enrolled elderly subjects (at least 70 years old; mean age 76 years). Data from Narayan et al. (25) suggest a plateau in the cumulative development of diabetes by 80 years of age, which many of these patients would have exceeded by the end of five to six years of follow-up. Additionally, the STOP-2 and SCOPE patients were all Caucasian and relatively lean (mean body mass index 27 kg/m2) and did not have significant CV or other co-morbidities, so they were at the lowest risk of developing diabetes (20,25,26). Furthermore, in the STOP-2 study, the doses of antihypertensive agents in each of the three treatment groups were relatively low in comparison to the trials subsequently published, which may have contributed to the non-significant differences in new diabetes when comparing the groups. These limitations notwithstanding, we believe that the total data set is very compelling, but clearly more research needs to be done in this important area.
These data suggesting a significant reduction in the risk of new diabetes, ranging from 15% to 87%, resulting from drug therapies containing ACE inhibitor, ARB, or CA agents, should not be ignored. In addition, recent data suggesting there are patient characteristics that may be associated with an increase in the risk of new diabetes (i.e., chronic heart failure, left ventricular hypertrophy, U.S. residency, Hispanic ethnicity, black race, previous stroke, increased body mass index, low serum high-density lipoprotein, high non-fasting serum glucose, elevated systolic blood pressure, increased age, female gender, and history of antihypertensive drug use) (25,27,28) make it imperative for physicians to take notice of these factors when selecting long-term CV medications for patients, particularly those at high risk of developing diabetes. Once patients develop diabetes, the CV risk implications are significant, and after a mean of six years, patients who develop diabetes after antihypertensive medication use are at the same increased risk of CV adverse outcomes as established diabetics (29).
- Abbreviations and Acronyms
- angiotensin-converting enzyme
- angiotensin receptor blocker
- calcium antagonist
- coronary artery disease
- Heart Outcomes Prevention Evaluation
- Study on Cognition and Prognosis in the Elderly
- Swedish Trial in Old Patients with Hypertension-2
- Received February 6, 2004.
- Revision received March 19, 2004.
- Accepted March 23, 2004.
- American College of Cardiology Foundation
- Lithell H.O.
- Samuelsson O.,
- Hedner T.,
- Berglund G.,
- Persson B.,
- Andersson O.K.,
- Wilhelmsen L.
- Holzgreve H.,
- Nakov R.,
- Beck K.,
- Janka H.U.
- Tatti P.,
- Pahor M.,
- Byington R.P.,
- et al.
- Brown M.J.,
- Palmer C.R.,
- Castaigne A.,
- et al.
- ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research Group
- Reid C.M.,
- Johnston C.I.,
- Ryan P.,
- Willson K.,
- Wing L.M.
- Lindholm L.H.,
- Persson M.,
- Alaupovic P.,
- Carlberg B.,
- Svensson A.,
- Samuelsson O.
- Pepine C.J.,
- Handberg E.M.,
- Cooper-DeHoff R.M.,
- et al.
- Cooper-DeHoff R.M.,
- Bakris G.L.,
- Cohen J.D.,
- Messerli F.H.,
- Erdine S.,
- Pepine C.J.