Reflections on Diabetes, Clinical Trials, and Cardiovascular Morbidity and Mortality
Abstract
Hypertension and diabetes are commonly found conditions, which predispose to premature cardiovascular morbidity and mortality. A strong consensus has emerged in support of aggressive blood pressure reduction in order to forestall the almost inevitable complications that follow from being a hypertensive diabetic. As of yet though it is not clearly determined as to what represents the best class(es) of antihypertensive medications to effect such blood pressure reduction. In this regard, considerable debate has arisen as to the cost/benefit ratio of dihydropyridine calcium channel blockers in the hypertensive diabetic. Although studies such as the Fosinopril vs. Amlodipine Cardiovascular Events Trial and the Appropriate Blood Pressure Control in Diabetes study would seem to argue against the use of dihydropyridine calcium channel blockers in the diabetic hypertensive, other studies such as the subset analyses of the Syst-Eur and the Syst-China and the Hypertension Optimal Treatment study provide almost compelling evidence for the safety of low to moderate doses of a dihydropyridine calcium channel blockers in this population. Safety issues of dihydropyridine calcium channel blockers will remain unresolved until the release of the Antihypersensitive and Lipid Lowering Study to Prevent Heart Attack results at which time a resolution to this question should be forthcoming.
Hypertension and diabetes are mutually reinforcing conditions that predispose to premature cardiovascular morbidity and mortality and renal disease. The prevalence of hypertension in type II diabetes is considerable. For example, among type II diabetics, approximately 40% are hypertensive by 45 years of age, and by age 75 years at least 60% are hypertensive.1 In patients with diabetes, up to 75% of the excess cardiovascular and renal risk can be attributed to hypertension.2 These observations underscore the importance of blood pressure (BP) control in the hypertensive diabetic. Unfortunately, the issue of BP reduction in the hypertensive diabetic extends beyond merely reducing BP. For example, despite mounting evidence in support of the concept “lower is better,” the optimal goal BP in the hypertensive diabetic is still actively debated. In addition, considerable controversy has surfaced in regards to which antihypertensive medication class is best suited for this highly vulnerable population. The basis for this controversy has been the implied cardiovascular risk, which allegedly accompanies the use of dihydropyridine calcium channel blockers (CCBs).
In a metaanalysis published in 1995, a hypothesis was raised that dihydropyridine CCBs might provoke rather than prevent myocardial infarction in patients with preexisting coronary artery disease.3 This metaanalysis heralded the start of a heated exchange, which still has not been resolved. In fact, this controversy was recently revived with the publication of a series of articles, which suggested that second generation CCBs, such as amlodipine4 and nisoldipine,5 might be harmful in diabetic patients with hypertension. The Isolated Systolic Hypertension in Europe Trial (Syst-Eur),6 Systolic Hypertension in China (Syst-China),7 and Hypertension Optimal Treatment (HOT) trial,8 strongly argue against this hypothesis. These five trials underscore the diversity of opinion surrounding CCB therapy in the treatment of the hypertensive diabetic.4–8 One concept that is no longer debated is that tight BP control is critical if end organ complications are to be warded off in the hypertensive diabetic.9,10
Establishing the risk benefit relationship for CCBs in the hypertensive diabetic is a work still in progress. The Antihypertensive and Lipid Lowering Study to Prevent Heart Attack (ALLHAT), the results of which are due in 2002, may answer this question. This study, involving in excess of 40,000 high risk, stage I and II hypertensive patients, contains an α1-adrenergic antagonist limb utilizing doxazosin, and will compare this drug class to three alternative treatments: the diuretic chlorthalidone, the CCB amlodipine, or the ACE inhibitor lisinopril.11
Calcium channel blockers
Negative trials.
Fosinopril vs. amlodipine cardiovascular events trial. The Fosinopril vs. Amlodipine Cardiovascular Events Trial (FACET) was an open label, randomized study in patients with hypertension and type II diabetes.4 Findings from FACET were originally presented as a poster at the 56th Meeting of the American Diabetes Association (San Francisco, CA, June 8–11, 1996). Its primary intent was to assess treatment related differences in serum lipid levels, diabetes control, and renal function. Patients were randomly assigned to either fosinopril (20 mg/daily) or amlodipine (10 mg/daily) as first line drugs. If goal BP was not reached, the alternative medication could be added; thus, a portion of the treatment population received fosinopril together with amlodipine. According to this presentation, the treatment population was comprised of 390 subjects fairly equally distributed among three treatment groups (amlodipine (n=140), fosinopril (n=130), and one treated with both drugs (n=110).
An intention to treat analysis was employed to assess cardiac events, which were defined as well documented acute myocardial infarction or new onset angina pectoris. In the amlodipine and fosinopril treated groups, 17 and seven events occurred, respectively. Three events occurred in the combination therapy group. The investigators were impressed by the low number of major cardiac events seen in the combination therapy group, which led them to suggest that the combination of an ACE inhibitor and a dihydropyridine CCB would be both a logical and therapeutically beneficial approach to the treatment of hypertension in the diabetic.
FACET has been roundly criticized in a number of circles and may well have provided an example of how not to conduct a clinical trial.12,13 A number of criticisms were directed towards FACET. For example, this was a single center, open label study, with six month intervals between visits. Events were monitored by asking the patient if they had been either hospitalized or had experienced any other event. Study conditions such as these, no doubt introduce considerable bias into study findings. The original intent of this study was not to identify between drug differences in the occurrence of cardiovascular end points and, therefore, was never powered for such a determination. In fact, the final study report limited its analysis to the two randomized groups whereas the preliminary findings described all three treatment groups. Furthermore, the preliminary analysis was revised so as to include end points other than morbid cardiac events, such as stroke. These issues severely limit the applicability of the FACET findings.
Appropriate blood pressure control in diabetes trial. The Appropriate Blood Pressure Control in Diabetes (ABCD) trial was a prospective, randomized study, in patients with type II diabetes. It was designed to test the primary hypothesis that two modes of treatment—intensive vs. moderate BP reduction—would either prevent cardiovascular events or slow the progression of nephropathy, neuropathy, and retinopathy. A secondary hypothesis of this study was that a long acting dihydropyridine CCB nisoldipine and an ACE inhibitor enalapril would have equivalent effects on the rate at which diabetic complications progressed.5
A total of 950 subjects with diabetes, both normotensive (n=480) and hypertensive (n=470), were randomly assigned to moderate (target diastolic BP, 80–89 mm Hg) or intensive (target diastolic, 75 mm Hg) antihypertensive treatment, administered in a double blind fashion. In the hypertensive cohort, patients were randomly assigned to either nisoldipine or enalapril as a primary antihypertensive medication. Nisoldipine was started at 10 mg with titration to 20, 40, or 60 mg/day as necessary, whereas enalapril treatment began at 5 mg with increases to 10, 20, or 40 mg/day, as warranted. In the normotensive cohort, the moderate treatment group received placebo. If study medication did not bring BP to goal, add on therapy was permissible. In this regard, open label metoprolol and hydrochlorothiazide could be added.
After 67 months, the Data Safety Monitoring Committee observed a significant difference in cardiovascular event rates between the two treatment groups in the hypertensive cohort. The enalapril treatment group had fewer cardiovascular events than the nisoldipine treatment cohort in both the intensive (12 fatal/nonfatal myocardial infarctions for nisoldipine vs. three for enalapril) and moderate (13 fatal/nonfatal myocardial infarctions vs. one for enalapril) BP treatment groups. These data led to a computed adjusted risk ratio of seven (95% CI: 2.3–21.4) for the combined end point of fatal and nonfatal myocardial infarctions (nisoldipine: enalapril). Because >50% of the subjects were not taking the original study medications by the end of the study, an additional analysis, according to actual drug exposure, was performed. This yielded a continued significant difference in the rate of myocardial infarctions between nisoldipine and enalapril treated patients. These findings suggest that patients taking nisoldipine were more apt to experience a myocardial infarction and to do so earlier than enalapril treated patients. Because of the presumed gravity of these findings, those in the hypertensive cohort having been randomized to nisoldipine were reassigned to treatment with enalapril.
The results of the ABCD trial require careful interpretation for several reasons. The decisions to prematurely stop the study and to report a secondary end point only in the hypertensive subgroup are open to critique. The current treatment environment for the hypertensive diabetic is highly charged, which conceivably could have compelled the Data and Safety Monitoring Committee to act on an extreme result. This result was not prespecified and not subject to monitoring boundaries, which may have inflated the risk of a false-positive finding. Other confounding variables of this interim analysis included the fact that diuretic and ß-blocker add on therapy was more common in the enalapril treatment group, and study medication was discontinued more frequently in the nisoldipine treatment group. Because of these differences, overall cardiovascular protection may have been unbalanced in favor of the enalapril treated group. Additional statistical analyses, which are more suited to this type of data set—such as a per protocol analysis—would have been instructive.
In the context of the ABCD results,5 the decisions of the monitoring board of ALLHAT are instructive since one of the treatment limbs of ALLHAT includes the long acting CCB, amlodipine.11 At the time of the ABCD publication,5 one-third of the 40,000 patients enrolled in the ALLHAT study were diabetic. The Data and Safety Monitoring Committee for this study periodically meets to review outcome data. An interval analysis was undertaken in the fall of 1997, when the CCB debate was first cresting, with more than 7000 patient years of experience in the diabetic patient subgroup. The ALLHAT committee saw no reason to alter the study at that time and recommended that the trial continue according to protocol. Furthermore, they did not feel compelled to advance the date of its next data and safety review.14
Positive trials.
Syst-Eur trial. The first of these is a short report from the Syst-Eur trial.6 These investigators performed a post hoc analysis in 492 type II diabetic patients with isolated systolic hypertension. They were compared to 4203 nondiabetic patients who had also entered the study. In this trial, CCB therapy was first line treatment, with ACE inhibitors and diuretics as add on therapy when necessary. In the diabetic subset of this study, noteworthy benefits were observed when systolic BP was reduced. After adjustment for possible confounding variables, active treatment was found to have reduced overall mortality by 55% and fatal and nonfatal strokes by 73%. Among the nondiabetic patients, active treatment decreased all cardiovascular events combined by 26% and fatal and nonfatal strokes by 38%. Thus, treatment of systolic hypertension in this diabetic cohort virtually abolished the excess risk for major cardiovascular events in diabetics compared to nondiabetics.
Hypertension optimal treatment study. The HOT Study was a multinational trial, which included 18,790 patients with hypertension and diastolic BP between 100–115 mm Hg who were randomly assigned to a spectrum of target diastolic BP. Patients were allocated to a target diastolic BP ≤90 mm Hg (n=6264), to goal diastolic BP values ≤85 mm Hg (n=6264), or to target diastolic values ≤80 mm Hg (n=6262). In addition, patients were randomized in a double blind fashion to a low dose, 75 mg/day of acetylsalicylic acid (n=9399) or identical appearing placebo tablets (n=9391). Patient randomization occurred on the basis of age, sex, prior antihypertensive therapy, smoking, prior myocardial infarction, diabetes mellitus, previous coronary artery disease, or stroke. Antihypertensive therapy, with the long acting calcium antagonist, felodipine (5 mg/daily) was given to all patients. Four additional treatment steps were utilized in a specific sequence (Table I). The average follow up time in these studies was 3.8 years.8,15
| Step 1 | 5 mg felodipine |
| Step 2 | 5 mg felodipine + low dose ACE inhibitor or ß-blocker |
| Step 3 | 10 mg felodipine + low dose ACE inhibitor or ß-blocker |
| Step 4 | 10 mg felodipine + high dose ACE inhibitor or ß-blocker |
| Step 5 | 10 mg felodipine + high dose ACE inhibitor or ß-blocker + low dose alternative addition of hydrochlorothiazide |
The main study drugs employed included felodipine, a dihydropyridine CCB (78%), and ACE inhibitors (41%) with ≤30% of the patients on either a ß-blocker or a diuretic. Unfortunately, no analysis of the relation between specific drug and cardiovascular events is provided; thus, the possibility that the predominant drugs employed in these studies, a dihydropyridine CCB and/or an ACE inhibitor contributed to the overall reduction in cardiovascular risk, can neither be excluded nor established. However, the results of this trial are reassuring in light of the growing concern that BP reduction with dihydropyridine CCB therapy is accompanied by an increased cardiovascular risk. In fact, the one marginally significant result on an intention to treat basis, was the trend to a lower risk of all myocardial infarctions in the ≤80 mm Hg treatment group. This group, in fact, was most intensively treated and received higher doses of felodipine (as much as 10 mg of felodipine by treatment protocol).
There were several positive outcomes from the HOT Study (Table II); the most important of which was the observation that there was no J shaped curve in the diabetic subset of this study. The HOT Study clearly documents that in diabetic patients with hypertension, lowering BP is beneficial and the lower the BP, the greater the benefits. Although the HOT Study has not solved all the therapeutic issues in the management of hypertension, it has shown that when used properly in clinical practice, antihypertensive drugs (often in combination) can normalize BP in the majority of hypertensive patients.
| • Normalization rate for diastolic blood pressure in 90% of subjects with low dose combination antihypertensive therapy |
| • Demonstration of cardiovascular safety with low dose dihydropyridine therapy |
| • Cardioprotection for the diabetic cohort in these studies |
| • Demonstration of a positive sense of well being at lower BP |
| • Reduction in major cardiovascular events as well as a decrease in fatal and nonfatal myocardial infarction rate with low dose aspirin therapy without an increased risk of intracerebral bleeding |
| • Demonstration that the lower the BP the slower the rate of decline in renal function16 |
| • Proof that a large BP mortality study can be completed without governmental support |
In the context of diabetic hypertensives, the HOT trial was a second study, which proved that BP could be safely reduced in hypertensive diabetics when a CCB based therapeutic regimen is used.8 Of the patients in this trial, 8% or 1501 were diabetic. In this study, the subgroup of patients with diabetes fared particularly well by having their BP tightly controlled. In these patients, a decline in major cardiovascular events was seen (p for trend=0.005). Among those with diastolic BP ≤80 mm Hg, the risk of major cardiovascular events was halved in comparison with the ≤90 mm Hg. These observations are in line with the Syst-Eur findings in patients with diabetes.6
Other diabetic trials.
United kingdom prospective diabetes study. The Hypertension Diabetes Study is part of the larger United Kingdom Prospective Diabetes Study (UKPDS) that focused on various approaches to glycemic control among 5102 newly diagnosed type II diabetics patients, age 25–65 years. The 1148 subjects with hypertension within the larger trial were randomized to conventional vs. more aggressive BP goals (tight control=BP <150/<85 mm Hg; conventional control=<180/<105 mm Hg). The patients allocated to tight control goals were further randomized to receive captopril (n=400) or atenolol (n=358) in an unblinded comparison. Captopril was administered twice daily, up to 50 mg per dose and atenolol once daily, up to 100 mg per dose. There was a choice of four step up drugs from different drug classes.10
It was observed that captopril and atenolol were equally effective in reducing BP to a mean of 144/83 and 143/81 mm Hg, respectively, with a similar proportion of patients (27% and 31%, respectively) requiring ≥3 antihypertensive drugs. Patients randomized to tight BP control goals showed a highly significant 24% reduction in any diabetes related end point, with a 32% reduction in mortality from “deaths related to diabetes” (e.g., myocardial infarction, stroke, peripheral vascular disease, and microvascular disease). No difference was observed between captopril and atenolol in the incidence of single or aggregate macrovascular (e.g., myocardial infarction, heart failure, or angina) or microvascular (e.g., retinopathy progression, amputation, renal failure, or progression of albuminuria) clinical end points. Mean BP was 144/82 mm Hg in the tight control group and 154/87 mm Hg in the less tight control group. Even this seemingly minor difference of 10/5 mm Hg between treatment groups was associated with the observed significant reductions in risk.
These two UKPDS studies indicate that tight BP control goals are extremely important in the prevention of macro and microvascular complications in patients with type II diabetes and hypertension. These benefits are the same whether a ß-blocker or an ACE inhibitor is used as the primary antihypertensive agent. The findings found that captopril and atenolol were equally effective in reducing BP to a mean of 144/83 mm Hg and 143/81 mm Hg, respectively, with a similar proportion of patients (27% and 31%, respectively) requiring ≥3 antihypertensive drugs. Similar to the ABCD trial, wherein most physicians would have predicted superiority of the ACE inhibitor in terms of organ protection, surprisingly this was not the case. The dosing regimen employed in this study can be criticized in that both the ACE inhibitor captopril and the ß-blocker atenolol were administered once daily and this may have precluded a between treatment difference, in favor of the ACE inhibitor, from being expressed.
Finally, a recent substudy publication of the Syst-China study7 and a subgroup analysis of the SHEP trial9 further corroborate both the safety of dihydropyridine CCB as well as the importance of BP reduction in the diabetic hypertensive. Collectively, these studies argue very strongly for aggressive lowering of BP in patients with diabetes and hypertension. They indicate that the current goal in the ADA and the Sixth Report of the Joint National Committee (JNC VI) guidelines of ≤ 130/85 mm Hg is a correct one, which in now supported by prospective clinical trial evidence. BP reduction in hypertensive diabetic patients has now emerged as a major goal of preventive therapy in the diabetic.
Conclusions
Perhaps the major message now being delivered by these recently completed clinical trials is, what JNC VI and World Health Organization-International Society of Hypertension say: lower the BP more than the usual goal of 140/90 mm Hg in the treatment of the patient with diabetes.17,18 How low to go remains uncertain, but 130/80 mm Hg seems appropriate until more data becomes available to direct us otherwise. Obviously, it takes multiple drugs to adequately control hypertensive patients, and, in particular, the hypertensive diabetic. Multidrug therapy for the treatment of hypertension can occur with either sequential drug addition or by the use of fixed dose combination therapy. The latter is experiencing a resurgence in use with the arrival of effective and well tolerated, fixed dose combination therapies comprised of low dose amounts of either a ß-blocker, ACE inhibitor, or angiotensin receptor blocker with a diuretic. In addition to the above, the use of an ACE inhibitor together with a CCB in a fixed dose combination offers a number of benefits which may go beyond simply BP reduction.
Current information appears inadequate for a pharmacologic ban on the use of dihydropyridine CCBs in the hypertensive diabetic. If a dihydropyridine CCB does in fact increase sympathetic nervous system activity—and with it the risk of myocardial infarction—it is no doubt a dose related event. To circumvent this dose related phenomenon, careful consideration should be given to the use of the smallest dose of dihydropyridine CCB, which might still then achieve BP control, typically in combination with other complementary antihypertensives. In fact, the majority of the response to any antihypertensive compound is seen to occur at its lower doses. The practice of dose titration of an antihypertensive compound, in order to obtain BP control, is seldom defensible beyond possibly one dose titration step. Increases in dose amount much beyond one serial doubling typically results in a deterioration of the cost benefit ratio.19
Commentary
The ALLHAT study recently terminated the doxazosin (Cardura Ó) treatment limb because this treatment group was observed to have 25% more cardiovascular events and were twice as likely to be hospitalized for congestive heart failure as users of chlorthalidone. The drugs were similarly effective in preventing heart attacks and in reducing the risk of death from all causes. The remainder of the ALLHAT study, which began in 1994, will continue as scheduled and is expected to end in 2002. Other findings about doxazosin in comparison to chlorthalidone are: those in the doxazosin group had slightly higher blood pressures than the chlorthalidone group, although the diastolic pressures were the same. The doxazosin group also had poorer compliance with treatment, only 75% were still on the drug or another α-blocker after four years, compared with 86% still taking chlorthalidone or another diuretic. Dr. Jeff Cutler, Director of the NHLBI Clinical Applications and Prevention Program and ALLHAT Project Officer states “patients on an α-blocker for high blood pressure should see their doctor and not just stop taking it.” He further emphasized that “we cannot conclude that the drug was harmful. Rather it did not work as well as the diuretic in reducing cardiovascular disease.” The future of peripheral α-blockers although not entirely bleak, does appear much cloudier.
Domenic A. Sica, MD James Pool, MD
