Pharmacologic Therapy of Mild to Moderate Hypertension: Possible Generalizability to Diabetics1

Transcription

1 Pharmacologic Therapy of Mild to Moderate Hypertension: Possible Generalizability to Diabetics1 Patricia R. Hebert,2 JoAnn E. Manson, and Charles H. Hennekens P.R. Hebert, J.E. Manson, OH. Hennekens, Departments of Medicine and Preventive Medicine, Channing Laboratory. Harvard Medical School and Brigham and Women s Hospital, Boston, MA (J. Am. Soc. Nephrol. 1992; 3: ) required to determine the efficacy of pharmacologic treatment of hypertension among diabetics could only derive from randomized trials of adequate sampie size to detect the most plausible small to moderate benefits. Key Words: Hypertension, diabetes, ant/hypertensive drugs ABSTRACT This article reviews the evidence on pharmacologic therapy of hypertension in reducing morbidity and mortality from stroke and coronary heart disease (CHD) and considers the possible generalizability of these findings to diabetics. For malignant hypertension, benefits are large and obvious from uncontrolled case series. For severe hypertension, conclusive benefits have been shown in several randomized trials. For mild to moderate hypertension, however, it is necessary to consider meta-analyses of all individual trials. The most comprehensive of these shows reductions of 42% for total stroke (95% CI, -33 to -50%; P< ) and 14% for all CHD (95% CI, -4 to -22%; P< 0.01). The applicability to diabetics is unclear because they were excluded from most of the trials. The Hypertension Detection and Follow-Up Program included diabetics and reported subgroup analyses. The reduction in mortality among the actively treated diabetics of 5% was less than the I 7% achieved in nondiabetics. It is unclear, however, whether the mortality reductions are truly different or reflect the play of chance. Because of the higher incidence of CHD events among diabetics with hypertension, a similar relative benefit would result in a much greater absolute risk reduction. Further, the drugs used adversely affect lipid and glucose metabolism. New antihypertensive drugs without these side effects may further improve the risk-to-benefit ratio of antihypertensive treatment, especially in diabetics, who are at a several-fold absolute increased risk or cardiovascular disease. The data I Presented in part at the Third International Symposium on Hypertension Associated with Diabetes Meltitus. 2 Correspondence to Dr. P.R. Hebert, 900 Commonwealth Avenue East. Boston, MA /0304-SI 35$03.00/0 Journal of the American Society of Nephrology Copyright C 1992 by the American Society of Nephrology O bscrvational epidemiologic studies, both case control and cohort, have consistently shown that hypertension is a major and independent risk factor for both stroke and coronary heart disease (CHD) (1). Nobody would disagree with the public health importance of hypertension. Hypertension affects approximately 60 million Americans, or one in three adults (2). About half of the hypcrtensivcs, or 1 5% of the general adult population, have mild to moderate elevations, which account for the majority of premature deaths (3). Thus, for a condition as common as CHD, even a small decrease in mortality associated with treatment, as for example, a 1 0 to 1 5% reduction, would have a large public health effect. This article will review the evidence on the robe of pharmacologic treatment of hypertension in reducing morbidity and mortality from stroke and CHD and will consider the possible gencralizabibity of these findings to diabetics who have about a twofold to fivefold increased risk of cardiovascular disease (4,5). In malignant hypertension, the benefits of pharmacobogic therapy appeared so great from early uncontrolled case series that randomized trials were neither necessary nor ethical and therapy for this condition quickly became widely accepted (6-9). With respect to severe hypertension (diastolic blood prcssure (BP) mm Hg), three randomized trials (10-1 2) taken together demonstrated an 80% reduction in a!! hypertensive events in treated patients, which was highly significant (13). For mild to moderate hypertension, defined as diastolic BP between 90 and mm Hg, data from individual randomized trials (3,14-19) on the benefits of treatment on various cardiovascular endpoints have demonstrated that pharmacologic treatment clearly reduces the incidence of stroke, but possible benefits on coronary disease remain controversial. This situation may have resulted from the fact that data from any single trial may have failed to reach conventional levels of statistical significance simply Journal of the American Society of Nephrology SI 35

2 Pharmacologic Therapy of Mild to Moderate Hypertension because of inadequate sample size. In such circumstances, an overview of meta-analysis can be performed so that beneficial (or harmful) effects of treatment may become evident that might not have been detected in individual randomized trials because they lacked sufficient statistical power to detect the most plausible small to moderate but clinically important differences between the treated and comparison groups. Alternatively, apparently null results, which arc uninformative from single trials, will become more informative as a result of the narrowing of the 95% confidence intervals (CI) as well as the increased precision of the point estimates, both because of increased sample size. For these reasons, the most precise estimates of the effect of antihypertensive treatment on stroke and CHD have come from recent overviews of individual randomized trials that have been conducted over the last 30 years (13,20,21). SUBJECTS AND METHODS The latest and most comprehensive of these overviews (2 1 ) included 1 4 randomized trials of pharmacobogic therapy conducted from roughly the mid- 1960s to the mid-1980s and included about 37,000 individuals. Entry diastolic BP ranged from 85 to 130 mm Hg; however. in eight of the trials, all participants had entry diastolic BP less than mm Hg. The antihypertensive drugs used were chiefly diuretics or beta-blockers. The mean duration of pharmacologic treatment in these trials was 5 yr and as compared with controls, the average reduction in diastolic BP achieved was 5 to 6 mm Hg. For the overview, each trial was treated as an individual stratum and standard statistical methods for combining information from 2 x 2 tables were used to obtain the summary relative risk estimates. Data from individual trials were weighted by their sample sizes, so large trials were given proportionally greater weights. Endpoints of interest in the overview included cardiovascular mortality, stroke, and all coronary events. For each endpoint of interest, the reduction in risk associated with pharmacologic treatment was calculated as was the corresponding 95% CI. RESULTS For total cardiovascular mortality, when a!! 14 trials were considered together, there was a highly significant 2 1 % reduction (95% CI, to -30% ; P < ). With regard to specific cardiovascular endpoints, the largest risk reductions were seen for stroke. For the combined endpoint of fatal and nonfatal stroke, there was a highly significant 42% reduction (95% CI, -33 to -50%; p < ). For fatal stroke, there was a highly significant reduction in risk of 45% (95% CI, -30 to -58%; P < ). The magnitude of these reductions in stroke from the randomized trials Is consistent with those from observational studies. For the combined endpoint of fatal plus nonfatal coronary events, there was a statistically significant 14% reduction (95% CI, -4 to -22%; P < 0.01). For fatal coronary events, the reduction in risk was 1 1 % (95% CI, +4 to -24%), a reduction that did not achieve statistical significance. Further, three trials were completed after the publication of the overview. Their findings contribute important relevant information concerning the possib!c benefits of pharmacologic therapy in reducing CHD. The Systolic Hypertension in the Elderly Program (SHEP) (22), conducted among 4,736 persons age 60 yr or older, reported a 27% reduction in fatal plus nonfatal coronary events in those assigned to active antihypertensive therapy as compared with those given placebo. The much smaller (N = 1,627) Swedish Trial in Old Patients with Hypertension (STOP-Hypertension) (23) reported 30 deaths from CHD, 1 0 among the treated group and 20 among the controls. Finally, the Medical Research Council Trial of Hypertension In Older Adults (24) also supports a benefit on CHD. The addition of these three trials to the overview data shows a reduction in CHD of 17% (P < 0.001) (R. Peto, personal communication). The observed reduction in risk of coronary events seen in the overview of randomized trials is only about half the magnitude of the effect predicted from observational studies (1). The discrepancy between the magnitude of the effect on coronary events found in randomized trials and that predicted from observational studies has several plausible explanations. First, it is possible that the discrepancy is due simply to the play of chance. Second, the relatively short duration of 3 to 5 yr of the BP lowering may yield benefits on stroke due to acute effects on BP bowering. On the other hand, to detect a decrease in coronary events that may involve chronic effects on athcrosc!crosis could require BP lowering over a decade or more. Third, there may be cardiotoxic side effects of specific antihypertensive drugs that may at least partially counterbalance the benefits of BP reduction. Specifically, the first-line drugs used In these trials were generally thiazide diuretics, which raise lowdensity lipoproteins by 5% (2 1,25). This could correspond to an increase in CHD, at least in theory, of 10 to 15%. POSSIBLE GENERALIZABILITY OF FINDINGS TO DIABETICS The majority of randomized trials of pharmacologic therapy of mild to moderate hypertension either cxcluded diabetics entirely or enrolled only small samplc sizes. In considering the implications of these SI 36 Volume 3. Supplement I 1992

3 Hebert et al findings for diabetics with hypertension, one possibi!ity is that the relative benefits of treatment arc similar. If so, the findings from nondiabetic popubations can be generalized to diabetics. Further, under this assumption, the absolute benefits of treatment would be expected to be far greater in diabetics, because of their increased incidence of hypertensionrebated complications, including coronary disease, stroke, and congestive heart failure. The data from observational cohort studies tend to support the deleterious interaction between hypertension and diabetes (4,5,26). For example, in the Nurses Health Study, rates of total CHD (nonfatal myocardial infarction and fatal CHD combined) were about four times greater in those with both hypertension and diabetes, specifically 434 per 1 00,000 person yr in hypertensives with diabetes as compared with per 1 00,000 in those with hypertension alone (26). Assuming a 14% decrease in the incidence of CHD with pharmacologic therapy, treating 1 00,000 hypertensives with diabetes for a year would result in the prevention of about 50 coronary events, whereas treating the same number of hypertensives without diabetes would prevent 1 4 coronary events. Even if antihypertensive treatment was of lower relative benefit in diabetics, given the greater severity of the pathologic sequclac of hypertension in this population, the absolute benefits of treatment could still prove to be substantially higher than In nondiabetics. However, this remains mainly a matter of speculation because so few data are currently availabbe on the direct effects of antihypertensive therapy in diabetics. On the other hand, effects of antihypertensive therapy may be different in diabetics compared with nondiabetics. Support for this hypothesis is provided by subgroup analyses, which arc useful to formulate but not to test hypotheses. The Hypertension Detection and Follow-up Program (HDFP) was one of the largest trials in the overview, contributed the greatest number of endpoints, and Included a subgroup of diabetics. This community-based trial compared aggressive antihypertensive therapy in a stepped-care regimen with usual care received in the community with respect to reducing 5-yr mortality among 10,940 individuals, ages 30 to 69 yr. Further, participants were randomized in strata according to the bevel of their entry diastolic pressure, which permitted a finer analysis of treatment effects by BP levels. Among nondiabetics, the mortality rate was reduced by 17% in the active group (27). Among diabetics, a smaller 5% reduction in mortality was observed, although because of the small number of diabetics, It is unclear whether the mortality reductions arc truly different or reflect the play of chance. Among the subgroup of diabetics with mild hypertension (entry diastobic BP from 90 to 1 04 mm Hg) who were not on antihypertensive drugs at baseline, those assigned to aggressive treatment had a 27% bower mortality than that observed among the usual-care controls. This decrease in mortality was similar to the 2 1 % reduction observed in the aggressively treated mild hypertensives without diabetes. These results raise the possibibity that, overall, treatment for hypertension may have a smaller relative benefit in diabetics than in nondiabetics. These considerations are compatible with the possibility that the clock for CHD may have been ticking far too long both before and after the onset of clinical diabetes for BP Interventions to have any significant effect (28). These findings, however, also raise the possibility that treatment begun early, when BP elevations arc likely to be milder, may be more beneficial than therapy administered later when BP elevations arc more severe. The issue of side effects of antihypertensive mcdications is of particular concern in treating hypertensives who arc also diabetic. Both thiazide diuretics and beta-blockers, the drugs that have been most widely used in randomized trials, adversely affect glucose and lipid metabolism. An alternative in treating diabetics would be to use some of the newer antihypertensive drugs that do not have these side effects, including a!pha-adrencrgic inhibitors such as prazosin, calcium channel blockers, or angiotensin-converting enzyme inhibitors. Short-term trials of BP lowering in relatively small samples sizes with these drugs suggest that adequate reductions can be achieved and adverse effects on glucose and lipid metabolism can be avoided. For example. in a study of the calcium channel blocker verapamib and/or the angiotensin-converting enzyme inhibitor enalaprib in patients with non-insulin-dependent diabetes with mild to moderate hypertension, within 30 wk of activc treatment with one or both of these agents, BP levels had been lowered to desired levels in 87% of the 47 patients (29). These drugs did not seem to adversely affect levels of fasting plasma glucose or total lipids and lipoproteins. CONCLUSIONS Pharmacologic therapy of malignant, severe, and mild to moderate hypertension seems to clearly reduce risks of stroke and, to a lesser extent, CHD (30). Virtually all available data, however, have been accumulated on nondiabetics. Although there is little direct evidence from randomized trials on the effectivcncss of antihypertensive treatment in diabetics as compared with nondiabetics, we have no compelbing evidence that diabetics would experience less benefit. Further, if the relative benefits of treatment arc similar, the absolute benefits of treatment would be expected to be much greater in diabetics because of their increased incidence of complications of hy- Journal of the American Society of Nephrology 5137

4 Pharmacologic Therapy of Mild to Moderate Hypertension pertension, including coronary events and stroke. On the other hand, to get such benefits might require early diagnosis and initiation of therapy in hypertensives with diabetes. The choice of drug therapy is of particular concern because of the adverse effects of diuretics and beta-blockers, frequently used as firstline therapies, on glucose and lipid metabolism. Newly introduced antihypertensivc drugs may have advantages over thiazide diuretics and beta-blockers. However, their efficacy, in diabetics as well as in nondiabetics, has not been tested directly in bargescale, randomized trials. For most hypotheses, randomized trials are neither necessary nor desirable. For example, lifelong cigarette smokers have about 20 times the risk of lung cancer as those who were never smokers (31). This hypothesis has been adequately tested in case-control and observational cohort studies. Further, current smokers have about 70% higher rates of CHD mortality (3 1.32). These hypotheses have also been tested in case-control and observational cohort studics. If the treatment of hypertension in diabetics reduces mortality by about 20%, however, this would have important clinical implications and public health effect. Nevertheless, the small to moderate size of the reduction in risk is about as large as the uncontrolled confounding inherent in a!! observational designs, including both case-control and cohort studies. In such circumstances, the only reliable design strategy is the randomized trial, but such trials may need to involve some tens of thousands of participants. In fact, trials of far smaller sample size may yield results of no benefits when they are simply too small to detect the most plausible small to modcrate treatment effects. It is only by the conduct of large-scale. randomized trials of adequate methodologic rigor that direct and reliable evidence will emerge concerning the balance of risk and benefits among hypertensives with diabetes of pharmacologic therapy of hypertension, particularly mild to moderate elevations. ACKNOWLEDGMENTS Supported in part by research grant HL REFERENCES 1. MacMahon S, Peto R, Cutler J, et at. : Blood pressure. stroke, and coronary heart disease: Part 1, prolonged differences in blood pressure: prospective observational studies corrected for the regression dilution bias. Lancet 1 990; Subcommittee on Definition and Prevalence of the 1984 Joint National Committee: Hypertension prevalence and the status of awareness, treatment, and control in the United States. Hypertension 1 985;7: Hypertension Detection and Follow-Up Program Cooperative Group (J. Taylor): The Hypertension Detection and Follow-Up Program: A progress report. Circ Res 1 978;40(supp! I): Fuller JH, Shipley MJ, Rose G, Jarrett RJ, Keen H: Mortality from coronary heart disease and stroke in relation to degree of g!ycaemia: The Whitehall Study. Br Med J 1 983;287: Rosengren A, Welin L, Tsipogianni A, Wilhelmsen L: Impact of cardiovascular risk factors on coronary heart disease and mortality among middle aged diabetic men: a general population study. BrMcdJ 1989:299: Harington M, Kincaid-Smith P. McMichael J: Results of treatment in malignant hypertension. Br Mcd J 1959;2: Sokolow M, Perloff D: Five-year survival of consecutive patients with malignant hypertension treated with antihypertensive agents. Am J Cardiob 1 960;6: Bjork 5, Sannerstedt R, Falkheden T, Hood B: The effect of active drug treatment in severe hypertensive disease. Acta Mcd Scand 1961 ;169: Mohler ER, Freis ED: Five-year survival of patients with malignant hypertension treated with antihypertensive agents. Am Heart J 1 960;60: Hamilton M, Thompson EN, Wisniewski TKM: The role of blood pressure control in preventing complications of hypertension. Lanc#{235}t 1 964; 1: Wolff FW, Lindeman RD: Effects of treatment in hypertension: Results of a controlled study. J Chronic Dis 1966;19: Veteran s Administration Cooperative Study Group on Antihypertensive Agents: Effects of treatment on morbidity in hypertension: Results in patients with diastolic blood pressure averaging through 1 29 mmhg. JAMA 1967;202: Cutler JA, Furberg CD: Drug treatment trials in hypertension: A review. Prey Med 1 985; 14: Hypertension Detection and Follow-up Program Cooperative Group: Five-year findings of the Hypertension Detection and Follow-up Program: I. Reduction in mortality of persons with high blood pressure, including mild hypertension. JAMA 1979;242: Medical Research Council Working Party: MRC trial of treatment of mild hypertension: Principal results. Br Mcd J 1985;291 : Australian National Blood Pressure Management Committee: The Australian therapeutic trial in mild hypertension. Lancet 1 980; 1: Helgeland A: Treatment of mild hypertension: A five-year controlled drug trial: The Oslo Study. Am J Mcd 1980:69: Amery A, Birkenhager W, Brixko P, et at. : Mortabity and morbidity results from the European Working Party on High Blood Pressure in the Elderly trial. Lancet 1 985; 1: Veterans Administration Cooperative Study Group on Antihypertensive Agents: Effects of treatment on morbidity in hypertension: II. Results in patients with diastolic blood pressure SI 38 Volume 3. Supplement I. 1992

5 Hebert et al averaging 90 through mmhg. JAMA 1970; 21: Hebert PR, Fiebach NH, Eberlein KA, Taylor JO, Hennekens CH: The community-based randomized trials of pharmacologic treatment of mild-to-moderate hypertension. Am J Epidemiob 1988; 127: Collins R, Peto R, MacMahon 5, et at.: Blood pressure, stroke and coronary heart disease: Part 2, short-term reductions in blood pressure: Overview of randomised drug trials in their epidcmio!ogical context. Lancet 1 990;335: SHEP Cooperative Research Group: Prevention of stroke by antihypertensive drug treatment in older persons with isolated systolic hypertension: Final results of the Systolic Hypertension in the Elderly Program (SHEP). JAMA 1991 ;265: Dahiof B, Lindholm LH, Hansson L, Schersten B, Ekbom T, Wester P-O: Morbidity and mortality in the Swedish Trial In Old Patients with Hypertension (STOP-Hypertension). Lancet 1991;338: The MRC Working Party: Medical Research Council (MRC) trial of hypertension in older adults: Principal results. Br Med J 1992; 304: Pollare T, Lithell H, Berne C: A comparison of the effects of hydrochborothiazide and captoprib on glucose and lipid metabolism in patients with hypertension. N Eng! J Med 1989;321: Manson JE, Colditz GA, Stampfer MJ, et at.: A prospective study of maturity onset diabetes meblitus and risk of coronary heart disease and stroke in women. Arch Intern Mcd 1991; 151: The Hypertension Detection and Follow-U Program Cooperative Research Group: Morta - ity findings for stepped-care and referred-care participants in the Hypertension Detection and Follow-up Program, stratified by other risk factors. Prey MccFicine 1985;14: Haffner SM, Stern MP, Hazuda HP, Mitchell BD, Patterson JK: Cardiovascular risk factors in confirmed prediabetic individuals: Does the clock start ticking before the onset of clinical diabetes? JAMA 190;263: Ferrier C, Ferrari P, Weidmann P, et at: Antihypertensive therapy with Ca2+: Antagonist verapamib and/or ACE inhibitor enalapril In NIDDM patients. Diabetes Care 1991;14: Moser M, Hebert P, Hennekens CH: An overview of the mcta-analyscs of the hypertension treatment trials. Arch Intern Mcd ; 151: US Department of Health and Human Services: Reducing the health consequences of smoking: 25 years of progress. A report of the Surgeon General. US Department of Health and Human Services, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. DHHS Publication No. (CDC) , Doll R, Peto R: Mortality in relation to smoking: 20 years observation on male British doctors. Br Mcd J 1976;2: Journal of the American Society of Nephrology SI 39