Reduced Cardiovascular Morbidity and Mortality in Hypertensive Diabetic Patients on First-Line Therapy With an ACE Inhibitor Compared With a Diuretic/β-Blocker–Based Treatment Regimen

A subanalysis of the Captopril Prevention Project

  1. Leo Niskanen, MD,
  2. Thomas Hedner, MD,
  3. Lennart Hansson, MD,
  4. Jan Lanke, MD,
  5. Anders Niklason, MB and
  6. for the CAPPP study group
  1. Department of Medicine, University of Kuopio, Kuopio, Finland


    OBJECTIVE—The Captopril Prevention Project (CAPPP) evaluated the effects of an ACE inhibitor–based therapeutic regimen on cardiovascular mortality and morbidity in hypertension. One planned subanalysis of the CAPPP was to evaluate the outcome in the diabetic patient group.

    RESEARCH DESIGN AND METHODS—In the CAPPP, 572 (4.9% of 10,985 hypertensive patients) had diabetes at baseline and were studied according to a prospective, randomized, open, blinded, end point trial design. Patients aged 25–66 years with diastolic blood pressure ≥100 mmHg were included and randomized to receive either captopril or conventional antihypertensive treatment (diuretics and/or β-blockers).

    RESULTS—The primary end point, fatal and nonfatal myocardial infarction and stroke as well as other cardiovascular deaths, was markedly lower in the captopril than in the conventional therapy group (relative risk [RR] = 0.59; P = 0.018). Specifically, cardiovascular mortality, defined as fatal stroke and myocardial infarction, sudden death, and other cardiovascular death, tended to be lower in the captopril group (RR = 0.48; P = 0.084), and no difference was observed between the study groups for stroke (RR = 1.02; P = 0.96). Myocardial infarctions were less frequent in the captopril group than in the conventional therapy group (RR = 0.34; P = 0.002). Furthermore, total mortality was lower in the captopril as compared with the conventional therapy group (RR = 0.54; P = 0.034). Patients with impaired metabolic control seemed to benefit the most from ACE inhibitor–based therapy.

    CONCLUSIONS—Captopril is superior to a diuretic/β-blocker antihypertensive treatment regimen in preventing cardiovascular events in hypertensive diabetic patients, especially in those with metabolic decompensation.

    Cardiovascular complications represent the major burden of death and disease in patients with diabetes (1,2,3,4,5). Hypertension is a common finding in patients with type 2 diabetes; when present, it is an ominous sign that doubles the already elevated risk of future cardiovascular events (2,3). The results from the recent U.K. Prospective Diabetes Study (UKPDS) Hypertension study (6,7) and the Hypertension Optimal Treatment (HOT) trial (8) demonstrated that aggressive lowering of diastolic blood pressure in diabetic patients to levels <85 and <80 mmHg, respectively, was accompanied by reductions of macrovascular events by about one-third and one-half, respectively. In addition, results from the Systolic Hypertension in the Elderly Program (SHEP) (9) and Systolic Hypertension in Europe (Syst-Eur) trials (10) favored the aggressive antihypertensive treatment of diabetic patients with isolated systolic hypertension.

    The choice of antihypertensive agents in diabetic patients without renal impairment has been a matter of controversy. In particular, ACE inhibitors have favorable metabolic, renal, cardiovascular, and quality-of-life effects as compared with other regimens (11,12,13,14,15,16,17). Therefore, ACE inhibitors may be considered as first-line agents in hypertensive diabetic patients. The 1999 Guidelines for the Management of Hypertension, issued by the World Health Organization (WHO) and the International Society of Hypertension, recommended ACE inhibitors, along with diuretics, β-blocking agents, calcium-channel blockers, and angiotensin receptor II antagonists, as suitable first-line agents for diabetic patients (18). However, β-blocking agents or diuretics are preferred as initial therapy in the absence of renal disease according to the recommendations given by the Sixth Joint National Committee in the U.S. (19).

    What is the current evidence of ACE inhibitor use in essential hypertension treatment in diabetic patients? The results from the Appropriate Blood Pressure Control in Diabetes (ABCD) trial (20), comprised of 470 patients, showed higher incidence of fatal and nonfatal myocardial infarction in the group assigned to calcium-channel blocker (nisoldipine) treatment than in those assigned to an ACE inhibitor (enalapril). However, the results were based on only 30 incident cases; furthermore, myocardial infarction was a secondary end point in the trial. Similarly, the Fosinopril Versus Amlodipine Cardiovascular Events Randomized Trial (FACET) (21) showed lower incidence of acute myocardial infaction and stroke as well as a lower rate of hospitalization due to angina pectoris in the group randomized to fosinopril compared with the group assigned to amlodipine. The UKPDS–hypertension tight control arm also provided a comparison (7) between two regimens, captopril 25–50 mg twice daily and atenolol 50–100 mg daily. The incidence of diabetic complications was reduced to a similar extent in the two groups during follow-up, suggesting that blood pressure reduction in itself may be more important than the type of antihypertensive treatment used. However, compliance with the drug regimen in the UKPDS was superior in the captopril group compared with the atenolol group. In addition, the increase in body weight was approximately twice as high in subjects receiving β-blockers compared with patients receiving an ACE inhibitor.

    The Captopril Prevention Project (CAPPP) was the first intervention study (22,23) in hypertension to evaluate the effects of an ACE inhibitor–based therapeutic regimen on cardiovascular morbidity and mortality. In the present study, cardiovascular mortality and morbidity in the diabetic subpopulation is further analyzed.


    The CAPPP was conducted at 536 health centers in Sweden and Finland using the PROBE (Prospective Randomized Open Blinded Endpoint) study design (22,24), which has great similarity to routine clinical practice. Patients were treated and followed for an average of 6.1 years (23). Male and female patients with treated or untreated primary hypertension, aged 25–66 years, were included if their untreated diastolic blood pressure was ≥100 mmHg on two different occasions. Patients with secondary hypertension, serum creatinine >150 μmol/l, and indications requiring β-blockade treatment were excluded. Eligible patients were randomized to receive either a captopril-based regimen or conventional antihypertensive treatment (diuretics and/or β-blockers).

    The treatment goal was a supine diastolic blood pressure ≤90 mmHg. To reach this goal, the dose could be increased to 100 mg daily (once or twice daily) in the captopril-treated group; if necessary, a diuretic could be added. In the conventionally treated group, an optimal dose of a β-blocker or a diuretic was used. The investigators were free to choose the agents, and in a second step, the diuretic and the β-blocker could be combined. As a last step, a calcium-channel blocker could be added in both treatment groups.

    The primary end point was the combination of fatal and nonfatal myocardial infarction and stroke as well as other cardiovascular deaths. Events were evaluated by an independent and blinded end point committee before being accepted. Secondary end points were new or deteriorated ischemic heart disease (IHD) and congestive heart failure (CHF), atrial fibrillation, diabetes, transient ischemic attacks, and noncardiovascular deaths as previously defined (22,23).

    The diagnosis of diabetes was made according to WHO criteria (25) and required at least two abnormal fasting glucose values that were unequivocal (i.e., between 5.5 and 6.7 mmol/l). If they were not unequivocal, diagnosis was confirmed by an oral glucose tolerance test.

    The diabetic patients were, for the most part, treated with diet and/or oral hypoglycemic agents at baseline (Table 1), and thus they had type 2 diabetes. The CAPPP study was approved by ethics committees at participating centers in Sweden and Finland.


    Analysis was by intention to treat. Cox regression analyses were performed, with time since randomization as the nonparametrically modeled time variable; age at randomization was therefore introduced as a covariate. Further covariates used were sex and systolic blood pressure at randomization. Both systolic blood pressure at randomization and previously untreated patients as well as their product were introduced into the model (23). All computations were performed in the commercial software Stata. P < 0.05 (two-sided) was considered statistically significant.


    Among the 10,985 patients in the CAPPP, there were 309 patients with diabetes in the captopril group and 263 in the conventional therapy group. These 572 patients provided 3,445 patient-years, and only one of them was lost to follow-up. Blood pressure at baseline was similar in the captopril compared with the conventional therapy group (163.6/97.1 vs. 163.3/97.3 mmHg). The baseline characteristics were comparable between the two study groups (Table 1). Blood pressure levels during the study are shown in Fig. 1. If the in-study blood pressure for every diabetic patient is defined as the average of all blood pressure recordings from 12 months onwards, then these in-study blood pressures were somewhat higher in the captopril than in the conventional group (155.5/89.0 mmHg; P = 0.091 vs. 153.5/88.0 mmHg; P = 0.040).

    In diabetic patients, the incidence of the primary end point, fatal and nonfatal myocardial infarction and stroke as well as other cardiovascular deaths, was significantly lower in the captopril than in the conventional therapy group (relative risk [RR] = 0.59; P = 0.018) (Tables 2 and 3 and Fig. 2). When the events comprising the primary end point were analyzed separately, the following results were obtained: 1) cardiovascular mortality, defined as fatal stroke and myocardial infarction, sudden death, and other cardiovascular death, tended to be lower in the captopril group (RR = 0.48; P = 0.084) (Table 2); 2) fatal and nonfatal strokes did not differ between the study groups (RR = 1.02; P = 0.959) (Table 2); and 3) fatal and nonfatal myocardial infarctions were markedly less frequent in the captopril than in the conventional therapy group (RR = 0.34; P = 0.002) (Table 2).

    Analyses of secondary events showed that 1) total mortality was lower in the captopril than in the conventional therapy group (RR = 0.54; P = 0.034) (Table 2), and 2) all cardiac events (fatal and nonfatal myocardial infarction, other cardiovascular deaths, and sudden deaths as well as IHD, CHF, and atrial fibrillation) were lower in the captopril than in the conventional therapy group (RR = 0.67; P = 0.029) (Table 2).

    In addition, we investigated whether the observed benefits of ACE inhibitor–based therapy in diabetic patients could be related to baseline clinical and metabolic characteristics. Thus, the effects of captopril versus conventional treatment on the RRs of the primary end point were further analyzed in various subgroups. These were sex and dichotomized continuous variables, including BMI, fasting glucose, blood pressure, and serum cholesterol, as well as HDL cholesterol and serum triglycerides which were dichotomized at their median values (Table 3). The benefits of captopril therapy were more evident in male diabetic patients (P = 0.004) and in those with poor glycemic control (P = 0.033) and tended to be more pronounced in those with elevated serum cholesterol (P = 0.052) and lowered HDL cholesterol (P = 0.052). In these analyses, P values were adjusted for the effects of age, sex, previously untreated patients, and systolic blood pressure.

    The incidence of new-onset diabetes in an originally nondiabetic study population was 337 in the captopril group and 380 in the conventional group (RR 0.79; P = 0.007) (23).


    In regard to the effects on cardiovascular morbidity and mortality, the CAPPP study (23) was the first intervention trial in hypertensive patients to compare an ACE inhibitor–based therapy with conventional antihypertensive therapy that was based on diuretics and/or β-blockers. For the entire study population, diabetic and nondiabetic subjects combined, both regimens were equally effective in preventing the primary end point. However, the effects of the two regimens in the diabetic subpopulation showed a clear difference in the risk of developing a primary end point in favor of a captopril-based regimen. Importantly, captopril treatment in the diabetic subpopulation reduced fatal cardiovascular events by about one-half. However, as a subgroup analysis, the data should be interpreted with caution.

    At baseline, blood pressure levels and other characteristics were comparable between the captopril and conventional groups. During the study, the achieved blood pressure levels were, if anything, somewhat higher in the captopril than in the conventional regimen; therefore, the favorable outcomes are evidently not explained by differences in blood pressure reductions.

    At first sight, these findings may seem to deviate from those of the UKPDS (7), which showed no advantage for captopril over atenolol in reducing the risk of macro- and microvascular diabetic complications. The divergent findings may be explained to a great extent by the fact that the blood pressure treatment goal was lower in the UKPDS (<150/85 mmHg) than in the CAPPP study (<90 mmHg). These findings from the UKPDS suggest that blood pressure lowering may be more important per se than the choice of antihypertensive agent, although captopril was better tolerated by most patients (7). Because the effective lowering of blood pressure in the UKPDS was highly beneficial in diabetic patients, this confirms the observations in the HOT study in which the excess risk of having diabetes was almost nullified in patients whose diastolic blood pressure was reduced to ∼80 mmHg (8). Furthermore, the diabetic patients recruited in the UKPDS generally had milder disturbances in glucose metabolism (inclusion criteria: UKPDS, fasting plasma glucose >6.0 mmol/l; WHO, plasma glucose ≥7.8 mmol/l corresponding to blood glucose ≥6.7 mmol/l). In addition, patients with symptomatic cardiovascular disease were not included in the UKPDS; therefore, the patients in that study were probably at lower risk than the diabetic patients in the CAPPP study.

    The results from the analysis of diabetic patients in the CAPPP study further suggest that the choice of antihypertensive drug regimen may be of importance when there is metabolic decompensation, e.g., in those patients with poor glycemic control and perhaps in those with more unfavorable lipid abnormalities. Thus, these results support those of the Heart Outcomes Prevention Evaluation (HOPE) study (26,27), which showed that subjects with multiple risk factors are likely to benefit from ACE inhibitor–based therapy, because the protective effect is well beyond that obtained by blood pressure reduction per se. The favorable effects of ACE inhibitor–based therapy may be a result of the improvement of whole-body glucose uptake observed with captopril treatment during euglycemic clamp measurements (11). There are also a variety of other disturbances in diabetic patients that are improved by ACE inhibition, e.g., attenuation of sympathetic nerve activity (28), enhancement of nitric oxide formation (29), improvement of endothelial dysfunction (30), improvement in coronary reserve (31), and even the direct inhibition of the atherosclerosis process, as suggested by animal experiments (32).

    In conclusion, analysis of the outcome of the diabetic patient group in the CAPPP showed that the ACE inhibitor is an obvious first-choice drug for hypertensive patients with diabetes, especially when there is metabolic decompensation. However, because elevated blood pressure levels should be aggressively treated in diabetic patients, a multiple drug regimen is commonly required to achieve optimal treatment targets.


    CAPPP study investigators

    Steering committee: Lennart Hansson, Jan-Erik Björck, Björn Dahlöf, Ulf de Faire, Thomas Hedner, Bengt E. Karlberg, Jan Lanke, Lars H. Lindholm, Kimmo Luomanmäki, Claes Mörlin, Leo Niskanen, P. O. Wester. End point committee: Gottfried Härtel, Pentti Siltanen, Karl Swedberg, Andreas Terent. End point secretariat, Gothenburg: Anita Holmner. Safety committee: Leif Erhardt, Markku Kupari, Thomas Thulin, Pär-Ola Bendahl (statistician). Data management, Lund: Björn Slaug. Coordinators: Sweden: Per-Olof Andersson, Johan Asplund, Anders Dahlqvist, Björn Fagerberg, Göran Frithz, Bengt-Göran Hansson, Anders Hägg, Svend Aage Jonsen, Thomas Kahan, Bo Carlberg, Bo Erik Kristensson, Jörgen Kuylenstierna, Hans Larsson, Bengt Lindström, Ingrid Mattiasson, Bengt-Hjalmar Möller, Ove R. Nilsson, Håkan Stakeberg, Anders Svensson, Kerstin Tolagen, Leif Weiner, Brage Åström, Jan Östergren; Finland: Aapo Lehtonen, Mauno Lilja, Matti Nikkilä, Juhani Partanen; and all CAPPP investigators (23).

    Figure 1—

    Blood pressure levels in diabetic patients during the CAPPP study. ♦, Captopril treatment group; •, conventional treatment group.

    Figure 2—

    Comparison by intervention group of diabetic and nondiabetic patients who remained free from primary end point (fatal or nonfatal stroke, myocardial infarction, or other fatal cardiovascular events) in the CAPPP cohort.

    Table 1—

    Baseline characteristics of diabetic patients in CAPPP randomized to captopril or conventional treatment

    Table 2—

    Number of patients with events during follow-up and RR for diabetic patients in the captopril vs. conventional therapy groups

    Table 3—

    Effect of captopril treatment on the primary end points in diabetic patients by sex, baseline BMI, fasting glucose, systolic and diastolic blood pressure, and serum lipids


    This study was supported by a project grant from Bristol-Myers Squibb.


    • Address correspondence and reprint requests to Leo Niskanen, MD, Associate Professor, Department of Medicine, University of Kuopio, Box 1777, FIN-70211 Kuopio, Finland. E-mail: leo.niskanen{at}

      Received for publication 24 March 2001 and accepted in revised form 4 September 2001.

      The names of CAPPP investigators appear in an appendix at the end of this article.

      A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.


    | Table of Contents