Trial profile. BP indicates blood pressure; GITS, gastrointestinal therapeutic system; and SR, sustained release.
Mean daytime (6 AM-6 PM) systolic (A) and diastolic (B) blood pressure at randomization and during follow-up. At 2 months all patients were still receiving monotherapy with the randomized first-line medications; thereafter 1 or more additional drugs could be added to the regimen (see Table 2). n refers to the total number of patients in follow-up. GITS indicates gastrointestinal therapeutic system; SR, sustained release.
Proportion of patients whose blood pressure was controlled (mean daytime diastolic blood pressure [DBP] <90 mm Hg) at various stages of follow-up. For further explanation, see Figure 2. n refers to the total number of patients in follow-up. GITS indicates gastrointestinal therapeutic system; SR, sustained release.
Proportion of patients who continued receiving monotherapy with randomized first-line medication. The curves represent Kaplan-Meier estimates9 in which the denominator is the number of patients available for analysis at each time point. GITS indicates gastrointestinal therapeutic system; SR, sustained release.
Mean echocardiographic left ventricular mass index at baseline and during follow-up. At 4 and 13 months 131 (38.1%) of the 344 patients and 130 (50.6%) of the 257 patients, respectively, were receiving combination therapy. n refers to the number of patients in follow-up. GITS indicates gastrointestinal therapeutic system; and SR, sustained release.
Sareli P, Radevski IV, Valtchanova ZP, Libhaber E, Candy GP, Den Hond E, Libhaber C, Skudicky D, Wang JG, Staessen JA. Efficacy of Different Drug Classes Used to Initiate Antihypertensive Treatment in Black SubjectsResults of a Randomized Trial in Johannesburg, South Africa. Arch Intern Med. 2001;161(7):965-971. doi:10.1001/archinte.161.7.965
Thiazides are recommended to initiate antihypertensive drug treatment in black subjects.
To test the efficacy of this recommendation in a South African black cohort.
Men and women (N = 409), aged 18 to 70 years, with a mean ambulatory daytime diastolic blood pressure between 90 and 114 mm Hg, were randomized to 13 months of open-label treatment starting with the nifedipine gastrointestinal therapeutic system (30 mg/d, n = 233), sustained-release verapamil hydrochloride (240 mg/d, n = 58), hydrochlorothiazide (12.5 mg/d, n = 58), or enalapril maleate (10 mg/d, n = 60). If the target of reducing daytime diastolic blood pressure below 90 mm Hg was not attained, the first-line drugs were titrated up and after 2 months other medications were added to the regimen.
While receiving monotherapy (2 months, n = 366), the patients' systolic and diastolic decreases in daytime blood pressure averaged 22/14 mm Hg for nifedipine, 17/11 mm Hg for verapamil, 12/8 mm Hg for hydrochlorothiazide, and 5/3 mm Hg for enalapril. At 2 months the blood pressure of more patients treated with nifedipine was controlled: 133 (63.3%, P≤.03) vs 20 (39.9%) receiving verapamil, 21 (40.4%) receiving hydrochlorothiazide, and 11 (20.8%) receiving enalapril. At 13 months (n = 257), more patients (P<.001) continued receiving monotherapy with nifedipine (94/154 [61.0%]) or verapamil (22/35 [62.9%]) than hydrochlorothiazide (10/39 [25.6%]) or enalapril (1/29 [3.4%]). A sustained decrease of left ventricular mass (P<.001) with no between-group differences was achieved at 4 and 13 months.
In contrast to current recommendations, calcium channel blockers are more effective than thiazides as initial treatment in black subjects with hypertension. If treatment is started with thiazides or converting-enzyme inhibitors, combination therapy is more likely to be required to control blood pressure and reduce left ventricular mass.
SIMILAR TO the recommendations of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure,1 the South African guidelines for antihypertensive therapy2 recommend the use of low-dose thiazides as the preferred first-line drug, to which calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, β-blockers, or reserpine may be added to achieve blood pressure (BP) control. Although many patients with mild to moderate hypertension require more than 1 drug to reach normal BP,3 it remains unclear what proportion of black patients with hypertension can be controlled in the long-term if thiazides or other drug classes are used to initiate treatment.
The Baragwanath Hypertension Ambulatory Monitoring Study was a randomized open-label trial conducted in black patients with hypertension. Its objective was to compare the BP-lowering effect of a thiazide, the recommended first-line treatment,1,2 with those of calcium channel blockers and an ACE inhibitor. To exclude patients with white-coat hypertension and to avoid observer bias4 in an open-label trial, ambulatory BP monitoring was used for the selection of patients and the adjustment of treatment. The patients were followed up for 13 months, so that the long-term efficacy of monotherapy could be evaluated as well as the need for additional medications to achieve lasting BP control. We also assessed echocardiographic left ventricular mass as an index of the efficacy of antihypertensive treatment.
This trial was a randomized, open-label study conducted at the Chris Hani-Baragwanath Hospital, Johannesburg, South Africa from September 1, 1994, through November 30, 1997. The protocol was approved by the Committee for Research on Human Subjects of the University of the Witwatersrand. Black men and women could be enrolled, if they were between the ages of 18 and 70 years and were free of clinically significant cardiovascular or noncardiovascular disorders. Women of reproductive age had to use adequate contraception. All patients gave informed written consent.
To screen patients for hypertension, we measured their sitting diastolic BP 10 times consecutively at 3-minute intervals, using calibrated oscillometric monitors5 (Dinamap 1846 SX; Critikon Inc, Tampa, Fla). If the mean diastolic BP was 90 mm Hg or higher, the patients underwent 24-hour ambulatory BP monitoring. Oscillometric devices6 (SpaceLabs 90207; SpaceLabs Inc, Redmond, Wash) were programmed to obtain readings every 15 minutes from 6 AM to 6 PM. Patients whose mean daytime diastolic BP was 90 mm Hg or higher then proceeded to a 2-week placebo run-in period, after which the ambulatory recording was repeated. If on the second set of measurements the mean daytime diastolic BP ranged from 90 to 114 mm Hg and if the count of the returned placebo tablets was within 80% to 120% of the expected number, the patients qualified for enrollment in the study.
Eligible patients were randomized to the nifedipine gastrointestinal therapeutic system (GITS), 30 mg/d, or 1 of 3 available reference treatments starting with verapamil hydrochloride sustained release (SR), 240 mg/d; hydrochlorothiazide, 12.5 mg/d; or enalapril maleate, 10 mg/d; as indicated in the current guidelines.1 Patients were followed up at monthly intervals for 13 months with the goal being to lower the daytime diastolic BP below 90 mm Hg. The ambulatory BP recordings were systematically repeated at monthly intervals for 4 months and at the final 13-month visit. From the 4-month visit onward, only the patients whose BP remained uncontrolled at a previous visit underwent ambulatory BP monitoring. At baseline and at 4 and 13 months, M-mode and 2-dimensional echcardiograms (Sonos 2500; Hewlett Packard, Andover, Mass) were obtained using a 2.5-MHz transducer. The echocardiograms were analyzed according to the American Society of Echocardiography convention.7 Left ventricular mass index was derived according to an anatomically validated regression method that corrects left ventricular mass estimates obtained from the recommended measurements.8
If at the first monthly follow-up visit the target BP was not attained, the daily dose of the first-line drug was increased as follows: nifedipine GITS to 60 mg/d, verapamil hydrochloride SR to 360 mg/d, hydrochlorothiazide to 25 mg/d, and enalapril maleate to 20 mg/d. If at 2 months the target BP was not attained, the patients in the nifedipine GITS group were randomized to 1 of the following 4 treatment strategies: the addition of enalapril maleate (10-20 mg/d), carvedilol (25 mg/d), or verapamil hydrochloride SR (120-240 mg/d), or the increase of the daily dose of nifedipine GITS to 90 mg. In the uncontrolled patients of the verapamil SR group, the daily dose of the calcium channel blocker could be increased to 480 mg, and subsequently hydrochlorothiazide, 12.5 mg/d, could be added to the regimen. Patients whose BP was uncontrolled while taking hydrochlorothiazide, 25 mg/d, received reserpine, 0.125 mg/d, and those whose BP was uncontrolled while taking enalapril maleate, 20 mg/d, received hydrochlorothiazide, 12.5 mg/d. Patients whose BP was still uncontrolled at 4 months could receive any study drug in combination with the treatment already instituted. Patients were withdrawn from the study if at any follow-up visit after 4 months while receiving combination therapy, the mean 24-hour systolic BP exceeded 200 mm Hg, or the mean daytime diastolic BP exceeded 114 mm Hg, or they experienced any serious adverse event.
Database management and statistical analysis were performed with SAS software, version 6.12 (SAS Institute Inc, Cary, NC). Between-group and within-group differences in continuous measurements were tested with repeated measures analysis of variance and the Scheffé test for multiple comparisons. Proportions were compared using the Fisher exact test with Bonferroni correction for multiple comparisons. Longitudinal changes in treatment status were evaluated by Kaplan-Meier survival function estimates9 and the log rank test.
As shown in Figure 1, of 591 patients enrolled in the placebo run-in phase, 409 patients were randomized. At baseline no statistically significant differences were noted between the patients in the 4 treatment groups (Table 1). The analysis included 366 patients at 2 months, 344 at 4 months, and 257 at 13 months (Table 2). The women in particular had high body mass indices (BMIs) (32.4 [SE, 0.4] kg/m2) compared with the men (26.9 [SE, 0.5] kg/m2, P<.001).
At 2 months, monotherapy with nifedipine GITS had lowered the daytime BP more than monotherapy with hydrochlorothiazide or enalapril (Figure 2). The mean systolic and diastolic decreases in the daytime BP at 2 months for the various treatment groups were as follows: 22/14 mm Hg for nifedipine GITS, 17/11 mm Hg for verapamil SR, 12/8 mm Hg for hydrochlorothiazide, and 5/3 mm Hg for enalapril. At 2 months (Figure 3) the BP of significantly more patients treated with nifedipine GITS was controlled: 133 (63.3%, P≤.03) vs 20 (39.2%) receiving verapamil SR, vs 21 (40.4%) receiving hydrochlorothiazide, and vs 11 (20.8%) receiving enalapril; BP control while receiving monotherapy was not different among patients randomized to verapamil SR, hydrochlorothiazide, or enalapril (Figure 3). At 2 months the decrease in the mean (SE) daytime heart rate in the verapamil SR–treated group (−5.6 [1.3] /min) was significantly different (P = .02) from the heart rate changes for those receiving nifedipine GITS (+2.7 [0.6] /min), hydrochlorothiazide (−0.5 [1.0] /min), or enalapril (+0.8 [1.2] /min). The heart rate changes among the latter 3 groups were not significantly different. There was no difference in control rates while receiving monotherapy between patients with a higher BMI (>25 kg/m2) and those with a BMI of 25 kg/m2 or less throughout the trial period.
At 4 months fewer patients (P<.004) in the nifedipine GITS–treated and verapamil SR–treated groups had required the addition of second-line medication than did those in the hydrochlorothiazide-treated and enalapril-treated groups (Table 2). Similarly, at 13 months (Table 2 and Figure 4), more patients (P<.001) continued to receive monotherapy with nifedipine GITS (94/154 [61.0%]) or verapamil SR (22/35 [62.9%]) than with hydrochlorothiazide (10/39 [25.6%]) or enalapril (1/29 [3.4%]). Because many patients proceeded to combination therapy (Table 2), the BP differences observed on monotherapy at 2 months disappeared at 13 months (Figure 2).
Within 4 months of randomization (Figure 5), left ventricular mass index had decreased by 15% (P<.001) from 118 (34) g/m2 (mean [SD]) to 100 (25) g/m2 with no further decline between 4 and 13 months when the proportions of patients receiving combination therapy were 131 (38.1%) of the 344 patients and 130 (50.5%) of the 257 patients, respectively. There were no between-group differences in the changes in echocardiographic left ventricular mass (Figure 5, P>.43).
Uncontrolled BP lead to the withdrawal of 6 patients in the enalapril-treated group (11.3%), 2 patients in the nifedipine GITS–treated group (0.95%, P<.001 vs enalapril), and no patients in either of the other 2 groups. The major adverse events were myocardial infarction (1 patient) and unstable angina pectoris (1 patient) in the nifedipine GITS–treated group; fatal pneumonia (1 patient) and second-degree atrioventricular block complicated with left ventricular failure (1 patient) in the verapamil SR–treated group; fatal bowel obstruction (1 patient) and grand mal epilepsy (1 patient) in the hydrochlorothiazide-treated group; and angioneurotic edema (5 patients) in the enalapril-treated group. No patient experienced gastrointestinal bleeding.
At 2 months the rate of BP control was higher for the patients receiving nifedipine GITS treatment than for those patients receiving verapamil SR, hydrochlorothiazide, or enalapril. Monotherapy with nifedipine GITS reduced BP significantly more than hydrochlorothiazide and enalapril. At 13 months more patients continued receiving monotherapy with nifedipine GITS or verapamil SR than hydrochlorothiazide or enalapril.
This study must be interpreted within the context of its limitations. Of 409 randomized patients, the numbers lost to follow-up were 32 (7.8%) at 2 months, 45 (11.0%) at 4 months, and 115 (28.1%) at 13 months. The patients enrolled in the trial were recruited from the Soweto district, an urban black community burdened by unemployment (52.9%), shortage of housing, deprivation, social divide, and broken households. These conditions explain why people frequently relocated to find new jobs and why many of our patients withdrew their consent or dropped out of the study. However, the black inhabitants of Soweto are representative of many modern African cities, characterized by a high prevalence and incidence of hypertension.10,11 The patients characteristically also had a high BMI that may be explained by inactivity or by their high-energy, low-protein maize-based diet. Our study had an open-label design, but we used 24-hour ambulatory BP monitoring to determine objectively the eligibility of the patients and to evaluate the effects of treatment. Indeed, ambulatory BP measurements are characterized by high reproducibility, are not subject to digit preference and observer bias, and avoid the transient rise of a patient's BP in response to the clinic surroundings or the presence of the observer, the so-called white-coat effect.4
To the best of our knowledge, no other study used ambulatory BP monitoring on a large scale in black African patients with hypertension. In a previous single-blind randomized trial (n = 41), we demonstrated that over a 12-week period treatment twice with nifedipine (20-40 mg twice daily) reduced the 24-hour systolic and diastolic BP by 28/17 mm Hg, whereas treatment with captopril (50 mg twice daily) did not change BP (+2/+1 mm Hg).12 Materson and coworkers13,14 randomized 1292 American men with a clinic diastolic BP of 95 to 109 mm Hg to placebo or 1 of the following 6 drugs: hydrochlorothiazide (12.5-50 mg/d), diltiazem SR (120-360 mg/d), atenolol (25-100 mg/d), clonidine (0.2-0.6 mg/d), captopril (25-100 mg/d), or prazosin (4-20 mg/d). This trial included 291 black patients,13,14 aged 21 to 60 years, in whom the nondihydropyridine calcium channel blocker diltiazem after 1 year achieved the best control rate (56.8%),13 defined as a clinic diastolic BP of less than 90 mm Hg. In agreement with our 2 months' observations, hydrochlorothiazide treatment titrated up from 12.5 to 50 mg/d controlled BP in 41.7% of the black patients.14
In keeping with other studies,3,15 most of our black patients with mild to moderate hypertension required multiple drug treatment to achieve adequate control, especially if treatment was initiated with a thiazide or an ACE inhibitor. Low rates of BP control have been previously documented in black patients with hypertension who were receiving treatment with captopril,12 enalapril,16,17 or hydrochlorothiazide.18 On balance, the evidence available suggests that ACE inhibitors should not be used as monotherapy in black patients with hypertension, but only in combination with other medications.14 At 13 months, the addition of hydrochlorothiazide and/or other drugs (nifedipine GITS in 15 of 29 patients) achieved BP control in 79.3% of patients randomized to enalapril treatment. Furthermore, at the end of the study, the addition of reserpine and/or enalapril to hydrochlorothiazide treatment resulted in a control of 66.7%.
Two research groups19,20 have suggested that to reduce left ventricular hypertrophy ACE inhibitors would be most efficient. However, a third meta-analyst21 considered only prospective, randomized, and properly controlled comparative studies. This analysis revealed that thiazides, β-blockers, calcium channel blockers,and ACE inhibitors reduced left ventricular mass to the same degree as the 3 other classes statistically combined, and that ACE inhibitors were not superior to calcium channel blockers.21 Our study showed a significant decrease in left ventricular mass index at 4 months paralleling the decline in BP, with no differences between the treatment groups. This observation suggests that BP control rather than drug class is the predominant factor determining left ventricular mass regression.
In contrast with the modified South African guidelines,1,2 long-acting calcium channel blockers are more effective than thiazides as initial treatment in African blacks with hypertension. Outcome trials in hypertension were largely conducted in white22,23 and Asian subjects.24- 26 Because such information is not yet available for African blacks, the present findings should be subject to further investigation in prospective morbidity and mortality trials in African blacks.
Accepted for publication November 7, 2000.
The Baragwanath Hypertension Ambulatory Monitoring Study was supported by Bayer (Pty) Ltd, Johannesburg.
We gratefully acknowledge the expert assistance of Elizabeth Tshele, RN, and Margaret Hlatswayo, RN. William H. Birkenhäger, MD, Erasmus University, Rotterdam, the Netherlands, provided helpful comments on the manuscript.
Corresponding author: Pinhas Sareli, MD, Department of Cardiology, Chris Hani-Baragwanath Hospital, University of the Witwatersrand, PO Bertsham, Johannesburg 2013, South Africa (e-mail: firstname.lastname@example.org).