Comparison of Prazosin with Hydralazine in Patients Receiving Hydrochlorothiazide A Randomized, Double-blind Clinical Trial VETERANS ADMINISTRATION COOPERATIVE STUDY GROUP ON ANTIHYPERTENSIVE AGENTS SUMMARY The antihypertensive efficacy and the incidence of side effects of prazosin and hydralazine were compared in a randomized, double-blind trial in 232 adult male hypertensives who could not be controlled with hydrochlorothiazide alone. There were no signifiant differences between regimens in the percentage of patients who attained goal blood pressure (reduction of diastolic blood pressure to below 90 mm Hg and at least 5 mm less than the baseline randomization pressure), effect on pulse rate or the incidence or reasons for termina- tions. Absolute reduction of blood pressure was similar for both drugs except for sitting systolic pressure at 3 and 6 months, when prazosin effected a 3.7- and 3.6-mm Hg greater response (p < 0.05). Orthostatic dizziness (p < O.OOS), sexual dysfunction @ < 0.02), and nightmares @ < 0.02) were more frequent with prazosin than with hydralazine; nevertheless, patient compliance was similar for both drugs. An unexpected finding was the lack of pulse rate increase associated with hydralazine, particularly in older patients. HYDRALAZINE AND PRAZOSIN are generally considered as alternative choices within the same levels of stepped-care management of hypertension. The successful combination of either of these drugs with diuretic and @-adrenergic inhibitory compounds has provided a rationale for more widespread use.`-lo The antihypertensive mechanisms of hydralazine are not well established.' Although prazosin inhibits phosphodiesterase, l1 it is not known if this action is clinically important. The antihypertensive properties of the medication are attributable primarily to blockade of postsynaptic a-adrenergic receptors (ai) of the vascular smooth muscle.Dv lo Both medications decrease total peripheral resistance and increase vas- cular cyclic AMP.`, s-11 Neither one manifests central, vagal, @-adrenergic receptor, neuronal or ganglionic blocking activity. l* g* lo Prazosin differs from hydrala- zine in that it dilates capacitance as well as resistance vessels,1z and by an apparent lack of induction of marked tachycardia and hyperreninemia.B* lo* la, l4 No large scale, double-blind systematic studies have compared the antihypertensive efficacy of hydrala- zine and prazosin. The available data, however, suggest that while the drugs are of similar potency, the incidence of side effects, some of them necessitating withdrawal, may be more pronounced with hydral- azine.`6-11 The purpose of this study was to compare the anti- hypertensive efficacy and the incidence of side effects of prazosin and hydralazine through a randomized, double-blind clinical trial in patients whose blood From participating Veterans Administration Medical Centers in Allen Park, Michigan, Jackson, Mississippi, Memphis, Tennessee, Miami, Florida, San Juan, Puerto Rico, and Washington, DC Supported by the Cooperative Studies Program of the Veterans Administration Research Service. Address for correspondence: Eli A. Ramirez, M.D., Veterans Administration Hospital, G.P.O. Box 4867, San Juan, Puerto Rico 00936. Received August 4, 1980; revision accepted January 21, 1981. Circulation 64, No. 4, 1981. pressure was not successfully controlled with hydro- chlorothiazide alone. Methods Objectives of the Study The study was designed to determine (1) the anti- hypertensive efficacy of prazosin compared with hydralazine (both given in addition to hydrochloro- thiazide) on the basis of the percentage of patients in each group who at the 5-month and 6-month postran- domization visits attained an average reduction of sit- ting diastolic pressure to below 90 mm Hg and at least 5 mm less than the baseline randomization pressure, and the mean changes in blood pressure between the randomization visit (hydrochlorothiazide alone) and the average of the 5-month and 6-month post- randomization visits (hydrochlorothiazide plus either prazosin or hydralazine); (2) the acceptability of both regimens over a 6-month experience based on the in- cidence of toxic reactions and side effects; and (3) the degree of chronotropic effect upon the heart of both regimens as measured by the change in pulse rate at the same levels of blood pressure response. Selection of Patients Male patients, 21-74 years of age, whose average diastolic pressure at two successive clinic visits was 95-114 mm Hg, were recruited from the admitting room, outpatient clinics and hospital. Patients were excluded from the study if they had severe complica- tions of hypertension, serious systemic disease or con- ditions that would contraindicate the drug regimens used. A complete list of exclusions is presented in appendix A. Patients receiving antihypertensive therapy were allowed to enter the study, provided the diastolic blood pressure was 95-114 mm Hg after medication was discontinued for 4 weeks. The blood pressures were measured at all visits after the patients lay undisturbed in a quiet room for lo-15 minutes. Three readings were taken in the right arm in 772 PRAZOSIN VS HYDRALAZINE TRIAL/VA Study Group 173 the sitting position at l-minute intervals using a stan- dard mercury sphygmomanometer. The average of these three readings was recorded, followed by the pulse rate counted for 1 minute. A large cuff was used when the circumference of the arm exceeded 32 cm. On the visit when hydrochlorothiazide was started, the randomization visit, the 4- or 6-week postrandomiza- tion visit and the last study visit, the blood pressure average was similarly determined in the supine posi- tion and in the erect position after standing 2 minutes, followed by counting the pulse for 1 minute. The measurements used throughout the study to determine therapeutic decisions and end points were the average of three diastolic fifth-phase (Korotkoff) readings in the right arm in the sitting position. Prerandomization Trial Period The nature of the study was explained to the patient and a signed informed consent was obtained.* A his- tory and physical examination were performed. Lab- oratory studies included chest roentgenogram, ECG, CBC, urinalysis, fasting serum sugar, SGOT, alkaline phosphatase, and serum determinations of potassium, uric acid, cholesterol, triglycerides, creatinine and antinuclear antibody. At all visits, a checklist of the known or suspected side effects of the drugs was reviewed with the patient. If the patient's diastolic blood pressure was in the range of 95-l 14 mm Hg and there were no exclusion factors, he was given placebo capsules and was in- structed to take them three times daily. He was further instructed to return the remaining capsules at each clinic visit, and pill counts were made to test compliance. Compliance was recorded as satisfactory if the patient took from 10% less to 10 pills more than the exact prescribed number of each medication. The patients were allowed a maximum of four biweekly visits to qualify for hydrochlorothiazide treatment or were dropped from the study. Within this "placebo period," if the diastolic blood pressure averaged 95-l 14 mm Hg without a pill count violation on two successive clinic visits, the patient was placed on hydrochlorothiazide, 25 mg three times daily, and continued on placebo. Two weeks later, if the average of three diastolic blood pressure readings was 90-l 14 mm Hg and the pulse rate below 95 beats/min without pill count violation, the patient was randomized; otherwise he was excluded from the study. The average pressure of the randomization visit was designated as the baseline pressure for the patient. Postrandomization Period In substitution of the placebo look-alike capsules, patients were randomly assigned in a double-blind fashion to one capsule three times daily of "Zarpine," the code name for the capsules that contained either prazosin or hydralazine. Hydrochlorothiazide was *The study was approved by the Human Use Committee at each hospital and conformed to the Helsinki declaration. continued at the same dose. Zarpine No. 1 contained either 1 mg of prazosin or 10 mg of hydralazine, Zar- pine No. 2 either 2 mg of prazosin or 25 mg of hydral- azine, and Zarpine No. 3 either 5 mg of prazosin or 50 mg of hydralazine. The patients were seen biweekly for the first four visits after randomization and then every 4 weeks for the final four visits of the postrandomization period. If the .pill count indicated that the patient had taken at least 80% of the expected dose, the dose of Zarpine was increased from No. 1 to No. 2 and then to No. 3, with the purpose of achieving goal blood pressure, defined as a sitting diastolic pressure less than 90 mm Hg and 5 mm Hg below the baseline pressure for the patient. If hypotensive symptoms developed or if the heart rate was greater than 99 beats/min, the strength of the Zarpine capsules was decreased to the next lower strength, or if on the weakest strength, decreased to twice daily, to once daily and then dis- continued if symptoms persisted. Laboratory tests performed at prerandomization were repeated at 4, 12 and 24 weeks after randomiza- tion. ECGs were obtained at the initial study visit, at randomization and 12 and 24 weeks later. Patients were terminated from the study and placed on appropriate antihypertensive treatment if they de- veloped severe complications of hypertension, condi- tions that would interdict the drug regimens used, in- adequate control of hypertension, or failed to comply with clinic appointments without proper excuse. Ap- pendix B is a complete list of reasons for termination. Statistical Methods The chi-square test was used for comparison of the discrete variables between the two drug regimens, the paired t test for individual changes in continuous variables, and the two-sample t test for comparison of continuous variables between the two regimens. Results Three hundred ninety-two patients were entered into the study at the six participating hospitals; 232 were eventually properly randomized, 111 to prazosin and 121 to hydralazine. Prerandomization Losses One hundred sixty patients were excluded before randomization for reasons specified in the protocol. Fifty-five (34%) were excluded because either blood pressure or pulse rate was out of the acceptable range; 21 of these responded to hydrochlorothiazide alone with a diastolic blood pressure below the acceptable lower limit of 90 mm Hg. Seventy-one (44%) were ex- cluded because they were noncompliant, 13 (8%) at their own request, and the remaining 2 1 (13%) for mis- cellaneous reasons. Baseline Data The baseline data at randomization of the 232 patients randomized and of the 198 patients who com- 774 CIRCULATION TABLE 1. Baseline Data at Randomization AI1 randomized Prazosin Hydralazine n 111 121 Sitting pulse (beats/min) 77.5 780.5 Standing pulse (beatslmin) 82.6 *84.9 Sitting DP (mm Hg) 100.2 98.9 Standing DP (mm Hg) 100.5 99.5 Sitting SP (mm Hg) 137.3 137.6 Standing SP (mm Hg) 135.1 137.3 Weight (kg) 84.1 86.3 Serum K (mEq/`I) 3.76 3.72 Uric acid (mg/dl) 8.05 7.91 Creatinine (mg/dl) 1.20 1.14 Cholesterol (mg/dl) 230.5 228.8 Glucose (mg/dl) 110.0 112.8 Triglycerides (mg/dl) 177.4 183.1 Age (ye-9 50.7 52.4 Race White (%) 50 60 Black (%) 50 40 Significance between regimens: *p < 0.1. tp < 0.01. Abbreviations: DP = diastolic pressure; SP = systolic pressure. VOL 64, No 4, OCTOBER 1981 Completed study Prazosin Hydralazine 92 106 77.5 79.9 82.1 84.2 99.6 98.7 100.0 99.4 137.6 137.3 135.3 137.3 83.5 85.3 3.75 3.72 8.12 7.84 1.19 1.14 228.9 222.7 107.7 112.9 177.0 178.9 51.5 52.1 50 59 50 41 pleted 6 months of treatment are shown in table 1. At randomization, the sitting pulse was significantly different between both groups (p < O.Ol), while it was borderline (p = 0.06) for the standing pulse. None of the other differences in this table are significant. Postrandomization Losses Thirty-four patients were lost after randomization, 19 were receiving prazosin and 15 hydralazine. The principal reasons for postrandomization loss were failure to return to the clinic in 13 instances and at the patient's otin request in five. Other causes were angina pectoris in four patients (three of whom were taking hydralazine), blood pressure exceeding protocol criteria in three patients using prazosin, and mis- cellaneous reasons in nine patients. Only one patient was dropped for rapid heart rate; he was receiving prazosin. The records of the 13 patients who failed to return to the clinic and of the five who requested to be dis- continued from the study were examined in detail retrospectively. Eleven had been randomized to prazosin and seven to hydralazine. Eleven were lost within the first 2 months after randomization, five of whom had only one or no postrandomization visit. No drug-related reason could account for discontinuing treatment except for one patient, who had discon- tinued prazosin for 1 week to undergo a prostatec- tomy. Upon resumption of the same dosage, he had an episode of syncope and thereafter refused to take the drug. This was the only "first-dose" effect observed during the study. Including this patient, two patients were lost because of dizziness and both were on prazosin. Effects on Blood Pressure The antihypertensive effects of the two drug regimens in the patients who completed 6 months on treatment are shown in table 2. The l- and 3-month blood pressure figures correspond to the readings of those clinic visits, but the &month figures represent the average of the 5- and 6-month clinic readings. The greatest difference between regimens in attainment of goal blood pressure was at 6 months, when it was 44.6% of prazosin-treated patients and 39.690 of hydralazine-treated patients. Even then, the difference was not significant (p > 0.05). The mean reductions in sitting systolic and diastolic pressures after starting prazosin were 6.7/6.7 mm Hg at 1 month, 9.6/9.7 mm Hg at 3 months, and 8.7/8.9 mm Hg at 6 months. For hydralazine, the corresponding figures were 5.0/6.4 mm Hg at 1 month, 5.9/8.8 mm Hg at 3 months, and 5.1/8.2 mm Hg at 6 months. These are significant decrements from baseline pressures for both drugs at all periods @ < 0.001); however, between drugs, the only significant differences are in systolic pressure at the third and sixth months in favor of prazosin (p < 0.05). For standing blood pressures, all reduc- tions from baseline are also significant for both drugs (p < O.OOl), but between drugs, the only difference that approaches significance is the systolic blood pressure at 1 month, which is in favor of prazosin cp = 0.055). The racial composition of the patient samples at PRAZOSIN VS HYDRALAZINE TRIAL/VA study Group 175 TABLE 2. Blood Pressure Effects Prazosin Hydralazine n 92 Attained goal blood pressure 1 month 34.8% 3 months 48.9% 6 months 44.6% Mean blood pressure reduction from baseline (mm Hg)* Sitting 1 month: Systolic 6.7 zk 1.2 Diastolic 6.7 f 0.7 3 months: 106 32.1% 47.2% 39.6% 5.0 * 1.2 6.4 f 0.7 Systolic 9.6 zt 1.3 : 5.9 + 1.2 Diastolic 9.7 3~ 0.8 8.8 3~ 0.8 6 months: Systolic 8.7 f 1.2 1 5.1 f 1.0 Diastolic 8.9 f 0.7 8.2 f 0.7 Standing 1 month: n 86 97 Systolic 7.7f1.2 t 4.3f1.3 Diastolic 7.8 f 0.8 6.4 f 0.7 6 months: n 91 104 Systolic 10.4 + 1.4 7.3 * 1.3 Diastolic 9.5 * 0.9 8.3 k 0.8 *Values are mean k SEM. Significance between regimens: tp < 0.1. $p < 0.05. randomization @ = 0.09) and upon completion of the study (p = 0.18) was not significantly different between regimens (table 1). Nevertheless, because of the possibility that the relatively higher proportion of blacks in the prazosin-treated patients may have affected blood pressure response adversely in this group, an analysis of blood pressure distribution and responses by race was performed. There was no significant difference in the baseline average diastolic blood pressure between whites and blacks of both groups. Furthermore, the percentage of blacks who at- tained goal blood pressure (41.3%) with prazosin was almost the same as for those who received hydrala- zine (41.9%). Therefore, there is no basis to suspect that the racial composition of the samples may have influenced the blood pressure response. Drug Dosage The patients at each dosage level were grouped ac- cording to whether they reached goal blood pressure at 6 months (table 3). The average dose at 6 months for all prazosin-treated patients was 10.6 mg/day; for the hydralazine-treated patients the corresponding average dose was 115.6 mg/day. The average daily dose of the patients on prazosin who reached goal blood pressure was 8.5 mg/day, compared with 12.5 mg/day for those who did not. In the case of hydralazine, the average daily dose for those patients who reached goal blood pressure was 94.3 mg/day, compared with 129.5 mg/day for those who did not. For both drugs a larger proportion of patients received the highest dosage among those who did not reach goal blood pressure than among those who did, reflect- ing the effort made in the clinics to achieve goal blood pressure. However, 12 patients who received prazosin and 13 who received hydralazine, although they did not attain goal blood pressure, were not at the max- imal dosage at the end of the study. Twelve had at- tained goal pressure previously, but a blood pressure reading higher than usual nudged the 5-6-month average over the goal; in nine the clinic was reluctant to increase the dosage because of pill count violations, and in the four others there were a considerable number of side effects. The reasons for failure to in- crease dosage were not noticeably different between the drugs. Side Effects The analyses of side effects were done both by add- ing elicited and volunteered side effects and by assessing them separately for each category, but there was no remarkable difference between these two ap- proaches. Table 4 shows the percentage incidence of the sum of elicited and/or volunteered post- randomization side effects. The data were analyzed counting all patients with postrandomization side effects and also excluding those who had the concerned side effect before ran- domization. The average percentage of clinic visits at which side effects were noted and the number of side effects manifested at two consecutive clinic visits were also analyzed. Orthostatic dizziness (p < O.OOS), nightmares (JJ < 0.02), sexual dysfunction (p < 0.02), and possibly lethargy (p < 0.08) were more frequent with prazosin than with hydralazine on one or more of these analyses. The incidence of these side effects on a month-by- month basis is shown in table 5. Patients treated with prazosin had a significantly higher incidence of side effects than those treated with hydralazine during the first month, but between the first and third months, only the incidence of orthostatic dizziness remained significantly higher. Between the third and sixth months, more patients continued to complain of these side effects with prazosin than with hydralazine, but the differences were less notable. Despite the differences in side effects, patient com- pliance, determined according to pill counts, was similar for both drugs at 1, 3 and 6 months. For prazosin, 66.3% of the patients were compliant at 1 month, 68.5% at 3 months and 68.5% at 6 months. For hydralazine, 76.4% were compliant at 1 month, 69.8% at 3 months, and 74.5% at 6 months. Efiect on Pulse Rate Pulse rates during the study are shown in table 6. As shown in table 1 the sitting pulse rate was significantly 776 CIRCULATION VOL 64, No 4, OCTOBER 1981 TABLE 3. Number of Patients by Dosage and Goal Attainment at 6 Months Prazosin Hydralazine Goal blood pressure Goal blood pressure mg/day Total Yes No mg/day Total Yes No 1 3 2 1 10 1 0 1 2 2 2 0 20 3 3 0 3 15 11 4 30 16 10 6 6 17 10 7 75 16 10 6 15 55 16 39 150 70 19 51 Total 92 41 (44.6%) 51 Total 106 42 (39.6%) 64 different between the groups before taking hydro- chlorothiazide; for the patients eventually random- ized to prazosin, the average pre-hydrochlorothiazide pulse was 74.5 beats/min and for patients eventually randomized to hydralazine it was 77.4 beats/min (p < 0.05). This difference was present before ran- domization, so is attributable to chance variation. Pulse rate also increased significantly during treat- ment with hydrochlorothiazide, by 3 beats/min in the group later randomized to prazosin @I < 0.05) and by 2.5 beats/min in the group later randomized to hydralazine (p < 0.05). However, after randomiza- tion, the changes in average pulse rates were not significant for either drug or between drugs at any period. On the assumption that pulse rate changes would be more noticeable in the standing position, the pertinent data were analyzed. Standing pulses were available for the periods indicated on the numbers of patients shown in parentheses in the table. There was an in- crease in standing pulse rate 1 month after randomiza- tion in patients who received prazosin (p < 0.05), but the other differences were not significant within or between regimens. Other Changes No significant differences were noted between regimens in serum potassium, uric acid, creatinine, cholesterol, sugar and triglycerides from randomiza- tion as compared to 6 months after randomization. However, body weight increased an average of 0.5 kg in the group treated with prazosin and decreased an average of 0.6 kg in the group treated with hydrala- zine 0, = 0.009). This difference might have been in- fluenced by regression toward the mean, because although there was no significant difference between the initial average weights of both groups, the figure was higher for the hydralazine-treated group than for the prazosin-treated group (table 1). TABLE 4. Percentage Incidence of Postrandomization Patient Side Effects Elicited or Volunteered By patients, By patients, By avg. % all postrandom. new only of visits Side effects P H P H P H At 2 consec. visits P H Anorexia Weakness Orthostatic dizziness Lethargy Headaches Dyspnea on effort Angina Palpitations Skin rash Arthritis Sexual dysfunction Depression Nightmares Ulcer symptoms Other 5.4 4.7 2.3 4.8 34.8 34.9 31.3 32.3 47.8 f. 26.4 37.8 t 22.4 37.0 * 25.5 31.7 20.8 40.2 39.6 31.9 31.2 47.8 41.5 39.0 32.9 12.0 8.5 9.0 5.9 28.3 27.4 17.7 18.5 8.7 10.4 6.7 6.1 25.0 18.9 20.0 17.0 32.6 t 19.8 27.7 17.8 14.1 9.4 10.2 7.7 19.6 7 7.5 17.8 t 6.8 14.1 11.3 12.4 10.0 53.3 53.8 46.7 44.4 1.7 0.9 5.4 * 1.0 12.9 9.1 16.3 12.3 17.7 1 8.1 22.8 $ 9.4 14.0 * 8.1 16.3 8.5 12.2 12.2 13.0 16.0 17.5 15.0 21.7 22.6 2.3 4.1 2.2 4.7 9.5 11.9 14.1 15.1 2.7 3.9 3.3 5.7 9.5 6.4 9.8 9.4 13.9 t 5.7 16.3 * 8.5 5.2 3.0 8.7 3.8 5.1 2.3 8.7 * 2.8 2.0 3.4 1.0 4.7 14.0 15.6 17.4 16.0 Significance between regimens: *p 104 mm Hg on two consecu- tive visits 2 weeks apart or > 114 mm Hg on two con- secutive visits 1 week apart on maximal dose of medi- cation Hypotensive symptoms persisting after reduction of Zarpine No. 1 to once daily Heart rate > 99 beats/min persisting after reduction of Zarpine No. 1 to once daily, or on two consecutive visits in a symptomatic patient Angina pectoris or myocardial infarction Congestive heart failure Atrioventricular block greater than first degree Grade III or IV hypertensive retinopathy Cerebral hemorrhage. subarachnoid hemorrhage. cere bra1 thrombosis, orhypertensive encephalopaihy Dissecting aneurysm Pulmonary embolism or infarction Arthritis or dermatitis associated with lupus cells in the blood or positive fluorescent ANA test -_ . (12) `I'hrombocytopenic purpura or agranulocytosis (13) Serum creatinine greater than 2.0 mg/dl and 50% higher than the baseline (14) Failure to meet clinic appointments without legiti- mate excuse