Reprinted From The Journal of the American Medical Association December 17, 1967 Volume 202 Copyright 7967, American Medical Association Effects of Treatment on Morbidity in Hypertension Results in Patients With Diastolic Blood Pressures Averaging 115 Through 129 mm Hg Veterans Administration Cooperative Study Group on Antihypertensive Agents A group of 143 male hypertensive patients with di- astolic blood pressures (at the clinic) averaging between 115 and 129 mm Hg were randomly assigned to either active (hydrochlorothiazide plus reserpine plus hydrala- zine hydrochloride) or placebo treatment. Twenty-seven severe, complicating events developed in the placebo- treated patients as compared to two in the active group. Four deaths occurred in the placebo-treated group and none in the actively treated patients. Other complications in the placebo group included grade 3 or 4 hypertensive retinopathy, congestive heart failure, increasing azotemia, cerebrovascular thrombosis, transient ischemic attacks, cerebral hemorrhage, myocardial infarction, and severely elevated blood pressure. Severe complications in the active-treatment group were one cerebrovascular throm- bosis and one case of multiple drug toxicity. Male pa- tients with diastolic blood pressures averaging 115 mm Hg or above represent a high-risk group in which antihy- pertensive therapy exerts a significant beneficial effect. T he value of antihypertensive drug treatment in malignant hypertension has been amply dem- onstrated."" However, its effectiveness in preverit- ing morbidity and mortality in less severe forms of hypertension has been disputed.`." Adequately con- trolled, prospective studies are needed to evaluate this question in patients with essential hyperten- sion.' An investigation of this type was initiated b.y the Veterans Administration Cooperative Study Group on Antihypertensive Agents in 1963. The design of the study, including the precautions em- ployed to maintain adherence to protocol and avoidance of dropouts, has been described in a pre- vious communication.`" The present report is con- cerned with the results obtained in the patients For complete list of participants, see page 1033. Reprint requested to 50 Irving St NW, Washington, DC 20422 (Dr. Edward I>. Freis). 116 without signs of accelerated hypertension at admis- sion whose diastolic blood pressures prior to treat- ment averaged 115 through 129 mm Hg. Plan of Investigation All patients were hospitalized for the initial work- up. Male patients whose diastolic blood pressures from the fourth through the sixth day of hospital- ization averaged 90 through 129 mm Hg without treatment were considered for admission to the prerandomization trial period. Severity was evaluated in five categories. These were the average diastolic blood pressure during hospitalization and the degree of clinically detect- able hypertensive damage in the following four tar- get organs: the optic fundi, the brain, heart, and kidneys. Severity of damage in each category was graded on a scale from 0 (no detectable abnormal- ity) to 4 (most severe changes). The criteria used for grading severity have been described in detail elsewhere." After doubling the scores for the aver- age diastolic blood pressure and the severity of damage to the optic fundi, the scores obtained in each category were summed to obtain a total sever- ity index. Patients with total scores of 2 through 7 were classified as mild, 8 through 15 as moderate, and 16 or above as severe, the latter being excluded from the trial. Also excluded from the study were patients with surgically curable hypertension, uremia, and con- comitant fatal diseases such as carcinoma. Patients with hemorrhages, exudates, or papilledema in the optic fundi, history of cerebral or subarachnoid hem- orrhage, dissecting aneurysm, or congestive heart failure resistant to digitalis and mercurial diuretics were excluded. Additional exclusions included pa- tients who wished to return to the care of their private physicians, those who for geographical or JAMA, Dee 11, 1967 o Vol202, No 11 MORBIDITY IN HYPERTENSION 1029 other reasons would be unable to attend clinic regu- larly, and patients of dubious reliability such as alcoholics, vagrants, and poorly motivated patients. Table l.-Background of Randomized Patients Prerandomization Trial Period.-Following dis- charge from the hospital, the patients entered a prerandomization trial period of two to four months' duration. They received two placebos, known to the physician but not to the patient, and were seen in the clinic at monthly intervals. Characteristic NCL No. Placebo Treated pv:!J Riboflavin (5 mg) was incorporated into one of the placebos. Riboflavin produces a yellow fluores- cence of the urine when the latter is viewed under ultraviolet light. At each visit, a urine specimen was examined for fluorescence. Patients were required to return all bottles of medication, at which time the tablets were counted to assess the patient's reliability. To qualify for admission to the study, the patients were required to have no "violations" on two successive clinic visits. A violation consisted either of failure to appear at the regularly scheduled clinic appointment, or failure of the urine to ex- hibit fluorescence, or a tablet count (of either of the two types of placebos) which ws outside the ac- ceptable range. The upper limit of the acceptable range was defined as the return of no more than a 10% excess of the calculated number of tablets re- maining if all doses had been taken as prescribed, while the lower limit was a return of five tablets less than the same calculated number. Nearly one half of the patients accepted into the trial period were excluded prior to randomization because of failure to pass the above tests of reliability. Total randomized White N&!g;rCl Family history of hypertension NO"2 Present Unkncwn Cardiac symptoms Non.2 Present Heart size by roentgenogram Ungerleider normal Ungerleider enlarged Electrocardiogram Left ventricular hypertrophy absent Left ventricular hypertrophy present Prior cardiovascular thrombosis Occipital headaches Diabetes absent Diabetes present 70 73 35 31 35 42 19 23 48 49 3 1 48 52 22 21 39 44 31 29 48 49 22 24 5 6 12 10 65 65 5 a Table 2.-Measurement Data Characteristic Age (yr) Height (in) Weight (lb) Duratlo" known hypertension (yr) Average hospital diastolic pressure (mm Hg) Average clinic systolic (mm Hg) Average clinic diastolic (mm Hg) Severity grades (0.4)*' Fundi (hvaertensive) Fundi (&erotic) Cardiac Central nervous system Re"al Blood glucose, fasting (mg/lOO cc) Blood glucose, 2.hr postprandial Cholesterol (mg/lOO cc) Placebo ----\ / Mean SD 51.4 10.8 69 2.6 182.9 34.5 5.4 4.4 105.8 a.4 Active __;\ Mean SD 50 a.7 68.7 2.9 185.2 36.7 5.3 4.7 186.8 17.2 121 4.7 106.5 a.4 185.6 15.4 121.2 5 1.3 . . . 1.4 . . . 1 . . . Also excluded were patients who during the trial period exhibited diastolic blood pressures (while in the sitting position) averaging below 90 or abovt 129 mm Hg. Thus, the final decision to accept the patient into the study was based on both reliability and the Ieve of diastolic blood pressure determined in the clinic during the trial period. 1.2 . . . 1.3 . . 1.2 . . . 0.4 . 0.5 . 96.8 20.6 0.3 . . . 0.3 . . . 97.7 la.9 ii8 41.5 116.1 54.3 251.3 59.5 242 51.5 *Detailed criteria for grades 0 through 4 are given in reference 11. Table S.-Trends of Diastolic Blood Pressure Placebo Postrandomization Period.-At the time of ran- domization, a sealed envelope was opened which assigned the patient to one of two possible regi- mens-active antihypertensive medications or their placebos. A table of random numbers was utilized by the statistician in determining the assignments. Patients classified by severity scores as having mild hypertension were randomized in a separate strati- fication from those with moderate hypertension. The double-blind technique was employed by uti- lizing a series of complex code numbers to disguise the identity of the randomized treatments and by making active drugs and placebos identical in ap- pearance. It is realized, however, that blood pres- sure levels and side effects made the maintenance of such a double-blind study difficult and imperfect. -- Time of No. Patients Observation Observed" Prerandomization 70 At 4 months 57 a 50 12 44 16 33 20 27 24 23 Average ' Diastolic, mm Hg 121 118.5 120.3 i 18.8 118.5 115.2 119.7 Active ----, Average No. Patients Diastolic, Observed mm Hg 73 121.2 68 93.1 64 92.2 58 91.6 47 92.1 40 89.4 32 91.5 &The decline I" the number of patients observed is due primarily to the fact that patients were admitted over a 2%year period. Hence. many were not in the study long enough to be observed at the longer time periods. The active drugs were incorporated in the two tablets as follows: tablet A contained 50 mg hydro- chlorothiazide plus 0.1 mg reserpine, and tablet B contained hydralazine hydrochloride. Tablet B was available in two strengths, 25 and 50 mg. Placebos were made up to correspond with each tablet of the active drugs. Treatment was begun with either tablet A, one twice daily, plus the 25-mg t.ablet B, one tablet three times daily, or else with the placebos of both of these tablets. At the next visit, tablet B was increased to the 50-mg strength. Thus, the regular maintenance daily dose to patients randomized to the active regimen was 100 mg of hydrochlorothia- zide, 0.2 mg of reserpine, and 150 mg of hydrala- zine hydrochloride. However, if there were hypo- tensive reactions or other severe side effects, doses could be reduced to a tolerable level. In order to further minimize losses due to drug JAMA, Dee 11, 1967 o Vol 202, No 11 117 1030 MORBIDITY IN HYPERTENSION IO CHANGI IN sY*ToLI` BLOOD PRESSURI ^ mm M CH*NCf IH OIISToc1C BLCOD PRlSsURt ml n9 Changes in systolic (left) and diastolic blood pressure (right) after four months of treat- ment in 57 oatients eiven Dlacebos (above) and 68 Datients treated with hydrochloro- thiazide plus' reserpine plus'hydralazine (below). . group have since been given active treatment. The total number of high-risk patients ran- domized was 143. Of this total, 70 received pla- cebos, and 73 received active antihypertensive drugs during the random- ization trial. Their aver- age age was 51 years with a range of 30 to 73 years. The average weight was 83.5 kg (184 lb). Seven- ty-seven patients were Negroes and 66 whites. There were no significant differences with regard to age, weight, duration of known hypertension, or family history of hyper- tension, between the pla- cebo- and active-treat- ment groups (Tables 1 toxicity, two special A tablets were available on request. One contained hydrochlorothiazide with- out reserpine for patients becoming depressed or having active peptic ulcer or other reserpine-associ- ated severe side effects. The other contained reser- pine without hydrochlorothiazide for toxicity asso- ciated with the latter, such as hyperglycemia or acute gout. These special A tablets were available in both active and placebo forms. The patients visited the clinic at monthly inter- vals for the first two months following randomiza- tion and at bimonthly intervals henceforth. Additional interim visits were scheduled if needed. Tablet counts were made at all clinic visits and fluorescence tests of the urine were made at alter- nate visits. Annual examinations included complete physical examination, roentgenogram of the chest, electrocardiogram, blood chemistry, and renal func- tion tests. General medical care measures and symptomatic treatment, with the exception of known active antihypertensive drugs, were em- ployed in all patients. Characteristics of Patients Only the patients whose diastolic blood pressures averaged 115 to 129 mm Hg during the last two prerandomization clinic visits are included in the present report. For this high-risk group the study was ended in May 1967. The patients with lower prerandomization diastolic readings are continuing in the study and will be reported on later. The study was terminated in the 115 mm Hg and above group at, an earlier date than expected when it became apparent that the risk rate increased sharp- ly at these levels of diastolic blood pressure and that the clinical course of such patients appeared to be favorably influenced by antihypertensive drug treatment. Therefore, all patients in this high-risk and 2). There were more Negro and diabetic pa- tients in the actively treated than in the placebo group, but the differences were not significant. The various indices of severity such as hospital and clinic blood pressure; funduscopic, cardiac, central nervous system, and renal abnormalities were es- sentially similar in the two groups. The 143 patients were admitted into the study from April 1964 to December 1966. Observations on all these patients ended in May 1967. Thus, there is considerable variation in the duration of observation (Table 3). Nevertheless, 38ye of the patients were observed for two years or more. The duration of the postrandomization phase of the study averaged 15.7 months for the placebo-treated patients and 20.7 for the active-drug group. Twen- ty-six of the placebo patients were in the postran- domization period of the study for less than one year as compared to 15 of the active group. Twen- ty-three placebo-treated patients exceeded a two- year period of postrandomization follow-up as opposed to 32 of the actively treated patients. The briefer period of follow-up of the patients in the placebo group was caused by the larger number of terminating events which developed in this group. Modification in Treatm.ent Regimens.-Of the 73 Table 4.-Incidence of Mortality and Morbidity Placebo-Treated Actively Treated Patients Patients Deaths 4 0 Class A events 10 0 Subtotal 14 0 Other treatment failures 7 1 Total terminating events 21 1 Class 6 events (nonterminattng) 6 1 Total 27 2 118 JAMA. Dee 11, 1967 o Vol 202. Ng 11 . MORBIDITY IN HYPERTENSION 1031 patients randomized to active antihypertensive drugs, 45 received the standard maintenance doses of hydrochlorothiazide, 50 mg plus reserpine 0.1 mg twice daily, and hydralazine hydrochloride, 50 mg three times daily, throughout their participation in the study. Dosages were reduced in the remaining patients because of low blood pressure levels with the standard regimen or because of side effects such as severe headache or weakness. There were no cases of systemic lupus erythematosus. Two patients were transferred to the special A tablets, one because of depression and the other because of a hyperglycemic reaction. Sixty-six of the 70 placebo-treated patients re- ceived standard maintenance dosages. Reduced doses of tablet B were given to two patients be- cause of presumed hydralazine-induced side effects. Special tablet A placebos were substituted in two other patients because of depression. Dropouts.-The total number of dropouts was 12 or 8.4%. Nine occurred during the first two months following randomization. Seven had been randomized to placebos and fivt: to active drugs. Thus, the dropout rate was small and was approxi- mately equally divided between the active- and placebo-treated patients. Changes in Blood Pressure.-Systolic and dia- stolic blood pressures fell promptly and significant- ly in the actively treated patients and remained at these reduced levels throughout the trial. The pro- spective trends for diastolic blood pressure recorded with the patients in the sitting position are shown in Table 5. After 24 months of active treatment. the reduction from prerandomization levels of clinic blood pressure averages 43 mm Hg systolic and 29.7 mm Hg diastolic. By contrast,, the placebo- treated patients showed no significant changes in average blood pressure levels following random- ization (Table 3). The distribution oi: individual changes are shown in the Figure. There is a marked shift to the left into the "decrease" zone for th