"22 Iieprivted from the Archives of IltterR A14gust 1962, Vol. 110, pp. m Copyright 1962, by Anlerican Medical Double Blind Control Study of Antihypertensive Agents II. Further Report on the Comparative EffectivelzeJs of Reserpine, Reserpine and Hydralazine, and Three Ganglion Blocking Agents, Chlorisondamine, . Mecamylamine, and Pentolinium Tartrate sociation VETERANS ADMINISTRATION COOPF.RATIVE STUDY ON ANTIHYPERTENSIVE AGENTS* WASHINGTON, D. C. This report presents the results after one year of follow-up of 759 hypertensive pa- tients who were begun on one of a number of randomized, double blind treatment regi- mens using various antihypertensive agents. The plan of the study and the methods of procedure have been published previous1y.l In brief, only patients whose diastolic blood pres- sure levels averaged 90 mm. Hg or above from the fourth through the sixth hospital days were in- cluded in the study. These patients were divided into categories of "mild, moderate, and severe" on the basis of the scores obtained in 5 panels: basal Submitted for publication Feb. 26, 1962. Complete tabular data will be included in the reprints. * Participants : Mark L. Armstrong, M.D. ; John L. Bakke, M.D. ; Harold Dodge, M.D. ; Loyal L. Conrad, M.D. ; Edward D. Freis, M.D. (Chair- man) ; Rudolph E. Fremont, M.D.; Walter M. Kirkendall, M.D.; Clifford G. Pilz, M.D.; Eli A. Ramirez, M.D.; David W. Richardson, M.D., and John H. Williams, Jr., B.A. These included the following Veterans Adminis- tration Hospitals : Brooklyn ; Chicago West Side ; Towa City; Oklahoma City; Richmond, Va. ; San Juan, P.R. ; Seattle, and Washington, D.C. diastolic pressure, optic fundi, and abnormalities of the cardiac, renal, and central nervous systems. Treatment for patients in the mild and half of the moderately severe group consisted of 1 of 3 possi- ble regimens as follows: (1) reserpine 0.5 mg. twice daily for 2 weeks followed by 0.25 mg. twice daily thereafter plus hydralazine 25 mg. 4 times daily for 2 weeks followed by 50 mg. 4 times daily thereafter, (2) reserpine as above plus placebo of hydralazine as above, and (3) placebo of reserpine plus placebo of hydralazine. Assignment of a given regimen to an individual patient was carried using a double blind, randomization technique.' Each regimen was continued unchanged for one year ex- cept for death, default, or other special reasons described below. Treatment for the remaining half of the moder- ately severe cases and for all of the severe group consisted of reserpine in the dosages stated above plus 1 of 3 ganglion blocking agents : mecamylamine, chlorisondamine, or pentolinium. The blocking agents were specially prepared in approximately cquipotent tablets so that titration of dosage could be carried out in each case without revealing the identity of the particular ganglion blocking drug being administered. In addition to determining clinic recordings of blood pressure, each patient was given a blood pres- sure apparatus for twice daily measurements in the 126 .)`TU/)Y 01; Ah'TIHYPEKTEn'SIVE AGENTS 223 l~omc. Folloa-up reports were sun uuxri~etl ever? groups may be due to the fact that the pa- 2 months, and a complete examination including laboratory studies was carried out at the end of 1 tients tended to have more severe and symp- year of follow-up. tomatic disease and, hence, were more willing to tolerate side-effects. Results Deaths: Since all deaths among veterans 1. Loss of Patients from Given Treatment are reported routinely to Veterans Adminis- lieginten.-Of the 759 patients who were be- tration Central Office, a complete follow-up gun on one of the randomized regimens, 426 as to mortality was obtained. Death accounted were treated without change of therapy for for 14% of the losses (6% of the total the full year. started) during the first year of treatment. Of the 333 lost to the study, 155, or 47% As would be expected because of the in- (20% of the total started), were patients creased severity of disease the percentage who did not return or could not or would not dying was higher in the patients treated with cooperate. The majority of these dropped out blocking agents than in the groups on the in the first 2 months of treatment. Such losses other regimens. The causes of death were as are inevitable when an effort is made to avoid follows: renal failure 11, cerebrovascular ac- selection, since a number of deteriorated, cident 7, myocardial infarction 6, congestive alcoholic, or uncooperative individuals will be heart failure 2, sudden death 5, unknown 4, included. The drop-outs were greater in large and 1 each from ruptured duodenal ulcer, cities, Brooklyn, Chicago, Washington, than dissecting aneurysm, acute alcoholism, and in a primarily rural area (Iowa City). postoperative complications. When the losses were classified according Major Side-Effects: Of the losses, 9% were due to side-effects which were sufficient- to treatment regimen there were no great differences between the various groups. The ly troublesome to warrant discontinuation of percentage lost ranged from 37% in the therapy. The incidence was 15% of the reserpine-hydralazine group to 54% in the losses in patients taking reserpine plus hy- chlorisondamine-treated patients. The rea- dralazine, 6% in those receiving reserpine sons for the losses, however, varied in inci- alone, and 2% in the placebo-treated group. dence in the different treatment categories. It is probable that losses due to this cause For example, 40 or 64% of the 62 losses in the groups receiving active agents were in in the hydralazine-reserpine group were due reality higher than reported, since some of to failure of the patient to return or to lack those who failed to return may have done so because of side-effects. Somewhat fewer of cooperation as compared to only 17, or 37%, of the 46 losses in the placebo group. of the patients receiving pentolinium tartrate were lost because of side-effects than in those The difference is probably explained by receiving mecamylamine or chlorisondamine. the higher incidence of side-effects on the Further data on these cases are presented former treatment regimen. The losses due to below in the section on side-effects. failure to return or to lack of cooperation Treatment Failures: Of the losses, 12% also was high (61%) in the cases treated were due to failure of the regimens to pre- with reserpine. vent severe elevations of blood pressure or Losses due to poor cooperation were lower other menacing complications such as hyper- in the patients treated with ganglion blocking tensive encephalopathy. The criteria for dis- agents, ranging from 32% of the total losses continuing the double blind regimens were as of patients taking mecamylamine to 41% in follows: (1) home diastolic average 130 mm. the case of pentolinium. Nevertheless, the Hg or about for 3 weeks or longer confirmed incidence of reported side-effects was high by clinic readings; (2) home diastolic aver- in the patients treated with ganglion blocking age 140 or above for 1 week or longer; (3) drugs. The reason for the smaller percentage evidence of serious organic progression such of voluntary drop-outs in the latter treatment a~ an attack of acute hypertensive encepha- 127 V&runs Administration 224 ARCHIVES OF INTERNAL MEDICINE lopathy (accompanied by high diastolic pres- sures), advancing congestive heart failure despite digitalis and periodic diuretics, de- velopment of hemorrhages, exudates, and/or papilledema in the optic fundi. The incidence of treatment failure was highest in the placebo group (26%) and low- est in the patients treated with reserpine plus hydralazine (2%). The most common com- plication necessitating a change in therapy was resistant congestive heart failure. In ad- dition there were 3 cases of cerebrovascular accident and myocardial infarction in which a change in treatment was considered neces- sary. II. Antihypertensive Effffectiveness.-Un- less stated to the contrary, the average blood pressure levels reported in a post-treatment period represent the average for 3-month periods of twice daily recordings of blood pressure taken by each patient or a member of his family in the home. The pretreatment control values are based on the averages of recordings taken 4 times daily in the hospital ward from the fourth through the sixth days of hospitalization1 In general, the patients who did not com- plete the full year exhibited a lesser antihy- pertensive response than those who did. This was probably due to the fact that treatment failures and deaths were included in the former group. Post-treatment data for those patients lost during the first 2 months of treatment are not included despite the fact that nearly 50% of all losses occurred during that interval. The reason for not including them is that the results were considered to be untrustworthy, since they,were constituted predominantly from the uncooperative pa- tients who failed to return for follow-up; nor does it seem of great practical importance to know which regimen was more effective for patients who would not accept treatment. Their prevalence in this study varied between 13% and 27% among the various regimens. Average Changes in Blood Pressure on Various Regimens : Since the post-treatment averages were essentially unchanged from quarter to quarter the third month results were used; this included all patients (most treatment failures and deaths) except those who dropped out before completion of the first quarter. The results are similar in most respects to those reported previously in a smaller series of cases.l Reserpine was slightly su- perior to placebos producing an average reduction of 315.1 mm. Hg (systolic/diastol- ic) as compared to an average elevation of 3.7/2.0 mm. Hg in the placebo-treated group. Reserpine plus hydralazine resulted in an average fall of 4.9/10.5 mm. Hg. The results after the 3 ganglion blocking agents did not differ greatly one from the other. In the 88 patients who received me- camylamine, the reduction of blood pressure averaged 16.3/14.1 mm. Hg, in 91 patients given chlorisondamine it was 15.2/12.3, and in 65 given pentolinium tartrate the reduction averaged 15.4/13.4 mm. Hg. In the moderate group of hypertensive patients the average reduction was approximately ll/lO mm. Hg in each of the 3 regimens. In the severe group the reductions were greater, averaging ap- proximately 24/20 mm. Hg. Whereas reser- pine plus hydralazine affected diastolic blood pressure primarily, reserpine plus ganglion blocking agents affected systolic slightly more than diastolic pressure. If the reductions of blood pressure seem rather small for all regimens, it should be remembered that "basal" hospital blood pressures were used as the pretreatment control values. "Casual" pretreatment blood pressures probably would average considerably higher. Percentage Distribution of Response on Various Regimens: Distributions of the per cent of patients with indicated amounts of blood pressure change are shown in Figures 1 through 4. Figure 2 indicates the per- centage distribution of changes in diastolic pressure in the patients taking reserpine, hydralazine, or their placebos. A reduction of 5 mm. Hg or more was obtained in 70% of the patients taking reserpine plus hydrala- zine, 54% of those on reserpine alone, and 33% of the placebo group. The placebo group displays a nearly normal distribution curve for diastolic blood pressure with a mean close to zero change in blood pressure. Corre- Vol. 110, Aug., 1962 12x bV$UcN.OC D OASTQLOC PERCENT oF CPlSES RESERPINE + HYDRALAZINE RESERPINE + HYDRALAZINE 30 0 44 36 26 20 12 404 12 26' 28 36 44 44 :6 ,`8 2'0 ,2 I 4 I 0 k 12 I 80 I 26 I I 36 44 I RESERPINE RESERPINE I I I I , I I 44 36 26 20 I2 4 0 4 I; I I I 1 20 28 36 44 PLACE50 0 414 :6 :6 :O I: 4! 0 : I: :o :* ?!6 14 Fig. I.-Percentage distributions of systolic pres- sure responses with the reserpine-hydraIa~~ne (above), the reserpine-placebo (middle), and the placebo-placebo regimens. Per cent oi patients is indicated on the ordinate and change in systolic pressure on the abscissa with decrease of pressure plotted to the right, 20 IO 0 I I I I I I1 I I I , I 44 36 26 20 12 404 12 20 26 36 44 PLACEBO 44 36 26 20 12 4 04 12 20 is 36 44 Fig. Z.-Percentage distributions of diastolic prcs- sure responses. Other notations as in Figure 1. mm. or more in diastolic blood pressure. The distribution of both systolic and diastolic sponding data for the ganglion blocking pressure as shown in Figures 3 and 4 is quite drugs shown in Figure 4 indicated that 69% similar comparing one blocking agent with to 77% of the patients exhibited a fall of 5 the others. The wide distribution of systolic pressure shows the great variability of re- PERCENl' OF CASES DnAsuQnrlC MECAMYLAMINE 36 26 20 404 12 20 26 36 44 4'4 :6 2'6 2'0 ,: :0: 8: :O / I I 28 36 44 CHLORIS~NDAMINE 30 r- CHLORISONDAMINE 0 10 i I I I I I I 1: t I , I 44 36 28 0 20 $2 4 0 4 20 28 36 44 44 36 :6 :O :8 3'6 44 PENTOLlNlUM TARTRATE 30 F PENTOLINIUM TARTRAT f 30 P T 20 E?m 0 d4 $6 :6 2'0 1: ::a k 1: :o 28 & 44 Fig. 3.--Percentage distributions 44 :6 of systolic :8 :O I: I 4 0: I: :O :z, ;6 pres- 44 sure responses of patients taking 1 of the 3 ganglion blocking agents plus reserpine. Fig. 4.-Percentage distributions of diastolic pres- sure of same patients as in Figure 3. I29 Veterans Administration 226 ARCHIVES OF INTERNAL MEDICINE sponse to all of the ganglion blocking drugs (Fig. 3). Differences Between Hospitals : Tabula- tion of the data by individual hospitals dis- closed that results usually were comparable within the various institutions. The hospitals which treated more Negroes reported higher average blood pressures under treatment for patients on all agents including placebos. There were no hospitals where one blocking agent appeared more effective than the oth- ers. In one hospital, reserpine alone was as effective as reserpine plus hydralazine, but there were only 12 patients on each of these regimens. Effects of Age and Race: With almost all regimens the average reduction in systolic blood pressure was greater in patients under as compared to those over 50 years of age. There were no other significant differences. Similar comparisons were made between Negro and Caucasian patients. No significant differences were observed. Negroes appeared to show no response to reserpine, but the elevation with placebos was greater so that the difference between the reserpine and placebo groups was about the same in Ne- groes as in Caucasians. Influence of Pretreatment Level of Blood Pressure: The degree of change in diastolic blood pressure also was analyzed in regard to the level of pretreatment blood pressure and severity score for the various regimens. The results indicated that regardless of the antihypertensive agent used, the average re- duction in diastolic pressure was greater in the patients with high initial pressures. The higher the pretreatment pressure, the greater was the potential fall. It is probable that greater reductions of blood pressure could have been achieved by more aggressive dos- age with the ganglion blocking drugs in pa- tients with the lowest initial diastolic blood pressure. The average daily dose in the mod- erately severe group was about half that administered to the patients with severe hy- pertension. The reductions of blood pressure were more than twice as great in the severe group. It would appear that when the choice of the dose is left to the discretion of the clinic physician, he will treat most aggressive- ly the patients with the highest blood pres- sure levels. Home Versus Clinic Levels of Blood Pres- sure: The averages of blood pressure levels recorded after a 15minute rest period in the clinic 1 were consistently higher than the averages of recordings made at home. This held true for all regimens including placebos. The relative antihypertensive effectiveness of one regimen as compared to another was about the same whether home or clinic blood pressure averages were used for determining the results. Since the home recordings rep- resent a far greater number of readings (twice daily) and since the average home recordings in the placebo group were much closer to the average of the pretreatment hos- pital recordings, they probably reflect more accurately the general levels of blood pres- sure during the treatment period than do the clinic readings. Stability of Average Blood Pressures : The average of home recordings in a given pa- tient usually varied little from one quarter to another. There often was, however, consid- erable variation from one home recording to the next or from one clinic reading to the next. Thus, the blood pressures of the hyper- tensive patients seemed to fluctuate more or less widely about a mean, the average gen- erally being fairly stable although subject to change depending on the treatment regi- men imposed. Although generally fairly stable, the averages for individuals usually were not as constant as the averages for groups of individuals. The variation of group average blood pressure from one quarterly period as compared to the other quarters never exceeded 4 mm. Hg systolic or 3 mm. Hg diastolic in any of the treatment groups. Effects on Organic Changes: A prelimi- nary analysis of the effects of treatment on organic changes was attempted in the patients who completed 1 year of treatment. These patients were subdivided into 4 groups ac- cording to height of initial blood pressure and then further divided as to whether they did or did not obtain a significant reduction of blood pressure. The average severity Vol. 110, Aug., 1962 130 S-TUDY Ot' A!\:TIHPPI~KTENSIV~ AGENTS 227 scores in the fundic, cardiac, central nervous system, and renal panels as well as their totals were then compared before and after 1 year of treatment. This method of lumping severity scores into averages is inadequate, since it makes the incorrect assumption that 2 grades tif 2 equals 1 of Grade 4. For a preliminary sur- vey of directional changes, however, it of- fered a convenient method of comparison with the understanding that the table should be interpreted only in the most general terms. In all groups there was a decrease in severity scores. For example, in the group treated with placebos the cardiac and central nervous system panel averages showed significant im- provement at the end of 1 year, and even the renal panel was slightly improved. Some of this apparent evidence of im- provement might be accounted for by the fact that data only on the patients treated throughout the year on the same regimen were tabulated. Thus, the unfavorable cases of patients who died or became treatment failures were not included. Since the inci- dence of treatment failures was higher in the groups who did not show a significant reduction of blood pressure, the bias intro- duced would tend to favor these subdivisions, particularly the placebo group. With the pos- sible exception of the patients with initial diastolic pressure levels above 120 mm. Hg the percentage improvement in total panel scores did not indicate significant differences after 1 year between those whose pressures were reduced and those exhibiting no reduc- tion. A longer period of follow-up will be required, however, before any judgment can be made as to the effects of antihypertensive drugs on the progression of organic disease. III. Side-Effects.-Major Side-Effects: A further breakdown of the major reactions in which the side-effects were judged to be suffi- ciently troublesome to warrant change of treatment is as follows: Of the 9 cases in the hydralazine-reserpine group there were 2 instances of depression, 1 of sleep disturb- ance with severe nightmares, 1 of edema of the hands and feet; 1 complained of im- potence, 1 of severe headaches, 1 of nausea 131 Veterans Administration `eutaq~h &aunq paJ!q!yxa I `ssauyzz!p pun but without other evidences of lupus, and 1 had a gastrointestinal hemorrhage of unde- termined cause. Not all of these reactions can be ascribed definitely to drug effects. More appeared to be those associated with reser- pine than with hydralazine. In the 4 cases taking reserpine alone, treatment was dis- continued in 2 because of depression and in the other 2 because of impotence. The latter cases are of interest, since reserpine general- ly is not believed to have effects on sexual function. Treatment was discontinued in one case taking placebos because of a generalized dermatitis. Treatment was changed because of side- effects in 5 patients receiving mecamylamine. The reactions consisted of 1 each of the fol- lowing disturbances: orthostatic hypotension, dizziness and constipation, paralytic ileus, impotence and gastrointestinal complaints, and upper abdominal pain with vomiting. In 10 patients receiving chlorisondamine, ortho- static hypotension accounted for 3 losses, difficult voiding for 1, abdominal pain for 1, diarrhea for 1, and multiple complaints for 4. Treatment was discontinued in the pento- linium group because of orthostatic hypo- tension in 1 and multiple complaints in 1. Minor Side-Effects: An attempt was made to determine the incidence of minor side- effects by a questionnaire-interview method.' Each patient was interviewed by the clinic secretary or physician at regular intervals. The interview was made from a check list containing standard questions relating to side-effects generally believed to be asso- ciated with the agents under study. Unfortunately, the interviews carried out prior to treatment revealed a high incidence of the side-effects usually attributed to reser- pine and hydralazine. For most of the side- effects the changes from the pretreatment incidence were insignificant at both 1 month and 1 year after treatment and also during the intervening quarterly periods. Headaches decreased in the drug-treated and placebo groups. The incidence of nasal stuffiness in- creased in the patients receiving reserpine either alone or with hydralazine as compared 228 ARCHIVES OF INTERNAL MEDICINE to the placebo group. Increased appetite was noted in the reserpine-treated groups but was equally prominent in the placebo-treated pa- tients. For the remaining side-effects of joint pains (hydralazine), lethargy and fatigue, depressed feeling, and nightmares (reser- pine), there were no significant changes after as compared to before treatment in either the drug-treated or placebo groups. The number of patients failing to return because of side-effects was, of course, un- known. Side-effects related to the gastrointestinal tract were common after the ganglion block- ing drugs, the frequency of complaints be- ing definitely higher during treatment than before. As compared to the pretreatment period the incidence of constipation increased after treatment by 21% to 36%, depending on the blocking agent. Other gastrointestinal side-effects including bloating, distention, and anorexia were increased by 16% to 40% and dryness of the mouth by 14% to 27%. The incidence of orthostatic faintness was not in- creased in the period between 1 and 12 months after beginning treatment but may have been higher in the earlier period of dosage regulation, Complaints of impairment of visual accommodation increased after treatment especially in the patients taking chlorisondamine where 19% were required to wear dark glasses when exposed to bright sunlight. Difficult voiding appeared to be an uncommon side-effect. Comment The results of the present study confirm in general the previous preliminary report1 It was not possible from the preliminary results in a smaller series of patients to determine whether reserpine exerted a significant anti- hypertensive effect as compared to placebo. The more extensive data reported here indi- cate that oral reserpine does have an anti- hypertensive effect, although it is not very great. The constancy of results obtained from the first 3-month period to the next for each of the regimens used indicated that 3 months was a sufficient period for the evaluation of antihypertensive effectiveness of these par- ticular agents. Such a short period, however, may not reveal chronic toxicity. The double blind method as employed here did not appear to be a sensitive method for detecting minor side-effects. Part of the dif- ficulty may be due to the fact that side-effects were reported at 1 month and then quarterly when most of the acute disturbances such as headache and palpitation from hydralazine, and constipation and dry mouth, etc., from the ganglion blocking drugs had abated. In addition, some of the early drop-outs may have occurred because of side-effects which, of course, would not be reported. In the pre- treatment period, the high incidence of com- plaints similar to the expected side-effects indicates the need for caution in interpreting the true prevalence of minor side-reactions in studies when such pretreatment evalua- tions have not been made. It is of interest that in patients with mod- erate hypertension, reserpine plus hydralazine resulted in a slightly greater reduction in diastolic pressure than that produced by re- serpine plus ganglion blocking agents, al- though the latter produced a slightly greater fall in systolic pressure. As mentioned pre- viously, the reduction of blood pressure in the patients treated with ganglion blocking agents might have been greater if more ag- gressive dosage had been employed. On the other hand, the participating physicians were experienced in the use of these agents and administered doses which resulted in a sig- nificant incidence of the side-effects associ- ated with ganglionic blockade. The reserpine and hydralazine regimen employed an un- complicated, fixed dose schedule; side-effects ascribable to drug action were lower than in the patients treated with ganglion blocking agents, and hydralazine toxicity was not a prominent feature with the doses of the drug employed in this study. These considerations favor the use of reserpine plus hydralazine over ganglion blocking drugs in the manage- ment of patients with moderately severe hy- pertension, at least as an initial trial before resorting to the blocking agents. Vol. 110. Aug., 1962 132 \-TIJDY OF ANTIHYPERTENSIVE AGENTS 229 Summary Various antihypertensive agents were ad- ministered using a randomized double blind technique to male patients with elevations of basal diastolic pressure above 90 mm. Hg with the following results. Of 759 patients entering the study, 426 completed one year of treatment. Of the losses, 47% were due to failure to return or to lack of cooperation, the majority occurring within the first 2 months of treatment. Deaths accounted for 14% of the losses, severe eleva- tion of blood pressure or acute hypertensive complications for 12%, and side-reactions for 9%. The average blood pressures in the various treatment groups remained essentially un- changed from the third month to the end of the,year. In 86 patients with "mild" and "moderate" hypertension receiving placebos, there was a slight elevation of blood pressure averaging 3.7/2.0 mm. Hg. In 136 similar patients given reserpine in a maintenance dose of 0.5 mg. daily, there was a slight reduction averaging 3.0/5.4 mm. Hg. In 139 patients given the same amount of reserpine plus hydralazine in a maintenance dose of 200 mg. per day, the average reduction of blood pressure was 4.9/10.5 mm. Hg. Reserpine 0.5 mg. daily plus 1 of 3 gan- glion blocking agents-mecamylamine, chlori- sondamine or pentolinium-was tested in treatment groups each comprising 65 to 91 "severe" and "moderately severe" hyperten- sive patients. The antihypertensive effects of the 3 ganglion blocking agents were essen- tially similar, averaging approximately 15.5/ 13 mm. Hg. With all of the antihypertensive regimens, the higher the pretreatment level of blood pressure, the greater was the degree of fall. Age or race did not seem to influence anti- hypertensive effectiveness significantly. In general, inter-regimen differences were similar in each of the participating hospitals. Llie wish to thank Dr. Edward Dunner, Research and Education Section, Central Office, Veterans Administration, for efficient management and co- ordination of administrative matters connected with this investigation. We also wish to thank the mem- bers of the Departments of Clinical Investigation of Ciba Pharmaceutical Products, Merck Sharp & Dohme, and Wyeth Laboratories for manufacture of special tablets, bottling, and code labeling of the various pharmaceutical products used in this study. Finally, the helpful advice and criticism of our consultants, Dr. Earle Estes and Dr. Eugene Stead, are gratefully acknowledged. Edward D. Freis, M.D., Chairman, Antihyper- tensive Drug Study, Veterans Administration Hos- pital, 2650 Wisconsin Ave., N.W., Washington 7, D.C. REFERENCE 1. Veterans Administration Cooperative Study on Antihypertensive Agents : A Double Blind Control Study of Antihypertensive Agents, Arch. Intern. Med. 106 :81-96, 1960. 133 Veterans Administration Printed and Published in the United States of America