Antihypertensive Efficacy of Treatment Regimens Used in Veterans Administration Hypertension Clinics
Abstract—There is continuing uncertainty about whether morbidity and mortality of treated hypertensive patients depends on the drug(s) used to treat or only on the level of blood pressure achieved. This study was undertaken in a sample of special Veterans Administration hypertension clinics to determine which antihypertensive drugs were selected by the involved healthcare providers and how effective they were in achieving normotension. Hypertensive veterans (n=6100) were followed in six VA Hypertension Screening and Treatment Program clinics for 46 months beginning in May 1989. Their average age was 60.7 years; 53% lived in the Stroke Belt; 46% had target organ damage, 36% were black, 23% smoked, and 10% had diabetes mellitus. Antihypertensive regimens were divided into 12 all-inclusive categories. Blood pressures were averaged at the last study visit for all patients on a regimen. The regimens of diuretic or diuretic plus β-blocker gave the lowest average pressures (140.6/82.3 mm Hg) and calcium antagonist the highest (149.0/86.5 mm Hg). ANOVA indicated that differences between seven common regimens and also between the four single drug regimens were highly significant (P<.0001). This pattern of low treated pressure with the “old” agents and higher treated pressure with newer agents was reflected in the percentage of patients controlled below 140/90 mm Hg and the percentage uncontrolled above 159/94 mm Hg. Blacks and patients with target organ damage resembled the entire cohort in average treated diastolic blood pressure, but the former had lower and the latter had higher treated systolic blood pressure than the entire cohort.
This report examines the overall results of treating hypertension in six Hypertension Screening and Treatment Program (HSTP) clinics of the Department of Veterans Affairs. The period covered was from May 1989 through February 1993. The choice of antihypertensive regimen and the extent of the effort to achieve the HSTP diastolic blood pressure (DBP) goal of 95 mm Hg were left to the discretion of the physician-coordinator of the clinic and the clinic staff. For this report, the various pharmacological treatments that were used have been divided into 12 regimens.
In 1974, the Veterans Administration (VA) initiated a special HSTP consisting of physician-supervised clinics in which primary care was provided by nurses and physician assistants intensively trained in hypertension.1 2 Initially they followed the treatment guidelines, including step-care, as described by Task Force I of the NHBPEP.3 These guidelines were periodically revised in accordance with the Joint National Committee (JNC) Reports, culminating in JNC-IV.4
For the first 3 years of its existence from 1974 through 1976, the adequacy of the HSTP was monitored by having clinical data collected at all patient visits and entered into a computerized database. Approximately 10 000 previously untreated hypertensive veterans were carefully followed during this period. Combination of their baseline and early treatment data with 15- to 20-year mortality data has yielded useful information on the natural history of treated hypertension.5 6 7 8 9 10
The data collected here were obtained in a “real world” VA outpatient setting rather than in a randomized double-blind research setting. The cohort studied was unusual in being 50% minority, economically deprived, and drawn from throughout the United States. In addition, it was primarily male and largely limited to those with mild and moderate hypertension because more severely hypertensive subjects were referred to more intensive care. This data collection project was designed to provide both quality control of antihypertensive treatment for the individual and information about the efficacy of treatment for individual clinics as well as for the entire HSTP.
This report examines the efficacy with which blood pressure was controlled by 12 mutually exclusive but all-inclusive pharmacological antihypertensive regimens as they were used at six clinics. A separate report deals with adverse effects, including morbid and mortal events.
A total of 6100 veterans were treated in six HSTP clinics (Indianapolis, Ind; Iowa City; Jackson, Miss; Memphis, Tenn; St Louis, Mo; and San Juan, PR) during the 46 months between May 1, 1989, and February 28, 1993. The clinical information determined at each patient visit was computerized with a program developed for the HSTP. A study chairman’s office at the St Louis, Mo, VA Medical Center supervised and facilitated the data collection at the HSTP clinics and data transmission to the VA Coordinating Center at Perry Point, Md.
At the Coordinating Center, the data were entered into a master file and analyzed. All clinics began entering clinic visit data in May of 1989 except for Indianapolis, which began in September of 1989; all continued to enter data through February of 1993. Every patient with a visit at one of the six HSTP clinics during this period was included in this report.
Two one-page forms were used to record data. Form 1 was completed once when a veteran entered the HSTP; it included demographic and pre-HSTP data including age, race, sex, height, past and family histories, years of education, and marital status. Form 2 (except for the “annual visit section”) was completed at each visit; it included the average of three blood pressures, antihypertensive medication prescribed at that visit (drug and dose), self-reported compliance with the medication prescribed at the prior visit (>80% of pills taken, 50% to 80%, 20% to 50%, or <20%), individually volunteered symptoms and their severity (easily tolerated, tolerated with difficulty, or intolerable), interval history (including any morbid cardiovascular event since last visit), ancillary medication, and weight.
The “annual visit section” of Form 2 was designed to be completed every 12 to 18 months. It included alcohol intake (none, <3 drinks/d, or ≥3 drinks/d), cigarette usage (number/d), abnormal physical findings (with emphasis on the cardiovascular system), and evidences of target organ damage (in retina, heart, kidney, nervous system, and vasculature). It also included ECG abnormalities, as defined by the Minnesota Code11 ; its specifically listed diagnoses were left ventricular hypertrophy, old myocardial infarction, premature ventricular contractions, ST-T changes, arrhythmias, and conduction defects. The annual visit section also included laboratory data, specifically serum glucose (with notation if nonfasting), serum cholesterol (HDL cholesterol optional), serum creatinine, potassium, uric acid, SGOT, and finally urinary protein and sugar by dipstick. Entries in the “annual visit section” of Form 2 were accepted into the master file whether or not all items were completed.
The characteristics of the cohort are tabulated in Table 1⇓. There were 6025 men and 75 women. Fifty-nine percent of the entire cohort was older than 60 years of age, with the percentage of patients over 60 varying among the six clinics from 56% to 64%. The racial distribution for the entire 6100 patients was 36% black, 50% white, and 14% Hispanic; however, it varied markedly among the six clinics: 98% of the Iowa City cohort was white, 72% of the St Louis cohort was black, and 99% of the San Juan cohort was Hispanic.
Target organ damage had been recognized in 46% of patients before their last visit during the observation period. On this basis, these patients were designated as having “complicated hypertension” or as being “complicated.” The incidence of damage to individual target organs was retina, 27%; heart, 14%; peripheral vasculature, 6%; kidney, 4%; and central nervous system, 4%. Some ECG abnormality was recorded for 46% of patients, with 12% having left ventricular hypertrophy. The cohort with target organ damage differed from the entire cohort in age and racial composition. Only 41% of blacks but 48% of nonblacks had target organ damage. Likewise, only 37% of patients 60 years of age or younger but 51% of those older than 60 had target organ damage.
As of February 28, 1993, the 6100 patients had been followed in the HSTP for an average of 6.4±5.7 years. Eleven percent of all patients had been followed for >15 years; 17% for 11 to 15 years; 20% for 6 to 10 years; and 52% for 0 to 5 years. The percentage followed for >15 years differed markedly among clinics, with a range of 1.5% to 40%.
On entering the HSTP, the average blood pressure for the entire 6100 patient cohort was 148/92 mm Hg; for 19% (1092 patients) who denied taking antihypertensive drugs at that time, the average was 158/100 mm Hg. The first blood pressure recorded in this 46-month study period averaged 144/86 mm Hg for the entire cohort; the last one recorded averaged 144/85 mm Hg.
A master file was compiled for all 6100 patients consisting of (1) the demographic and baseline data obtained at the first HSTP visit after April of 1989, (2) blood pressure and medication data obtained at all of the 41 498 visits made by these 6100 patients during the 46 months of study follow-up, and (3) physical findings, laboratory work, alcohol intake, and cigarette usage for the 3421 of these patients who had annual visit data collected during the last year of follow-up. The number of visits averaged 6.8 per patient. Eleven percent of patients had only 1 visit, 19% had 2 or 3 visits, 25% had 4 to 6 visits, 25% had 7 to 10 visits, and 20% had >10 visits. The interval between visits averaged 118 days (median, 96 days). The total follow-up time equaled 11 444 patient-years.
Determination of Blood Pressure
At all visits, three blood pressures were measured on seated subjects under minimally stressful conditions, with standardized mercury sphygmomanometers. The three pressures were averaged to obtain a “visit pressure.” In compiling Table 3⇓, the blood pressure at the last visit is coupled with the antihypertensive regimen prescribed at the next-to-last visit because that medication was considered to be primarily responsible for blood pressure at the following visit.*
Antihypertensive Treatment Regimens
In Table 2⇓ and subsequent tables, the patients are divided into 12 groups on the basis of the antihypertensive regimen prescribed at their next-to-last visit. These 12 regimens are defined as ND, no antihypertensive drug; D, diuretic; B, β-blocker; A, ACE inhibitor; C, calcium antagonist; OD, other single antihypertensive drug; D+B, diuretic+β-blocker; D+S, diuretic+non–β-blocker sympatholytic agent; D+A, diuretic+ACE inhibitor; D+C, diuretic+calcium antagonist; D+OD, diuretic+single antihypertensive agent not listed above; and Other, any regimen not listed above. All two-drug regimens without a diuretic and all regimens with three or more drugs are included under “Other.”
Sixty-one percent of the patients were receiving one of the seven most frequently prescribed, that is, common regimens. These included four one-drug regimens (D, B, A, and C) and three two-drug regimens consisting of diuretic plus a second drug (D+B, D+A, and D+C). These seven regimens are emphasized in the presentation of “Results” and in the “Discussion.”
ANOVA and χ2 techniques were used to assess the significance of differences between treatment regimens. Confidence intervals were calculated in two different ways. Standard 95% confidence intervals for average systolic blood pressure (SBP) and DBP were calculated independently for each treatment regimen. Confidence intervals for the differences between regimens were determined from Tukey’s studentized range test. Two-way ANOVA with medical center as a blocking factor was also used to assess the significance of differences between treatment regimens with regard to SBP and DBP.
Prescribed Antihypertensive Agents
Sixty-one percent of the 6100 patients studied used one of the seven common regimens, that is, four single drug regimens and three regimens involving diuretic plus a second drug. Two of the seven, the two single-drug regimens consisting of either diuretic (D) or calcium antagonist (C), were used much more frequently than any of the other five (Table 2⇑). The 39% of patients not following any of these seven common regimens were following a variety of regimens (Table 2⇑). Two-drug regimens without diuretic (7.6%, n=463), three-drug regimens (11.9%, n=724), and four- or five-drug regimens (0.12%, n=62) have been combined into “Other Treatment Regimens” in Table 2⇑. The six clinics differed widely in their patterns of regimen usage (Table 2⇑).
Total Usage of ‘Early’ Drugs
An examination of the overall usage of antihypertensive agents revealed that 3265 patients across all treatment regimens (54% of the cohort) took diuretics. Of those taking diuretics, 55% used hydrochlorothiazide (exclusive of combination pills), 25% used the combination of hydrochlorothiazide and triamterene, 8% used furosemide, and 5% used chlorthalidone. Twenty-one percent of patients took a β-blocker, with atenolol being responsible for 44%, metoprolol for 22%, propranolol for 15%, and nadolol for 13% of their usage. A total of 11.8% of patients took one of the older sympatholytic agents, with clonidine accounting for 63% and reserpine for 29% of their usage.
Total Usage of ‘Newer’ Drugs
Of the “newer” antihypertensive agents, calcium antagonists were taken by 33% of the patients, ACE inhibitors by 25%, and α-blockers by 4.9%. For calcium antagonists, the percentage used in individual clinics varied substantially, ranging from 19% to 49%. Nifedipine was responsible for 33% of the calcium antagonist usage, verapamil for 53%, and diltiazam for 13%. For ACE inhibitors, the percentage used in individual clinics ranged from 15% to 58%. Enalapril was responsible for 60% of the ACE inhibitor usage, lisinopril for 26%, and captopril for 14%.
In Fig 1⇓, the treatment regimens for the entire 6100 patients are arranged in ascending order of average SBP at last visit, and the 95% confidence intervals are indicated. The two-drug regimen consisting of diuretic+β-blocker had the lowest average SBP (140.1 mm Hg), whereas diuretic alone and β-blocker alone had the next two lowest average SBPs. In contrast, calcium antagonist had the highest average SBP (149.0 mm Hg). The 95% simultaneous confidence interval for the difference between SBP associated with diuretic and the SBP associated with calcium antagonist was 5.1 to 11.3 mm Hg.
Fig 1⇑ gives the pattern for DBP; this was similar to but not identical with that for SBP. Here the treatment regimens are arranged in ascending order of average DBP, with the extreme contrast being between diuretic (81.9 mm Hg) and calcium antagonist (86.5 mm Hg). Again the 95% simultaneous confidence interval for the difference between the DBPs associated with diuretic and calcium antagonist was 3.0 to 6.3 mm Hg.
The average blood pressure was lower for the entire cohort in the three clinics with high diuretic usage (≈21%) and low calcium antagonist usage (≈9%) than it was for the entire cohort in the other three clinics, which had an almost reverse usage pattern of “early” versus “newer” drugs* (141.8/82.0 versus 147.0/87.7 mm Hg) (Table 3⇓). Thus the average difference in blood pressure between these two groups of clinics was 5.2/5.7 mm Hg. However, for each of the six clinics individually, the averages of both SBP and DBP for the patients on diuretic alone were lower than the averages for patients receiving calcium antagonist alone (Table 3⇓). The average difference in blood pressure between these two single-drug regimens was 5.3/2.2 mm Hg (Table 3⇓).
A two-factor ANOVA model was used with clinic as a blocking factor to evaluate the differences between the seven common regimens as well as the differences between the four single-drug regimens of Table 3⇑. The results were similar for both the seven common and the four single-drug regimens and indicated that the differences between regimens were significant (P<.0001) and did not depend on interclinic differences, although remaining interactions indicated that regimen efficacy did differ among clinics.
SBP ≥160 mm Hg
SBP at last visit is given by treatment regimen in Table 4⇓. The treated SBP was <140 mm Hg for 43% of the entire population and <160 mm Hg for 79%. At the other extreme, 5% of the cohort had treated SBP of ≥180 mm Hg (Table 4⇓). At the time these data were collected, the HSTP had no stated SBP goal. Diuretic plus β-blocker had the smallest percentage of patients whose treated SBP was >159 mm Hg (14.3%), whereas calcium antagonist had the largest percentage (28.0%) (Fig 2⇓).
DBP ≥95 mm Hg
For the six individual clinics, the fraction of patients with DBP <95 mm Hg ranged from 75% to 93%. Reflecting the average DBP pattern, diuretic was associated with the smallest percentage of patients with treated DBP of >94 mm Hg (8.6%), whereas calcium antagonist was associated with the largest percentage (20.4%) (Table 5⇓).
Overall, the 2175 black patients had better SBP control than the entire cohort (Table 4⇑). For each of the 12 regimens, blacks had lower average treated SBP than the entire cohort. Moreover, for 11 of the 12 regimens, the percentage of black patients with SBP >159 mm Hg was less than the percentage for the entire cohort (Fig 2⇑). In contrast, blacks did not differ from the total cohort in average treated DBP or in percentage of patients with treated DBP >94 mm Hg (Table 5⇑).
Considering only the seven common regimens, blacks who took a diuretic plus β-blocker had the lowest average SBP (135.4 mm Hg), and those who took a calcium antagonist had the highest (144.5 mm Hg). These two regimens also produced the lowest and highest percentages of patients with SBP> 159 mm Hg (10.6% and 19.2%) (Fig 2⇑).
Patients With Target Organ Damage (Complicated Hypertension)
The 2776 patients with target organ damage (both black and nonblack) had higher average treated SBP (2.6 mm Hg) and a higher percentage of patients with SBP >159 mm Hg (3.9%) than the entire cohort (Table 4⇑). However, like the blacks, the complicated hypertension subjects had treated DBP similar to the treated DBP of the entire cohort (Table 5⇑).
Considering only the seven common regimens, complicated hypertension subjects who took diuretics had the lowest average SBP (141.8 mm Hg), and those who took calcium antagonists had the highest (151.1 mm Hg). Diuretics and ACE inhibitors produced the lowest and highest average DBPs, respectively (81.0 and 86.4 mm Hg).
Change in Treatment Regimen
Changes in treatment regimen (drug and/or dose) during the follow-up are summarized in Table 6⇓. For 23.6% of the 5440 patients with more than one visit, treatment was not changed at all during the 44-month period of observation; for 26.6% it was changed only once; for 19.7% it was changed twice; for 12.3% it was changed thrice; and for 17.7% it was changed more than three times. One third of the changes (34%) were because of unsatisfactory blood pressure control; 15% were to reduce dosage; 12% were because of symptoms; and 15% were made outside of the HSTP clinic and hence for uncertain reasons. The remaining one fourth of the changes were made for miscellaneous or unspecified reasons.
Forty-six percent of patients who were receiving a diuretic at the next-to-last visit had not changed medication during the observation period. In contrast, only 18% of patients receiving a calcium antagonist or an ACE inhibitor and only 13% receiving diuretic+ACE inhibitor had no change (Table 6⇑).
The data reported here indicate significant differences in the blood pressure control associated with various antihypertensive regimens at six HSTP clinics. Diuretics and β-blockers were associated with both lower average blood pressures (both SBP and DBP) and lower percentages of uncontrolled blood pressures than calcium antagonists (Tables 4⇑ and 5⇑). For DBP, the ACE inhibitor pattern resembled the calcium antagonist pattern. In contrast, for SBP, patients receiving an ACE inhibitor did better and their values resembled those of patients receiving a diuretic more closely than those of patients receiving a calcium antagonist.
Finally, the difference in blood pressure response to the various regimens was reflected in how often the regimen was changed, with half of those receiving a diuretic having no change during follow-up versus one fourth of those receiving a calcium antagonist or ACE inhibitor (Table 6⇑).
A simple explanation is that for the elderly cohort and the clinical conditions described here, the diuretic was a more potent antihypertensive agent than was the calcium antagonist. However, other possible explanations for this apparent difference in efficacy cannot be excluded. One is that clinics prescribing a calcium antagonist for large numbers of patients (St Louis, 23.0%; Iowa City, 22.8%) did not set their goal pressures as low as clinics prescribing a diuretic for large numbers of patients (San Juan, 23.6%; Jackson, 20.1%; and Memphis, 18.5%). A second explanation is that patients who were begun on diuretic and who were uncontrolled by that regimen were then switched to a calcium antagonist or had a second drug added to the regimen. A third explanation is that patients originally selected for diuretics differed from those selected for calcium antagonists, that is, that the latter group had higher pretreatment blood pressures, had more complicated hypertension, or differed in other ways that might affect treated blood pressure.
In blacks, the average treated SBP differed from that achieved by the entire cohort (141.2 versus 144.2 mm Hg); however, there was no difference in DBP. A partial explanation for this dissociation of SBP and DBP may well be the absence of an SBP goal and the presence of a DBP goal that was attained by both blacks and nonblacks. The dissociation in blacks might be associated with their younger age (1.4 years) and therefore a less predominantly systolic hypertension. In support of this, the 16% of blacks for whom pretreatment blood pressures were available had lower pretreatment SBP (1.8 mm Hg) and higher pretreatment DBP (1.5 mm Hg) than the rest of the cohort (Table 1⇑).
The level of achieved average SBP in hypertensive subjects with target organ damage also differed from that of the entire cohort; for them it was higher, again possibly reflecting the absence of an SBP goal and an age difference, the average age of the complicated hypertension subjects being 2.1 years older than that of the entire cohort.
Whether this better control of blood pressure with early drugs than with newer agents that was observed in an unblinded “real-world” study of elderly veterans will be reflected in the morbidity and mortality results of a randomized, blinded study like the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Disease Trial (ALLHAT), which is currently under way,12 remains to be seen. Certainly the relative antihypertensive efficacy of the various classes of drugs observed in the VA Cooperative Study of Single Drug Therapy13 was different and apparently below what was observed here. The conditions, however, were very different, and any comparison would be comparing two completely different types of trials.
The importance of including a diuretic in antihypertensive drug regimens when more than one agent is needed was demonstrated in a recent report from the VA Cooperative Study Group on Antihypertensive Agents. Patients who failed to be controlled with either of two randomly selected antihypertensive drugs as monotherapy (from six study drugs) were then treated with a combination of the two drugs. Two-drug regimens containing the diuretic hydrochlorothiazide as one of the drugs controlled 69% of these “monotherapy-resistant” patients; regimens not containing a diuretic only controlled 51% of patients.14
The above data from six VA HSTP clinics do not represent the results of a randomized clinical study. Rather, choice of drug is subject to physician preference, formulary restrictions, and regional differences in patient cohort. However, the data reflect overall prescribing patterns in a sample of VA clinics and in the resulting blood pressure response. These results support the efficacy of a diuretic-based or β-blocker–based regimen in treating hypertension. The regimens with diuretic or β-blocker or with the two combined resulted in the lowest SBP and DBP (Tables 4⇑ and 5⇑), although these three regimens represent only 43% of the seven commonly used one- or two-drug regimens (Table 2⇑). Both diuretics and β-blockers are the only antihypertensive agents shown to reduce cardiovascular morbidity and mortality in hypertensive patients in randomized clinical studies. Whether the newer agents that are now widely prescribed will be just as effective in reducing cardiovascular disease is currently being tested in ongoing randomized clinical studies such as ALLHAT.12 In the interim, surveillance of prescribing patterns within the VA system emphasizes wide regional variation but confirms the effectiveness of diuretics and β-blockers as first-line agents in treating hypertension.
This study was supported by the Cooperative Studies Program of the Department of Veterans Affairs Medical Research Service (Chairman’s Office, St Louis, Mo: H. Mitchell Perry, Jr, Chairman; Sharon E. Carmody, Administrator; Mohinder P. Sambhi Consultant; Coordinating Center, Perry Point, Md: Joseph Collins, Director; Stephen Bingham, Biostatistician; Anne Horney, Statistician). Clinician members of the Department of Veterans Affairs Cooperative Study Group for the Hypertension Screening and Treatment Program were as follows: Indianapolis, Ind: J. Howard Pratt, Mary Jane Spencer; Iowa City: Sharlene Hartman, George Uhl, Mary Wall; Jackson, Miss: Richard Horne, Kent Kirchner, Martha Hilliburton, Betty Townsend; Memphis, Tenn: William Cushman, John J. Davis, Gale Rutan, Marion Taylor; St Louis, Mo: Karen Bobo, Greta H. Camel, Eleanor Janson, Cecilia Pectol, Kathleen Wenzel; San Juan, PR: Jose Ciancini, Jean DaMore, Eli Ramirez; Washington, DC: Grace Meyer.
Reprint requests to Dr Perry, Veterans Affairs Medical Center (111DJC), 915 N Grand Blvd, St Louis, MO 63106.
1 Eleven percent of the cohort (n=660) had only one visit during study follow-up. For 411 of them, the regimen was not changed at that visit and hence was the same as the regimen at the next-to-last visit. Because there is some uncertainty about the regimen at the next-to-last visit for the remaining 249, the regimen at the last visit was used for them.
2 For Iowa City and St Louis, the heavily used “newer” drug was a calcium antagonist; for Indianapolis it was an ACE inhibitor.
- Received January 9, 1997.
- Revision received February 14, 1997.
- Accepted October 9, 1997.
Carlson JH, Perry AG. Hypertension Specialist Training Program. 1979. US Government Printing Office 0–311-644/3253.
Perry HM Jr, Belton ED, Moses C, Moss WG, Page IR, Smith WM, Sokolow M, Witwer BD. National High Blood Pressure Education Program. Report to the Hypertension Information and Education Advisory Committee, Task Force I Data Base. 1973. US Dept of Health, Education, and Welfare Publication No. (NIH) 74–593.
Miller JP, Perry HM Jr, Rossiter JE, Baty JD, Carmody SE, Sambhi MP. Regional differences in mortality during 15-year follow-up of 11 936 hypertensive veterans. Hypertension. 1994;23:431–438.
Perry HM Jr, Carmody SE, Virgo K. Relation of ten-year mortality to early visit compliance and early treated diastolic pressure for 5522 hypertensive men. Hypertension. 1991;17:441. Abstract.
Perry HM, Miller JP, Fornoff JE, Baty JD, Sambhi MP, Rutan G, Moskowitz DW, Carmody SE. Pretreatment characteristics and early treated blood pressures associated with development of end-stage renal disease in hypertensive men during 15 years of follow-up. Hypertension. 1995;25:587–594.
Prineas RJ, Crow R, Blackburn H. Minnesota Code Manual of Electrocardiographic Findings. Littlejohn, MS: John Wright-PSG; 1982.
Davis BR, Cutler JA, Gordon DJ, Furberg CD, Wright JT, Cushman WC, Grimm RH, LaRosa J, Whelton PK, Perry HM, Alderman MH, Ford CE, Oparil S, Francis C, Proschan M, Pressel S, Black HR, Hawkins CM, for the ALLHAT Research Group. Rationale and design for the antihypertensive and lipid lowering treatment to prevent heart attack trial (ALLHAT). Am J Hypertens. 1996;9:342–360.
Materson BJ, Reda DJ, Cushman WC, Massie BM, Freis ED, Kochar MS, Hamburger RJ, Fye C, Lakshman R, Gottdiener J, Ramirez EA, Henderson WG, for the Department of Veterans Affairs Cooperative Study Group on Antihypertensive Agents. Single-drug therapy for hypertension in men: a comparison of six antihypertensive agents with placebo. N Engl J Med.. 1993;328:914–921.