| The ALLHAT investigators believe... the take-home message is... that doctors should begin drug treatment for hypertension with a diuretic [but] acknowledge that most patients will need more than 1 drug to adequately control their blood pressure. |
The major results of ALLHAT, the largest hypertension clinical trial ever conducted, show that the thiazide diuretics are more effective than angiotensin converting enzyme (ACE) inhibitors or calcium channel blockers (CCBs) with regard to blood pressure lowering, prevention of hypertension-related complications, and side effects. Consequently, the trial investigators recommend that, for most people who are diagnosed with hypertension, drug treatment should begin with a diuretic. The investigators also commend the use of these diuretics for their possible cost savings compared with the newer, more expensive antihypertensive agents from other drug classes.
The eagerly anticipated findings from ALLHAT were revealed on December 17th, in Washington, DC, at a press conference organized by the National Heart, Lung, and Blood Institute (NHLBI), which supported the study. Individual results from the antihypertensive and lipid components of the study were published in 2 respective articles in the December 18, 2002, issue of JAMA.
"ALLHAT shows that diuretics are the best choice to treat hypertension and reduce the risk of its complications, both medically and economically," said NHLBI Director Dr. Claude Lenfant. "Many of the newer drugs were approved because they reduce blood pressure and the risk of heart disease compared with a placebo," he continued, "but they were not tested against each other. Yet, these more costly medications were often promoted as having advantages over older drugs, which contributed to the rapid escalation of their use. Now, at last, we can make those needed comparisons and know which blood pressure drug to choose to begin therapy."
The ALLHAT investigators point out that between 1982 and 1992, diuretic use fell from 56% to 27% of antihypertensive prescriptions, and that had this decrease not occurred, antihypertensive drug costs for that period would have been lower by about US $3.1 billion.
Within hours of announcement of the ALLHAT results, widely divergent opinions about their implications were being expressed in the medical community, ranging from predictions that they would completely change medical practice in hypertension to rejection of the findings based on criticisms of the study design. Comments are still coming in, and many future reviews and comments on the trial results are expected in the medical press, including Medscape Cardiology.
The ALLHAT blood pressure study was a randomized, double-blind, active-controlled clinical trial that compared a thiazide diuretic with newer antihypertensive drugs as first-choice blood pressure lowering treatment. The primary outcome of the study was combined fatal coronary heart disease (CHD) or nonfatal myocardial infarction (MI). Secondary outcomes were all-cause mortality, stroke, combined CHD, and combined cardiovascular disease.
Enrollment of high-risk hypertensive patients aged ≥ 55 years began in February 1994, and follow-up was completed in March 2002 at 623 centers in Canada, Puerto Rico, the United States, and the US Virgin Islands. All participants underwent medical checkups at 3, 6, 9, and 12 months after entry into the study and every 4 months thereafter.
The original trial population comprised 42,418 patients with SBP ≥ 140 mm Hg and/or DBP ≥ 90 mm Hg or who were taking antihypertensive medication, with at least 1 other CHD risk factor, including cigarette smoking or type 2 diabetes.
Patients were randomized to receive 1 of 4 antihypertensive agents:
Chlorthalidone (diuretic),
Lisinopril (ACE inhibitor),
Amlodipine (CCB), or
Doxazosin (alpha-adrenergic blocker).
The doxazosin arm of the study was stopped in March 2000 because of a 25% higher rate of combined cardiovascular events and a 2-fold higher rate of heart failure compared with chlorthalidone.
The remaining 33,357 patients stayed on their study drugs through the end of the study, for an average follow-up of 4.9 years. Of this population:
15,255 were randomized to chlorthalidone (12.5-25.0 mg/day),
9048 were randomized to amlodipine (2.5-10.0 mg/day),
9054 were randomized to lisinopril (10-40 mg/day).
At the discretion of the patients' physicians, doses of the study drugs were optimized and then additional drugs from other classes (other than the study drugs) were prescribed as needed to achieve blood pressure control.
The mean age of this population was 67 years, and the mean body mass index (BMI) was 30. Forty-seven percent of the study patients were women, 35% were black, and 19% were Hispanic. Thirty-six percent had diabetes, 90% had been on prior antihypertensive drug treatment, 22% were smokers, and 3% had left ventricular hypertrophy by ECG.
After about 5 years of follow-up, there were no significant differences in primary outcome, fatal CHD, or nonfatal MI among the amlodipine, lisinopril, and chlorthalidone groups. However, there were differences in secondary outcomes for both of the newer drugs compared with the diuretic.
Compared with participants who were taking the diuretic, those on amlodipine had:
On average, about a 1 mm Hg higher SBP
38% higher risk of developing heart failure (P < .001) (6-year absolute risk, 2.5%)
35% higher risk of hospitalization for fatal heart failure (P < .001).
Importantly, however, with a null difference in primary outcome, the safety of the CCB was reaffirmed in this very large sample of patients, in contradistinction to some previous controversy in this regard.
Compared with participants who were taking the diuretic, those on lisinopril had:
On average, about a 2 mm Hg higher SBP (4 mm Hg higher in blacks)
15% higher risk of stroke (P < .02); 40% higher risk for blacks (P = .01)
10% higher risk of combined cardiovascular disease (P < .001) (6-year absolute risk, 2.4%); 19% higher risk in blacks (P = .04)
19% higher risk of developing heart failure (P < .001)
10% higher risk of hospitalization for fatal heart failure (P = .11)
11% higher risk of hospitalization for treated angina (P = .01)
10% higher risk of coronary revascularization (P =.05)
Side effects with chlorthalidone were, as expected, hypokalemia, slightly higher mean serum cholesterol and glucose levels, and a slightly high incidence of new cases of diabetes at 4 years' follow-up. However, the investigators stated that these side effects did not result in more adverse clinical outcomes when patients' serum potassium was monitored and when potassium supplements were provided where necessary. Some creatinine-based measures of renal function were more favorable in the amlodipine group, although rates of end-stage renal disease were not significantly different compared with the chlorthalidone group.
The ALLHAT investigators believe that the take-home message from these results is that doctors should begin drug treatment for hypertension with a diuretic, and that for those few patients who cannot tolerate a diuretic, a beta-blocker, an ACE-inhibitor, or a CCB may be used to start treatment. The investigators also acknowledge that most patients will need more than 1 drug to adequately control their blood pressure (63% of patients in ALLHAT were on ≥ 2 drugs by the end of the study), but insist that 1 of the drugs used should be a diuretic. However, as emphasized by Jackson T. Wright Jr, MD, PhD (Professor of Medicine Case Western Reserve University, Cleveland, Ohio) in answer to a question at the press conference, ALLHAT was not designed to investigate which class of antihypertensive drug should be the preferred second-line treatment in any group of patients.
They recommend that patients who are now on a CCB or an ACE inhibitor or another hypertension drug other than a diuretic should not stop taking their medication, but that they should talk with their physician about adding or switching to a diuretic for their treatment if they are not already on one.
The investigators believe that the ALLHAT's findings refine the current US clinical guidelines that recommend starting therapy for hypertension with a diuretic or a beta-blocker. The Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC-VI), made available by the NHLBI in 1997, recommends a diuretic or a beta-blocker for initial treatment of uncomplicated hypertension. A new committee has just been appointed to prepare the Seventh Report (JNC-VII), Dr. Lenfant announced, and it is expected that the ALLHAT findings will be incorporated. Dr. Lenfant further told Medscape that it is hoped that JNC-VII can be released during the next meeting of the American Society of Hypertension, to be held May 2003 in New York City.
The first comments on the study appeared alongside the ALLHAT results in JAMA, in an editorial by Lawrence K. Appel, MD, MPH (Johns Hopkins University, Baltimore, Maryland). He queried the increased risk of heart failure associated with lisinopril in the trial (statistically significant for total heart failure, although nonsignificant for hospitalization for or death from heart failure), noting that this result is difficult to reconcile with the well-documented benefits of ACE inhibitors in heart failure, such as shown in the Studies Of Left Ventricular Dysfunction (SOLVD) and the Heart Outcomes Prevention Evaluation (HOPE). He also pointed out that chlorthalidone, although frequently used in clinical trials, is not commonly used in clinical practice in the United States, and that other thiazide diuretics may not be comparable, although they might be reasonably considered to be so. However, he concluded that despite the "striking and unequivocal null result" of the trial with regard to the primary outcome, the results provide compelling evidence that thiazide diuretics should be the first choice for patients with hypertension, and that they may even be more effective than beta-blockers.
ALLHAT also included a cholesterol-lowering study that compared the effects of a statin drug with "usual care." Both groups had a substantial decrease in cholesterol levels. The difference in cholesterol levels between the groups was too small to show a difference in death rates and produced only a small, nonsignificant decrease in the rates of heart attack and stroke in the statin group.
This trial was the first to be performed exclusively in patients with high blood pressure and involved 10,355 of the hypertension trial's participants. At the start of the trial, patients were included if they had moderately elevated blood cholesterol but were judged by their physicians not to need cholesterol-lowering medication. All had at least 1 heart disease risk factor in addition to high blood pressure and elevated cholesterol. About 49% were women, 38% black, and 23% Hispanic. About 35% had type 2 diabetes, and approximately 14% had heart diseaseat baseline. They were followed for an average of 4.8 years. Participants were assigned to receive either pravastatin or usual care, which, at the start of the study, involved no cholesterol-lowering drug. Both groups followed a cholesterol-lowering diet. The study was not blinded, and participants and their healthcare providers knew what treatment was administered.
Pravastatin was chosen for the trial because it had been shown in prior studies to safely yield long-term total cholesterol reductions of 20% or more, an improvement necessary to gauge the therapy's effects on heart disease and overall deaths in ALLHAT's relatively short span (average of 5 years).
During the trial, those in the usual care group were prescribed a cholesterol-lowering drug (not provided by the study) when their doctor felt it was warranted by changes in their condition, such as a heart attack or marked cholesterol increase. Of those in the usual care group, 32% who had heart disease at the start of the study and 29% of those without heart disease at the outset used a cholesterol-lowering drug.
After 4 years, both the statin and usual care groups had reductions in total and low-density lipoprotein (LDL) cholesterol. Total cholesterol dropped by 17.2% in the pravastatin group and 7.6% in the usual care group, a modest 9.6% difference. This was probably because many of those in the usual care group received a cholesterol-lowering drug, reflecting the trend toward increasing use of these drugs in usual care during the 8 years of the trial.
There were no significant differences between the pravastatin and usual care groups with regard to overall mortality or mortality from any single cause. There were 631 deaths reported in the pravastatin group and 641 in the usual care group.
There also were only modest, nonsignificant differences in the rates of fatal and nonfatal heart attack or stroke between the pravastatin and usual care groups. There were 380 CHD events and 209 strokes in the pravastatin group, compared with 421 heart disease events and 231 strokes in the usual care group.
Rates of heart failure and cancer also were similar in the 2 groups. Results for death rates did not differ by age, gender, race, or the presence of type 2 diabetes. The ALLHAT findings were consistent with the results of other statin trials in which larger reductions in total and LDL cholesterol were associated with proportionally greater reductions in the rates of heart attack and mortality, and the results were consistent with the current national guidelines that stress the need to aggressively lower high cholesterol levels.