Review of the Beta Blocker Heart Attack Trial (BHAT)

JAMA 1982; 247:1707-14

Background Beta blockers were commonly used in the treatment of coronary artery disease by the mid-1970s, predominately for symptomatic relief of angina. Lower quality evidence was also emerging that they reduced myocardial oxygen demand, could limit infarct size and improve survival in patients suffering from an acute myocardial infarction. National Heart, Lung, and Blood Institute (NHLBI) sponsored the beta blocker heart attack trial (BHAT) trial. It sought to test the hypothesis that oral administration of the beta blocker propranolol to patients hospitalized with an acute heart attack would reduce death over a 2 to 4 year follow up period.

Patients Patients were recruited at 31 centers with 134 participating hospitals. Eligible participants were men and women between the ages of 30 through 69 who were hospitalized with an acute MI documented by appropriate symptoms, ECG changes, and enzymatic changes. Patients were excluded if; they had a medical contraindication to propranolol such as marked sinus bradycardia or severe congestive heart failure or asthma as an adult, a life-threatening illness other than CHD, had or were likely to undergo cardiac surgery or were already taking a beta blocker or were likely to have beta blockers prescribed to them.

Baseline characteristics The number of patients who had an MI that met study criteria, were age eligible and survived at least 5 days after admission was 16,400 and 3,837 (23%) were ultimately enrolled. The cohort was composed primarily of young (average age 55 years), white (89%), men (84%) and 57% were current smokers. At the time of enrollment, the average BP was 112/72 mmHg and heart rate was 76 beats per minute. Only 14% developed heart failure symptoms prior to randomization and 17% were on a diuretic at the time of randomization, 23% experienced ventricular tachycardia. The mean number of days in the hospital prior to randomization was 14. Patients with anterior and inferior MIs were nearly equally represented in the cohort (30%) and non-transmural MIs accounted for 23%.

Procedures Patients were randomized 5 to 21 days after hospital admission and while still hospitalized. A regimen of assigned study medication was begun immediately after randomization with either 20 mg of propranolol or matching placebo. If no adverse reactions were noted, the dose was increased to 40 mg of propranolol or placebo every 8 hours. Patients then had to be monitored for a minimum of 6 consecutive doses (at least 2 days) on the 40 mg schedule and then 8 hours after the last dose, blood samples were drawn to test blood propranolol levels. Based on the level, patients were prescribed either 60 or 80 mg three times a day for either 180 or 240 mg/day, respectively.

Patients were asked to report to their coordinating centers at 4 weeks and 6 weeks and then every 3 months thereafter. At each visit, adherence to study drug, side effects, health status, the use of non-study medication, and occurrence of morbid events were monitored and additional study medication was dispensed. The clinic physician, at his or her discretion, could reduce the prescribed dose of medication. Furthermore, patients who were prescribed non-study beta blockers had those medications withdrawn.

Endpoints All-cause mortality, cardiovascular mortality, death due to arteriosclerotic heart disease, sudden death and non-sudden death.

Results The mean follow-up was 2.1 years. At the time of last study visit, approximately 85% of the propranolol group and 13% of the placebo group were taking a beta blocker (57% in the propranolol group were receiving a full protocol dose). Discontinuation of propranolol compared to placebo was relatively uncommon and most reported side effects were evenly distributed. Several complaints were statistically greater in the propranolol group but the difference is small from a clinical standpoint. 

There were 188 deaths in placebo group (10%; 5 per 100 patient years) and 138 in the propranolol group (7%; 3 per 100 patient years) and this was statistically significant (P<.005). Statistically significant differences were also observed for cardiovascular death and death due to arteriosclerotic heart disease overall and for the component of sudden death but not non-sudden death; however, a numerical reduction was observed that was consistent with overall observed mortality reductions. Only a small number of patients (<1% in each group) died from non-cardiovascular causes and the difference was not significant. The results of 19 subgroup analyses are presented and evidence of an interaction is observed for only those with nontransmural MIs, who did not appear to benefit from propranolol.

At one year follow-up, the mean heart rates and blood pressure in the propranolol and placebo groups were 65 vs 73 bpm and 127/80 vs 130/81 mmHg, respectively. These results are based on the intention to treat analysis (ITT).

Conclusions

Propranolol modestly reduced death in patients hospitalized with an MI. Differences in side effects were small and do not appear clinically relevant with the exception of ventricular arrhythmias, which were more common in the placebo group and associated with a higher rate of antiarrhythmic drug use during the trial.

The internal validity of the study is high but the external validity is limited by the fact that administration of the study drug would be challenging to replicate in clinical practice due to the intensity of monitoring and follow up.

Furthermore, the population is highly selected. Only a minority of heart attack patients who met study criteria were enrolled and another large group of patients (those >70) were not eligible for consideration.

Finally, results from this study provide no guidance on the management of heart attack patients in the inpatient setting between 0 to 5 days and it’s ultimately unclear what made someone become eligible for enrollment after an extended hospitalization. 

Tomorrow morning we will publish our review on the ISIS-1 trial that tested intravenous atenolol in patients with acute MI and tomorrow afternoon will be our first podcast featuring discussion of the BHAT and ISIS-1 trials.