[Skip to Content]
[Skip to Content Landing]
December 3, 1982

Propranolol After Myocardial Infarction-Reply

JAMA. 1982;248(21):2834. doi:10.1001/jama.1982.03330210025014

In Reply.—  The previous letters raise a very important question. In any study that demonstrates treatment benefit, for how long should that treatment be continued? Because studies do not go on indefinitely, there are no easy answers. Several points, however, can be made.First, the BHAT study was designed so that the entire period of patient follow-up, not a time segment, would be analyzed. Therefore, any conclusions based on time segments defined post hoc are risky. Furthermore, because of differential mortality in the first year, any analyses performed on time periods, eg, in the second or third years, may not be appropriate. For example, a number of presumably high-risk patients in the group receiving propranolol were saved in the first year. This most likely resulted in a patient population who received that treatment that was, on the average, more sick than that receiving the placebo at one year. Thus, only