What do subgroup analyses reveal about differential response to beta-blocker therapy? The Beta-Blocker Heart Attack Trial experience.

Results based on small subgroups are prone to much random variation and firm conclusions based on subgroup analyses should be avoided. Caution is advisable, especially with regard to post hoc and multiple analyses. Subgroup findings for which a reasonable biologic explanation can be given and those supported by independent results from other subgroups within the trial are more likely to be real. The strongest support for a subgroup finding in one trial is a replication from another trial. Post hoc analyses from the Beta-Blocker Heart Attack Trial suggest, and are supported by the Norwegian Timolol Study, that myocardial infarction patients older than age 60 years and those with arrhythmic and other complications not severe enough to preclude treatment benefit the most in absolute terms from beta blockers.