Encouragement of subgroup assessment by the FDA.

Responses to a drug treatment can differ among subgroups of the population, including demographic subgroups, groups with different disease characteristics, and groups with different metabolic or excretory functions. The differences can be pharmacokinetic (PK) or pharmacodynamic (PD) affecting toxicity or effectiveness. In recent years wide availability of drug blood level data, including metabolite data, has made PK differences in subgroups very readily detectable, but PD differences are more difficult to detect and need to be carefully examined by assessment of effectiveness and toxicity in demographic and other subgroups, as illustrated by the forest plots examining multiple subgroups typically presented for cardiovascular outcome studies. Such examinations and presentations would probably be more broadly useful and informative if used to examine pooled outcome and toxicity data for symptomatic treatments.