Knowing why patients fail to complete a clinical trial should help in planning such trials and in estimating their costs. We analyzed reasons why subjects failed to complete two similar acute stroke trials of calcium-channel antagonists. The trials were double-blind, placebo-controlled, and lasted 2 and 3 weeks, respectively. We enrolled 43 patients in the first trial and 37% were withdrawn; 30 patients were enrolled in the second trial and 47% were withdrawn. Reasons for withdrawal were similar in the two trials. When data from the two trials were combined, reasons for withdrawal were errors in administering medication (11%), use of a proscribed drug (7%), early death (7%), a disqualifying illness discovered after enrollment (8%), withdrawal of consent (4%), and enrollment beyond the time limit (<1%). Presence of stroke risk factors, age, sex, and even stroke severity did not significantly increase withdrawal frequency. Since most patients were lost early in the trial, its length did not contribute substantially to premature withdrawal.