Treatment of variceal hemorrhage is one of the most controversial subjects in medicine. The resurgence of old therapies (endoscopic sclerotherapy) and the introduction of new modalities (obliterative angiotherapy) has exacerbated the controversy. No widely accepted controlled therapeutic trial is available. The problem related to survival analysis has been studied in the light of the available information concerning the natural history of variceal bleeding. It is believed that controlled trials can be designed which will prove the efficacy, or lack of it, for any proposed treatment; however, valid conclusions based on previous studies are limited, largely because of the many confounding variables. Time, as a variable factor both for randomization and therapeutic intervention, has been largely ignored, yet, we believe it is the major variable in this setting. For the population of variceal bleeders, risk of rebleeding or death rapidly diminishes over the first few days after a bleed, and early survival may be the best marker for later survival. Neither presentation nor treatment seems to alter this fundamental behavior. Variceal hemorrhage may serve as a prototype for problems of survival analysis of diseases with early high mortality.