Pentoxifylline has been shown to improve treadmill walking distances under blinded, controlled conditions in patients with intermittent claudication. From the pooled data of a blinded, controlled, randomized, multicenter trial, the data from all enrolled patients with severe claudication (less than 70 m on treadmill at baseline) were evaluated. The treadmill data from these more severely ill patients were analyzed separately as a "severe subset" (placebo n = 17; pentoxifylline n = 21). No differences between the two treatment groups were observed in demography, history, or baseline treadmill walking distances. The initial claudication distance (ICD) improved 68% over baseline with pentoxifylline and 12% with placebo (p = .012) after twenty-four weeks of treatment. A new, derived efficacy variable was developed, "minimum distance walked," which tended to minimize psychological effects on treadmill performance. Over sixteen to twenty-four weeks of treatment, the pentoxifylline group improved 49% over baseline and the placebo group 3% (p = .019), when the "minimum distance walked" measurement was used. In this controlled trial the subset of patients with severe intermittent claudication benefited from pentoxifylline therapy.