We sought to develop an enrichment crossover study design that would allow us to efficiently evaluate and compare promising candidate neuropathic pain drugs. We evaluated the efficacy of gabapentin or tramadol vs. active placebo (diphenhydramine) in subjects with biopsy-proven painful idiopathic small fiber neuropathy (SFN) who were self-reported gabapentin responders. Eligible subjects entered two single blind run-in phases. In the first phase (Period A), subjects were treated with single blinded gabapentin at their prestudy dose followed by a second run-in phase (Period B) in which they were treated with diphenhydramine active placebo. Subjects with >or=3 pain and a >or=30% increase in pain intensity in Period B compared to Period A were then randomized to a double-blind three period cross over trial of gabapentin at pre study dosage, tramadol 50mg QID and diphenhydramine 50mgqhs. Of the 59 subjects enrolled, 41 subjects were excluded: Twenty-three had an insufficient rise in pain intensity in Period B; eight had skin biopsies that did not confirm SFN. Eighteen subjects were randomized into the double-blind, crossover phase. There was a significant treatment effect of gabapentin vs. diphenhydramine (p=0.001) and tramadol vs. diphenhydramine (p=0.018) by the before-bed daily pain score averaged over the final 7 days of each treatment period. We conclude that gabapentin and tramadol were effective in the treatment of painful SFN and that this experimental enrichment paradigm is attractive to screen potential neuropathic pain compounds for efficacy in proof-of-concept studies.