The identification of effective treatments for uremic pruritus used a literature search of MEDLINE (1966 to March 1991), Index Medicus (1978 to March 1991), bibliographic reviews of textbooks, and pertinent articles. Only randomized controlled clinical trials were selected for analysis. Assessment of study quality was performed independently by two observers with specific methodologic criteria. A clinically significant therapeutic effect was defined as a statistically significant result in the proportion of patients improving and a reduction in pruritus score of at least 50%. Clinically significant outcomes were obtained for two of three whole-body ultraviolet B trials. Meta-analysis of the ultraviolet B trials retained the significant effect in analysis of proportions (pooled odds ratio 18; 95% confidence interval 4, 161). Trials of lidocaine, charcoal, and nicergoline demonstrated either statistically significant improvement in pruritus score or in proportions, but not both. Ultraviolet B phototherapy is the treatment of choice in moderate to severe uremic pruritus.