N-acetylcysteine has been recommended for patients with renal insufficiency who are to receive radiocontrast media. However, trials of oral N-acetylcysteine for the prevention of radiocontrast-induced nephropathy have yielded inconsistent results. A systematic review of patient and study characteristics was undertaken to discover possible explanations of the inconsistencies. The databases MEDLINE, EMBASE, and CENTRAL (1966 to March 2003) were searched in all languages, and conference proceedings from several professional societies from the years 1999 to 2003 were also searched. Only prospective controlled trials of oral N-acetylcysteine were included. Risk difference estimates and 95% confidence intervals were calculated. The estimates were examined for evidence of publication bias and heterogeneity. Stratified and meta-regression analyses were used to compare estimates by study and patient characteristics. Identified were 16 studies, 15 published and 1 unpublished. There was no evidence of publication bias, but there was substantial evidence of heterogeneity, thus precluding reliance on a meaningful summary effect estimate. Meta-regression identified several patient and study characteristics, with some evidence of association with study-specific estimates. None of these characteristics, however, formed subsets of studies with results that were homogeneous enough to aggregate. Research on N-acetylcysteine and the incidence of radiocontrast nephropathy is too inconsistent at present to warrant a conclusion on efficacy or a recommendation for its routine use. Identified patient and study characteristics may be responsible for some, but not all, of this inconsistency. A large, randomized, placebo-controlled trial, a pooled analysis of patient-level data, or both may resolve this issue.