Background: To compare the effectiveness of defined-formula diets versus steroids for the treatment of active Crohn's disease, we conducted a meta-analysis of the published clinical trials. Methods: Standard techniques for literature search were used to identify the pertinent trials. All studies included in our meta-analysis (n = 7) were aimed at comparing defined-formula diets versus steroids, using a randomized design. The patient-specific end-point of the meta-analysis was the occurrence of a treatment failure. Results: Our meta-analysis indicated that steroids are more effective than defined-formula diets for inducing remission in active Crohn's disease. In fact, the relative risk of treatment failure (RTF) was significantly lower in the steroid group than in the diet group (risk values for patients given steroids compared with patients given diet: a) method of Mantel-Haenszel: RTF = 0.35; 95% confidence interval, 0.23-0.53; p < 0.001; b) method of Der Simonian & Laird: RTF = 0.43; 95% confidence interval, <0-0.94; p = 0.03). A separate analysis was carried out in which only the subgroup of patients who were not intolerant to diet were evaluated; this analysis also showed a superiority of steroids over diet. Conclusions: The data examined in this meta-analysis do not support the use of diets as primary treatment for acute exacerbations of Crohn's disease in adults.