Gastric ulcer is relatively infrequent, and clinical trails are often based on small-sized samples. The aim of this study was to define the "gold standard" therapy of active gastric ulcer. We included all single- or double-blind clinical trials on the short-term treatment of gastric ulcer. All the articles published over the period 1977-1994 were reviewed. Meta-analysis was done with both fixed and random effect models; results were shown using Galbraith's radial plot. Forty-eight papers comprising 52 studies were evaluated. Cimetidine, ranitidine, and famotidine proved significantly better than placebo [odds ratio (OR) and 95% confidence interval (CI 95%) at four to six weeks were: 2.67 (2.03-3.52), 3.94 (2.28-6.80), 1.76 (1.08-2.88), respectively]. Cimetidine and ranitidine had results comparable with the newer H2 blockers [OR (CI 95%) at four weeks: 1.16 (0.91-1.47), 1.11 (0.80-1.55), respectively]. H2 blockers were proved comparable with either sucralfate [OR (CI 95%) at eight weeks: 0.81 (0.37-1.79)] or bismuth [OR (CI 95%) at four to six weeks: 0.67 (0.37-1.20)]. Omeprazole is more effective than H2 blockers [OR (CI 95%) at four weeks: 2.00 (1.57-2.55)]. It is concluded that H2 blockers are preferred to either a placebo or sucralfate for short-term gastric ulcer treatment; the newer H2 blockers do not have significant advantages over the older types; omeprazole can be regarded as the "gold standard" for active gastric ulcer treatment.