Twelve studies were eligible for inclusion: 9 studies (ITT, n=677; PP, n=842) compared PPI versus RBC plus clarithromycin and amoxycillin, and 5 studies (ITT, n=493; PP, n=550) compared PPI versus RBC plus clarithromycin and a nitroimidazole.
PPI versus RBC plus clarithromycin and amoxycillin. Mean H. pylori eradication with RBC was 76.6% (95% confidence interval, CI: 72, 81) by ITT and 83.3% (95% CI: 79, 86) by PP analysis. With a PPI, the mean H. pylori eradication rates were 73.7% (95% CI: 69, 78) and 81% (95% CI: 77, 85), respectively. The corresponding ORs were 1.15 (95% CI: 0.80, 1.64) by ITT, and 1.08 (95% CI: 0.75, 1.55) on a PP basis. The chi-squared homogeneity test was 12.2 (p=0.1) by ITT and 14.3 (p=0.07) by PP analysis; there was evidence of significant heterogeneity in both cases.
PPI versus RBC plus clarithromycin and a nitroimidazole. Mean H. pylori eradication with RBC was 87.2% (95% CI: 83, 91) by ITT and 93.5% (95% CI: 91, 96) by PP analysis. With a PPI, the mean eradication rates were 79.4% (95% CI: 74, 84) and 85% (95% CI: 81, 89) by ITT and PP analysis, respectively. The corresponding ORs were 1.76 (95% CI: 1.08, 2.85) by ITT, and 2.43 (95% CI: 1.42, 4.18) on a PP basis. The chi-squared homogeneity test was 6.24 (p=0.18) by ITT and 1.6 (p=0.8) by PP analysis.
The overall efficacy of the two regimens was not shown to be different by either ITT (OR 1.37, 95% CI: 0.93, 2.05) or PP (OR 1.37, 95% CI: 0.88, 2.02) analysis. However, RBC-clarithromycin-nitroimidazole combinations were more effective than RBC-clarithromycin-amoxycillin regimens, both by ITT (87.2 versus 76.6%; OR 2.08, 95% CI: 1.33, 3.26) and PP (93.5 versus 83.3%; OR 2.88, 95% CI: 1.67, 4.97).