Abstract
BACKGROUND: Proton pump inhibitors (PPIs) are widely used for the long-term management of gastroesophageal reflux disease (GERD). However, concerns about the cost and/or inconvenience of continuous maintenance PPI treatment have led to the evaluation of various alternative approaches. AIM: To assess the effectiveness of on-demand PPI therapy in the maintenance treatment of nonerosive reflux disease (NERD) or mild erosive esophagitis (EE). METHODS: We searched MEDLINE, EMBASE, Web of Science, and Cochrane Library from inception until October 2, 2017, for randomized controlled trials (RCTs) comparing on-demand PPI versus placebo or daily PPI in the management of NERD or mild EE (Savary-Miller grade 1). Discontinuation of therapy during the trial was used as a surrogate for patient dissatisfaction and failure of symptomatic control. We calculated pooled odds ratios (OR) to evaluate the efficacy of on-demand PPI treatment. Separate analyses were conducted for studies comparing on-demand PPI with daily PPI and with placebo. Subgroup analysis was done based on NERD studies alone and on studies of both NERD and mild EE. These were analyzed using a random effects model. RESULTS: We included 10 RCTs with 4574 patients. On-demand PPI was superior to daily PPI (pooled OR = 0.50; 95% confidence interval (CI) = 0.35, 0.72). On subgroup analysis in NERD patients only, pooled OR was 0.44 (0.29, 0.66). In studies including patients with NERD and mild EE, pooled OR was 0.76 (0.36, 1.60). For studies comparing on-demand PPI with placebo, pooled OR was 0.21 (0.15, 0.29); subgroup analyses of studies evaluating NERD only and studies conducted in NERD and mild EE showed similar results (pooled OR was 0.22 (0.13, 0.36) and 0.18 (0.11, 0.31), resp.). CONCLUSIONS: On-demand PPI treatment is effective for many patients with NERD or mild EE. Although not FDA-approved, it may be adequate for those patients whose symptoms are controlled to their satisfaction.