To assess (i) the efficacy of short-term proton pump inhibitors (PPIs) in non-cardiac chest pain (NCCP) and (ii) the performance of an empirical short-term treatment with PPI (PPI test) to establish a diagnosis of abnormal acid reflux in NCCP. Metaanalysis of English language studies identified by searching MEDLINE (1966-May 2004), EMBASE (1980-May 2004), Cochrane Controlled Trials Register, and abstract books from major gastroenterology meetings (1993-2004). For the metaanalysis of PPI efficacy in NCCP, we selected randomized controlled trials (parallel group and crossover designs) comparing PPI therapy with placebo. For the metaanalysis of PPI test performance, we selected uncontrolled studies comparing the test with a standard reference. Eight studies were included in the PPI efficacy analysis. The pooled risk ratio for continued chest pain after PPI therapy was 0.54 (95% CI 0.41-0.71). The overall number needed to treat was 3 (95% CI 2-4). The pooled sensitivity, specificity, and diagnostic odds ratio for the PPI test versus 24-h pH monitoring and endoscopy were 80%, 74%, and 13.83 (95% CI 5.48-34.91), respectively. All studies were small and there was evidence of publication bias or other small study effects. PPI therapy reduces symptoms in NCCP and may be useful as a diagnostic test in identifying abnormal esophageal acid reflux.