The Pharmaceutical Journal
Vol 275 No 7369 p402
1 October 2005

This article
Reprint   Photocopy

PDF 77K, Acrobat Reader

PPIs effective for diagnosis and treatment in non-cardiac chest pain

POEM series

Clinical question How accurate is a trial of a proton pump inhibitor for the diagnosis of acid reflux and how effective are PPIs for the treatment of non-cardiac chest pain?

Bottom line A PPI is useful in the diagnosis of gastro-oesophageal reflux disease (GORD) and an effective treatment for patients with non-cardiac chest pain. Because some smaller studies with negative results may not have been published, the estimate of the degree of benefit of PPIs in this study may be on the high side.

Synopsis This was actually two studies in one and it did a good job of clarifying the accuracy and extent of benefit of PPIs in the diagnosis and treatment of non-cardiac chest pain. The authors performed an adequate search of the literature, including not only MEDLINE, but Embase, the Cochrane Controlled Trials Register, and a hand search of meeting abstracts. For the meta-analysis of the diagnosis of GORD using PPIs, acid reflux was diagnosed using 24-hour pH monitoring as the reference standard test and heart disease was excluded using appropriate tests (eight studies with 321 patients). Studies were not heterogeneous, and the pooled sensitivities and specificities for the outcome of "greater than 50 per cent response" were 83 per cent and 75 per cent. Given the overall prevalence of oesophagitis using the reference standard test of 22 per cent, this corresponds to positive and negative predictive values of 48 per cent and 94 per cent, respectively. The sensitivity was much lower (46 per cent) when the outcome used was “any response”. For the meta-analysis of treatment of non-cardiac chest pain, only randomised trials that blinded patients to treatment assignment, used intention-to-treat analysis, and had a placebo control group were included (seven studies with 232 patients). Using the outcome of “greater than 50 per cent response” as the definition of success, the number needed to treat for PPIs in patients with non-cardiac chest pain was three (pooled risk ratio=0.54; 95 per cent confidence interval, 0.41–0.71). There was no significant benefit if the stricter criterion of “complete resolution” of chest pain was used. There was evidence of publication bias, with an absence of small studies that showed less benefit.

Level of evidence 1a (systematic review of randomised controlled trials).

Reference Cremonini F, Wise J, Moayyedi P, Talley N. Diagnostic and therapeutic use of proton pump inhibitors in non-cardiac chest pain. American Journal of Gastroenterology 2005;100:1226–32.

POEM (Patient Oriented Evidence that Matters) is a registered trademark of InfoPOEMs
© infoPOEMs 1995–2004