Short-Term Treatment with Proton-Pump Inhibitors as a Test for Gastroesophageal Reflux Disease

A Meta-Analysis of Diagnostic Test Characteristics

6 April 2004 | Volume 140 Issue 7 | Pages 518-527

Background: A response to proton-pump inhibitors (PPIs) is commonly considered to support the diagnosis of gastroesophageal reflux disease (GERD). However, the accuracy of this diagnostic strategy has not been well established.

Objective: To estimate the diagnostic test characteristics of successful PPI treatment with objective measures of GERD by performing a meta-analysis based on the published literature.

Data Sources: English-language studies were identified by searching the Cochrane Clinical Trial Register and MEDLINE from 1 January 1980 through 1 July 2003.

Study Selection: Studies in which the clinical response to a short course (1 to 4 weeks) of normal- or high-dose PPI therapy could be compared with objective measures of GERD (24-hour pH monitoring, endoscopy findings, symptom questionnaires).

Data Extraction: Studies were screened for inclusion by 1 author. Final decisions on exclusion were made by consensus with 2 of the other authors. Two investigators independently extracted the data. Information extracted included patient characteristics, study design, setting, specific type and dose of medication, duration of treatment, and definitions of outcomes.

Data Synthesis: Sensitivity and specificity were determined by comparing a clinical response to PPIs with objective measures for GERD. The summary receiver-operating characteristic curve method was used to summarize test characteristics across studies. Sensitivity and specificity were also combined independently by using a random-effects model. Fifteen studies met the inclusion criteria and provided sufficient data. With 24-hour pH monitoring as the reference standard, the positive likelihood ratio ranged from 1.63 to 1.87, and combined estimates of sensitivity and specificity were 0.78 (95% CI, 0.66 to 0.86) and 0.54 (CI, 0.44 to 0.65), respectively. These values were lower with the other reference standards.

Limitations: Data were insufficient to determine the effect of various doses of PPIs and duration of therapy on test characteristics.

Conclusion: Successful short-term treatment with a PPI in patients suspected of having GERD does not confidently establish the diagnosis when GERD is defined by currently accepted reference standards.

Editors' Notes Context Does symptomatic response to a proton-pump inhibitor (PPI) diagnose gastroesophageal reflux disease (GERD)? Contribution This meta-analysis included 15 studies that compared clinical response to a short course of a PPI with an objective measure of GERD, such as 24-hour pH monitoring. The positive likelihood of a symptomatic response detecting GERD ranged from 1.63 to 1.87. Sensitivity was 0.78 (95% CI, 0.66 to 0.86) and specificity was 0.54 (CI, 0.44 to 0.65). Implications Symptomatic response to short-term treatment with a PPI does not confidently diagnose GERD. –The Editors

 

Author and Article Information

From University Medical Center Utrecht, Utrecht, the Netherlands, and Tufts-New England Medical Center, Boston, Massachusetts.

Acknowledgment: This work is part of the GERD-guidelines project of the European Society of Primary Care Gastroenterology (ESPCG).

Grant Support: Mattijs Numans, MD, PhD, was financially supported for this study in the United States by the Netherlands Organization for Scientific Research, the University Medical Centre, Utrecht, the Netherlands (Foundation Girard de Mielet van Coehoorn), and by unrestricted grants from the Dutch divisions of Altana Pharma, Jansen Cilag, and AstraZeneca. Joseph Lau, MD, is supported in part by a grant (R01 HS013328) from the Agency for Healthcare Research and Quality.

Potential Financial Conflicts of Interest:Stock ownership or options (other than mutual funds): J. Lau (Merck, Pfizer); Expert testimony: N.J. de Wit (Janssen Cilag); Grants received: M.E. Numans (AstraZeneca, Altana Pharma, Janssen Cilag), N.J. de Wit (Altana Pharma, Novartis, Janssen Cilag).

Requests for Single Reprints: Mattijs E. Numans, MD, PhD, Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, PO Box 85060, 3508 AB Utrecht, the Netherlands; e-mail, m.e.numans{at}med.uu.nl.

Current Author Addresses: Drs. Numans and de Wit: Julius Center for Health Sciences and Primary Care, University Medical Center Utrecht, PO Box 85060, 3508 AB Utrecht, the Netherlands.

Drs. Lau and Bonis: Center for Clinical Evidence Synthesis, Tufts-New England Medical Center, Kneeland Drive, Boston MA 02111.

Author Contributions: Conception and design: M.E. Numans, J. Lau, N.J. de Wit, P.A. Bonis.

Analysis and interpretation of the data: M.E. Numans, J. Lau, P.A. Bonis.

Drafting of the article: M.E. Numans, P.A. Bonis.

Critical revision of the article for important intellectual content: M.E. Numans, J. Lau, N.J. de Wit, P.A. Bonis.

Final approval of the article: M.E. Numans, J. Lau, P.A. Bonis.

Provision of study materials or patients: M.E. Numans.

Statistical expertise: M.E. Numans, J. Lau, P.A. Bonis.

Obtaining of funding: M.E. Numans, N.J. de Wit.

Administrative, technical, or logistic support: J. Lau.

Collection and assembly of data: M.E. Numans, P.A. Bonis.

Read all Rapid Responses

Is diagnosis all we need?

Agostino Colli, et al.

Annals Online, 20 Jul 2004 [Full text]