Loading...
Clinicians sometimes diagnose gastroesophageal reflux disease by evaluating symptom response to empirical therapy with a proton-pump inhibitor (PPI), but is this so-called “PPI test” reliable?
To evaluate the test’s usefulness in primary care settings, Dutch investigators studied 74 adults with symptoms suggestive of GERD; all were recruited through primary care practices or advertisements. At baseline, each participant underwent a 24-hour pH study to determine the likelihood of a temporal association between symptoms and reflux. If the likelihood of an association was greater than 95%, the patient was considered to have GERD.
Each study participant then underwent a PPI test (receiving esomeprazole, 40 mg daily) for 13 days. If the subject reported overall symptom relief, the test was considered positive for GERD. Researchers then calculated the sensitivity and specificity of the PPI test on each of the 13 days, using the pH study results as the gold standard. The pH results were not revealed to the patient or treating physician until after completion of the PPI test. The study was supported by the maker of esomeprazole.
The results of the pH studies indicated that 70% of the study population had GERD. The mean sensitivity of the PPI test during the 13-day treatment period was 91% (95% CI, 78%–96%), but the mean specificity was only 26% (95% CI, 10%–49%). The positive likelihood ratio for the test was 1.2 (95% CI, 0.9–1.6) and varied little during the treatment period.
Aanen MC et al. Diagnostic value of the proton pump inhibitor test for gastro-oesophageal reflux disease in primary care. Aliment Pharmacol Ther 2006 Nov 1; 24:1377-84.
Comment
Sensitivity and specificity are relatively poor measures of the value of a diagnostic test, because they vary with disease prevalence in the population. Obviously, any test for GERD in a population of patients selected for symptoms would have high sensitivity and low specificity. A much more robust measure of the value of a test is the likelihood ratio, which measures whether a positive test influences the odds that a patient has the disease. In this study, the positive likelihood ratio was constant at 1.2 throughout 13 days, suggesting that the PPI test was of very little value in distinguishing patients who had GERD from those who did not in this selected population.