The aim of this study was to determine the main diagnostic validity parameters of a quantitative real-time polymerase chain reaction (PCR) system for detecting Helicobacter pylori in gastric biopsies. Prospective study. The real-time PCR has an internal control for eliminating the false negatives. Our system has a good diagnostic capacity compared with the gold standard and was superior in antral mucosa: area under the curve was 0.91 for antrum (95% confidence interval [CI] 0.87 to 0.96) and 0.83 for corpus (95% CI 0.77 to 0.9). The optimum cut-off point was 3.56 microorganisms/cell for antrum (sensitivity 83.5% [95% CI 74.2 to 89.9]; specificity 91.3% [95% CI 82.3 to 96.0]; positive predictive value 92.2%; negative predictive value 81.8%). The positive likelihood ratios were 9.61 and 8.52 for antrum and corpus, respectively. With the cut-off point that maximises the Youden index, 8.7% false positives were obtained. Our methodology is useful for diagnosing infection due to H. pylori and the false positives detected probably correspond to patients who were actually infected but the infection was not detected by traditional techniques. The clinical importance of these cases should be studied in greater detail since they may involve colonisations unrelated to the patient's digestive pathology.
Copyright © 2012 Elsevier Inc. All rights reserved.