BackgroundThe Johnson–DeMeester composite score (DMS) is the historical gold standard for diagnosing gastroesophageal reflux disease (GERD). The Lyon Consensus outlines criteria for diagnosing GERD by pH monitoring, defining normal acid exposure time (AET) as < 4% and pathological as > 6%, presenting diagnostic uncertainty from 4 to 6%. We aimed to (i) calculate the proportion of borderline studies defined by total AET alone that are reclassified as normal or pathological by the DMS, (ii) determine the importance of supine AET for reclassification, and (iii) propose a new classification system using a composite score that considers positional changes.MethodsThis single-center, retrospective, observational study analyzed data from patients with an overall total AET from 2 to 6% on 48-h pH monitoring (Bravo pH capsule). Preselected predictors (supine and upright AET) were included in a model to create a composite score (i.e., pHoenix score) using the regression coefficients. The model was internally validated, and discriminative ability was tested against the DMS and compared to the total AET.ResultsWe identified 114 patients (80 [70.2%] women; median age, 55 years). Using the total AET, 26 (22.8%) were classified as normal and 88 (77.2%) as borderline; however, using the DMS, 45 (39.5%) were classified as normal and 69 (60.5%) as pathological. The new pHoenix score demonstrated strong discriminative ability (AUC: 0.957 [95% CI 0.917, 0.998]) with high sensitivity and specificity (lower threshold, 94.4% and 79.2%; upper threshold, 87 and 95.8%). Compared to the total AET alone, the pHoenix score significantly decreased the proportion of inconclusive cases (77.2% vs. 13.2%, p < 0.001).ConclusionTotal AET has low sensitivity to identify pathological reflux as it disregards supine versus upright reflux. The pHoenix score improves the distinction between normal and pathological cases and reduces ambiguity, offering an alternative approach to diagnosing GERD that addresses the limitations of using total AET alone or the DMS.Graphical abstract
Read full abstract