Response to Vittori et al

We appreciate the comments and feedback from Vittori et al (1) regarding our report describing the Lyon Score based on the Lyon Consensus 2.0. At the outset, the Lyon Score was devised to condense data from multiple esophageal testing metrics recommended by the Lyon Consensus into a single, easily calculated score for clinical use (2).

The Lyon Score incorporates a composite of metrics rather than acid exposure time (AET) alone in the determination of gastroesophageal reflux disease (GERD). Thus, the score is purposely geared toward the use of measures beyond AET from prolonged wireless pH monitoring, in recognition of the fact that wireless monitoring is not universally available, and that 24-hour pH-impedance monitoring is widely used in motility laboratories worldwide. As Vittori et al have indicated, factors other than AET can be important in determining reflux burden, and inclusion of additional metrics from 24-hour pH-impedance monitoring in the Lyon Score is precisely the reason for discordance with AET alone and with the DeMeester score, which are based on acid-based metrics from pH monitoring. As endoscopy is an integral part of esophageal evaluation, endoscopic evidence of reflux should be incorporated and scored, which we did using a separate cohort with both endoscopic and reflux monitoring findings, so the 2 could be correlated (3). We therefore agree with Vittori et al that the Lyon Score captures elements from reflux testing other than AET—these include endoscopic findings, reflux episode counts (more accurate with pH impedance compared with pH alone), mean nocturnal baseline impedance for mucosal integrity, and reflux-symptom association. We consider this a strength of the Lyon Score rather than a weakness.

Most of the elements in the TRIPOD checklist were indeed fulfilled by our description of methodology and results. As the cohorts from which the Lyon Score was developed and tested were convenience retrospective cohorts with existing data, further adjudication of how and why each patient was treated with specific management options was not part of this initial analysis. However, a “real-world” approach to GERD management is reflected in the cohorts, and use of medical management directed at GERD when GERD did not exist per the Lyon score actually helped identify lack of symptom improvement with these approaches.

Visual analog scale, on which the global symptom severity (GSS)score is based, has been widely used and validated in the assessment of several global symptomatic states, including esophageal symptoms (4,5). GSS has been successfully used in demonstrating GERD symptom improvement with antireflux therapy (6). The Lyon Score was tested and validated in convenience cohorts from 3 different continents with a wide spectrum of primary languages. A visual instrument such as GSS that does not rely on comprehension of written questions may provide advantage in reducing heterogeneity among cohorts because of language and translational differences. Moreover, while the primary GERD questionnaires used were not similar across the different international centers, GSS was available for all cohorts.

Since its first publication, the Lyon Consensus has become a widely used standard in the diagnosis of GERD worldwide (7,8). The Lyon Score represents an initial step toward simplifying the Lyon Consensus thresholds for a general audience. We agree with the overall conclusion of Vittori et al that the Lyon Score shows promise in discriminating GERD phenotypes. We demonstrate that such a score, in spite of limitations in methodology, actually predicts outcome from GERD management, but as we alluded to in the report, the proof lies in the pudding. Prospective testing of the Lyon Score is already underway in different populations all around the world. We maintain the view that critique and discussion can only happen after a new paradigm is introduced, and collective literature going forward will serve to improve and validate or reject the Lyon Score as new research gets published.