Combined Lyon and Milan Scores Predict Gerd Management Outcome Better Than Either Score Alone or Their Individual Components

Background: Confidence in gastro-oesophageal reflux disease (GERD) diagnosis is crucial to improve outcomes from escalation of treatment. The Lyon score phenotypes patients through endoscopy and pH-impedance (MII-pH). The Milan score quantifies antireflux barrier through high-resolution manometry (HRM) parameters. Aim: To explore the relationship between the Lyon and Milan scores and their combined performance in predicting clinical outcomes. Methods: We collected clinical and follow-up data of consecutive patients with HRM and MII-pH from nine centres. Clinical improvement was defined as a 50% reduction in global symptoms. The relationship between Lyon and Milan scores and the rate of patients improved in Lyon phenotypes and Milan categories were explored. The ability of the Lyon, Milan, DeMeester scores and acid exposure time (AET) in predicting outcomes was assessed through receiver operating characteristics (ROC) analysis. Results: Among 532 patients (50.6% female, age 50 years), 47.7% had pathologic GERD. A stepwise increase in the Milan score in Lyon phenotypes was observed. Sixty-three patients had surgical treatment, and 131 medical. Clinical improvement in Lyon phenotypes Conclusive and Severe was 81% and 83%; in Milan categories very likely and extremely likely was 88.5% and 100%. If Lyon and Milan scores were positive, improvement was 89%; if discordant, 63%; if both negative, 19% (p < 0.001). ROC analysis showed an AUC of 0.790 for Lyon score, 0.835 for Milan score, 0.736 for DeMeester score and 0.741 for AET. Conclusions: The Lyon and Milan scores outperformed AET and DeMeester scores in predicting outcomes in GERD patients. When concordant, they provide optimal predictive accuracy, guiding escalation of therapy.