Validation of Obesity Surgery Mortality Risk Score in patients undergoing gastric bypass in a Canadian center

Abstract Background The Obesity Surgery Mortality Risk Score (OS-MRS) has been proposed as a user-friendly tool for the assessment and risk stratification of patients undergoing Roux-en-Y gastric bypass (RYGB). We assessed the validity of the OS-MRS in 2121 primary RYGB procedures performed at our center during a 25-year period. Methods A retrospective study of the patients who had undergone primary RYGB since 1983 was performed. The 90-day mortality and all mortalities related to complications of the RYGB were determined. For every patient, we assigned the relevant risk score according to their co-morbidities and relevant demographics. Each patient was assigned to a class (A, B, or C) according to the OS-MRS. We used the Z test to estimate whether the difference between the actual and predicted risk using the OS-MRS was statistically significant. Results We identified 2121 patients who had undergone primary RYGB, of which 1254 (59%) were open (ORYGB) and 867 (41%) were laparoscopic (LRYGB). The mean body mass index was 50.7 ± 8.6 kg/m 2 , and the mean age was 39.7 ± 9.9 years. The mortality rate for ORYGB was 1% (13 patients) and for LRYGB was .4% (4 patients). The overall mortality rate was .8% (17 patients). Of the 2121 patients, 1385 (65%) were in class A, 671 (32%) were in class B, and 65 (3%) were in class C. The expected versus observed mortality rate was .3% versus .3% for class A, 1.9% versus 1.5% for class B, and 7.5% versus 3% for class C, respectively. The difference between the mortality expected from applying the OS-MRS in our cohort and the observed mortality was assessed for statistical significance using Flora's Z statistic. No significant difference was found between the observed and expected mortality, suggesting that the OS-MRS was a valid tool for predicting mortality in our cohort. Conclusion In our bariatric center with >25 years' experience, the OS-MRS accurately predicted the postoperative mortality for RYGB surgery. It appears to be a user-friendly scoring system that could facilitate the informed consent process. Before the system is unequivocally adopted, additional validation trials of a prospective nature are required.