5-mFI is more accurate than ASA score in predicting postoperative mortality in rectal cancer: A case series of 109 patients.

Introduction: The aim of this study was to compare the accuracy of 5-mFI (modified frailty index) to ASA score (American Society of Anesthesiologists score) in predicting postoperative mortality in patients with rectal cancer. Materials and methods: The ability of each parameter to predict postoperative mortality was attested in 2 ways: Area under the curve (AUC) was determined using ROC curves analysis. A comparison of AUC was performed using Delong test and Henley-McNeil test.-Multivariate analysis to determine the weight of each variable in predicting postoperative mortality. Results: The records of 109 patients undergoing surgical resection, for curative intent, for rectal cancer, were analyzed. Nine patients died during the 30-day postoperative period (8.25%). The optimum cutoff for 5-mFI to predict mortality using the ROC analysis was 1.5. The AUC at the cut-off point was 0.93. The optimum cutoff for ASA score to predict mortality was 1.5 and the AUC at the cut-off point was 0.81. The AUC of 5-mFI was significantly higher than the AUC of ASA score (p < 0.0001 using Delong test and p = 0.0024 using Hanley and McNeil test).On univariate analysis, predictive factors of mortality were: age (p = 0.002), ASA score≥2 (p = 0.0001) and 5-mFI≥2 (p = 0.0001). On multivariate analysis, 5-mFI≥2 was the only factor significantly associated with increased odds of postoperative mortality (OR = 1.73; 95% CI 1.05-2.01). Conclusion: 5-mFI was more accurate than ASA score in predicting postoperative mortality in patients with rectal cancer.

Acute Gangrenous Cholecystitis: Proposal of a Score and Comparison with Previous Published Scores.

BACKGROUND: Gangrenous cholecystitis (GC) is a particularly severe form of acute cholecystitis (AC) and is associated with an increased risk of postoperative morbidity and mortality. Recent reports show that surgeons are remarkably unsuccessful in diagnosing GC. METHODS: We conducted a retrospective study involving 587 patients with AC. Logistic regression analysis was used to identify independent predictive factors of GC. We assigned points for the score according to the regression coefficient. The area under the curve (AUC) was determined using receiver operating characteristic (ROC) curves. The scoring system was then prospectively validated on a second population. We validated 2 previously published scoring models. RESULTS: Six independent predictive factors of GC were identified: [3-]4 ASA score, temperature, duration of symptoms, WBC, male gender, and pericholecystic fluid. A predictive score of GC was established based on these independent predictive factors. Sensitivity was 81.4%; specificity was 70%. The AUC of this clinicoradiological score was 0.83. The AUC of our score was higher than that of the first published score (the AUC was 0.75 in the original report and 0.78 in the validation model using our dataset) and that of the second published score (the AUC was 0.77 in the original report and 0.72 in the validation model using our dataset). CONCLUSIONS: The AUC of our score exceeded 0.80, indicating that this score can help in diagnosing patients with GC, and thus in prioritizing these patients for surgery or choosing the adapted technique of drainage in critically ill patients.