Severe morbidity after pancreatectomy is accurately predicted by preoperative pancreatic resection score (PREPARE): A prospective validation analysis from a medium-volume center.

ABSTRACT

BACKGROUND: Major morbidity in pancreatic surgery remains high. Different scores for predicting complications have been described. Preoperative pancreatic resection (PREPARE) score is based on objective preoperative variables and offers good predictive accuracy for Clavien ≥ III complications. This study aimed to validate this score and analyze other preoperative variables in a prospective study performed in a medium-volume center. METHODS: A total of 50 pancreatic resections were included. Preoperative variables were registered and PREPARE was calculated. The main outcome was severe morbidity (Clavien ≥ III) up to 30 days after discharge. The secondary outcomes were length of stay (LOS) and readmission. Statistical validation was performed to compare severe morbidity rate among the scores categories. Association with other preoperative variables (not included in PREPARE) was also tested. RESULTS: Of the 50 pancreatic resections, the severe morbidity was 34.0%, with median LOS of 11 days. Readmission rate was 25.5%. Severe morbidity rates according to PREPARE categories were 18.5% in low-risk group, 41.7% in intermediate-risk group, and 63.6% in high-risk group, respectively (P = 0.023). The accuracy was 72% (Hosmer-Lemeshow, P = 0.86). ROC curve was obtained both for PREPARE score expressed as incremental values and categorized as the three risk groups, showing an area under curve (AUC) of 0.736 (95% CI 0.586-0.887; P = 0.007) and 0.712 (95% CI 0.555-0.869; P = 0.015), respectively. PREPARE was significant in multivariate analysis. Median LOS was statistically higher as PREPARE category increases (9, 11 and 15 days in low-, intermediate- and high-risk groups, respectively; P = 0.009). Readmission was not associated with any variables. CONCLUSIONS: PREPARE behaves as an independent risk factor for severe morbidity after pancreatic surgery. Score validation shows good accuracy prediction. Increasing PREPARE category is also associated with longer LOS.