N. Amini1,2, C. R. D’Adamo1, R. Dodoson1, M. R. Katlic1, J. Wolf1, V. Ahuja3, D. Khashchuk1, A. A. Mavanur1 1Sinai Hospital of Baltimore,Surgery,Baltimore, MD, USA 2Johns Hopkins University School Of Medicine,Surgery,Baltimore, MD, USA 3Yale University School Of Medicine,Surgery,New Haven, CT, USA
Introduction: Pancreatectomy is an operation with high mortality and morbidity especially in elderly patients. NSQIP database has been used as a risk calculator to help surgeons and patients to make decision regarding surgical treatment. The purpose of this study was to determine the accuracy of current risk calculator and its correlation with age for patients who underwent pancreatic resection.
Methods: Using NSQIP database, patients who underwent pancreatic resection (whipple, distal pancreatectomy and total pancreatectomy) between 2012 and 2015 were included in the study. Analysis of variance was performed to assess differences between age categories in predicted and actual mortality and Pearson correlation coefficients were calculated. Logistic regression models were constructed to evaluate associations adjusted for key covariates.
Results: A total of 17906 patients were included with age distribution of 18-64 (n=9060,50.6%), 65-79 years (7537,42.1%), and 80-89 (n=1309,7.3%). About two-third (n=11567, 64.6%) of patients underwent whipple procedure, 32.8% (n=5881) underwent distal pancreatectomy, and 2.6% (n=458) underwent total pancreatectomy. Both actual and predicted mortality increased with age (P<0.001) (Figure). The overall correlation between actual and predicted mortality was low (r=0.14, P<0.001). The correlation was lowest among patients who underwent whipple procedure, especially with overestimation of mortality (actual mortality:3.2% vs predicted mortality 5.6%) in age group 80-89 (r=0.05). After adjusting for covariates (BMI, race, smoking, functional status, and major hepatectomy) the interaction term between age and predicted mortality was statistically significant in multivariable analysis model, indicating the varying accuracy of the actual and predicted mortality by age. Regarding morbidity, the correlation between actual and predicted one was also lowest among patients who underwent whipple procedure compared to distal and total pancreatectomy (age 18-64 r=0.09; age 65-79 r=0.11; age 80-89 r=0.10, All P values<0.001). In addition, the morbidity was overestimated in 80-89 age group and had the weakest correlation among patients underwent distal or total pancreatectomy (r for both groups:0.13; P<0.05)
Conclusion: The ACS NSQIP risk calculator in older patients appears to overestimate overall mortality and morbidity risk, especially for patients undergoing whipple procedure. In elderly patients undergoing pancreatic resection, functional assessment in addition to NSQIP calculator may aid in a more accurate risk prediction.
