J. R. Bergquist1,2, C. R. Shubert1,2, D. S. Ubl2, C. A. Thiels1,2, M. L. Kendrick1, M. J. Truty1, E. B. Habermann2 1Mayo Clinic,General Surgery,Rochester, MN, USA 2Mayo Clinic,Center For The Science Of Health Care Delivery,Rochester, MN, USA
Introduction: Expected mortality after elective pancreaticoduodenectomy (PD) in contemporary series is less than 5% even in older patients (>80). The perioperative risk in these older patients has not been reported with consideration of the specific indication for PD. We hypothesized that 30-day mortality, major morbidity, and prolonged length of stay (PLOS) following PD varies by diagnosis risk group in patients over 80, and that those elderly patients with high risk diagnoses may have higher than expected peri-operative risk.
Methods: ACS-NSQIP was reviewed for all PDs from 2005-2012. ICD-9 diagnoses (indication for PD) were categorized into high and low diagnosis risk groups based on incidence of 30-day major morbidity. Univariate and multivariate analyses compared PD outcomes (1) by diagnosis risk among patients over 80 and (2) by age group (80+ vs 18-79 and vs 70-79) among patients in the same diagnosis risk group.
Results: Of 7192 total patients, pancreas cancer (N=4200) and chronic pancreatitis (N=608) experienced similar major morbidity (p=0.64) and were grouped as “low risk”. Bile duct and ampullary neoplasm (N=1503), duodenal neoplasm (N=686), and neuroendocrine tumor (N=195) experienced similar major morbidity (p=0.69) and were grouped as “high risk”. The 30 day mortality risk for patients aged 80+ (N=749) undergoing PD with high risk diagnosis was found to be 7.0% vs 4.1% for those with low risk diagnosis (p<0.001). Of patients with high-risk diagnoses, patients 80+ had greater mortality risk (7.0%) than those 70-79 (3.9%, p=0.037) or all patients aged 18-79 (2.9%, p<0.001). Risk of major morbidity and prolonged length of stay was also increased in patients 80 and older (see table). On multivariate analysis, controlling for diagnosis risk, patients over 80 had greater odds of 30 day mortality (OR 2.155, 95% CI 1.242-3.741, p=0.0063), any major complication (OR 1.658, 95% CI 1.312-2.095, p<0.001), and PLOS (OR 1.448, 95% CI 1.140-1.838, p=0.0024), and when compared with patients 18-79.
Conclusion: In patients over 80 undergoing PD, high-risk diagnoses are independently associated with increased 30-day mortality compared to those with low-risk diagnoses and younger age groups. Risk of 30-day mortality following PD in patients 80+ with high risk diagnoses exceeds the expected threshold of 5%; those with low risk diagnoses however do not. For 80+ patients with duodenal, neuroendocrine, or bile duct and ampullary neoplasm, pre-operative counseling and shared decision making should reflect the increased 30-day mortality risk for pancreaticoduodenectomy.
