When Should We Operate on the Elderly with Acute Pancreatitis Requiring Emergency Surgery? A 10-Year Study of a National Inpatient Sample Using a Generalized Additive Model of 110,289 Patients.

Introduction: Elderly patients with acute pancreatitis have longer hospital length of stay (HLOS) and higher mortality compared to adult patients. We aimed to assess the optimal timing to operate for acute pancreatitis and to evaluate the relationship between HLOS and mortality.

Materials And Methods: This was a retrospective cohort study of 110,289 elderly patients diagnosed with acute pancreatitis requiring emergency admission using the National Inpatient Sample (NIS) between 2005-2014. The ICD9 code 577.0 was used to select patients with a diagnosis of acute pancreatitis. Stratified analysis was performed to compare male versus female, survived versus deceased, and no operation versus operation. Multivariable logistic regression models were created to assess independent risk factors of mortality. Generalized additive models (GAM) were created to assess the linearity of the relationship between HLOS and in-hospital mortality.

Results: The mean age of the cohort was 76 years old, and 56.3% were female. The mean frailty index was 1.65. Twenty-five percent of patients underwent an operation, with a mean time to operation being 3.44 days for females and 3.77 days for males. Overall mortality was 2.3%. For patients who had an operation, each additional day of delay until operation increased the odds of mortality by 8.8%. Each additional point for the modified frailty index increased the odds of mortality by 30.2%. HLOS had a non-linear relationship with mortality, with an estimated degree of freedom of 22.05 and a nadir at three to seven days. Each additional day in hospital after day seven increased the odds of mortality by 6.7%.

Conclusions: In those who required an operation, every day of delay in operation increased the odds of mortality by almost 9%. The lowest mortality for elderly patients with acute pancreatitis occurred with a hospital length of stay of three to seven days. After seven days, each additional day increased the odds of mortality by 6.7%.