S105 Acute Pancreatitis in a Los Angeles Safety Net Hospital System: Patient Characteristics and Predictors of Readmission

Abstract

Introduction: Acute pancreatitis (AP) is the most common gastrointestinal cause for hospital admission in the United States and may disproportionately affect minority populations. We aimed to characterize patients with AP and identify predictors of readmission in hospitals belonging to the Los Angeles County Department of Health Services, a large safety net hospital system. Methods: A database of patient encounters at Olive View-UCLA, Harbor-UCLA, and LAC+USC Medical Centers was queried for patients seen in 2017 with an elevated lipase or diagnosis matching ICD-10-AM codes K85.0, K85.1, K85.2, K85.3, K85.8, or K85.9. Patients were included in the study if they met at least 2 of the following 3 criteria: 1) lipase >3x the upper limit of normal, 2) radiographic findings suggestive of AP, or 3) abdominal pain. Patients were excluded if radiographic evidence of chronic pancreatitis was present. Outcomes included overall length of stay (LOS), mortality due to AP during the index admission, and readmission for AP or AP-related complications. A multiple logistic regression model was used to identify predictors of readmission for AP. Results: Six hundred and twenty-three patients with median age 45 years, 83% Hispanic, and 51% male met inclusion criteria for the study. The most common AP etiology was gallstones, seen in 251 patients (40%), followed by alcohol, seen in 144 (23%). Radiographic evidence of acute interstitial edematous (IEP) and necrotizing pancreatitis (NP) was seen in 61% and 7% of patients, respectively. Greater than half of all patients with IEP developed an acute peripancreatic fluid collection, while 48% and 39% of patients with NP developed an acute necrotic collection or walled-off necrosis, respectively. The median LOS was 4 days (IQR 2 – 6). Five patients (1%) died during the index admission due to an AP-related complication, while 48 (8%) had a readmission for AP at a median of 91 days (IQR 15 – 178) from discharge. Increasing age (OR = 0.97, P = 0.038) was associated with decreased odds of readmission for AP when controlling for sex, BMI, and presence of any local complication on imaging. (Table) Conclusion: In our safety net patient population, gallstones and alcohol were the most common AP etiologies and increasing age at admission was associated with decreased odds of readmission for AP, mirroring findings reported in the literature. Our mortality rate was comparatively low at just 1%, however this may be because we did not limit our study cohort to patients with a primary diagnosis of AP. Table 1. - Predictors of readmission for AP modeled using multiple logistic regression Predictor Odds Ratio Std. Err. P-Value 95% Confidence Interval Male 1.578 0.634 0.256 0.718 – 3.470 Age 0.972 0.013 0.038 0.947 – 0.998 BMI 1.009 0.021 0.678 0.968 – 1.050 Presence of local complications 1.560 0.618 0.261 0.718 – 3.391