LAUSR

Introduction/Hypothesis: Patients with severe acute pancreatitis (AP) are generally triaged to ICU care and have high risk of death. However, there are sparse population-level data on the contemporary epidemiology, characteristics, and outcomes of patients with AP admitted to ICU. Methods: The Texas Public Use Data File was used to identify hospitalizations aged ≥18 years with a primary diagnosis of AP (ICD-9 code 577.0) and those with ICU admission during 20092014. The characteristics and outcomes of ICU admissions were examined. Logistic and linear regression models were used to explore the temporal trends of ICU admission rates and APR-DRG severity of illness (SOI) [converted to a 4-level quantitative score (e.g., “mild”=1, “extreme”=4)] among ICU admissions, respectively. Multivariate logistic regression was used to examine temporal trends of adjusted hospital mortality among ICU admissions. Results: Among 120,491 hospitalizations with AP, 25,943 (21.5%) were admitted to ICU. ICU admission rates increased from 18.4% in 2009 to 23% in 2014 (odds ratio [OR] 1.063/ year [95% CI 1.054-1.072]). The volumes of ICU and hospital admissions rose 34.8% and 8.7%, respectively during the corresponding period. Among ICU admissions, 28.6% were aged ≥65 years, 52.5% were female, and 43.9% had ≥1 organ dysfunction (OD); mechanical ventilation (MV) and hemodialysis (HD) were used in 6.4% and 6.3%, respectively. APR-DRG SOI decreased over time (beta [SE] = -0.05/year [0.003]; p<0.0001). There were 14,511 (56%) ICU admissions without reported OD, MV, or HD (group 1). Hospital mortality decreased among all ICU admissions with AP from 3.8% to 2% between 2009 and 2014 (adjusted OR 0.89/year [95% CI 0.84-0.93]) and was 0.6% among group 1 ICU admissions. Conclusions: Admission to ICU was increasingly common among hospitalizations with AP. This rising demand for critical care services occurred despite the corresponding decrease in SOI and low and decreasing mortality among those admitted to ICU. The majority of ICU admissions with AP had extremely low risk death. Further studies are needed using more granular data to inform triage decisions and to reduce ICU utilization when equivalent care may be provided in non-ICU settings.