Abstract
S188 Low BMI Identifies a High-Risk Nutritional-Metabolic Phenotype in Acute Pancreatitis Patients With Alcohol Use Disorder
The American journal of gastroenterology, Vol.121(10S1), pp.S41-S41
10/01/2026
DOI: 10.14309/01.ajg.0001210004.99648.64
Abstract
Introduction:
Alcohol use disorder (AUD) is a clinically important comorbidity in acute pancreatitis (AP), but predictors of poor outcomes within this subgroup remain poorly defined. We evaluated whether nutritional-metabolic factors identify AP patients with AUD at risk for prolonged hospitalization and readmission.
Methods:
We used a prospectively maintained cohort of 200 consecutive AP patients screened for trial enrollment and performed a retrospective analysis. AUD status was determined through structured chart review of clinician-documented diagnoses and alcohol-use history. Forty-two patients were classified as having AUD. Key exposures were diabetes mellitus (DM) and body mass index (BMI). Outcomes included length of stay (LOS), 30-day and 90-day readmission. Multivariable models included linear regression for LOS and logistic regression for readmission, adjusting for age, sex, etiology, bedside index for severity in acute pancreatitis score, and chronic pancreatitis. Intensive care unit admission was included only in LOS models to avoid collinearity with severity. Receiver operating characteristic (ROC) analysis evaluated BMI discrimination for 30-day readmission and identify an exploratory BMI threshold.
Results:
Among 42 AP patients with AUD, patients readmitted within 30 days had lower BMI than those not readmitted (21.6 vs 26.9 kg/m2, P = 0.015), as did patients readmitted within 90 days (21.5 vs 26.6 kg/m2, P = 0.028). In multivariable models, higher BMI was independently associated with lower odds of both 30-day readmission (OR 0.79 per 1 kg/m2 increase, P = 0.023) and 90-day readmission (OR 0.74 per 1 kg/m2 increase, P = 0.026). ROC analysis demonstrated good discrimination for BMI predicting 30-day readmission (area under the curve 0.75). An exploratory BMI cutoff of <24.7 kg/m2 identified high-risk patients with 82% sensitivity and 90% negative predictive value. DM was present in 8 patients (19%). Patients with DM had longer unadjusted LOS than those without DM (15.4 vs 6.7 days, P = 0.038) and numerically higher 30-day readmission rates (50% vs 20.6%, P = 0.21). In multivariable models, DM remained directionally associated with longer LOS (+6.8 days, P = 0.41).
Conclusion:
Low BMI may identify a vulnerable subgroup of AP patients with AUD who experience higher 30 and 90-day readmission rates. DM was associated with longer hospitalization but remained underpowered after adjustment, supporting further evaluation of metabolic risk in this subgroup. These findings support optimizing nutritional assessment and diabetes management as part of in-patient and post-discharge care.
Details
- Title: Subtitle
- S188 Low BMI Identifies a High-Risk Nutritional-Metabolic Phenotype in Acute Pancreatitis Patients With Alcohol Use Disorder
- Creators
- Suryansh SuryanshAbhiram KannanSweta BhandariTejeshwar JainNikhil Bush JayaramSrikanth IyerVikas Dudeja
- Resource Type
- Abstract
- Publication Details
- The American journal of gastroenterology, Vol.121(10S1), pp.S41-S41
- DOI
- 10.14309/01.ajg.0001210004.99648.64
- ISSN
- 1572-0241
- eISSN
- 1572-0241
- Publisher
- Wolters Kluwer Health Medical Research, Lippincott Williams & Wilkins
- Language
- English
- Date published
- 10/01/2026
- Academic Unit
- Anatomy and Cell Biology; Surgery
- Record Identifier
- 9985244681502771
Metrics
1 Record Views