
Nonalcoholic Fatty Liver Disease After Pancreaticoduodenectomy for a Cancer Diagnosis
Author(s) -
Amy McGhee-Jez,
Inna Chervoneva,
Misung Yi,
Amisha Ahuja,
Ritu Nahar,
Sushrut D. Shah,
Rebecca Loh,
Sarah Houtmann,
Rohan Shah,
Charles J. Yeo,
Harish Lavu,
Steven J. Cohen,
Dina Halegoua-DeMarzio,
Atrayee Basu Mallick
Publication year - 2021
Publication title -
journal of pancreatic cancer
Language(s) - English
Resource type - Journals
ISSN - 2475-3246
DOI - 10.1089/pancan.2020.0006
Subject(s) - medicine , nonalcoholic fatty liver disease , gastroenterology , body mass index , hazard ratio , perioperative , diabetes mellitus , incidence (geometry) , fatty liver , surgery , disease , endocrinology , confidence interval , physics , optics
Purpose: Current literature reports increased incidence of postpancreaticoduodenectomy (PD) nonalcoholic fatty liver disease (NAFLD), a precursor for nonalcoholic steatohepatitis and cirrhosis. The incidence of and risk factors (RFs) for NAFLD in the PD population, however, are not well elucidated. Methods: A cohort of 421 patients from a single institution who underwent PD for carcinoma and followed for at least 6 months were assessed retrospectively for age, gender, pathology, surgical complications (operative blood loss and length of stay [LOS]), comorbidities (diabetes, hypertension, hyperlipidemia, obesity), tobacco use, pre- and postoperative nutritional status (albumin and body mass index [BMI]), use of pancreatic enzyme replacement, and perioperative laboratory values (hemoglobin and liver function test). Cox proportional hazards model was used to examine these potential RFs as predictors of time to development of post-PD NAFLD. Results: Sixty (14.3%) patients developed post-PD NAFLD. Patients with NAFLD were younger (61.10 vs. 65.01 years old) and had higher preoperative BMI (28.92 vs. 26.61). Multivariate Cox proportional hazard model identified higher preoperative BMI, shorter postoperative LOS, and female gender as RFs for post-PD NAFLD. After excluding 12 patients with rare histology, there was a lower unadjusted hazard of developing NAFLD ( p -value = 0.018) in the adenocarcinoma group than in the neuroendocrine and periampullary tumor groups. There was no statistically significant association between post-PD NAFLD and other characteristics. Conclusion: Female gender, higher preoperative BMI, and shorter LOS deserve closer monitoring for earlier detection and management of NAFLD.