What question did this study set out to answer?

This research aims to determine if an anesthesiology-led ERAS workflow minimizes postoperative pulmonary complications following laparoscopic hepatectomy while identifying influential risk factors.

June 20, 2026Open Access

Impact of a hepatic enhanced recovery after surgery workflow on pulmonary complications after laparoscopic hepatectomy: machine learning-assisted analysis

SLSibei LiGuangzhou University of Chinese Medicine XWXinyuan WuNingxia University TLTongsen LuoSun Yat-sen University

Key Points

This research aims to determine if an anesthesiology-led ERAS workflow minimizes postoperative pulmonary complications following laparoscopic hepatectomy while identifying influential risk factors.
Retrospective review of 1057 patients from 2015 to 2021
Stratification into pre-ERAS, transitional, and full ERAS phases
Use of propensity score matching and causal inference analyses to assess associations
PPC incidence decreased significantly from pre-ERAS (1.3%) to full ERAS (76.2%), p < 0.001
ERAS implementation correlates with reduced PPC risk (OR 0.506; 95% CI 0.307–0.834; p = 0.007)
Machine learning identified nutritional status and inflammatory markers as key influential factors in later ERAS phases.

Abstract

To evaluate whether an anesthesiology-led hepatic enhanced recovery after surgery (ERAS) workflow reduces postoperative pulmonary complications (PPCs) following laparoscopic hepatectomy and to identify influential risk factors via machine learning. PPCs are a leading cause of postoperative morbidity and mortality. Although ERAS protocols reduce complications in many surgical fields, their effect on PPCs after laparoscopic hepatectomy—using standardized definitions—remains unclear. We retrospectively reviewed consecutive patients who underwent laparoscopic hepatectomy (2015–2021) at a tertiary academic center. Patients were stratified into three periods: pre-ERAS (2015–2016), transitional (2017–2019), and full ERAS implementation (2020–2021). PPCs were defined per the Standardized Endpoints for Perioperative Medicine (StEP) consensus. Propensity score matching (PSM) and causal inference analyses were used to assess the associations between ERAS and PPCs. The Boruta algorithm ranked influential factors across the ERAS phases. A total of 1057 patients were analyzed (pre-ERAS, 1.3%; transitional, 34.7%; full ERAS, 76.2%). The PPC incidence decreased significantly over time ( p < 0.001), with reductions in pneumonia, atelectasis, surgery duration, and postoperative length of stay. After Least Absolute Shrinkage and Selection Operator (LASSO)-Firth penalized logistic regression, ERAS implementation was independently associated with reduced PPC risk (OR 0.506; 95% CI 0.307–0.834; p = 0.007). Propensity score matching confirmed this association (OR 0.428; 95% CI 0.243–0.729; p = 0.002), with consistent findings across conditional logistic regression and inverse probability-of-treatment weighting. Machine learning revealed a shift in influential factors toward nutritional status, inflammatory markers, and liver function metrics in later ERAS phases. A structured hepatic ERAS workflow significantly reduces the number of PPCs after laparoscopic hepatectomy. The optimization of nutrition, inflammation control, and liver function may further increase the benefits of ERAS.

Ask AI

Helpful

Bookmark

View Full Paper