Impact of corticosteroids on postoperative outcomes after hepatopancreatobiliary surgery.

Abstract

Phase 3 clinical trial data have suggested that perioperative administration of corticosteroids may improve postoperative outcomes by suppressing inflammatory and immune responses associated with surgery. We sought to investigate the real-world effectiveness of corticosteroid administration to improve postoperative outcomes among patients undergoing hepatopancreatobiliary surgery. Patients who underwent hepatopancreatobiliary surgery (excluding simple cholecystectomy) between 2016 and 2024 were identified using the Epic Cosmos database. The primary exposure was the administration of a single shot of corticosteroids (methylprednisolone or hydrocortisone) on the day of surgery. Stabilized inverse probability of treatment weighting was used to examine the association between administration of corticosteroids and postoperative outcomes (ie, complications, length of stay, and 90-day mortality). Among 125,269 patients, a majority underwent pancreatic surgery (n = 65,663, 52.4%) for a nonmalignant indication (n = 79,145, 63.2%). Overall, 2.2% of patients (n = 2,782) were administered corticosteroids on the day of surgery. Patients who did versus did not receive corticosteroids were more likely to be younger (61.9 years [standard deviation ±14.7 years] vs 63.8 years [standard deviation ±14.8 years]), have a higher Charlson Comorbidity Index (>2) (79.5% vs 72.2%), and undergo hepatic surgery (42.7% vs 31.0%) via an open surgical approach (93.4% vs 86.5%) for a nonmalignant indication (72.9% vs 63.0%) (all P < .05). On inverse probability of treatment weighting analysis after adjusting for relevant clinicodemographic factors, administration of corticosteroids was associated with higher odds of postoperative complications (odds ratio: 2.16 [95% confidence interval: 1.93-2.42]), extended length of stay (>75th percentile) (odds ratio: 1.56 [95% confidence interval: 1.40-1.73]), and 90-day mortality (odds ratio: 1.26 [95% confidence interval: 1.11-1.41]) (all P < .001). On subgroup analyses, the administration of corticosteroids did not impact the incidence of bile leak among patients undergoing hepatobiliary surgery (odds ratio: 1.15 [95% confidence interval: 0.92-1.43]), nor did it affect the incidence of pancreatic fistula among patients undergoing pancreatic surgery (odds ratio: 1.11 [95% confidence interval: 0.83-1.47]) (both P > .05). Based on real-world electronic health record data, administration of corticosteroids was not associated with improved outcomes after a hepatopancreatobiliary surgical procedure. Rather, use of same-day corticosteroids had a detrimental effect on postoperative outcomes, suggesting that steroids should not be routinely used in the perioperative hepatopancreatobiliary setting.