A Multi-variable, Veteran-Centered Study of PONV Risk Factors in Separate Regional versus General Anesthesia Contexts

Christopher A Schumacher¹Ridhi Choragudi¹Joseph M Mikolic²Monique Y Boudreaux-Kelly²Brian A. Williams^1,2*

• ¹University of Pittsburgh, Pittsburgh, United States
• ²Veterans Affairs Pittsburgh Health System, Pittsburgh, United States

Background: Currently, perioperative stakeholders are guided to provide general anesthesia (GA) patients with either 2 or 4 antiemetic prophylaxis (AEPPx) medications based on 1990s legacy risk factors (RFs). There are no Veteran-centric regional anesthesia (RA) or GA postoperative nausea-vomiting (PONV) risk factor (RF) studies, and few race/ethnicity-based-studies. Thus, currently-accepted AEPPx in Veterans may be escalating symptoms, costs, and lengths of stay. Methods: We first conducted IRB-approved secondary analyses from a prospective Veteran-specific randomized trial to assess for RA-specific PONV RFs. We then conducted IRB-approved retrospective analyses of observational quality improvement (QI) data regarding Veterans receiving GA with or without intrathecal morphine (ITM) pre-operatively (with ITM cases accompanied by 5-drug AEPPx). The goal was to assess Veteran-specific, and RA-then-GA-specific PONV RFs. For RA-specific PONV RF analyses in 115 Veterans, we queried electronic medical records (EMR) along with database-archived study data from case report forms. For GA-specific PONV risk factor analyses in 468 Veterans, we analyzed EMR data to compare PONV-free patients with PONV-positive patients, both for postoperative days 0-1 (POD#0-1 after surgery), and for POD#2. Results: Postoperative opioids were associated with increased PONV in both analyses. For RA, African-American Veterans were found to have more PONV despite lower overall opioid consumption than in the race-referent group, while diabetic Veterans overall showed less PONV. For GA-specific analyses (informed by the risks and signals identified in RA analyses), African-American Veterans again had more PONV. Consensus-guided RFs added to the models were often non-predictive, particularly (i) smoking status and past PONV (RA-specific), and (ii) gender and past PONV (GA-specific). This may suggest underpowering in both limited sample sizes, or, instead indicate race as a profoundly overriding RF. RFs associated with POD#2 PONV after GA (after no PONV on POD#0-1) notably differed from factors driving POD#0-1 PONV. Conclusion: Consensus-guided AEPPx may require reevaluation, particularly in Veterans undergoing RA or GA, if not population-wide. All Veterans could benefit from our 2023-described off-patent 5-drug AEPPx before any anesthetic drug is administered, described herein and elsewhere. Emerging RFs may have pharmacoequity and race-based implications.