Association between early postoperative acetaminophen exposure and acute kidney injury across surgical cohorts: a retrospective analysis of the MIMIC-IV database

The association between early postoperative acetaminophen exposure and acute kidney injury (AKI) remains uncertain. We evaluated this association across major surgical cohorts using the MIMIC-IV database. We conducted a retrospective cohort study including 19,938 adult ICU patients undergoing cardiovascular, gastrointestinal, hepatobiliary, or neurosurgical procedures. Acetaminophen exposure was defined as at least one oral or intravenous administration within 0–48 h after surgery, and the primary outcome was AKI within 7 postoperative days. Multivariable logistic regression and Cox proportional hazards models were adjusted for demographic characteristics, surgical type, illness severity, renal function, comorbidities, and concomitant nephrotoxic medication exposure. Interaction analyses and sensitivity analyses, including a 48-hour landmark analysis and multiple imputation, were additionally performed. Acetaminophen exposure was associated with lower odds of postoperative AKI overall (OR = 0.76, 95% CI: 0.71–0.82; P < 0.001) and in cardiovascular (OR = 0.75, 95% CI: 0.68–0.83), gastrointestinal (OR = 0.88, 95% CI: 0.78–0.99), and hepatobiliary surgery (OR = 0.69, 95% CI: 0.54–0.88), but not neurosurgery (OR = 0.81, 95% CI: 0.55–1.18). The overall Cox model yielded a similar association (HR = 0.77, 95% CI: 0.73–0.82; P < 0.001). Associations differed by surgical type (global interaction P < 0.001), while exploratory interaction analysis suggested heterogeneity by SOFA category ( P = 0.021). In the 48-hour landmark analysis, the association remained evident overall and in the gastrointestinal and hepatobiliary cohorts, but not in the cardiovascular or neurosurgical cohorts. Early postoperative acetaminophen exposure was associated with lower observed AKI occurrence, although the findings varied by surgical context and were sensitive to exposure timing. The gastrointestinal association persisted in the landmark analysis but was no longer statistically significant after multiple imputation, whereas the hepatobiliary association was more consistent across sensitivity analyses.