Association between fever burden and delirium risk after cardiac surgery: A retrospective cohort study based on the MIMIC database.

BACKGROUND: Postoperative delirium (POD) represents a prevalent adverse complication following cardiac surgical procedures. Although postoperative fever is frequently observed, its cumulative effect has not been quantified in relation to POD risk. We therefore examined this association using the MIMIC‑IV database. OBJECTIVES: We quantified the standardized fever burden within 72 h after cardiac surgery and assessed its independent association with POD. METHODS: We retrospectively enrolled 894 patients, grouped into delirium (n = 154) and non-delirium (n = 740) cohorts. The standard fever burden (SFB, °C·h/day) was calculated using the trapezoidal method. The association between SFB and POD was analyzed using univariate and multivariate Logistic regression. Subgroup analysis was conducted to explore heterogeneity, and sensitivity analysis was performed to verify the robustness of the conclusion. RESULTS: SFB was significantly higher in the delirium group (0.49 vs. 0.13 °C·h/day, p < 0.001). Univariate analysis showed that SFB was significantly associated with the risk of POD (OR = 3.58, 95% CI: 2.28∼5.62, p < 0.001). After multivariate stepwise adjustment for confounding factors, SFB remained an independent associated factor for POD (OR = 2.13, 95% CI: 1.22∼3.70, p = 0.007). Subgroup analysis showed no significant interaction with age or mechanical ventilation duration. Sensitivity and propensity score matching analyses confirmed the stability of results. CONCLUSIONS: Fever burden (≥37.9 °C) within 72 h after cardiac surgery was independently associated with POD. These findings were robust across sensitivity analyses and suggest that fever burden may help identify high risk patients.