Abstract
Background/Objectives: Postoperative body temperature abnormalities are common yet underrecognized, and their association with acute kidney injury (AKI) remains unclear. In this study, we aimed to identify early postoperative temperature trajectories and evaluate their associations with AKI. Methods: This retrospective cohort study included 3274 adults undergoing valvular heart surgery with cardiopulmonary bypass. The patients' temperatures were continuously measured using a pulmonary artery catheter for 12 postoperative hours, and temperature trajectories were identified using data-driven latrend class modeling. The primary outcome was severe AKI (KDIGO stage ≥ 2), and the secondary outcome was non-recovery AKI (≥72 h). Multivariable logistic regression and E-value sensitivity analysis were performed. Results: Four distinct temperature trajectories were identified: Class 1 (32.8%), initial normothermia progressing to mild hyperthermia (37.5-38.0 °C); Class 2 (27.4%), mild hypothermia (36.0-36.5 °C) with rapid normalization; Class 3 (24.4%), stable normothermia; and Class 4 (15.4%), lower-range mild hypothermia (35.5-36.0 °C) with delayed recovery. Severe AKI and non-recovery AKI occurred most frequently in Class 4 patients (15.1% vs. 2.9%, 3.9%, and 4.8% in Classes 1-3, p < 0.001; 15.1% vs. 1.7%, 4.0%, and 4.4%, p < 0.001, respectively). After adjusting for key clinical variables, Class 4 remained independently associated with severe AKI (OR 2.44, 95% CI: 1.69-3.57; E-value 4.33) and non-recovery AKI (OR 2.78, 95% CI: 1.89-4.00; E-value 4.97). Conclusions: Early postoperative temperature trajectories were significantly associated with severe AKI, with the highest risks in patients exhibiting lower-range mild hypothermia with delayed recovery. These findings suggest that early postoperative temperature patterns may be useful for risk stratification for severe AKI after cardiac surgery.