Mechanical Power and Acute Kidney Injury in Mechanically Ventilated Critically Ill Patients: A Single-Center Retrospective Cohort Study

Background and Objectives: Acute kidney injury (AKI) is highly prevalent among critically ill patients receiving invasive mechanical ventilation (IMV) and is associated with adverse outcomes. Mechanical power (MP) reflects the total energy delivered to the respiratory system, but its association with AKI in mechanically ventilated critically ill patients remains incompletely understood. Few studies have examined the relationship between MP and AKI development. Materials and Methods: Adult patients with MP data available for each of the first five consecutive days of IMV constituted the study population. MP and ventilatory parameters were recorded during this period. Factors independently associated with AKI were identified using multivariable logistic regression. The discriminative performance of MP for AKI and continuous renal replacement therapy (CRRT) use was evaluated. Results: Among 924 patients, 605 (65.5%) developed AKI. MP was consistently higher in patients who developed AKI during the first five days of IMV (11.45 ± 2.88 vs. 9.48 ± 2.50 J/min; p < 0.001). Although PEEP, peak inspiratory pressure, driving pressure, respiratory rate, and FiO2 were higher in the AKI group, tidal volume and respiratory system compliance did not differ significantly. In multivariable analysis, higher mean MP (OR = 1.330, 95% CI 1.235–1.432; p < 0.001), greater mean fluid balance (OR = 1.477, 95% CI 1.225–1.780; p < 0.001), older age, and trauma were independently associated with AKI. MP showed moderate discrimination for AKI (AUC = 0.681), with an optimal cutoff of 10.38 J/min. For CRRT use, MP showed significant but limited discrimination (AUC = 0.624; 95% CI, 0.584–0.664; p < 0.001). Conclusions: In critically ill patients receiving IMV, higher MP remained associated with AKI after adjustment for measured covariates. The 10.38 J/min threshold is hypothesis-generating and requires external validation.