Characteristics and predictors of acute kidney injury in patients with nonmalignant pleural effusions

Background Previous small-sample studies have suggested that pleural effusion may be a risk factor for acute kidney injury (AKI). However, the incidence and predictors of AKI in nonmalignant pleural effusion (NMPE) patients remain unclear.

Objective To assess the incidence and predictors of AKI in NMPE inpatients.

Methods A retrospective study of inpatients with pleural effusions admitted to Chinese PLA General Hospital from 2018 to 2021 was conducted based on the electronic medical records. All patients with pleural effusions confirmed by chest radiography (CT or X-ray) were included, excluding diagnosis of malignancy, maintenance dialysis, end-stage renal disease, AKI before chest radiography and those with serum creatinine tests being less than twice. The incidence of hospital-acquired AKI (HA-AKI) was described according to different departments and diagnoses. LASSO-logistic regression model was used to identify predictors of HA-AKI.

Results Totally 3047 cases met the inclusion criteria. Among them, there were 360 cases (11.8%) of HA-AKI. After selected by LASSO-logistic regression, pleural effusion volume, mechanical ventilation, ICU admission, vancomycin/teicoplanin, vasoactive drugs, spironolactone, loop diuretics, non-steroidal anti-inflammatory drugs, low eGFR, low platelets, high white blood cell count, and proteinuria were independent predictors of HA-AKI. Patients with moderate and large pleural effusions were more likely to develop AKI in hospital compared to patients with small effusions (moderate: OR=1.47, 95% CI: 1.11-1.94, P=0.006; large: OR=1.86, 95% CI: 1.05-3.20, P=0.028).

Conclusion The incidence of HA-AKI is 11.8% among NMPE patients, and large effusion volume is independently associated with HA-AKI. In addition, mechanical ventilation, ICU admission, vancomycin/teicoplanin, vasoactive drugs, spironolactone, loop diuretics, non-steroidal anti-inflammatory drugs, low eGFR, low platelets, high white blood cell count, and proteinuria are independent predictors of HA-AKI during hospitalization.