Association between volume status assessed by bioelectrical impedance analysis, lung ultrasound, or weight change and mortality in patients with sepsisassociated acute kidney injury receiving continuous kidney replacement therapy.

Abstract

Fluid overload is an independent risk factor of mortality in patients with acute kidney injury (AKI) receiving continuous kidney replacement therapy (CKRT). However, the association between fluid status, as assessed by bioelectrical impedance analysis (BIA) or lung ultrasound, and survival in patients with AKI requiring CKRT has not been established. We analyzed 36 participants with sepsis-associated AKI who received CKRT at a tertiary hospital. The main exposures were volume surrogates: 1) overhydration normalized by extracellular water (OH/ECW, L/L) assessed by BIA, 2) the number of B-lines measured by lung ultrasound, and 3) weight change ([body weight at CKRT initiation - body weight at admission] × 100/body weight at admission). The primary outcome was the 28-day mortality. Seventeen participants (47.2%) died within 28 days. There were no significant correlations between OH/ECW and weight change (R2 = 0.040, p = 0.24), number of B-lines and OH/ECW (R2 = 0.056, p = 0.16), or weight change and number of B-lines (R2 = 0.014, p = 0.49). Kaplan-Meier analyses revealed that patients in the highest tertile of OH/ECW showed a significantly lower cumulative 28-day survival probability than the others (the lowest + middle tertiles). The survival probability of participants in the highest tertile of the number of B-lines or weight change did not differ from that of their counterparts. In a multivariate Cox proportional hazard model, the hazard ratio for the highest tertile of OH/ECW was 3.83 (95% confidence interval, 1.04-14.03). Volume overload assessed using BIA (OH/ECW) was associated with the 28-day survival rate in patients with sepsis-associated AKI who received CKRT.