[Dose-response association between fluid overload and hospital mortality in patients with sepsis].

Objective: To investigate dose-response associations between fluid overload (FO) and hospital mortality in patients with sepsis. Methods: The current cohort study was prospective and multicenter. Data were derived from the China Critical Care Sepsis Trial, which was conducted from January 2013 to August 2014. Patients aged≥18 years who were admitted to intensive care units (ICUs) for at least 3 days were included. Fluid input/output, fluid balance, fluid overload (FO), and maximum FO (MFO) were calculated during the first 3 days of ICU admission. The patients were divided into three groups based on MFO values: MFO<5%L/kg, MFO 5%-10%L/kg, and MFO≥10% L/kg. Kaplan-Meier analysis was used to predict time to death in hospital in the three groups. Associations between MFO and in-hospital mortality were evaluated via multivariable Cox regression models with restricted cubic splines. Results: A total of 2 070 patients were included in the study, of which 1 339 were male and 731 were female, and the mean age was (62.6±17.9) years. Of 696 (33.6%) who died in hospital, 968 (46.8%) were in the MFO<5%L/kg group, 530 (25.6%) were in the MFO 5%-10%L/kg group, and 572 (27.6%) were in the MFO≥10%L/kg group. Deceased patients had significantly higher fluid input than surviving patients during the first 3 days [7 642.0 (2 874.3, 13 639.5) ml vs. 5 738.0 (1 489.0, 7 153.5)ml], and lower fluid output [4 086.0 (1 367.0, 6 354.5) ml vs. 6 130.0 (2 046.0, 11 762.0) ml]. The cumulative survival rates in the three groups gradually decreased with length of ICU stay, and they were 74.9% (725/968) in the MFO<5% L/kg group, 67.7% (359/530) in the MFO 5%-10%L/kg group, and 51.6% (295/572) in the MFO≥10%L/kg group. Compared with the MFO<5%L/kg group, the MFO≥10%L/kg group had a 49% increased risk of inhospital mortality (HR=1.49, 95%CI 1.28-1.73). For each 1% L/kg increase in MFO, the risk of in-hospital mortality increased by 7% (HR=1.07, 95% CI 1.05-1.09). There was a"J-shaped"non-linear association between MFO and in-hospital mortality with a nadir of 4.1% L/kg. Conclusion: Higher and lower optimum fluid balance levels were associated with an increased risk of in-hospital mortality, as reflected by the observed J-shaped non-linear association between fluid overload and inhospital mortality.