| Objective:Sepsis is a life-threatening organ dysfunction caused by abnormal response of the host after being infected by pathogenic microorganisms.It is a common fatal diseases in the field of emergency and critical medicine.Intravenous fluid resuscitation is one of the core links of bundle therapy for sepsis.Sepsis patients with heart failure have a higher risk of fluid volume overload,and these patients have worse clinical outcomes.At present,the research on fluid management in sepsis patients with heart failure is mainly focused on the efficacy of bundle therapy in sepsis.In these patients,the method of evaluating fluid overload and how fluid accumulation is related to prognosis is still unknown.In these patients,the method of evaluating fluid overload and how fluid accumulation is related to prognosis is still unknown.The purpose of this study was to evaluate the association of fluid management with in-hospital mortality of sepsis patients with heart failure.Methods:This study is a retrospective cohort study with data from the MIMIC-III database.The study included sepsis patients with heart failure older than 18 years old who had an ICU hospital stay of more than 48 hours.Patients who lacked fluid intake and output,received renal replacement therapy within 48 hours and were admitted to ICU due to cardiopulmonary bypass were excluded.The general data of patients,types of intensive care unit,chronic coexisting diseases,vital signs within 24 hours after admission,laboratory indexes and fluid management indexes within 24 hours after admission were collected.The main end point of the study was hospital mortality.In this study,the restricted cubic spline(RCS)model was used to explore the relationship between each variable and hospital mortality,and then linear spline function(LSP)and rank variables were used to establish Logistic model to explore the relationship between fluid balance(FB),liquid intake(FI),fluid accumulation index(FAI)(FB/FI)and in-hospital mortality.Results:(1)A total of 1801 sepsis patients with heart failure were included in this study,and were divided into survival group(n=1302)and non-survival group(n=499)according to their survival at discharge.The in-hospital mortality rate was 27.7%.FB(3.2L/48 h vs.1.7L/48 h,P<0.001),FI(5.8L/48 h vs.5.2L/48 h,P<0.001)and FAI(0.55 vs.0.34,P<0.001)in the non-survival group were significantly higher than those in the survival group.Urine volume in the non-survival group(2.2L/48 h vs.3L/48 h,P<0.001)was significantly lower than that in the survival group.(2)In this study,we used LSP function conversion and rank variable conversion for FI,FB and FAI in univariable Logistic regression.The results of rank variables showed that compared with moderate FI(45-100ml/kg/48h),low FI(<45m L/kg/48h)was negatively correlated with in-hospital mortality[Odds Ratio(OR): 0.682,95%Confidence Interval(CI): 0.523-0.887,P=0.005].Logistic regression using LSP function conversion showed no significant correlation between FI and in-hospital mortality.Using medium FB(0-60ml/kg/48h)as reference,rank variable results showed that low FB(≤0m L/kg/48h)was significantly associated with lower in-hospital mortality(OR: 0.694;95%CI: 0.529-0.905;P=0.007),while high FB(>60m L/kg/48 h)was associated with higher in-hospital mortality(OR: 1.427;95%CI:1.123-1.813;P=0.004);The Logistic regression result after LSP function conversion showed low FB(≤0m L/kg/48h)(OR: 1.012;95%CI: 1.002-1.023;P=0.025)and medium FB(0-60 m L/kg/48h)(OR: 1.007;95%CI: 1.002-1.013;P=0.012)was significantly associated with higher in-hospital mortality.The results of rank variables showed that FAI(>0.42/48h)was significantly correlated with high in-hospital mortality(OR: 1.725;95%CI: 1.322-2.266;P<0.001);After LSP function conversion,FAI(≤0/48h)(OR: 1.454;95%CI: 1.117-1.989;P=0.011)and FAI(>0.42/48h)(OR:8.859;95%CI: 4.002-19.775;P<0.001)was significantly associated with higher in-hospital mortality.(3)In the multivariate Logistic regression model,after adjusting for confounding factors,FAI was significantly correlated with in-hospital mortality,while FB and FI were not.Multivariate Logistic regression analysis with rank variable conversion showed that FAI(>0.42/48h)was significantly correlated with in-hospital mortality(OR: 1.472;95%CI: 1.111-1.963;P=0.008).In multivariate Logistic regression analysis after LSP function conversion,the study found a similar association between FAI and in-hospital mortality,with FAI(≤0/48h)(OR: 1.408;95%CI: 1.088-1.915;P=0.017)and FAI(>0.42/48h)(OR: 4.683;95%CI: 2.018-10.927;P=0.048)was significantly positively associated with higher in-hospital mortality.Conclusion:FAI(> 0.42/48h)was associated with higher in-hospital mortality in patients with sepsis and heart failure.Appropriate fluid management in these patients may improve outcomes. |
