Fluid balance and mortality in critically ill patients with acute kidney injury: a multicenter prospective epidemiological study

In this study, we used a database from a prospective, multicenter, observational study in which investigators examined the epidemiology of AKI in critically ill patients at 30 ICUs among 28 tertiary hospitals in Beijing, China, from 1 March to 31 August 2012 (the Beijing Acute Kidney Injury Trial [BAKIT]) [19]. For a complete list of these hospitals and the persons responsible for the data acquisition, see Additional file 1. Study subjects included all adult patients (age ≥18 years) admitted consecutively to ICUs. Only the initial ICU admission was considered in this study. The following patients were excluded: patients with preexisting end-stage chronic kidney disease, patients already undergoing RRT before admission to the ICU, and patients who had received kidney transplants in the previous 3 months. Preexisting comorbidities were diagnosed based on International Classification of Diseases, Tenth Revision, codes. Among the 3107 patients who were admitted consecutively, 2526 patients with their first 3 days of sequential data were included in our study (Fig. 1).

The collected data included demographics, anthropometrics, admission diagnosis, comorbidities, daily vital signs, and laboratory values, which were to calculate automatically the Acute Physiology and Chronic Health Evaluation II (APACHE II) [20], the Simplified Acute Physiology Score II (SAPS II) [21], and the Sequential Organ Failure Assessment (SOFA) score [22]. RRT and mortality data were also reported. Furthermore, hourly urine output, fluid balance, and use of diuretics were recorded daily. Fluid balance was calculated daily as the difference between fluid intake and fluid output. Fluid output included all body fluids, including urine, excrement, and, if applicable, dialysis ultrafiltrate. The patients were followed until death, hospital discharge, or for 28 days.

This study was approved by the institutional review board of the ethics committee of the lead study center (Fu Xing Hospital, Capital Medical University, Beijing China) and all other participating hospitals (see Additional file 2). The requirement for informed consent was waived for this observational survey. The patient records and information were anonymized and deidentified before analysis.

The AKI cohort consisted of patients who developed AKI within the first 72 h. Patients who developed AKI after the third day (e.g., on day 4) were classified as non-AKI. AKI severity was classified according to the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines [23]. AKI was staged for severity (stages 1, 2, and 3) based on serum creatinine (SCr) or urine output or both. Baseline creatinine was defined as the lowest known SCr value during the previous 3 months [24]. For patients without these values or without renal failure, the baseline SCr was estimated by using the Modification of Diet in Renal Disease equation [25], assuming a GFR of 75 ml/min/1.73 m² [26]. For patients with chronic renal failure but not receiving dialysis, the initial SCr value upon admission was used as the baseline value [26]. We regarded the worst stage during the first 3 days as the AKI stage.

Oliguria was defined as urine output less than 500 ml/day. The daily fluid balance was recorded, and the cumulative fluid balance was registered at 24, 48, and 72 h. We calculated the daily AKI numbers when we compared daily fluid balance between the AKI and non-AKI groups, but we included all patients with AKI in the first 3 days when we compared the cumulative fluid balance. To quantify the cumulative fluid balance over 3 days in relation to body weight, we used the following formula: sum of daily (fluid intake [liters] − total output [liters])/body weight (kilograms). We used the term percentage of fluid accumulation to define the percentage of cumulative fluid balance adjusted for body weight. Baseline body weight was recorded at initial hospital admission. We defined FO as fluid accumulation greater than 10 % of the baseline weight [27].

Non-normally distributed continuous variables were expressed as median with interquartile range (IQR) and were compared using the Mann–Whitney U test or the Kruskal–Wallis analysis of variance test with the Bonferroni correction. Categorical variables were expressed as the number of cases and proportions and were compared using the Mantel–Haenszel χ² test. We used a logistic regression model to evaluate the effect of FO on the incidence of AKI. We conducted exploratory univariate analysis of several variables to identify possible confounders associated with 28-day mortality and to assess the influence of cumulative fluid balance on the survival time of patients with AKI. A multivariate Cox regression analysis was then performed using a backward stepwise selection method, with a P value less than 0.05 as the entry criterion and a P value 0.10 or higher as the removal criterion. The assumption of proportional hazards was checked graphically using log (−log [survival probability) plots and was found to be appropriate. Variables considered for multivariable analysis included age, comorbid diseases, diagnosis, illness severity scores, cumulative fluid balance at 72 h, oliguria, and sepsis (similarly to a previous study) [28]. We tested for colinearity among all variables using a Cox regression analysis to generate hazard ratios and 95 % confidence intervals (CIs). The 28-day survival stratified by the presence or absence of FO was additionally evaluated graphically using the Kaplan–Meier product limit survival plot. All statistical analyses were performed using IBM SPSS 17.0 software package (IBM, Armonk, NY, USA), with a two-sided P value less than 0.05 considered statistically significant.