SUN-063 MORTALITY RISK OF FLUID OVERLOAD IN DIABETIC ESRD PATIENTS

Abstract

Chronic fluid overload has been related to severely increased mortality in ESRD patients. In addition, the increasing prevalence of diabetes along with concomitant diseases make fluid management more difficult and calls for individualized fluid targets, taking co-morbidities into account. Aim of this work was to assess the mortality risk of diabetes in different degrees of chronic fluid overload, and to identify potential target ranges for optimal outcome. Data from more than 90,000 patients treated in the Fresenius NephroCare clinic network from 26 countries were included in the analysis, including 1.3 million measurements of fluid status (determined as fluid overload, FO [L]) by bioimpedance spectroscopy using the BCM Body Composition Monitor. Hazard ratios (HR) were calculated by time dependent cox analysis (adjusted for 30 different parameters) over deciles in both diabetic and non-diabetic patients. In both groups, the decile with best outcome served as reference (HR=1). Prevalence of diabetes was strongly related with pre-dialysis fluid status (Figure 1). In patients who were normally hydrated (or even slightly dehydrated), prevalence of diabetes was around 15%, rising to threefold (>45%) in states of severe fluid overload. The HR of pre-dialysis fluid overload in the highest decile was 2.7 in non-diabetics, and 2.0 in diabetics (Figure 2). The HR of FO was lower in diabetics since the underlying mortality is generally higher in this more co-morbid group, and fluid overload (while still being a risk factor) adds less relative hazard on top as compared with non-diabetics. Dehydration was found to be a significant risk factor in both groups. The pre-dialysis fluid overload range in which the mortality risk is increased by no more than 50% compared with the reference was relFO<15% (relFO=FO/ECW, ECW: extracellular water) in non-diabetics, and relFO<22% in diabetics.Figure 2Adjusted hazard ratios for all-cause mortality in non-diabetic (left) and diabetic patients (right). In both diagrams patients were separated into 10 deciles of relative pre-dialysis fluid overload. The third decile (corresponding with the best outcome) was used as reference in both cox models.View Large Image Figure ViewerDownload Hi-res image Download (PPT) Both the prevalence of diabetes and the risk of mortality in diabetic patients strongly depend on degree of long-term chronic fluid overload. Whether there is a causal relationship or this is just an association needs to be further investigated.