Association of extracellular water/total body water ratio with protein-energy wasting and mortality in patients on hemodialysis

Bioimpedance analysis-assessed extracellular water/total body water (ECW/TBW) ratio may be a marker for mortality and poor nutritional status in hemodialysis patients. In 193 maintenance hemodialysis patients, we retrospectively investigated the relationships among ECW/TBW ratio, mortality, and protein-energy wasting (PEW). Four components—body mass index, normalized protein catabolic rate, normalized serum creatinine level, and serum albumin level—constitute the simple PEW score; this score was calculated based on the positive number of items concerning malnutrition among these four components. A score ≥ 3 indicated PEW. Patients were stratified by an ECW/TBW ratio cut-off value (0.40) and by PEW versus non-PEW status. The simple PEW score, cardiothoracic ratio, and log-transformed C-reactive protein level were independently correlated with the ECW/TBW ratio. Eighty-four patients died during follow-up (median 4.3 years). After adjustments for sex, age, hemodialysis vintage, histories of cardiovascular events and diabetes, and C-reactive protein level, a higher ECW/TBW ratio and PEW were independently related to elevated risks of all-cause death. Adding the ECW/TBW ratio to a baseline risk model including PEW significantly increased C-statistics from 0.788 to 0.835. In conclusion, the ECW/TBW ratio may be an indicator of PEW and may be a predictor of death even accounting for PEW, in hemodialysis patients.


Model discrimination.
The C-index predicting all-cause death improved significantly from 0.788 to 0.835 (p = 0.038) when adding ECW/TBW ratio to the baseline risk model, which included age, sex, hemodialysis vintage, history of diabetes and CVD events, CRP, and PEW (Table 4).

Discussion
In patients receiving hemodialysis, the ECW/TBW ratio was independently correlated with simple PEW score, Log-transformed CRP, and CTR.In this population, those with greater ECW/TBW ratios and those who had PEW had higher likelihoods of dying from any cause.Those who had PEW and an elevated ECW/TBW ratio had the highest probability of dying from any cause.Additionally, the prediction of mortality was much improved when adding the ECW/TBW ratio to the simple PEW score-included risk model.The ECW/TBW ratio-a marker of PEW, inflammation, and fluid overload-may therefore be a strong predictor of mortality.
PEW, a wasting of muscle and fat, is common and highly related to mortality in patients undergoing hemodialysis.We have recently proposed that the ECW/ICW ratio, a novel but minor BIA-measured parameter, may be a marker of PEW with the ECW/ICW ratio being independently associated with the GNRI 11 .In addition, the repeated measurements of the ECW/ICW ratio and the GNRI increased the predictability for mortality 17,18 .Moreover, the combination of the ECW/ICW ratio and the GNRI improved the predictive accuracy of mortality 11 .The current study evaluated the association between the ECW/TBW ratio, one of the major established BIAmeasured parameters, and PEW.PEW was diagnosed using established diagnostic criteria and not a surrogate marker.The usefulness of the WCW/TBW ratio for predicting all-cause mortality with adjustment for PEW was also investigated.
In the present study, the ECW/TBW ratio had independent and positive associations with the simple PEW score, log-transformed CRP, and CTR in patients on hemodialysis.To our knowledge, this study is the first to directly examine the association between the ECW/TBW ratio and PEW.In this present study, the ECW/ TBW ratio had a significant negative correlation with each item of the simple PEW score except for BMI: Alb, r = − 0.508; Cre/BSA, r = − 0.446; nPCR, r = − 0.416; and BMI, r = 0.031 (p = 0.66).According to the original PEW criteria 1 , body mass is assessed using the BMI, body fat percentage, and unintentional weight loss.As ECW/TBW ratio had a significant association with Cre/BSA but not BMI, this might be an indicator of muscle wasting and www.nature.com/scientificreports/ each other.The possible pathophysiology may be that fluid overload can introduce bowel edema; thereafter, the translocation of bowel endotoxin into the circulation may induce inflammation 21 .The inflammation can lead to malnutritional status owing to protein catabolism 22,23 , and then hypoalbuminemia may exacerbate the systemic edema.Thus, ECW/TBW ratio may indicate PEW, inflammation, and fluid overload.However, the present study was performed in a cross-sectional design; therefore, establishing causal relationships among variables is difficult.
A longitudinal study is needed to confirm these findings in the future.www.nature.com/scientificreports/In this study, PEW was evaluated using a simple PEW score, and the ARNOS prospective study validated its utility for predicting mortality 4 .The study divided patients into four groups as follows: (G1) severe wasting, simple PEW score of 0 and 1; (G2) moderate wasting, simple PEW score of 2; (G3) slight wasting, simple PEW score of 3; and (G4) normal nutritional status, simple PEW score of 4. Compared with G1, G3 and G4 (simple PEW score ≥ 3) had an independent correlation with an increased mortality risk.Additionally, the original diagnostic criteria for PEW required at least three of the four components for a diagnosis of PEW.Thus, in this study, patients with a simple PEW score ≥ 3 were diagnosed as having PEW.In the present study, PEW was independently associated with an elevated all-cause death risk.Moreover, patients with a higher ECW/TBW ratio and PEW (G4) were at the highest risk of death; therefore, the combination of the ECW/TBW ratio and PEW may be useful to stratify the risks of death.Interestingly, compared with G1 and G2, G3 and G4 were associated with elevated mortality risks.The predictability of all-cause death significantly improved after the ECW/TBW ratio was added to the simple PEW score-contained baseline risk model.Therefore, the ECW/TBW ratio may be a powerful predictor of all-cause mortality in hemodialysis patients with PEW and without PEW.This might be because ECW/TBW ratio is not merely an indicator of PEW but also an indicator of inflammation and volume overload.
There were some limitations to consider.First, we included only Japanese patients receiving hemodialysis; therefore, our results may not apply to hemodialysis patients of other ethnic groups.Second, this study was a retrospective single-center study with a relatively small number of hemodialysis patients.Third, the ECW/TBW ratio was measured during study enrollment; therefore, the changes in the ECW/TBW ratio during the followup periods were not evaluated.To validate our study findings, multicenter prospective studies with large sample sizes and other ethnic groups may be needed in the future.
In conclusion, for hemodialysis patients, the simple PEW score, CRP, and CTR were independently correlated with the ECW/TBW ratio.An elevated ECW/TBW ratio was linked to greater risks of all-cause death.Moreover, adding the ECW/TBW ratio to the simple PEW score-containing baseline risk model significantly improved the prediction of death.The ECW/TBW ratio may not only be useful to predict PEW but also to predict inflammation and volume overload and, therefore, may be a powerful predictor of mortality.

Methods
Participants included in this study.Patients who had stably undergone hemodialysis (4 h/session, 3 times a week) for more than 6 months were retrospectively included.At the Matsunami General Hospital, body composition was measured using BIA between January 2008 and December 2015.Because the current investigation was carried out retrospectively by evaluating data gathered from routine clinical practice, the necessity to obtain prior informed consent was waived by the "Ethics Committee of Matsunami General Hospital"."Ethics Committee of Matsunami General Hospital" approved the study plan, which was performed adhering to the principles of the Declaration of Helsinki (Approval No. 541).
Data collection.The baseline information of each participant obtained from medical records was: sex; age; kidney disease that led to end-stage kidney disease, such as diabetic nephropathy, nephrosclerosis, chronic glomerulonephritis, and others; hemodialysis duration; smoking and alcohol habits; and histories of diabetes, hypertension, and cardiovascular disease (CVD) events.Diabetes was defined based on having received a previous diagnosis of diabetic retinopathy and/or a history of using glucose-lowering medication.Hypertension was defined by two parameters: (1) using some anti-hypertensive drugs and/or (2) a pre-hemodialysis blood pressure ≥ 140/90 mmHg.The CVD events included peripheral artery disease, myocardial infarction, heart failure, angina pectoris, and stroke (hemorrhage or infarction).At the start of the week, blood tests were taken in the supine position before and after a hemodialysis session.BIA was performed in the supine position about 30 min after the hemodialysis session, and ICW, ECW, and TBW values were obtained by a body composition analyzer with multiple frequencies (2.5-300 kHz) using the ankle and wrist method (MLT-550N; Sekisui Medical, Siga, Japan).The measurement was performed on Wednesday or Thursday.The CTR was calculated based on chest radiography which was performed on the same day.BMI was calculated from dry weight and height: BMI = dry weight/height 2 (kg/m 2 ).
Definition and calculation of the simple PEW score.The simple PEW score was defined as 4 the positive number of items regarding malnutrition among the four components (i.e., Alb, Cre/BSA, BMI, and nPCR).The cut-off values of Alb, Cre/BSA, BMI, and nPCR were defined as 3.8 g/dL, 380 μmol/L/m 2 , 23 kg/m 2 , and 0.8 g/kg/day, respectively.In this study, accounting for the original version of PEW 1,24 , PEW was diagnosed when at least three of the four components were applied.

Table 1 .
Baseline data of the study participants.G1 non-PEW and low ECW/TBW ratio group, G2 PEW and low ECW/TBW ratio group, G3 non-PEW and high ECW/TBW ratio group, G4 PEW and high ECW/TBW ratio group, G1-G4 groups 1-4, BMI body mass index, BUN blood urea nitrogen, Ca calcium, Cre/BSA serum creatinine level adjusted by body surface area, Alb serum albumin, CRP C-reactive protein, nPCR normalized protein catabolic rate, CTR cardiothoracic ratio, iPTH intact PTH, P phosphorous, PEW protein-energy wasting, TBW total body water, ICW intracellular water, ECW extracellular water.

Table 2 .
Regression analyses of the associations of the ECW/TBW ratio with baseline parameters.CVD cardiovascular disease, CTR cardiothoracic ratio, CRP C-reactive protein, Alb serum albumin, BMI body mass index, Cre serum creatinine, Cre/BSA serum creatinine level adjusted by body surface area, nPCR normalized protein catabolic rate, TBW total body water, ECW extracellular water, PEW protein-energy wasting.
Figure1.Kaplan-Meier analyses for all-cause mortality.The survival rates for the (a) two groups with a lower ECW/TBW ratio (ECW/TBW ratio ≤ 0.40) and a higher ECW/TBW ratio (ECW/TBW ratio > 0.40); (b) two groups with PEW and without PEW; and (c) four groups with a lower ECW/TBW ratio and no PEW (group 1: G1), a lower ECW/TBW ratio and PEW (group 2: G2), a higher ECW/TBW ratio and no PEW (group 3: G3), and a higher ECW/TBW ratio and PEW (group 4: G4).PEW protein-energy wasting, ECW extracellular water, TBW total body water.

Table 3 .
The Cox analysis of the ECW/TBW ratio and PEW with all-cause mortality.ECW extracellular water, TBW total body water, PEW protein-energy wasting.*Adjustments for age, sex, hemodialysis vintage, history of diabetes and cardiovascular disease, and C-reactive protein.

Table 4 .
The predictability of the ECW/TBW ratio with all-cause mortality.NRI net reclassification improvement, ECW extracellular water, TBW total body water, CVD cardiovascular disease.*Included age, sex, hemodialysis vintage, history of diabetes and CVD, C-reactive protein, and simple protein-energy wasting score.