Interactions Between Malnutrition, Inflammation, and Fluid Overload and Their Associations With Survival in Prevalent Hemodialysis Patients

ORIGINAL RESEARCH

Interactions Between Malnutrition, Inflammation, and Fluid Overload and Their Associations With Survival in Prevalent Hemodialysis Patients Marijke J. E. Dekker, MD,*,† Constantijn Konings, MD, PhD,† Bernard Canaud, MD, PhD,‡ Frank M. van der Sande, MD, PhD,* Stefano Stuard, MD, PhD,‡ Jochen G. Raimann, MD, PhD,§ € urk, MD,* Len Usvyat, PhD,§,{ Peter Kotanko, MD, FASN,§,** and Elife Ozt€ Jeroen P. Kooman, MD, PhD* Objective: Predialysis fluid overload (FO) in hemodialysis (HD) patients is associated with an increased risk of death, further increased by the presence of inflammation. Malnutrition is also associated with outcome. Study objectives were, firstly, to investigate if the presence of FO is associated with malnutrition and whether this association is influenced by the presence of inflammation. Second, we assessed the associations of FO, malnutrition, and inflammation with outcome individually and in combination. Design: International cohort study. Setting: European patients of the Monitoring Dialysis Outcome Initiative cohort where bioimpedance and C-reactive protein measurements are performed as standard of care. Subjects: 8883 prevalent HD patients. Main Outcome Measure: Body composition, nutritional and inflammation status were assessed during a 3-month baseline period, and all-cause mortality was noted during 1 year follow-up. Malnutrition was defined as a lean tissue index ,10th percentile (of age and gender matched healthy controls), FO as a predialysis overhydration .11.1 L and inflammation as a C-reactive protein . 6.0 mg/L. We used Cox models to investigate the association with outcome. Results: The presence of malnutrition was associated with higher levels of FO, this amount further increased when inflammation was present. Only 11.6% of the patients did not have any of the 3 risk factors and only 6.5% of the patients were only malnourished, which was not associated with an increased risk of death (Hazard Ratio [HR] 1.22 [95% Confidence Interval [CI]: 0.75-1.97]), whereas the combination of severe malnutrition, FO, and inflammation comprised the highest risk of death (HR 5.89 [95% CI: 2.28-8.01]). Conclusion: In HD patients, predialysis FO associates with both malnutrition and the presence of inflammation, with the highest levels of FO observed when both are present. Malnutrition as singular risk factor was not associated with increased mortality risk. The highest mortality risk was observed in patients where all 3 risk factors were present. Ó 2018 by the National Kidney Foundation, Inc. All rights reserved.

Introduction

R *

Maastricht University Medical Center, Maastricht, The Netherlands. Catharina Hospital Eindhoven, Eindhoven, The Netherlands. ‡ Fresenius Medical Care, Bad Homburg, Germany. § Renal Research Institute, New York, New York. { Fresenius Medical Care North America, Waltham, Massachusetts. ** Mount Sinai Hospital, New York, New York. Support: No financial support was received for this study. Financial Disclosure: B.C., B.C., S.S., and L.U. are employees of Fresenius Medical Care and hold shares in the company. P.K. holds stocks in Fresenius Medical Care. All the other authors declared no competing interests. Address correspondence to: Marijke J.E. Dekker, MD, Division of Nephrology, Department of Internal Medicine, Maastricht University Medical Center, Post Box 5800, Maastricht, 6202 AZ, The Netherlands. E-mail: †

[email protected] Ó

2018 by the National Kidney Foundation, Inc. All rights reserved. 1051-2276/$36.00 https://doi.org/10.1053/j.jrn.2018.06.005

Journal of Renal Nutrition, Vol -, No - (-), 2018: pp 1-10

ECENT STUDIES HAVE shown a strong prognostic impact of fluid overload (FO), assessed by multifrequency bioimpedance spectroscopy (MF-BIS).1-4 It has been long known that malnutrition, defined by various parameters such as hypoalbuminemia, low subjective global assessment, and a low body mass index (BMI), is also an important risk factor for adverse outcomes,5 which was recently also confirmed for body composition assessed by MF-BIS.6 Previous studies showed the strength of an integrative interpretation of risk factors in dialysis patients, with as most pronounced example the malnutrition-inflammation-atherosclerosis complex.7 However, although the relation between FO and outcome in dialysis is well accepted, this parameter is often interpreted in isolation, whereas it may be part of a wider syndrome of adverse risk factors, such as protein energy wasting or inflammation, which may warrant additional interventions.8-10 Moreover, FO may influence the risk profile 1

2

DEKKER ET AL

associated with inflammation and malnutrition. Recently, our study group showed that depletion of lean tissue mass (LTM), assessed by MF-BIS, and especially when combined with a depletion in fat tissue mass (FTM), was associated with adverse outcomes.6 In a recent study, we showed that predialysis FO .11.1 L in hemodialysis (HD) patients was already associated with an adverse prognosis and that the adverse effect of FO on outcome was multiplied by the concomitant presence of inflammation.4 Another study showed that FO, malnutrition (defined as a depletion of LTM below the 10th percentile of a reference population) were all predictors of mortality.4 However, the relation between depletion in LTM and/or FTM and FO and the combined effects on outcome were not assessed in this study. Another recent study showed the protective effect of high BMI only held true for patients with concomitant inflammation, but they did not distinguish between different body compartments.11 To the best of our knowledge, the combined effects of FO, inflammation, and malnutrition on outcome have not yet been investigated. As most of the studies assessing the relation between FO and markers of malnutrition were carried in single center studies and their prognostic interaction was not assessed, we sought to explore these matters in an international database containing detailed information on fluid status, body composition, as well as inflammatory markers. The aim of this study was to test the following hypotheses, firstly, that the presence of malnutrition is associated with FO, and this association is further influenced by the presence of inflammation. Secondly, that the associations of FO, malnutrition, and inflammation with outcome add to a cumulative risk profile.

Methods Patient Selection For this observational study, the European subset of the international Monitoring Dialysis Outcome Initiative database was used because in these patients, MF-BIS measurements were performed as routine care at least quarterly. This database comprises HD patients data of Fresenius Medical Care clinics in 16 European countries. Only patients with both a C-reactive protein (CRP) and an MF-BIS assessment in 2011 were included (Fig. 1). The 90 days before the last MF-BIS assessment of the year 2011 were defined as the baseline period, of which averaged of laboratory and clinical data were used for the analyses. The baseline period was directly followed by 1-year follow-up period during which the outcome all-cause mortality was noted (Fig. 2). Patients based on country of dialysis, patients were stratified into 4 regions (north, east, south, and west) based on the United Nations geographical scheme.12 Clinical and laboratory data, including comorbidities were derived from electronic health records. Patients gave informed consent for the use of their data in anonymized form, for data collection and analysis all standards are followed.13,14 This research was exempt from IRB review (for detailed information see supplemental material). Determination and Characterization of Body Composition Factors Body composition was assessed by using a multifrequency bioimpedance spectroscopy (MF-BIS) device (Body Composition Monitor; Fresenius Medical Care, Bad Homburg, GermanyÒ). The device presents body

Figure 1. Study flow chart. CRP, C-reactive protein; HD, hemodialysis; MF-BIS, multifrequency bioimpedance spectroscopy.

3

NUTRITION, INFLAMMATION, AND FLUID IN HD

Figure 2. Study design. Alle paitnes of the European subset of the MONDO initiative database with at least 1 MF-BIS and 1 CRP measurement in 2011 were included. All document data points, available of the 3 months baseline period, were used to determine averaged of laboratory and clinical parameters. MF-BIS, multifrequency bioimpedance spectroscopy; MONDO, Monitoring Dialysis Outcome.

composition as a 3-compartment model with LTM, FTM, and a separate compartment of overhydration (OH) (as a surrogate marker of extracellular FO). The LTM and FTM are also corrected for BMI and presented as a lean tissue index (LTI) and a fat tissue index (FTI) and by using an age and gender matched healthy control population patients with the lowest percentile, below the 10th percentile of the reference population, can be identified. Malnutrition was therefore classified as an LTI below the 10th percentile of the reference population and in subset analysis as a combined LTI and FTI below the 10th percentile. FO was classified as OH.11.1 L based on previous studies and the upper reference limit of the nomogram used by the device.15 The majority of the body composition measurements were performed midweek (81%), and all were performed predialysis in seating position in the dialysis chair.

Categorization of Other Risk Factors and Confounders The presence or absence of inflammation was based on, the baseline periods average CRP values .6.0 mg/L and #6.0 mg/L, respectively. Ultrafiltration volume (L) was calculated by subtracting postdialysis weight from predialysis weight. Statistical Analyses Continuous variables are reported as mean 6 standard deviation (SD) or median and the 25th to 75th percentile depending on their distribution. Comparison of continuous variables was performed via 1-way Analysis of Variance (ANOVA), and categorical variables were compared with chi-square analyses. We used Cox models with

different levels of adjustment. Model I: unadjusted. Model II: adjusted for age (years), gender (male/female), vintage (years), region (north, east, south, west), access type (catheter/fistula). Model III: adjusted for Model II and interdialytic weight gain (kg), BMI (kg/m2), serum albumin (g/ dL), predialysis systolic blood pressure (mm Hg), serum sodium (mEq/L), ultrafiltration rate (mL/h/kg) (stratified in ,10, 10-13 and . 13), and the presence of the comorbidities diabetes mellitus, congestive heart failure, cerebrovascular disease, peripheral vascular disease, and the presence of cancer. All analyses were performed with SPSS statistics, version 23.0 (IBM Corp., Armonk, NY).

Outcome The recorded outcome of this study was all-cause mortality based on ICD-9 codes.

Results Baseline/Cohort Characteristics In this cohort, 8,883 patients were included, see Figure 1 for the flow chart. Patients’ baseline characteristics of the total study cohort are displayed in Table 1. Patients’ characteristics for patients stratified on nutritional and predialysis fluid status level are presented in Table 2. In all the different subgroups, a mean dialysis vintage above 3 years was observed with the highest dialysis vintage observed in patients with FO and inflammation. Patients age was comparable across the different subgroups with the lowest age, 61.10 years, observed in patients with only malnutrition and the highest, 67.70 years, in patients with FO and inflammation. In patients with no risk factors,

4

DEKKER ET AL

Table 1. Patients Characteristics at Baseline for the Total Cohort (N 5 8883) Parameters Males (%) Age (years) Dialysis vintage (years) Diabetes mellitus present (%) Catheter access (%) Region* North (%) East (%) South (%) West (%) Dialysis treatment characteristics Predialysis weight (kg) Postdialysis weight (kg) Body mass index (kg/m2) Ultrafiltration rate (mL/kg/h) Interdialytic weight gain (kg) Predialysis systolic blood pressure (mmHg) Predialysis diasystolic blood pressure (mmHg) Body composition parameters Normohydration (NH) weight (kg) Predialysis overhydration (OH) (L) Predialysis OH corrected for NH-weight (%) Postdialysis OH (L)# Lean tissue index ,10th percentile (%) Fat tissue index ,10th percentile (%) Laboratory values Normalized protein catabolic rate (g/kg/day) Serum albumin (g/dL) C-reactive protein (CRP) (mg/L) Log(10) CRP (mg/L) Creatinine (mg/dL)

Mean/Median 57.2 63.47 3.60 18.6 18.6

SD/25th-75th Percentile

14.77 1.63-6.94

0.2 46.7 49.4 3.7 72.67 70.61 25.96 7.36 2.10 139.06 71.16

16.01 15.69 5.23 2.85 0.84 19.88 12.16

70.79 1.66 2.41 20.33 45.5 19.2

16.05 0.78 to 2.62 1.10 to 3.95 21.32 to 0.66

1.07 3.85 5.7 1.71 7.95

0.13 0.39 2.4-13.0 1.27 2.27

*Countries contributing to the different regions North: United Kingdom; East: Bosnia, Czech Republic, Hungary, Romani, Russia, Slovaki; South: Italy, Portugal, Serbia, Slovenia, Spain, Turkey; West: France. # Calculated: postdialysis fluid status (L) 5 (postdialysis weight [kg] 2 predialysis weight [kg]) 1 predialysis fluid status (L).

a remarkable lower incidence of diabetes mellitus was observed. Many other differences between the different parameters (as weight, blood pressure and laboratory parameters [albumin, sodium]) can partly be explained by the categorization of the patients in these different subgroups.

Predialysis Fluid Status in Patients With Malnutrition and Inflammation Levels of predialysis FO were higher in patients with malnutrition. This difference in predialysis fluid status between patients with or without malnutrition was further enlarged when inflammation was present (Fig. 3, supplemental material Table 1). The highest levels of FO predialysis (mean 3.06 L [95% CI: 2.79-3.34]) were observed in patients with both LTI and FTI ,10th percentile and inflammation present. Prevalence of Malnutrition, Fluid Overload and Inflammation Of the total cohort of this study, no risk factor was present in only 11.9% of the patients. Predialysis FO was the

most observed solitary present risk factor (18.6%) (Fig. 4). In 17.6% of the patients, all 3 risk factors were present.

Association of Malnutrition, Fluid Overload, and Inflammation With Survival The presence of more than one risk factor was associated with an incremental increasing risk of mortality. When only malnutrition was present as a risk factor, we did not find a significant association with survival in all Cox models with different levels of adjustment (HR 1.22 [95% CI 0.751.97]) (Fig. 5, supplemental material Table 2). The highest risk of death was observed when all 3 risk factors were present (HR 5.89[95% CI 4.28-8.10]) (Fig. 5, supplemental material Table 2). Subanalyses When all analyses were performed with a more extended adjusted model, the results were comparable with the results mentioned above (supplemental material Table 2). Owing to the remarkable finding that malnutrition (defined as LTI,10th percentile) as single risk factor was not associated with an increased mortality risk, we also performed all the

Table 2. Patients Characteristics of the Study Cohort Stratified by Predialysis Fluid Status No Risk Factor Parameters

1060 (11.9) 42.7 61.49 3.18 10.1 12.5

SD/25th-75th Percentile

16.05 1.56-5.93

0.1 43.2 52.4 4.3

Only Fluid Overload Mean/ Median 1652 (18.6) 54.8 63.42 3.71 15.6 14.0

Only Inflammation

SD/25th-75th Percentile

Mean/ Median

15.37 1.72-7.10

799 (8.8) 41.9 65.29 3.03 13.9 20.2

0.1 55.1 42.4 2.3

SD/25th-75th Percentile

14.63 1.56-5.78

0.3 41.8 54.7 3.3

Only Malnutrition Mean/ Median

SD/25th-75th Percentile

573 (6.5) 49.9 61.10 15.24 3.40 1.56-7.13 13.6 13.6 0.0 38.0 55.8 6.1

71.07 69.17 26.29 6.91 1.92 136.82

15.27 15.04 5.11 2.96 0.81 18.09

71.25 69.12 25.19 7.66 2.15 142.99

14.72 14.45 4.37 2.86 0.80 18.19

76.27 74.31 28.58 6.66 2.00 137.18

17.17 16.85 6.20 2.91 0.88 20.22

72.05 70.10 26.42 7.06 1.98 132.28

15.94 15.60 5.36 3.05 0.89 19.00

70.78

11.88

72.40

12.15

70.29

11.91

69.54

11.81

14.64 1.57-2.83 2.31-4.24

76.13 0.30 0.40

17.31 20.39-0.76 20.51-1.04

70.78 0.36 0.52

15.33 20.22-0.77 20.31-1.13

68.81 2.10 3.13

21.59 22.9

22.40 to 20.89

0.12 33.1

20.58-0.83

21.74 15.5

22.65 to 21.05

71.68 16.01 0.52 0.01-0.82 0.73 0.01-1.20 21.53 22.26 to 20.87 5.4

1.08

0.12

1.06

0.12

1.07

0.13

1.09

0.12

3.97 2.40 0.71 8.50

0.33 1.24-4.00 0.85 2.36

3.95 2.49 0.72 8.20

0.34 1.20-3.95 0.81 2.22

3.86 12.00 2.62 7.93

0.39 8.00-12.00 0.65 2.27

3.89 2.60 0.80 8.13

0.34 1.50-4.10 0.80 2.39 (Continued )

NUTRITION, INFLAMMATION, AND FLUID IN HD

Number of patients (% of total) Males (%) Age (years) Dialysis vintage (years) Diabetes mellitus present (%) Catheter access (%) Region* North (%) East (%) South (%) West (%) Dialysis treatment characteristics Predialysis weight (kg) Postdialysis weight (kg) Body mass index (kg/m2) Ultrafiltration rate (mL/kg/h) Interdialytic weight gain (kg) Predialysis systolic blood pressure (mm Hg) Predialysis diasystolic blood pressure (mm Hg) Body composition parameters Normohydration (NH) weight (kg) Predialysis overhydration (OH) (L) Predialysis OH corrected for NHweight (%) Postdialysis OH (L)† Fat tissue index ,10th percentile (%) Laboratory values Normalized protein catabolic rate (g/kg/d) Serum albumin (g/dL) C-reactive protein (CRP) (mg/L) Log(10) CRP (mg/L) Creatinine (mg/dL)

Mean/ Median

5

No Risk Factor Parameters

SD/25th-75th Percentile

Mean/ Median

SD/25th-75th Percentile

Fluid Overload and Malnutrition

Fluid Overload and Inflammation

1333 (15) 69.7 59.66 4.01 26.5 17.6

1333 (15) 57.8 67.70 3.65 16.4 22.2

14.73 1.85-8.07

0.4 47.6 47.0 5.0

13.78 1.51-7.12

0.2 50.6 46.6 2.6

Only Inflammation Mean/ Median

SD/25th-75th Percentile

Inflammation and Malnutrition 568 (6.4) 47.7 65.68 3.11 18.0 25.7

13.47 1.40-5.91

0.4 31.2 64.6 3.9

Only Malnutrition Mean/ Median

SD/25th-75th Percentile

All 3 Risk Factors 1565 (17.6) 72.5 63.64 13.06 3.90 1.71-7.60 27.0 23.8 0.4 47.1 48.4 4.2

71.41 69.20 24.57 8.00 2.23 140.71

15.14 14.85 4.55 2.81 0.82 20.62

73.72 71.63 26.28 7.24 2.10 141.00

16.13 15.77 5.19 2.71 0.83 19.06

73.56 71.58 27.47 6.98 2.01 132.85

17.48 17.14 6.03 2.72 0.83 21.41

73.59 71.42 25.44 7.52 2.18 139.05

16.87 16.47 5.14 2.71 0.86 20.90

72.56

11.64

70.74

12.29

68.77

12.61

71.15

12.51

68.76 2.28 3.49

15.07 1.73-3.17 2.51-4.86

70.89 2.25 3.35

15.96 1.66-3.17 2.35-4.71

73.10 0.53 0.72

17.58 0.01 to 20.85 0.01-1.19

70.66 16.86 2.46 1.75-3.40 3.58 2.52-5.06

0.32 29.7

20.39-1.20

21.54 2.8

22.30 to 20.79

0.49 20.36-1.36 12.0

1.06

0.15

1.09

0.13

3.80 12.47 2.72 7.78

0.42 8.30-25.00 0.71 2.30

3.74 14.1 3.77 7.55

0.36 8.90-24.38 0.69 2.23

0.24 12.0

20.40-1.12

1.07

0.12

3.87 2.67 0.77 7.99

0.36 1.46-4.10 0.86 2.21

1.06

0.12

3.71 0.43 14.90 9.18-28.90 2.86 0.76 7.52 2.13

FO, fluid overload; LTI, lean tissue index. Data presented as mean (SD), or median (25th and 75th percentile). *Countries contributing to the different regions North: United Kingdom; East: Bosnia, Czech Republic, Hungary, Romani, Russia, Slovaki; South: Italy, Portugal, Serbia, Slovenia, Spain, Turkey; West: France. †Calculated: postdialysis fluid status (L) 5 (postdialysis weight [kg] 2 predialysis weight [kg]) 1 predialysis fluid status (L). Definitions of the risk factors subgroups: fluid overload: predialysis OH .11.1 L; inflammation: CRP-level above 6 mg/L; malnutrition: LTI ,10th percentile of an age and gender matched control population.

DEKKER ET AL

Number of patients (% of total) Males (%) Age (years) Dialysis vintage (years) Diabetes mellitus present (%) Catheter access (%) Region* North (%) East (%) South (%) West (%) Dialysis treatment characteristics Predialysis weight (kg) Postdialysis weight (kg) Body mass index (kg/m2) Ultrafiltration rate (mL/kg/h) Interdialytic weight gain (kg) Predialysis systolic blood pressure (mmHg) Predialysis diasystolic blood pressure (mmHg) Body composition parameters Normohydration (NH) weight (kg) Predialysis overhydration (OH) (L) Predialysis OH corrected for NHweight (%) Postdialysis OH (L)† Fat tissue index ,10th percentile (%) Laboratory values Normalized protein catabolic rate (g/kg/day) Serum albumin (g/dL) CRP (mg/L) Log(10) CRP (mg/L) Creatinine (mg/dL)

Mean/ Median

Only Fluid Overload

6

Table 2. Patients Characteristics of the Study Cohort Stratified by Predialysis Fluid Status (Continued )

NUTRITION, INFLAMMATION, AND FLUID IN HD

7

Figure 3. Levels of predialysis fluid overload (L) for patients stratified based on nutritional and inflammatory status. Levels of predialysis fluid overload (L) in patients with or without malnutrition, defined as an LTI or FTI or both LTI/FTI below the 10th percentile of an age and gender matched control population and with or without the presence of inflammation, defined as a CRP level above 6 mg/L, with 95% confidence intervals. See Supplemental material Table 1 for the exact numbers. CRP, C-reactive protein; FTI, fat tissue index; LTI, lean tissue index.

analyses with malnutrition defined as both an LTI and FTI below the 10th percentile. With these analyses, the same results were found (supplemental material Table 3), with again no increased mortality risk observed when only malnutrition was present (HR 1.86 [95% CI: 0.46-7.61]).

Discussion This study firstly showed that FO was associated with malnutrition. The highest levels of FO were observed in those patients with a depletion of both fat and LTM and concomitant inflammation. Moreover, the presence of FO and/or inflammation strongly modulates the mortality risk associated with malnutrition. Remarkably, in patients with low LTI levels, and even in those with where a com-

bination of both low LTI and FTI levels was present, the increased risk of mortality only became significant when accompanied with FO and/or inflammation. FO is very common in dialysis patients. According to MFBIS, 25% of peritoneal dialysis (PD) patients was classified as severely fluid overloaded, defined by an OH:extracellular water (ECW) index .15%, corresponding to an absolute FO of . 12.5L.1 Comparable observations have been done in HD patients.16 In agreement, in our cohort, 27.4% of patients had predialysis FO levels . 12.5L, whereas 39% of patients had moderate FO (predialysis FO .11.1L to 12.5L).4 There are several risk factors for FO in HD patients, such as the loss of residual renal function and high interdialytic

Figure 4. Prevalence numbers of the risk factors malnutrition, inflammation and fluid overload, and the percentages of patients in which these risk factors combined are present. CRP, C-reactive protein; FTI, fat tissue index.

8

DEKKER ET AL

Figure 5. Associations of the different risk factors alone and combined with mortality. Results of a Cox proportional hazard model adjusted for age (years), gender (male/female), vintage (years), region (north, east, south, west), and access type (catheter/fistula). See Supplemental material Table 2 for the exact numbers, and the results of the different Cox models and the additional subanalysis with malnutrition defined as both LTI and FTI ,10th percentile. CRP, C-reactive protein; LTI, lean tissue index; FTI, fat tissue index.

weight gain.17 In addition, results from smaller studies showed a relation between malnutrition and FO, both in HD and PD patients, as well as in patients with chronic renal failure not on dialysis.10,18,19 In 2 studies in PD patients, FO expressed as the extracellular to total body water ratio (ECW:total body water (TBW) ratio) was inversely related to nutritional markers such handgrip strength and subjective global assessment18 as well as body fat mass, serum albumin, and serum creatinine.9 However, as TBW is the sum of both intracellular and extracellular water, the ratio in these studies can be directly influenced by malnutrition itself.8 In a study of patients with chronic kidney disease (CKD) stages 3-5, using the OH index provided by the body composition monitor (BCM), FO assessed was associated with various cardiovascular risk factors, but also with markers of inflammation such as IL6 and tumor necrosis factor (TNF)-alfa, and inversely with serum albumin and LTI, assessed by MF-BIS.10 Our study expands on these earlier studies by providing a large international cohort, and detailed parameters of body composition, including not only LTI but also FTI. In previous studies in the same cohort, we showed that, serum albumin levels, as well as normalized protein catabolic rate (nPCR), but also LTI and FTI were incrementally lower in patients with mild and severe FO as compared to patients with normal predialysis fluid status. Also other studies have shown that, systemic inflammation, as reflected by increased CRP levels, was significantly higher in patients with FO.18,20 In our study, FO was increased in the presence of depletion of LTI, even most pronounced in the subgroup of patients with both low LTI levels and FTI levels, and highest in the subgroup of inflamed and malnourished patients, likely reflecting a subgroup with cachexia. Patients without inflammation and with both LTI and FTI above 10th percentile of the reference population had mean predialysis

FO levels of 11.40L, whereas this was 13.06L for patients with both LTI and FTI ,10th percentile in the presence of inflammation. The relation between FO and the depletion of body compartments might be due to several factors. The first is a progressive decline in body mass, which may go partly undetected and thus lead to incorrect prescription of dry weight. Alternatively, in patients with underlying illness, contributing to malnutrition, dry weight may be more difficult to achieve due to the increased risk of hemodynamic instability. In addition, hypoalbuminemia, related to malnutrition and/or inflammation may also lead to fluid redistribution to the interstitial compartment hampering fluid removal during dialysis.21,22 It should also be mentioned that the model used to assess FO in the 3-compartment model of the BCM is based on the theoretical assumption of a fixed hydration state of LTM and FTM, whereas it is not formally known whether hydration levels of FTM and LTM might differ in inflamed or malnourished patients.23,24 However, in this study, the MF-BIS appeared to be high predictive of outcome in the 3 different dimensions (LTI, FTI, and FO), adding information next to, but also independent of each other. To the best of our knowledge, our study is the first addressing the interaction between FO, malnutrition, and inflammation in detail. Interestingly, whereas the association with protein energy waisting (PEW) and mortality is well known, low LTI levels were not associated with a significant increase in 12 months mortality in normovolemic patients, even when accompanied by low FTI levels. In comparison, in patients with FO and/or inflammation, low LTI levels were associated with adverse outcomes with the highest risk in those patients in whom all 3 risk factors were present. This suggests a modulating effect of FO and/or inflammation with the outcome associated with malnutrition. This might firstly

NUTRITION, INFLAMMATION, AND FLUID IN HD

be explained by a direct causal effect of FO. Next to its effect cardiac dilatation and failure, but also enhance translocation of endotoxins into the circulation by bowel edema. Alternatively, given the fact that causality cannot be inferred from the present data, FO might, next to its cardiac effects, identify a phenotype of patients at higher risk for mortality. The absence of a direct association of malnutrition with increased mortality is in line with a recent study by Stenvinkel et al.,11 where the authors found that the effects of BMI on mortality was absent in HD patients without inflammation. Despite the high predictive value of the MF-BIS in this study, it cannot be used to assess its usefulness in clinical practice beyond its predictive power. For instance, it is not known whether attempts to achieve the so-called normohydration weight in inflamed and malnourished patients would lead to better outcomes or also carry dangers by exposing patients at higher risk for intradialytic hypotension. This should be assessed in properly powered clinical trials. In a randomized study by Hur et al, prescription of dry weight according to MF-BIS resulted in a reduction in blood pressure, arterial stiffness left ventricular mass, whereas 1 study showed a reduction in mortality.2 However, in this study, the mean age of the patients in the intervention and control groups was 51 and 52 years, respectively, with CRP levels of 1.05 and 1.16 mg/dL and serum albumin levels of 4.13 and 4.16 g/dL. In the randomized controlled study of Onofriescu, prescription of dry weight according to MFBIS resulted in a reduction in systolic blood pressure and arterial stiffness, whereas also a reduction in mortality in the intervention group was observed after a follow-up time of 3.5 years.25 The mean age in the intervention and control groups were 52 and 54 years, respectively, with mean serum albumin levels of 5 and 4 g/dL. Therefore, these studies might not be representative for the elderly, malnourished, and inflamed population in the different subgroups of this study, and we suggest that additional studies including high-risk patients are needed to definitely show the efficacy and safety of MF-BIS guided fluid management in this population. This study has the strength of assessing the relation and prognostic significance of 4 different dimensions (FO, depletion FTM and LTM, and inflammation), which may be amendable for therapeutic intervention, in a large international cohort. However, the database used was built primarily for clinical purposes, and therefore, some selection bias, for example, due to the fact that patients with more severe illness might have received more measurements, cannot be excluded. Moreover, data on residual renal function, which may be an important modifier of both inflammation and FO, are missing. Nevertheless, we suggest that this may not have had important effects on the main findings of the study, given the relatively high dialysis vintages of the cohort, which did not greatly deviate between sub-

9

groups with different levels of fluid status. Moreover, body composition was based on MF-BIS data and not on gold standard methods of body composition. However, given the excellent prognostic stratification, both in the dimension of body tissue as well as FO compartments, in which differences of 1.5-2 L were shown to have prognostic significance, we suggest that MF-BIS is a valid research tool in this respect. Concluding, this study shows that the risk of FO increases incrementally with more severe depletion of lean tissue and/or fat tissue compartments according to MFBIS and is further aggravated by the presence of inflammation. Whereas these dimensions have independent prognostic significance, as well as in combination, these different dimensions may carry a multiplicative hazard in the dialysis population.

Practical Application This study shows that the presence of FO is often part of a wider syndrome of adverse risk factors. We therefore suggest, based on these results, to not interpret FO as a singular risk factor and always address nutritional and inflammatory status. We also believe this approach can assist the clinician to help identify patients with an adverse prognosis. Supplementary Data

Supplementary data related to this article can be found at https://doi.org/10.1053/j.jrn.2018.06.005.

References 1. Wizemann V, Wabel P, Chamney P, et al. The mortality risk of overhydration in haemodialysis patients. Nephrol Dial Transplant. 2009;24:15741579. 2. Hur E, Usta M, Toz H, et al. Effect of fluid management guided by bioimpedance spectroscopy on cardiovascular parameters in hemodialysis patients: a randomized controlled trial. Am J Kidney Dis. 2013;61:957-965. 3. Chazot C, Wabel P, Chamney P, Moissl U, Wieskotten S, Wizemann V. Importance of normohydration for the long-term survival of haemodialysis patients. Nephrol Dial Transplant. 2012;27:2404-2410. 4. Dekker M, Marcelli D, Canaud B, et al. Unraveling the relationship between mortality, hyponatremia, inflammation and malnutrition in hemodialysis patients: results from the international MONDO initiative. Eur J Clin Nutr. 2016;70:779-784. 5. Chung SH, Lindholm B, Lee HB. Is malnutrition an independent predictor of mortality in peritoneal dialysis patients? Nephrol Dial Transplant. 2003;18:2134-2140. 6. Marcelli D, Usvyat LA, Kotanko P, et al. Body composition and survival in dialysis patients: results from an international cohort study. Clin J Am Soc Nephrol. 2015;10:1192-1200. 7. Stenvinkel P, Chung SH, Heimb€ urger O, Lindholm B. Malnutrition, inflammation, and atherosclerosis in peritoneal dialysis patients. Perit Dial Int. 2001;21(Suppl 3):S157-S162. 8. Demirci MS, Demirci C, Ozdogan O, et al. Relations between malnutrition-inflammation-atherosclerosis and volume status. The usefulness of bioimpedance analysis in peritoneal dialysis patients. Nephrol Dial Transpl. 2011;26:1708-1716. 9. Guo Q, Yi C, Li J, et al. Of fluid overload in southern Chinese continuous Ambulatory peritoneal dialysis patients. Sen U, ed. PLoS One. 2013;8:e53294.

10

DEKKER ET AL

10. Hung S-C, Lai Y-S, Kuo K-L, Tarng D-C. Volume overloa d and adverse outcomes in chronic Kidney disease: clinical observational and Animal studies. J Am Heart Assoc. 2015;4:1-12. https://doi.org/10.1161/ (ISSN)2047-9980. 11. Stenvinkel P, Gillespie IA, Tunks J, et al. Inflammation Modifies the Paradoxical association between body mass index and mortality in hemodialysis patients. J Am Soc Nephrol. 2016;27:1479-1486. 12. United Nations statistics Division- standard country and Area codes Classifications (M49). Available at: http://unstats.un.org/unsd/methods/ m49/m49regin.htm#europe. Accessed January 23, 2016. 13. Von Gersdorff GD, Usvyat L, Marcelli D, et al. Monitoring dialysis outcomes across the world-The MONDO global database consortium. Blood Purif. 2014;36:165-172. 14. Usvyat LA, Haviv YS, Etter M, et al. The MONitoring dialysis outcomes (MONDO) initiative. Blood Purif. 2013;35:37-48. 15. Wabel P, Moissl U, Chamney P, et al. Towards improved cardiovascular management: the necessity of combining blood pressure and fluid overload. Nephrol Dial Transpl. 2008;23:2965-2971. 16. van Biesen W, Williams JD, Covic AC, et al. Fluid status in peritoneal dialysis patients: the European body composition monitoring (EuroBCM) study cohort. Zoccali C, ed. PLoS One. 2011;6:e17148. 17. Hecking M, Karaboyas A, Saran R, et al. Predialysis serum sodium level, dialysate sodium, and mortality in maintenance hemodialysis patients: the Dialysis Outcomes and Practice Patterns Study (DOPPS). Am J Kidney Dis. 2012;59:238-248.

18. Cheng LT, Tang W, Wang T. Strong association between volume status and nutritional status in peritoneal dialysis patients. Am J Kidney Dis. 2005;45:891-902. 19. Zhou X, Luo YC, Ji WJ, et al. Modulation of mononuclear phagocyte inflammatory response by liposome-encapsulated voltage gated sodium channel inhibitor ameliorates myocardial ischemia/reperfusion injury in rats. PLoS One. 2013;8:e74390. 20. Konings CJAM, Kooman JP, Schonck M, et al. Fluid status in CAPD patients is related to peritoneal transport and residual renal function: Evidence from a longitudinal study. Nephrol Dial Transpl. 2003;18:797-803. 21. John B, Tan BK, Dainty S, Spanel P, Smith D, Davies SJ. Plasma volume, albumin, and fluid status in peritoneal dialysis patients. Clin J Am Soc Nephrol. 2010;5:1463-1470. 22. Yu Z, Tan BK, Dainty S, Mattey DL, Davies SJ. Hypoalbuminaemia, systemic albumin leak and endothelial dysfunction in peritoneal dialysis patients. Nephrol Dial Transpl. 2012;27:4437-4445. 23. Chiao JJ, Jones WG 2nd, Shires GT 3rd, Barber AE, Shires GT. Effect of sepsis on intracellular sodium activity, sodium concentration, and water content in thermal injured rat. Circ Shock. 1992;38:42-49. 24. Faintuch J, Soriano FG, Ladeira JP, Janiszewski M, Velasco IT, GamaRodrigues JJ. Changes in body fluid and energy compartments during prolonged hunger strike. Rev Hosp Clin Fac Med Sao Paulo. 2000;55:47-54. 25. Onofriescu M, Hogas S, Voroneanu L, et al. Bioimpedance-guided fluid management in maintenance hemodialysis: a Pilot randomized controlled trial. Am J Kidney Dis. 2014;64:111-118.