Bioelectrical impedance analysis for assessing hydration status in patients undergoing maintenance hemodialysis

Abstract

BACKGROUND

Adequate hydration status is a concern in patients receiving hemodialysis, as fluid overload increases the risk of complications and mortality. Fluid overload in this population is associated with complications, including arterial hypertension, intradialytic hypotension, pulmonary and peripheral edema, cardiovascular events, reduced body mass index, malnutrition, and increased mortality. Several methodologies have been proposed to evaluate hydration status in hemodialysis patients using indices derived from Bioelectrical impedance analysis parameters.

AIM

To assess the hydration status of patients receiving hemodialysis using different indices.

METHODS

Patients receiving hemodialysis for at least 3 months were included. Hydration status was evaluated before and after the hemodialysis session. The data provided by bioelectrical impedance analysis were total body water (TBW), intracellular water, and extracellular water (ECW). Fluid overload [overhydration (OH)] was calculated. Hydration status was classified according to the three indices: OH, OH/ECW, and ECW/body weight. Interdialytic weight gain (IDWG) and blood pressure (BP) were collected. N-terminal pro-brain natriuretic peptide were measured before and after the hemodialysis.

RESULTS

A total of 54 patients were evaluated, with mean age of 57.6 ± 13.1 years, 59.3% male. Before hemodialysis session, 23 patients were classified as having fluid overload, identified by at least one of the indices. IDWG and ECW showed significant positive correlation with all evaluated indices. BP, TBW, and intracellular water correlated only with the OH and OH/ECW indices. After hemodialysis session, 5 patients were identified with fluid overload by the OH and ECW/body weight indices. ECW was the only parameter that showed correlation with all indices. N-terminal pro-brain natriuretic peptide correlated with the OH and ECW/body weight indices. BP, IDWG, and TBW remained correlated with OH and OH/ECW.

CONCLUSION

The OH index identified a greater number of patients with fluid overload and showed better correlation with hydration parameters provided by the bioelectrical impedance analysis.

Key Words: Hydration; Renal dialysis; Total body water; Volume overload; Bioelectrical impedance

Core Tip: Adequate hydration status is a concern in patients undergoing hemodialysis, as fluid overload increases the risk of complications and mortality. In our study, we evaluated hydration status before and after the hemodialysis session using data provided by bioelectrical impedance. Hydration status was classified according to the three indices: Overhydration (OH), OH/extracellular water, and extracellular water/body weight. The OH index was the most effective in detected more patients with fluid overload before and after the dialysis session, showed better correlation with hydration parameters provided by bioimpedance, blood pressure, and N-terminal pro-brain natriuretic peptide.

INTRODUCTION

Adequate hydration status is a major concern in patients on maintenance hemodialysis, due to fluid volume and distribution alterations related to the reduced kidney function[1]. Fluid overload in this population is associated with complications, including arterial hypertension, intradialytic hypotension, pulmonary and peripheral edema, cardiovascular events, reduced body mass index, malnutrition, and increased mortality[2-5].

Despite advances in technology, accurately estimating fluid overload remains a challenge[2]. Gold standard dilution methods for measuring extracellular volume (sodium bromide), intracellular volume (total body potassium), and total body water (TBW; deuterium or tritium) provide precise results. Nevertheless, they are too complex for routine clinical use[6]. Bioelectrical impedance analysis, a simpler, non-invasive and cost-effective technique[7,8], estimates TBW, intracellular water (ICW), and extracellular water (ECW) based on resistance and reactance measurements[9]. Several methodologies have been proposed to evaluate hydration status in hemodialysis patients using indices derived from Bioelectrical impedance analysis parameters, including those developed by Chen et al[7], Zoccali et al[10], and Dekker et al[11]. Therefore, this study aimed to assess the hydration status of patients receiving hemodialysis using different indices derived from bioimpedance analysis.

MATERIALS AND METHODS

Study design and patients

This was an observational study analysis that enrolled participants from a previously published study[12]. Maintenance hemodialysis patients from the Clinical Hospital of the Federal University of Uberlândia, were included. Patients aged ≥ 18 years and undergoing hemodialysis for at least 3 months were eligible for the study. The exclusion criteria were bedridden patients, wheelchair users, patients with amputated limbs, those with edema or ascites, pacemakers, and metallic implants. These exclusion criteria were adopted because bioimpedance analysis assumes stable body geometry and homogeneous electrical current distribution, which may be substantially altered in individuals with limb amputation, marked edema or ascites, implanted electrical devices, metallic implants, or severe mobility limitations, potentially leading to inaccurate estimation of body water compartments. The study was approved by the Human Research Ethics Committee of Federal University of Uberlândia. All participants provided written informed consent.

The following clinical and demographic data were collected from medical records: Age, gender, dialysis vintage, interdialytic weight gain (IDWG), and systolic and diastolic blood pressure (BP) before and after a hemodialysis session. Results from the patients’ routine monthly laboratory tests, performed in the same month as the hydration assessment, were also collected. Serum concentrations of pre-dialysis creatinine, urea, albumin, and C-reactive protein were considered.

Assessments

Hydration status and body weight were assessed before and after the hemodialysis session. Anthropometric assessment included measuring body weight and height using standardized methodology[13]. Body mass index was calculated by dividing the post-dialysis body weight by height squared. Hydration status was assessed using single-frequency bioelectrical impedance analysis (biodynamics model 450), performed before and after the hemodialysis session, without standardization of the day of the week. The assessments were performed after a 5-minute resting period in the supine position for fluid redistribution. The tests were conducted regardless of whether it was the first, second, or third hemodialysis session of the week. Patients were evaluated in the supine position, with electrodes placed on the hand and foot of the arm without the arteriovenous fistula.

The bioimpedance parameters considered were TBW, ICW, and ECW. To calculate fluid overload [overhydration (OH)], the Watson et al[14] formula was first used to estimate TBW under normal physiological conditions: For men: TBW = 2.447 - 0.09516 × age + 0.1074 × height + 0.3362 × body weight. For women: TBW = -2.097 + 0.1069 × height + 0.2466 × body weight. The TBW value obtained from the Watson et al[14] formula was subtracted from the TBW measured by bioimpedance, according to the following equation: OH = TBW_bioimpedance - TBW_Watson. Fluid overload was classified according to three indices: (1) OH > 1.1 L[11]; (2) OH/ECW × 100 > 15% in men and > 13% in women[10]; and (3) ECW/body weight > 28% in men or > 25% in women[7].

Serum N-terminal pro-brain natriuretic peptide (NT-proBNP) concentrations were determined from blood samples collected before and after the hemodialysis session. Samples were centrifuged at 3000 rpm for 15 minutes at 4 °C. The serum was then separated, stored in eppendorf tubes, and frozen at -80 °C until analysis. NT-proBNP levels were later measured using the enzyme-linked immunosorbent assay method (Elabscience Biotechnology Co., Ltd, Wuhan, Hubei Province, China), following the manufacturer’s instructions.

Statistical analysis

The normality of variable distributions was tested using the Kolmogorov-Smirnov test. Data were expressed as mean ± SD or as median (interquartile range), depending on variable distribution, and as n (%).

Comparisons between before and after dialysis moments were performed using the paired t test or the Wilcoxon signed-rank test. Correlations between hydration assessment parameters and BP, body water compartments (from bioimpedance), and NT-proBNP levels were tested using Pearson or Spearman coefficients. The strength of the correlations was classified as: Very strong (r > 0.8), moderately strong (r = 0.6-0.8), fair (r = 0.3-0.5), and poor (r = 0-0.3)[15]. Statistical significance was set at P < 0.05. Analyses were performed using SPSS 20.

RESULTS

A total of 54 patients were enrolled, with a mean age of 57.6 ± 13.1 years; 32 (59.3%) patients were male. The other characteristics of the sample are presented in Table 1.

Table 1 General characteristics of the patients, n (%)/mean ± SD/median (interquartile rage).

Variable	n = 54
Sex (male)	32 (59.3)
Age (years)	57.6 ± 13.1
Hemodialysis vintage (months)	21.5 (5.0-62.2)
Diabetic	21 (38.9)
Vascular access
Arteriovenous fistula	37 (68.6)
Double-lumen catheter	17 (31.5)
Creatinine (mg/dL)	8.91 ± 3.56
Urea (mg/dL)	111.5 (97.0-122.7)
C-reactive protein (mg/dL)	0.48 (0.17-1.21)
Albumin (g/dL)	3.90 ± 0.55
Body mass index (kg/m2)	26.25 ± 5.75
Interdialytic weight gain (kg)	1.94 ± 0.95
Urea clearance index	1.30 ± 0.26

In the assessment before dialysis session, 9 patients (16.7%) were classified as having fluid overload according to the OH/ECW index, while 14 (25.9%) and 15 (27.8%) patients were classified as having fluid overload by the OH index and the ECW/body weight index, respectively. Some patients were identified by more than one index (Figure 1). Thirty-one patients were classified as normohydrated according to all three indices. In the assessment after dialysis session, four patients (7.4%) were considered to have fluid overload according to OH or ECW/body weight indices. Three patients were classified by both indices. According to OH/ECW index, no patients presented fluid overload (Figure 1).

Figure 1 Fluid overload according to the overhydration, overhydration/extracellular water, and extracellular water/body weight indices before and after the hemodialysis session. OH: Overhydration; ECW: Extracellular water.

Table 2 presents the hydration parameters assessed before and after the dialysis session. As expected, all parameters measured by bioimpedance, as well as NT-proBNP levels, showed a significant reduction after the dialysis session, except for the percentage of ICW, which significantly increased. However, BP parameters were not significantly altered after the dialysis session.

Table 2 Hydration parameters before and after the hemodialysis session (n = 54), median (interquartile rage)/ mean ± SD.

Parameters	Before dialysis	After dialysis	P value
SBP (mmHg)	137 (129-161)	143 (124-158)	1.000
DBP (mmHg)	76 (67-88)	78 (70-98)	0.608
Total body water (L)	36.94 ± 7.02	33.27 ± 6.42	< 0.001
Intracellular water (L)	19.06 ± 4.15	18.12 ± 3.92	< 0.001
Intracellular water (%)	51.50 ± 4.48	53.87 ± 5.41	< 0.001
Extracellular water (L)	17.89 ± 3.62	15.15 ± 3.26	< 0.001
Extracellular water (%)	48.49 ± 4.48	45.56 ± 4.68	< 0.001
NT pro-BNP (ng/mL)	0.677 (0.182-2.422)	0.549 (0.138-2.322)	0.002
Overhydration (L)	-0.36 ± 2.93	-3.43 ± 2.57	< 0.001
Overhydration/extracellular water (%)	-3.95 ± 16.80	-25.34 ± 19.88	< 0.001
Extracellular water/weight (%)	24.81 ± 3.36	21.55 ± 2.99	< 0.001

SBP: Systolic blood pressure; DBP: Diastolic blood pressure; NT pro-BNP: N-terminal pro-brain natriuretic peptide.

In the moment before dialysis session, the OH, OH/ECW, and ECW/body weight were significantly and positively correlated with IDWG and ECW (liters). Additionally, OH and OH/ECW were positively correlated with systolic and diastolic BP, TBW, and ICW (liters). No significant correlations were observed between any of the indices and NT-proBNP (Table 3).

Table 3 Correlations of hydration assessment indices with blood pressure and body water parameters before dialysis session (n = 54).

Pre-dialysis parameter		OH	OH/ECW	ECW/body weight
SBP (mmHg)	r	0.3741	0.3521	0.217
	P value	0.0051	0.0091	0.115
DBP (mmHg)	r	0.3161	0.2973	0.132
	P value	0.0201	0.0293	0.342
IDWG (kg)	r	0.4821	0.5052	0.3541
	P value	< 0.0011	< 0.0012	0.0091
TBW (L)	r	0.5172	0.5532	0.089
	P value	< 0.0012	< 0.0012	0.524
ICW (L)	r	0.3951	0.4321	-0.104
	P value	0.0031	< 0.0011	0.453
ICW (%)	r	-0.115	-0.106	-0.429
	P value	0.409	0.446	0.001
ECW (L)	r	0.5402	0.5642	0.2893
	P value	< 0.0012	< 0.0012	0.0343
ECW (%)	r	0.113	0.104	0.4291
	P value	0.414	0.452	0.0011
NT pro BNP (ng/mL)	r	0.181	0.180	0.257
	P value	0.228	0.232	0.084

¹Poor correlation.

²Fair correlation.

³Negligible correlation.

SBP: Systolic blood pressure; DBP: Diastolic blood pressure; IDWG: Interdialytic weight gain; TBW: Total body water; ICW: Intracellular water; ECW: Extracellular water; NT pro-BNP: N-terminal pro-brain natriuretic peptide; OH: Overhydration.

In the moment after dialysis session, both the OH and OH/ECW indices were positively correlated with systolic and diastolic BP, IDWG, TBW, and ECW. Furthermore, the OH index was positively correlated with NT-proBNP, and OH/ECW was correlated with ICW. The ECW/body weight index was positively correlated with ICW, ECW, and NT-proBNP (Table 4).

Table 4 Correlations of hydration assessment indices with blood pressure and body water parameters after dialysis session (n = 54).

Post-dialysis parameter		OH	OH/ECW	ECW/body weight
SBP (mmHg)	r	0.5202	0.4951	0.252
	P value	< 0.0012	< 0.0011	0.066
DBP (mmHg)	r	0.3931	0.3431	0.055
	P value	0.0031	0.0111	0.695
IDWG (kg)	r	0.3391	0.3671	0.184
	P value	0.0121	0.0061	0.183
TBW (L)	r	0.3641	0.5482	0.062
	P value	0.0071	< 0.0012	0.657
ICW (L)	r	0.235	0.3721	-0.172
	P value	0.087	0.0061	0.214
ICW (%)	r	-0.122	-0.162	-0.3681
	P value	0.378	0.241	0.0061
ECW (L)	r	0.4351	0.6322	0.3291
	P value	0.0011	< 0.0012	0.0151
ECW (%)	r	0.216	0.2903	0.5592
	P value	0.117	0.0343	< 0.0012
NT pro BNP (ng/mL)	r	0.3101	0.269	0.4271
	P value	0.0361	0.070	0.0031

¹Poor correlation.

²Fair correlation.

³Negligible correlation.

SBP: Systolic blood pressure; DBP: Diastolic blood pressure; IDWG: Interdialytic weight gain; TBW: Total body water; ICW: Intracellular water; ECW: Extracellular water; NT pro-BNP: N-terminal pro-brain natriuretic peptide; OH: Overhydration.

DISCUSSION

In this study, we evaluated the agreement between three bioimpedance-derived indices[7,10,11] for classifying fluid overload in patients receiving hemodialysis. Some discrepancies were found among the classifications produced by the indices. In the analysis prior to the hemodialysis session, all indices identified individuals with fluid overload. In the moment after the dialysis session, only the OH[11] and ECW/body weight[7] indices identified patients with volume overload. Although fewer cases of fluid overload were expected after dialysis session, identifying these patients remains crucial for dry weight adjustment.

It is important to recognize that the three bioimpedance-derived indices used in this study do not represent identical physiological constructs. OH reflects absolute fluid excess, whereas OH/ECW expresses fluid overload relative to the extracellular compartment, and ECW/body weight reflects the proportional expansion of ECW relative to body mass. Therefore, discrepancies among indices may reflect conceptual and physiological differences rather than simple measurement disagreement. Nevertheless, we also analyzed the correlations between these indices and clinical and laboratorial variables. The OH/ECW[10] and OH[11] indices showed stronger correlations with parameters related to fluid status. The OH index[11] was the most effective for detecting hyperhydration in hemodialysis patients, as it not only correlated with hydration parameters but also identified a greater number of individuals with fluid overload.

The finding that, after the dialysis session, fluid overload was detected only by the OH and ECW/body weight indices, but not by OH/ECW, suggests that these indices capture different aspects of post-dialysis volume status. OH reflects absolute residual fluid excess, which may persist despite ultrafiltration, particularly in patients with chronically expanded extracellular volume. In contrast, OH/ECW depends on proportional ECW expansion and may normalize more rapidly after dialysis, even if a small absolute excess volume remains. Alternatively, ECW/body weight may be more sensitive to subtle proportional changes in extracellular volume when TBW is reduced after dialysis, highlighting residual extracellular expansion in a lower-volume state.

Bioimpedance is increasingly used as a non-invasive, practical tool for assessing fluid status in patients receiving hemodialysis[11]. Onofriescu et al[16] showed that bioimpedance-guided volume control in patients on hemodialysis led to a significantly lower fluid overload, systolic BP, and pulse wave velocity compared to volume control based on clinical parameters only.

Fluid overload is a strong predictor of complications and mortality. The indices used in our study have previously been associated with worse outcomes in patients receiving hemodialysis. Dekker et al[11], in a cohort study of 8833 patients receiving hemodialysis, found that pre-dialysis fluid overload was associated with an increased risk of mortality. Similarly, Zoccali et al[10], in a cohort study using bioimpedance in 39566 patients receiving hemodialysis, demonstrated that fluid overload was a strong predictor of mortality in these individuals. Chen et al[7] assessed ECW and ICW in 121 patients receiving hemodialysis using bioimpedance. Subsequently, dry weight was adjusted according to the assessment results, and bioimpedance analysis was repeated after 4 months. After this adjustment, improvements in BP were observed. Despite current guidelines recommending multifrequency bioimpedance[17], the single-frequency device effectively detected fluid overload using the proposed indices. As expected, few patients were overloaded after dialysis; however, those identified still require clinical attention, highlighting bioimpedance role in optimizing dry weight.

Relying solely on clinical parameters to identify fluid overload in patients receiving hemodialysis may be insufficient. Our findings showed no significant changes in BP after the dialysis session, meaning some overloaded patients would not have been detected without bioimpedance. Early detection and correction of fluid overload in CKD patients is crucial, as this condition is associated with worse clinical outcomes. Wang and Gu[18], in a meta-analysis of eight studies, identified that an OH/ECW ratio greater than 15%, the same parameter used in our study, is an independent risk factor for mortality. Fluid overload contributes to cardiovascular damage and events[19,20], and higher all-cause hospitalization rates[16].

In addition to BP and body water parameters, our study also included NT-proBNP measurements from samples collected at the same time as bioimpedance assessments. NT-proBNP levels were correlated with fluid overload only in the moment after hemodialysis, as measured by the ECW/body weight[7] and OH[11] indices. NT-proBNP levels significantly declined post-dialysis, consistent with reduced body water. This biomarker is released by cardiac muscle cells, especially in response to significant extracellular volume expansion and is validated in patients receiving hemodialysis[21].

This study has some limitations. The sample size was relatively small, and this study was not originally designed specifically to assess hydration status. Some factors that could influence fluid overload were not evaluated, such as dialysate sodium concentration, clinical signs and symptoms (e.g., edema, cramps), and the absence of standardization for the day of the week when assessments were performed. Evaluations were conducted during a single dialysis session without control for weekly variations. Additionally, NT-proBNP levels, although useful as a volume marker, can be influenced by cardiac comorbidities[22]. Finally, although multifrequency bioimpedance and bioimpedance spectroscopy are currently considered the reference method for assessing fluid overload in dialysis patients[17] and were used in the validation of some of the analyzed indices[10,11], a single-frequency device was used in this study due to its wider availability in routine clinical practice. Single-frequency body impedance analysis may be less precise in differentiating ECW and ICW, particularly in states of marked fluid imbalance, which should be considered when interpreting our findings. Nonetheless, our results align with other studies using similar devices[4,23]. A key strength was the inclusion of NT-proBNP as a complementary volume marker.

CONCLUSION

In conclusion, the OH index, proposed by Dekker et al[11], identified more patients with fluid overload detected before and after dialysis session, and showed better correlation with hydration parameters provided by bioimpedance, BP, and NT-proBNP. Although the OH index detected a greater number of patients with fluid overload and showed stronger correlations with clinical and biochemical markers, our findings indicate that the three indices provide complementary information. Their combined interpretation may offer a more physiologically comprehensive assessment of hydration status in hemodialysis patients.