Abstract
Background:
Fluid overload (FO) is frequent in critically ill patients and is linked to adverse outcomes. Conventional assessment methods – clinical signs, body weight variation, and cumulative fluid balance (cumFB) – have limitations during prolonged intensive care unit (ICU) stays. Bioelectrical impedance analysis (BIA) is a non-invasive technique that may provide a more accurate evaluation of hydration status. This study assessed the ability of BIA to detect FO in ICU patients compared with standard clinical and ultrasound methods.
Methods:
We performed a prospective observational study in a medical ICU, including adults expected to stay ≥7 days. Hydration status was monitored daily by clinical exam, ultrasound, weight change, and cumFB. From Day 3, BIA measurements [total body water (TBW), extracellular water (ECW), intracellular water (ICW), third-space volume, ECW/TBW ratio, ECW/ICW ratio, excess volume, and phase angle] were obtained. FO was defined by BIA-derived criteria. Agreements and correlations were analyzed.
Results:
Twenty-one patients (median age 69.5 years; 66.7% male) underwent 88 BIA assessments. Clinical FO was present in 55.7% of cases, whereas BIA identified FO in 82.8% (excess volume >1 L). In patients without clinical signs, BIA still revealed hidden fluid excess. Weight change correlated with cumFB (ρ = 0.8), but conventional indicators showed weak or no correlation with BIA parameters. Patients with significant weight loss (<–5%) often remained overhydrated according to BIA.
Conclusion:
BIA detects subclinical FO overlooked by standard methods, supporting its integration into ICU fluid management and de-resuscitation strategies. Larger studies are needed to confirm its prognostic value.
Keywords: bioimpedance analysis, fluid overload, intensive care
Introduction
Most critically ill patients develop a positive fluid balance (FB) within the first few days of intensive care unit (ICU) admission, leading to fluid overload (FO). This condition is typically defined as a weight gain of more than 5% from baseline. FO is associated with increased mortality[1] and organ failure, particularly acute respiratory failure (acute pulmonary edema), and acute renal injury (renal congestion)[2–4]. It can also contribute to multiple organ failures[5]. Minimizing both the severity and duration of FO is crucial for improving the prognosis of ICU patients. Therefore, it is essential, especially after initial stabilization, to assess FO early. Evaluating hydration status is key to guiding fluid removal (de-resuscitation)[6] while preventing hypovolemia[7,8].
HIGHLIGHTS
Bioelectrical impedance analysis (BIA) identified subclinical fluid overload (FO) in the majority of critically ill patients.
Clinical assessment showed poor concordance with BIA-derived hydration indices.
BIA frequently indicated persistent FO, despite the normalization of clinical signs.
BIA may provide complementary guidance for intensive care unit fluid management and de-resuscitation strategies.
Several methods are available to assess hydration status in ICU patients. Clinical signs, such as lower limb edema, effusions (ascites or pleural), and weight fluctuations, are commonly used but lack sensitivity and may become unreliable over time[9,10]. Weight monitoring remains the most reliable method for longer ICU stays. However, ICU patients often experience catabolism and muscle wasting, losing up to 2% of lean body mass per day or 1.5–3 kg per week[11]. These changes, along with inflammation-induced shifts in water distribution and altered vascular permeability, can make weight-based assessments less dependable over extended stays.
Alternative methods, such as thermodilution[12], can assess extravascular pulmonary water but require invasive procedures. Imaging techniques, like ultrasound and chest X-rays, can detect pleural effusions but only when FO is significant.
Bioelectrical impedance analysis (BIA) presents a promising, non-invasive alternative for assessing FO[7]. BIA, which is easily performed at the ICU bedside, evaluates body composition and fluid status using four electrodes to send a low-intensity electrical current through the body. By measuring impedance, reactance, resistance, and phase angle, BIA estimates total body water (TBW), extracellular water (ECW), and intracellular water (ICW) volumes. Studies have shown its effectiveness in assessing hydration status in critically ill patients[13,14] and guiding fluid management[15,16] (Supplemental Digital Content Table S1, available at: http://links.lww.com/MS9/B196). This technology can assist in determining dry weight for de-resuscitation[17,18]. However, BIA’s accuracy may be compromised in patients with severe FO or systemic inflammation[7,19,20].
However, most available studies have focused either on early resuscitation or on single-time-point assessments, with limited data on dynamic changes in hydration status during the recovery and deresuscitation phases of ICU stay. In addition, discrepancies between BIA-derived FO and conventional clinical assessment remain insufficiently explored, particularly in patients who have returned to or fallen below their baseline body weight[8,19,21,22].
Given these limitations, BIA’s reliability in detecting residual FO – particularly at the end of the de-resuscitation phase, when patients return to their baseline weight – remains uncertain.
The primary objective of this study was to assess the agreement between bedside clinical evaluation of FO and BIA-derived FO across repeated measurements.
Materials and methods
Type of study
This single-center prospective study was carried out in the medical ICU of Clermont-Ferrand in France between July 2024 and September 2024.
Study population
Adult patients (≥18 years) admitted to the ICU with an expected length of stay (LOS) greater than 7 days were eligible for inclusion. Patients with an ICU stay exceeding 5 days prior to inclusion were excluded to ensure longitudinal assessment of hydration status throughout the ICU course. Additional exclusion criteria included refusal to participate, legal guardianship, limb amputation, pregnancy or breastfeeding, and the presence of implantable electronic devices (pacemaker or defibrillator). BIA measurements were initiated on Day 3 to reduce interference from the early resuscitation phase, which is characterized by rapid fluid shifts and hemodynamic instability.
Collection of the data
Upon admission, the patient’s age, weight, height, main comorbidities, reason for admission, and severity scores (SAPS II – Simplified Acute Physiology Score and SOFA – Sepsis-related Organ Failure Assessment) were recorded. Clinical data, including 24-h diuresis, daily weight, daily and cumulative fluid balance (cumFB), and SOFA scores, were collected from the day of admission (D1) to D14, or until discharge from the ICU if it occurred before D14. The change in weight (delta weight) compared to the baseline weight was calculated in both kilograms and percentages. Biological data were obtained from blood and urine ionograms and hemograms (sodium, potassium, blood glucose, blood and urine urea, creatinine, proteinuria, osmolarity, and hematocrit).
FO, assessed by clinical examination, was defined by the presence of at least one of the following signs: ascites, peripheral edema, pulmonary crackles, jugular venous distension, hepatojugular reflux, or pleural effusion.
Ultrasound assessments for edema, pleural effusions, and ascites were conducted. Acute pulmonary edema was identified with B-profiles in the four quadrants.
BIA measurements (impedance, reactance, resistance, and phase angle), along with derived parameters [TBW, ECW, ICW, third space, ECW/ICW ratio, ECW/TBW ratio, phase angle, fat-free mass (FFM)], were collected starting from D3, as close as possible to the collection of biological data.
BIA and ultrasound measurements were performed simultaneously under standardized conditions in the morning (around 10:00 a.m.), before patient mobilization and routine care, and after a stabilization period of approximately 2 hours without fluid administration or changes in caloric intake, in order to minimize variability related to posture, physical activity, and recent fluid infusion.
Vital status was recorded upon discharge from the ICU, from the hospital, and/or on Day 28.
Definitions of FO by BIA
Several complementary BIA-derived indices were analyzed to characterize hydration status and fluid distribution. A concise summary of the indices and cutoff values used in the study is provided in Table 1. Detailed physiological background is provided in Supplemental Digital Content, available at: http://links.lww.com/MS9/B196.
Table 1.
Bioimpedance-derived indices of fluid overload.
| Index | Definition | Cutoff/reference value | Interpretation |
|---|---|---|---|
| ECW/TBW ratio | Ratio of extracellular water to total body water | >0.40 | Increased extracellular water fraction |
| CLI | Ratio of ECW to ICW | >0.877 | Altered fluid distribution/capillary leak |
| EV, BIS-OH | Absolute excess extracellular water compared with expected ECW | >1 L | Bioimpedance-defined fluid overload |
| Third space | Device-derived estimate of fluid in non-functional compartments | Device-specific | Additional marker of fluid accumulation |
| TBW/FFM ratio | Total body water relative to fat-free mass | >0.73 | Relative fluid overload |
| Phase angle | Derived from resistance and reactance | Low values (context-dependent) | Impaired cellular integrity/altered hydration |
BIA, bioimpedance analysis; BIS-OH, bioimpedance spectroscopy–derived overhydration; CLI, capillary leak index; ECW, extracellular water; EV, excess volume; FFM, fat-free mass; ICW, intracellular water; TBW, total body water.
For the primary analyses, BIA-defined FO was identified as excess volume [BIS-OH (bioimpedance spectroscopy–overhydration)] >1 L, as this parameter provides a quantitative estimate of extracellular fluid excess. The remaining BIA-derived indices were analyzed as secondary or exploratory markers to characterize fluid distribution, cellular health, and body composition.
Endpoints
The primary endpoint was the agreement between clinical FO (yes/no) and BIA-defined FO (yes/no), defined as excess volume (BIS-OH) >1 L.
Secondary endpoints included:
Agreement between clinical FO (yes/no) and ultrasound signs of congestion (yes/no) is defined by the presence of B-lines and/or pleural effusion.
Agreement between BIA-defined FO (BIS-OH >1 L) and ultrasound signs of congestion.
- Comparison of BIA-derived hydration indices between measurements with and without clinical FO, including:
- ECW/TBW ratio
- Capillary leak index (CLI, ECW/ICW)
- Excess volume (BIS-OH), analyzed as a continuous variable
- Third space (device-derived output)
- TBW/FFM ratio
- Phase angle
Comparison of BIA-derived hydration indices across weight-change categories (−5%, −5 to 0%, 0–5%, >5%).
Correlation between BIA-derived hydration indices and clinically derived hydration indices, including delta weights and cumFB.
Sample size considerations
No formal a priori sample-size calculation was performed. This study was conceived as an exploratory feasibility study, based on the number of patients and repeated assessments achievable during the study period. Given the presence of repeated measurements within patients, with expected strong within-patient correlation, and the proportion of missing bioimpedance data, the study was not powered to demonstrate a predefined level of agreement.
Statistical analysis
Patient characteristics were presented as numbers (percentages) for categorical data and medians [interquartile range (IQR)] for continuous data.
All analyses were exploratory. Effect sizes (e.g., agreement estimates, differences between groups) and their confidence intervals were prioritized for interpretation, while P-values were considered descriptive and not used for formal hypothesis testing.
Agreement was evaluated using Cohen’s kappa coefficient and percent agreement based on complete measurement pairs.
The correlations between the various continuous variables were assessed using Spearman’s correlation coefficient. Comparisons according to FO, assessed by clinical examination and delta weights (−5%, −5 to 0%, 0–5%, >5%), were achieved.
Comparisons were achieved with Chi², Wilcoxon, or ANOVA as appropriate. For each test, a two-sided 5% alpha risk was considered. There was no imputation. Statistical analyses were performed using SAS software, Version 9.4 (SAS Institute, Cary, NC), and R (Version 3.6.3).
A systematic literature search was conducted in PubMed using a Boolean search strategy based on the terms “hydration status,” “monitoring,” “bioimpedance,” and “intensive care,” as detailed in the Supplemental Digital Content, available at: http://links.lww.com/MS9/B196.
Results
General characteristics of the study population
This study included 21 patients with a median age of 69.5 years, of whom 66.7% were male with a median body mass index of 27.3 kg/m2 (Table 2). The most common reasons for admission were sepsis (38.1%), cardiac arrest (33.3%), and acute respiratory failure (28.6%). Nearly half (47.6%) had cardiovascular comorbidities, while 28.6% had renal disease or immunosuppression, and 23.8% had respiratory conditions. Severity scores indicated a median SAPS II of 54 and a SOFA of 9. The median length of ICU stay was 10 days, and ICU mortality was 38.1%.
Table 2.
General characteristics of the study population.
| Parameters | N (%)/Med [IQR] |
|---|---|
| Number of patients | 21 |
| Reason for admission | |
| Cardiac arrest | 7 (33.3) |
| Sepsis | 8 (38.1) |
| Acute respiratory failure | 6 (28.6) |
| Age | 69.5 [52.4; 73.8] |
| Sex (M) | 14 (66.7) |
| BMI (kg/m2) | 27.3 [23.1; 29] |
| Comorbidities | |
| Cardiovascular | 10 (47.6) |
| Respiratory | 5 (23.8) |
| Coagulation | 0 |
| Liver | 1 (4.8) |
| Renal | 6 (28.6) |
| Immunosuppression | 6 (28.6) |
| Diabetes | 4 (19) |
| Severity admission | |
| SAPS II | 54 [42; 64] |
| SOFA | 9 [4; 12] |
| SOFA respiratory >2 | 9 (42.9) |
| SOFA coagulation >2 | 1 (4.8) |
| SOFA liver >2 | 0 |
| SOFA cardiovascular >2 | 11 (52.4) |
| SOFA neurology >2 | 6 (28.6) |
| SOFA kidney >2 | 7 (33.3) |
| Length of stay in intensive care | 10 [7; 11] |
| Deaths in intensive care | 8 (38.1) |
ANOVA: analysis of variance; BMI, body mass index; ICU, intensive care unit; SAPS II, Simplified Acute Physiology Score II; SOFA, Sequential Organ Failure Assessment.
Characteristics of hydration assessments during ICU stay
A total of 88 measurements were finally achieved (Table 3). The median SOFA score on the day of measurement was 4.5. Clinical FO was observed in 55.7% of cases. Weight fluctuations were notable, with a median delta weight of 0.2 kg, and 27.3% of patients experienced a weight change exceeding 5%. The median cumFB was 0.9 L, with 48.9% exceeding 1 L. TBW was 57% on average, while ECW and ICW were 28.4% and 30.2%, respectively. Third-space volume accumulation was observed in 60.6% of cases, with a median of 1.7 L. FO, defined by an ECW/TBW >0.4, was prevalent in 97.3% of cases, and 83.6% of the measurements of EV were over 1 L (BIS-OH). The TBW/FFM ratio was 1.2, and the median phase angle was 4.5.
Table 3.
Severity and hydration status characteristics at the time of assessment in the overall cohort.
| Parameters | N (%)/Med [IQR] |
|---|---|
| Number of measurements | 88 (100) |
| SOFA | 4.5 [1; 8] |
| SOFA respiratory >2 | 14 (15.9) |
| SOFA coagulation >2 | 2 (2.3) |
| SOFA liver >2 | 0 |
| SOFA cardiovascular >2 | 24 (27.3) |
| SOFA neurology >2 | 13 (14.8) |
| SOFA kidney >2 | 30 (34.1) |
| Hematocrit (miss = 3) | 29 [25; 36] |
| Osmolality (miss = 7) | 303.7 [293.3; 311.3] |
| Ascites | 7 (8) |
| Edema | 45 (51.1) |
| Crackling | 18 (20.5) |
| Jugular turgor | 3 (3.4) |
| Hepatojugular reflux (miss = 1) | 1 (1.1) |
| Pleural effusion (miss = 2) | 7 (8.1) |
| Pleural B (echo) (miss = 2) | 33 (38.4) |
| Clinical overload | 49 (55.7) |
| Delta weight (kg) | 0.2 [−4.9; 4.2] |
| Delta weight (%) | 0.3 [−4.9; 5.8] |
| Delta weight >10% | 12 (13.6) |
| Delta weight >5% | 24 (27.3) |
| Cumulative fluid balance (L) | 0.9 [−1.8; 4.4] |
| Cumulative fluid balance >1 L | 43 (48.9) |
| Total body water (%) | 57 [52.6; 66.6] |
| Extracellular water (%) (miss = 15) | 28.4 [23.7; 36.7] |
| Intracellular water (%) (miss = 17) | 30.2 [25.7; 32.8] |
| 3rd space (L) (miss = 17) | 1.7 [0.6; 3.4] |
| 3rd space >1 L (miss = 17) | 43 (60.6) |
| ECW/ICW ratio (miss = 20) = CLI | 0.9 [0.8; 1.1] |
| ECW/TBW ratio (miss = 15) | 0.4 [0.4; 0.6] |
| Overload (ECW/TBW >0.4) (miss = 15) | 71 (97.3) |
| Excess volume (miss = 15) | 2.5 [1.5; 9.4] |
| Excess volume >1 L (miss = 15) = BIS OH | 61 (83.6) |
| Total body weight/lean body mass = TBW/FFM | 1.2 [0.9; 1.9] |
| Phase angle | 4.5 [3.8; 5.7] |
ECW, extracellular water; ICW, intracellular water; TBW, total body water; CLI, capillary leak index; BIS OH, bioimpedance spectroscopy–overhydration; FFM, fat-free mass; miss, missing.
During the ICU stay, improvement in body weight and cumFB was frequently accompanied by a reduction in clinically assessed FO, whereas bioimpedance-defined overhydration often persisted (Fig. 1 and Supplemental Digital Content Figure S1, available at: http://links.lww.com/MS9/B196).
Figure 1.
Evolution of clinical and bioimpedance-defined fluid overload during the first week. Daily assessments were classified into four mutually exclusive categories: concurrent clinical and BIA-defined fluid overload, clinical fluid overload only, BIA-defined fluid overload only, and absence of both. Bars represent proportions per day (J0–J7). Missing data are displayed in light grey. The number of analyzable pairs per day (n) is indicated above each bar. FO, fluid overload; BIA, bioimpedance analysis
Agreement between BIA, clinical assessment, and ultrasound signs of FO
Agreement between BIA, clinical assessment, and ultrasound signs of congestion was limited (Fig. 2). Agreement between BIA-defined FO and clinical FO was poor (Po = 0.575; N = 73), with no agreement beyond chance (κ = 0.015). Similarly, agreement between BIA-defined FO and ultrasound congestion was low (Po = 0.437; N = 71; κ = −0.077). In contrast, agreement between clinical FO and ultrasound congestion was higher (Po = 0.628; N = 86), with fair agreement beyond chance (κ = 0.265).
Figure 2.
Agreement between bioimpedance analysis (BIA), clinical assessment, and ultrasound signs of fluid overload. (A) Agreement between BIA-defined fluid overload (BIS-OH >1 L) and clinical assessment. (B) Agreement between BIA-defined fluid overload and ultrasound congestion, defined by the presence of B-lines and/or pleural effusion. (C) Agreement between clinical fluid overload and ultrasound congestion. Numbers represent counts for each category. Cohen’s κ and percent agreement (Po) are reported for each comparison. FO, fluid overload; US, ultrasound.
Comparison according to FO assessed by clinical examination
Patients with clinical FO had significantly higher SOFA scores (7 vs. 3, P < 0.01) and lower hematocrit levels (29 vs. 31%, P = 0.01). Almost all the indicators obtained by BIA were higher in the subgroup of patients with clinical FO, except for ICW, FO (ECW/TBW >0.4), BIS-OH, and TBW/FFM. Other parameters, such as cumFB, excess volume, and weight changes, did not show statistically significant differences between groups. Of note, almost all the patients without clinical FO presented “BIA FO” as defined by BIS-OH (82.8%) or ECW/TBW >0.4 (93.1%) (Table 4).
Table 4.
Comparison of severity and hydration status characteristics according to clinical overload.
| Parameters – N (%)/Med [IQR] | No clinical overload | Clinical overload | P-value |
|---|---|---|---|
| Number of measurements | 39 | 49 | – |
| SOFA respiratory >2 | 3 (7.7) | 11 (22.4) | 0.06 |
| SOFA coagulation >2 | 1 (2.6) | 1 (2) | 0.87 |
| SOFA liver >2 | . | ||
| SOFA cardiovascular >2 | 2 (5.1) | 22 (44.9) | <0.01 |
| SOFA neurology >2 | 4 (10.3) | 9 (18.4) | 0.29 |
| SOFA kidney >2 | 14 (35.9) | 16 (32.7) | 0.75 |
| Pleural B-lines (ultrasound) (miss = 2) | 11 (29.7%) | 22 (44.9%) | 0.15 |
| Δ Weight (kg) | –0.2 [–7.5; 4.7] | 0.7 [–3.8; 3.6] | 0.17 |
| Δ Weight (%) | –0.2 [–7.5; 6.6] | 0.8 [–4.3; 5.1] | 0.14 |
| Cumulative fluid balance (L) | 0 [–7; 4.6] | 2 [–0.3; 4.4] | 0.11 |
| >1 L cumulative fluid balance | 16 (41%) | 27 (55.1%) | 0.19 |
| TBW (%) | 54.8 [52.6; 63.2] | 61.1 [52; 72.4] | 0.05 |
| ECW (%) (miss = 15) | 23.5 [22.3; 29] | 31.7 [26.5; 38.1] | <0.01 |
| ICW (%) (miss = 17) | 30.1 [27.5; 31.5] | 30.2 [25.1; 33.5] | 0.31 |
| 3rd space (L) (miss = 17) | 0.8 [0; 1.9] | 2.7 [1.1; 6.1] | <0.01 |
| 3rd space >1 L (miss = 17) | 13 (41.9%) | 30 (75%) | <0.01 |
| ECW/ICW ratio = CLI (miss = 20) | 0.8 [0.7; 0.9] | 0.9 [0.8; 1.4] | <0.01 |
| ECW/TBW ratio (miss = 15) | 0.4 [0.4; 0.5] | 0.5 [0.4; 0.6] | 0.06 |
| Overload (ECW/TBW >0.4) (miss = 15) | 27 (93.1%) | 44 (100%) | 0.08 |
| Excess volume (L) (miss = 15) | 1.6 [1.3; 4.4] | 4.1 [1.7; 13.1] | 0.02 |
| Excess volume >1 L (BIS-OH) (miss = 15) | 24 (82.8%) | 37 (84.1%) | 0.88 |
| TBW/FFM | 1.3 [0.9; 2] | 1.1 [0.9; 1.4] | 0.22 |
| Phase angle | 5.1 [4.2; 6.5] | 4.2 [3.5; 5] | <0.01 |
Δ, change; BIA, bioelectrical impedance analysis; TBW, total body water; ECW, extracellular water; ICW, intracellular water; BIS OH, bioimpedance spectroscopy–overhydration; FFM, fat-free mass; miss, missing.
Distribution of the variables over time
The distributions of the main variables are depicted in Supplemental Digital Content Figure S2, available at: http://links.lww.com/MS9/B196.
During the first week in the ICU, we could first see that the SOFA score decreased. Delta weight, cumFB, and third-space volume tended to decrease at the end of this first week. It was not evident that a trend could be observed for the other indices.
Comparisons according to delta weight
This analysis categorizes 88 measurements into four groups based on percentage changes in body weight (Table 5). The median SOFA score increased with greater weight gain, rising from 4 in the most weight-loss group to 8 in the highest weight-gain group (P = 0.016). Hematocrit decreased significantly with increasing weight gain (P = 0.027), while pleural effusion was more frequent in the highest weight-gain group (P = 0.028). CumFB varied significantly across groups, with the lowest weight-loss group having a median of −10.1 L and the highest weight-gain group reaching 5.9 L (P < 0.001). Similarly, those with a cumFB >1 L increased from 14.3% in the highest weight-loss group to 87.5% in the highest weight-gain group (P < 0.001). The TBW and CLI remained relatively stable across groups. Third space was significantly more common in the moderate weight-gain group (P = 0.036). TBW/FFM significantly increased with greater weight gain (P < 0.001), while the phase angle remained unchanged.
Table 5.
Variables according to quartile of delta weight.
| Characteristics – N(%)/Med [IQR] | Delta <−5% | Delta −5 to 0% | Delta 0–5% | Delta >5% | P-value |
|---|---|---|---|---|---|
| Number of measurements | 21 (100) | 21 (100) | 22 (100) | 24 (100) | . |
| SOFA | 4 [3; 7] | 1 [1; 5] | 5 [1; 8] | 8 [3; 10] | 0.016 |
| SOFA respiratory >2 | 2 (9.5) | 2 (9.5) | 5 (22.7) | 5 (20.8) | 0.479 |
| SOFA coagulation >2 | 1 (4.8) | 1 (4.8) | 0 | 0 | 0.524 |
| SOFA liver >2 | 0 | 0 | 0 | 0 | . |
| SOFA cardiovascular >2 | 4 (19) | 0 | 8 (36.4) | 12 (50) | 0.001 |
| SOFA neurology >2 | 3 (14.3) | 2 (9.5) | 2 (9.1) | 6 (25) | 0.388 |
| SOFA kidney >2 | 6 (28.6) | 4 (19) | 10 (45.5) | 10 (41.7) | 0.233 |
| Hematocrit (miss = 3) | 33 [27.2; 36] | 30.1 [28.5; 38.7] | 28 [24; 37] | 27.7 [24.5; 29.8] | 0.027 |
| Osmolality (miss = 7) | 299.3 [295; 324.2] | 305.9 [291.7; 316.5] | 297.6 [293.1; 305.1] | 306.4 [301; 310.3] | 0.327 |
| Ascites | 0 | 2 (9.5) | 4 (18.2) | 1 (4.2) | 0.139 |
| Edema | 9 (42.9) | 12 (57.1) | 12 (54.5) | 12 (50) | 0.803 |
| Crackling | 2 (9.5) | 7 (33.3) | 4 (18.2) | 5 (20.8) | 0.289 |
| Jugular turgor | 1 (4.8) | 0 | 1 (4.5) | 1 (4.2) | 0.805 |
| Hepatojugular reflux (miss = 1) | 1 (4.8) | 0 | 0 | 0 | 0.365 |
| Pleural effusion (miss = 2) | 2 (10.5) | 0 | 0 | 5 (20.8) | 0.028 |
| Pleural B (echo) (miss = 2) | 9 (47.4) | 7 (33.3) | 8 (36.4) | 9 (37.5) | 0.820 |
| Clinical overload | 10 (47.6) | 12 (57.1) | 13 (59.1) | 14 (58.3) | 0.863 |
| Delta weight (kg) | −11.5 [−14; −9.5] | −2.7 [−4; −2] | 1.2 [0.7; 2.5] | 7.1 [5.9; 8.9] | <0.001 |
| Delta weight (%) | −14.2 [−17.8; −10.7] | −3.2 [−4.3; −2.4] | 1.6 [0.8; 2.9] | 10.3 [8.7; 14.2] | <0.001 |
| Cumulative fluid balance (L) | −10.1 [−16; 0] | 0 [−1.5; 1.6] | 2.3 [0.3; 4.4] | 5.9 [3.1; 6.5] | <0.001 |
| Cumulative fluid balance >1 L | 3 (14.3) | 6 (28.6) | 13 (59.1) | 21 (87.5) | <0.001 |
| Total body water (%) | 55.6 [52.7; 69.8] | 54.8 [52.5; 58.3] | 61.6 [53.8; 66.7] | 55.6 [50.8; 67.3] | 0.363 |
| Extracellular water (%) (miss = 15) | 25 [23.5; 29] | 29.3 [22.1; 38.5] | 28.9 [26.4; 39.6] | 31.1 [25.6; 36.2] | 0.163 |
| Intracellular water (%) (miss = 17) | 31.4 [26.3; 32.6] | 29.4 [17.6; 30.2] | 31.2 [22.3; 33.1] | 30.2 [26.6; 33] | 0.266 |
| 3rd space (L) (miss = 17) | 1.6 [0.7; 6.3] | 0.8 [0.2; 1.8] | 2.7 [1.5; 4] | 2.2 [0.8; 3.4] | 0.129 |
| 3rd space >1 L (miss = 17) | 9 (52.9) | 7 (38.9) | 16 (84.2) | 11 (64.7) | 0.036 |
| ECW/ICW ratio (miss = 20) = CLI | 0.8 [0.7; 1] | 0.8 [0.7; 2.1] | 0.9 [0.8; 1.4] | 1 [0.8; 1] | 0.280 |
| ECW/TBW ratio (miss = 15) | 0.4 [0.4; 0.5] | 0.4 [0.4; 0.7] | 0.4 [0.4; 0.7] | 0.5 [0.4; 0.6] | 0.281 |
| Overload (ECW/TBW > 0.4) (miss = 15) | 18 (100) | 18 (90) | 19 (100) | 16 (100) | 0.142 |
| Excess volume (miss = 15) | 1.8 [1.3; 4.1] | 2.1 [1.3; 15.2] | 2.3 [1.5; 19.6] | 4.3 [3; 10.2] | 0.206 |
| Excess volume >1 L (miss = 15) = BIS-OH | 15 (83.3) | 17 (85) | 15 (78.9) | 14 (87.5) | 0.918 |
| TBW/FFM | 1.5 [1.3; 4.6] | 0.9 [0.8; 0.9] | 1 [0.9; 1.3] | 1.2 [1.1; 3.6] | <0.001 |
| Phase angle | 4.6 [3.9; 6.5] | 4.9 [3.8; 5.6] | 4.7 [3.9; 5.8] | 4.3 [2.9; 5] | 0.454 |
SOFA, Sequential Organ Failure Assessment; TBW, total body water; ECW, extracellular water; ICW, intracellular water; BIS-OH, bioimpedance spectroscopy–overhydration; FFM, fat-free mass; miss, missing.
Correlations between indices of hydration in the whole cohort and at different levels of delta weight
In the whole cohort, and independently of delta weight, there were always strong correlations between ECW/TBW, CLI, and EV (0.89–1) (Fig. 3, Supplemental Digital Content Figures S3–-S5, available at: http://links.Lww.Com/MS9/B196).
Figure 3.
Correlation in the whole cohort of the different parameters related to fluid status. FB, fluid balance; ECW, extracellular water; ICW, intracellular water; TBW, total body water.
There was a strong positive correlation between delta weight and cumFB (0.8) in the whole cohort; however, this correlation was only moderate when considering delta-weight categories, with only 0.44 when delta weight was less than −5%.
Phase angle was moderately correlated with third-space volume in the whole cohort (−0.37) and when delta weight was above −5% (−0.55), but no correlation was observed when delta weight was under −5%.
In the whole cohort, no correlation was observed between delta weight and/or cumFB and indices obtained by BIA, but some correlation could be observed depending on delta weight.
For the patients with delta weight <−5%, TBW/FFM shows a strong negative correlation with cumFB (−0.8) and delta weight (−0.8). A moderate negative correlation was found between cumFB and third-space volume.
For the patients with a delta weight between −5 and 5%, there were moderate correlations between delta weight, third-space volume, and cumFB (0.44–0.58).
Finally, for the patients with delta weight >5%, a strong correlation could also be observed between third-space volume and TBW/FFM (0.75). There was no correlation between delta weight and/or cumFB and indices obtained by BIA.
Discussion
This prospective study is among the first to evaluate dynamic changes in hydration status in critically ill patients using BIA, in parallel with conventional bedside tools, including clinical assessment, body weight variation, cumFB, and ultrasound. Our findings highlight substantial discrepancies between these commonly used approaches and suggest a complementary role for BIA in the assessment of FO in the ICU.
We observed a moderate correlation between delta weight and cumFB, consistent with prior reports showing only weak-to-moderate agreement between these two surrogates of fluid status[23,24]. Both methods have well-known limitations: FB calculations are prone to inaccuracies and omissions, while daily weighing is often technically challenging in critically ill patients[24,25]. Our results reinforce the notion that neither measure alone reliably reflects tissue-level hydration, and that their combined use remains necessary but insufficient.
In contrast, correlations between weight-based measures, cumFB, and BIA-derived indices were weak. A substantial proportion of patients without clinical signs of FO, including those with significant weight loss (>5%), continued to exhibit BIA-defined FO. This discordance suggests persistent extravascular or interstitial fluid accumulation that is not captured by changes in body weight or FB alone. Similar dissociations have been reported in ICU cohorts, particularly in the context of early fat-free mass loss due to critical illness–related proteolysis[11], which may mask residual extracellular fluid excess despite apparent weight normalization.
Our findings underscore the limited sensitivity of clinical examination and ultrasound for detecting early or moderate FO. Clinical signs typically appear only after substantial extracellular volume expansion, which may remain clinically silent until extracellular volume increases by more than 30% (approximately 4–5 kg)[26]. Several ICU studies have shown that BIA-defined FO may be present despite acceptable cumFB, absence of clinical signs, or normal central venous pressure, highlighting the poor correlation between pressure-based hemodynamic surrogates and tissue-level hydration, which predominantly involves the interstitial rather than the intravascular compartment[7,19,27,28]. Accordingly, the combined use of BIA and biomarkers such as NT-proBNP has been proposed in acute decompensated heart failure, supporting a complementary rather than substitutive role for BIA in fluid assessment[29,30]. Similar dissociations between weight change and BIA-derived hydration indices have been reported in ICU cohorts, with persistent ECW/TBW elevation and declining phase angle despite apparent weight normalization, underscoring the limitations of weight-based metrics for fluid assessment[20,31]. Ultrasound, while valuable for detecting pulmonary congestion, pleural effusions, or ascites, primarily reflects localized or advanced manifestations of FO and remains operator-dependent, with limited sensitivity for diffuse or moderate interstitial fluid accumulation. In this context, BIA provides a global and quantitative assessment of whole-body fluid distribution, enabling earlier detection of tissue-level extracellular fluid expansion that may not yet be apparent on ultrasound[32,33].
Nevertheless, the sensitivity of BIA for moderate fluid shifts remains debated. In heart failure populations, studies have shown correlations between ECW reduction and weight loss[34], but the performance of BIA in detecting moderate changes (<2 L) remains inconsistent. Forni et al[35] and Jones et al[36] each reported that while BIA reliably identified marked overhydration and dehydration, its sensitivity to moderate fluid shifts was limited. Similarly, Ruste et al[21] demonstrated that reductions in ECW and ICW during ultrafiltration strongly correlated with cumFB (ρ = 0.63–0.68), though Bland–Altman analysis indicated poor agreement for fluid shifts under 2 L, suggesting that BIA may be better suited for tracking trends than for precise quantification.
The inverse correlation between third-space volume and phase angle underscores the close link between fluid imbalance and cellular health. As a marker of cell membrane integrity and body cell mass, phase angle provides information complementary to volume-based indices, and values below 4.5° in critically ill patients are consistently associated with worse outcomes[20,37,38]. In patients with persistent BIA-defined FO but preserved phase angle, residual extracellular fluid may be more amenable to cautious removal, whereas a low phase angle likely reflects inflammation or cellular dysfunction and may identify patients at higher risk of hemodynamic intolerance or poor outcomes. Similarly, an elevated CLI suggests ongoing vascular permeability, indicating that aggressive deresuscitation could be poorly tolerated or followed by rapid fluid redistribution.
From a clinical perspective, BIA should therefore be viewed as a risk-stratification tool rather than a trigger for automatic fluid removal. Serial BIA assessments during stabilization and recovery may help identify persistent extracellular FO or unfavorable cellular trajectories, even when conventional markers appear reassuring. In this context, an excess volume or third-space volume greater than approximately 2 L appears to indicate clinically relevant residual FO, whereas isolated abnormalities of redistribution indices (e.g., ECW/TBW >40%) should prompt only cautious fluid removal under close hemodynamic surveillance.
These measurements should be interpreted within a multimodal framework, integrating clinical examination, cumFB, hemodynamic monitoring, and ultrasound findings. Together, these observations generate important hypotheses for future studies, including whether phase angle or CLIs can predict tolerance to deresuscitation or guide its timing and intensity. Although not designed to address these questions definitively, our study highlights the potential value of BIA as a complementary decision-support tool in clinically ambiguous situations where conventional markers of fluid status are inconclusive.
This study has several methodological strengths. Its prospective design allowed standardized, repeated assessment of hydration status using complementary modalities, including clinical examination, ultrasound, and BIA, enabling direct comparison between conventional tools and BIA-derived indices.
However, several limitations must be acknowledged. The study was monocentric and exploratory, with no formal sample size calculation. The small, heterogeneous cohort and repeated measurements within patients limited statistical power, particularly for longitudinal analyses, and precluded robust assessment of associations with clinical outcomes. In addition, systematic hemodynamic, biomarker, and cardiac evaluations were not available, including central venous pressure, NT-proBNP, echocardiographic markers, and indices of capillary leak, such as the CRP/albumin ratio. Consequently, intravascular volume status could not be reliably distinguished from extravascular fluid accumulation, and the prognostic value of BIA-derived parameters for outcomes such as mortality or ICU length of stay could not be evaluated. Larger, multicenter studies with comprehensive hemodynamic and biomarker profiling are therefore warranted.
Conclusion
This study demonstrates that BIA can identify FO in ICU patients who show no clinical or ultrasound evidence of overhydration. These findings suggest that BIA may provide additional diagnostic value in detecting early or subclinical fluid accumulation. Beyond diagnosis, BIA may contribute to fluid management by helping define the patient’s dry weight and guiding the decision to pursue or stop fluid removal strategies.
Further large-scale, prospective, and interventional studies are needed to validate the routine use of BIA in critical care, to define clinically meaningful thresholds or trajectories, and to determine whether BIA-guided fluid management improves patient-centered outcomes.
Acknowledgements
We thank all the patients who participated in this clinical trial, as well as Frédéric Duée and Marine Bereiziat for their active involvement in the conduct of the study.
Footnotes
Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.
Supplemental Digital Content is available for this article. Direct URL citations are provided in the HTML and PDF versions of this article on the journal’s website, www.lww.com/annals-of-medicine-and-surgery.
Contributor Information
Hugo Moiroux, Email: hugo.moiroux63@gmail.com.
Tala Salman, Email: tala.salman@etu.uca.fr.
Julien Aniort, Email: janiort@chu-clermontferrand.fr.
Rémi Andral, Email: Remi.ANDRAL@etu.uca.fr.
Edouard Dugat, Email: edugat@chu-clermontferrand.fr.
Kevin Grapin, Email: kgrapin@chu-clermontferrand.fr.
Jean Christophe Bouennec, Email: jcbouenec@chy-clermontferrand.fr.
Maxime Dumesnil, Email: mdusmenil@chu-clermontferrand.fr.
Bertrand Burel, Email: bburel@chu-clermontferrand.fr.
Laure Calvet, Email: lcalvet@chu-clermontferrand.fr.
Bertrand Souweine, Email: bsouweine@chu-clermontferrand.fr.
Claire Dupuis, Email: cdupuis1@chu-clermontferrand.fr.
Ethical approval
Ethical approval and consent to participate: In line with French regulations, the EBaHIR study was approved by the ethics committee of Clermont-Ferrand University Hospital (IRB00013412, “CHU de Clermont-Ferrand IRB #1,” IRB 2024-CF348) on 25 June 2024. This study was conducted in accordance with the Helsinki Declaration.
Consent
Only the patient’s non-opposition, after informed consent, was required.
Sources of funding
Not applicable.
Author contributions
C.D. and H.M. assisted in data collection and analysis, drafted, and edited the manuscript. T.S., R.A., J.A., E.D., K.G., J.C.B., M.D., B.B., L.C., and B.S. assisted in data collection and edited the manuscript.
Conflicts of interest disclosure
The authors declare that they have no competing interests.
Research registration unique identifying number (UIN)
Not applicable.
Guarantor
Claire Dupuis.
Provenance and peer review
Not applicable.
Data availability statement
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.
References
- [1].Da R, Hise AC, Gonzalez MC. Assessment of hydration status using bioelectrical impedance vector analysis in critical patients with acute kidney injury. Clin Nutr Edinb Scotl 2018;37:695–700. [DOI] [PubMed] [Google Scholar]
- [2].Boyd JH, Forbes J, Nakada T, et al. Fluid resuscitation in septic shock: a positive fluid balance and elevated central venous pressure are associated with increased mortality. Crit Care Med 2011;39:259–65. [DOI] [PubMed] [Google Scholar]
- [3].Kim IY, Kim JH, Lee DW, et al. Fluid overload and survival in critically ill patients with acute kidney injury receiving continuous renal replacement therapy. PLoS ONE 2017;12:e0172137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Wiedemann HP, Wheeler AP, Bernard GR. Comparison of two fluid-management strategies in acute lung injury. J Vasc Surg 2006;44:909. [DOI] [PubMed] [Google Scholar]
- [5].Malbrain MLNG, Marik PE, Witters I, et al. Fluid overload, de-resuscitation, and outcomes in critically ill or injured patients: a systematic review with suggestions for clinical practice. Anaesthesiol Intensive Ther 2014;46:361–80. [DOI] [PubMed] [Google Scholar]
- [6].Hoste EA, Maitland K, Brudney CS, et al. ADQI XII Investigators Group, Four phases of intravenous fluid therapy: a conceptual model. Br J Anaesth 2014;113:740–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Cleymaet R, D’Hondt M, Scheinok T, et al. Comparison of Bioelectrical Impedance Analysis (Bia)-derived Parameters in Healthy Volunteers and Critically Ill Patients. Life Basel Switz 2023;14:27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Basso F, Berdin G, Virzì GM, et al. Fluid management in the intensive care unit: bioelectrical impedance vector analysis as a tool to assess hydration status and optimal fluid balance in critically ill patients. Blood Purif 2013;36:192–99. [DOI] [PubMed] [Google Scholar]
- [9].Köster M, Dennhardt S, Jüttner F, et al. Cumulative changes in weight but not fluid volume balances reflect fluid accumulation in ICU patients. Acta Anaesthesiol Scand 2017;61:205–15. [DOI] [PubMed] [Google Scholar]
- [10].Tolstrup J, Brandstrup B. Clinical assessment of fluid balance is incomplete for colorectal surgical patients. Scand J Surg SJS Off Organ Finn Surg Soc Scand Surg Soc 2015;104:161–68. [DOI] [PubMed] [Google Scholar]
- [11].Puthucheary ZA, Rawal J, McPhail M, et al. Acute skeletal muscle wasting in critical illness. Jama 2013;310:1591–600. [DOI] [PubMed] [Google Scholar]
- [12].Beurton A, Teboul J-L, Monnet X. Transpulmonary thermodilution techniques in the haemodynamically unstable patient. Curr Opin Crit Care 2019;25:273–79. [DOI] [PubMed] [Google Scholar]
- [13].Malbrain MLNG, Huygh J, Dabrowski W, et al. The use of bio-electrical impedance analysis (BIA) to guide fluid management, resuscitation and deresuscitation in critically ill patients: a bench-to-bedside review. Anaesthesiol Intensive Ther 2014;46:381–91. [DOI] [PubMed] [Google Scholar]
- [14].Meguid MM, Lukaski HC, Tripp MD, et al. Rapid bedside method to assess changes in postoperative fluid status with bioelectrical impedance analysis. Surgery 1992;112:502–08. [PubMed] [Google Scholar]
- [15].Denneman N, Hessels L, Broens B, et al. Fluid balance and phase angle as assessed by bioelectrical impedance analysis in critically ill patients: a multicenter prospective cohort study. Eur J Clin Nutr 2020;74:1410–19. [DOI] [PubMed] [Google Scholar]
- [16].Costa D, Muzzio M, Saglietti L, et al. Fluid status after cardiac surgery assessed by bioelectrical impedance vector analysis and the effects of extracorporeal circulation. J Cardiothorac Vasc Anesth 2021;35:2385–91. [DOI] [PubMed] [Google Scholar]
- [17].Moonen HPFX, Van Zanten ARH. Bioelectric impedance analysis for body composition measurement and other potential clinical applications in critical illness. Curr Opin Crit Care 2021;27:344–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Roos AN, Westendorp RG, Frölich M, et al. Weight changes in critically ill patients evaluated by fluid balances and impedance measurements. Crit Care Med 1993;21:871–77. [DOI] [PubMed] [Google Scholar]
- [19].Samoni S, Vigo V, Reséndiz LIB, et al. Impact of hyperhydration on the mortality risk in critically ill patients admitted in intensive care units: comparison between bioelectrical impedance vector analysis and cumulative fluid balance recording. Crit Care Lond Engl 2016;20:95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Ellegård LH, Petersen P, Öhrn L, et al. Longitudinal changes in phase angle by bioimpedance in intensive care patients differ between survivors and non-survivors. Clin Nutr ESPEN 2018;24:170–72. [DOI] [PubMed] [Google Scholar]
- [21].Ruste M, Chabanol C, Fellahi J-L, et al. Bioelectrical impedance analysis during deresuscitation: correlation and agreement with cumulative fluid balance in ICU patients. J Clin Monit Comput 2023;37:679–87. [DOI] [PubMed] [Google Scholar]
- [22].Rodríguez-Moguel N, Osuna-Padilla IA, Piekarska KB, et al. Fluid status assessment in critically Ill patients with COVID-19: a retrospective cohort study. J Clin Med 2024;13:540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Mensink RSM, Paans W, Renes MH, et al. Fluid balance versus weighing: a comparison in ICU patients: a single center observational study. PloS One 2024;19:e0299474. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Davies H, Leslie G, Jacob ER, et al. Estimation of body fluid status by fluid balance and body weight in critically Ill adult patients: a systematic review. Worldviews Evid Based Nurs 2019;16:470–77. [DOI] [PubMed] [Google Scholar]
- [25].Schneider AG, Baldwin I, Freitag E, et al. Estimation of fluid status changes in critically ill patients: fluid balance chart or electronic bed weight? J Crit Care 2012;27:745.e7–12. [DOI] [PubMed] [Google Scholar]
- [26].Piccoli A. Bioelectric impedance measurement for fluid status assessment. Contrib Nephrol 2010;164:143–52. [DOI] [PubMed] [Google Scholar]
- [27].Piccoli A, Pittoni G, Facco E, et al. Relationship between central venous pressure and bioimpedance vector analysis in critically ill patients. Crit Care Med 2025;28:132. [DOI] [PubMed] [Google Scholar]
- [28].Karpavičiūtė J, Skarupskienė I, Balčiuvienė V, et al. Assessment of fluid status by bioimpedance analysis and central venous pressure measurement and their association with the outcomes of severe acute kidney injury. Med Kaunas Lith 2021;57:518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Valle R, Aspromonte N, Milani L, et al. Optimizing fluid management in patients with acute decompensated heart failure (ADHF): the emerging role of combined measurement of body hydration status and brain natriuretic peptide (BNP) levels. Heart Fail Rev 2011;16:519–29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Lee K-S, Kim J-H, Kang J, et al. Association between changes in Bioelectrical Impedance Analysis (BIA) parameter and the clinical outcomes in patients with acute heart failure. J Korean Med Sci 2023;38:e276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Peng J, Yang J, Li F. Association of bioimpedance analysis parameters trajectories with clinical outcomes in neurocritical patients. Heliyon 2024;10:e32948. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Ciumanghel A, Siriopol I, Blaj M, et al. B-lines score on lung ultrasound as a direct measure of respiratory dysfunction in ICU patients with acute kidney injury. Int Urol Nephrol 2018;50:113–19. [DOI] [PubMed] [Google Scholar]
- [33].Beaubien-Souligny W, Galarza L, Buchannan B, et al. Prospective study of ultrasound markers of organ congestion in critically Ill patients with acute kidney injury. Kidney Int Rep 2024;9:694–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [34].Sakaguchi T, Yasumura K, Nishida H, et al. Quantitative assessment of fluid accumulation using bioelectrical impedance analysis in patients with acute decompensated heart failure. Circ J 2015;79:2616–22. [DOI] [PubMed] [Google Scholar]
- [35].Forni LG, Hasslacher J, Joannidis M. Bioelectrical impedance vector analysis in the critically ill: cool tool or just another “toy”? Crit Care Lond Engl 2015;19:387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].Jones SL, Tanaka A, Eastwood GM, et al. Bioelectrical impedance vector analysis in critically ill patients: a prospective, clinician-blinded investigation. Crit Care Lond Engl 2015;19:290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [37].Thibault R, Makhlouf A-M, Mulliez A, et al. Phase Angle Project Investigators, Fat-free mass at admission predicts 28-day mortality in intensive care unit patients: the international prospective observational study Phase Angle Project. Intensive Care Med 2016;42:1445–53. [DOI] [PubMed] [Google Scholar]
- [38].Lima J, Eckert I, Gonzalez MC, et al. Prognostic value of phase angle and bioelectrical impedance vector in critically ill patients: a systematic review and meta-analysis of observational studies. Clin Nutr 2022;41:2801–16. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.