Prognostic value of bioelectrical impedance analysis in heart failure: A systematic review of congestion assessment and clinical outcomes

Abstract

Congestion is a major driver of adverse outcomes in heart failure, yet its objective assessment in clinical practice remains challenging. Bioelectrical impedance analysis (BIA) offers a non-invasive evaluation of fluid distribution and cellular integrity. We conducted a systematic review in accordance with PRISMA 2020, searching PubMed/MEDLINE, Embase, Cochrane CENTRAL and Web of Science from database inception to 30 November 2025 for studies evaluating BIA-derived parameters and clinically relevant outcomes in adult patients with acute or chronic heart failure. Thirteen studies met the inclusion criteria; all were observational and most were single-centre. In acute heart failure, elevated extracellular water, abnormal impedance vectors and increased edema indices were associated with rehospitalization and mortality, particularly when congestion persisted at discharge, although in one cohort this association did not persist after adjustment for age and natriuretic peptide. In chronic heart failure, BIA detected subclinical fluid overload associated with worsening heart failure events and cardiovascular death. Among the studies reporting multivariable-adjusted hazard ratios, values for adverse outcomes ranged from 1.96 to 2.67. Lower phase angle was also associated with increased mortality, although phase angle reflects cellular integrity and nutritional status rather than congestion specifically. The evidence base is heterogeneous with respect to devices, parameters, cut-offs, timing of measurement and outcome definitions, and quantitative pooling was not appropriate; the review was not prospectively registered and small-study effects and selective publication cannot be excluded. BIA-derived parameters are associated with clinically relevant outcomes in heart failure and may complement conventional congestion assessment, but no interventional evidence currently supports BIA-guided management, which remains to be tested.

Graphical abstract

1. Introduction

Heart failure (HF) remains one of the leading causes of hospitalization worldwide and continues to carry substantial morbidity and mortality despite advances in pharmacological and device-based therapies [1], [2]. Episodes of acute decompensation are particularly relevant, as each hospitalization is associated with incremental short- and long-term risk [1]. At the center of this clinical trajectory lies congestion.

Congestion in HF reflects a complex interplay of elevated cardiac filling pressures, neurohormonal activation, and renal sodium and water retention [2], [3], [4]. Activation of the renin–angiotensin–aldosterone system and sympathetic pathways promotes extracellular fluid expansion, which in turn contributes to pulmonary edema, renal dysfunction, and impaired tissue perfusion [3], [4], [5]. Importantly, a substantial proportion of patients are discharged with residual congestion, and this residual burden has been repeatedly linked to early readmission and mortality [6], [7].

Yet, accurately defining and quantifying congestion remains challenging. Clinical examination lacks both sensitivity and reproducibility [8]. Natriuretic peptides are indispensable for diagnosis and prognostic but reflect myocardial wall stress rather than direct measurement of extracellular volume [2], [9]. Imaging techniques such as lung ultrasound provide useful information about pulmonary interstitial edema [10], [11], although they do not offer a comprehensive assessment of systemic fluid distribution.

Bioelectrical impedance analysis (BIA) represents an alternative physiological approach. By measuring tissue resistance and reactance to a low-amplitude alternating current, BIA estimates total body water (TBW) and extracellular water (ECW) [12]. BIA allows graphical evaluation of hydration status without reliance on regression-based assumptions [13]. Because HF-related congestion is largely characterized by extracellular fluid expansion, ECW-based indices appear conceptually well suited to capture congestion burden.

Over the past decade, several observational studies have explored the relationship between BIA-derived parameters and outcomes in HF. In acute settings, abnormal impedance measurements at discharge have been associated with rehospitalization and mortality [14], [15], [16], while other studies have related impedance-derived indices to the volume of fluid removal required during hospitalization [17] and to hospital length of stay [18]; serial measurements during admission appear to track the response to decongestive therapy [19], [20]. In chronic and ambulatory populations, subclinical volume overload detected by BIA has been linked to worsening HF events and death [21], [22], [23]. More recently, phase angle has been proposed as a complementary prognostic marker reflecting cellular integrity and systemic vulnerability [24], [25], [26].

Despite this expanding literature, methodological variability and differences in reported endpoints have limited integration of BIA into routine clinical decision-making. A structured synthesis focusing specifically on congestion quantification and clinically meaningful outcomes across acute and chronic HF is therefore warranted.

Given the limitations of current congestion assessment in routine clinical practice, tools that allow objective and reproducible quantification of fluid status may have direct implications for patient management.

The aim of the present systematic review is to evaluate the role of BIA in adult patients with HF, with particular attention to its ability to quantify congestion and predict rehospitalization and mortality.

2. Methods

2.1. Study design and objectives

This systematic review was conducted and reported in accordance with the PRISMA 2020 statement; the completed PRISMA 2020 checklist is provided in Supplementary Appendix S2. The objective was to evaluate the role of BIA in the assessment of congestion and prediction of clinically relevant outcomes in adult patients with HF.

2.2. Search strategy

A comprehensive literature search was performed in PubMed/MEDLINE, Embase, the Cochrane Central Register of Controlled Trials (CENTRAL) and Web of Science. The search was carried out between 1 September and 30 November 2025 and covered the period from database inception to 30 November 2025, which was the pre-specified closing date of the search. Study selection, data extraction and the drafting of this manuscript were completed within that same period, and studies published after 30 November 2025 were therefore not eligible for inclusion. The search strategy combined controlled vocabulary (MeSH terms) and free-text keywords related to “heart failure,” “acute heart failure,” “chronic heart failure,” “bioelectrical impedance,” “bioimpedance,” “impedance vector analysis,” and “phase angle.” No restriction on publication year was applied. Only studies published in English were considered eligible. Reference lists of relevant articles were manually screened, and no additional eligible studies were identified by this route. The full Boolean search string for each database, together with the exact dates searched and the limits applied, is provided in Supplementary Appendix S1.

2.3. Study selection

All records were imported into a reference management database and duplicates were removed. Two reviewers independently screened titles and abstracts. Full texts were retrieved for potentially eligible studies and assessed for inclusion. Disagreements were resolved by consensus or consultation with a third reviewer.

Studies were included if they met the following criteria: (1) adult patients with HF; (2) assessment of BIA-derived parameters related to fluid distribution or cellular integrity (including ECW, TBW, ECW/TBW ratio, impedance vectors or phase angle); and (3) reporting of at least one clinically relevant outcome (e.g., mortality, HF rehospitalization, or worsening HF events).

Exclusion criteria included case reports, small case series (<10 patients), narrative reviews, editorials, conference abstracts without full data, and studies focused exclusively on dialysis populations without specific HF analysis.

2.4. Data extraction

Data extraction was performed independently by two reviewers using a predefined template. Extracted variables included study design, sample size, patient characteristics, HF phenotype, BIA methodology, parameters analysed, timing of measurements, follow-up duration, and reported outcomes. When available, multivariable-adjusted effect estimates (hazard ratios or odds ratios) were preferentially recorded.

2.5. Quality assessment

Methodological quality was assessed using the Newcastle–Ottawa Scale (NOS) for cohort studies across three domains: Selection (maximum 4 stars), Comparability (maximum 2 stars) and Outcome (maximum 3 stars), giving a maximum of 9 stars. A score of ≥7 was considered to indicate low risk of bias. Each study was assessed independently by two reviewers, and disagreements were resolved by discussion and, where necessary, by consultation with a third reviewer.

The NOS was not applied to four of the included studies (Yamazoe et al., Massari et al., De Ieso et al. and Castillo-Martínez et al.), none of which reported a mortality or heart-failure event outcome assessed over a defined follow-up period. Yamazoe et al. reported an association with a weight-loss requirement rather than a clinical event; the outcome in Massari et al. was hospital length of stay rather than mortality or a post-discharge event; De Ieso et al. was a method-validation study correlating BIA-derived fluid changes with weight loss during admission, without follow-up towards a subsequent clinical outcome; and Castillo-Martínez et al. was a single-timepoint cross-sectional comparison of BIA parameters across NYHA class, with no follow-up period. Because the Comparability and Outcome domains of the NOS presuppose an exposure–outcome relationship assessed over follow-up, these four studies were marked as not applicable rather than assigned a partial score, so as not to imply a degree of comparability with the prognostic studies that the scale was not designed to capture.

Among the nine studies to which the scale was applied, NOS scores ranged from 6 to 9 (median 8), and eight of the nine scored ≥7. Domain-level scores and the justification for each are reported in Supplementary Table S1.

2.6. Data synthesis

Given the heterogeneity across studies in terms of patient populations, BIA techniques, timing of measurements, and outcome definitions, a formal meta-analysis was not performed. Instead, results were synthesized qualitatively and grouped according to clinical context (acute HF and chronic HF) and, separately, according to studies in which phase angle was the parameter of interest. Because phase angle was evaluated in both the acute and the chronic setting, these groupings are descriptive and are not mutually exclusive; the clinical setting of each individual study is reported separately in Table 1. Quantitative findings were reported descriptively when comparable data were available.

Table 1.

Characteristics and outcomes of included studies evaluating bioimpedance analysis (BIA) parameters in acute and chronic heart failure.

Studies are grouped by outcome type: prognostic (mortality/HF events), intermediate clinical outcomes, and severity-association (cross-sectional).

Study	Design/Setting	N	HF phenotype/setting	BIA parameter & device	Timing	Outcome (definition, follow-up)	Adjusted effect estimate (95% CI)
Santarelli et al. 2017a [14]	Prospective, multicentre (3 Italian ED), observational, AHF vs non-AHF comparison	n = 336 (221 AHF, 115 non-AHF)	Acute HF, hospitalized (Emergency Dept.)	BIVA (bioelectrical impedance vector analysis)	Admission (t0) and discharge (tdis)	Composite clinical events at 90 days (rehospitalization/death), telephone follow-up	AUC 0.56 (t0), 0.57 (discharge); combined with clinical signs AUC 0.97. No multivariate model including BIVA variables was reported; univariate Cox analysis showed a coefficient of 0.417 (p < 0.01) for an R change >11 Ohm/m. Clinical signs (rales, oedema, JVD) significant unadjusted ORs on univariate analysis.
Santarelli et al. 2017b [15]	Prospective, multicentre (Italy + Brazil), observational	n = 292	Acute HF, hospitalized	BIVA (hydration index, reactance, resistance) + admission BNP	BNP at admission; BIVA at discharge	90-day cardiovascular mortality, telephone follow-up	AUC 0.715 (HI), 0.712 (Xc), 0.65 (R); combined with BNP AUC 0.74; a multivariate logistic regression was performed (age, BNP, hydration index, R, Xc, rales, entered together); only age (OR 1.05, 95% CI 1.01–1.11, p = 0.01) and BNP (OR 1.00, p = 0.01) remained independently significant; hydration index (OR 1.10, 95% CI 0.97–1.25, p = 0.12), R (OR 0.99, p = 0.42), Xc (OR 1.00, p = 0.90), and rales (OR 1.61, p = 0.36) were not independent predictors once adjusted.
Sakaguchi et al. 2015 [16]	Prospective, single-centre (Osaka National Hospital), observational, with matched non-HF control group for equation derivation	n = 130 ADHF (60 non-HF controls for reference equations)	Acute decompensated HF, hospitalized	Multi-frequency BIA (BioScan 920–2), ECW measured/predicted (M/P) ratio	Admission and discharge	Cardiac death or ADHF readmission within 6 months post-discharge	Multivariate HR 1.483 (95% CI 1.200–1.834) per 0.1-unit increase in ECW M/P ratio at discharge, adjusted for prior ADHF admission (p < 0.001).
Yamazoe et al. 2015 [17]	Prospective, single-centre (St. Luke's International Hospital, Tokyo), observational	n = 60	Acute HF, hospitalized	8-polar multifrequency BIA (InBody S10); Edema Index (EI = ECW/TBW)	Admission only (weight-reduction target modeled for the hospitalization period)	Reports the amount of weight reduction (kg) needed during hospitalization to reach “dry weight”/decongestion (a decongestive-therapy-guidance target)	Multivariate linear regression for required weight reduction: EI coefficient 100.62 (95% CI 57.64–143.59), p < 0.001; adjusted R2 = 0.631; other independent predictors: male sex, IVC diameter, admission weight.
Massari et al. 2019 [18]	Retrospective, multicentre, Italy	n = 706	Acute HF, hospitalized	BIVA (hydration index)	At admission	Hospital length of stay (LOS)	Significant association (BNP, BUN, and hydration index independently associated with length of stay on multivariate analysis)
De Ieso et al. 2021 [19]	Observational, single-centre (University Hospital Basel, Switzerland)	n = 142 (867 repeated BCA/BIVA assessments)	Acute HF, hospitalized, undergoing intensified diuretic therapy	8-electrode segmental BIVA/body composition analyser	Daily, repeated during hospitalization	Correlation of BIVA-derived TBW/ECW change with daily body-weight change during admission	Repeated-measures correlation: rho 0.43 (weight loss vs TBW loss) and rho 0.71 (weight loss vs ECW loss), both p < 0.01.
Lee et al. 2023 [20]	Retrospective, single-centre (Seoul National University Hospital, Korea)	n = 116 (of 310 screened)	Acute decompensated HF, hospitalized; 55.8% HFrEF	BIA (InBody-type), serial measurement: edema index (ECW/TBW)	Admission and discharge (change/delta calculated)	Composite: in-hospital mortality or rehospitalization for worsening HF within 1 month	Fully adjusted model (age, sex, hypertension, diabetes, CKD, prior revascularization, BMI): adjusted OR 4.8 (95% CI 1.2–19.3, p = 0.029) for high/positive vs low/negative delta-ECW/TBW group (23.2% vs 8.3% event rate). Alternative continuous-variable model (adjusted for CKD, diuretics, NT-proBNP): OR 12.55 (95% CI 1.25–97.03, p = 0.018). AUC for delta-ECW/TBW: 0.71 (95% CI 0.61–0.79).
Sasmita et al. 2024 [21]	Prospective, single-centre (Chongqing Medical University, China), observational	n = 401	Chronic HF, outpatient/ambulatory	BIA (InBody S10), Edema Index (EI = ECW/TBW)	At enrollment (admission EI)	Long-term all-cause and cardiovascular mortality (median follow-up 1200 days, ∼3.3 years)	Multivariate Cox: EI >0.397 independent predictor of all-cause mortality, HR 1.959 (95% CI 1.304–2.944, p = 0.001), and cardiovascular mortality, HR 2.051 (95% CI 1.276–3.296, p = 0.003).
Bragança et al. 2025 [22]	Prospective, single-centre (Unidade Local de Saúde Tâmega e Sousa, Portugal), observational	n = 83	Chronic HF, HFrEF outpatients	Segmental multi-frequency BIA (InBody BWA 2.0), ECW/TBW z-score (subclinical congestion = z-score > 2 without clinical signs; clinical congestion = z-score > 2 with clinical signs)	At enrollment (ambulatory visit)	Composite: all-cause death + worsening HF (WHF), median follow-up 10 months	Multivariate (adjusted for CKD, NYHA class, haemoglobin, age, sex, diabetes, hypertension, AF, HF aetiology, LVEF, and NT-proBNP): subclinical congestion HR 9.4 (95% CI 1.04–85.1, p = 0.046); clinical congestion HR 17 (95% CI 1.11–261, p = 0.042). Composite outcome occurred in 22/83 (27%) patients, mostly WHF events
Rodríguez-López et al. 2024 [23]	Prospective, single-centre (Fundación Jiménez Díaz University Hospital, Madrid), observational	n = 100	Chronic HF, HFrEF outpatients (LVEF <40%)	BIA dry-weight method; volume overload = measured weight minus BIA-derived dry weight (optimal cut-off 1.2 L)	At enrollment, followed 3 months	Worsening HF (WHF): ED visit or hospitalization for HF, 3-month follow-up (11/100 events)	Multivariate adjusted OR 2.70 (95% CI 1.30–5.63, p = 0.008); adjusted HR 2.03 (95% CI 1.37–3.02, p < 0.001). AUC 0.88 (95% CI 0.79–0.97, p < 0.001), optimal cut-off 1.2 L (91% sensitivity, 75% specificity). Final model via backward stepwise selection (p < 0.05) from a candidate set (E/e′, LVEF, signs of congestion, IVC dilatation, NT-proBNP)
Alves et al. 2016 [24]	Prospective, single-centre (Hospital de Clínicas de Porto Alegre, Brazil), observational	n = 71	Acute HF (ADHF), LVEF <45%	Tetrapolar BIA, phase angle (cut-off <4.8° by ROC)	At admission (first 36 h)	All-cause mortality, average 24 months after discharge	Multivariate (adjusted for age, LVEF, urea): HR 2.67 (95% CI 1.21–5.89), p = 0.015. Mortality 49% (35/71) at average 24-month follow-up; cohort also required Boston criteria ≥8 points (acute HF severity score) for inclusion
Scicchitano et al. 2022 [25]	Retrospective, single-centre (F. Perinei Hospital, Altamura, Bari, Italy), consecutive AHF patients enrolled Jan 2010-Nov 2013	n = 252	Acute HF (ADHF)	Bioimpedance vector analysis; phase angle (cut-off ≤4.9°) as one component of a 4-variable multiparametric score (BNP, BUN, PaO2, phase angle)	At admission	All-cause mortality, median follow-up 447 days	Multivariate Cox (adjusted for BNP, BUN, PaO2 — age tested but not retained in the final model) HR 0.72 (95% CI 0.57–0.91, p = 0.008) per 1-degree increase (i.e., lower phase angle = higher risk; cut-off ≤4.9°, AUC 0.68 [0.62–0.74]). Separately, HR 2.2 (95% CI 1.8–2.8, p < 0.0001) is for a 1-unit increase in a derived 4-variable composite score (BNP + BUN+PaO2 + phase angle, each dichotomized)
Castillo-Martínez et al. 2007 [26]	Cross-sectional, single-centre (Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico), observational	n = 243 (140 HFrEF +103 HFpEF)	Chronic HF, HFrEF and HFpEF, outpatient	Tetrapolar multi-frequency BIA (BodyStat QuadScan 4000): phase angle, reactance, impedance ratio, BIVA vector	Single time point	Cross-sectional comparison of BIA parameters across NYHA class (I–II vs III–IV); no follow-up, no clinical endpoint.	Significant findings (NYHA I-II vs III-IV, by gender, age/sex-adjusted p-values): in both HF groups, reactance (Xc) and phase angle (PA) were significantly lower, and the impedance ratio Z200/Z5 was significantly higher, in NYHA III-IV. HFrEF: Xc p = 0.03, PA p = 0.04, Z200/Z5 p = 0.01. HFpEF: Xc p = 0.008, PA p = 0.01, Z200/Z5 p = 0.001. Impedance vector position differed significantly between NYHA groups in women (HFrEF p < 0.02; HFpEF p < 0.001) with a trend in men (p = 0.06); NYHA III-IV vectors were shorter and downsloping in both HF categories/genders, and in men with HFrEF fell outside the lower pole of the 95% tolerance ellipse. Handgrip strength was also significantly lower in NYHA III-IV (HFrEF, p = 0.001)

2.7. Protocol registration

The review protocol was not prospectively registered.

3. Results

3.1. Study characteristics

The study selection process is summarized in Fig. 1. The main characteristics of the included studies are summarized in Table 1. The final set of studies included a mix of clinical contexts. Seven were conducted in patients hospitalized with acute heart failure (AHF) [14], [15], [16], [17], [18], [19], [20] and three in ambulatory populations with chronic heart failure (CHF) [21], [22], [23]; a further three examined phase angle as the parameter of interest [24], [25], [26]. This last group does not represent a separate clinical setting: two of these studies were conducted in acute HF [24], [25] and one was a cross-sectional study in ambulatory patients [26].

Fig. 1.

PRISMA 2020 flow diagram of study identification, screening and inclusion. Databases were searched from inception to 30 November 2025. This figure was prepared with the assistance of generative artificial intelligence tools (ChatGPT, OpenAI; Claude, Anthropic) and was reviewed and edited by the authors, who take full responsibility for its content.

Most investigations followed a prospective observational design. Sample sizes varied widely, ranging from 60 to 706 patients. Follow-up duration also differed across studies, spanning from in-hospital outcomes to a median follow-up of approximately 3.3 years.

In terms of methodology, there was notable heterogeneity. Some studies used multifrequency BIA and several derived secondary indices such as ECW, ECW/total body water ratios, edema index, or vector displacement.

Despite these methodological differences, most studies shared a common focus: the relationship between impedance-derived measures and clinically meaningful outcomes, including mortality, rehospitalization, or composite endpoints.

3.2. Bioelectrical impedance in acute heart failure

Among patients admitted with AHF, a broadly concordant pattern emerged across studies [14], [15], [16], [17], [18], [19], [20]. Measures reflecting extracellular fluid expansion—whether expressed as ECW, ECW/TBW ratio, or impedance vector displacement—were repeatedly associated with subsequent adverse events.

Several studies focused on the time of discharge. Patients who remained congested according to bioimpedance parameters, even when clinically stable, showed higher rates of rehospitalization and mortality during follow-up [14], [15], [16], although in one of these cohorts the association between BIVA parameters and 90-day cardiovascular mortality did not persist after adjustment for age and BNP [15]. This observation was not limited to a single technique or index but appeared across different methodological approaches.

Longitudinal assessments added further nuance. In cohorts where BIA was measured serially during hospitalization, failure to normalize ECW or ECW/TBW was associated with worse long-term outcomes [19], [20]. In other words, clinical improvement did not always translate into physiological decongestion.

In addition, impedance-derived indices showed a relationship with the intensity of decongestive therapy. Higher baseline congestion was associated with a greater amount of fluid removal required to achieve clinical stability [17]. This finding suggests a potential, although still theoretical, role for BIA in guiding treatment.

Phase angle also showed prognostic relevance in the acute setting. Lower values were associated with increased mortality risk, with reported hazard ratios around 2.67 after multivariable adjustment [24]. Similarly, in studies incorporating phase angle into multiparametric models, the association with long-term mortality remained robust, with a hazard ratio of 2.2 per 1-unit increase in that composite score; in the same study the estimate for phase angle alone was an adjusted hazard ratio of 0.72 (95% CI 0.57–0.91) per 1-degree increase, indicating a higher risk at lower phase angle values [25].

Taken together, these findings indicate that bioimpedance-derived parameters capture aspects of congestion that are not fully reflected by clinical assessment alone. The direction of the signal is concordant across studies, although thresholds and methodologies differ.

3.3. Bioelectrical impedance in chronic heart failure

In ambulatory populations with CHF, the role of BIA appears to shift slightly. Rather than identifying overt congestion, it seems to detect earlier, subclinical stages of fluid accumulation [21], [22], [23].

Across studies, patients with higher ECW or elevated ECW/TBW ratios experienced more frequent worsening HF events and higher mortality. Notably, these associations were observed in patients who were otherwise considered clinically stable.

In one cohort of patients with HF reduced ejection fraction, bioimpedance-derived volume overload emerged as the only independent predictor of worsening HF at short-term follow-up, with adjusted hazard ratios exceeding 2 [22]. In a chronic HF cohort followed for a median of 1200 days, an edema index above 0.397 was associated with an approximately twofold increase in all-cause mortality (adjusted HR 1.96, 95% CI 1.30–2.94) and in cardiovascular mortality (adjusted HR 2.05, 95% CI 1.28–3.30) [23].

These observations suggest that extracellular fluid expansion may develop progressively before becoming clinically evident. In this context, BIA may offer a window into early pathophysiological changes that precede overt decompensation.

3.4. Phase angle and prognosis

A smaller group of studies specifically examined phase angle [24], [25], [26]. Although fewer in number, their findings pointed in the same direction. One of these studies was cross-sectional and reported an association with NYHA class rather than with a clinical event [26].

Lower phase angle values were associated with increased mortality in both AHF and CHF settings. The strength of association was comparable to that observed for fluid-related indices, with adjusted hazard ratios typically exceeding 2.

Unlike ECW-based parameters, phase angle does not directly reflect fluid distribution. Instead, it appears to capture broader aspects of physiological status, including cellular integrity and metabolic reserve. This may explain why its prognostic value extends across different clinical contexts.

3.5. Overall quantitative patterns

The included studies reported different effect measures, so the estimates below are not directly comparable and were not pooled. Among the studies reporting multivariable-adjusted hazard ratios, values for mortality or worsening HF ranged from 1.96 to 2.67 [22], [23], [24]. Adjusted odds ratios were 2.70 (95% CI 1.30–5.63) for worsening HF in ambulatory patients [22] and 4.8 (95% CI 1.2–19.3) for a positive change in ECW/TBW during admission [20]. Phase angle was analysed as a continuous variable, with an adjusted hazard ratio of 0.72 per 1-degree increase (95% CI 0.57–0.91), that is, a higher risk at lower values [25]. Two studies reported only discrimination statistics or unadjusted estimates for the impedance parameters [14], [15].

The direction of the reported associations was broadly concordant, with higher extracellular fluid indices and lower phase angle values linked to worse outcomes. This should nevertheless be interpreted with caution. The included studies used different effect measures and different endpoints, and in one cohort the association between BIVA parameters and mortality did not persist after multivariable adjustment [15]. Given the small number of studies, the predominance of single-centre cohorts and the likelihood of selective publication, the apparent uniformity of direction may partly reflect what has been published rather than the underlying evidence.

Devices, thresholds and study designs differed substantially, which limits the extent to which these findings can be compared directly or attributed to a single underlying mechanism.

These findings were observed in both acute and chronic HF settings.

4. Discussion

This review brings together a set of studies that, although diverse in design and methodology, point in a similar direction. BIA does not simply mirror clinical congestion; rather, it appears to capture aspects of fluid distribution and physiological vulnerability that are not fully accessible through routine assessment. When the evidence is considered as a whole, the association between impedance-derived parameters and clinically relevant outcomes is apparent across most, although not all, of the included studies [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26].

Previous reviews of congestion assessment in heart failure have focused predominantly on natriuretic peptides or on lung ultrasound. This review addresses BIA-derived parameters specifically, across both the acute and the chronic setting, and considers phase angle separately from extracellular-water indices because the two capture different constructs. It also reports, at study level, whether each association was adjusted and for which covariates, so that the strength of the underlying evidence can be judged rather than assumed.

Three observations can be outlined, although their boundaries are not entirely rigid.

In the AHF setting, the most consistent signal relates to residual congestion at discharge. Several studies showed that patients who remained “wet” according to bioimpedance parameters—despite clinical stabilization—had a higher risk of early rehospitalization and mortality [14], [15], [16], although in one of these cohorts this association did not persist after multivariable adjustment [15]. Congestion has long been recognized as a central driver of outcomes in HF [1], [2], [3] and residual congestion, in particular, has been repeatedly linked to early events after discharge [6], [7]. What BIA adds is a way of quantifying this state more objectively. It makes visible what clinical examination often misses.

At the same time, the data suggest that clinical improvement and physiological decongestion do not always move in parallel. In studies where impedance was measured over time, some patients improved symptomatically while their ECW remained elevated [19], [20]. Those patients tended to do worse. This raises a practical question: are we discharging patients too early, guided more by symptoms than by underlying fluid status? The available evidence does not provide a definitive answer, but it clearly challenges the assumption that clinical stability equals euvolemia.

In CHF, the picture is slightly different. Here, BIA seems to identify patients before overt congestion develops. Individuals classified as having higher ECW—often without clear clinical signs—showed increased rates of worsening HF and death [21], [22], [23]. This aligns with the idea that fluid accumulation is a gradual process. By the time edema or weight gain becomes obvious, the underlying pathophysiology has already been evolving for some time. In that sense, BIA may act as an early signal rather than a confirmatory tool.

The magnitude of these associations is also worth noting. Across the studies reporting adjusted hazard ratios, values for adverse outcomes fell between 1.96 and 2.67 [22], [23], [24]. Associations of this size are unlikely to be trivial, but they derive from different populations, parameters and endpoints and should not be read as a single pooled effect; residual confounding remains a plausible alternative explanation.

Phase angle adds another layer to the interpretation. Unlike ECW or ECW/TBW, it does not measure fluid directly. Instead, it reflects cellular integrity and body composition [12]. Lower values have been linked to worse outcomes in both acute and chronic settings [24], [25], [26]. One way to interpret this is that congestion is only part of the story. Patients with HF often exhibit features of systemic illness—sarcopenia, inflammation, nutritional decline—and phase angle may be capturing that broader vulnerability. In practice, this means that two patients with similar fluid status may not carry the same risk if their underlying physiological reserve differs.

When placed alongside traditional tools, BIA occupies an interesting position. Clinical examination is simple but often imprecise [8]. Natriuretic peptides are well validated but reflect pressure rather than volume [9]. Imaging techniques such as lung ultrasound provide valuable information, yet remain focused on the pulmonary compartment [10], [11]. BIA, by contrast, offers a whole-body perspective. None of the included studies provided a direct head-to-head comparison of BIA against natriuretic peptides or lung ultrasound, so the suggestion that BIA might complement these methods in situations where the clinical picture is ambiguous remains a hypothesis rather than a comparative finding.

That said, caution is necessary. All studies included in this review are observational [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26]. Even when adjusted analyses are performed, confounding cannot be fully excluded. Patients with higher ECW may simply be sicker in ways that are not fully captured by measured variables. In addition, methodological variability is considerable. Devices differ, thresholds are not standardized, and the timing of measurements varies across studies. These issues limit comparability and make it difficult to define clear clinical cut-offs.

Perhaps the most important unanswered question is whether BIA can guide therapy in a way that improves outcomes. At present, there are no randomized trials demonstrating that adjusting treatment based on impedance measurements reduces mortality or rehospitalization. The concept is attractive—especially in light of the associations observed—but it remains hypothetical.

From a clinical perspective, the current evidence supports a cautious but interested stance. BIA appears to provide information that is not redundant with standard assessment. It may be particularly useful in identifying residual congestion at discharge and in detecting subclinical fluid accumulation in ambulatory patients. Whether this information should change management decisions is less clear.

Looking ahead, several areas deserve attention. Standardization of measurement protocols would be an important first step. Establishing clinically meaningful thresholds for ECW, ECW/TBW and phase angle is equally critical. Most importantly, interventional studies are needed. Without them, BIA will remain a promising marker rather than a tool that actively shapes patient care.

From a clinical perspective, BIA may represent a pragmatic and non-invasive tool to refine congestion assessment in routine practice. Its ability to detect residual congestion at discharge and identify subclinical fluid overload in ambulatory patients might in principle support closer monitoring in high-risk populations, although any therapeutic application remains hypothesis-generating in the absence of interventional data.

In summary, BIA offers a physiologically grounded way of assessing fluid status and systemic vulnerability in HF. The reported associations with outcomes are broadly concordant in direction and of potentially relevant magnitude, although they derive exclusively from observational data. The challenge now is to determine how this information can be translated into better clinical decisions.

Three constructs should finally be distinguished, since they are not equivalent: the quantification of congestion, the prognostic association between an impedance-derived parameter and a clinical outcome, and the use of such a parameter to guide treatment. The evidence reviewed here addresses the first two. It does not address the third, as no interventional study has tested whether management guided by bioimpedance improves outcomes.

4.1. Limitations

All of the included studies were observational, so the reported associations cannot be interpreted as causal and residual confounding cannot be excluded even where multivariable adjustment was performed. Two studies did not report a multivariable-adjusted estimate for the impedance parameters themselves [14], [15], and in one of these the association lost statistical significance after adjustment for age and BNP [15].

The review protocol was not prospectively registered in PROSPERO or an equivalent registry. This offers no formal protection against post hoc modification of the eligibility criteria, outcomes or analysis plan, and the possibility of selective reporting at review level therefore cannot be excluded.

Formal assessment of small-study effects was not undertaken. Funnel-plot asymmetry and Egger's test are unreliable with fewer than ten comparable studies, and the included studies reported heterogeneous effect measures (adjusted hazard ratios, odds ratios, discrimination statistics and correlation coefficients). Publication bias and selective outcome reporting are nevertheless likely to have shaped this evidence base: small single-centre cohorts reporting positive associations are more readily published than comparable studies reporting null findings, and several included studies reported multiple candidate parameters without a pre-specified primary analysis. The apparent concordance of direction across studies should be read in that light.

The search was restricted to studies published in English, which introduces the possibility of language bias. Neither grey literature nor trial registries were searched. The search was closed on 30 November 2025 and was not subsequently updated; studies published after that date, including cohorts appearing during the editorial process, are therefore not represented in this synthesis, which reflects the evidence available up to that date.

Quantitative pooling was not appropriate. The sources of heterogeneity included the BIA devices and measurement frequencies used; the parameters and indices reported (ECW, ECW/TBW, edema index, impedance vectors and phase angle); the absence of standardized cut-offs; differences in HF phenotype and clinical setting; the timing of measurement, from admission to discharge to ambulatory assessment; and heterogeneous outcome definitions, follow-up durations and effect measures. In addition, the Newcastle-Ottawa Scale could not be applied to four studies whose designs did not involve an exposure-outcome relationship assessed over a defined follow-up period, so the risk of bias for approximately one third of the included studies is described narratively rather than scored.

5. Conclusion

BIA-derived markers of extracellular fluid expansion and cellular integrity are associated with adverse clinical outcomes in both acute and chronic HF. Across the available evidence [14], [15], [16], [17], [18], [19], [20], [21], [22], [23], [24], [25], [26], elevated ECW indices, abnormal impedance vectors and reduced phase angle were associated with an approximately twofold increase in the risk of mortality and worsening HF in those analyses that reported a multivariable-adjusted estimate; two studies did not provide such an estimate for the impedance parameters themselves [14], [15].

In AHF, objective quantification of residual congestion at discharge identifies patients at increased risk of early readmission and death. In chronic ambulatory populations, BIA detects subclinical volume overload preceding overt decompensation. Phase angle further provides complementary prognostic information, reflecting systemic vulnerability beyond congestion alone.

Because the current evidence is exclusively observational, the direction and magnitude of the reported associations can suggest, but not establish, the clinical relevance of impedance-derived parameters. Standardization of methodology and randomized trials evaluating BIA-guided management strategies are now warranted to determine whether integrating fluid compartment assessment into routine care improves outcomes in HF.

CRediT authorship contribution statement

Raquel Rodil-Fraile: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Conceptualization. Andres Blanco-Di Matteo: Writing – review & editing. Julio Oteiza-Olaso: Writing – review & editing. Francisco Javier Anniccherico-Sánchez: Writing – review & editing. Diego Aguiar-Cano: Writing – review & editing. Maria Ruiz-Castellano: Writing – review & editing.

Ethical statement

This article is a systematic review of previously published studies and did not involve any new study of human or animal subjects performed by any of the authors. Ethical approval and informed consent were therefore not required. The authors confirm that the manuscript is original, has not been published previously and is not under consideration for publication elsewhere.

Declaration of Generative AI and AI-assisted technologies in the writing process

During the preparation of this work the authors used ChatGPT (OpenAI) and Claude (Anthropic) solely for the preparation of figures. After using these tools, the authors reviewed and edited the figures as needed and take full responsibility for the content of the publication. No generative artificial intelligence was used to write or edit the text of the manuscript, or to produce, analyse or interpret any scientific data.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary Appendix S1

Full reproducible search strategy, with the complete Boolean string for each database, the exact dates searched and the limits applied.

Supplementary Appendix S2

Completed PRISMA 2020 checklist.

Supplementary Table S1

Newcastle-Ottawa Scale (NOS) quality assessment of the included studies.

Data availability

All data analysed in this review were extracted from the published articles cited in the reference list. No new data were generated. The data extraction table and the risk-of-bias assessments are available in Table 1 and Supplementary Table S1, and further details are available from the corresponding author on reasonable request.

All data analysed in this review were extracted from the published articles cited in the reference list. No new data were generated. The data extraction table and the risk-of-bias assessments are available in Table 1 and Supplementary Table S1, and further details are available from the corresponding author on reasonable request.