Classification of hydration in clinical conditions: indirect and direct approaches using bioimpedance

nutrients Review Classification of Hydration in Clinical Conditions: Indirect and Direct Approaches Using Bioimpedance Henry C. Lukaski 1, Nicanor Vega Diaz 2, Antonio Talluri 3and Lexa Nescolarde 4,* 1Department of Kinesiology and Public Health Education, University of North Dakota, Grand Forks, ND 58202-7166, USA; henry[email protected] 2Nephrology Service, University Hospital of Grand Canary and Faculty of Science, University Los Palmas, 35019 Los Palmas, Grand Canary, Spain; [email protected] 3Antonio Talluri, Fatbyte, Inc., 50012 Bagno a Ripoli, Florence, Italy; [email protected] 4Department of Electronic Engineering, Universitat Politècnica de Catalunya, 08034 Barcelona, Spain *Correspondence: [email protected]; Tel.: +34-93-4137286 Received: 5 March 2019; Accepted: 8 April 2019; Published: 10 April 2019   Abstract: Although the need to assess hydration is well recognized, laboratory tests and clinical impressions are impractical and lack sensitivity, respectively, to be clinically meaningful. Different approaches use bioelectrical impedance measurements to overcome some of these limitations and aid in the classification of hydration status. One indirect approach utilizes single or multiple frequency bioimpedance in regression equations and theoretical models, respectively, with anthropometric measurements to predict fluid volumes (bioelectrical impedance spectroscopy—BIS) and estimate fluid overload based on the deviation of calculated to reference extracellular fluid volume. Alternatively, bioimpedance vector analysis (BIVA) uses direct phase-sensitive measurements of resistance and reactance, measured at 50 kHz, normalized for standing height, then plotted on a bivariate graph, resulting in a vector with length related to fluid content, and direction with phase angle that indexes hydration status. Comparison with healthy population norms enables BIVA to classify (normal, under-, and over-) and rank (change relative to pre-treatment) hydration independent of body weight. Each approach has wide-ranging uses in evaluation and management of clinical groups with over-hydration with an evolving emphasis on prognosis. This review discusses the advantages and limitations of BIS and BIVA for hydration assessment with comments on future applications. Keywords: fluid overload; resistance; reactance; bioelectrical impedance vector analysis; bioelectrical impedance spectroscopy; malnutrition 1. Introduction Routine assessment of hydration status to support patient care persists as a challenge. The preponderance of methods suffers from impracticality and insensitivity [ 1 , 2 ]. Historically, hydration assessment was synonymous with the use of tracer dilution methods to estimate total body water (TBW), extracellular water (ECW), and plasma volume (PV) with the calculation of intracellular water (ICW, ICW = TBW − ECW). This approach is invasive, time-consuming, costly, and prone to ambiguous interpretations of hydration status because of reliance on the assumption of constant TBW to body weight (% TBW), which is prejudiced by inter-individual differences in body composition (adipose, muscle mass, or cachexia) , or TBW to fat-free mass (FFM) (% TBW/FFM) that requires an independent determination of FFM [ 3 – 6 ]. Thus, the measurement of fluid volumes to assess human hydration status has limited value. Laboratory tests of plasma or serum constituents (osmolality, sodium, natriuretic peptide, and other promising proteins), urine characteristics (osmolality, specific gravity, conductivity, and 24-h Nutrients 2019,11, 809; doi:10.3390/nu11040809 www.mdpi.com/journal/nutrients Nutrients 2019,11, 809 2 of 22 output), and saliva and tear composition (osmolality) have inadequate feasibility and lack sensitivity in a clinical setting [ 2 ]. Imaging techniques (radiography and ultrasound) and invasive procedures (catherization) are costly and fail to identify early disturbances in hydration. Simple assessments such as self-reported thirst, patient history, and physical examination are subjective and lack sensitivity [ 2 , 7 ]. Thus, the need persists for a method that is non-invasive, valid (accurate and sensitive), practical, cost-effective, and provides real-time discrimination within the spectrum of the hydration status. Classification of hydration as normal, less, or greater than normal is a practical and valid approach that can overcome these limitations. Burgeoning evidence supports bioelectrical impedance as a unique and promising method to classify the hydration status of an individual. Bioimpedance (BI) is the general term describing the safe, non-invasive measurement of the passive electrical characteristics of an organism after introduction of a painless, low-level alternating current into the body [ 8 – 10 ]. Bioimpedance analysis (BIA), however, augments BI measurements with theoretical biophysical models and statistical (regression) equations contributing to “black box” predictions of fluid volumes that can be unreliable and imprecise. Bioimpedance per se offers direct, uncomplicated measurements that facilitate a practical and valid approach to monitor hydration. This review emphasizes the use of BI to classify hydration. It describes the biophysical basis of the BI method, summarizes different technical approaches and their limitations to estimate fluid volumes, discusses advantages of BI measurements compared to calculated fluid volumes, highlights clinical applications of BI emphasizing its emerging role in prognosis, and provides recommendations to improve and expand the use of BI in clinical applications. 2. Bioelectrical Impedance 2.1. Bioimpedance Basics The physical basis of the BI method is the awareness that the human body is a network of resistors and capacitors [ 8 ]. Physiological fluids behave as resistors and cell membranes act as capacitors (Figure 1). Thus, the body may be represented as a parallel resistor-capacitor (RC) equivalent circuit (Figure 2) in which the introduced alternating current divides into resistive (fluid and electrolytes) and capacitive (cell membranes and tissue interfaces) pathways. The lower specific resistivity of body fluids and tissues containing water and electrolytes compared to intact lipid-laden cell membranes [ 11 ] enables frequency-dependent impedance measurements [ 12 ]. A low-level alternating current continuously passes predominantly through the resistive component but concurrently delayed (temporarily stored) by capacitive elements in tissues and tissue interfaces. Importantly, a variable amount of very low-frequency current, regardless at which frequency the current is introduced, can penetrate the membranes of muscle cells, particularly when the current is parallel to the muscle fiber [ 13 ]. Cell membranes are capacitive elements surrounding the intracellular fluid (ICF) in a series circuit that exists in parallel with the water-containing interstitial gel or extracellular fluid (ECF). Any alternating current will penetrate the capacitive component (reactance, Xc) of the cell membrane in proportion to the frequency of the applied current. Reactance is inversely related to frequency (f) and capacitance (C, Farads): Xc (ohm, Ω ) = 1/[(2 π• f • C)]. Capacitance is the ability of a system or circuit to store an electrical charge. Nutrients 2019,11, 809 3 of 22 Nutrients 2019, 11, x FOR PEER REVIEW 3 of 21 Figure 1. Illustration of the body as a network of resistors and capacitors in a parallel configuration. The alternating current usually exceeds 1 kHz and typically is 50 kHz. C M is membrane capacitance and R e and R i is extracellular and intracellular resistance, respectively. Figure 2. Representation of the body as a parallel resistor-capacitor (RC) equivalent circuit. Delay of the current penetration at the cell membrane causes an out-phasing of current. At low frequencies (e.g., 1–5 kHz), alternating current flows largely in the ECF but not in a direct or fixed proportion relative to the ICF. Similarly, at 50 kHz alternating current does not distribute in proportion to fluid distribution (extrato intra-cellular fluid volume) but relative to capacitive elements. Thus, the phase angle (PhA) is frequency-dependent, principally due to the amount of Xc, and is an index of the amount of applied current that penetrates the capacitive element or cell membranes. Current that is restricted or delayed at (intact) cell membranes becomes out-phased from the voltage drop that occurs at the cell. Impedance (Z) is the broad term describing the opposition to the flow of alternating current by any biological conductor, and is defined by the type of electrical current introduced to a circuit. When direct current is applied, the total conductor is termed resistance (R) whereas when alternating current is introduced, it is called Z. Direct current passes only through resistive elements but alternating current flows through resistive and capacitive elements. The resistive (R) component of Z is independent of frequency so it has the same measurement value (Ω) when either direct or alternating current is used. The capacitor in a biological circuit serves as an insulator and may be envisioned as an imaginary component of Z when direct current is introduced. However, when alternating current is present, the imaginary component acts as a poor conductor with capacitive Figure 1. Illustration of the body as a network of resistors and capacitors in a parallel configuration. The alternating current usually exceeds 1 kHz and typically is 50 kHz. C M is membrane capacitance and Reand Riis extracellular and intracellular resistance, respectively. Nutrients 2019, 11, x FOR PEER REVIEW 3 of 21 Figure 1. Illustration of the body as a network of resistors and capacitors in a parallel configuration. The alternating current usually exceeds 1 kHz and typically is 50 kHz. C M is membrane capacitance and R e and R i is extracellular and intracellular resistance, respectively. Figure 2. Representation of the body as a parallel resistor-capacitor (RC) equivalent circuit. Delay of the current penetration at the cell membrane causes an out-phasing of current. At low frequencies (e.g., 1–5 kHz), alternating current flows largely in the ECF but not in a direct or fixed proportion relative to the ICF. Similarly, at 50 kHz alternating current does not distribute in proportion to fluid distribution (extrato intra-cellular fluid volume) but relative to capacitive elements. Thus, the phase angle (PhA) is frequency-dependent, principally due to the amount of Xc, and is an index of the amount of applied current that penetrates the capacitive element or cell membranes. Current that is restricted or delayed at (intact) cell membranes becomes out-phased from the voltage drop that occurs at the cell. Impedance (Z) is the broad term describing the opposition to the flow of alternating current by any biological conductor, and is defined by the type of electrical current introduced to a circuit. When direct current is applied, the total conductor is termed resistance (R) whereas when alternating current is introduced, it is called Z. Direct current passes only through resistive elements but alternating current flows through resistive and capacitive elements. The resistive (R) component of Z is independent of frequency so it has the same measurement value (Ω) when either direct or alternating current is used. The capacitor in a biological circuit serves as an insulator and may be envisioned as an imaginary component of Z when direct current is introduced. However, when alternating current is present, the imaginary component acts as a poor conductor with capacitive Figure 2. Representation of the body as a parallel resistor-capacitor (RC) equivalent circuit. Delay of the current penetration at the cell membrane causes an out-phasing of current. At low frequencies (e.g., 1–5 kHz), alternating current flows largely in the ECF but not in a direct or fixed proportion relative to the ICF. Similarly, at 50 kHz alternating current does not distribute in proportion to fluid distribution (extrato intra-cellular fluid volume) but relative to capacitive elements. Thus, the phase angle (PhA) is frequency-dependent, principally due to the amount of Xc, and is an index of the amount of applied current that penetrates the capacitive element or cell membranes. Current that is restricted or delayed at (intact) cell membranes becomes out-phased from the voltage drop that occurs at the cell. Impedance (Z) is the broad term describing the opposition to the flow of alternating current by any biological conductor, and is defined by the type of electrical current introduced to a circuit. When direct current is applied, the total conductor is termed resistance (R) whereas when alternating current is introduced, it is called Z. Direct current passes only through resistive elements but alternating current flows through resistive and capacitive elements. The resistive (R) component of Z is independent of frequency so it has the same measurement value ( Ω ) when either direct or alternating current is used. The capacitor in a biological circuit serves as an insulator and may be envisioned as an imaginary component of Z when direct current is introduced. However, when alternating current is present, the imaginary component acts as a poor conductor with capacitive resistance and is termed reactance Nutrients 2019,11, 809 4 of 22 (Xc), which is frequency-dependent. Thus, Z is a complex number, Z 2 = R 2 + Xc 2 , and characterizes the specific fluid and cellular components of an organism. Impedance is a vector with length and position on the bivariate RXc plot (Figure 3). The PhA describes the position of Z and is the angular transformation of Xc to R [arc tangent (Xc/R) •(180◦/π)] expressed in radian degrees. Nutrients 2019, 11, x FOR PEER REVIEW 4 of 21 resistance and is termed reactance (Xc), which is frequency-dependent. Thus, Z is a complex number, Z 2 = R 2 + Xc 2 , and characterizes the specific fluid and cellular components of an organism. Impedance is a vector with length and position on the bivariate RXc plot (Figure 3). The PhA describes the position of Z and is the angular transformation of Xc to R [arc tangent (Xc/R) • (180°/𝜋)] expressed in radian degrees. Figure 3. Geometric relationships among the resistance, reactance (capacitance, C M ), impedance, and phase angle. 2.2. Measurement of Bioelectrical Impedance Assessment of whole-body hydration uses surface, gel (silver-silver chloride) or modern hydrogel electrodes, and tetrapolar electrode placements. Whole-body BI measurements employ traditional limb placements with paired current-introducing (source) and voltage-drop-sensing (detector) electrodes placed on the distal wrist and ankle [8,10]. 3. Volume Quantification in Hydration Assessment: Limitations and Imprecision A primary factor for reliable hydration assessment is the technical validity of the BI measurement. It includes the accuracy of the instrument, which is determined with a precision (<1%) circuit of a resistor and a capacitor in a parallel circuit, and precision or reproducibility that requires repeated measurements of a validation circuit and repetitive in vivo measurements. The technical validity should be <2% for all BI measurements [9]. Similarly, the validity of the reference tracer dilution methods must be assessed with technical accuracy and precision determined to be <2%. Inter-individual biological sources of error for the tracer dilution methods, including hydrogen exchange and anion (bromine-chlorine) shifts that are >4% [14,15], directly affect the variability of fluid volume estimates. 3.1. Single-Frequency Bioimpedance Clinical investigators applied different BI approaches to estimate fluid volumes as indicators of hydration. They mostly used phase-sensitive single-frequency (50 kHz) BI because of the high signalto-noise ratio [9,16] or multiple-frequency (5 kHz to 1 MHz) measurements coupled with either multiple regression equations or theoretical biophysical models to predict fluid volumes [10,17–19]. Each of these approaches has notable concerns that limit their general application in the estimation of fluid volumes. The electrical volume model adapted from Ohm’s Law is the physical basis for estimation of fluid volumes with singleand multiple-frequency BI measurements [10]. The volume (V) of a conductor (fluid plus electrolytes) depends on the length (L) and cross-sectional area (A) of a cylindrical conductor with constant geometry and composition (specific resistivity or ρ), so that V = ρ(L 2 /R). Standing height (Ht) is a biological surrogate for L, so V = ρ(Ht 2 /R). The assumptions of this model, which include the body is a uniform cylinder of constant chemical composition (e.g., water Figure 3. Geometric relationships among the resistance, reactance (capacitance, CM), impedance, and phase angle. 2.2. Measurement of Bioelectrical Impedance Assessment of whole-body hydration uses surface, gel (silver-silver chloride) or modern hydrogel electrodes, and tetrapolar electrode placements. Whole-body BI measurements employ traditional limb placements with paired current-introducing (source) and voltage-drop-sensing (detector) electrodes placed on the distal wrist and ankle [8,10]. 3. Volume Quantification in Hydration Assessment: Limitations and Imprecision A primary factor for reliable hydration assessment is the technical validity of the BI measurement. It includes the accuracy of the instrument, which is determined with a precision (<1%) circuit of a resistor and a capacitor in a parallel circuit, and precision or reproducibility that requires repeated measurements of a validation circuit and repetitive in vivo measurements. The technical validity should be <2% for all BI measurements [ 9 ]. Similarly, the validity of the reference tracer dilution methods must be assessed with technical accuracy and precision determined to be <2%. Inter-individual biological sources of error for the tracer dilution methods, including hydrogen exchange and anion (bromine-chlorine) shifts that are >4% [ 14 , 15 ], directly affect the variability of fluid volume estimates. 3.1. Single-Frequency Bioimpedance Clinical investigators applied different BI approaches to estimate fluid volumes as indicators of hydration. They mostly used phase-sensitive single-frequency (50 kHz) BI because of the high signal-to-noise ratio [ 9 , 16 ] or multiple-frequency (5 kHz to 1 MHz) measurements coupled with either multiple regression equations or theoretical biophysical models to predict fluid volumes [ 10 , 17 – 19 ]. Each of these approaches has notable concerns that limit their general application in the estimation of fluid volumes. The electrical volume model adapted from Ohm’s Law is the physical basis for estimation of fluid volumes with singleand multiple-frequency BI measurements [ 10 ]. The volume (V) of a conductor (fluid plus electrolytes) depends on the length (L) and cross-sectional area (A) of a cylindrical conductor with constant geometry and composition (specific resistivity or ρ ), so that V = ρ(L2/R) . Standing height (Ht) is a biological surrogate for L, so V = ρ (Ht 2 /R). The assumptions of this model, which include the body is a uniform cylinder of constant chemical composition (e.g., water content) of the conductor, are open to criticism [ 18 ]. Specifically, the body consists of five cylinders Nutrients 2019,11, 809 5 of 22 (two arms, two legs, and the thorax) with different geometry (varying length and diameter) and different chemical composition (specific resistivities of different tissues) including a disputed constant hydration of the fat-free body (73%). Researchers exploited this bioelectrical volume model, together with various anthropometric information (Ht, weight (Wt), age, and gender), to develop and validate multiple regression prediction equations using singleand multiple-frequency measurements for TBW estimates in healthy and ill adults [18]. A technical issue exists with any application of these regression models. The independent variable is R and not Z. Thus, only phase-sensitive BI instruments should be used in the application of specific single-frequency prediction models using R as an independent variable. Although Z and R are highly correlated (R 2 > 0.9), the magnitude of Z is greater than R (1% to 2%) because Z includes the value of Xc, which is not negligible (~10% of R) and includes biologically important information for hydration classification [ 9 , 10 ]. Similarly, predictions based on multi-frequency BI measurements should ensure that appropriate devices are used to obtain the required Z or R measurements. Critical evaluation of a population-based model to predict TBW highlights a practical limitation to the use of BIA to estimate the fluid volume. Sun et al. [ 20 ] derived and validated TBW prediction equations in a large sample of healthy adults of diverse ethnicity: Males: TBW (L) = 0.87 + 0.43 (Ht2/R) + 0.20 Wt Females: TBW (L) = 3.27 + 0.45 (Ht2/R) + 0.12 Wt Cross-validation of these equations revealed a large predictive error (standard error of estimate (SEE)) of 3.8 L and 2.6 L, for the males and females, respectively. The magnitude of these prediction errors is too large (e.g., imprecise) to estimate TBW or to identify change in TBW of an individual and, therefore, limits these equations to observational or epidemiological studies. 3.2. Bioelectrical Impedance Spectroscopy (BIS) Some BI instruments measure Z over a wide range of frequencies (e.g., 5 kHz up to 1 MHz), report various measurements and then use programmed software to estimate fluid volumes [ 17 ]. Certain non-phase-sensitive devices only measure Z and provide the ratio of Z at low and high frequencies (Z low /Z high ) that is equivalent to PhA at 50 kHz. Similarly, they also use the Z low /Z high to compute the phase and cannot perform valid calculations to form the impedance modulus for the Cole plot. The BIS method relies on the assumption that low-frequency current flows through the ECF and high-frequency current penetrates ECF and ICF. These assumptions are open to criticism largely because they were derived from in vitro studies of cells suspended in fluid and ignore the cell-cell interfaces that occur in tissues [ 10 ]. Additionally, investigators agree that at the lowest frequencies some current penetrates cell membranes (e.g., extracellular and intracellular paths) and that at the highest frequencies current does cross all cells (e.g., extracellular and intracellular paths) [11,21–23]. The BIS method uses non-linear least-square (polynomial) modeling of limited phase-sensitive multi-frequency measurements of Z and Xc, calculated from PhA, then extrapolates these values to generate a Cole plot (Figure 4). This mathematical model [ 24 ] calculates theoretical resistivity values that are used to approximate fluid volumes. The extrapolated (modeled) variables are R 0 , also termed as Re (the resistance of the extracellular fluid or interstitial resistivity), and R ∞ (resistance associated with the sum of extracellular and intracellular fluids), critical frequency (f C ) or the frequency at which Xc is maximal, and membrane capacitance (Cm). The value of calculated Cm depends on another fitted variable, time delay (Td), which represents the effect of cell membrane and orientation on the current transmission in the body. These resistance parameters are used to calculate intracellular resistance (Ri) as 1/Ri = 1/R∞−1/Re. Mixture theory uses these calculated resistance parameters with anthropometric measurements and some wide-ranging assumptions to estimate fluid volumes [ 17 , 25 , 26 ]. It relies on a theoretical model that estimates apparent conductivity in a heterogeneous entity composed of conductive (water and ions) and non-conductive (anhydrous materials such as bone, fat, and cell membranes) components such as the body. Fluid volumes, ECW, and ICW are predicted separately and summed to obtain Nutrients 2019,11, 809 6 of 22 the TBW. Calculation of the ECW and ICW depends on the estimated Cole resistance parameters, body scalar factors calculated from the Ht and Wt of an individual, an assumed body density value derived from body mass index (BMI), empirically derived gender-specific resistivity values, and other assumptions. Extracellular fluid volume (V ECF ) is calculated as V ECF = k ECF • (Ht 2• Wt 0.5 /Re) 2/3 with k ECF = 10 -3 • (K B2•ρECF2 /D b ) 1/3 , where K B is a body geometry factor that relates relative volume of the legs, arms, and trunk, ρECF is the resistivity of the extracellular fluid, and D b is total body density. Intracellular fluid volume (V ICF ) also is calculated from the derived Cole resistance variables and gender-dependent resistivities of the ECW and ICW. These calculations are performed with software provided by the manufacturers of the different BIS instruments, and are subject to change. Nutrients 2019, 11, x FOR PEER REVIEW 6 of 21 to obtain the TBW. Calculation of the ECW and ICW depends on the estimated Cole resistance parameters, body scalar factors calculated from the Ht and Wt of an individual, an assumed body density value derived from body mass index (BMI), empirically derived gender-specific resistivity values, and other assumptions. Extracellular fluid volume (V ECF ) is calculated as V ECF = k ECF • (Ht 2 • Wt 0.5 /Re) 2/3 with k ECF = 10 -3 • (K B2 • ρ ECF2 /D b ) 1/3 , where K B is a body geometry factor that relates relative volume of the legs, arms, and trunk, ρ ECF is the resistivity of the extracellular fluid, and D b is total body density. Intracellular fluid volume (V ICF ) also is calculated from the derived Cole resistance variables and gender-dependent resistivities of the ECW and ICW. These calculations are performed with software provided by the manufacturers of the different BIS instruments, and are subject to change. Figure 4. Plot of reactance and resistance of a healthy male obtained with a Xitron 4200 and derived using non-linear curve-fitting software based on the Cole model. Note that the majority of values (dashed lines) were estimated. R 0 and R ∞ were calculated and they approximate resistance at 0 and the highest frequency, respectively. Uncertain basic assumptions and constants of the original BIS approach [26–29] resulted in errors in the predictions of fluid volumes. Specifically, significant errors in the estimation of TBW (2 L) and ECW (~1 L) in individuals with increasing adiposity, characterized using BMI, were explained as the effect of increasing adipose tissue on the assumed resistivity of the ECW [30–32]. A proposed remedy to this limitation was the use of BMI as a proxy for adiposity [33]. A multi-center validation trial [33] of this revised BIS approach included healthy adults and dialysis patients and found that this change improved the sensitivity of the revised ECW predictions by decreasing the variability of the ECW estimate by 24% (0.6 L) in healthy and dialysis patients and reducing the variability of the TBW prediction in adults with BMI values of less than 20 and greater than 30 kg/m 2 . Although no significant differences between tracer dilution reference measurements and BIS predictions of the TBW, ECW, and ICW were found, there were wide limits of agreement (95% confidence intervals—95% CI). Specifically, the mean differences were small with wide limits of agreement for prediction of the TBW (−0.2 L (95% CI: 4.6 L)), ECW (−0.4 L; 95% CI: 2.9 L)), and ICW (0.2 L (95% CI: 4.1 L)). A limitation of the use of BMI as a surrogate for body fatness is the insensitivity of BMI to reliably differentiate the body composition (body fat and muscle mass) of an individual, healthy or ill [34]. These wide limits of agreement contribute to poor sensitivity of the ECW and TBW estimates estimated with BIS for individuals. 3.3. Comparison of Fluid Volumes Estimated with Single-Frequency Bioimpedance and Bioelectrical Impedance Spectroscopy Although a common opinion is that BIS, compared to a single-frequency BIA, provides more accurate estimates of fluid volume [23], experimental data fail to confirm this supposition. Impedance predictions of the TBW and ICW were equally accurate in comparison to isotope dilution estimates Figure 4. Plot of reactance and resistance of a healthy male obtained with a Xitron 4200 and derived using non-linear curve-fitting software based on the Cole model. Note that the majority of values (dashed lines) were estimated. R 0 and R ∞ were calculated and they approximate resistance at 0 and the highest frequency, respectively. Uncertain basic assumptions and constants of the original BIS approach [ 26 – 29 ] resulted in errors in the predictions of fluid volumes. Specifically, significant errors in the estimation of TBW (2 L) and ECW (~1 L) in individuals with increasing adiposity, characterized using BMI, were explained as the effect of increasing adipose tissue on the assumed resistivity of the ECW [ 30 – 32 ]. A proposed remedy to this limitation was the use of BMI as a proxy for adiposity [33]. A multi-center validation trial [ 33 ] of this revised BIS approach included healthy adults and dialysis patients and found that this change improved the sensitivity of the revised ECW predictions by decreasing the variability of the ECW estimate by 24% (0.6 L) in healthy and dialysis patients and reducing the variability of the TBW prediction in adults with BMI values of less than 20 and greater than 30 kg/m 2 . Although no significant differences between tracer dilution reference measurements and BIS predictions of the TBW, ECW, and ICW were found, there were wide limits of agreement (95% confidence intervals—95% CI). Specifically, the mean differences were small with wide limits of agreement for prediction of the TBW ( − 0.2 L (95% CI: 4.6 L)), ECW ( − 0.4 L; 95% CI: 2.9 L)), and ICW (0.2 L (95% CI: 4.1 L)). A limitation of the use of BMI as a surrogate for body fatness is the insensitivity of BMI to reliably differentiate the body composition (body fat and muscle mass) of an individual, healthy or ill [ 34 ]. These wide limits of agreement contribute to poor sensitivity of the ECW and TBW estimates estimated with BIS for individuals. 3.3. Comparison of Fluid Volumes Estimated with Single-Frequency Bioimpedance and Bioelectrical Impedance Spectroscopy Although a common opinion is that BIS, compared to a single-frequency BIA, provides more accurate estimates of fluid volume [ 23 ], experimental data fail to confirm this supposition. Impedance predictions of the TBW and ICW were equally accurate in comparison to isotope dilution estimates but Nutrients 2019,11, 809 7 of 22 included a significant bias (error) in single-frequency and BIS prediction of the ECW in steady-state hemodialysis patients [ 35 ]. Both impedance methods had a proportional error in estimation of the ICW largely due to assumptions of the BIS model and the single frequency equation used to predict the ICW with calculations from total body potassium, a specific marker for body cell mass. Similar findings came from another study of hemodialysis patients [ 36 ]. A consistent observation is the wide limits of agreement between the impedance and reference methods that cautions against the use of these methods for individual assessment of fluid volumes [37]. Other research supports these findings. In adults receiving growth hormone replacement therapy, the TBW was determined using isotope dilution and predicted from BI measurements, collected with the same multi-frequency instrument, using 50 kHz values and four published equations and the Hanai mixture [ 25 ] model with and without adjustment for BMI [ 38 ]. Measured and predicted TBW values were significantly and similarly correlated for linearity and concordance (r > 0.9). Comparisons of measured and predicted TBW values revealed limits of agreement with slightly greater variability (0.9 L to 2.7 L or 4% to 7%) for the 50 kHz predictions compared to variability estimates (0.6 L to 0.9 L or 1% to 2%) from the BIS predictions. Thus, both the single frequency prediction equations and BIS models performed equally well in predicting the TBW at the population level. Follow up analyses determined the comparability of these different approaches to estimate the TBW at the level of an individual [ 38 ]. The mean absolute error of BIS was slightly improved compared to multiple regression equations using 50 kHz values (5% vs. 7% or 2 L vs. 3 L). These findings indicate that the precision of these BIA approaches may be appropriate for group comparisons but is too large for monitoring changes in the TBW in clinical conditions. Any comparison of BIA-predicted fluid volumes with tracer dilution reference measurements suffers from technical and biological limitations. Sources of error include the validity (accuracy and precision) of the impedance measurement, error of prediction from the regression equation, intrinsic error of the reference method (technical accuracy and reproducibility, biological errors in assumptions of the dilution method), electrical-volume errors (anisotropy and body geometry), and biological variability (inter-individual differences in the diameter of body segments, limb lengths, and variation in body fatness) [ 37 ]. Use of the Cole model depends on the reliability of the fitting of the impedance data. Review of recent BIS publications reveals that individual subjects are removed from analyses because of failure to achieve acceptable tolerances (>99% reproducibility) of the repeated time delay estimates derived from curve fitting [ 35 , 36 ]. This observation indicates significant imprecision in the calculation of the current delay at membranes. Finally, use of different single-frequency mathematical prediction equations complicates comparisons. As suggested by Seone et al. [ 38 , 39 ], volume prediction models that include more independent variables than Ht 2 /R and weight tend to produce greater errors in estimated TBW that indicates a specificity of that regression equation and, thus, excludes its application in any sample or individual differing in characteristics from the original sample in which the regression model was developed. These factors as well as the imprecision of the prediction models, single-frequency and BIS, impedes their use in point-of-care individual assessments of hydration. 4. Classification of Hydration with Bioelectrical Impedance Vector Analysis (BIVA) In contrast to the use of mathematical modeling of limited BI measurements, theoretical and regression prediction models, classification of hydration only requires precise BI measurements and reference values [ 40 ]. A 50 kHz phase-sensitive BI instrument directly measures R and Xc, which are normalized for standing height to achieve a standard resistivity and plotted on the RXc graph (Figure 5). The vector (Z) has a length that is inversely related to TBW and its position, described as the PhA, and provides information on tissue hydration. Nutrients 2019,11, 809 8 of 22 Nutrients 2019, 11, x FOR PEER REVIEW 8 of 21 Figure 5. Resistance-reactance (RXc) plot with tolerance ellipses from healthy Caucasian males. Bioelectrical impedance vector analysis (BIVA) enables classification (under-, normal, and overhydration) and ranking of change in hydration (more or less than before treatment), as well as classification (more or less) of soft tissue mass by comparing vector position of an individual or a group to a healthy ethnicity-, age-, and gender-matched population [40]. Implementation of BIVA follows the work of Moore and Boyden [3] in which hydration is a function of fluid distribution in the extracellular and intracellular spaces, notably BCM. Clinically, over-hydration is expressed as an expansion of the extracellular fluid (interstitial fluid plus plasma), also termed fluid overload, relative to the BCM (intracellular fluid); patients with renal disease and heart failure are characterized as over-hydrated. Malnutrition, including cachexia, also is associated with an increase in ECW and a decrease in BCM. Hypo-hydration is seen with a decrease in ECW and little or no change in BCM. It can be acute as with excess fluid removal therapy or short-term with diseases causing excessive fluid loss (diarrhea or emesis). Compared to single-frequency BIA and BIS predictions of fluid volumes, BIVA only has minimal error associated with BI measurement and reproducibility (1–2%) whereas fluid volume predictions include additional sources of error including regression error of the prediction equation (~10%), technical error in the reference method (~4%), limitations of the bioelectrical volume model (i.e., anisotropy of tissues and geometry), and biological variability (i.e., inter-individual body composition differences) that propagate. Therefore, for an individual, classification and ranking of hydration is more precise and accurate than is the quantification of fluid volume because BIVA is independent of regression equations and theoretical models that are acquired with limited and specific samples and, thus, are not robust in the assessment of hydration outside of the group in which they were developed, and are adversely affected by illness. The RXc graph (Figure 5) is a probability distribution that classifies a vector according to its distance from the mean healthy vector. The variability of the impedance vector is represented in the bivariate normal distribution with elliptical probability areas (50%, 75%, and 95%) in the tolerance or reference ellipses. Vectors displacements parallel to the major axis indicate changes in hydration (more or less fluids). Vectors within the 50% tolerance ellipse are considered to be at normal hydration whereas lengthening of vectors from the 51% to 75% and >76% percentile of the upper range of percentiles indicate moderate and severe dehydration, respectively. Conversely, shortening of vectors from the 51% to 75% and >76% percentile reference ellipses in the lower range indicate increasing fluid overload. Vectors positioned to the left of the major axis reflect increasing cell mass and vectors to the right indicate decreasing cell mass, respectively. Thus, BIVA uses patterns of Figure 5. Resistance-reactance (RXc) plot with tolerance ellipses from healthy Caucasian males. Bioelectrical impedance vector analysis (BIVA) enables classification (under-, normal, and over-hydration) and ranking of change in hydration (more or less than before treatment), as well as classification (more or less) of soft tissue mass by comparing vector position of an individual or a group to a healthy ethnicity-, age-, and gender-matched population [ 40 ]. Implementation of BIVA follows the work of Moore and Boyden [ 3 ] in which hydration is a function of fluid distribution in the extracellular and intracellular spaces, notably BCM. Clinically, over-hydration is expressed as an expansion of the extracellular fluid (interstitial fluid plus plasma), also termed fluid overload, relative to the BCM (intracellular fluid); patients with renal disease and heart failure are characterized as over-hydrated. Malnutrition, including cachexia, also is associated with an increase in ECW and a decrease in BCM. Hypo-hydration is seen with a decrease in ECW and little or no change in BCM. It can be acute as with excess fluid removal therapy or short-term with diseases causing excessive fluid loss (diarrhea or emesis). Compared to single-frequency BIA and BIS predictions of fluid volumes, BIVA only has minimal error associated with BI measurement and reproducibility (1–2%) whereas fluid volume predictions include additional sources of error including regression error of the prediction equation (~10%), technical error in the reference method (~4%), limitations of the bioelectrical volume model (i.e., anisotropy of tissues and geometry), and biological variability (i.e., inter-individual body composition differences) that propagate. Therefore, for an individual, classification and ranking of hydration is more precise and accurate than is the quantification of fluid volume because BIVA is independent of regression equations and theoretical models that are acquired with limited and specific samples and, thus, are not robust in the assessment of hydration outside of the group in which they were developed, and are adversely affected by illness. The RXc graph (Figure 5) is a probability distribution that classifies a vector according to its distance from the mean healthy vector. The variability of the impedance vector is represented in the bivariate normal distribution with elliptical probability areas (50%, 75%, and 95%) in the tolerance or reference ellipses. Vectors displacements parallel to the major axis indicate changes in hydration (more or less fluids). Vectors within the 50% tolerance ellipse are considered to be at normal hydration whereas lengthening of vectors from the 51% to 75% and >76% percentile of the upper range of percentiles indicate moderate and severe dehydration, respectively. Conversely, shortening of vectors from the 51% to 75% and >76% percentile reference ellipses in the lower range indicate increasing fluid overload. Vectors positioned to the left of the major axis reflect increasing cell mass and vectors to the right indicate decreasing cell mass, respectively. Thus, BIVA uses patterns of impedance vector distribution without the need for prediction equations, body weight or reliance on the assumption Nutrients 2019,11, 809 9 of 22 of stable composition (water content) of the FFM. BIVA enables detection and ranking of changes in tissue hydration status of <500 mL in real-time [40]. Appropriate derivation and implementation of the RXc graph to classify hydration requires some important technical considerations. All measurements should be acquired with a phase-sensitive (e.g., phase angle directly measured) BI instrument. Failure to use a phase-sensitive device results in significant errors in R (10 Ω ) and Xc (10–12 Ω ) measurements [ 41 ] contributing to 8–10% repositioning of reference and patient vectors. Additionally, use of high-impedance electrodes can lead to misclassification of hydration [42]. 4.1. Clinical Applications of BIVA A focal application of BIVA is the assessment of the hydration status in patients with fluid overload and evaluation of effects of therapeutic procedures designed to achieve homeostasis without undesirable side effects. Kidney disease and hemodynamic impairments are the principal conditions in which BIVA is used to classify hydration and to monitor changes in response to treatments. Traditional approaches to assess hydration in these patients are unreliable because they rely on body weight to designate excess fluid and are insensitive to fluid overload in the presence of edema [43]. 4.1.1. Hydration Assessment in Hemodialysis (HD) Dialysis, particularly hemodialysis (HD), is amenable to BIVA because of the need for fluid removal and post-dialytic fluid retention. Routine evaluation of hydration includes monitoring of body weight and blood pressure changes that are not reliably determined by fluid volume. Edema is not usually detectable until interstitial fluid volume increases 30% over normal levels (4–5 kg gain in body weight) and severe dehydration can occur before appearance of clinical signs. Thus, traditional indicators of overand under-hydration in patients with renal disease are insensitive and inadequate [40]. 4.1.2. Hydration Changes and Vector Patterns During the HD Cycle The objective of HD is to remove excess fluid without complications for the patient. Estimation of the volume of fluid for removal is subjective and generally related to previous post-dialytic body weight, which may be unreliable [ 43 ]. Fluid removal occurs during HD. It can be symptomatic or uncompensated with episodes of hypotension, malaise, and cramps or asymptomatic (compensated). Fluid overload can occur during the inter-dialytic period and is symptomatic with edema, exacerbated hypertension, and pulmonary congestion. Dry weight is a general term associated with the post-dialytic body weight at which most of the excess fluid has been removed. The optimal “dry weight” is determined clinically and operationally as the weight at which a patient can tolerate HD without adverse intra-dialytic symptoms, notably hypotension. Body fluid volumes, however, change continuously during the inter-dialytic period so that euhydration occurs only for a brief period. Serial BI measurements and examination of vectors before and after HD reveal a classic pattern of fluid removal and repletion during a standard 3-d HD cycle. Fluid removal via ultrafiltration (UF) accompanies vector lengthening parallel to the mean vector of healthy adults [ 44 ]. Within the first 2 post HD, vectors are relatively stable near the post HD vector. During the next 24, 48, and 72 h however, vectors progressively shorten within the 50% to 75% tolerance ellipse and reflect fluid repletion of 1.4, 2.6, and 3.4 L. Thus, BIVA is specific and sensitive to monitor fluid during the wet and dry cycle of HD. 4.1.3. Vector Trajectories: Adequacy of Ultrafiltration (UF) Measurement of the electrical properties of tissues enables the classification of vectors as abnormal and identification of HD patients at risk of intra-dialytic problems. Observational findings indicate that BIVA can identify potential symptomatic HD patients. Before dialysis, HD patients, compared to healthy controls, exhibit evidence of fluid overload with shorter vectors and lesser phase angles [ 44 , 45 ]. Pre-HD vector position discriminated unstable (symptomatic), compared to stable (asymptomatic), Nutrients 2019,11, 809 16 of 22 significantly affected vector position on the RXc graph resulting in significant differences in hydration classification of healthy adults. Standardization of BI measurements remains a challenge for clinical users of BI for hydration classification. The lack of international manufacturing standards, coordination of technology, and cross-calibration of electrical accuracy hinders growth in clinical applications of BI. Concurrence on appropriate measurements conditions in hydration assessment is imminent but clinical investigators need to be aware of elevated core and peripheral temperatures as well as high impact BI measurements. Use of artifact-free contact electrodes is recommended for reliable classification of hydration. Small differences in BI measurements incurred by measurement influences can alter hydration classification in clinical conditions. 8. Summary and Conclusions Bioimpedance measurements are gaining interest to assess and monitor hydration status in dialytic and critically ill patients because they overcome many of the limitations of traditional methods of hydration assessment. Whereas two approaches use of BI measurements each has specific advantages and certain limitations. Use of direct 50 kHz, phase-sensitive BI measurements in the BIVA model for hydration assessment is progressing in a variety of clinical conditions. BIVA illustrates a vector whose position is evaluated relative to healthy reference ranges and is interpreted as normal or abnormal hydration based on distance from the mean healthy vector. Migration of the vector (shortening or lengthening) in response to progression of a physiological process, pathology, or an intervention indicates changes in hydration (gain or loss of fluids). Classification of hydration status (normal or abnormal) avoids insensitivities (>10% variability and imprecision for individual estimations) associated with regression equations and unproven biophysical models, and does not rely on body weight to assess hydration. Use of BIVA improves the prescription of UF in dialysis by monitoring the backward-forward displacement of vectors in relation to the wet-dry cycle of HD. It further enhances decision-making in dialysis by facilitating the interpretation of alterations in blood pressure relative to the hydration status and, thus, adjusting UF. Among critically ill patients, BIVA is significantly and inversely correlated with CVP. On the RXc graph, impedance vectors of patients with low, normal, and high CVP move downward and outside of the 75% tolerance level with increasing CVP, which indicates excess fluid accumulation. Identification of hypohydration in stable CAD patients before angiographic procedures enabled rehydration with appropriate volumes of saline to attenuate risk of CI-AKI. Use of BIVA in assessment of malnutrition, notably in illnesses with muscle loss when hydration is altered, is increasing. Phase-sensitive multi-frequency BIS measurements coupled with biophysical models indirectly estimate ECW that is compared with projections of normal ECW to calculate excess ECW described as FO. The estimates of ECW, adjusted for BMI, are derived from regression equations using gender, height, and body weight as independent variables with wide limits of agreement and, hence imprecision for an individual ECW estimate. Estimates of FO >2.5 L or >15% predicted ECW are associated with increased morbidity in dialysis patients but the prognostic value for mortality is lacking. 9. Future Directions Advancement of the use of BI measurements to aid in fluid management of dialysis patients requires prospective controlled studies with more focus on outcomes [ 79 , 90 – 92 ]. Early efforts were descriptive emphasizing proof-of-principle research. Adequate confirmation of positive findings related to validity BIVA and BIS is available. New research should critically evaluate the benefit of these BI approaches in patient care. Importantly, there is a need for more consistency and scrutiny in research design with appropriate sample sizes to test hypotheses and comparability in reporting of results. Some topics include the ability to predict intraand inter-dialytic complications and mortality Nutrients 2019,11, 809 17 of 22 of HD patients, measure the effects of therapeutic interventions on changes in hydration and fluid status on preservation of renal function, evaluate the specificity and sensitivity of available and putative clinical markers of hydration, determination of changes in body composition as well as nutritional status/intake relative to longevity of dialysis, and the effect of interventions to preserve muscle mass on survival. The goal is to provide findings that can be applied practically to improve patient outcomes. Author Contributions: Each author contributed equally in the preparation of this manuscript. Funding: This research received no external funding. Conflicts of Interest: The authors report no conflicts of interest. A. Talluri was founder and president of Akern Srl, and is no longer involved in this company. References 1. Armstrong, L.E. Assessing hydration status: The elusive gold standard. J. Am. Coll. Nutr. 2007 ,26 (Suppl. 5), 575S–584S. [CrossRef] [PubMed] 2. Di Somma, S.; Navarin, S.; Giordano, S.; Spadini, F.; Lippi, G.; Cervellin, G.; Dieffenbach, B.V.; Maisel, A.S. The emerging role of biomarkers and bio-impedance in evaluating hydration status in patients with acute heart failure. Clin. Chem. Lab. Med. 2012,50, 2093–2105. [CrossRef] 3. Moore, F.D.; Boyden, C.M. Body cell mass and limits of hydration of the fat-free body: Their relation to estimated skeletal weight. Ann. N. Y. Acad. Sci. 1963,110, 62–71. [CrossRef] [PubMed] 4. Snyder, W.S.; Cook, M.J.; Nasset, E.S.; Karhausen, L.R.; Howells, G.P.; Tipton, I.H. Report of the Task Group on Reference Man; Pergamon Press: Oxford, UK, 1975; pp. 27–32. 5. Wang, Z.; Deurenberg, P.; Wang, W.; Pietrobelli, A.; Baumgartner, R.N.; Heymsfield, S.B. Hydration of fat-free body mass: Review and critique of a classic body-composition constant. Am. J. Clin. Nutr. 1999 ,69, 833–841. [CrossRef] [PubMed] 6. Chumlea, W.C.; Schubert, C.M.; Sun, S.S.; Demerath, E.; Towne, B.; Siervogel, R.M. A review of body water status and the effects of age and body fatness in children and adults. J. Nutr. Health Aging 2007 ,11, 111–118. 7. Armstrong, L.E.; Kavouras, S.A.; Walsh, N.P.; Roberts, W.O. Diagnosing dehydration? Blend evidence with clinical observations. Curr. Opin. Clin. Nutr. Metab. Care 2016,19, 434–438. [CrossRef] 8. Lukaski, H.C. Biological indexes considered in the derivation of the bioelectrical impedance analysis. Am. J. Clin. Nutr. 1996,64 (Suppl. 3), 397S–404S. [CrossRef] 9. Kyle, U.G.; Bosaeus, I.; De Lorenzo, A.D.; Deurenberg, P.; Elia, M.; Gómez, J.M.; Heitmann, B.L.; Kent-Smith, L.; Melchior, J.C.; Pirlich, M.; et al. Composition of the ESPEN Working Group. Bioelectrical impedance analysis-part I: Review of principles and methods. Eur. J. Clin. Nutr. 2004,23, 1226–1243. [CrossRef] 10. Lukaski, H.C. Evolution of bioimpedance: A circuitous journey from estimation of physiological function to assessment of body composition and a return to clinical research. Eur. J. Clin. Nutr. 2013 ,67 (Suppl. 1), S2–S9. [CrossRef] [PubMed] 11. Faes, T.J.; van der Meij, H.A.; de Munck, J.C.; Heethaar, R.M. The electric resistivity of human tissues (100 Hz–10 MHz): A meta-analysis of review studies. Physiol. Meas. 1999,20, R1–R10. [CrossRef] 12. Foster, K.R.; Lukaski, H.C. Whole-body impedance—What does it measure? Am. J. Clin. Nutr. 1996 , 64 (Suppl. 3), 388S–396S. [CrossRef] [PubMed] 13. Grimnes, S.; Martinsen, O. Bioimpedance and Bioelectricity Basics, 3rd ed.; Elsevier Ltd.: London, UK, 2015; pp. 1–8. 14. Racette, S.B.; Schoeller, D.A.; Luke, A.H.; Shay, K.; Hnilicka, J.; Kushner, R.F. Relative dilution spaces of 2H-and 18O-labeled water in humans. Am. J. Physiol. 1994,267, E585–E590. [CrossRef] 15. Schoeller, D.A. Hydrometry. In Human Body Composition, 2nd ed.; Heymsfield, S.B., Lohman, T.G., Wang, Z.M., Going, S.B., Eds.; Human Kinetics: Champaign, IL, USA, 2005; pp. 35–50. Nutrients 2019,11, 809 18 of 22 16. Piccoli, A.; Pastori, G.; Guizzo, M.; Rebeschini, M.; Naso, A.; Cascone, C. Equivalence of information from single versus multiple frequency bioimpedance vector analysis in hemodialysis. Kidney Int. 2005 ,67, 301–313. [CrossRef] 17. Matthie, J.R. Bioimpedance measurements of human body composition: Critical analysis and outlook. Expert Rev. Med. Devices 2008,5, 239–261. [CrossRef] [PubMed] 18. Kyle, U.G.; Bosaeus, I.; De Lorenzo, A.D.; Deurenberg, P.; Elia, M.; Gómez, J.M.; Lilienthal Heitmann, B.; Kent-Smith, L.; Melchior, J.C.; Pirlich, M.; et al. Bioelectrical impedance analysis-part II: Utilization in clinical practice. Eur. J. Clin. Nutr. 2004,23, 1430–1453. [CrossRef] [PubMed] 19. Mulasi, U.; Kuchnia, A.J.; Cole, A.J.; Earthman, C.P. Bioimpedance at the bedside: Current applications, limitations, and opportunities. Nutr. Clin. Pract. 2015,30, 180–193. [CrossRef] [PubMed] 20. Sun, S.S.; Chumlea, W.C.; Heymsfield, S.B.; Lukaski, H.C.; Schoeller, D.; Friedl, K.; Kuczmarski, R.J.; Flegal, K.M.; Johnson, C.L.; Hubbard, V.S. Development of bioelectrical impedance analysis prediction equations for body composition with the use of a multicomponent model for use in epidemiological surveys. Am. J. Clin. Nutr. 2003,77, 331–340. [CrossRef] 21. Ellis, K.J.; Wong, W.W. Human hydrometry: Comparison of multifrequency bioelectrical impedance with 2H2O and bromine dilution. J. Appl. Physiol. 1998,85, 1056–1062. [CrossRef] 22. Ellis, K.J.; Bell, S.J.; Chertow, G.M.; Chumlea, W.C.; Knox, T.A.; Kotler, D.P.; Lukaski, H.C.; Schoeller, D.A. Bioelectrical impedance methods in clinical research: A follow-up to the NIH Technology Assessment Conference. Nutrition 1999,15, 874–880. [CrossRef] 23. Ellis, K.J. Human body composition: In vivo methods. Physiol. Rev. 2000 ,80, 649–680. [CrossRef] [PubMed] 24. Cole, K.S. Permeability and impermeability of cell membranes for ions. In Cold Spring Harbor Symposia on Quantitative Biology; Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY, USA, 1940; Volume 8, pp. 110–122. 25. Hanai, T. Electrical properties of emulsions. In Emulsion Science; Sherman, P.H., Ed.; Academic Press: London, UK, 1968; pp. 354–477. 26. De Lorenzo, A.; Andreoli, A.; Matthie, J.; Withers, P. Predicting body cell mass with bioimpedance by using theoretical methods: A technological review. J. Appl. Physiol. 1997,82, 1542–1558. [CrossRef] 27. Bolton, M.P.; Ward, L.C.; Khan, A.; Campbell, I.; Nightingale, P.; Dewit, O.; Elia, M. Sources of error in bioimpedance spectroscopy. Physiol. Meas. 1998,19, 235–246. [CrossRef] 28. Buendia, R.; Seoane, F.; Lindecrantz, K.; Bosaeus, I.; Gil-Pita, R.; Johannsson, G.; Ellegárd, L.; Ward, L.C. Estimation of body fluids with bioimpedance spectroscopy: State of the art methods and proposal of novel methods. Physiol. Meas. 2015,36, 2171–2187. [CrossRef] 29. Ward, L.C.; Isenring, E.; Dyer, J.M.; Kagawa, M.; Essex, T. Resistivity coefficients for body composition analysis using bioimpedance spectroscopy: Effects of body dominance and mixture theory algorithm. Physiol. Meas. 2015,36, 1529–1549. [CrossRef] 30. Cox-Reijven, P.L.; Soeters, P.B. Validation of bio-impedance spectroscopy: Effects of degree of obesity and ways of calculating volumes from measured resistance values. Int. J. Obes. Relat. Metab. Disord. 2000 , 24, 271–280. [CrossRef] 31. Cox-Reijven, P.L.; van Kreel, B.; Soeters, P.B. Accuracy of bioelectrical impedance spectroscopy in measuring changes in body composition during severe weight loss. J. Parenter. Enter. Nutr. 2002 ,26, 120–127. [CrossRef] 32. Mager, J.R.; Sibley, S.D.; Beckman, T.R.; Kellogg, T.A.; Earthman, C.P. Multifrequency bioelectrical impedance analysis and bioimpedance spectroscopy for monitoring fluid and body cell mass changes after gastric bypass surgery. Clin. Nutr. 2008,27, 832–841. [CrossRef] 33. Moissl, U.M.; Wabel, P.; Chamney, P.W.; Bosaeus, I.; Levin, N.W.; Bosy-Westphal, A.; Korth, O.; Müller, M.J.; Ellegård, L.; Malmros, V.; et al. Body fluid volume determination via body composition spectroscopy in health and disease. Physiol. Meas. 2006,27, 921–933. [CrossRef] 34. Gonzalez, M.C.; Correia, M.I.T.D.; Heymsfield, S.B. A requiem for BMI in the clinical setting. Curr. Opin. Clin. Nutr. Metab. Care 2017,20, 314–321. [CrossRef] 35. Raimann, J.G.; Zhu, F.; Wang, J.; Thijssen, S.; Kuhlmann, M.K.; Kotanko, P.; Levin, N.W.; Kaysen, G.A. Comparison of fluid volume estimates in chronic hemodialysis patients by bioimpedance, direct isotopic, and dilution methods. Kidney Int. 2014,85, 898–908. [CrossRef] Nutrients 2019,11, 809 19 of 22 36. Raimann, J.G.; Abbas, S.R.; Liu, L.; Zhu, F.; Larive, B.; Kotanko, P.; Levin, N.W.; Kaysen, G.A.; FHN Trial. Agreement of single-and multi-frequency bioimpedance measurements in hemodialysis patients: An ancillary study of the Frequent Hemodialysis Network Daily Trial. Nephron Clin. Pract. 2014,128, 115–126. [CrossRef] 37. Piccoli, A. Estimation of fluid volumes in hemodialysis patients: Comparing bioimpedance with isotopic dilution methods. Kidney Int. 2014,85, 738–741. [CrossRef] 38. Seoane, F.; Abtahi, S.; Abtahi, F.; Ellegárd, L.; Johannsson, G.; Bosaeus, I.; Ward, L.C. Slightly superior performance of bioimpedance spectroscopy over single frequency regression equations for assessment of total body water. In Proceedings of the IEEE Engineering in Medicine and Biology Society (EMBC), Milan, Italy, 25–29 August 2015; pp. 3707–3710. 39. Seoane, F.; Abtahi, S.; Abtahi, F.; Ellegárd, L.; Johannsson, G.; Bosaeus, I.; Ward, L.C. Mean expected error in prediction of total body water: A true accuracy comparison between bioimpedance spectroscopy and single frequency regression equations. Biomed. Res. Int. 2015,11, 656323. [CrossRef] 40. Piccoli, A.; Rossi, B.; Pillon, L.; Bucciante, G. A new method for monitoring body fluid variation by bioimpedance analysis: The RXc graph. Kidney Int. 1994,46, 534–539. [CrossRef] 41. Chumlea, W.C.; Guo, S.S.; Kuczmarski, R.J.; Johnson, C.L.; Flegel, K.; Heymsfield, S.; Lukaski, H.; Schoeller, D.; Friedl, K.; Hubbard, V.S. Body composition estimates from NHANES III bioelectrical impedance data. Int. J. Obes. 2002,26, 1596–1609. [CrossRef] 42. Nescolarde, L.; Lukaski, H.; De Lorenzo, A.; de-Mateo-Silleras, B.; Redondo-Del-Río, M.P.; Camina-Martín, M.A. Different displacement of bioimpedance vector due to Ag/AgCl electrode effect. Eur. J. Clin. Nutr. 2016 , 70, 1401–1407. [CrossRef] 43. Pillon, L.; Piccoli, A.; Lowrie, E.G.; Lazarus, J.M.; Chertow, G.M. Vector length as a proxy for the adequacy of ultrafiltration in hemodialysis. Kidney Int. 2004,66, 1266–1271. [CrossRef] 44. Piccoli, A.; Codognotto, M. Bioimpedance vector migration up to three days after the hemodialysis session. Kidney Int. 2004,66, 2091–2092. [CrossRef] 45. Piccoli, A. Whole body–single frequency bioimpedance. Contrib. Nephrol. 2005,149, 150–161. 46. Piccoli, A. Identification of operational clues to dry weight prescription in hemodialysis using bioimpedance vector analysis. The Italian Hemodialysis-Bioelectrical Impedance Analysis (HD-BIA) Study Group. Kidney Int. 1998,53, 1036–1043. [CrossRef] 47. Piccoli, A.; Italian CAPD-BIA Study Group. Bioelectric impedance vector distribution in peritoneal dialysis patients with different hydration status. Kidney Int. 2004,65, 1050–1063. [CrossRef] 48. Guyton, A.C. Textbook of Medical Physiology; Saunders: Philadelphia, PA, USA, 1991; p. 282. 49. Piccoli, A.; Pittoni, G.; Facco, E.; Favaro, E.; Pillon, L. Relationship between central venous pressure and bioimpedance vector analysis in critically ill patients. Crit. Care Med. 2000,28, 132–137. [CrossRef] 50. Basso, F.; Berdin, G.; Virzì, G.M.; Mason, G.; Piccinni, P.; Day, S.; Cruz, D.N.; Wjewodzka, M.; Giuliani, A.; Brendolan, A.; 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, 192–199. [CrossRef] 51. Samoni, S.; Vigo, V.; Reséndiz, L.I.; Villa, G.; De Rosa, S.; Nalesso, F.; Ferrari, F.; Meola, M.; Brendolan, A.; Malacarne, P.; 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 2016,20, 95. [CrossRef] 52. Valle, R.; Aspromonte, N.; Milani, L.; Peacock, F.W.; Maisel, A.S.; Santini, M.; Ronco, C. 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, 519–559. [CrossRef] 53. Bouchard, J.; Soroko, S.B.; Chertow, G.M.; Himmelfarb, J.; Ikizler, T.A.; Paganini, E.P.; Mehta, R.L.; Program to Improve Care in Acute Renal Disease (PICARD) Study Group. Fluid accumulation, survival and recovery of kidney function in critically ill patients with acute kidney injury. Kidney Int. 2009 ,76, 422–427. [CrossRef] 54. Hise, A.C.D.R.; Gonzalez, M.C. Assessment of hydration status using bioelectrical impedance vector analysis in critical patients with acute kidney injury. Clin. Nutr. 2018,37, 695–700. [CrossRef] Nutrients 2019,11, 809 20 of 22 55. Maioli, M.; Toso, A.; Leoncini, M.; Musilli, N.; Bellandi, F.; Rosner, M.H.; McCullough, P.A.; Ronco, C. Pre-procedural bioimpedance vectorial analysis of fluid status and prediction of contrast-induced acute kidney injury. J. Am. Coll. Cardiol. 2014,63, 1387–1394. [CrossRef] 56. Maioli, M.; Toso, A.; Leoncini, M.; Musilli, N.; Grippo, G.; Ronco, C.; McCullough, P.A.; Bellandi, F. Bioimpedance-Guided Hydration for the Prevention of Contrast-Induced Kidney Injury: The HYDRA Study. J. Am. Coll. Cardiol. 2018,71, 2880–2889. [CrossRef] 57. Di Somma, S.; Lalle, I.; Magrini, L.; Russo, V.; Navarin, S.; Castello, L.; Avanzi, G.C.; Di Stasio, E.; Maisel, A. Additive diagnostic and prognostic value of bioelectrical impedance vector analysis (BIVA) to brain natriuretic peptide ‘grey-zone’ in patients with acute heart failure in the emergency department. Eur. Heart J. Acute Cardiovasc. Care 2014,3, 167–175. [CrossRef] 58. Massari, F.; Scicchitano, P.; Ciccone, M.M.; Caldarola, P.; Aspromonte, N.; Iacoviello, M.; Barro, S.; Pantano, I.; Valle, R. Bioimpedance vector analysis predicts hospital length of stay in acute heart failure. Nutrition 2018 , 61, 56–60. [CrossRef] 59. Núñez, J.; Mascarell, B.; Stubbe, H.; Ventura, S.; Bonanad, C.; Bodí, V.; Núñez, E.; Miñana, G.; Fácila, L.; Bayés-Genis, A.; et al. Bioelectrical impedance vector analysis and clinical outcomes in patients with acute heart failure. J. Cardiovasc. Med. 2016,17, 283–290. [CrossRef] 60. Santarelli, S.; Russo, V.; Lalle, I.; De Berardinis, B.; Navarin, S.; Magrini, L.; Piccoli, A.; Codognotto, M.; Castello, L.M.; Avanzi, G.C.; et al. Usefulness of combining admission brain natriuretic peptide (BNP) plus hospital discharge bioelectrical impedance vector analysis (BIVA) in predicting 90 days cardiovascular mortality in patients with acute heart failure. Intern. Emerg. Med. 2017,12, 445–451. [CrossRef] 61. Piccoli, A.; Codognotto, M.; Cianci, V.; Vettore, G.; Zaninotto, M.; Plebani, M.; Maisel, A.; Peacock, W.F. Differentiation of cardiac and noncardiac dyspnea using bioelectrical impedance vector analysis (BIVA). J. Card. Fail. 2012,18, 226–232. [CrossRef] 62. Davies, S.J.; Davenport, A. The role of bioimpedance and biomarkers in helping to aid clinical decision-making of volume assessments in dialysis patients. Kidney Int. 2014,86, 489–496. [CrossRef] 63. Chamney, P.W.; Wabel, P.; Moissl, U.M.; Müller, M.J.; Bosy-Westphal, A.; Korth, O.; Fuller, N.J. A whole-body model to distinguish excess fluid from the hydration of major body tissues. Am. J. Clin. Nutr. 2007 ,85, 80–89. [CrossRef] 64. Wabel, P.; Moissl, U.; Chamney, P.; Jirka, T.; Machek, P.; Ponce, P.; Taborsky, P.; Tetta, C.; Velasco, N.; Vlasak, J.; et al. Towards improved cardiovascular management: The necessity of combining blood pressure and fluid overload. Nephrol. Dial. Transplant. 2008,23, 2965–2971. [CrossRef] 65. Oei, E.L.; Fan, S.L. Practical aspects of volume control in chronic kidney disease using whole body bioimpedance. Blood Purif. 2015,39, 32–36. [CrossRef] 66. O’Lone, E.L.; Visser, A.; Finney, H.; Fan, S.L. Clinical significance of multi-frequency bioimpedance spectroscopy in peritoneal dialysis patients: Independent predictor of patient survival. Nephrol. Dial. Transplant. 2014,29, 1430–1437. [CrossRef] 67. Wabel, P.; Chamney, P.; Moissl, U.; Jirka, T. Importance of whole-body bioimpedance spectroscopy for the management of fluid balance. Blood Purif. 2009,7, 75–80. [CrossRef] 68. Wizemann, V.; Wabel, P.; Chamney, P.; Zaluska, W.; Moissl, U.; Rode, C.; Malecka-Masalska, T.; Marcelli, D. The mortality risk of overhydration in haemodialysis patients. Nephrol. Dial. Transplant. 2009 ,24, 1574–1579. [CrossRef] 69. Tabinor, M.; Elphick, E.; Dudson, M.; Kwok, C.S.; Lambie, M.; Davies, S.J. Bioimpedance-defined overhydration predicts survival in end stage kidney failure (ESKF): Systematic review and subgroup meta-analysis. Sci. Rep. 2018,13, 4441. [CrossRef] 70. Hur, E.; Usta, M.; Toz, H.; Asci, G.; Wabel, P.; Kahvecioglu, S.; Kayikcioglu, M.; Demirci, M.S.; Ozkahya, M.; Duman, S.; et al. Effect of fluid management guided by bioimpedance spectroscopy on cardiovascular parameters in hemodialysis patients: A randomized controlled trial. Am. J. Kidney Dis. 2013,61, 957–965. [CrossRef] 71. Onofriescu, M.; Hogas, S.; Voroneanu, L.; Apetrii, M.; Nistor, I.; Kanbay, M.; Covic, A.C. Bioimpedanceguided fluid management in maintenance hemodialysis: A pilot randomized controlled trial. Am. J. Kidney Dis. 2014,64, 111–118. [CrossRef] Nutrients 2019,11, 809 21 of 22 72. Van Biesen, W.; Williams, J.D.; Covic, A.C.; Fan, S.; Claes, K.; Lichodziejewska-Niemierko, M.; Verger, C.; Steiger, J.; Schoder, V.; Wabel, P.; et al. Fluid status in peritoneal dialysis patients: The European Body Composition Monitoring (EuroBCM) study cohort. PLoS ONE 2011,6, e17148. [CrossRef] 73. Ronco, C.; Verger, C.; Crepaldi, C.; Pham, J.; De Los Ríos, T.; Gauly, A.; Wabel, P.; Van Biesen, W.; IPOD-PD Study Group. Baseline hydration status in incident peritoneal dialysis patients: The initiative of patient outcomes in dialysis (IPOD-PD study). Nephrol. Dial. Transplant. 2015,30, 849–858. [CrossRef] 74. Organ, L.W.; Bradham, G.B.; Gore, D.T.; Lozier, S.L. Segmental bioelectrical impedance analysis: Theory and application of a new technique. J. Appl. Physiol. 1994,77, 98–112. [CrossRef] 75. Sipahi, S.; Hur, E.; Demirtas, S.; Kocayigit, I.; Bozkurt, D.; Tamer, A.; Gunduz, H.; Duman, S. Body composition monitor measurement technique for the detection of volume status in peritoneal dialysis patients: The effect of abdominal fullness. Int. Urol. Nephrol. 2011,43, 1195–1199. [CrossRef] 76. Parmentier, S.P.; Schirutschke, H.; Schmitt, B.; Schewe, J.; Herbrig, K.; Pistrosch, F.; Passauer, J. Influence of peritoneal dialysis solution on measurements of fluid status by bioimpedance spectroscopy. Int. Urol. Nephrol. 2013,45, 229–232. [CrossRef] 77. Arroyo, D.; Panizo, N.; Abad, S.; Vega, A.; Rincón, A.; de José, A.P.; López-Gómez, J.M. Intraperitoneal fluid overestimates hydration status assessment by bioimpedance spectroscopy. Perit. Dial. Int. 2015 ,35, 85–89. [CrossRef] 78. Oh, K.H.; Baek, S.H.; Joo, K.W.; Kim, D.K.; Kim, Y.S.; Kim, S.; Oh, Y.K.; Han, B.G.; Chang, J.H.; Chung, W.; et al. Control of fluid balance guided by body composition monitoring in patients on peritoneal dialysis (COMPASS) study. Does routine bioimpedance-guided fluid management provide additional benefit to non-anuric peritoneal dialysis patients? Results from COMPASS clinical trial. Perit. Dial. Int. 2018 , 38, 131–138. 79. Tabinor, M.; Davies, S.J. The use of bioimpedance spectroscopy to guide fluid management in patients receiving dialysis. Curr. Opin. Nephrol. Hypertens. 2018,27, 406–412. [CrossRef] 80. Zha, Y.; Qian, Q. Protein nutrition and malnutrition in CKD and ESRD. Nutrients 2017,9, 208. [CrossRef] 81. Cupisti, A.; Brunori, G.; Di Iorio, B.R.; D’Alessandro, C.; Pasticci, F.; Cosola, C.; Bellizzi, V.; Bolasco, P.; Capitanini, A.; Fantuzzi, A.L.; et al. Nutritional treatment of advanced CKD: Twenty consensus statements. J. Nephrol. 2018,31, 457–473. [CrossRef] 82. Bellizzi, V.; Scalfi, L.; Terracciano, V.; De Nicola, L.; Minutolo, R.; Marra, M.; Guida, B.; Cianciaruso, B.; Conte, G.; Di Iorio, B.R. Early changes in bioelectrical estimates of body composition in chronic kidney disease. J. Am. Soc. Nephrol. 2006,17, 1481–1487. [CrossRef] 83. Lukaski, H.C.; Kyle, U.G.; Kondrup, J. Assessment of adult malnutrition and prognosis with bioelectrical impedance analysis: Phase angle and impedance ratio. Curr. Opin. Clin. Nutr. Metab. Care 2017 ,20, 330–339. [CrossRef] 84. Piccoli, A.; Codognotto, M.; Piasentin, P.; Naso, A. Combined evaluation of nutrition and hydration in dialysis patients with bioelectrical impedance vector analysis (BIVA). Clin. Nutr. 2014 ,33, 673–677. [CrossRef] 85. Piccoli, A.; Pastori, G.; Codognotto, M.; Paoli, A. Equivalence of information from single frequency v. bioimpedance spectroscopy in bodybuilders. Br. J. Nutr. 2007,97, 182–192. [CrossRef] 86. Teruel-Briones, J.L.; Fernández-Lucas, M.; Ruiz-Roso, G.; Sánchez-Ramírez, H.; Rivera-Gorrin, M.; Gomis-Couto, A.; Rodríguez-Mendiola, N.; Quereda, C. Analysis of concordance between the bioelectrical impedance vector analysis and the bioelectrical impedance spectroscopy in haemodialysis patients. Nefrologia 2012,32, 389–395. 87. Gonzales-Correa, C.H.; Caicedo-Eraso, J.C. Bioelectrical impedance analysis (BIA): A proposal for the standardization of the classical method for adults. J. Phys. Conf. Ser. 2012,407, 012018. [CrossRef] 88. Genton, L.; Herrmann, F.R.; Spörri, A.; Graf, C.E. Association of mortality and phase angle measured by different bioelectrical impedance analysis (BIA) devices. Clin. Nutr. 2018,37, 1066–1069. [CrossRef] 89. Silva, A.M.; Matias, C.N.; Nunes, C.L.; Santos, D.A.; Marini, E.; Lukaski, H.C.; Sardinha, L.B. Lack of agreement of in vivo raw bioimpedance measurements obtained from two single and multi-frequency bioelectrical impedance devices. Eur. J. Clin. Nutr. 2018,1, 22. [CrossRef] 90. Di Somma, S.; Lukaski, H.C.; Codognotto, M.; Peacock, W.F.; Fiorini, F.; Aspromonte, N.; Ronco, C.; Santarelli, S.; Lalle, I.; Autunno, A.; et al. Consensus paper on the use of BIVA (Bioeletrical Impendance Vector Analysis) in medicine for the management of body hydration. Emerg. Care J. 2011 ,4, 6–14. [CrossRef] Nutrients 2019,11, 809 22 of 22 91. Covic, A.; Ciumanghel, A.I.; Siriopol, D.; Kanbay, M.; Dumea, R.; Gavrilovici, C.; Nistor, I. Value of bioimpedance analysis estimated “dry weight” in maintenance dialysis patients: A systematic review and meta-analysis. Int. Urol. Nephrol. 2017,49, 2231–2245. [CrossRef] 92. Scotland, G.; Cruickshank, M.; Jacobsen, E.; Cooper, D.; Fraser, C.; Shimonovich, M.; Marks, A.; Brazzelli, M. Multiple-frequency bioimpedance devices for fluid management in people with chronic kidney disease receiving dialysis: A systematic review and economic evaluation. Health Technol. Assess. 2018,22, 1–138. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).