Prevalence of extracellular fluid volume expansion has been previously reported to be high.

The present study confirms the high prevalence of volume expansion and indicates that this cannot be estimated by clinical parameters. Moreover, the study adds that volume contraction is also relatively common. Last, volume status is not related to intradialytic weight gain and is unable to explain the high incidence of paradoxical hypertension.

Accurate assessment of volume status (VS) remains one of the greatest challenges in the treatment of hemodialysis (HD) patients [1]. Chronic volume expansion contributes to hypertension, left ventricular hypertrophy, and heart failure in HD patients [2–6]. Conversely, volume contraction could predispose the HD patient to intradialytic hypotension, cramps, arrhythmias, cardiac stunning, and reduced well-being after treatment [2]. Therefore, adequate extracellular fluid volume (ECFV) control is crucial for blood pressure regulation and for prevention of cardiovascular complications in this population [7].

Clinical assessment of volume expansion by assessing blood pressure, edema, and central venous pressure has limitations [8–11]. Interdialytic weight gain (IDWG) is not sufficient to assess volume expansion since it does not necessarily correlate with ECFV expansion [12]. The inferior vena cava collapse index [2], ultrasound assisted assessment of pulmonary fluid content [13], and echocardiography [2] are helpful to assess volume status, but they do not provide an estimate of total ECFV expansion or depletion and are impractical for usage in clinical practice. Multifrequency bio-impedance spectroscopy is a convenient method to assess extracellular and intracellular fluid volume [14]. This method has been validated against gold standard dilution methods [14, 15] and is applicable in the setting of HD [16].

Given the mortality associated with ECFV expansion [6, 17] and the knowledge that adequate volume control provides better control of blood pressure [18], more accurate objective methods for volume assessment would be valuable. In the current study, we tested the hypothesis that volume expansion is highly prevalent in HD patients, likely due to the inability to judge volume status from clinical parameters. Bio-impedance spectroscopy was applied to assess volume status in an in-center HD population and compared with clinical volume assessment. Our aims were (1) to assess the prevalence of volume expansion and volume depletion in our HD population using bio-impedance spectroscopy measurements, (2) to investigate the association between clinical parameters and volume status as assessed by bio-impedance, and (3) to search for a set of clinical parameters that best predict volume status in HD patients.

In the current study, ECFV expansion as assessed by bio-impedance in in-center HD patients was highly prevalent. ECFV contraction was also observed and was more frequently associated with intradialytic hypotension. Paradoxical hypertension was highly prevalent and not associated with volume status. Several clinical parameters were more prevalent in HD patients with ECFV expansion, however, were not sufficiently sensitive and specific to be applied to robustly assess volume status.

Fluid volume expansion in HD patients is a well-recognized problem [12, 24] associated with hypertension [2, 3], left ventricular hypertrophy, dilated cardiomyopathy [4], heart failure, and eventually with high mortality [5, 6]. Strikingly, about 50 % of our patients displayed mild to severe ECFV expansion of >7 % of ECFV. Moderate to severe ECFV expansion of >15 % of ECFV was observed in 23 % of the patients, despite frequent routine clinical volume assessment. Previous studies using bio-impedance for quantification of volume status reported slightly lower [25] or similarly high prevalence of ECFV expansion [4, 5]. Of note, volume status and interdialytic weight gain were not correlated in the current study. This implies that strategies to improve volume status need to separately address dry weight and dietary salt intake to achieve good volume regulation. Underscoring the relevance of IDWG, one study reported that reducing the interdialytic weight gain without changing the dry weight reduced ventricular hypertrophy [26]. Therefore, although the debate as to whether chronic ECFV expansion or IDWG is the most important factor determining cardiovascular outcome is not settled, controlling fluid overload can reduce hypertension and LVH and improve outcome [27]; controlling the latter could decrease cardiac stunning [28, 29].

Pre-dialysis volume depletion of >7 % of ECFV (comparable to 1.1 L for average ECFV) was demonstrated in 9 % of patients, compared to >1.1 L in 5 % of patients in a previous report [25]. Volume-depleted patients only differed from normo- or hypervolemic patients regarding more frequent intradialytic hypotension. Intradialytic hypotension occurred in 17 % of all patients. Previous studies reported the incidence of intradialytic hypotension to be between 15 and 25 % of HD sessions [30, 31]. Since volume status was assessed pre-dialysis and several hypovolemic patients had substantial interdialytic weight gains of >25 % of ECFV, these patients would be severely hypovolemic post-dialysis. Although several patients with volume depletion had very substantial interdialytic weight gain, no correlation could be established between interdialytic weight gain and volume contraction. Fluid overload has previously been inversely associated with IDWG, in about 15000 patients from 60 dialysis centers [12]. Taken together, our study indicates that a relevant fraction of patients showed fluid depletion and a higher risk of intradialytic hypotension.

Hypertension was present in 54 % (27 % of total) of hypervolemic and in 36 % (19 % of total) of normovolemic patients. Fluid status and systolic blood pressure were correlated, in contrast to ECFV expansion and diastolic and pulse pressure. This shows that the relationship between blood pressure and volume status is complex. Wabel et al. analyzed the relation between blood pressure and volume expansion using comparable methodology in 500 HD patients [32]. Volume-dependent hypertension was found in 15 % of patients; the majority of patients (27 %) were normotensive and normovolemic. Only 10 % of patients had normal blood pressure despite volume expansion. This suggests that volume expansion is responsible for hypertension in HD patients; yet, the use of antihypertensives might obscure the importance of fluid overload for BP control.

Paradoxical hypertension in HD patients has been associated with increased risk of mortality [33]. Paradoxical hypertension was common in our study population (31 %) and was not significantly different among the three patient groups. Limited information is available about the prevalence of paradoxical hypertension and its relation to volume status in a large HD population. A previous study using similar definition of paradoxical hypertension reported prevalence in 21 % of patients [34]. UF rate was significantly lower in patients with paradoxical hypertension; however, no assessment of ECFV was performed [34]. Using several different definitions for paradoxical hypertension, other studies report incidence of 10–15 % [23]. A recent study reported a decline in the incidence of paradoxical hypertension with excessive ultrafiltration concluding that intradialytic hypertension may be a sign of volume expansion [35]; ECFV was not assessed in that study. Altogether, our data indicate a high frequency of paradoxical hypertension but no clear association with volume expansion. Moreover, it underscores the need to understand this important problem with consequences for outcome better [36].

In search for clinical and laboratory characteristics associated with volume status, multivariate regression analysis identified only four predictors for volume expansion (edema, lower BMI, higher SBP, and smoking). None of these parameters displayed both a good sensitivity and specificity. In our study, the presence of edema had a good positive predictive value. A previous study reported that pedal edema correlates well with cardiovascular risk factors and left ventricular mass, but it did not reflect volume in HD patients as assessed by cardiac biomarkers and echocardiography [37]. Using a similar methodology in PD patients, multiple regression analysis revealed that DM, higher SBP, older age, male gender, lower serum albumin, and lower BMI were significant predictors for volume expansion [38]. It seems therefore that a set of parameters that can assist to better assess fluid volume in HD patients on clinical grounds cannot be formulated. In addition to the evaluation of specific clinical indicators of fluid overload, we related the clinically assessed volume status to the estimated normal ECFV and compared this with the volume status measured by the BCM. This yields the interesting observation that although there is a positive correlation between clinically assessed and BCM assessed volume status, clinical assessment indicated fluid overload at normovolemia assessed by BCM. Moreover, the slope between clinically assessed and BCM assessed VS was far below unity. This indicates that a clinical assessment is not precise and that severe fluid overload is underestimated.

Our study has several limitations. First of all, bio-impedance spectroscopy-based extracellular fluid volume assessment has an error of −0.4 ± 1.4 L compared to gold standard dilution methods [19]. The study did not separately assess the validity on the 120 healthy subjects in the study and the 32 HD patients. However, a recent review reported that radioisotope methods, previously considered the gold standard, have similar accuracy compared to bio-impedance in dialysis patients [39]. Moreover, bio-impedance is highly reproducible with inter-observer and intra-observer errors of less than 2 % [39]. The measurement was performed once, prior to the second dialysis session of the week. Although this could introduce some noise, it is not likely that it would affect our interpretation of the data. Finally, we cannot assure that different devices to assess fluid status and different methodology to calculate the normal ECFV or an individual would not lead to slightly different numerical results. It is considered not likely that the overall interpretation of the study would have been different.

In summary, using bio-impedance spectroscopy, we found that volume expansion is highly prevalent in our HD patients and volume contraction was also not uncommon. Intradialytic hypotension was more common in hypovolemic patients. Fluid volume expansion or contraction could not be reliably identified by clinical parameters, except that edema predicted fluid volume expansion. This study shows that bio-impedance can assist to determine volume status. Volume status, in turn, is not related to intradialytic weight gain and is unable to explain the high incidence of paradoxical hypertension. Whether volume expansion or contraction determined using bio-impedance is associated with worse outcome still needs to be determined.