Comparison of vector and conventional bioelectrical impedance analysis in the optimal dry weight prescription in hemodialysis

Background: Dry weight prescription is commonly based on symptoms induced by inappropriate fluid removal by hemodialysis (HD). Aim of this study was to compare the assessment of volume status by conventional bioelectrical impedance analysis (BIA) and the resistance-reactance (RXc) graph method in HD patients achieving their target dry weight determined on clinical criteria. Methods: We studied 39 HD patients (23 males and 16 females, mean age 52 +/- 17 years, dialytic age 41.2 +/- 37 months). Dry weight, prescribed according to the standard clinical criteria, was constantly achieved in the last 3 months. Patients symptom-free over the last 3 months were defined as asymptomatic. Patients with either muscular cramps or hypotensive episodes were defined as symptomatic. Thirty-three healthy volunteers (11 males, 22 females, mean age 50 +/- 11 years) constituted the control group. Standard, single frequency (50 kHz), tetrapolar, BIA measurements were obtained in controls, and in patients before, every 60 min, and 30 min after one HD session. Total body water (TBW), and extracellular water (ECW) were calculated using conventional BIA regression equations. In both groups, tissue hydration was also assessed by the RXc graph method. Results: On the basis of 95% tolerance interval (mean +/- 2 SD) for the ECW (%) calculated in healthy subjects (ECW = 35-44%), HD patients were divided into 3 groups according to their post-HD ECW: 72% normohydrated with ECW 35-44%, 10% overhydrated with ECW >44%, and 18% underhydrated with ECW <35%. Patients were also classified into 3 categories according to the RXc graph method: 38% normohydrated with vectors within the reference 75% tolerance ellipse, 0% overhydrated with short vectors below the lower pole of the 75% tolerance ellipse, and 62% underhydrated with long vectors above the upper pole of the 75% tolerance ellipse. The progressive removal of body fluid during HD treatment was associated with a progressive increase in both impedance vector components, R and Xc. Eleven of thirty-nine patients (28%) were symptomatic during HD treatment in the last 3 months. The majority of these (73%) were classified as normohydrated according to ECW estimates, while 9 and 18% were classified as over- and underhydrated, respectively. This frequency distribution was significantly different from that obtained with the RXc graph method (chi(2) = 6.9, p = 0.03) where the majority (73%) were classified as underhydrated, while 0 and 27% were classified as over- and normohydrated, respectively. The frequency distribution of the 28 asymptomatic patients also significantly differed between conventional BIA and RXc graph hydration categories (chi(2) = 10.8, P = 0.005), since 11, 71 and 18% vs. 0, 43 and 57% of patients were classified as over-, normo-, and underhydrated, respectively. Conclusions: The classification of volume status based on conventional BIA was insensitive to either clinical situation (presence or absence of symptoms). in contrast, the classification based on the RXc graph was consistent with the clinical course in symptomatic patients (73% dehydrated, and 27% normohydrated), while it did not reflect the clinical course in asymptomatic patients, 57% of whom were classified as (already) underhydrated. A longitudinal study will establish the clinical usefulness of RXc graph indications in asymptomatic patients. Copyright (C) 2000 S. Karger AG, Basel.