Measurement of intercompartmental fluid shifts during haemodialysis in children

Seven children (age range 12-19 years, post-dialysis weights 23-43 kg) were studied during 20 haemodialysis sessions. Impedance between wrist and ankle (on the non-fistula side) was recorded using the Xitron 4000B analyser. A 2 ml sample of blood was taken for total protein and haematocrit from the arterial line at the start of dialysis. At approximately 20 minute intervals during dialysis, the time and volume of ultrafiltrate removed were recorded, and a simultaneous measurement of whole body impedance made over 25 logarithmically spaced frequencies in the range 5-500 kHz. A 2 ml sample of blood was also taken, from which serum protein and haematocrit were calculated. Hypotensive episodes occurred during four haemodialysis sessions. The percentage change in extracellular fluid (ECF) volume was calculated, at each sample time for each session, using the impedance measurements and ultrafiltration measurements (denoted delta Vi and delta U respectively). Changes in the intravascular volume were estimated using measurements of haematocrit and serum protein (and denoted delta Vh and delta Vp respectively). Least-squares regression gave delta Vi = 3.77 delta Vh, 1.33 delta Vp and 0.39 delta U, and r2 = 0.72, 0.94 and 0.95 respectively (p < 0.0001 in each case) for the 16 dialysis sessions without hypotensive episodes. Similar analysis of four dialysis sessions with hypotensive episodes gave similar relationships with correlation coefficients 0.64, 0.92 and 0.94. These relationships may not be accounted for by the anthropometric terms alone in the impedance equations. Impedance measurements also detected the addition of 300 ml isotonic saline given at the onset of each of the four hypovolaemic episodes. The regression equations support the following hypothesis: during haemodialysis, ultrafiltrate is removed from the intravascular volume but is replenished by fluid from the interstitial volume. The reduction in ECF volume measured by impedance (where the ECF comprises the intravascular and interstitial volumes) delta Vi is therefore greater than delta Vh and delta Vp, which only measure intravascular volume, but less than delta U since the ECF is replenished by fluid from the interstitial space. That delta Vh is greater than delta Vp may be due to protein loss during dialysis. The results suggest that whole body impedance measurements reflect changing body water distribution during dialysis in children.