What volume to choose to assess online Kt/V?

• Published in: Journal of nephrology Vol. 33; no. 1; pp. 137 - 146
• Main Authors: Casino, Francesco Gaetano, Mancini, Elena, Santarsia, Giovanni, Mostacci, Salvatore Domenico, D’Elia, Filomena, Di Carlo, Maria, Iannuzzella, Francesco, Rossi, Luigi, Vernaglione, Luigi, Grimaldi, Daniela, Rapanà, Renato, Basile, Carlo
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.02.2020
• Subjects: Medicine; Medicine & Public Health; Nephrology; Original Article; Urology; Urea distribution volume; Watson; Kt/V; Double pool urea kinetic model; Hemodialysis; Ionic dialysance
• ISSN: 1121-8428; 1724-6059; 1724-6059
• DOI: 10.1007/s40620-019-00636-9

Introduction Urea distribution volume (V) can be assessed in different ways, among them the anthropometric Watson Volume (V W ). However, many studies have shown that V W does not coincide with V and that the latter can be more accurately estimated with other methods. The present multicentre study was designed to answer the question: what V to choose to assess online Kt/V? Materials and methods Pre- and postdialysis blood urea nitrogen concentrations and the usual input data set for urea kinetic modelling were obtained for a single dialysis session in 201 Caucasian patients treated in 9 Italian dialysis units. Only dialysis machines measuring ionic dialysance (ID) were utilized. ID reflects very accurately the mean effective dialyser urea clearance (Kd). Six different V values were obtained: the first one was V W ; the second one was computed from the equation established by the HEMO Study to predict the single pool-adjusted modelled V from V W (V H ) (Daugirdas JT et al. KI 64: 1108, 2003); the others were estimated kinetically as: 1. V_ID, in which ID is direct input in the in the double pool variable volume (dpVV) calculation by means of the Solute-solver software; 2. V_Kd, in which the estimated Kd is direct input in the dpVV calculation by means of the Solute-solver software; 3. V_KTV, in which V is calculated by means of the second generation Daugirdas equation; 4. V_SPEEDY, in which ID is direct input in the dpVV calculation by means of the SPEEDY software able to provide results quite similar to those provided by Solute-solver. Results Mean± SD of the main data are reported: measured ID was 190.6 ± 29.6 mL/min, estimated Kd was 211.6 ± 29.0 mL/min. The relationship between paired data was poor (R 2  = 0.34) and their difference at the Bland–Altman plot was large (21 ± 27 mL/min). V W was 35.3 ± 6.3 L, V H 29.5 ± 5.5, V _ ID 28.99 ± 7.6 L, V _ SPEEDY 29.4 ± 7.6 L, V_KTV 29.7 ± 7.0 L. The mean ratio V W /V_ID was 1.22, (i.e. V W overestimated V_ID by about 22%). The mean ratio V H /V_ID was 1.02 (i.e. V H overestimated V_ID by only 2%). The relationship between paired data of V_ID and V W was poor (R 2  = 0.48) and their mean difference at the Bland–Altman plot was very large (−  6.39 ± 5.59 L). The relationship between paired data of V_ID and V H was poor (R 2  = 47) and their mean difference was small but with a large SD (− 0.59 ± 5.53 L). The relationship between paired data of V_ID and V_SPEEDY was excellent (R 2  = 0.993) and their mean difference at the Bland–Altman plot was very small (− 0.54 ± 0.64 L). The relationship between paired data of V_ID and V_KTV was excellent (R 2  = 0.985) and their mean difference at the Bland–Altman plot was small (− 0.85 ± 1.06 L). Conclusions V_ID can be considered the reference method to estimate the modelled V and then the first choice to assess Kt/V. V_SPEEDY is a valuable alternative to V_ID. V_KTV can be utilized in the daily practice, taking also into account its simple way of calculation. V W is not advisable because it leads to underestimation of Kt/V by about 20%.