Abstract
Objective
A feasible method has recently been proposed for estimating absolute blood volume (ABV) in hemodialysis (HD) patients based on intradialytic infusion of a dialysate bolus and visual assessment of the subsequent increase in relative blood volume (RBV) tracked by the dialysis machine. The aim of this study was to develop a method for more objective determination of such RBV increase to improve the accuracy of ABV estimation.
Methods
We proposed a numerical algorithm consisting of interpolation and polynomial fitting of RBV signals, which we evaluated on data from 64 HD sessions in 48 patients with 240 mL of dialysate infused approximately 1 h into HD. The estimated ABV values were compared with those from a simple two-point method described previously as well as with the values calculated using the Nadler formula and the Lemmens-Bernstein-Brodsky formula.
Results
Compared to the simple method, the improved method provided higher (more plausible) estimates of ABV (median 4.79 vs 4.53 L, p < 0.001) and specific ABV (median 68.5 vs 66.4 mL/kg, p < 0.001). The improved method also provided much lower intra-patient variability of ABV estimated in different sessions of the same week (median spread 180 vs 462 mL, p < 0.001) and showed narrower limits of agreement with both Nadler and Lemmens-Bernstein-Brodsky formulae.
Conclusion
The proposed numerical method constitutes a substantial improvement over the simple method by averaging the noise and short-term variability in RBV signals.
Significance
More accurate estimates of ABV in HD patients could aid in managing their fluid status.
A feasible method has recently been proposed for estimating absolute blood volume (ABV) in hemodialysis (HD) patients based on intradialytic infusion of a dialysate bolus and visual assessment of the subsequent increase in relative blood volume (RBV) tracked by the dialysis machine. The aim of this study was to develop a method for more objective determination of such RBV increase to improve the accuracy of ABV estimation.
Methods
We proposed a numerical algorithm consisting of interpolation and polynomial fitting of RBV signals, which we evaluated on data from 64 HD sessions in 48 patients with 240 mL of dialysate infused approximately 1 h into HD. The estimated ABV values were compared with those from a simple two-point method described previously as well as with the values calculated using the Nadler formula and the Lemmens-Bernstein-Brodsky formula.
Results
Compared to the simple method, the improved method provided higher (more plausible) estimates of ABV (median 4.79 vs 4.53 L, p < 0.001) and specific ABV (median 68.5 vs 66.4 mL/kg, p < 0.001). The improved method also provided much lower intra-patient variability of ABV estimated in different sessions of the same week (median spread 180 vs 462 mL, p < 0.001) and showed narrower limits of agreement with both Nadler and Lemmens-Bernstein-Brodsky formulae.
Conclusion
The proposed numerical method constitutes a substantial improvement over the simple method by averaging the noise and short-term variability in RBV signals.
Significance
More accurate estimates of ABV in HD patients could aid in managing their fluid status.
Original language | English |
---|---|
Article number | 105440 |
Pages (from-to) | 1-9 |
Number of pages | 9 |
Journal | Biomedical Signal Processing and Control |
Volume | 88 |
Issue number | Part A |
Early online date | 10 Oct 2023 |
DOIs | |
Publication status | Published - Feb 2024 |
Research Field
- Medical Signal Analysis
Keywords
- Absolute blood volume
- Fluid status
- Hemodilution
- Relative blood volume
- Ultrafiltration