Linear decay of relative blood volume during ultrafiltration predicts hemodynamic instability

• Published in: American journal of kidney diseases Vol. 40; no. 3; pp. 556 - 565
• Main Authors: Mitra, Sandip, Chamney, Paul, Greenwood, Roger, Farrington, Ken
• Format: Journal Article
• Language: English
• Published: Orlando, FL Elsevier Inc 01-09-2002; Elsevier
• Subjects: Adult; Aged; Anesthesia. Intensive care medicine. Transfusions. Cell therapy and gene therapy; Antihypertensive Agents - pharmacology; Biological and medical sciences; Blood Circulation - drug effects; Blood Pressure - drug effects; Blood Pressure Monitoring, Ambulatory - statistics & numerical data; blood volume; Blood Volume - physiology; Body Water - drug effects; Body Water - metabolism; Catheters, Indwelling - adverse effects; dry weight; Emergency and intensive care: renal failure. Dialysis management; Hemodialysis (HD); Hemodynamics - drug effects; Humans; hypotension; Hypotension - blood; Hypotension - metabolism; Hypotension - physiopathology; Intensive care medicine; Linear Models; Medical sciences; Middle Aged; Multivariate Analysis; perturbation analysis; Predictive Value of Tests; Renal Dialysis - adverse effects; Renal Dialysis - statistics & numerical data; ultrafiltration (UF); Ultrafiltration - methods; hypotension; ultrafiltration (UF); blood volume; Hemodialysis (HD); perturbation analysis; dry weight; Human; Extrarenal dialysis; Blood volume; Hemodialysis; Arterial hypotension; Hemofiltration; Prediction; Cardiovascular disease; Hemodynamics; Dry weight
• ISSN: 0272-6386; 1523-6838
• DOI: 10.1053/ajkd.2002.34914

Background: Relative blood volume (RBV) changes during hemodialysis (HD) are poorly understood. We wish to define characteristics of RBV profiles at different hydration states predictive of hemodynamic instability. Methods: Thirty patients underwent online RBV monitoring during an HD session with intermittent ultrafiltration (UF) pulses administered until the onset of hypotension. The RBV decay constant (τ) was derived from curve fitting. Linear divergence, the net deviation of the RBV curve during UF from predicted linear decay, was computed from initial 1-minute slopes. Results: The best correlation with proximity to dry weight (PDW) was provided by linear divergence (r = 0.817; P < 0.001), its major determinant in multiple regression analysis. Other predictors were RBV at initiation of UF pulse, UF pulse volume, and UF decay constant (τUF). These parameters were significantly different in UF pulses within 1 kg and 1 kg or greater of dry weight. There were no correlations with refill parameters. The occurrence of hypotension was not different at RBVs less than 90% (7.4%) or 90% or greater (5.3%). τUF, linear divergence, RBV at initiation of UF pulse (all P < 0.001), and UF decay amplitude (P < 0.01) were different between hypotensive and normotensive UF pulses. Hypotension was the only independent predictor of τUF (R2 = 0.40; P < 0.001). The only independent predictor of linear divergence was PDW (R2 = 0.667; P < 0.001). Conclusion: Approaching dry weight, the RBV decline during UF switched from exponential to linear decay, probably indicating failing vascular refill. Monitoring deviation from linearity may allow improved hemodynamic stability and attainment of optimal post-HD weight. © 2002 by the National Kidney Foundation, Inc.