Diagnostic utility of blood volume monitoring in hemodialysis patients.

BACKGROUND Assessment of volume state is difficult in hemodialysis patients. Whether continuous blood volume monitoring can improve the assessment of volume state is unclear. STUDY DESIGN Diagnostic test study. SETTINGS & PARTICIPANTS Asymptomatic long-term hemodialysis patients (n = 150) in 4 university-affiliated hemodialysis units. INDEX TESTS Ultrafiltration rate (UFR) divided by postdialysis weight (UFR index), slopes of relative blood volume (RBV), RBV slope corrected for UFR and weight (volume index). REFERENCE TESTS Dialysis-related symptoms and echocardiographic signs of volume excess and volume depletion, assessed by using inferior vena cava (IVC) diameter after dialysis and its collapse on inspiration. Volume excess was defined as values in the upper third of IVC diameter or lower third of IVC collapse on inspiration. Volume depletion was defined as values in the lower third of IVC diameter or upper third of IVC collapse on inspiration. RESULTS Mean UFR was 8.3 +/- 3.8 (SD) mL/h/kg. Mean RBV slope was -2.32% +/- 1.50%/h. Mean volume index was -0.25% +/- 0.17%/h/mL/h ultrafiltration/kg. Volume index provided the best fit of observed RBV slopes. Volume index was related to dizziness, the need to decrease UFR, and placement in Trendelenburg position. RBV and volume index, but not UFR index, were related to echocardiographic markers of volume excess and depletion. Areas under the receiver operating characteristic curve to predict volume excess were 0.48 (95% confidence interval [CI], 0.33 to 0.63) for UFR index, 0.71 (95% CI, 0.60 to 0.83) for RBV slope, and 0.73 (95% CI, 0.59 to 0.86) for volume index. Areas under the receiver operating characteristic curve to predict volume depletion were 0.56 (95% CI, 0.38 to 0.74) for UFR index, 0.55 (95% CI, 0.38 to 0.72) for RBV slope, and 0.62 (95% CI, 0.48 to 0.76) for volume index. LIMITATIONS Dialysis-related symptoms and echocardiographic findings are not validated measures of volume. Our results were not adjusted for demographic or clinical characteristics; performance characteristics of the indices may differ across populations. CONCLUSIONS Volume index appears to be a novel marker of volume, but requires validation studies, and its utility needs to be tested in clinical trials.