Estimating the plasma volume by infusing albumin: a retrospective feasibility study

The combined changes in plasma albumin and blood hemoglobin can probably be used to estimate the plasma volume (PV) when albumin is infused. However, the optimal setup, timing of the blood sampling, and the importance of capillary leakage to the calculations are unclear.

In this technical vignette, we estimated the PV using retrospective data on plasma albumin and blood hemoglobin obtained during intravenous infusion of 3 mL/kg of 20% albumin over 30 min in 41 volunteers and 45 patients. We used a manual and a kinetic correction for capillary leakage of albumin. The results were compared to the mean of two anthropometric equations derived via tracer methods.

The anthropometric PV was 3.00 ± 0.63 L (mean ± SD). The strongest linearity between the albumin-derived and anthropometric PV was obtained at the end, and 10 min after the end, of the 30-min infusions; the correlation coefficient was 0.75 over this time frame. The difference between the two measures (the prediction error) was 0.31 ± 0.56 L but the SD was only half as high for PVs< 2.5 L than for larger PVs. There was slightly stronger linearity and better accuracy, but no better precision, when data were corrected for capillary leakage.

This study suggests how an evaluation of this method using isotopes can be conducted. Changes in plasma albumin and blood hemoglobin have the best chance to accurately indicate the PV at the end of, or 10 min after, a 30-min infusion of albumin. Subtraction of 0.3 L from the PV is sufficient to correct for capillary leakage of albumin.