To date, the quantification of the anticoagulant (ACD-A) in plasma units has been based on theoretical calculations. An accurate quantification could help minimize the risks associated with plasmapheresis, given that the total ACD-A used during the procedure is distributed between the donor and the plasma unit. Our aim was to experimentally quantify the volume of ACD-A in units collected by plasmapheresis.We used proton nuclear magnetic resonance spectroscopy to measure the ACD-A volume in 295 plasma units collected by the Azienda USL-IRCCS of Reggio Emilia, Italy. We analyzed the determinants of the differences between estimated and measured ACD-A through multivariate regression models.The experimentally measured ACD-A in plasma units was variable, with 45% of the samples showing a discrepancy of more than 15 mL compared to the manufacturer's estimate. ACD-A was underestimated for higher density of the units (p < 0.0005); a weak association was also observed with triglycerides (underestimated for higher levels, p = 0.015) and sex (overestimated in females, p = 0.008), but our model explained only 35% of the individual variability.The manufacturer's algorithms do not accurately estimate the ACD-A in units collected by plasmapheresis. Donor-related characteristics may affect ACD-A distribution between donor and plasma unit, thereby explaining the discrepancies between estimate and measurement. Errors in the estimate of the ACD-A actually received by donors could hamper studies on dose-response relationship between anticoagulant and adverse reactions. Our work should stimulate research on tailored procedures aimed at minimizing the anticoagulant received by donors and increasing plasmapheresis safety.
