Abstract
Background: The needs for sensitive coagulation factor assays able to measure factor VIII (FVIII) and factor IX (FIX) in the range of 0.0 to 1.0 %, are continuously growing with diversification of hemophilia management. However, practical methods with sufficient analytical sensitivity available in clinical laboratory have not yet been introduced. We developed new coagulation factor assays applying various parameters derived from a turbidity based coagulometer and examined their ability to measure low-level FVIII and FIX and analytical resolution in that range.
Method: We prepared 12 spiked samples with FVIII and FIX levels from 0.0 to 2.4 % and conducted conventional one-stage coagulation factor assays in repeat. We collected measured values of APTT, velocity and acceleration peaks of coagulation (peak 1 and peak 2) from each measurement. We also calculated values of peak 1 and peak 2 from the mathematical model of turbidity curves. From the measured values of these parameters we derived calibration formulae for coagulation factor assays, FVIIICT, FVIIIpeak1, FVIIIpeak2, FVIIIcalc1, FVIIIcalc2, FIXCT, FIXpeak1, FIXpeak2, FIXcalc1, and FIXcalc2.
Results: The reliability interval (range of FVIII levels producing unequivocal results) of FVIIICT (the conventional FVIII assay) covered only 9 % of 0.0 to 1.0 % range. For new assays, the coverages were 54, 31, 55, and 65 % for FVIIIpeak1, FVIIIpeak2, FVIIIcalc1, and FVIIIcalc2 respectively. The resolution between immediate levels of spiked samples could be determined from modeled distributions or be checked simply by inspecting the actual assay result distributions. For FVIIIpeak1, 0.2 % and 0.6 % results stood apart from each other. For FVIIIcalc1 and FVIIIcalc2, 0.2, 0.4, and 0.6 % were distinguished from each other. When we measured recombinant human (rh) FVIII, the coverages were 7, 64, 52, 73, and 79 % for rhFVIIICT, rhFVIIIpeak1, rhFVIIIpeak2, rhFVIIIcalc1, and rhFVIIIcalc2 respectively. (rh)FVIIIpeak1, (rh)FVIIIcalc1, and (rh)FVIIIcalc2 particularly showed wide measurable ranges of guarantee. For FVIIIpeak1, 0.2 % and 0.6 % results stood apart from each other. For FVIIIcalc1 and FVIIIcalc2, 0.2, 0.4, and 0.6 % were distinguished from each other. rhFVIIIpeak1 and rhFVIIIcalc1 showed slightly better resolution than the former. rhFVIIIcalc2 was notable in that every 0.1, 0.2, 0.4, 0.6, 0.8 % result stood apart from each immediate level result. We could not determine certainty interval (the range of unequivocal values) of FIXCT and FIXpeak2 because the 0.0 % and 1.0 % ranges overlapped. Thus, the conventional FIX assay cannot measure between 0.0 and 1.0 %. FIXpeak1, FIXcalc1, and FIXcalc2 worked better and the certainty interval of unequivocal results could be determined between 0.0 and 1.0 %. The reliability interval was not available for any FIX assay. Results from rhFIX measurements were similar those of plasma FIX assays.
Conclusion: We introduce new FVIII and FIX assays with superior analytical resolution in the range of 0.0 to 1.0 % in comparison to the conventional assays.
No relevant conflicts of interest to declare.
