Scientific Publications by FDA Staff

Thromb Res 2013 Sep;132(3):374-80

Determining the impact of instrument variation and automated software algorithms on the TGT in hemophilia and normalized plasma.

Woodle SA, Shibeko AM, Lee TK, Ovanesov MV

BACKGROUND: Despite increasing recognition as a more precise test of in vivo hemostatic conditions, standardization of the thrombin generation test (TGT) continues to hinder its development as routine clinical practice. Prior efforts largely focused on comparing the effects of experimental conditions and different reagents. Commercialized kits, instruments and software have been introduced to calculate the TG curve and its parameters. However, modified versions of the TGT continue to be used worldwide on a variety of microplate reader instruments and processed using individualized algorithms. No prior study has compared the effect of instrument choice and its inherent noise profile on the processing of the TG curve and its common endpoint parameters. MATERIALS AND METHODS: Hemophilia A plasma supplemented with buffer or Factor VIII, mimicking hemophilic or normalized samples respectively, was monitored for thrombin generation after activation with TF on six different fluorescent microplate readers. Each instrument was optimized for TGT signal recording prior to testing. An automated software package containing various mathematical algorithms was utilized to compute the TG curves and parameters, and compare different TG processing approaches. RESULTS: Instruments produced unique noise profiles and end-point parameters that were incomparable in absolute signal terms. Similar relative hemophilic responses were obtained across various instruments when the normalized plasma sample was used as an internal standard. Smoothing algorithms corrected destructive instrument noise. CONCLUSIONS: Instrument-induced errors from numerical differentiation during TG curve processing cannot be eliminated by external calibrators, and require careful qualification of the instrument and implementation of noise-reducing software algorithms.