Inhibitor activity of patient samples is read in NBU mL−1 from a semi logarithmic plot representing the correlation between residual FVIII activity (logarithmic) and inhibitor activity (linear) [15]. The regression line is fully defined by 100% residual FVIII activity with 0 NBU mL−1 inhibitor and 50% residual FVIII activity with 1 NBU mL−1 inhibitor (Fig. 2). Dose–response curves of test plasma need to show parallelism with this calibration curve. If not, inhibitor data are not reliable and an alternative

strategy needs to be followed (e.g. type II FVIII inhibitors). When the residual FVIII activity of undiluted sample is below 25%, retesting of more diluted samples is recommended because of non-linearity of inhibitor concentration and residual FVIII activity with high inhibitor titres. Dilutions Cell Cycle inhibitor have to be made with FVIII-deficient plasma. The internationally accepted cut-off value for Bethesda-based inhibitor assays is 0.6 BU mL−1. This value is rather

high, for it has been derived from the results with the classical Bethesda assay and is a reflection of the low sensitivity LY294002 cost and specificity of this method. However, sensitivity and specificity, including the cut-off value, have been improved in the Nijmegen assay [14] although clinical studies comparing inhibitor titres and kinetic parameters are still lacking. Therefore, every individual laboratory has to assign the laboratory-specific cut-off value by assaying positive and negative inhibitor samples from haemophilia patients. The FVIII inhibitor assay is rather complicated and includes critical analytical stages and variables that need careful handling to get reliable results. The inactivation of FVIII by inhibitors is pH-, temperature- and time-dependent. The

pH stability of the incubation mixtures is an essential feature of the Nijmegen assay. Incubation of insufficiently buffered plasma mixtures will give rise to increasing pH resulting in uncontrolled and non-specific inactivation of FVIII [13,16]. pH stabilization of MCE the incubation mixture by buffering the normal pooled plasma will overcome this problem and will increase the specificity and sensitivity of the method. The effect of incubation time and temperature on the measured inhibitor titre is shown in Fig. 3a,b. The experiments were performed using a purified inhibitor directed towards A2 and C2 domain [17] diluted in FVIII-deficient plasma. At 37°C, an optimal inhibitor titre is reached after 120 min of incubation (Fig. 3a). At incubation times more than 180 min, a marked decrease of FVIII activity is noticed even in the control sample (Fig. 3b) rendering the inhibitor data unreliable at longer incubation times. In contrast, at room temperature the FVIII activity in the control mixture remains stable up to 240 min (Fig. 3b) whereas, in the test mixture, the remaining FVIII activity does not reach a stable level in this period because of slow-acting progressive inhibitor activity.