SummaryPatients with haemophilia requires different amounts of FVIII to prevent and treat bleeds. We hypothesise that this is because FVIII has variable effects on individual patients' global haemostasis. Twelve patients with severe haemophilia A were infused with 50 IU/kg FVIII and thrombin generation in platelet rich (PRP) and platelet poor plasma (PPP) and velocity of changing clot elasticity were measured preinfusion and at nine subsequent time points over 72 h. Despite a close correlation between median FVIII and median initial rate of thrombin generation (R 2 0AE94), endogenous thrombin potential (ETP; R 2 0AE94) and peak thrombin (R 2 0AE91) in PPP, there was wide inter-patient variability at each time point. There was, however, a highly predictable intra-patient relationship between FVIII level and thrombin generation. Inter-patient variability was due to both differences in FVIII levels and the variable effect FVIII had on an individuals' thrombin generation. The utility of PRP was limited because, at low-FVIII levels, only rate of thrombin generation was measurable. At low-FVIII concentrations, the rate of thrombin generation in PPP was the most useful test whilst at higher levels ETP and peak thrombin could also be used.

Global haemostatic assays (GHA) may be useful in the management of patients with severe haemophilia A. We studied 12 patients with baseline FVIII <1 IU/dL. After a 72 hour washout patients were infused with 50IU/kg recombinant FVIII and samples taken into citrated plasma, with or without corn trypsin inhibitor (CTI), pre- and 0.25, 1, 3, 6, 24, 34, 48, 54 and 72 hours post-infusion. At each time point FVIII was measured using a one stage and an antigenic assay. Blood taken into CTI was used to perform thrombin generation assays (TGA) in platelet rich (PRP) and poor plasma (PPP) and whole blood low-dose tissue factor thromboelastography (Rotem) (LD-TF TG). Tissue factor (TF) concentrations in PRP were 0.1 and 0.25pM and in PPP 1 and 2.5pM. At the pre-infusion time point FI, FII, FV, FVII, FIX, FX, FXI, FXII, VWF:Ag, TFPI and TAFI were measured. In PRP TGA were unrecordable pre-infusion in most patients with both a 0.1 and 0.25pM TF trigger, hence this assay was unsuitable for assessing low concentrations of FVIII. Post infusion in PRP the endogenous thrombin potential (ETP) and peak thrombin (PT) increased towards the normal range (mean ± 1SD of 30 controls) and then decreased with time. The ETP could not be measured in the majority of patients after 30 hours. The only TGA parameter in PRP that could be measured at all time points in the majority of patients was the initial rate of thrombin formation calculated by linear regression. In PPP a similar pattern was observed except that ETP and PT could be measured in most patients to a FVIII level of 1 IU/dL, whilst initial rate could be measured at levels <1IU/dL . Variability of TGA was such that no parameters correlated with FVIII level between patients, however, within an individual there was a very strong correlation between FVIII level and lag time, ETP, PT and initial rate. This suggests that underlying global haemostatic potential varies widely between patients and that a certain level of FVIII will correct haemostasis to a different level in each patient but that within patients FVIII correlates closely with TGA. This was further demonstrated by mathematically modelled TGA parameters to predefined FVIII levels of 1, 10, 20, 50, 75 and 100IU/dL and, although there was again a wide variability between patients, there was a reproducible relationship between ETP, PT and initial rate with FVIII level. LD-TF TG could be measured in almost all patients even at baseline FVIII. The parameters that were most useful for following FVIII levels were maximum velocity of clot formation (Vmax) and time to Vmax. Similar to TGA there was a wide inter-patient variation and FVIII corrected assay parameters to variable levels. There was poor correlation between parameters of thrombin generation and the LD-TF TG assays suggesting that the two assays were measuring different aspects of haemostasis. In conclusion we have demonstrated that a defined FVIII level in individual patients corrected GHA to a variable extent but that within an individual there is a close correlation between FVIII and GHA parameters suggesting that these assays may be useful for monitoring patients during FVIII replacement therapy and give additional information to FVIII levels. At FVIII levels around 1IU/dL only the initial rate of thrombin generation could be measured in all patients and this may be a useful parameter to assess trough levels during prophylaxis. The clinical utility of these assays needs to be confirmed in studies that link bleeding endpoints to assay parameters.