Introduction: Recently published guidelines on antithrombotic treatment in persons with hemophilia
(PWH) appraised the literature on whether PWH are naturally anticoagulated: Only two
studies compared thrombin generation (TG) in PWH and patients on vitamin K antagonists
(VKA), both analyzing endogenous thrombin potential (ETP) and limited by arbitrary
stratifications [1]. Here, we aimed to compare TG in PWHA at varying factor (F) VIII concentrations
with patients on VKA and rivaroxaban.
Method: We sampled plasma from 1) PWHA of all severities participating in our biobank and
2) from participants on VKA or rivaroxaban of our hospital-based registry on patients
with non-valvular atrial fibrillation (AF). TG was measured by a commercially available
assay (Technothrombin, Technoclone) and FVIII levels by chromogenic substrate assay
(Biophen, Hyphen Biomed). Rivaroxaban concentrations were quantified by LC-MS/MS [2]. We regressed 1) TG parameters on FVIII and 2) INR or rivaroxaban concentrations
on TG parameters in PWHA and VKA or rivaroxaban samples, respectively. Estimating
TG parameters in 1) and INR or rivaroxaban concentrations at those values in 2), respectively,
allowed us to compare INR or rivaroxaban concentrations and FVIII in a continuous
fashion, computing 95% bootstrap confidence intervals (95% CI).
Results: We collected a total of 510 samples from 112 PWHA and 189 AF patients treated with
VKA (n=121) or rivaroxaban (n=68). The median (IQR) FVIII activity and INR values
in the respective cohorts was 13 IU/dL (3-38) and 2.0 (1.7-2.4). The median rivaroxaban
concentration measured among AF patients treated with rivaroxaban was 92.6 ng/mL (33.0-192.2).
TG curve parameters showed substantial variability among all three cohorts, and, accordingly,
their relationship with FVIII, INR, and rivaroxaban and ability to differentiate between
changes was only modest. Importantly, TG curves clustered distinctly between cohorts
([Fig. 1]). As such, different TG parameters lead to varying apparent comparison estimates.
Comparing INR and FVIII by ETP revealed an estimated mean INR equivalence of 2.1 (95%
CI, 1.9-2.3) in PWHA<1 IU/dL followed by an immediate decline and plateau ([Fig. 2]). In contrast, a comparison by thrombin peak height (TPH) lead to a mean INR equivalence
of 2.5 (2.3-2.7) at<1 IU/dL with a less steep decline. Comparing PWHA and AF patients
on rivaroxaban by ETP and TPH lead to an estimated mean rixaroxaban concentration
equivalence of 284.8 (210.0-357.7) and 229.7 ng/mL (180.6-280.4) at<1 IU/dL FVIII
activity, respectively.
Fig. 1
Thrombin generation parameters by sample cohort. Thrombin generation parameters showed distinct clustering between the three cohorts
of PWHA (blue), AF patients treated with VKA (red), and AF patients treated with rivaroxaban
(yellow). Abbreviations: AF, atrial fibrillation; ETP, endogenous thrombin potential; PC (%),
principal component (explained variation); PWHA, persons with hemophilia A; VI, velocity
index, VKA, vitamin K antagonists; TPH, thrombin peak height; TTP, time-to-peak.
Fig. 2
Relationship of ETP and TPH with FVIII, INR, and rivaroxaban. Scatter plots displaying the relationship of ETP and TPH with FVIII in PWHA (blue);
with INR in VKA-treated AF patients (red); and with drug concentration in rivaroxaban-treated
AF patients (yellow). The rightmost column shows mean INR- (purple) and rivaroxaban-equivalence
(green) for increasing FVIII activity levels. Error bars/bands display 95% confidence
intervals (CI). Abbreviations: AF, atrial fibrillation; ETP, endogenous thrombin potential; FVIII,
factor VIII; INR, international normalized ratio; TPH, thrombin peak height; ρ, Spearman’s rank correlation; τ, Kendall rank correlation.
Conclusion: TG showed substantial variability and differed distinctly between PWHA, VKA-treated,
and rivaroxaban-treated AF patients. Our data was incompatible with a guideline-concluded
threshold of 10 IU/dL for indicating natural anticoagulation in PWHA similar to therapeutic
INR in VKA patients. Our study suggests comparisons by individual TG parameters between
PWHA and anticoagulated patients to be overly simplistic.
