Anticoagulant therapy in patients with congenital FXI deficiency

The relevance of factor XI (FXI ) deficiency has undergone significant change since its discovery in 1953 by Rosenthal et al¹  in a Jewish family in the United States. FXI deficiency was thought to be similar to hemophilia and therefore was named “hemophilia C.”¹  Nevertheless, further evaluation revealed that bleeding is usually mild and more commonly occurs at sites of injury and/or highly fibrinolytic tissues.^2,3  Moreover, unlike hemophilia, the bleeding phenotype of FXI deficiency is unpredictable. Neither FXI antigen nor activity is a prognostic marker.^4-6 

On the other hand, FXI deficiency confers antithrombotic protection. This fact has been proved in animal models and human cohorts.^7-9 Preis et al⁹  identified 1235 FXI-deficient subjects in Israel and demonstrated that they had lower incidence of thrombosis than nondeficient individuals. These observations had encouraged the development of FXI inhibitors (anti-FXI), which may offer the opportunity to prevent major bleeding by dissociating “pathological” from “hemostatic” coagulation.^10,11 

Given the aforementioned considerations, it has been suggested that anticoagulant therapy might not represent such a high-risk factor for bleeding in FXI deficiency. Indeed, a review suggests that FXI-deficient patients can be safely treated with vitamin K antagonists (VKA).¹²  However, no systematic study has investigated bleeding in FXI-deficient individuals on anticoagulant therapy. Besides, there is no information on the effect of FXI deficiency in VKA dosing or management, nor on the use of direct oral anticoagulants (DOACs) in this setting either.

This study was conducted in a cohort of 269 white subjects with congenital FXI deficiency (FXI:C<70%) from 2 Spanish centers (Yecla/Murcia: N = 238, Lugo: N = 31) collected from 1995-2021. The Yecla study cohort is described elsewhere.¹³ 

Plasma activated partial thromboplastin time and FXI:C were determined in an automatic coagulometer (SynthASil™ and FXI-deficient plasma; Instrumentation Laboratory). Sodium dodecyl sulfate–polyacrylamide gel electrophoresis and western blot of plasma FXI were conducted with a specific antibody (GAFXI-AP190R1; Enzyme Research Laboratories). Sequencing of the exons and flanking regions of the F11 gene by Sanger or next-generation sequencing and analysis of structural variants by MLPA were carried out in cases displaying FXI deficiency.

We selected patients with chart-based evidence of anticoagulation during the study period. Subjects treated exclusively with prophylactic-dose heparin or antiplatelets were excluded. All included subjects gave their informed consent, which was approved by the Ethics Committee of Hospital Morales Meseguer and performed in accordance with the Declaration of Helsinki.

Four investigators (J.E., M.J.S., E.F.-M., and A.L.) independently extracted data from medical records using a standardized form. Such data were independently checked by 2 investigators (C.B.-P. and M.E.d.l.M.-B.). Final follow-up date was June 15, 2021. Variables regarding anticoagulants comprised indication, drug, dosing, and length of therapy. Bleeding was graded according to World Health Organization (WHO) scale. In the case of VKA, the international normalized ratio (INR) and the time in therapeutic range (TTR) were collected. Statistical analysis was performed with STATAv.16 (StataCorp-LLC).

Overall, we identified 269 FXI-deficient subjects carrying 22 different mutations. Only 15 had severe deficiency (FXI:C < 20%). This cohort exhibited a mild bleeding tendency. Fifty-two cases had bleeding history (19.3%; 95% confidence interval [CI]: 15.1% to 24.5%), and a high proportion of episodes occurred at sites of injury (19/52, 36.5%; 95% CI: 24.8% to 50.1%).

Thirteen subjects (4.8%; 95% CI: 2.8% to 8.1%) developed a cardiovascular event (myocardial infarction: N = 8, stroke: N = 3; transient ischemic attack: N = 2), whereas only 1 subject (0.3%, 95% CI: 0.1% to 2.1%) had a venous thrombosis. Prior works indicate that FXI-deficient patients seem to be best protected from venous thromboembolism.⁹ 

Fifteen cases with FXI deficiency required anticoagulant treatment (Table 1). Seven patients (46.7%) were women. Median age of this group was 70 years (interquartile range [IQR]: 64.5 to 79 years). All subjects had mild to moderate deficiency, with a median FXI:C of 40% (IQR: 33% to 48.5%).

We identified 8 different F11 variants in these individuals. All mutations caused quantitative FXI deficiency (cross-reacting material-negative, CRM ⁻) in heterozygous state, except for p.P538L, which caused a qualitative deficiency (CRM⁺) in homozygosis. Two of 15 patients had positive bleeding history before starting anticoagulation (13.3%, 95% CI: 3.7% to 37.9%), in both cases, these were mild events secondary to injury.

The main indication for anticoagulants was atrial fibrillation (AF, 12 of 15). Fourteen patients started therapy with VKA, all with acenocoumarol, but 4 subjects were on DOAC at the end of follow-up (3 sequentially, after the VKA, and 1 as frontline). All DOACs were anti-FXa (Table 1).

Median length of anticoagulation per patient was 40.4 months (IQR: 11.4-90 months). Over 1020.2 months of cumulative therapy, only 2 bleeding episodes in 2 patients (13.3%; 95% CI: 3.7% to 37.9%) were recorded. Both events were mild (WHO grade 1). Remarkably, no major/fatal bleeding was reported. It is also important to note that 2 patients received short courses of treatment-dose heparin prior to initiating oral anticoagulants, and that 4 cases were simultaneously taking antiplatelets, all of them with no evidence of bleeding. Interestingly, the “pre-post” frequency and severity of hemorrhage did not change despite the start of anticoagulation. We mention P4, 1 of the 2 subjects who bled after anticoagulant start, because he had low FXI:C (24%) and prior hemorrhagic history. The “pre-post” characteristics of bleeding were similar in this patient.

As regards VKA, median dose of acenocoumarol was 10.75 mg/wk (IQR: 8-16 mg/wk), and median TTR was 70% (IQR: 60% to 81%). Only 2 patients had TTR < 65%; in both cases it was due to INR < 2, and this was the reason for switching to a DOAC. VKA dosing and TTR did not differ significantly from the standards and seemed to be unaltered by FXI deficiency.

Patients with congenital coagulopathies are relatively protected against thrombosis, but they can eventually develop age-related thrombotic complications or AF.¹⁴  Albeit crucial, anticoagulation might result in exacerbated hemostatic disequilibrium toward major hemorrhage. Because there is a lack of evidence-based guidelines on this scenario, the experience of exceptional cases can be of great value.¹⁵ 

In some bleeding disorders, beyond hemophilia A/B, the evidence is practically absent. This is the case of FXI deficiency. In fact, to the best of our knowledge, there are only 2 recent reports on FXI-deficient patients treated with anticoagulants, a single case and a small series (N = 5), all with VKA.^16,17  Interestingly, in parallel with our results, the only patient from these reports who had bleeding on anticoagulation also had prior history of hemorrhage, and both “pre” and “post” episodes had similar features.¹⁷ 

Our observations may be highly relevant to the fascinating possibilities opened by anti-FXI.^18,19  Noteworthy, only 15 subjects from our cohort (5.6%), and none of the 15 patients with severe deficiency, had thrombotic events requiring anticoagulants. Moreover, because the grade of suppression achieved by anti-FXI is compatible with moderate deficiency, and it may not substantially increase the bleeding risk over that conferred by VKA or DOAC, the addition of anti-FXI on standard anticoagulation would be beneficial when required without accentuating bleeding risk.^10,11 

In conclusion, despite the retrospective analysis and the low sample size owing to a rare disease, we report the largest descriptive study on anticoagulation in FXI deficiency, and the first cases receiving DOAC. Our clinical observations suggest, in consonance with prior reports, that moderate FXI deficiency does not seem to significantly interfere with anticoagulant management nor bleeding risk. Future works should focus on this issue, particularly to evaluate patients with severe deficiency, in whom our results cannot be extrapolated, and confirm if some variables, such as prior history of bleeding, can guide clinical decisions. Finally, our findings provide further evidence that may be connected to the fascinating possibilities opened by FXI inhibition.

This work was supported by Instituto de Salud Carlos III and FEDER (PI21/00137) and Fundación Séneca (19873/GERM/15).

M.E.d.l.M.-B. has a postdoctoral contract from University of Murcia, Spain. C.B.-P. has a Río Hortega fellowship from Instituto de Salud Carlos III (CM20/00094).

Contribution: J.E., M.J.S., E.F.-M., E.F.-T., A.L., and M.C.-S. recruited all patients, were responsible for clinical and coagulation data, and reviewed medical charts; C.B.-P. and M.E.d.l.M.-B. independently revised extracted data; C.B.-P., M.E.d.l.M.-B., J.P., and A.M. were responsible for molecular, functional, and biochemical characterization of FXI deficiency; J.E., V.V., and J.C. designed the research plan and coordinated the experimental work; C.B.-P., M.E.d.l.M.-B., V.V., and J.C. interpreted the results and drafted the manuscript; and V.R. contributed to the critical revision of anticoagulant data.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Javier Corral, Centro Regional de Hemodonación, Universidad de Murcia, Calle Ronda de Garay s/n, 30003 Murcia, Spain; e-mail: javier.corral@carm.es; and María Eugenia de la Morena-Barrio, Centro Regional de Hemodonación, Universidad de Murcia, Calle Ronda de Garay s/n, 30003 Murcia, Spain; e-mail: uge2985@hotmail.com.