Elevated FVIII levels in hereditary hemorrhagic telangiectasia: Implications for clinical management

1 INTRODUCTION

Hereditary hemorrhagic telangiectasia (HHT) is an autosomal dominant inherited vascular dysplasia with an approximate prevalence of 1 in 6000 individuals.^{1, 2} It is characterized by mucocutaneous telangiectasias and arteriovenous malformations (AVMs) in internal organs. The telangiectasias in HHT increase in number and distribution with age, and characteristic locations are the face, fingers, and mucous membranes of the oral cavity, nose, and gastrointestinal tract (GIT). AVMs most often occur in the pulmonary, cerebral, and hepatic circulations.³

Epistaxis is the most common and debilitating symptom for patients with HHT^4-6 and varies in frequency, intensity, and duration. Chronic blood loss from the GIT is also a prominent feature in HHT, as is iron deficiency anemia.⁷ Most HHT cases are associated with genetic mutations in one of three genes—ENG, ACVRL1—and, more rarely, SMAD4.^{8, 9} The diagnosis of HHT is clinical and based on four criteria: (1) first-degree relative with a diagnosis of HHT, (2) mucocutaneous telangiectasias, (3) frequently recurring epistaxis, and (4) AVM(s). Three or four positive criteria are conclusive, two positive criteria are suspected, and one or zero is unlikely.⁴

Although HHT causes chronic bleeding, it is not a coagulation disorder and does not protect against thromboembolism. On the contrary, patients with HHT may have an increased rate of venous thromboembolism (VTE), including deep vein thrombosis (DVT) and pulmonary embolism (PE), compared to the general population.¹⁰ Previous studies^{10, 11} have revealed that many patients with HHT have elevated coagulation factor VIII (FVIII) levels, especially if concurrent low iron levels are present. Elevated levels of FVIII are a significant predictor for first and recurrent VTE,^{12, 13} and the thromboembolic risk increases with higher levels of FVIII.^12-15 Causes of transient and sustained high levels of FVIII are shown in Table 1.^16-25

The treatment for epistaxis in HHT is both surgical and non-surgical. Among the non-surgical options are: Tranexamic acid (TXA), Thalidomide, hormonal therapy such as anti-estrogens (e.g., Tamoxifen and Raloxifene), and Bevacizumab, a vascular endothelial growth factor inhibitor, all of which list thromboembolic risk as a possible side effect.^26-30

The Second International Guidelines for Diagnosing and Managing HHT³¹ recommends TXA (up to 1.5 g three times daily) as a second-line treatment for HHT-related epistaxis if topical moisturizing therapy is inadequate. The Guideline supplementary document cites absolute contraindications to this treatment³¹ as recent VTE or arterial thrombosis, while atrial fibrillation, thrombophilia, or other pro-coagulant tendencies (e.g., elevated FVIII) are considered relative contraindications. In spite of elevated levels of FVIII being proposed as a relative contraindication to this treatment, no screening strategy for FVIII has been proposed.

The objective of this study was twofold: to determine the prevalence of arterial and venous thromboembolic events in the Norwegian HHT population, and to explore potential factors linked to such events, with particular emphasis on FVIII.

2 METHODS AND STUDY PARTICIPANTS

Patients with an HHT diagnosis attending the Otorhinolaryngology Department at Oslo University Hospital—Rikshospitalet were included consecutively between April 1, 2021, and November 30, 2022. This Department forms part of the national HHT Center of Excellence in Norway, receiving referrals from the whole country. The inclusion criterion was a diagnosis of HHT according to the diagnostic criteria.⁴ Genetic testing was performed. Patients suspected of having HHT (two positive criteria) in addition to a pathogenic gene variant were included.

The participant's medical history was recorded and included previous or active thromboembolic disease and smoking. Patients were asked explicitly about DVT and PE, myocardial infarction (MI), and thromboembolic stroke. Thromboembolic disorders other than these were recorded as “other.” The participants' recollection of the time of disease was noted, and their medical record was reviewed for accuracy. Thus, the patient's electronic medical record confirmed all recorded thromboembolic events. Recent events that could be a risk factor for thrombosis or high FVIII were stressed.

The epistaxis grade was evaluated using the Epistaxis Severity Score (ESS)³² and the Epistaxis Intensity, Frequency, and Need for Blood Transfusion Score (IFT).^{33, 34}

Blood tests were collected with each consultation or treatment and screened for coagulation factors VIII and IX, vWF antigen and activity, hemoglobin, ferritin, iron, and C-reactive protein (CRP). However, the measurements of FIX and vWF were not performed for all patients. Hospital laboratory blood testing and analysis were not set to a specific time during the day as patients travel from all parts of the country, and compliance would not have been feasible. Testing and analysis were therefore performed whenever possible but not more than 2 days after the consultation or treatment. Measurements of consecutive blood samples from the same participants were included in the study material if obtained within the study period. Screening for pro-coagulant factors other than those mentioned above was not performed.

The participant's medical prescription history was reviewed through the national prescription database, and prescriptions for TXA and iron supplements were noted if prescribed within the last 3 years.

3 RESULTS

4 DISCUSSION

Epistaxis is the most common and debilitating aspect of HHT. The challenge for an otolaryngologist is in providing the most optimal treatment for epistaxis after considering the severity of the episodes, quality of life, blood work, and ongoing treatment for co-existing illnesses. The high prevalence of thromboembolic events in HHT patients raises concern about treatment safety, keeping in mind that most medications used to treat epistaxis in HHT may give thrombosis as a potential side effect. The Second International Guidelines for Diagnosing and Managing HHT patients recommend treatment with TXA at high dosages for epistaxis after excluding the risk of thromboembolic disease, including high FVIII levels³¹ (see Guideline supplement).

TXA is an antifibrinolytic agent that stabilizes and prevents premature dissolution of the fibrin clot. TXA is a competitive inhibitor to the lysine receptor found on plasminogen, reducing the conversion of plasminogen into plasmin. Plasmin degrades fibrin clots, fibrinogen, and other plasma proteins, including the pro-coagulant factors V and VIII.³⁵

Two randomized controlled trials^{36, 37} have demonstrated a significant decrease in the severity and frequency of epistaxis episodes in HHT patients treated with TXA. No increased risk of thromboembolic events using TXA was reported. However, the trial period of both clinical studies was limited to 3 months. Information about the long-term use of TXA in HHT patients is lacking. However, a recent study involving 24 patients treated with TXA or aminocaproic acid (an antifibrinolytic agent) for 13 months did not report any thromboembolic events.³⁸ A retrospective study³⁹ involving 29 patients using TXA reported no difference in the thrombosis rate before and during treatment. Although none of these studies measured FVIII, there is no indication that elevated FVIII levels, TXA use, and thromboembolic events are correlated. In addition, TXA is shown neither to increase thrombin generation nor to lead to hypercoagulability.^{40, 41}

Despite the evidence that TXA is a safe treatment for epistaxis in HHT patients, skepticism about its possible side effects remains.^{31, 42} Although absolute and relative contraindications to this treatment are suggested, an individual assessment of thromboembolic risk before initiating treatment is not explicitly required or recommended. Many patients with HHT have elevated levels of FVIII, a known risk factor for thromboembolic disease, especially VTE.^{12, 13} The higher the level of FVIII, the higher the risk of developing VTE.^14-17 We aimed in this study to identify factors associated with increased risk of thromboembolic events in HHT patients, including FVIII.

The risk of VTE in HHT patients has been comprehensively described by Shovlin et al.¹¹ However, the total prevalence of thromboembolic events (both arterial and venous) in HHT patients has not been previously investigated. Our study showed that the prevalence of VTE events among our HHT population was 12.7%, while 16.5% were arterial events. These prevalence numbers are much higher than in the general population, which for VTE in Norway, is estimated to be 0.24%.⁴³ Of the registered thromboembolic events in our study, 9.0% were thromboembolic strokes, and 3.0% were MI. The period prevalence of stroke (including hemorrhagic stroke) and MI in the general population in Norway between 2012 and 2020 was 0.26% (95% CI: 0.26–0.26) and 0.30% (95% CI: 0.30–0.30), respectively.⁴⁴ These observations confirm that thromboembolic events are a significant complication in HHT and that arterial thromboembolic events are an important concern.

Over two-thirds (68.3%) of our patients had high levels of FVIII, and most HHT patients with a history of thromboembolic events had high levels of FVIII (93.3%). A previous study by Livesey et al.¹⁰ showed that low serum iron levels correlate with high FVIII levels and VTE. Elevation of FVIII to promote thrombosis in the setting of iron deficiency was proposed as a plausible mechanism.¹⁰ Although our data initially suggested a correlation between low iron levels and thromboembolic events, this was not reflected in the regression analysis. Rather, FVIII and age were the factors that significantly associated with thromboembolic events. However, it should be borne in mind that iron levels are influenced by iron supplementation as well as diurnal changes. An alternative explanation for elevated levels of FVIII in HHT patients could be that chronic vascular inflammation causes endothelial cell activation. This is because Pentraxin 3 is shown to be significantly elevated in HHT patients, reflecting systemic inflammation and, at least partly, vascular inflammation.⁴⁵

HHT patients with PAVM are known to be at a higher risk of thromboembolic stroke due to the lack of first-pass filtration by the lungs,⁴⁶ and our results are in line with this concept.

The level of FVIII can fluctuate over time. However, high levels of FVIII are known to persist in patients with thrombosis and are generally not caused by acute-phase reactions.^{14, 15, 47, 48} The acute phase reactants CRP and ferritin were not correlated with the FVIII level in our study. Among our patients whose FVIII levels were measured at regular intervals, 75% had consistently high or normal levels of FVIII. The remaining 25% had fluctuating levels of FVIII. Of those with consistently high levels, 33% had experienced thromboembolic events, while 18% of those with fluctuating levels had experienced the same. No patients with consistently normal levels of FVIII reported having a thromboembolic event. This suggests that relying on a single measurement of FVIII in individuals with HHT may not be sufficient and that the test should be repeated consecutively. The intention of this study was not to prove causality, and the study design would not be sufficient for this purpose. However, we see a clear tendency where the patients with previously reported thromboembolic event(s) also had sustained high FVIII levels.

Among the limitations of our study, screening for other pro-thrombotic factors beyond those already mentioned was not performed. Thus, whether co-existing pro-thrombotic factors influenced the recorded thromboembolic events in those included remains unknown. Two patients with Factor V Leiden mutation were included in this study. This decision to include these was made to reflect the general population in which this mutation is estimated to be present in 3%–8% of caucasians.⁴⁹ The only potentially pro-thrombotic medications our subjects used were TXA and intranasal Avastin (Bevacizumab). Intranasal Avastin is not known to cause thrombosis mainly due to the low concentration used and the longer intervals between treatments.^{50, 51} The retrospective nature of the assessment of thromboembolic events is a potential weak point. ABO typing is relevant to FVIII measurements as non-O blood group patients have higher levels of FVIII than those with type O.^23-25 Blood typing was not performed in this study and, therefore, not adjusted for. The timing of blood sampling was not standardized, which is a limit because iron levels vary during the day. Our overview of the participants' prescription history for TXA and iron supplements was insufficient due to limited access to the patient's medical and prescription history.

5 CONCLUSION

Both arterial and venous thromboembolic events are highly prevalent in the Norwegian HHT population. High FVIII levels and age are important risk factors for thromboembolic events in HHT. Hemoglobin, iron, and ferritin are unreliable indicators for the level of FVIII and the thromboembolic risk. FVIII levels can fluctuate over time. Therefore, repeated measurements of FVIII are important to assess present levels accurately. However, it is unclear whether the FVIII level should be taken into consideration when considering the use of medications that carry a potential risk of thrombosis.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.