Factor XIII deficiency: Lessons from two patients with unusual bleeding

Patient 1

Mrs A, a 76-year-old woman presented with swelling and discolouration over her left axilla for a week. She was evaluated elsewhere and found to have a large area of ecchymosis over the axilla and medial aspect of the left arm. An ultrasound showed features of subcutaneous haematoma, along with reduced flow in the axillary vein, leading to the diagnosis of deep venous thrombosis. She was started on enoxaparin but in view of worsening pain and swelling presented to us for further management. Examination revealed a 20 cm × 18 cm ecchymotic area over the axilla and arm, with local fullness. An ultrasound showed subcutaneous collection with no evidence of deep venous thrombosis. Enoxaparin was stopped and she was evaluated for any underlying bleeding disorder. She had a history of two deliveries in the past and a hip replacement surgery 1 year ago without any abnormal bleeding. Investigations revealed a normal platelet count, prothrombin time (PT), activated partial thromboplastin time (aPTT) and fibrinogen. A thromboelastogram showed accelerated clot lysis at 8.2%. A FXIII assay revealed undetectable FXIII activity, indicating severe deficiency. In view of her age, she was worked up for co-existent malignancy and autoimmune disorders, both of which were negative. She was given cryoprecipitates at a dose of 1 unit per 10 kg body weight, following which her bleeding stopped.

Patient 2

Mr K, a 50-year-old man presented with a history of sudden onset abdominal pain 1 year before admission. On evaluation elsewhere, he was found to have a peri-splenic haematoma with active bleeding. He underwent an emergency CT angiography, followed by splenic artery embolization. No further evaluation was done and he was subsequently well for 6 months. He presented again with discolouration over the right axilla, associated with pain. Initial examination revealed the presence of ecchymosis, and an ultrasound showed a subcutaneous haematoma. A CT angiography revealed active bleeding from a branch of the axillary artery, which was embolized by intervention radiology. Investigations revealed a normal platelet count, PT, aPTT and fibrinogen. A FXIII assay was advised, but he left against medical advice and was on follow-up locally. After 3 months, he presented to us with sudden onset of headache, vomiting and giddiness. A CT scan of the brain revealed a large cerebellar haematoma with peri-lesional oedema. A repeat work-up revealed the presence of low FXIII activity, for which he was given cryoprecipitates. He had no further episodes of bleeding and was well after that.

FXIII deficiency with no history of bleeding needs a high index of suspicion

FXIII deficiency, a rare disease, has an incidence of 1–2 per million population. The most common manifestation described is umbilical stump bleeding at birth, occurring in >80% of affected patients, followed by skin and muscle bleeds.⁷ FXIII deficiency may be either inherited or acquired, and more than 150 mutations leading to inherited deficiency have been described.⁸ Delay in diagnosis, especially in acquired forms, appears to be the norm, with data from a tertiary centre documenting a median delay of 1.5 month before a diagnosis is reached.²

Both our patients had multiple encounters with other healthcare systems before contacting us. The recommended screening test for FXIII deficiency, the urea clot lysis, was not available for the first patient, necessitating a thromboelastography for further work-up. The urea clot solubility test, which has traditionally been recommended for diagnosing FXIII deficiency, has now been reported to have low sensitivity and specificity. A number of pre-analytical variables, including the choice of lysing agents (urea, acetic acid) or clot-inducing agent (thrombin, calcium) are also known to significantly affect the test performance.⁹ The current consensus recommends FXIII activity assay or molecular characterization for a definitive diagnosis of FXIII deficiency and advises against the use of urea clot lysis.⁹ Even after initial suspicion, a universal algorithm cannot be recommended, as most mutations noted with congenital FXIII deficiency are endemic and vary across geographical locations.¹⁰ Each laboratory must therefore have its own algorithm based on the common endemic genetic markers.¹¹ Moreover, the molecular laboratory evaluating the FXIII gene mutations should remember that FXIII has multiple polymorphic variants consisting of over 25 haplotypes.¹²

Acquired forms of FXIII deficiency

In an adult patient with unexplained skin, soft tissue or muscle bleeding, acquired haemophilia is often the first suspicion, with acquired haemophilia A being the most common.¹³ Acquired deficiency of FXIII was first described in the 1970s by Lorand, who classified inhibition of FXIII based on the underlying mechanism.¹⁴ Since then, acquired FXIII deficiency has been reported to occur in immune- or non-immune-mediated settings. Immune-mediated FXIII inhibition is described in the setting of malignancies, autoimmune diseases or drugs. These patients usually present with major bleeding manifestations. On the other hand, non-immune mechanisms such as increased consumption or reduced synthesis are less likely to lead to major abnormal bleeding.¹⁵ However, it has been seen that over 50% of acquired FXIII deficiency may be idiopathic, with no specific underlying cause being identified.¹⁶

Our patients probably represent an acquired pathology, as both had a history of multiple haemostatic challenges without any abnormal bleeding. Considering their age, work-up for malignancy was particularly important and was negative for both patients. Both patients are being regularly followed up for development of any malignancy or autoimmune disease.

Contribution of FXIII to arterial and stromal function is being identified, and may explain the pattern of bleeding with FXIII deficiency

FXIII has multiple well-defined roles beyond haemostasis. The association of poor wound healing with FXIII deficiency was noted in the first patient with FXIII deficiency ever described.¹⁷ Knock-out mice for FXIII genes display impaired wound healing and abnormal scar formation.¹⁸ It has since been elucidated that FXIII acts on a number of substrates vital to stromal function, including thrombospondin, von Willebrand factor (vWF), vitronectin and osteopontin.¹⁹ These proteins are essential for the maintenance of stromal structure, wound healing and attachment of cells to connective tissue matrix. Initiation of wound healing by migration of fibroblasts to site of injury is also mediated by FXIII.²⁰ A similar effect is noted on endothelial cell function, proliferation and survival.²¹ These pleiotropic effects of FXIII may explain the unique patterns of bleeding in FXIII and FIX deficiencies. FXIII deficiency perhaps impairs matrix integrity, leading to skin and connective tissue bleeding. This is in keeping with roles of FXIII in stromal function, angiogenesis and vascular permeability.^22,23 Presentation of FXIII late in life with major arterial bleeding continues to be uncommon and may be related to the effects of FXIII on arterial function. Certain FXIII polymorphisms reduce the risk for coronary artery disease with a concurrent rise in the risk for intracerebral bleeding, indicating a fine balance of FXIII function, which may be tipped in either direction.²⁴

FXIII deficiency, though rare, is a healthcare burden

FXIII deficiency is associated with increased financial burden and reduced QoL. Registry data indicate that congenital FXIII deficiency is associated with relatively high rates of severe bleeding, seen in >48% of patients. Intracerebral bleeding, seen in >30% of patients, is the leading cause of mortality in this cohort.²⁵ The data on QoL or cost of treatment of FXIII deficiency are sparse, but both the above factors putatively lead to impaired QoL.²⁶ Patients with serious or life-threatening bleeding require prophylactic therapy to prevent further such episodes. FXIII concentrates, both plasma-derived and recombinant, have been available for the past few years. Plasma-derived FXIII (pdFXIII) has been shown to be effective in reducing bleeding rates.²⁷ Recombinant FXIII was made available in 2014 but remains in sparse use due to prohibitive costs.²⁸ Both these can be administered every 3–4 weeks, owing to the long half-life of FXIII.²⁹ However, cost and availability continue to be a concern, and most patients on prophylactic treatment have to depend on cryoprecipitate or plasma infusions. Our patients have done well on cryoprecipitate infusions every 4 weeks. The cost of extended cryoprecipitate prophylaxis has to be kept in mind, with the cost reaching upwards of US$ 100 for each infusion for a 60 kg adult in a non-subsidized setting. This is in addition to the baseline risk of infection and adverse effects associated with each transfusion.

Conclusion

FXIII is an under-recognized cause of serious or life-threatening bleeding, especially in adult patients with no history of bleeding. Considering that a majority of patients with acquired FXIII deficiency may not have an immediately recognizable aetiology, a high index of suspicion is required for diagnosis. This is more so in the setting of unusual sites of bleeding such as skin, connective tissue or other major arterial networks. Major arterial bleeding appears to be increasingly described in this setting, possibly owing to the pleiotropic effects of FXIII. More importantly, urea clot lysis should best be used as a screening tool and should be replaced by FXIII activity assay and molecular testing wherever possible.