A case of acquired factor XI deficiency and severe bleeding tendency associated with Streptococcus pyogenes cellulitis

Factor XI (FXI) is a 160-kDa glycoprotein, circulating in the plasma as a homodimer zymogen, complexed to H-kininogen (HK).1 It plays an early role in the contact activation phase of blood coagulation, mainly after being proteolyzed by thrombin on platelet membrane surfaces. FXI deficiency can be either congenital or seldom acquired.2 The acquired form of FXI deficiency occurs when an inhibitory antibody is developed by a dysregulated immune system. To our knowledge, acquired FXI deficiency has been associated with malignancy,2, 3 autoimmune diseases2, 4, 5 and coronavirus infection,6 but not previously reported with bacterial infections. Therefore, we report an acquired FXI deficiency case associated with a Streptococcus pyogenes infection, resulting in severe hematomas, and the corresponding management in the absence of FXI concentrate in our country. We report herein the case of a 77-year-old man with hypertension and atrioventricular block, who was admitted to our hospital with right upper extremity cellulitis, originating from a traumatic wound of the thumb (Figure 1A). Initial laboratory results showed left-shift leucocytosis, thrombocytosis and high C-reactive protein (CRP) level. A minor surgical intervention for debridement and decompression was performed; bacterial culture grew S. pyogenes (A) for which the treatment with penicillin G was initiated. A few days later, there was a continuous drop in haemoglobin (Hg) level and CT scan showed a large upper extremity haematoma measuring 15 × 3 × 5 cm. Furthermore, the patient's laboratory results showed de novo 1.5-fold prolongation of activated partial thromboplastin time (aPTT) of 51 s from a normal baseline of 34 s (28–37 s). Further investigation revealed low FXI level of 13% (60%–120%), FXII of 27% (60%–150%) and an elevated FVIII of 192% (60%–160%). In the absence of bleeding history and a previously normal aPTT, we postulated that bleeding occurred due to acquired FXI deficiency. FXI deficiency was further confirmed by in vitro mixing studies compatible with aPTT correction and increase in FXI levels (Table 1). A 2-h incubation period at 37°C failed to demonstrate the presence of in vitro neutralizing antibodies, but the persistently prolonged aPTT and the low levels of FXI, despite plasma infusions, supported the presence of an inhibitor. All other coagulation tests including antiphospholipid antibodies and platelet function analysis were within normal limits. The patient received targeted antibiotic therapy to control the infection and plasma infusion (Octaplas® 8 IU in total) to control the bleed. To eradicate the FXI inhibitor, steroids at 1 mg/kg/day were initiated. Subsequent daily measurements of aPTT and FXI levels revealed little response to plasma infusion, suggesting a rapid in vivo clearance of FXI. The addition of two doses of recombinant activated factor VII (rFVIIa) at 90 μg/kg was required to achieve haemostasis. Three days later, the patient developed a new large retroperitoneal hematoma measuring 25 cm with multiple foci of active bleeds, requiring urgent angioembolization twice. Thus, rFVIIa was stopped and activated prothrombin complex concentrate (aPCC) was started at 50 IU/kg/day and increased to 200 IU/kg/day (total dose 43 000 IU) for haemostasis control. Intravenous immunoglobulin (IVIG) at 1 g/kg/day for 3 days was also introduced to suppress the inhibitor formation. Unfortunately, while receiving aPCC, D-Dimer levels suddenly increased from 1.7 to 68 (<0.5 mg/L), explained by multiple small venous thrombi along the subclavian and superficial veins of the left arm. Since extensive haemostatic support is known to carry a risk for thrombosis, aPCC was discontinued. Low molecular weight heparin (LMWH) was introduced cautiously with a starting dose of 17 IU/kg increased to 45 IU/kg ×3 tinzaparin, for a total of 8 days, until a complete resolution of thrombosis. During follow-up, steroids and IVIG doses were tapered, rituximab was added at a dose of 375 mg/m2/week for 3 weeks (Figure 1B). Acquired FXI deficiency in association with a bacterial infection caused by S. pyogenes is a new observation. Takeyama et al. reported an acquired FVIII deficiency in a child after a streptococcus infection treated with penicillin.7 The role of S. pyogenes in coagulation and fibrinolysis was extensively reviewed by Loof et al. Briefly, S. pyogenes autoactivates the contact pathway, FXII, leading to the activation of FXI on the bacterial surface, triggering the coagulation process.8 Indeed, our patient had low FXII in addition to low FXI. In terms of molecular mimicry, an antibody directed against a streptococcal antigen may cross-react with a structural protein, aka FXI in our case, causing destruction or alteration of its function.9 Extensive studies on rheumatic fever and post-infectious glomerulonephritis showed that autoantibodies target surface endothelium, laminin and N-acetyl-glucosamine. FXI is a highly glycosylated protein of the contact system, with a strong interaction with laminin, along with HK and FXII, leading to thrombin generation after blood vessel injury.10 Similarity in amino acid sequences of S. pyogenes and FXI may explain the anti-FXI inhibitor formation. Interestingly, with an FXI of 13%, the bleeding could have been more extensive, but haemostasis was likely compensated by the FXI content in platelet α granules,11 and the reactively elevated VWF and FVIII levels.1 Moreover, when the patient was given a high-dose aPCC, he developed multiple small venous thrombi, but there was a rapid resolution of thrombosis, in less than 10 days, under low doses of LMWH, possibly explained by the potent fibrinolytic capacity of streptokinase, secreted by S. pyogenes,8 even after infection resolution. From a laboratory perspective, mixing studies did not demonstrate the presence of an inhibitor. Such in vitro results are not unusual, especially since aPTT measurement is a functional assay affected by various factor levels.12 The absence of the in vitro correction of FXI levels/aPTT could be explained by different epitopes on FXI, elevated FVIII and an unexpected antibody clearance. Unfortunately, an FXI inhibitor level assay was not available for us for further investigation. In conclusion, treatment of any acquired coagulopathy requires immediate haemostatic control of the bleed, eradication of the inhibitor and effective immunomodulation. Treating such complex cases is challenging, especially that FXI replacement therapy is not widely available, and the administration of other therapeutic alternatives may carry a thrombotic risk. Timea Szanto, Mirka Sivula and Anna-Elina Lehtinen treated the patient and followed the clinical case. Nancy El Beayni wrote the manuscript, and all the authors contributed to the final version. Riitta Lassila oversaw the management, and Timea Szanto supervised the clinical/manuscript work. This work has been made possible by support of an ISTH Reach-the-World fellowship. We acknowledge Professor Jan Voorberg and Professor Joost Meijer for their valuable insights. No conflicts of interest declared. As a case report, no formal ethical approval was necessary. However, informed, written consent was obtained from the patient. All data and materials are available from the corresponding author upon reasonable request.