Haemophilia A is an X-linked disorder characterized by congenital deficiency of factor VIII (FVIII). The FVIII protein is secreted as a mixture of a single chain and heterodimer consisting of one heavy and one light chain. The heavy chain is comprised of the A1, A2 and B domains, whereas the light chain contains the activation peptide, A3, C1 and C2 domains [1]. Severe haemophilia A is defined by a FVIII activity level <1% of normal. Moderate and mild disease are defined by FVIII activity levels between 1% and 5%, and 5% and 40% of normal respectively [2]. Haemorrhage in severe haemophilia A may be spontaneous, whereas in mild and moderate disease, bleeding is usually a result of surgery or trauma. Patients are treated with either recombinant or plasma-derived FVIII concentrates which can induce the formation of inhibitory antibodies. The crude incidence of inhibitors in patients with severe haemophilia A is approximately 23%, whereas inhibitors in mild and moderate disease occur much less frequently [3]. When patients with mild and moderate haemophilia A (MMHA) develop inhibitors, antibodies are produced against the exogenous FVIII and occasionally against endogenous FVIII as well [4]. When both exogenous and endogenous FVIII are inhibited, the FVIII activity level falls to <1% which can increase the baseline bleeding frequency.



Two adult patients with moderate haemophilia A seen at Vanderbilt University Haemostasis clinic who developed inhibitors were selected. Information on the clinical course during the presence of an inhibitor was obtained from the electronic medical record.

Domain mapping ELISA

Plasma specimens were available in the IRB approved inhibitor bank at The Emory University Haemophilia Treatment Center. Domain-specific anti-FVIII antibody mapping was carried out by direct ELISA in half-area 96-well plates using purified single human domain hybrid FVIII proteins as test antigens and B domain deleted (BDD) human FVIII and BDD porcine FVIII as positive control antigens. Patient plasma was diluted 1–20 in blocking buffer (0.15 m NaCl/20 mm HEPES/5 mm CaCl2/0.05% sodium azide/0.05% Tween-80/0.25% bovine serum albumin, pH 7.4) and then serially diluted down the ELISA plate. Domain specificity was evident by visual inspection of colour [5].


Patient 1

A 56-year-old man presented to establish care in comprehensive haemophilia clinic. He was diagnosed with FVIII deficiency in the 1980s and subsequently received factor replacement prior to invasive procedures. He had chronic hepatitis C infection presumed from blood product exposure. One month prior to presentation to clinic, he underwent a radical prostatectomy at another facility for a recently diagnosed prostate adenocarcinoma. Preoperative labs were significant for a prothrombin time (PT) of 16.4 s, INR 1.5 and partial thromboplastin time (PTT) of 63 s. He did not receive factor replacement prior to surgery. His postoperative course was complicated by hypovolemic shock felt secondary to mechanical bleeding. Following surgical intervention, the patient experienced continued oozing of blood from the surgical site. A haematology consultant confirmed his congenital FVIII deficiency. The patient's FVIII activity level at that time was not reported in the records received from the outside facility. He was treated with recombinant FVIII and blood products without cessation of bleeding. A FVIII inhibitor was suspected and confirmed. Recombinant activated FVII (rFVIIa) successfully treated his bleeding.

Five weeks postoperatively, the FVIII activity level was 2% and the factor VIII inhibitor was measured at 16 Bethesda units (BU). Twelve weeks later, the inhibitor titre had fallen to 1 BU. He was taking rFVIIa for continued minor oozing from the prostate bed, but was not suffering from any spontaneous bleeds.

Factor VIII gene sequencing revealed a missense mutation, Arg593Cys, in the A2 domain. There are approximately 50 reported cases of the Arg593Cys mutation in the haemophilia A mutation database ( and it has been previously associated with inhibitors. Evaluation of this patient's inhibitor showed a polyclonal response with the largest quantity of antibodies being directed against the A2 and C2 domains of FVIII (Fig. 1).


Figure 1. Domain mapping ELISA. Plasma from patient 1 was serially diluted and the binding to human FVIII, porcine FVIII and human/porcine hybrid proteins was compared. The human/porcine hybrids each contain a single human domain (A1, A2, activation peptide (ap), A3, C1 and C2) with the remaining domains being porcine in sequence.

Download figure to PowerPoint

Patient 2

A 20-year-old man with moderate haemophilia A (FVIII activity 3%) had previously received factor infusions only for trauma-related bleeds. Approximately 18 months prior to presentation, a FVIII inhibitor of 10.2 Nijmegen-BU was detected on routine screening labs. The patient was not experiencing any change in the pattern of bleeding and on repeat evaluation 6 months later, the inhibitor was not detected. The patient presented with complaints of right-sided abdominal pain. CT scan was remarkable for a complex mass in the right lower quadrant and right hydronephrosis. Preoperative PTT was 54.3 s and the patient received 100% FVIII correction prior to surgery. Intraoperative exploration revealed localized perforation and inflammation of the right colon with haemorrhage into the wall. He underwent emergent right hemicolectomy and ureteral stenting. There was no histologic evidence of acute appendicitis, thus a pseudotumor with erosion into the colon was suspected. Postoperatively, the patient received 50% FVIII dose every 12 h without correction of his PTT. A mixing study was consistent with an FVIII inhibitor (Table 1), but titre was 0 by Bethesda assay. On postoperative day 5, the patient experienced oozing from the wound. Local pressure, rFVIIa and increased doses of FVIII were used and haemostasis was achieved. The patient was discharged, but returned 2 days later with a massive gastrointestinal haemorrhage from a vessel at the anastomotic site. Inhibitor titre was 32 BU. He was treated with endoscopically placed vascular clips, desmopressin (DDAVP) and activated prothrombin complex concentrate (aPCC) followed by rFVIIa. The bleeding resolved and the patient was discharged on prophylactic rFVIIa. Evaluation of the patient's inhibitor 5 months later revealed a titre of 3.6 BU; in 13 months the inhibitor titre was 0 BU.

Table 1. Mixing study at diagnosis of FVIII inhibitor
 PatientPooled normal plasma0 h at 37°C1 h at 37°C2 h at 37°C
  1. PTT partial thromboplastin time (seconds).


Factor VIII gene sequencing revealed a missense mutation, Arg1941Gln, in the A3 domain. According to a recent search of the haemophilia A mutation database, there have been more than 10 previous reports of this mutation. The Arg1941Gln mutation has not been associated with inhibitor formation in previous reports.


The presented cases highlight important treatment and patient-related risk factors for development of an inhibitor. As patients with MMHA are typically exposed to exogenous FVIII at an older age with intensive treatment related to surgery or trauma, they are more likely to develop inhibitors well into adulthood [6, 7]. The risk of inhibitor development after intensive treatment is greater in older patients [8]. This is in contrast with patients with severe haemophilia A that typically develop inhibitors in childhood. Development of an inhibitor in MMHA typically presents as a change in bleeding pattern or as bleeding not responsive to factor replacement.

The majority of the FVIII mutations in MMHA are missense mutations. The Arg593Cys mutation seen in our first patient is located in the A2 domain and has specifically been identified as a high-risk mutation for inhibitor formation [6]. In the haemophilia A mutation database, patients with Arg593Cys had similar FVIII antigen and activities levels pointing towards poor secretion of a functional protein. In addition, Roelse et al. using a heterologous expression system found that an Arg593Cys substitution led to elevated accumulation of intracellular functional FVIII relative to wild-type FVIII [9]. The patient with the Arg593Cys had a polyclonal response that cleared both endogenous and exogenous FVIII. In a recent case–control study, another mild to moderate mutation in haemophilia A patients N1922S showed a trend towards increased inhibitor development [8]. The N1922S mutation has been shown to lead to a defect that results in hyposecretion of a functionally intact FVIII molecule [10]. This study reported on one patient with this mutation who also had an immune response that cleared both endogenous and exogenous FVIII. Suggesting that low circulating FVIII antigen levels may be insufficient to maintain immunologic tolerance. In contrast, patient 2, with the Arg1941Gln mutation had an immune response isolated to the structural region of the FVIII affected by the mutation itself thus leaving his endogenous FVIII unaffected.

Many inhibitors will spontaneously regress, but will have a brisk anamnestic response on reexposure to exogenous FVIII, similar to patient 2. Bleeding episodes in patients with inhibitors are treated with bypassing agents such as rFVIIa and aPCC. In those MMHA patients in which the inhibitor does not cross-react with endogenous FVIII, DDAVP can be used. Eradication of inhibitors can be obtained via immune tolerance induction (ITI). In MMHA, ITI has been used with variable success. Hay et al. evaluated 26 patients with MMHA and inhibitors. There were only two of eight patients who had successful ITI. It was postulated that the low success rate was due to greater age and immunologic maturity [7]. The patient's bleeding phenotype must be taken into account when deciding whether to pursue ITI. In the two patients presented, there were no spontaneous bleeds when inhibitor titres were high, so an attempt at ITI was not indicated. Their inhibitor titres have returned to zero with no further FVIII exposure. Now, bleeding episodes can be controlled with use of DDAVP or rFVIIa.


  1. Top of page
  2. Acknowledgements
  3. Disclosures
  4. References

Authors gratefully acknowledge the Universal Data Collection Project for Blood Inhibitor Study from the CDC Foundation which performed the DNA sequencing of these patients' Factor VIII gene.


  1. Top of page
  2. Acknowledgements
  3. Disclosures
  4. References

The authors stated that they had no interests which might be perceived as posing a conflict or bias.


  1. Top of page
  2. Acknowledgements
  3. Disclosures
  4. References