Inhibitor incidence after intensive FVIII replacement for surgery in mild and moderate haemophilia A: a prospective national study in the Netherlands
Correspondence: Karin Fijnvandraat, Department of Paediatric Haematology, Academic Medical Centre, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands.
Inhibitor development is currently the most severe complication in mild/moderate haemophilia A patients, causing increased bleeding tendency, hospitalization and mortality. It has been suggested that receiving high doses of factor VIII (FVIII) concentrates for surgical procedures is an important risk factor for inhibitor development in these patients. The current multicentre study aimed to determine prospectively the incidence of inhibitor development after intensive FVIII replacement therapy for surgical procedures in patients with mild/moderate haemophilia A. All consecutive patients with mild/moderate haemophilia A were included when they required at least 10 000 iu of FVIII concentrates (or 250 iu/kg) for 5 or more days for a surgical procedure. Potential clinical risk factors for inhibitor development and results of inhibitor tests were collected. Forty-six patients with a median age of 54 years (interquartile range, 40–59 years) were included in the study. F8 genotyping revealed 20 different missense mutations. Patients received either recombinant (65%) or plasma-derived FVIII concentrates (35%) by intermittent bolus injections (41%) or continuous infusion (57%). Two patients developed a low titre inhibitor post-operatively. The incidence of inhibitor development following intensive treatment for surgery in this unselected prospective cohort of mild/moderate haemophilia A patients was 4% (95% confidence interval, 0·5–14·8).
Haemophilia A is an X-linked bleeding disorder caused by a mutation in the factor VIII (FVIII) gene (F8), resulting in a deficiency of plasma FVIII. Classically, haemophilia A is divided into three forms of severity, based on the level of plasma FVIII coagulant activity (FVIII:C): severe (FVIII:C < 2 iu/dl), moderate (FVIII:C 2–5 iu/dl) and mild (FVIII:C 6–40 iu/dl). Spontaneous bleeding rarely occurs in mild haemophilia A and is mostly provoked by (minor) trauma or surgery. Bleeding episodes are preferably treated or prevented by desmopressin (DDAVP) in responders (Mannucci, 2000), while major bleeds or invasive procedures requiring prolonged increased plasma levels of FVIII are treated with FVIII concentrates.
The development of inhibiting antibodies towards FVIII is the most severe complication of treatment with FVIII concentrates. Inhibitor formation represents a major challenge in the management of haemophilia A as it renders the administered FVIII concentrates ineffective, leading to increased complications and mortality. In patients with mild/moderate haemophilia, inhibitors can also cross-react with endogenous FVIII resulting in a more severe phenotype with spontaneous bleeding (Fijnvandraat et al, 1997; Thompson et al, 1997; Jacquemin et al, 2003). Bleeding symptoms are often severe and life threatening, forcing these patients to change their lifestyle completely (Hay et al, 1998).
Inhibitor eradication therapy by repeated high dose of intravenous FVIII administration (Immune Tolerance Induction) is only successful in 51–76% of severe haemophilia A patients with a high titre inhibitor and is very expensive (Coppola et al, 2010). Moreover, no standardized therapies for inhibitor eradication are available for mild/moderate haemophilia A (Peerlinck & Jacquemin, 2010).
Inhibitors occur more frequently in patients with severe haemophilia (incidence 25–30%) as compared to patients with mild/moderate haemophilia (incidence 3–13%) (Wight & Paisley, 2003). However, inhibitor development in patients with mild/moderate haemophilia A seems to be rising: in the Netherlands more than one-third of the newly diagnosed inhibitors were shown to occur in mild/moderate haemophilia (Colvin et al, 1995; Plug et al, 2004). Given that more than 50% of all haemophilia patients have a mild/moderate form, the clinical impact of the problem is substantial.
Previous studies have demonstrated that inhibitor development in mild/moderate haemophilia A is multi-factorial, involving both genetic and environmental risk factors. (Eckhardt et al, 2009; Astermark et al, 2010) The genetic predisposition for inhibitor development is associated with the underlying F8 mutation (Oldenburg et al, 2002) and a positive family history of inhibitor development (Astermark et al, 2001). Mild/moderate haemophilia A is generally caused by a missense mutation of F8. Previous research showed that specific mutations that alter the three-dimensional structure of the FVIII protein are associated with a higher incidence of inhibitor formation (e.g. Arg593Cys and Arg2150His) (Peerlinck & Jacquemin, 2010).
Especially in mild/moderate haemophilia A, concerns have been raised about the association between inhibitor development and intensive treatment with FVIII concentrates for surgical procedures (Sharathkumar et al, 2003; Von Auer et al, 2005; Kempton et al, 2010; Gouw & van den Berg, 2009), particularly in patients carrying high-risk F8 mutations (Eckhardt et al, 2009) and in older patients (Kempton et al, 2010). This could be explained immunologically by the combination of cytokine release through tissue damage resulting from surgery, combined with excessive amounts of exogenous FVIII antigen. Both may render the immune system more susceptible to develop inhibitors (Matzinger, 2002). The influence of other possible contributing factors, such as the type of administered FVIII concentrate, administration by continuous infusion instead of intermittent bolus injections and the occurrence of concomitant infection or inflammation, needs further investigation.
It is important to know the incidence of inhibitor development in patients with mild/moderate haemophilia A patients undergoing surgery, because this risk has to be considered when an elective surgical procedure is planned. However, prospective studies on this topic are lacking. Previous studies were all retrospective, performed in small populations (Sharathkumar et al, 2003; Von Auer et al, 2005) or in selected groups of patients with a high proportion of high-risk mutations (Sharathkumar et al, 2003; Von Auer et al, 2005; Kempton et al, 2010; Gouw & van den Berg, 2009). Although mild/moderate haemophilia A patients are not treated frequently, they will usually have received several doses of FVIII concentrate at the time of surgery. To generate data that are relevant for clinical practice, we designed a prospective multicentre study to determine the incidence of inhibitor development in a consecutive group of previously treated mild/moderate haemophilia A patients receiving intensive FVIII replacement therapy for surgical procedures.
Patients and methods
All mild (FVIII:C 6–40 iu/dl) and moderate (FVIII:C 2–5 iu/dl) haemophilia A patients, at least 12 years of age and requiring intensive treatment with FVIII concentrate for elective surgery were eligible for study entry. Intensive treatment of FVIII was defined as the cumulative use of at least 10 000 iu) or 250 iu/kg for 5 or more consecutive days. Patients were excluded if they had any other haemostatic disorder or a positive history of inhibitors.
Patients were recruited consecutively at seven Haemophilia Treatment Centres (HTCs) in the Netherlands during a prospective period of 4 years. The source population from which eligible patients were recruited comprised all patients with mild/moderate haemophilia A cared for by these seven HTCs (approximately 750 mild/moderate haemophilia A patients). The study protocol was conducted in accordance with the Declaration of Helsinki and ethical approval was obtained by each participating centre. Informed consent was obtained in writing from all subjects before study entry.
After inclusion, clinical data were collected from hospital databases and patient files by the patient's physician at least 1 week before surgery and included: date of birth, ethnicity, cumulative number of exposure days (ED) to FVIII concentrates, previous intensive treatments for bleeding or surgery, family history of haemophilia A and inhibitors, co-morbidity and medication.
During surgery, which took place in one of the participating HTCs, patients were treated according to national guidelines under supervision of their treating haemophilia specialist. Data recorded peri-operatively included: type of surgery and indication, type of FVIII concentrate, mode of FVIII administration (intermittent bolus injections or continuous infusion), number of ED, cumulative amount of FVIII concentrate administered, peri-operative medication and complications during the first week after surgery.
Patients were seen at a follow-up visit at four to 8 weeks after the surgical procedure. During this visit all post-operative complications were recorded and a blood sample was obtained.
Ethylenediaminetetraacetic acid (EDTA) anti-coagulated blood for F8 genotyping was sent to the Academic Medical Centre genomics laboratory, Amsterdam. F8 mutation was determined by sequencing of the F8 gene. Laboratory assessments that were performed at each local laboratory included: FVIII:C by one-stage clotting assay, von Willebrand factor (VWF) Ristocetin Cofactor activity (VWF:RCo), VWF antigen (VWF:Ag), anti-FVIII, antibodies against hepatitis A virus, hepatitis B virus, hepatitis C virus and Human Immunodeficiency Virus (HIV).
Inhibitors were locally tested by the Nijmegen modification of the Bethesda assay. (Verbruggen et al, 1995) Patients were tested for inhibitors before surgery and at the follow-up visit after surgery or earlier in case of a clinical indication (i.e. no response to FVIII treatment or increased bleeding tendency). A titre of 1–4 Bethesda Units (BU)/ml was defined as a low inhibitor titre and a titre of at least 5 BU/ml was defined as a high inhibitor titre. If the result of the Bethesda assay was between 0·6 and 1 BU/ml, the patient was classified as an inhibitor patient if he presented with spontaneous bleeding symptoms, or if the FVIII ratio (FVIII:C during inhibitor/FVIII:C before inhibitor) was 0·5 or less.
All data was collected prospectively. In the description of patient characteristics continuous data are presented as medians and interquartile ranges (IQR).
Patient characteristics and genotype
During the enrolment period, 54 patients were admitted to one of the participating HTCs for a surgical procedure; 48 (89%) fulfilled with the inclusion criteria. Of the six excluded patients, one patient was a female carrier of haemophilia A, three patients were found to have von Willebrand disease type 2N and two patients had a positive inhibitor history. Of the 48 included patients, two patients were excluded during study follow-up because they received less FVIII concentrate than initially estimated; both of these patients did not develop an inhibitor.
In total, 46 haemophilia A patients, 43 mild and three moderate, with a median FVIII baseline level of 16 iu/dl (IQR, 8–25 iu/dl) and a median age at inclusion of 54 years (IQR, 40–59 years) underwent intensive FVIII treatment for a surgical procedure and were included in the study. Baseline characteristics of the patients are listed in Table 1. Forty-three patients (93%) were genotyped, revealing 20 different F8 missense mutations, including four novel mutations not previously reported in the HaMSTERS database (http://hadb.org.uk) (Table 2).
Table 1. Baseline characteristics of all consecutive mild and moderate haemophilia A patients treated intensively for surgery (n = 46)
|Positive family history of inhibitors|
|Baseline laboratory results|
|Previous treatment for surgical procedures|
|Yes, covered by FVIII||28||(61%)|
|Yes, covered by cryo or DDAVP||5||(11%)|
Table 2. Characteristics of the 20 F8 gene missense mutations identified in 43 patients with mild/moderate haemophilia A
Medical history of patients
More than half of the patients (n = 26; 57%) had one or more co-morbidities besides haemophilia A. Eight patients (17%) were suffering from cardiovascular disease; five of them were taking anticoagulants (thrombocyte aggregation inhibitor (n = 4) or heparin (n = 1)). Fifteen patients (32%) were infected with hepatitis C, none were infected with HIV. Other co-morbidities were malignancy (n = 3), epilepsy (n = 1), non-insulin dependent diabetes mellitus (n = 1), hypertension (n = 2) and hypercholesterolaemia (n = 1).
There was a broad range in previous ED (Table 1). Seventy percent (n = 33) of the patients had undergone one or more surgical procedure before inclusion; eleven of them had received peri-operative treatment for a period of more than three consecutive ED in the past. Fourteen patients (38%) had been intensively treated for one or more bleeding episode prior to study participation.
Various surgical procedures were performed (Table 3). Thirty patients (65%) were treated with a recombinant FVIII product to cover the surgical procedure. The other 16 patients (35%) received plasma-derived FVIII products. Patients were exposed to FVIII concentrates for a median of nine cumulative ED (IQR 7–13) following surgery and received a median cumulative FVIII of 453 iu/kg (IQR 251–610). In 26 patients (59%) the administration of FVIII concentrates was initially by continuous infusion, followed after several days by intermittent bolus injections for a median of 9 ED (IQR 7–12). The cumulative use of FVIII in patients who received FVIII concentrates exclusively by bolus injections was slightly lower as compared to continuous infusion (median 430 iu/kg; IQR 217–633 vs. 472 iu/kg; IQR 337–571).
Table 3. Overview of the 46 surgical procedures and peri-operative characteristics
|Type of surgery|
|Cumulative amount of FVIII administered peri-op, iu||34 480||(20 000–49 780)|
|Total number consecutive ED post-op||9||(7–13)|
|Administration mode, patients (n)|
|Administration mode, days (n)|
|Blood transfusion performed peri-op||5||(10%)|
|Medication administered peri-op|
|Complications first week after surgery|
Two patients (4%; 95% confidence interval [CI], 0·5–14·8) developed an inhibitor post-operatively.
Case 1 was a 35-year-old male with mild haemophilia A (FVIII:C, 9 iu/dl) of Syrian origin. At the age of 25 years he had a spinal intramedullary haemorrhage following a traffic accident and has been paraplegic since then. Multiple surgical procedures have been covered by FVIII concentrates previously at another HTC; his previous exposure status was not exactly known but estimated to be at least over 50 ED.
On this occasion he underwent plastic surgery for decubitus ulcers as a result of his paraplegia in combination with the excision of an infected femur head. Surgery was covered by continuous infusion of recombinant FVIII concentrate for 14 d; he received a total cumulative amount of 100 000 units FVIII. During the post-operative period, his FVIII dose had to be increased daily to maintain his FVIII:C at an adequate level. Haemostasis was optimal and no bleeding complications occurred. Two days after continuous infusion was stopped, he tested positive for an inhibitor (2·5 BU/ml, also peak itire). His plasma FVIII level was similar to prior levels (FVIII:C 7 iu/dl). Immune tolerance induction was started on the same day with a regimen of 25 iu/kg per day three times per week. After 3 months the inhibitor disappeared. F8 mutation analysis indentified a missense mutation (Arg531Cys). His two brothers, who both have haemophilia A, have not developed an inhibitor; no further information was available about the number of ED that these brothers received.
Case two was a 58-year-old Caucasian male with mild haemophilia A (FVIII:C, 9 iu/dl). He developed a low-titre inhibitor (1 BU/ml) after vascular surgery. Prior to this procedure his lifetime cumulative number of ED was <10 and he had not previously been treated intensively with FVIII concentrates or for surgery. Besides haemophilia A he suffered from vascular disease, hypercholesterolaemia and hypertension.
His surgery was managed by continuous infusion for 5 d followed by daily bolus injections for 4 d (a total of nine consecutive ED); a total amount of 73 000 units recombinant FVIII concentrate was administered. Haemostasis was optimal and no bleeding or other complications occurred. An inhibitor (0·8 BU/ml) was detected 40 d after surgery. His baseline FVIII:C level remained stable (FVIII:C, 8 iu/dl). There were no bleeding complications and the inhibitor was followed by regular testing. The peak inhibitor was 1·0 BU/ml 3 months after surgery and the inhibitor spontaneously disappeared within several months. F8 mutation analysis indentified a missense mutation (Arg593Cys).
This is the first prospective nationwide study to measure the cumulative incidence of inhibitor development after intensive FVIII replacement therapy for surgery in mild/moderate haemophilia A. Two out of 46 patients (4%; 95% CI, 0·5–14·8) developed an inhibitor post-operatively. The frequency of inhibitor development after surgery that we found in this relative large cohort of consecutive patients was lower than that observed in a smaller study (Sharathkumar et al, 2003) (25%; 95% CI 7·3–52·4) and a previous retrospective cohort study (Eckhardt et al, 2009) at one of the participating HTCs (17%; 95% CI 7·2–32·1). The higher inhibitor incidence in those studies may be explained by the selection of high-risk patients for inclusion in these studies. Sharathkumar et al (2003) reported that four of the 16 patients treated intensively with FVIII concentrates developed an inhibitor. Three of these patients were previously untreated patients (PUPs) and one had only one previous ED to FVIII concentrate. In the cohort study (Eckhardt et al, 2009), which found an inhibitor incidence of 17% after surgery (7/41), a large proportion (38%) of the included patients carried the F8 Arg593Cys mutation which is associated with an increased risk of inhibitor development.
In the present study the majority of the surgical patients that were consecutively included (70%) had been challenged before in one or more periods of (intensive) FVIII treatment without developing inhibitors. Remarkably, one of the patients that developed an inhibitor did so after more than 50 ED while he had faced many prior challenges – including surgery, inflammation and intensive treatment – without developing an inhibitor. An unexpectedly late occurrence of inhibitors in mild/moderate patients has also been observed in a previous case-control study (Kempton et al, 2010). The authors reported that 42% of the case subjects developed an inhibitor after more than 50 cumulative ED, illustrating that previously treated (>50 ED) mild/moderate haemophilia A patients are still at risk for inhibitor development. This may also be the case for severe patients, as demonstrated recently in a UK national study comprising 2528 severe haemophilia A patients who were followed up for a median of 12 years (Hay et al, 2011). Although the incidence of inhibitors was highest among the patients below 5 years of age (64/1000 treatment years) there was a second peak of significantly increased inhibitor incidence among those aged >60 years (11/1000 treatment years).
As life expectancy increases, the rising incidence of inhibitor development in older patients with haemophilia A will become an important clinical challenge (Leebeek et al, 2004). Moreover, age-related health problems, such as cancer and arthrosis, will increase the need for intensive treatment with FVIII concentrates for surgery in these patients (Mauser- Bunschoten et al, 2009). In Case 1, intensive treatment for surgery was performed while infection was present (infected femur head). Environmental circumstances, such as high FVIII antigen load, tissue damage and inflammation may trigger the immune system to form inhibitor antibodies, irrespective of previous cumulative ED. Further research is required to explore the pathophysiology of late-onset inhibitors and to elucidate the role of other potentially synergistic risk factors for inhibitor development at advanced age, to enable preventive measures.
The genotype may be an important contributing factor to inhibitor development in the two cases of the present study. Both cases were carrying high-risk F8 missense mutations, caused by a cysteine replacement. The formation of disulphide bridges by cysteine replacement may alter the folding of the FVIII protein and thereby induce recognition of wild-type FVIII by the immune system (van den Brink et al, 1999; Bril et al, 2004). Among the patients with missense mutations that are associated with inhibitor development according to the HaMSTERS database (i.e. Arg531Cys, Arg593Cys, Asn618Ser, Arg1781His and Arg2150His) (http://hadb.org.uk), the incidence of inhibitors in our cohort was 10% (2/21).
The incidence of inhibitor development found in this observational study may be influenced by multiple factors including age, previous (intensive) FVIII exposure and the presence of high-risk F8 genotypes. The absence of a reference group in this study restricted the analysis to calculating the incidence of inhibitor development after surgery.
Nevertheless, the results emphasize the importance of being aware of inhibitor risk in these patients, irrespective of age and cumulative exposures to FVIII concentrates. Caution should be taken when treating mild/moderate haemophilia A patients intensively with FVIII concentrates for surgery, and alternative or additional use of desmopressin should be considered whenever possible. Routine inhibitor testing should be performed post-operatively to detect inhibitors at an early stage. Further study is needed to confirm these findings in mild/moderate patients with other characteristics (e.g. previously untreated patients) and to investigate potential risk factors that may play a role in post-operative inhibitor development. This will ultimately help to optimize inhibitor prevention in future treatment.
Haemophilia Treatment Centres at the following medical centres participated in this study: Academic Medical Centre, Amsterdam; Academic Medical Centre, Groningen; Van Creveldkliniek, Utrecht; Erasmus Medical Centre, Rotterdam; Radboud Medical Centre, Nijmegen; VU Medical Centre, Amsterdam; Leiden University Medical Centre, Leiden.
CLE performed research, collected, analysed and interpreted data and drafted and wrote the manuscript; EPMB, FWGL and FJMM supervized data collection and critically reviewed the paper; MP designed research, supervized data collection and writing of manuscript; and KF designed research, supervized data collection, analysis and interpretation, drafted the manuscript and supervized writing of the manuscript.
Conflict of interest
CL Eckhardt is financially supported by a grant from the Dutch Health Council (ZonMw Grant no 40-00703-98-8570); the other authors declare no competing financial interest.