Ensuring the Safety of Medicines for Haemophilia
A medicine’s journey from laboratory bench to bedside is a highly regulated process – one that takes years. Trials involve thousands of people who take a compound that may or may not become a medicine and for which the risks are largely unknown. The safety of volunteers and patients who participate in these trials is protected by legislation. Most countries have a regulatory agency that enforces these laws and supervises medicines development. In Europe, it is the European Commission acting via the European Medicines Agency (EMA); in the United States, it is the Food and Drug Administration.
The development of most drugs follows a similar pathway (Table 1). A drug company must obtain approval from the regulator before moving from one stage to the next. This is done by conducting clinical trials to provide the scientific evidence that the chemical has the activity and lack of toxicity that would be expected for its potential use. Companies and regulators work closely together throughout the development process and regulators often provide support by fast-tracking compounds that are innovative or offer new treatments for disorders that are rare or commercially neglected (orphan drugs).
Table 1. Summary of drug development pathway
|Laboratory screening or molecular manipulation to identify potentially active chemicals|
|Studies in animal models and cell cultures to determine pharmacological activity and toxicity (the preclinical phase)|
|Clinical trials in healthy volunteers and then in patients|
|Regulatory approval (licensing) to promote and sell a product as a treatment for a specific disorder (its indication)|
|Post-marketing surveillance to monitor safety in clinical use and adapt the dose or indication if needed|
|Ongoing development to find new indications|
On receiving regulatory approval, the drug company is granted a product licence specifically for their (now branded) compound. The licence defines the approved indication(s), dose, prescribing cautions, contraindications and side effects. The company can only promote its product within the terms of this licence. In many countries, new medicines are further evaluated by health technology agencies such as the National Institute for Health and Care Excellence or NICE (in England and Wales) and the Scottish Medicines Consortium – which seek to determine whether they offer value for money for patients and healthcare systems.
There are four phases of clinical trials (Table 2). As they become larger and longer, trials provide increasingly robust evidence for the safety and efficacy of the new compound. By the time Phase III trials are completed, the drug company and regulators have a good idea how effective the compound is, how it compares with an established alternative and a reasonable idea how safe it is. Phase III trials are known as pivotal trials because they provide the scientific evidence on which marketing authorisation depends. They reliably reveal and measure common safety risks but rare adverse effects are usually only identified after a medicine has been marketed, when many more patients will be treated for longer periods. Postmarketing safety is assessed by Phase IV studies.
Table 2. Clinical trial phases. As haemophilia is a rare disease, new treatments are tested on far fewer patients at each stage than for most other drugs
|Phase||Purpose||Number of participants||Duration||% of compounds moving to the next phase|
|I||Safety and pharmacokinetics in healthy volunteers||20 – 100||months||70|
|II||Safety and efficacy in patients||hundreds||months – 2 years||33|
|III||Safety and treatment benefit in patients in randomised double-blind trials||hundreds to thousands
|1 – 4 years||25 – 30
|IV||Usually safety||thousands||years – ongoing||–
Adapted from Drug development. Step 3. Clinical research. Food and Drug Administration. (www.fda.gov/ForPatients/Approvals/Drugs/ucm405622.htm)
New treatments for haemophilia follow a similar development pathway as products for other conditions, but special measures are needed for medicines obtained from human plasmaor industrial scale cell cultures. Sourcing medicines in this way requires additional measures to ensure the purity of the final product, freedom from pathogens, preservation of biological activity and additional storage measures to maintain its integrity.
Manufacturers of plasma products adhere to standards for obtaining and processing plasma donations (see Plasma Protein Therapeutics Association at www.pptaglobal.org) and regulatory authorities publish guidance on plasma collection and good manufacturing practice[CMP, 2011). Regulators periodically visit manufacturing sites to check on compliance with these standards.
The safety of plasma products begins with ensuring the quality of supply. This is most effectively achieved by maintaining a register of donors, who are normally voluntary. Registration makes it easier to monitor the health of donors and to keep records that link donations to individual donors, who must (in the EU countries) complete a high-risk behaviour questionnaire and undergo a complete medical examination and donation screening tests. Donations may, however, come from paid unregistered donors; in such cases, safety depends on effectively screening the donated plasma for pathogens (e.g. viruses). In the United States, potential donors who test positive for selected pathogens (hepatitis B and C viruses, human immunodeficiency virus, parvovirus B19) are added to a national database of individuals who are prohibited from donating plasma. Similar donor databases exist in European countries, although the information is generally shared only within the plasma centres network. If a donor tests positive once, he or she is included in this database and rejected from future donations in that network.
Overall, European plasma has the same level of quality and safety as US plasma: European plasma centres implement quality standards considered equivalent to the US standards. European plasma has to follow European pharmacopeia standards, and even products manufactured with US plasma sold in EU countries must follow these requirements.
Donations are collected by plasmapheresis, a procedure in which blood cells are separated from plasma and returned to the donor. A record of each donation links it to the final medicinal product throughout the manufacturing process so that safety concerns can be tracked back to a single batch or the donor, even years after the donation was made.
Once a donation has been shown to be free of selected viruses it is stored for at least 60 days before entering the production process. This delay allows time for the donor to provide a second donation, which can be checked against the first to determine whether pathogens have emerged in the period since the first donation was made.
In the case of products for haemophilia, donations must be pooled to provide a sufficient quantity of clotting factors (this contrasts with fresh frozen plasma for transfusion, each unit of which comes from a separate, single donor). The pooled plasma is retested for pathogens then undergoes fractionation, a process that separates out its component proteins. These undergo purification and a range of treatments to inactivate and remove viruses; depending on the product, this may include heat treatment, pasteurisation, solvents, detergents and nanofiltration. The product is then packaged and labelled with a batch number so that it can be tracked through the supply chain to the bedside. Manufacturers use additional seals or unique markings to guarantee the authenticity of their product and reduce the risk of counterfeiting.
Human Factor VIII and Factor IX were cloned in the early 1980s, introducing a new era of recombinant clotting factors manufactured from well established lines of cell cultures, initially using mammalian cells and, more recently, human cells. Subsequently, recombinant Factor VIIa and Factor XIII have been introduced.
Recombinant technology offers the advantages of controlling the origin of clotting factors and greater certainty about continuity of supply compared with plasma-derived products. Typically, the cell line is inoculated with a gene that causes the cell to synthesise clotting factor. It is then cultured in large-scale fermenters and, when this stage is complete, the clotting factor is extracted by filtration and purification. These latter stages also inactivate viruses but, because the ‘upstream’ stages of the manufacturing process include derivatives of animal or human albumin (and therefore possibly viral protein), a precautionary antiviral step is also required. The most recent generation of recombinant products is manufactured without human or animal-derived additives but, again, components of animal and human origin are used early in the development and production of cell lines and a viral inactivation step remains necessary. (The US National Library of Medicine maintains an online description of DNA technology).
Monoclonal antibodies are becoming more familiar in the treatment of many disorders – notably rheumatoid arthritis, where they have quickly become the gold standard. These molecules are antibodies produced by recombinant technology using a single line of mammalian cloned cells (i.e. monoclonal). Exposing the cell culture to selected antigens makes it produce antibodies targeted at specific proteins – examples include cytokines such as tumour necrosis factor, some of the many interleukins and, in the case of haemophilia, proteins that affect clotting.
Monoclonal antibodies have a molecular sequence very similar to the antibodies produced by the mammal from which the cell line was derived, so they are themselves antigenic. To reduce this problem, the antibody can be made closer to the human form (‘humanised’) by combining human and mammalian DNA in the cell culture; a version identical to a human antibody can be created with transgenic technology.
The monoclonal antibody for the treatment of haemophilia that is most advanced in development is emicizumab, a humanised antibody that binds to activated Factor IX and Factor X, forming a bridge between the molecules. This Factor IX/X-antibody complex mimics the actions of Factor VIII but is not affected by inhibitors (antibodies directed against Factor VIII – see below), so it reduces bleeding in people whose factor replacement therapy has become less effective due to this complication[Sima et al, 2016]. More monoclonal antibodies targeting other steps in the coagulation pathway are in development.
The process by which monoclonal antibodies are manufactured is, like that of recombinant products, a highly complex sequence of cell culture and purification. The antibody is harvested from the cell culture by centrifugationand filtration, then purified chromatographically. This step also removes viruses but, again, it is a regulatory requirement to ensure that such procedures have been effective for each batch. The manufacturing process, once established and shown to be safe, can be used to produce antibodies directed against different target molecules – this is known as platform manufacturing.
The processes by which compounds such as clotting factors and monoclonal antibodies are manufactured have an effect on how antigenic the medicine will be. These biological medicines do not contain completely identical molecules – unlike, for example, a blood pressure lowering drug that is a synthetic small molecule. Instead, they contain low concentrations of molecules with slightly different structures. These variants contribute to the antigenicityof the medicine and their frequency and nature are determined by the manufacturing process (for example, a change in pasteurisationmay increase the number of variant molecules). Different brands of a clotting factor therefore have different antigenicities and even changes to the process by which a company manufactures its product can alter its antigenic profile. This is important for the treatment of haemophilia because long term treatment is complicated by antibodies (or inhibitors) that develop when the immune system recognises the clotting factor as a foreign protein.
It is a general principle of drug development that the balance of risk and benefit should as much as possible be in a patient’s favour. Clinical trials in the early stages of a compound’s development, when less is known about its safety, will include patients who are approaching the end of the treatment pathway – that is, they have already received the usual medicines but they still have troublesome signs and symptoms. (It is not unusual to hear such individuals described as having ‘failed’ the available treatments; in fact, it is the treatments that have failed them.) These patients have the most to gain from a new treatment so, for them, the potential benefit can outweigh the risk: they need more than current options can deliver and a new medicine might help them. On the other hand, the frequency and nature of its adverse effects are not fully known.
Trials of new haemophilia treatments therefore first involve previously treated patients (PTPs), who are usually adults; once safety is better understood, previously untreated patients (PUPs), who are usually children can be included. The US National Institutes of Health explains the issues arising from participation in clinical trials in detail.
The purpose of a clinical trial is to provide statistically robust evidence about how a medicine affects a person’s medical condition (the endpoints). Different types of endpoints are used to assess efficacy and safety in trials of haemophilia treatments (Table 3). The usual primary endpoint for a trial of a clotting factor is the frequency of bleeding, adjusted to an annual rate. For many years it has been assumed by those who design and carry out clinical trials that it is the clinical endpoints that matter most – after all, that’s what tells us how safe and effective a medicine is. It is now acknowledged that the patient’s perspective is at least as important, if not more so: after all, it’s not haemophilia we’re treating, it’s a person who has haemophilia.
Table 3. Examples of endpoints in clinical trials of haemophilia treatments
|Annualised bleeding rate
Target joints (number, location)
Use of rescue medication
Development of inhibitors
|Activated partial thromboplastin time
Renal and hepatic function Imaging to assess joint damage
|Activities of daily living Disability
Quality of life
Accordingly, new endpoints – known as patient reported outcomes (PROs) or patient oriented endpoints – have assumed greater importance. Examples include what difference the treatment has made to an individual’s ability to carry out normal daily activities like going to school, work or the ability to play sport; their mental and physical wellbeing; and carers’ workload. Quality of life, which encompasses physical, mental and social wellbeing in a single concept, can be difficult to measure scientifically. Measures of health-related quality of life may be disease-specific (e.g. the HAEMO-QoLfor haemophilia) or generally applicable (e.g. Short Form-36, SF-36) and they must be scientifically validated in the relevant patient group. In haemophilia, self-reported measures of physical function, such as the HALand PedHAL, have been developed.
The number of people who have taken a new medicine by the time it is introduced into clinical practice is still relatively low – several thousand, at most – and may be too small to reliably measure the risk of rareadverse effects (Table 4). Further, the duration of clinical trials is relatively brief – at most 2 – 3 years by the time of marketing – which is too short to be sure about the impact of life-long treatment. And the new medicine will be used in the real world, a setting less closely supervised than a clinical trial, in patients likely to have comorbidities and other treatments that would exclude them from a clinical trial. It’s therefore clear that trials can provide an indication of the risks and benefits of a new medicine but not a completely reliable picture of what will happen in clinical practice.
Table 4. Rare adverse events detectable by postmarketing surveillance
|Viral contamination||Measures to remove viruses during manufacture were greatly strengthened after almost 5,000 NHS patients with haemophilia were infected with HIV or hepatitis C after treatment with contaminated blood products in the 1970s and 1980s. Two factors make this a continuing priority. First, blood products are routinely screened for a small number of viruses and it is possible that a different virus could escape detection and cause infection. Second, the consequences for the patient are so serious that ongoing surveillance is justified even when the absolute risk is very low.|
|Rare allergic reactions||These are unpredictable and potentially serious. It is therefore important to know whether they are more frequent with one product than an alternative, and to identify any factors that contribute to the risk. Further, the risk may increase if the manufacturing process is altered and this can only be detected by continued monitoring.|
|Thrombosis||Clotting factors may cause thrombosis if they raise Factor VIII levels too high. This is a very rare event (one review found only 20 such events reported in 5,500 patients participating in 71 studies – see Haemophilia 2012;18:e173–e187) but it has only recently been recognised that the risk is probably higher, though uncertain, in people with cardiovascular disease.|
For this reason, safety is therefore monitored for several years after a product is launched – a process known as pharmacovigilance(Table 5). A regulatory authority may require this as part of a structured pharmacovigilance plan if it believes there are still safety issues to be answered when a medicine is marketed but the process may also be driven by the manufacturer and clinicians. Of course, detecting adverse events relies heavily on patients reporting them, so they must be informed and encouraged to participate.
Table 5. Examples of postmarketing surveillance
|Phase IV studies||These may be observational studies (in which all patients prescribed a medicine are monitored and adverse events are reported to a central coordinator – for example, the European Haemophilia Safety Surveillance study, www.euhass.org) or interventional (to compare different treatments)|
|Disease registries||In the case of rare disorders like haemophilia, details about patients’ treatment and outcomes are maintained on a central register, providing a database that can be analysed to investigate risks; this approach overcomes the problem of low numbers by pooling patients from several countries. Examples in haemophilia include PedNet (https://pednet.eu/registry) and the Network for Rare Bleeding Disorders (http://eu.rbdd.org).|
|Spontaneous monitoring schemes||These include the Medicines and Healthcare regulatory Agency’s Yellow Cards (https://yellowcard.mhra.gov.uk). These schemes rely on the vigilance of clinicians and patients to identify a suspected adverse reaction and report it to a central agency. These are usually government schemes that cover a single country. They are useful for identifying very rare events because they cover an entire national population but under-reporting is a major problem.|
A typical pharmacovigilance plan will therefore include measures to monitor long-term treatment, assess efficacy and safety in groups of patients who were under-represented in clinical trials (e.g. older people), monitor the outcomes of treatment during pregnancy, and review practical aspects such as injection devices and catheterisation.
Replacement clotting factors are large molecules of biological origin. They may be recognised as foreign by the immune system, resulting in the formation of antibodies, known as inhibitors, that bind to the clotting factor and neutralise it. For the patient, this means an increased risk of bleeding, joint damage and pain. Some people are more likely to develop an inhibitor than others, depending on their genetic makeup, haemophilia severity and treatment (Table 6).
Table 6. Inhibitors are more likely to develop…
|In people with severe haemophilia A (30%) than mild or moderate haemophilia A (9%)|
|In haemophilia A than haemophilia B (3%)|
|In people genetically predisposed (family history, mutations of Factor VIII gene and immune response genes, some ethnic groups)|
|With younger age at first treatment|
|With higher dose intensity|
|During the first 20 treatment days in previously untreated patients (and in most cases within the first 50 treatment days)|
|If the patient undergoes surgery or has an infection or inflammatory state|
|With, according to some but not all studies, recombinant factor VIII products than plasma-derived products that contain von Willebrand factor|
Adapted from: Peyvandi F et al. Haemophilia 2017;23 Suppl 1:4-13. doi: 10.1111/hae.13137).
Given that there are genetic and environmental factors influencing the development of inhibitors, of which some are unmodifiable, strategies to minimise inhibitors need to focus on those that can be modified. These include consideration of the dose, frequency of exposure, and product choice [Iorio et al, 2017]. Most work has focused on product choice.
Early analyses based on pooled data from registration studies and observational studies suggested that recombinant Factor VIII was associated with a higher risk of developing inhibitors than plasma-derived Factor VIII, or that there was no difference in risk between the two types. However, more recently, a randomised controlled trial (the SIPPET study) showed the reverse may be the case [Peyvandi et al, 2017]. This trial was the first that was specifically designed to compare the incidence of inhibitors in patients treated with plasma-derived factor VIII containing von Willebrand factor or recombinant factor VIII. In a total of 251 patients, the incidence of all inhibitors was about 29% with plasma-derived factor VIII and 45% with recombinant factor VIII. The figures for high titre inhibitors were 19% and 28% respectively.
It is presently uncertain how to apply these findings to the clinical use of replacement clotting factors given the conflicting conclusions of the various published studies. The patients in the SIPPET study came mainly from the Middle East and India and many had gene mutations that increased their risk of inhibitors; no third generation recombinant clotting factors were used; some patients received very few doses; and the threshold for confirmed positive inhibitor status was lower than has been used in other studies. Nevertheless, based on the available evidencethat one particular factor may account for the difference observed, the World Federation of Hemophilia suggests it may be prudent to consider not prescribing the recombinant product octocog alfa (available as Kogenate from Bayer and Helixate from NexGen) for newly diagnosed PUPs with severe haemophilia A when other safe clotting factor concentrates are available.6It adds that there is no known increased risk for any other patients using these products. The European Medicines Agency recently reviewed the evidence in light of the SIPPET study and concluded there was no clear difference between recombinant and plasma-derived products overall; more likely, any difference is between specific products (this is consistent with the WHF recommendation). However, it is now re-examining the evidence.
People receiving clotting factor replacement therapy are monitored to ensure their blood level of Factor VIII is within the therapeutic range. An unexpectedly low level raises the possibility that an inhibitor may have developed. This is confirmed by a blood test to determine the level of inhibitor (its titre), which is measured in Bethesda Units (BU; the amount of inhibitor that will inactivate half of a coagulant in a set time).
Various authorities have different recommendations for inhibitor testing:
- The World Federation of Hemophilia recommends children are screened once every 5 exposure days until 20 exposure days, every 10 exposure days between 21 and 50 exposure days, and at least twice a year until 150 exposure days. For adults with more than 150 exposure days (apart from a review every 6 – 12 months), any failure to respond to adequate factor concentrate replacement therapy in a previously responsive patient is an indication to assess for an inhibitor. Inhibitor measurement should also be done in all patients who have been intensively treated for more than five days, within four weeks of the last infusion; prior to surgery or if recovery assays are not as expected; and when clinical response to treatment of bleeding is sub-optimal in the post-operative period.
- The European Haemophilia Consortium recommends that children and adults who are newly diagnosed with haemophilia should be tested regularly for inhibitors between the 1st and 50th days of treatment. After the 50th day of treatment, they should be checked at least twice a year until they have received 150 – 200 doses then at least once annually. Testing for inhibitors should also be done before any major surgery.
- The United Kingdom Haemophilia Centres Doctors Organisation (UKHCDO) recommends that infants and children with severe haemophilia should be tested for an inhibitor at least every day of treatment until the 20thday, then every 3 – 6 months until 150 treatment days have been completed.7
Inhibitor status is categorised as low titre (<5 BU) or high titre (≥5 BU); a high titre means greater inhibition of clotting factor activity and a higher risk of bleeding. For a PUP, the risk of developing an inhibitor is greatest early in treatment – highest during the period in which the first 20 days of treatment are given and diminishing with time until they are unlikely to develop after 50 treatment days. Inhibitors can develop in previously treated people but the risk if very small (about 3 per 1000 years of treatment), the titre is usually low and they may disappear spontaneously.
There are three options to overcome the effects of inhibitors. If the titre is low, it is possible to compensate for the reduced activity of clotting factor by increasing its dose. For people with a high titre, a bypassing agent can be used to circumvent the inhibitor and act on a different part of the coagulation pathway; this is usually an option to provide short term cover during surgery or to stop acute bleeding. The currently available bypassing agents are recombinant activated FVII and plasma-derived activated prothrombin complex concentrate, both of which are administered intravenously.
The preferred treatment to eradicate inhibitors is immune tolerance induction (ITI), started as soon as possible. The UKHCDO recommends this as first-line therapy for children regardless of the inhibitor titre.7ITI entails daily or several times weekly administration of high doses of Factor VIII until the immune system becomes tolerant to it, as indicated by the disappearance of inhibitors from the blood and the recovery of Factor VIII activity. The odds of success are greater in individuals with low levels of inhibitors (ideally <5 BU), have never had high levels (>50 BU), and have not had inhibitors for more than 5 years.
ITI is usually successful within a year but some people may need up to two years’ treatment. Bleeding episodes during this period are treated with a bypassing agent. Treatment should not be interrupted because this impairs tolerisation. This is an intensive approach that is demanding for patients and families but success rates are very high, with inhibitors eradicated in 60 – 80% of patients. ITI appears expensive but it likely to be cost-effective compared with on demand or prophylactic use of bypassing agents.
There is no clear evidence to guide the choice of Factor VIII preparation for ITI. The UKHCDO recommends recombinant FVIII concentrate for children; if this is not effective, the options are switching to a plasma-derived FVIII product with a high von Willebrand factor (vWF) content, increasing the dose of Factor VIII or immunosuppression with rituximab [UKHCDO, 2017]. US guidance makes no recommendation for first line therapy but suggests that if treatment with recombinant Factor VIII is unsuccessful, a switch to plasma-derived Factor VIII/von Willebrand factor should be considered [Valentino et al, 2015]. The evidence on choosing a Factor VIII product has been reviewed recently [Iorio et al, 2017].
vWF is a protein that occurs in the circulation bound in a stable complex with Factor VIII. In this state, Factor VIII is inactive but it is released from the complex with vWF by the action of thrombin; it then contributes to the coagulation cascade by interacting with Factor IXa to activate Factor X.
vWF is present in plasma-derived Factor VIII products, where it may help to reduce their antigenicity. It acts as a ‘chaperone and protector’ of Factor VIII and the circulating complex formed by the two molecules is non-antigenic. In recombinant products, about one-fifth of the Factor VIII does not bind with vWF and may therefore contribute to antigenicity. Other properties may be important. For example, recombinant Factor VIII products lack the immunosuppressive proteins that occur naturally in plasma-derived products and recombinant Factor VIII derived from mammalian rather than human cell lines may have small structural differences that increase antigenicity.
Medicines to treat haemophilia undergo rigorous appraisal of their efficacy and safety to provide evidence in support of marketing approval but this phase of their development is just part of a process of assessment that continues in everyday clinical practice. The range of options for factor replacement therapy is widening and making the right choices is becoming complex; more evidence is needed to help patients and clinicians navigate the treatment pathway.
Committee for Medicinal Products for Human Use. Guideline on plasma-derived medicinal products. EMA/CHMP/BWP/706271/2010. European Medicines Agency. July 2011 (www.ema.europa.eu/ema/pages/includes/document/open_document.jsp?webContentId=WC500109627; accessed June 2017)
European Medicines Agency. Research and development. (www.ema.europa.eu/ema/index.jsp?curl=pages/regulation/general/general_content_001768.jsp&mid=WC0b01ac0580b18a3a; accessed June 2017)
Iorio A, Fischer K, Makris M. Large scale studies assessing anti-factor VIII antibody development in previously untreated haemophilia A: what has been learned, what to believe and how to learn more. Br J Haematol 2017 Apr 7. doi:10.1111/bjh.14610.
National Institutes of Health. NIH research trials and you. The basics. (www.nih.gov/health-information/nih-clinical-research-trials-you/basics; accessed June 2017)
National Library of Medicine. From DNA to beer. (www.nlm.nih.gov/exhibition/fromdnatobeer; accessed June 2017)
Peyvandi F, Mannucci PM, Garagiola I, et al. A randomized trial of factor VIII and neutralizing antibodies in hemophilia A. N Engl J Med. 2016 May 26;374(21):2054-64. doi: 10.1056/NEJMoa1516437.
Peyvandi F et al. Haemophilia 2017;23 Suppl 1:4-13. doi: 10.1111/hae.13137).
Shima M, Hanabusa H, Taki M et al. Factor VIII-mimetic function of humanized bispecific antibody in hemophilia A. N Engl J Med 2016;374:2044-53.
United Kingdom Haemophilia Centres Doctors Organisation. UKHCDO protocol for first line immune tolerance induction for children with severe haemophilia A: A protocol from the UKHCDO Inhibitor and Paediatric Working Parties. February 2017. (www.ukhcdo.org/wp-content/uploads/2017/01/ITI-protocol-2017.pdf; accessed June 2017)
Valentino LA, Kempton CL, Kruse-Jarres R et al. US Guidelines for immune tolerance induction in patients with haemophilia A and inhibitors. Haemophilia 2015;21:559–567.
World Federation for Hemophilia. Third publication suggests recombinant FVIII product associated with higher risk of inhibitor development in newly diagnosed, previously untreated patients with severe hemophilia A. March 2015. (www.wfh.org/en/our-work/treatment-safety/inhibitors-pups-update-nov2014; accessed June 2017)