M.E. Mancuso, Angelo Bianchi Bonomi Hemophilia & Thrombosis Center, IRCCS Maggiore Hospital, Mangiagalli & Regina Elena Foundation, University of Milan, Milan, Italy. Tel.: +39 0255034072; fax: +39 02547074; e-mail: firstname.lastname@example.org
Summary. Haemophilia therapy is aimed at treating and preventing bleeding episodes and related complications and clinical studies have shown that regular prophylaxis, started at an early age, is able to reduce physical impairment from haemophilic arthropathy. Today, the development of anti-Factor VIII (FVIII) inhibitors is the most serious treatment-related complication of haemophilia therapy and a number of genetic and environmental risk factors have been identified in the past years. Clinical data show that early start of prophylaxis and the avoidance of intensive treatment periods may protect patients from inhibitor development. The mechanisms are not completely understood; yet, recent experimental data suggest that pro-inflammatory or ‘danger signals’ may be involved in inducing tolerance vs. an effector immune response. So, exposure to a factor concentrate by itself may not be enough to trigger an immune response, while an intensive exposure to FVIII in the presence of such ‘danger signals’ can activate antigen-presenting cells, up-regulating co-stimulatory signals for T lymphocytes and ultimately enhancing antibody production. The ‘optimal’ regimen for primary prophylaxis is still not identified and barriers to prophylaxis implementation remain relevant. Key issues include the optimal age at prophylaxis onset, the optimal dosage/schedule, the proper clinical and laboratory monitoring and patients’ compliance. Practical approaches to early prophylaxis as implemented in the haemophilia centres in Milan and Bremen are discussed in this respect.
Haemophilia therapy is aimed at treating and preventing bleeding episodes and related complications, to preserve and/or restore joint function and, very importantly, to integrate patients into a normal social life.
Joint bleeds, which typically start between 1 and 3 years of age are a significant problem in severe haemophilia. Because of the oxidative capacity of iron to cause ongoing cellular injury, repeated extravasation of even small amounts of blood into joint cavities and the ensuing depositing of iron from lysed red blood cells into the joint will cause adverse effects on synovium, cartilage and bone. While the exact pathogenesis of haemophilic arthropathy is still poorly understood, iron-induced apoptosis of chondrocytes and proliferation of the synovial membrane seem to play a central role . Joint bleeds thus destroy the cartilage in a vicious cycle of bleeding and re-bleeding, facilitated by the hypertrophic and highly vascularized synovial tissue and may result in contractures and total loss of function of the joint.
Based on the observation that a moderate form of haemophilia A is associated with a considerably lower risk of spontaneous joint bleeds and arthropathy than a severe form (i.e. FVIII <1 IU dL−1), the concept behind the development of prophylaxis regimens in haemophilia management was to convert a severe to a moderate form of the disease, thereby reducing the number of spontaneous bleeds and related complications.
Since the introduction of the Swedish protocol in the 1960′s, it has been shown that regular prophylaxis, when started at an early age, is able to reduce physical impairment from haemophilic arthropathy [2,3] and largely ameliorate the major threats to haemophilic people – to become crippled or to bleed to death.
Since the advent of safe factor concentrates with a negligible risk of pathogen transmission, development of inhibitors has become the most serious and challenging treatment-related complication in haemophilia therapy in the developed world, as it renders the treatment of bleeding episodes problematic and prophylaxis unfeasible in many cases.
Inhibitors develop after the first 10 to 15 exposure days (EDs) in up to 30% of patients with severe haemophilia  and represent mostly a paediatric issue. Inhibitor development is still considered a poorly predictable multi-factorial event including both genetic and environmental risk factors. It is well-known that the type of FVIII gene mutation is a major determinant of the individual risk of developing an inhibitor and a high risk is associated with the genetic defects that totally prevent the production of the protein, referred to as null mutations (large deletions, inversions and nonsense mutations) .
However, in the Malmö International Brother Study (MIBS) the inhibitor concordance rate between siblings with the same mutation type was only 40% suggesting the influence of other genetic risk factors . Indeed, further studies on the same cohort showed the association between an increased inhibitor risk and two specific polymorphisms in the promoter region of interleukin-10 and tumour necrosis factor-α genes [7,8] and the protective effect of a polymorphism in the promoter region of the cytotoxic T-lymphocyte associated protein-4 (CTLA-4) gene .
On the other hand, the impact of non-genetic factors had been pointed out by the observation of haemophilic monozygotic twins discordant for inhibitor status . Thus, the role of many environmental and treatment-related factors has been advocated and further investigated: the type of FVIII products (recombinant and plasma-derived), the age at first FVIII exposure, FVIII treatment in the presence of events associated with tissue damage and/or inflammation (i.e. surgery, severe bleeding or vaccinations), antenatal/perinatal exposure to maternal FVIII (i.e. villo/amniocentesis or breast-feeding), intensity of treatment and type of treatment regimens (i.e. prophylaxis or on demand) [11–15].
Mechanisms of immune tolerance in haemophilia: self recognition or danger model?
The efficient generation of protective immune responses towards non-self molecules has certainly been an important characteristic in terms of natural selection, by decreasing the likelihood of infection. As a consequence, the therapeutic use of potentially immunogenic non-self proteins face the enormous obstacle of immune mechanisms evolutionarily selected for thousands of years to detect and rapidly eliminate such non-self molecules. These immune responses are not only a problem for the generation of inhibitors to clotting factors in haemophilia patients but also an obstacle for other potentially immunogenic drugs, such as monoclonal antibodies or gene therapy.
In recent years, significant efforts have been made for the development of efficient tolerisation protocols. Experiments in animal models of disease have suggested that the immune system can be reprogrammed towards tolerance . Interestingly, it appears that several immune mechanisms can participate in tolerisation, such as an expansion of regulatory T cells with suppressive properties [17,18] or the elimination of aggressive T cell clones by activation-induced cell death [19–21]. Inflammatory signals seem to influence the local microenvironment dictating which mechanism plays the predominant role in tolerance induction [22,23]. A better understanding of immune mechanisms that induce tolerance of non-self proteins will lead to tolerisation protocols that enable an improvement in the efficacy of potentially immunogenic drugs.
Early prophylaxis and inhibitor development
Two cohort studies had suggested an inverse relationship between the age at first exposure and the likelihood of inhibitor development [11,12], but such findings were not confirmed by studies in which the influence of this risk factor was adjusted for other variables [14,15].
The possible impact of prophylaxis on inhibitor development was suggested by the low incidence of inhibitor development reported in Sweden  and by two small retrospective cohort studies carried out in Spain and the UK [13,25]. In a retrospective data collection on 50 children from one Spanish centre, 15 of the 19 children treated on demand developed inhibitors while none of the 31 children on prophylaxis did (78% vs 0%; P < 0.05) . Subgroup analysis of the 20 children with high-risk gene mutations showed that none of the eight children on prophylaxis, but 11 of the 12 high-risk children treated on demand, developed inhibitors (0% vs 92%; P < 0.05). However, some limitations of this study do not allow drawing definite conclusions, as the FVIII genotype was determined for only 28 of the 50 children, it was a retrospective data collection, the sample size was relatively small and no multivariate analysis was performed. Moreover, in this study, children were treated with different product types and no information on the frequency of inhibitor assessment or on the total number of EDs in non-inhibitor children was available.
Similarly, in the retrospective UK study on 41 children, none developed an inhibitor on prophylaxis and three patients who had low-titre inhibitors prior to prophylaxis had undetectable inhibitors after prophylaxis .
The case–control study by Santagostino et al.  first showed a protective role of prophylaxis on inhibitor development by including this variable in a multivariate model with other relevant risk factors as FVIII gene mutations and family history of inhibitors (adjusted Odds ratio 0.2, 95% confidence interval: 0.06–0.9). Moreover, a subgroup analysis performed in children who started prophylaxis within 3 years of age confirmed a 70% reduction of the risk of inhibitor development . The case–control design assessing children with uniform characteristics (ethnicity, age, disease severity, FVIII product, inhibitor assessment), the sample size and the multivariate analysis in different subgroups make the results of this study more convincing.
The protective effect of prophylaxis was then confirmed in the frame of the CANAL study , a large retrospective multicenter cohort study aimed at evaluating the relationship between treatment characteristics and inhibitor development in previously untreated patients with severe haemophilia A. This study included 366 unselected patients from 14 haemophilia centres in Europe and Canada. Eighty-seven patients (24%) developed a clinically significant inhibitor and 69 of these (79%) were high responders (>5 BU mL−1). The incidence of inhibitors was clearly associated with surgery and intensive treatment at first FVIII exposure, while children on regular prophylaxis showed a 60% lower risk of inhibitor development than those receiving on demand treatment (Fig. 1) .
The results of all of these studies suggest that FVIII exposure by itself is not enough to trigger the immune response and that the type of treatment regimen may play a crucial role. In fact, with the ‘danger model’, it has been postulated that an intensive exposure to FVIII in the presence of danger signals deriving from injured cells (i.e. during major bleedings or surgery) may activate antigen-presenting cells, up-regulating co-stimulatory signals for T lymphocytes and ultimately enhancing antibody production by B lymphocytes, while regular exposure to FVIII in the absence of danger signals, as during prophylaxis, may lead to a regulation of the immune response through peripheral anergy of FVIII-specific T lymphocytes .
While the increasing evidence on the protective role of prophylaxis on inhibitor development could further prompt its use, the principles of evidence-based medicine require that these effects of early prophylaxis should be evaluated in the frame of prospective studies, even if a randomized design in comparison with on demand treatment may not be accomplished for obvious ethical reasons.
Practical approaches to early prophylaxis – how do I treat?
Today primary prophylaxis is generally recommended in children with severe haemophilia. By definition, primary prophylaxis is started before the age of 2 years, prior to any clinically evident joint bleeding or prior to the onset of joint damage (presumptively, after the first and within the second joint bleed) [27,28].
However, the ‘optimal’ regimen for primary prophylaxis is still not identified. Key issues in the selection of the optimal treatment regimen are the best age at prophylaxis onset, the dose regimen and the most adequate treatment monitoring including efficacy assessment. Also, despite the undeniable advantages, there are still barriers that stand in the way of widespread use of primary prophylaxis, such as the need for an adequate venous access, adherence to treatment regimens and economic considerations.
The importance of starting prophylaxis before the onset of joint damage is well recognized; however, it is difficult to establish the right time to initiate this treatment regimen because of the variability of the bleeding tendency among children with severe haemophilia. In fact, at least 10–15% of patients with a laboratory phenotype of severe haemophilia seldom bleed spontaneously and behave clinically as those with moderate or mild disease .
In the recently published multicentric and randomized prospective US American study in patients with severe haemophilia A, the Joint Outcome Study, early prophylactic treatment resulted in 93% of patients having normal index-joint structure on MRI at 6 years of age compared to only 55% of patients in the on-demand-treatment group . In this study, the reduction of MRI detectable joint damage by 83% was achieved by a prophylaxis regimen started after up to two haemorrhages into each index joint. However, as even one or two joint bleeds may already have an adverse impact, defining the optimal onset of primary prophylaxis is crucial and starting prophylaxis before any detectable joint bleed occurs has been advocated.
PEDNET, the European Paediatric Network for Haemophilia Management based on the collaboration of 23 paediatricians from 16 European countries, have thus recently provided the following new definitions, precisely describing the differences between treatment schedules: Primary prophylaxis A is the regular continuous treatment started after the first joint bleed and before the age of 2 years while primary prophylaxis B refers to regular continuous treatment started before the age of 2 years without previous joint bleeds .
Apart from recommendations based on the number of joint bleeds, the early occurrence of a life- or limb-threatening haemorrhage is usually considered a valid reason to start regular prophylaxis and in individual cases, it may be decided to start continuous prophylaxis after non-joint bleeds such as recurrent massive haematoma. Understanding the social life and commitment of the family, based on discussions with parents and caregivers, should be seen as a prerequisite for the decision.
As discussed above, the protective effect of early prophylaxis on the inhibitor risk [14,15] may represent another important factor to influence the decision for anticipating prophylaxis onset, particularly in patients exposed to other risk factors for inhibitor development, such as a positive family history of inhibitors, early intensive treatment for severe bleeding or major surgery .
The Swedish regimen, based on the administration of 20–40 IU kg−1 of FVIII thrice weekly and started roughly at 1 year of age is considered the gold standard of prophylaxis . In this setting of frequent dosing at a young age, the need for an adequate and stable venous access represents the main barrier to treatment feasibility, as frequent venous access can often be obtained only by using various forms of central venous lines that are not without mishaps such as infections and thromboses. Two different approaches have been used to overcome such difficulties: an incremental schedule of prophylactic infusions and the use of alternative forms of venous access.
The incremental dosing approach proposed by the Stockholm group allows to gradually increase the number of infusions per week (from 1 to 3 times per week) to get the children familiar with the standard venepuncture technique [32,33]. Other individualized approaches tailor treatment to bleeding pattern rather than to trough factor levels. In the Netherlands, the dose regimen is usually adjusted according to the bleeding diathesis with the aim of preventing spontaneous joint bleeding and this individualized approach has provided satisfactory results on long-term outcome by using a significantly lower amount of concentrate . The Canadian dose-escalating protocol  was designed to avoid the unnecessary implementation of prophylaxis in patients with a mild clinical phenotype. This regimen is initially based upon once weekly infusion of a relatively large single dose of FVIII (50 IU kg−1), with escalation to more frequent infusions on every third and then every other day if breakthrough bleeding occurs. However, long-term data are not yet available for this protocol.
The approach of alternative venous access is restricted to children with peripheral veins that are not adequate to bear frequent venepunctures and implies the surgical insertion of central venous catheters or creation of arteriovenous fistula (AVF) [36–38]. In a systematic review and meta-analysis of 48 studies with a total of 2704 patients, difficult venous access and prophylaxis were the primary indications for central venous access devices (CVAD) in 31.8% and 29.1% of patients respectively. The pooled incidence of infection was 0.66 per 1000 CVAD-days. However, 31.3% of CVADs were removed and infection was the reason for removal in 69.9% of cases . Results from the Angelo Bianchi Bonomi Haemophilia & Thrombosis Center in Milan showed that ports represent an alternative venous access able to improve prophylaxis feasibility in the majority of non-inhibitor children without adequate peripheral veins , whereas their use is greatly affected by severe complications as infections in inhibitor children who undergo immune tolerance induction (ITI) regimens . In these cases, the use of AVF allows ITI completion without interruptions and with an acceptable rate of complications .
So what do current approaches to early prophylaxis look like in different European haemophilia centres?
In the Angelo Bianchi Bonomi Haemophilia & Thrombosis Center in Milan, Italy, continuous prophylaxis is started after the first (within the second) joint bleed, irrespective of age. Because of the wealth of solid data on this protocol, the standard Swedish dosing regimen is prescribed, if venous access allows. If not, prophylaxis is started with once-weekly infusions that are increased over a period of up to 6 months to the minimum frequency of two injections per week. As soon as a single joint bleed occurs, the regimen is switched to full-dose prophylaxis. In Milan, treatment monitoring to determine the trough factor levels that prevent bleeding in the individual patient includes measuring trough levels of factor activity every 3 months and if spontaneous bleeds occur. However, these data alone are not considered sufficient to guide dose adjustment. As there is a lack of data on the number of breakthrough bleeds that might be acceptable without causing long-term joint damage, the occurrence of a single joint bleed is sufficient to increase the dose regimen. In Milan, peripheral veins are considered the first-choice access. However, this requires a strong commitment by the healthcare personnel to perform adequate home treatment training. If peripheral veins are not adequate, port insertion or creation of AVF represent the alternatives [37,38]. To reduce the risk of inhibitor development intensive on demand, regimens are avoided and elective surgery deferred, if possible, and, if intensive treatment is required, prophylaxis is started soon after without interruption.
The Comprehensive Care Centre for Haemophilia and Thrombosis at the Professor-Hess-Children’s Hospital in Bremen and some other German paediatric haemophilia treating centres start prophylaxis before the first joint bleed to avoid inhibitor development and severe bleeding complications. Patients are started on a once weekly prophylactic regimen with the lowest dose available (usually 250 IU corresponding to 25–35 IU kg−1 bodyweight) and access via a peripheral vein is used if possible.
This dose will be kept as long as possible. According to the bleeding frequency of the individual patient, the frequency of treatment will be increased to twice weekly or – if necessary – to three times per week. With this treatment schedule, central lines can be avoided and the compliance of the parents is good. After 30 to 40 exposure days, when the children are no longer in a period of high risk for inhibitor development, the individual dose can be increased.
This protocol has resulted in a benefit concerning the rate of inhibitor formation. In 13 previously untreated patients with severe Haemophilia A treated with this regimen with different factor concentrates, no inhibitor could be detected in the last 10 years while in a group of 12 PUP’s treated first on demand for more than 5 exposure days and/or with higher dosages during the same time period, three patients developed high titre inhibitors (personal communication).
In patients with severe haemophilia A, regular prophylaxis started at an early age is able to reduce physical impairment from haemophilic arthropathy. Study data suggest that early prophylactic treatment and avoidance of intensive treatment periods may also protect patients from the risk of inhibitor development. In Bremen and other German haemophilia centres, the implementation of a protocol of early prophylaxis, started with once weekly injections before the first joint bleed, has resulted in a very low rate of inhibitor formation. If the benefits of early prophylaxis in term of inhibitor development will be confirmed in the frame of large prospective studies, the definition of prophylaxis may have to be rewritten to include an additional ‘immunological’ aim besides arthropathy prevention.
This manuscript is based on presentations given at the Bayer Schering Pharma Haematology Conference 2008 on September 19, 2008 in Berlin, Germany. Editorial support was provided by Dr Gisela Beyendorff-Hajda from Physicians World GmbH, Mannheim, Germany.
M.E. Mancuso has occasionally acted as a paid consultant in educational activities for Bayer Schering pharma and CSL Behring.
G. Auerswald has acted as a paid consultant for Novonordisk and CSL Behring; a speaker for Bayer, Baxter and Novonordisk, as well as receiving funds from Baxter, Novonordisk and CSL-Behring.
L. Graca and E. Santagostino have no interests which might be perceived as posing a conflict or bias.