Evolution of the European guidelines for the clinical development of factor VIII products: little progress towards improved patient management


Correspondence: Pier Mannuccio Mannucci, Scientific Direction, IRCCS Ca' Granda Maggiore Policlinico Hospital Foundation, Via Pace 9, 20122 Milano, Italy.

Tel.: +39 02 55038377; fax: +39 02 50320723;

e-mail pmmannucci@libero.it


In the process of clinical development and licensing of factor VIII (FVIII) products for treatment of haemophilia A, the safety concerns generated in the 1980s by the risk of pathogen transmission were tremendously reduced by the implementation of an array of methods for inactivation/removal of blood borne pathogens. The current focus on the risk of FVIII inhibitors does not stem from a new awareness, because this multifactorial complication has long been recognized. With this background, I believe that the current European regulatory guidelines for the clinical development and licensing of FVIII products fail to reflect the tremendous progress made in terms of clinical efficacy and safety, because they are witnessing a continuous increase in the demands from health agencies to the point that clinical studies have become more and more difficult to carry out. This article reviews the evolution of the European regulations on new FVIII products, lists a number of regulatory requirements whose scientific and/or clinical rationale is perhaps questionable and recommends keeping such requirements in reasonable limits of feasibility, without jeopardizing current high standards of efficacy and safety.


Haemophilia A is an inherited blood coagulation disorder, characterized by the deficiency of factor VIII (FVIII) that occurs almost exclusively in men at a rate of about 1 in 5000 births. The current treatment is mainly based upon replacement of the deficient factor to prevent or stop bleeding. Compared to the 1960s, when plasma and cryoprecipitate were the only products available for treatment of haemophilia A, continuous progress has been made from the 1970s through the manufacturing of efficacious concentrates from human plasma or by genetic engineering. In this technological evolution, manufacturers' expertise and regulators' requirements have operated until recently in a synergistic manner, with a sharp focus on efficacy and safety, but bearing in mind the limitations inherent to the rarity of haemophilia and allied coagulation disorders. However, in the last few years, regulatory requirements for clinical development and licensing of new coagulation factors have entered into a circle of continuous increase in the demand and burden of required investigations, as well as in the number of these rare patients to be included in pre-licensing trials. The aim of this article is to examine whether or not this increase in regulatory demands is scientifically justified and is likely to improve efficacy and safety in patients with haemophilia who actually enjoy, at least in high income countries, a life expectancy similar to that of the general male population [1, 2].

Historical overview


In the early 1980s, the dramatic onset of the AIDS epidemics among patients with haemophilia increased awareness of the risk of transmission by replacement therapy of blood borne viral pathogens (HIV, hepatitis B or C) and led to a cogent demand for regulatory actions meant to improve the viral safety of plasma products. This focus on viral safety generated a seminal position paper (III/5830/93) approved in 1993 by the Committee for Medicinal Products for Human use (CPMP) of the European Medicine Agency (EMEA), which recommended the adoption in the manufacturing process of coagulation factors of methods of viral inactivation/removal effective against enveloped viruses (HIV, HBV, HCV) and non-enveloped viruses such as the hepatitis A virus. Because any change in the manufacturing process owing to the implementation of virucidal methods may alter the physicochemical structure of coagulation factors, induce loss of coagulant activity and of immunological tolerance leading to the development of inhibitory alloantibodies, additional clinical data on safety and efficacy were clearly needed, which were detailed in the CPMP/198/95 guidelines approved in February 1996, following proposals made by the International Society of Thrombosis and Haemostasis (ISTH).

These regulatory requirements regarding the licensing of new products were as follows:

  • A pharmacokinetic study, including at least 12 patients (of 12 or more years of age) with severe haemophilia A and measuring half-life, recovery and safety parameters. Patients should continue treatment for 6 months, and at least five of them had been re-tested for half-life and recovery after 3–6 months;
  • Clinical efficacy on bleeding episodes had to be evaluated enrolling at least 30 previously treated patients (PTPs), followed up for at least 6 months with at least 10 cumulative exposure days (an exposure day is defined as a calendar day during which one or more infusions of coagulation factor product are administered). These patients were also monitored with laboratory tests for inhibitors (every 3 months) and various viral markers (HIV, hepatitis A, B and C, parvovirus B19);
  • At least 20 previously untreated patients (PUPs) had to be followed up for at least 2 years (to evaluate viral safety) and 100 exposure days or 5 years (to evaluate immunogenicity). This study was initiated prior to product licensing, but could be completed in the setting of post marketing surveillance;
  • At least 50 PTPs had to be followed up for at least 2 years during the post marketing surveillance period, monitoring clinical efficacy, immunogenicity and safety.
  • Regarding products already authorized, but with added changes in the manufacturing process, the main requirement was that the previous product had to be used as a control in the pharmacokinetic trial.

All in all, these early guidelines appear to us still valid and are able to guarantee with good likelihood full efficacy and safety of new FVIII products.

CPMP/198/95 revision 1

Until the early 1990s, general knowledge on the natural history of the development of FVIII inhibitory alloantibodies was that this complication develops afresh in up to 35% of newly diagnosed patients (PUPs) at an early age of 2 to 3 years, more frequently after 10 to 20 days of exposure to FVIII, less frequently between 20 and 50 exposures and seldom in multiply treated patients3. PTPs with severe haemophilia who had multiple exposures to FVIII (usually defined 100–150 lifetime or more) develop inhibitors at a small rate throughout life (less than 10 inhibitors per 1000 treatment years) [3-5].

This widely accepted knowledge on the natural history of FVIII inhibitors was challenged in the early 1990s, when an outbreak of inhibitors occurred in Belgium and the Netherlands in PTPs who had changed their routinely used plasma-derived FVIII for a newly manufactured product. A higher than expected incidence of clinically relevant FVIII inhibitors was independently detected in Belgian and Dutch patients, who previously had at least 200 days of FVIII exposure. In the Belgian cohort of 109 patients, the incidence was 66 per 1000 patient-years of observation, in the Dutch cohort of 144 patients 20 per 1000 patient-years. These incidences compare unfavourably with the historical incidences observed in PTPs, always smaller than 10 per 1000 [3-5].

This outbreak in PTPs remained isolated and was shown to be caused by that product with peculiar physicochemical features related to methods used for fractionation and viral inactivation, and inhibitors disappeared spontaneously or after immune tolerance induction when patients stopped the incriminated product [6, 7]. Yet, this observation marked a milestone in the history of development of clinical guidelines, because it did turn from pathogen to inhibitor-risk the focus of regulatory agencies, which became newly concerned that new fractionation and viral inactivation methods would trigger the development of FVIII inhibitors in tolerant patients previously treated multiple times.

The Belgian–Dutch epidemics led to the decision that PTPs were the most appropriate patient population to assess the immunogenicity of new FVIII products, and hence to a revision of the CPMP/BPWG/198/95, approved in October 2000. As compared to the former version, the main modifications were as follows:

  • There was no longer any requirement to enrol PUPs;
  • Patients included in the pharmacokinetic trial had to continue treatment with the same product for at least 6 months. As changes in pharmacokinetic parameters (incremental recovery, half-life, area under the curve, clearance and mean residence time) could give a clue about the occurrence of weak inhibitors, the same patients had to be retested for the same pharmacokinetic parameters after 6 months, using the same dose as in the initial study;
  • Clinical efficacy and overall safety should be reported in a minimum of five haemophilia A patients undergoing at least 10 surgical procedures;
  • At least 50 PTP should be followed up for at least 50 exposure days or 6 months for FVIII inhibitors (determined every 3 months);
  • A paediatric trial should be initiated before product submission for licensing and should include a minimum of 12 patients under the age of 6 years regardless of prior treatment. They had to be followed up for clinical efficacy, immunogenicity and safety parameters until they had received at least 50 exposures to the product or had been treated for 6 months (whichever came first);
  • The protocol of a post-marketing surveillance study should be submitted together with the licensing product file.


The aforementioned 1995 European guidelines dealt with regulatory requirements regarding both FVIII and factor IX, but it was subsequently decided to have separate protocols for FVIII and factor IX products. EMA/CHMP/BPWP/144533/2009 was devoted to clinical investigations required for FVIII, and dealt with both plasma-derived and recombinant products. The 2009 guidelines took into account the conclusions of an EMA expert meeting on the risk of inhibitor development, held in 2006 and gathering specialists from EU, USA, Japan, Canada, representatives from the ISTH, the World Health Organization, patient organizations and manufacturers.

As compared to previous guidelines, the main changes were as follows:

  • Pharmacokinetic data were now required for 12 paediatric patients under the age of 6 years;
  • Pharmacokinetic and safety data were also required for paediatric patients between 6 and 12 years age;
  • Clinical, immunogenicity and safety data were required for 50 haemophilic children, allocated in two cohorts of 25 patients each, one under the age of 6 years and the other between 6 and 12 years;
  • Still expected to include a total of 200 patients, the post marketing study should include 60 paediatric PTPs under the age of 12 years;
  • There was a new shift regarding PUPs, because: clinical pre-approval study was now presented as necessary in this population, including 50 patients and post approval follow-up was required in a total of 100 patients (allowing the inclusion of those investigated in the pre-approval study).


Over a period of time when no major discovery has been made regarding the viral, immunogenicity or safety hazards of FVIII products (barring the peculiar, unprecedented and isolated Belgian–Dutch inhibitor outbreak in PTPs), the regulatory requirements concerning the overall size of the severe haemophilia A population spread from 112 patients in 1995 up to 341 patients in 2009, of whom 134 should be less than 12 years of age with further stratification between less than 6 years and between 6 and 12 years. The duration of cumulative exposure was also extended from 10 to 50 days for clinical efficacy studies. It must be emphasized that these are requirements not only for new FVIII products but also for any change in the manufacturing process of licensed products, which in my opinion should be seen as a continuous quality improvement process of any medical product and should occur regularly over time. All in all it can be easily grasped how difficult is to enrol such a large number of rare voluntary patients within a reasonable period of time. This problem is of course much more dramatic for haemophilia B patients with factor IX deficiency, which occurs in males at a rate of 1 in 30 000.

It must be emphasized that the focus of current guidelines on FVIII inhibitors does not stem as a new concern, but rather as a shift from the primary safety concern due to viral transmission. The latter may be now viewed as basically solved, because the measures implemented for plasma selection, as well as the two or more inactivation/removal procedures currently adopted by most manufacturers, are highly effective to optimize the safety of plasma-derived products pertaining to enveloped viruses [8]. Viral inactivation methods are also added to the manufacturing process of recombinant products, even though no transmission of infectious pathogens has ever been documented. Potential transmission of emerging non-enveloped viruses highly resistant to such inactivation methods as heating and solvent/detergent cannot be adequately assessed by means of the clinical studies recommended by regulatory agency (hence the need for long-term post marketing follow-up). However, this theoretical risk cannot be reduced by an increase in the number of recruited patients.

A second objection concerns the rationale used by the regulators to select sample size. The recommended number of PTPs is definitely inadequate to establish whether or not new products carry a risk of inhibitor higher than that predicted in this population on the basis of current knowledge on the natural history of this complication [3-6]. Assuming, on the basis of recent epidemiological data [5], an incidence of new inhibitors of 5.5 per 1000 treatment years (95% confidence interval 4.6–6) a huge sample size of more than 14 000 PTPs would be required to have an 80% power to detect a 50% increase in inhibitor incidence, and more than 95 000 patients to demonstrate with a 20% boundary of equivalence that there is no increase in inhibitors (CR Hay, personal communication). This is the numerical penalty for a small effect on a low-incidence complication, so that adequately powered information is perhaps only possible by pooling results of long-term pharmaco-vigilance programmes such as, for instance, the European Haemophilia Safety Surveillance (EUHASS). The currently required sample size is definitely adequate to detect a very strong signal of heightened immunogenicity such as that detected in the frame of the Belgian–Dutch outbreak. However, such a strong signal would be equally detected by a much smaller number of patients, such as for instance those recommended in the CPMP/198/95.

A third objection concerns the risk of inhibitor in PUPs and the new CPMP decision of enrolling again these patients. It is now generally agreed that this multifactor risk, besides being potentially related to the product used, may also be related to an array of genetic or environmental factors such as the type of gene mutations [9], familial history of inhibitors [10], ethnic origin [11], HLA system [12] and intensity of treatment [13]. Recognizing that a multiplicity of patient-related factors play a significant role on the risk of inhibitors in PUPs, besides the type of FVIII products, one may wonder whether or not shifting towards additional investigations is scientifically justified or practically feasible in this rare population (a new child with severe haemophilia is expected yearly among as many as 3 million births). Most importantly, the number of PUPs actually required by CPMP is unable to provide useful information on the risk of inhibitor associated with new products, because the rate of inhibitors in PUPs ranges from as little as 0% to as much as 38% [14].


It must be borne in mind that all available FVIII products (plasma-derived or recombinant) still carry a small, but incompressible risk of transmission of infectious pathogens, because both are biological agents. However:

  • A major safety progress was made with methods of viral inactivation for plasma-derived FVIII products [8]. The current focus on the risk of inhibitors is a direct consequence of this revolution in pathogen safety, and does not come from a novel awareness of an immunological risk neglected by previous guidelines;
  • On the basis of the aforementioned data, the increase in the regulatory demands regarding the development of new FVIII products is an undue one and is not likely to carry a benefit in terms of patient safety. The rarity of the haemophilias makes it more and more difficult to recruit a number of patients sufficient to allow adequately powered statistical analyses, and this is true for both PTPs and PUPs.
  • The development of inhibitors in PTPs may occur very late after the introduction of a FVIII product, supporting post marketing studies and registries rather than huge, but relatively short-term clinical studies to monitor such long-terms hazards.
  • Due to the continuous improvement in manufacturing process, the lifecycle of the products for the treatment of haemophilia A is fairly short.

Rather than endlessly increasing their requirements in terms of sample size or duration of follow-up, regulatory agencies should be better inspired to work together on the harmonization of international guidelines, because the US requirements are somewhat different from European requirements. As suggested by Aledort [15] a move towards harmonization would permit meta-analyses of available data and represent a major step towards statistically and clinically valid assessment of the risk of haemophilia treatment, especially that of inhibitor development.


I am grateful to Françoise Rossi from the International Plasma Fractionation Association (IPFA) who supplied me with useful information on CPMP guidelines and useful criticism and advice on the text of this article.


In the last 2 years, the author has been acting as consultant in education activities for the Bayer Awards, and has received honoraria for speaking at meetings organized by Bayer, Biotest, Grifols, Novo Nordisk and Pfizer.