The potential for prophylaxis to alter the natural history of severe haemophilia has been demonstrated in retrospective cohort studies [23–26] and a prospective randomized study . However, debate continues regarding the exact timing and optimal prophylaxis regimen for patients with severe haemophilia A. The rationale for prophylaxis was originally devised following the observation that patients with moderate haemophilia (FVIII/IX 1–5 IU dL−1) had fewer haemarthroses and were less prone to arthropathy than patients with severe haemophilia [23,24,28]. These observations led to the hypothesis that maintaining FVIII/IX above 1 IU dL−1 would achieve the desired phenotypic changes in both bleed number and long-term preservation of musculoskeletal function.
Pharmacokinetic parameters and break through bleeds during prophylaxis
The proven success of prophylaxis may depend predominantly on maintaining an adequate trough level and limiting the time per week with a factor level below a certain level, as originally suggested. Hypothetically, however, there may also be a role for the area under the factor level vs. time curve (AUC), a measure of how much coagulation factor a person is exposed to, or for recurrent high peak levels in treating early subclinical bleeds. It is possible that the parameter that is most important for the efficacy of a regimen depends on whether prophylaxis is mainly aimed at preventing clinically evident spontaneous bleeds, preventing trauma or sport-induced bleeds or preventing subclinical bleeds. The importance of the trough, AUC and peak may differ depending on the circumstances.
The concept that the trough level is an important determinant of bleeding has been supported by observational data which have shown that the time per week with FVIII/IX levels less than 1 IU dL−1 is associated with an increased rate of bleeding [29,30]. One of these studies  examined and found no association between bleeding and area under the FVIII curve per week suggesting that once the FVIII level is above a certain threshold, no further benefit in bleed prevention is gained. Studies that investigate the effect of peak levels have not been reported. In addition, no data are available that investigate FIX prophylaxis specifically. These data do not imply that a FVIII/IX level of 1 IU dL−1 is a critical level in all patients. In a cohort of 34 children, e.g. 79% had a trough below 1 IU dL−1; but despite this, 59% had no clinical evidence of haemarthrosis during 1-year follow-up and there was no difference in the number of bleeds when comparing those with trough levels below or above 1 IU dL−1 . The data are best interpreted as support for the hypothesis that the longer a patient spends with a low FVIII/IX level, the higher their risk of bleeding, while at the same time recognizing that the event of a bleed depends on many other factors, e.g. physical activity, trauma, the state of the underlying joint and how the patient’s underlying haemostatic system responds to replacement therapy. There is also debate as to whether the same level of FVIII or FIX has an identical effect on the haemostatic system [32–35], and it is possible that adequate trough levels for prophylaxis may differ between the two disorders.
If it is accepted that the time per week with a low coagulation factor level plays a role in a patient’s response to prophylaxis then the inter-patient variation in PK is potentially very significant. However, it is important to recognize that a significant determinant of the time per week with low FVIII is adherence to the prescribed prophylactic regimen . Strategies to improve adherence would be expected to decrease the number of bleeds, whereas poor adherence make PK dose tailoring irrelevant.
Pharmacokinetic implications of dosing
The implications that PK has for prophylactic treatment have been previously reviewed [4,6,10–12]. Initially, simulations demonstrated the potential for more cost-effective dosing . Subsequently, a study on 21 patients with severe haemophilia A showed that prophylaxis aimed at targeting a trough level decided by the clinician, based on PK data (based on seven blood samples over 48 h), compared to standard dosing, resulted in a higher mean trough level (2.2 vs. 0.9 IU dL−1) and reduced FVIII usage (mean 85 000 vs. 124 000 IU in 6 months) . There was no observable difference in the number of bleeds; however, the study lacked statistical power to draw a firm conclusion in this respect. A study on eight patients with severe haemophilia B showed similar results with significantly decreased usage of pdFIX for maintaining an adequate trough level if patients were treated every third day rather than twice a week. Even more cost-effective treatment was possible if treatment was given on alternate days . A simulation study using data from 55 patients treated with rFIX (BeneFix®) implied that annual consumption to maintain a 1 IU dL−1 trough level could be decreased from on average 4700 IU kg−1–2400 IU kg−1 by changing from an every third day to an alternate day dosing schedule .
Building on these findings, a modelling study using representative FVIII PK data has been performed . These simulations demonstrate that the trough level and time per week with FVIII less than 1 IU dL−1 are affected more by half-life and frequency of infusions and less by recovery and dose kg−1. The data confirm that, if trough levels are clinically important, there will be a large difference in the amount of concentrate kg−1 that patients require for successful prophylaxis.
Using representative PK data , it was calculated that the time taken to reach 1 IU dL−1 following a standard infusion of 30 IU kg−1 in children aged 1–6 years would vary between 43 and 77 h (a difference of 44 h) when comparing children with the shortest and longest FVIII half-lives. In older patients, the difference would be 59 h, ranging from 51 to 110 h (Fig. 1). To look at these data in a different way, in patients on a prophylactic regimen of 30 IU kg−1 on alternate days, the trough FVIII level in the average 1–6 year-old would be 1.7 IU dL−1. In those with the longest half-lives, the trough would be 4.7 IU dL−1, whereas those with the shortest half-life would spend 17.5 h per week with FVIII less than 1 IU dL−1 . This suggests that standard prophylactic regimens may not be appropriate for all patients and that knowledge of half-life, in addition to observation of the bleeding pattern, may help tailor prophylaxis to individual patients. Similar calculations for recovery show that this parameter has a proportionally much smaller effect than half-life .
The frequency of infusions, whilst keeping the total dose of coagulation factor constant, has a large effect on trough levels in patients treated with prophylaxis for both FVIII and IX [5,7–9,13]. If the effect of the half-life and the frequency of dosing are combined, then widely variable amounts of FVIII would be required to maintain the trough FVIII above a predetermined level. Data presented in Table 1 are adapted from a previous publication  and summarize the amount of FVIII required in an average adult to maintain a trough FVIII between 1 and 1.5 IU dL−1 depending on half-life and dose frequency.
Table 1. Factor VIII requirements depending on dose schedule and half-life.
|Half-life*||Dose of factor VIII per infusion required to maintain trough level of 1–1.5 IU dL−1 in 70 kg man (IU)|
|Daily dosing||Alternate day dosing||Every third day dosing|
|Short: 7.5 h (5th percentile of normal range)||240||2420||22410*†|
|Median: 10.4 h||120||700||3570|
|Long: 16.5 h (95th percentile of normal range)||50||200||600|
In these simulations, the dose of FVIII required to maintain a trough level between 1 and 1.5 IU dL−1 in the average adult varied 30-fold when comparing daily with every third day dosing. The effect of half-life is the largest if every third day regimens are used with a 37-fold difference in the amount of FVIII required when comparing the shortest and longest half-lives. This is in contrast to a 12-fold and fivefold difference when alternate day or daily dosing is used. The effect of half-life is, therefore, exaggerated by less frequent dosing and knowledge of a patient’s FVIII half-life will potentially have a significant impact on the prescription of prophylactic regimens, especially in adult patients.
In contrast to changing the frequency of dosing, increasing the dose kg−1 of FVIII for each prophylactic infusion has a smaller effect on the trough level. For example, if a certain dose results in a trough of 1 IU dL−1 at 48 h, then doubling the dose would result in a trough of 2 IU dL−1.
To date, there is no corresponding simulation study on FIX, due to lack of data on the variance of PK (in particular on pdFIX) in a representative population of patients. In addition, FIX is characterized by marked ‘two-compartment PK’, with a rapid initial and a slow terminal half-life [10,36]. Thus, dose requirements cannot be calculated as simply as for FVIII with a single half-life as the main determinant. There is also a marked discrepancy in PK between pdFIX and rFIX. Comparisons between the methodologically most robust studies on either species indicate that the average CL of recombinant FIX is twice as high as that of pdFIX . This difference is confirmed in cross-over comparisons [37,38]. The typical elimination half-life of recombinant FIX is 17–23 h [9,37,39–41] as compared with approximately 30 h for pdFIX [5,10,36]. The difference in CL between rFIX and pdFIX appears to be greater and more consistent between studies, than the difference in half-life, which demonstrate that the comparison cannot be based on only a single PK parameter. These differences heavily influence calculated dose requirements for prophylaxis. The median dose to maintain a 1.5 IU dL−1 trough level of pdFIX in eight adult patients was 1000 IU every third day or 500 IU alternate days . As re-calculated from  the average doses of rFIX would be about 3800 IU every third day and 1250 IU alternate days – an approximately threefold difference over all. In fair agreement with these separate estimations, the two cross-over studies [37,38] demonstrated that FIX plasma levels 48 h after a dose of rFIX were only approximately 50–70% of those obtained with the same dose of pdFIX. This can be directly translated to a 1.5- to 2-fold increase in dose requirement during prophylactic treatment.