Randomized, prospective clinical trial of recombinant factor VIIa for secondary prophylaxis in hemophilia patients with inhibitors


Barbara A. Konkle, Penn Comprehensive Hemophilia Program, University of Pennsylvania, Medical Arts Bldg, Suite 103, PMC, 39th Market St, Philadelphia, PA 19104, USA.
Tel.: +001 215 662 9248; fax: +001 215 243 4621; e-mail: barbara.konkle@uphs.upenn.edu


Summary. Background: Hemophilic patients with factor VIII (FVIII) and FIX inhibitors suffer from frequent bleeding episodes and reduced quality of life. Objectives: To evaluate whether secondary prophylaxis with activated recombinant factor VII (rFVIIa) can safely and effectively reduce bleeding frequency as compared to conventional on-demand therapy. Methods: Thirty-eight male patients entered a 3-month preprophylaxis period to confirm high baseline bleeding frequency (mean ≥ 4 bleeds per month). Twenty-two patients were randomized 1:1 to receive daily rFVIIa prophylaxis with either 90 or 270 μg kg−1 for 3 months, followed by a 3-month postprophylaxis period. Results: Bleeding frequency was reduced by 45% and 59% during prophylaxis with 90 and 270 μg kg−1, respectively (< 0.0001); however, there was no significant difference detected between doses. The majority of this reduction was maintained during the postprophylaxis period. Although all types of bleed were similarly reduced, the effect was most pronounced for spontaneous joint bleeds. Patients reported significantly fewer hospital admissions and days absent from work/school during prophylaxis as compared to the preprophylaxis period. No thromboembolic events were reported during prophylaxis. Conclusion: Clinically relevant reductions in bleeding frequency during prophylaxis as compared to conventional on-demand therapy were achieved without raising safety concerns. These results provide evidence for the concept of secondary rFVIIa prophylaxis in inhibitor patients with frequent bleeds.


Patients with hemophilia and high-titer inhibitors against factor VIII (FVIII) or FIX experience severe morbidity, including persistent or recurrent episodes of joint, muscle and deep tissue bleeding, which can be life- and/or limb-threatening [1].

If inadequately treated, joint bleeds lead to synovial inflammation and hypertrophy, and predispose to recurrent bleeding with progressive damage to the cartilage and subchondral bone [2]. In the long term, inadequate treatment of joint bleeds may result in progressive severe arthropathy being painful and disabling [3]. General health status and quality of life have been found to be significantly reduced among inhibitor patients as compared to the general population, because of impaired orthopedic status [4]. Many such patients frequently experience pain, have frequent hospital visits and considerable absence from school or work, and require devices such as wheelchairs and crutches to facilitate mobility [5].

The main goal of hemophilia treatment is to decrease joint bleeds and to arrest the progression of joint destruction, in order to increase the quality of life. Conventional on-demand therapy administered to hemophilic patients is characterized by frequent bleeding episodes and the development of joint damage [3,6–8]. Secondary prophylaxis has been demonstrated to be a promising alternative to on-demand therapy in hemophilic patients without inhibitors [6–9] that allows the possibility of reducing bleeds and thereby slowing the rate of joint destruction [10]. To date, the evidence for the benefits of secondary prophylaxis as compared to on-demand treatment of hemophilic patients with inhibitors is derived solely from observational studies and clinical case reports assessing the efficacy and safety of prophylactic treatment with bypassing agents [i.e. activated recombinant FVII (rFVIIa) (Novo Nordisk A/S, Bagsvaerd, Denmark) and plasma-derived activated prothrombin complex concentrates (aPCCs) [9,11–16]. Although prophylactic treatment with rFVIIa and aPCCs suggests that this treatment paradigm may be efficacious, the evidence is based on very few patients, and no standard treatment is defined [9,11–16].

The present article presents results from the first prospective, multicenter, randomized, double-blind clinical trial designed to explore the efficacy and safety of secondary prophylaxis with rFVIIa in patients with congenital hemophilia A or B with inhibitors and high requirements for on-demand therapy.

Patients, materials and methods

Trial design

The present trial was a multicenter, randomized, double-blind, parallel-group trial investigating the efficacy and safety of secondary rFVIIa prophylaxis in patients with congenital hemophilia A or B with inhibitors and high requirements for on-demand therapy. The trial consisted of a preprophylaxis period, a prophylaxis (treatment) period, and a postprophylaxis period, each of 3 months’ duration. This trial was conducted in 20 sites in 11 countries: Argentina (one site), Brazil (two sites), Bulgaria (one site), the Philippines (one site), Poland (two sites), Romania (four sites), Russia (one site), South Africa (one site), Spain (three sites), Turkey (one site), and the USA (three sites). The trial protocol and the patient information sheet were approved by health authorities and/or ethics committees according to local regulations, prior to trial initiation. All patients were recruited from the investigators’ patient files. Written informed consent was obtained from the patients themselves, or the patient’s parents or legally acceptable representative (if applicable) before any trial-related activities. An independent, blinded Steering Committee consisting of external experts was established to provide scientific and medical monitoring during the entire trial. The trial was conducted in accordance with the Declaration of Helsinki [17] and International Conference on Harmonization/Good Clinical Practice guidelines for clinical investigations [18].


Males with severe congenital hemophilia A or B with a high historical inhibitor titer (with an inhibitor titer > 2 BU mL–1 in the preceding 12 months), a requirement for current treatment of bleeds with bypassing agents, and at least four bleeds requiring hemostatic drug treatment (except dental bleeds and bruises) within the previous month were eligible for inclusion. Key exclusion criteria prior to trial entry included: prophylaxis with any hemostatic drug within the last 3 months, immune tolerance induction (ITI) within the last month, known pseudotumors, platelet count < 50 000 μL–1, advanced atherosclerotic disease, and congenital or acquired coagulation disorders other than hemophilia A or B.


Following screening, 38 eligible patients underwent a 3-month preprophylaxis observation period to determine baseline bleeding rate and exclude non-frequent bleeding patients. All patients who experienced at least two bleeds per month, and a total of ≥ 12 bleeds requiring hemostatic drug-based treatment during the observation period, entered the 3-month prophylaxis period, provided that they were not scheduled for surgery in the subsequent 3 months.

A centralized, computer-generated randomization list was used to randomly allocate patients to receive either 90 or 270 μg kg−1 rFVIIa (Novoseven; Novo Nordisk A/S, Copenhagen, Denmark) once daily for 3 months. Each rFVIIa dose was to be self-administered before 11 am in a home setting as a slow bolus i.v. injection over a period of 2 min. Blinding was maintained by providing an equal volume of trial drug to be injected in both treatment groups. Concomitant administration of other hemostatic drugs was permitted during the entire trial period, except from 1 h prior to and until 2 h after rFVIIa administration during the prophylaxis period. The 3-month prophylaxis period was followed by a 3-month postprophylaxis period.

Patients and physicians were advised to continue on-demand treatment of acute/breakthrough bleeds according to normal practise throughout the trial. The hemostatic drugs used for on-demand treatment were recorded throughout the trial.


The primary efficacy endpoint was number of bleeds per month during the prophylaxis period as compared to the preprophylaxis period. A bleed was defined as rebleeding if it occurred at the same site within 6 h of treatment, whereas episodes beginning 6 h after treatment or occurring in another site were defined as a new episode. Secondary efficacy endpoints included the number of bleeds per month occurring in the postprophylaxis period as compared to those observed in the observation and prophylaxis period, at specific bleeding sites (target joint, joint, muscle, soft-tissue bleeds), and cause of bleed (traumatic, spontaneous and other) over the entire trial period. Target joints were defined as those joints into which bleeding had occurred ≥ 3 times in the last 6 months. To ensure that on-demand treatment with FVIII products during the trial did not have any impact on the primary endpoint, the endpoint was reassessed, excluding patients who had been treated with on-demand FVIII preparation without anamnestic response. The 24-h bleeding pattern for each treatment group during the entire trial period was explored.


Safety was evaluated by the number and type of adverse events reported during the 9-month trial period, and was graded by severity and seriousness as well as probable relation to the trial product.

Health-related quality of life

Health-related quality of life was evaluated on the basis of pharmacoeconomic endpoints [hospital days, days absent from school or work, days requiring mobility aids (e.g. wheelchair, crutches, and with cane)], and by the widely used self-administered questionnaire EuroQoL health profile (EQ-5D) [19]. The EQ evaluation is a validated generic questionnaire used in health economic calculations, but is not specific for patients with hemophilia; no hemophilia-specific questionnaires were available at the time of trial initiation. The EQ-5D profile, consisting of five domains (mobility, self-care, usual activities, pain/discomfort and anxiety/depression) and three levels (‘no problem’, ‘some or moderate problems’, and ‘extreme problems/impossible to do’), was adopted [19].

Orthopedic status

The orthopedic status was evaluated by the patient’s doctor using an orthopedic joint score system based on Gilbert’s joint scoring system [20]. Orthopedic joint scores (including pain) ranging from 0 (better orthopedic condition) to 15 (worse condition) were assigned following physical examination of joints (ankle, knee, and elbow) carried out at trial entry and at the end of each trial period. Amplitude of joint movement was determined with a goniometer. No X-ray examinations of the joints were performed.

FVII measurements

Blood samples were collected before (after 3 months of observation), during (after 1 month of rFVIIa exposure) and at the end of the prophylaxis period (after 3 months of rFVIIa exposure). FVII:C trough levels were determined by the FVII:C assay, which determines the FVII activity in a one-stage clotting assay [21], using recombinant human thromboplastin tissue factor (Innovin; Dade Behring, Marburg, Germany), immunodepleted FVII-deficient plasma (American Diagnostica; Greenwich, CT, USA) and a normal human plasma pool as calibrator.


Sample size calculations were based on statistical simulations using a Poisson distribution, in which four bleeds per month per patient during the preprophylaxis period and a 40% reduction in number of bleeds per month (for treatment groups combined) during treatment as compared to the preprophylaxis period were assumed. With an expected withdrawal rate < 30% during the prophylaxis period, 10 patients were calculated to be required to detect this difference with 80% power and 5% significance level. As the trial involved two treatment groups, a total number of 20 patients was planned. Statistical analyses were performed with the intention-to-treat population, defined as all patients randomized and exposed to at least one dose of trial product.

A logistic regression model was used to analyze changes in number of bleeds per month. The model included the ratio of number of days in each trial period as offset. To minimize interpatient variations, each patient served as his own control. The estimated changes were tested for statistical significance using Wald’s test. To compare the estimated changes between the two treatment groups, a likelihood ratio test was used. No formal analysis was applied to compare adverse events. Health economic variables were analyzed using a sign test combining treatment groups. For the overall development in the orthopedic joint score throughout the trial, a Wilcoxon signed rank test was applied. FVII:C over time was analyzed by an anova including visit and patient as factors. P-values < 0.05 were considered to be significant.


Patient characteristics and demographics

Thirty-eight patients were recruited for the trial; 37 patients entered the preprophylaxis period, of whom 22 were randomized and received trial product for 3 months (Fig. 1). One patient withdrew informed consent prior to entering the preprophylaxis period. Fifteen of the remaining 37 patients were withdrawn after the initial preprophylaxis period, 13 because of insufficient number of bleeds, one because of lack of treatment of bleeds, and one because of coagulopathy resulting from existing hepatitis that was reported as an adverse event before FVII prophylaxis treatment. All 22 rFVIIa-exposed patients (21 with hemophilia A and one with hemophilia B) completed the trial (Table 1). Considering the small sample size, baseline characteristics were roughly comparable, although more adults (four vs. two) and fewer children (two vs. four) were randomized to the 270 μg kg−1 treatment group than to the 90 μg kg−1 treatment group. Accordingly, median body weight was slightly higher in the 270 μg kg−1 rFVIIa treatment group (66.0 kg) than in the 90 μg kg−1 rFVIIa treatment group (51.4 kg).

Figure 1.

 Trial flow diagram. §One of the randomized patients received 270 μg kg−1 rFVIIa for three months, although he violated one of the exclusion criteria by having a pseudotumor. The tumor was evaluated to be inactive; therefore the Steering Committee considered this patient eligible for entering the prophylaxis period.

Table 1.   Patient characteristics
 90 μg kg−1 recombinant factor VII (rFVIIa) 270 μg kg−1 rFVIIaTotal
  1. Target joint: any joint into which bleeding had occurred ≥ 3 times in the last 6 months.

  2. The orthopedic joint score was based on pain and examinations of joints (ankle, knee, and elbow).

Median age (range) (years)13.0 (5.1–50.5)17.8 (10.6–56.1)15.7 (5.1–56.1)
 < 12 years [no. (%)]4 (36)2 (18)6 (27)
 At least 12 years, but < 18 years [no. (%)]5 (46)5 (46)10 (46)
 At least 18 years [no. (%)]2 (18)4 (36)6 (27)
Median body weight (range) (kg)51.4 (17.4–75.0)66.0 (26.0–79.2)54.0 (17.4–79.2)
Hemophilia type [no. (%)]
 A10 (91)11 (100)21 (95)
 B1 (9%)0 (0%)1 (5)
Target joint [no. (%)]
 Yes10 (91)11 (100)21 (95)
 No1 (9%)0 (0)1 (5)
Orthopedic joint score (all joints)
 Median (range)1.67 (0.00–3.80)1.92 (0.33–9.00)1.83 (0.00–9.00)
Orthopedic joint score (target joints)
 Median (range)2.75 (2.00–5.33)3.67 (0.00–12.00)3.00 (0.00–12.00)

Effect of prophylaxis: overall bleed frequency

During the entire 9-month trial period, 821 bleeds were reported; 558 of which were spontaneous episodes (see Table 2 for details). The majority of bleeds were into target joints (= 440). The percentage of treatment period days where trial product was reported as taken ranged from 84.5% to 100% in the 90 μg kg−1 treatment group and from 96.7% to 100% in the 270 μg kg−1 treatment group.

Table 2.   Overview of bleed counts
Trial periodBleeds90 μg kg−1 recombinant factor VII (rFVIIa)270 μg kg−1 rFVIIaTotal
n%BMedium (range)n%BMedium (range)n%BMedium (range)
  1. n, number of patients with bleeds; %, percentage of total number of bleeds; B, number of bleeds; Medium (range), the median and range of number of bleeds per subject.

PreprophylaxisTotal11100212 11100196 22100408 
Target joint105912611.5 (5–22)1142829 (1–10)215120810 (1–22)
Spontaneous116814515 (1–23)116713111 (1–32)226827612.5 (1–32)
ProphylaxisTotal11100106 1110075 22100181 
Target joint1064684 (1–20)1051382.5 (1–16)20591063 (1–20)
Spontaneous1066706 (1–16)1072544 (1–18)20691245 (1–18)
Post-prophylaxisTotal11100137 1110095 22100232 
Target joint1055755 (2–22)954515 (1–15)19541265 (1–22)
Spontaneous1169958 (1–26)1166634 (1–17)22681586.5 (1–26)

During the prophylaxis period, treatment with 90 μg kg−1 rFVIIa significantly reduced bleeds per month from 5.6 to 3.0 (Fig. 2). Treatment with 270 μg kg−1 rFVIIa reduced the bleeding frequency from 5.3 to 2.2 bleeds per month. A significant reduction in bleeding frequency was maintained after treatment with both 90 and 270 μg kg−1 rFVIIa (4.1 and 2.7, respectively) during the postprophylaxis period. The effective reduction in bleeding frequency with rFVIIa prophylaxis as compared to the preprophylaxis period was 45% with the 90 μg kg−1 dose (< 0.0001) and 59% with the 270 μg kg−1 dose (< 0.0001). During the postprophylaxis period, the reduction in the number of bleeds per month persisted, and the number of bleeds remained significantly reduced as compared to the preprophylaxis period for both treatment groups (90 μg kg−1 rFVIIa, 27%, < 0.01; 270 μg kg−1 rFVIIa, 50%, < 0.0001). Although a similar reduction was observed with all types of bleed, the effect was most pronounced for spontaneous joint bleeds (Figs 3 and 4).

Figure 2.

 Number of bleeds per month. The bracketed data are the estimated changes (percentage) in number of bleeds per month (defined as 28 days) for the 90 and 270 μg kg−1 recombinant factor VIIa treatment groups during the prophylaxis or postprophylaxis period as compared to the preprophylaxis period, and during the prophylaxis period as compared to the postprophylaxis period. *** 0.001; ** 0.01; * 0.05.

Figure 3.

 Number of bleeds by site per trial period. The bracketed data are the estimated changes (percentage) in number of joint bleeds per trial period for the 90 and 270 μg kg−1 recombinant factor VIIa treatment groups during the prophylaxis period (P) and postprophylaxis period (post) as compared to the preprophylaxis period (pre), and during the postprophylaxis period as compared to the prophylaxis period. *** 0.001.

Figure 4.

 Number of bleeds by type per trial period. The bracketed data are the estimated changes (percentage) in number of spontaneous bleeds per trial period for the 90 and 270 μg kg−1 recombinant factor VIIa treatment groups during the prophylaxis period (P) and postprophylaxis period (post) as compared to the preprophylaxis period (pre), and during the postprophylaxis period as compared to the prophylaxis period. *** 0.001; * 0.05.

Six patients (90 μg kg−1 rFVIIa, two patients; 270 μg kg−1 rFVIIa, four patients) were treated for bleeding episodes with FVIII preparations sporadically throughout the 9-month trial period without any documented anamnestic response. When these patients were excluded from the analysis, the results obtained were essentially the same as those of the full analysis set. Bleeding frequency during the prophylaxis period as compared to the preprophylaxis period was reduced by 48% in the 90 μg kg−1 dose group (< 0.0001) and by 64% in the 270 μg kg−1 dose group (< 0.0001) from the preprophylaxis to the prophylaxis period.

Effect of prophylaxis: target joint bleeds

During the preprophylaxis period, 408 joint bleeds were reported, 208 of which were target joint bleeds (Table 2). When compared using the same logistic regression model used for comparing changes in the overall bleed frequency (above), target joint bleeds were significantly reduced by 43% (< 0.001) and 61% (< 0.0001) during prophylaxis in the 90 and 270 μg kg−1 rFVIIa treatment groups, respectively, as compared to the preprophylaxis period. Similarly, the significant reduction in target joint bleeds persisted in the postprophylaxis period as compared to the preprophylaxis period for both treatment groups (90 μg kg−1 rFVIIa, 32%, < 0.01; 270 μg kg−1 rFVIIa, 46%, < 0.001). There was no obvious correlation between initial joint score and response to prophylaxis.

Effect of prophylaxis: spontaneous bleeds

The number of spontaneous bleeding episodes was significantly reduced from 276 in the preprophylaxis period to 124 in the prophylaxis period (Table 2 and Fig. 4). This reduction persisted during the postprophylaxis period (= 158).

Individual patient bleeding profiles

Individual patient bleeding profiles are illustrated in Fig. 5A,B. Three patients (90E, 90 J and 90 K) in the 90 μg kg−1 rFVIIa treatment group had a less pronounced overall reduction in bleeding frequency during the treatment and postprophylaxis periods (Fig. 5A). For patient 90E, the number of traumatic bleeds increased in the third month, but the number of spontaneous bleeds remained low. When the data for these three patients are removed from the analysis, the trend for greater efficacy with the 270 μg kg−1 dose in Fig. 2 is no longer apparent (data not shown).

Figure 5.

 Patient individual bleeding profiles. (A) Number of bleeds reported for the individual patients dosed with 90 μg kg−1 recombinant factor VIIa (rFVIIa) (90A–K) during each month (defined as 28 days) through the entire trial period. (B) Number of bleeds reported for the individual patients dosed with 270 μg kg−1 rFVIIa (270A–K) during each month (defined as 28 days) through the entire trial period.

Twenty-four-hour bleeding pattern

During the preprophylaxis period, the majority of bleeds were reported to occur in the morning (7–11 am), whereas during the prophylaxis period, bleeds were more equally distributed over the day. Patients were treated on average at 9:17 am. During the postprophylaxis period, the 24-h bleeding pattern tended to revert to the pattern seen in the preprophylaxis period, but with less frequent bleeds.


Overall, there were no apparent treatment-dependent patterns in number or types of adverse events reported during the 9-month trial period (Table 3). No thromboembolic adverse events occurred. Five serious adverse events were reported; all were judged by the investigator to be unlikely to be related to administration of rFVIIa.

Table 3.   Adverse events
Treatment groupsAdverse eventsPreprophylaxisProphylaxisPostprophylaxis
  1. n, number of patients with adverse events; %, proportion of patients having an adverse event; E, number of adverse events.

  2. The five non-serious adverse events judged as being possibly/probably related to prophylaxis treatment with 90 μg kg−1 rFVIIa were single events of moderate fever, headache and vertigo in one patient (none of which lasted for more than 24 h), and two episodes of allergic dermatitis in one patient – in both cases, the patient had recovered in < 1 week.

  3. §The four serious adverse events in the prophylaxis period were a gastrointestinal hemorrhage (required hospitalization), interstitial lung disease (pneumonia), a fracture of the ulna (required hospitalization), and a viral infection. The serious adverse event in the postprophylaxis period was an infected hematoma (in the patient with pneumonia).

90 μg kg−1 recombinant factor VII (rFVIIa)All adverse events873199823576413
Possibly/probably related0002185000
Serious adverse events000000000
270 μg kg−1 rFVIIaAll adverse events98224873163276
Possibly/probably related000000000
Serious adverse events0004364§191§

Health-related quality of life

Prophylaxis with rFVIIa resulted in a significant reduction in both days absent from school/work and days of hospitalization attributable to bleeding episodes. The effect of rFVIIa did not seem to differ between dose groups. The results of the two dose groups were pooled, because of the limited number of patients in either dose group. The mean proportion of absentee days decreased from 38.7% in the observation period to 16.7% in the treatment period (= 0.0127). Similarly, the mean proportion of days in hospital decreased from 13.5% to 5.9% (= 0.0026). Overall use of mobility aids remained unchanged. Two of the five domains evaluated in the ED-5D questionnaires showed trends towards change over time, with fewer patients reporting pain and mobility problems during the prophylaxis and postprophylaxis periods.

Orthopedic status

Orthopedic joint score remained unchanged throughout the trial (data not shown).

FVII:C measurements

As compared to baseline, higher preinfusion (trough) plasma concentrations of FVII:C were observed during prophylaxis (median concentrations of 0.5, 0.7 and 0.7 IU mL−1 at baseline, 1 month and 3 months, respectively). Eight patients were excluded from this analysis because of significant deviations from the planned blood sampling schedule.

Consumption of hemostatic drugs during the trial

Patients reported on-demand treatment of bleeds with one or more of the following standard hemostatic drugs as required throughout the study: rFVIIa, FEIBA, antifibrinolytics, FVIII, FIX and aPCC. The overall utilization of all hemostatic drugs for on-demand treatment decreased by between 42% and 100% with 90 μg kg−1 prophylaxis as compared to the preprophylaxis period, depending on the drug. Similar decreases were observed for 270 μg kg−1 prophylaxis.


This is the first randomized, double-blind, prospective clinical trial evaluating secondary prophylaxis in patients with congenital hemophilia A and B with inhibitors. The data presented show that secondary prophylaxis of bleeding with rFVIIa has benefits for hemophilic patients with inhibitors. Bleeding frequency was essentially halved during prophylaxis treatment with rFVIIa in both treatment groups as compared to the 3-month preprophylaxis period, during which only conventional on-demand hemostatic therapy was administered. Reductions in bleeding frequency persisted during the 3-month postprophylaxis period. The majority of bleeds were spontaneous bleeds, and a large proportion of these were into target joints. Bleeding frequency was similarly reduced for target joints and other joints. The reductions in bleeding frequency observed in this trial corresponded with important improvements in absenteeism (days absent from work/school) and hospital admission.

No safety concerns were raised in patients receiving daily prophylaxis with 90 or 270 μg kg−1 rFVIIa over 3 months. Although the number of subjects exposed in this trial was limited, no thromboembolic events were reported with either dose. This is in accordance with findings from case reports of severely affected hemophilic patients with inhibitors who suffered no thrombosis during short- or long-term secondary rFVIIa prophylaxis [22].

The observed benefits of rFVIIa prophylaxis in hemophilic patients with inhibitors are consistent with reports of secondary prophylactic treatment in patients without inhibitors [23–25]. Nilsson et al. [22] showed that prophylactic intervention in patients with severe hemophilia led to reductions in both the severity and frequency of bleeding episodes. With a marked reduction in the number and duration of hospital stays, and an overall reduced burden of the severe hemophilia, these patients were able to live almost normal lives. To date, three prospective studies in non-inhibitor patients have confirmed that secondary prophylaxis with factor concentrates significantly reduces both total and joint bleeds as compared to conventional on-demand treatment [6–8].

The effect of daily prophylactic infusions of rFVIIa in reducing bleeding episodes may be explained by sustained, albeit low, ambient levels of rFVIIa in the circulation. The diurnal distribution of bleeds was markedly altered during prophylaxis. With on-demand treatment, during both the observation and the postprophylaxis period, the majority of bleeds reported occurred between 7 and 11 am. By contrast, during the prophylaxis period, there was a reduction in the number of bleeds during this morning interval, which may indicate some degree of residual activity. Alternatively, given that patients were treated on average at 9:17 am, the reduction of bleeds during this period may represent early treatment of bleeding episodes that developed during the night.

Given that the intravascular half-life of rFVIIa is approximately 2.5 h [26], there may be alternative explanations. One hypothesis for the mechanism of persistent action relates to extravascular concentrations of rFVIIa obtained with daily dosing [27–29]. Such extravascular rFVIIa concentrations at the bleeding sites may be high enough to support complex formation with tissue factor, thereby providing enough thrombin to form fibrin plugs to stop minor bleeds in the microvascular compartment, as suggested by Hedner [27].

Interestingly, the reductions in the number of bleeds with treatment persisted throughout the 3-month postprophylaxis period, during which prophylactic infusions were not given and only on-demand therapy was given. This reduction in the number of joint bleeds may be due to a reduction in the inflammatory process associated with chronic synovitis [2,30].

This trial was designed as a clinical proof-of-concept trial, and was not powered to detect either clinically significant or clinically relevant differences between the dose levels. Daily prophylaxis with 90 and 270 μg kg−1 provided similar reductions in bleeding frequency. The observed reduction from five or six bleeds per month in the preprophylaxis period to two or three per month after 3 months of prophylaxis is not clinically optimal by any means. Nevertheless, it does represent a clinically relevant and promising improvement. A detailed evaluation of individual profiles revealed that three patients in the 90 μg kg−1 treatment group failed to show a significant reduction in bleeding episodes during the prophylaxis period. These three patients responded well to the 90 μg kg−1 dose during some of the 3 months, but there was less effect during other months, because of different circumstances (traumatic bleeds, increased activity). In addition, given that the post-treatment period was only 3 months, it is not known whether a higher dose would result in a more prolonged beneficial effect. It can be speculated whether different approaches to dosing (e.g. twice daily or once every other day) and extending the interventional period to more than 3 months would provide further reductions in bleeding frequency. However, no firm conclusions concerning the most appropriate dosing level or dosing regimen can be drawn on the basis of this trial.

In addition to daily administration of rFVIIa, acute bleeds were treated on demand with rFVIIa or other hemostatic drugs according to normal practise. Surprisingly, six patients reported occasional use of FVIII as on-demand therapy in spite of having confirmed requirements for bypassing therapy. Exclusion of these patients did not have a significant impact on bleeding frequency, indicating that on-demand treatment with FVIII did not confer any added benefits in terms of a reduction in the overall bleeding frequency.

The reductions in bleeding frequency and the beneficial safety profile during rFVIIa prophylaxis were associated with important improvements in the patients’ quality of life as compared to conventional on-demand therapy. In particular, rFVIIa prophylaxis was associated with significantly less time spent in hospital and being absent from work or school, and non-significant decreasing trends in the number of patients reporting pain and mobility problems. These findings are comparable to those reported in hemophilic patients without inhibitors during prophylaxis [16]. Although rFVIIa prophylaxis provided a clear reduction in bleeding frequency, the results of this trial do not provide direct evidence of the theoretical benefits of prevention and/or delayed progression of arthropathy. Although routine prophylaxis using factor replacement concentrates is known to prevent bleeding and resultant joint damage over the long term in patients with severe hemophilia without inhibitors [3,6,31–35], there were no changes in orthopedic joint scores during the study period. This is probably related to the severity of pre-existing joint damage and the short prophylaxis period.

The observational data available from published case studies suggest that prophylaxis with rFVIIa or aPCC is of potential use in patients with severe hemophilia A or B with high-responding inhibitors [9,11–14,22,36–41]. In general, there appears to be a clear decrease in bleed frequency, associated with few adverse events and little anamnesis. The dosing and dosing frequency reported in these studies vary widely, but, in general, rFVIIa was dosed more frequently.

Clinical assessment of joint function following prophylaxis with the aPCC FEIBA (Baxter Bioscience, Deerfield, IL, USA) has provided mixed results, and there are no data available for rFVIIa prophylaxis. In a retrospective review of medical records from seven patients with hemophilia A with inhibitors, Hilgartner et al. [39] found that whereas 3 years of prophylaxis with aPCC did not prevent the progression of arthropathy in patients with pre-existing joint disease, there was no development of arthropathy in patients who did not have pre-existing joint disease [39].

Although the present trial demonstrates the efficacy and safety of daily administration of rFVIIa as secondary prophylaxis, it should be noted that the conclusions are limited to patients with hemophilia A with inhibitors who experience frequent bleeding episodes. There is no indication that the effect of rFVIIa prophylaxis differs between patients with hemophilia A and B inhibitors, on the basis of our knowledge of the mode of action of rFVIIa and case reports in the literature [9,14–16,22,36,37,41]. However, because of the intensity of this new treatment modality, in practise prophylaxis will be most relevant for patients suffering from frequent bleeding.

Similarly, it is not possible to draw any conclusions regarding any relationships between age and potential benefits of prophylaxis, because of the small number of children and adolescents enrolled. Moreover, because all patients enrolled in this trial reported one or more target joint and > 15 bleeding episodes during the preprophylaxis period, we can only speculate on the possible outcome of secondary prophylaxis in frequently bleeding patients without target joint bleeds.

Secondary rFVIIa prophylaxis has a promising clinical potential in connection with ITI therapy. Regular dosing with rFVIIa prior to and during ITI therapy may prevent bleeding and joint deterioration prior to establishing tolerance and the resumption of prophylactic therapy with FVIII. This would also avoid exposure to FVIII prior to initiating ITI while the FVIII antibody titer is decreasing. However, further studies are necessary to evaluate this treatment approach.

In conclusion, the results of this trial provide evidence for the concept of secondary rFVIIa prophylaxis in inhibitor patients with frequent bleeds. Clinically relevant reductions in bleeding frequency were observed during prophylaxis as compared to conventional on-demand therapy, without any safety concerns being raised.


The authors gratefully acknowledge the following principal investigators for their excellent conduct of this trial: J. Mahlangu, Hemophilia Clinic Johannesburg Hospital, Johannesburg, South Africa; D. N. Colita, Clinic of Haematology, Bucharest, Romania; C. V. Arion, Institute Fundeni, Clinic of Pediatry, Bucharest, Romania; V. Vdovin, Izmailovskaya Children Municipal Clinical Hospital, Moscow, Russia; K. Kavakli, Ege University Medical Faculty Child Health and Disease Department, Paediatric Haematology Unit, Bornova Turkey; M. Castro Ozelo, Hemodentro de Dampinas, Campinas, Brazil; M. Hermida de Oliveira, Hemoria, Rio de Janeiro, Brazil; M. Ng Chua, National Haemophilia Center, Manila, Philippines; M. J. Matysiak, Klinika Hematologii, Warsaw, Poland; J. Windyga, Instytut Hematologii I Transfuzjologii, Warsaw, Poland; C. Serdano, Hospital Marqués de Valdecilla, Santander, Spain; F. Hernández, Hospital Universitario La Paz, Madrid, Spain; R Perez, Hospital General Edificio Principal, Sevilla, Spain, B. Lewis, Alta Bates Comprehensive Cancer Center, Berkley, USA; and J. Dipaola, University of Iowa Hospitals and Clinics, Iowa City, USA. In addition, the authors would also like to thank the Steering Committee, consisting of H. Roberts, University of North Carolina, School of Medicine, Chapel Hill, NC, USA, K. Hoots, University of Texas Medical School and University of Texas-MD Anderson Cancer Center, Houston, TX, USA, G. Auerswald, Professor-Hess-Children’s Hospital, Bremen, Germany, and R. Ljung, Department of Paediatrics, University of Lund, University Hospital, Lund, Sweden, for their input to the trial design, and interpretation and discussion of the data.

Disclosure of Conflict of Interests

The authors thank U. Friedrich, Novo Nordisk A/S, Bagsvaerd, Denmark for contribution to trial design and performance of logistic management of the trial, U. Hedner, Novo Nordisk A/S, Bagsvaerd, Denmark for input to trial design, M. Suntum, Novo Nordisk A/S, Bagsvaerd, Denmark for assistance in data analysis, and N. M. Albrektsen and M. White, Novo Nordisk A/S, Bagsvaerd, Denmark for their assistance with the preparation of this manuscript.