Comparative thrombotic event incidence after infusion of recombinant factor VIIa vs. factor VIII inhibitor bypass activity—a rebuttal


Sabah Sallah MD, International Medical Affairs—Biopharm, Novo Nordisk A/S, Krogshøjvej 53–55, 9C1.15, DK-2880 Bagsværd, Denmark.
Tel.: +45 44428973; fax: +45 44427114; e-mail:

Among the major complications of treatment with factor concentrates in haemophilia patients is the development of inhibitory antibodies against factor VIII (FVIII) or factor IX (FIX). The formation of such inhibitors further complicates the care of these patients. The introduction of products such as FVIII inhibitor bypassing activity (FEIBA) and recombinant activated FVII (rFVIIa) has provided a significant advancement in the treatment of patients with inhibitors [1,2]. By definition, FEIBA and rFVIIa achieve haemostasis through by-passing the need for FVIII or FIX. The prothrombinase complex appears to be the major target for FEIBA, while the haemostatic effect of rFVIIa is probably related to the binding to the surface of activated platelets and subsequent activation of FX and generation of thrombin [3–5]. Experimental evidence suggests that the haemostatic effect of rFVIIa is localized to the site of injury where activated platelets and tissue factor are available [6]. Not only may the mode of action of rFVIIa and FEIBA be different, but it is also important to distinguish between the clinical applications of these two agents. rFVIIa is indicated for patients with inhibitors against FVIII or IX in the United States and the European Union (EU), and acquired haemophilia, Glanzmann's thrombo-asthenia with platelet refractoriness and congenital FVII deficiency in the EU. It has also been used in a wide spectrum of medical and surgical disorders, including intracerebral haemorrhage, anticoagulation reversal, disseminated intravascular coagulation, trauma and liver disease [7–10]. In contrast, FEIBA is used almost exclusively in congenital and acquired haemophilia [1,2,11,12].

In a recent published report in the journal by Aledort [13], it was stated that the thrombotic adverse events (AEs) of rFVIIa were significantly higher than those associated with the administration of FEIBA. In this report, a total of 67 thrombotic AEs corresponding to an incidence rate of 24.6 per 105 rFVIIa infusions and 16 thrombotic AEs corresponding to an incidence rate of 8.24 per 105 FEIBA infusions were identified. These data were extracted from the MedWatch pharmacovigilance program of the United States Food and Drug Administration and supplemented by published case reports. Based on these data, the author concluded that the risk for thrombosis may be higher following the administration of rFVIIa compared to FEIBA.

There are several major concerns with the above conclusion. First, there are well-recognized limitations to using the MedWatch data for evaluating safety, which include the reliance on voluntary reporting of adverse events, and uncertainty in determining causality of the adverse event (drug vs. disease) because of inadequate detail and poor quality of the submitted reports (reviewed in [14]). It is well established that the reporting rate from the marketed use of a product is dependent on many factors, including the time since launch and the indication for the use of the drug. There is a well-known tendency to report on AEs of agents that are novel or relatively newly introduced to the market. This is referred to as the ‘Weber effect’, and in this particular case the Weber effect will be in favour of FEIBA as it has been in use for at least 30 years [15]. Although the paper acknowledges the role of the Weber effect in reporting on AEs, surprisingly, it concludes that such effect may have in fact contributed to a lower number of AEs reported with rFVIIa. Moreover, 24 of the rFVIIa case reports actually originate from clinical trials. Safety in clinical trials is monitored prospectively and intended to capture all adverse events. During the period covered by the Aledort report, rFVIIa was the subject of a large number of clinical trials. The use of spontaneously reported AEs in combination with AEs reported from clinical trials in very different patient populations is not scientifically meaningful.

Secondly, in order to state with confidence whether rFVIIa is more thrombogenic than FEIBA, the population of patients in question should be at least comparable in terms of the underlying conditions and risk factors. Unfortunately, because of the method by which data were derived, the majority of conditions that prompted the use of rFVIIa or FEIBA were not specified in the paper [13]. It is well known that FEIBA is rarely used to control bleeding episodes other than in patients with inhibitors to FVIII or FIX, while the investigational use of rFVIIa is common [9,11,12]. Among the 67 AEs reported in the paper, only 20 AEs occurred in inhibitor patients and 40 AEs involved investigational use of rFVIIa. The conditions that prompted the administration of rFVIIa on an investigational basis included: liver transplantation (11 patients), gastrointestinal haemorrhage (seven patients), liver resection (five patients), stem cell transplantation (three patients), FXI deficiency (three patients), splenectomy (one patient), blunt trauma (one patient), Bernard Soulier's syndrome (one patient), subarachnoid haemorrhage (one patient), lung transplantation (one patient), cardiac bypass (one patient), Marfan's syndrome (one patient), von Willebrand disease (one patient) and uraemic platelet dysfunction (one patient), and was unknown in two other patients. It is plausible that many of the patients receiving rFVIIa for trauma, liver disease or for stem cell transplantation are at inherent risk for clotting. In addition, it is very difficult to state whether a thrombotic event was present in a particular patient prior to receiving rFVIIa. Along the same line of discussion, it would be difficult to conclude whether activation of coagulation, for example, secondary to trauma or high dose chemotherapy, had occurred before the administration of rFVIIa.

Thirdly, data from the MedWatch pharmacovigilance program is usually used to either identify unusual or serious AEs or to compare the safety profile of agents of the same class in terms of indications, applications and mechanism of action [16,17]. This will ensure balance in reporting under similar test conditions. While both FEIBA and rFVIIa are considered bypassing agents, the proposed mechanism of action and, most importantly, the types of patients receiving rFVIIa differs greatly from FEIBA.

Fourthly, it is stated in the paper that the number of infusions for FEIBA was calculated based on a standard dose of 75 IU kg−1 for a body weight of 40 kg, while the number of rFVIIa infusions is given for 2001, and ‘adjusted’ for the time period. What is not stated is that the standard dose of rFVIIa is 90 µg kg−1 for a body weight of 70 kg. By this simple observation, one can appreciate that if the body weights used for the standard dose calculations for rFVIIa and FEIBA were the same there would be an almost 2-fold difference in the number of rFVIIa infusions, which would by itself significantly decrease the calculated incidence rate of associated thrombotic AEs.

Unfortunately, the editorial associated with the paper stated that the thrombotic AEs analysed occurred in inhibitor patients only [18]. Clearly, and based on the above discussion, this was not the case. The editorial has also concluded that the current dosing recommendations of rFVIIa in patients with factor VII deficiency, which is much less than the standard dose used in patients with inhibitors (15–30 µg kg−1 every 4–6 h vs. 90 µg kg−1 at 2–3-h intervals), may be associated with an unnecessary risk of clotting. Again, there is no clear scientific or medical reason that warrants this statement. To ensure comprehensive reporting of the adverse events of rFVIIa, registries for the use of rFVIIa in haemophilia, Glanzmann's Thrombasthenia and FVII deficiency have been established in collaboration with the relevant health authorities.

In brief, while collection of data from the MedWatch system may provide valuable information about the safety of several agents, such methodology does not ensure comprehensive or comparative reporting. The superiority of one agent over another in the same class can be demonstrated only in a randomized clinical trial setting where the conditions of testing are well balanced.