Immunogenic but effective: the HIT-fondaparinux brain puzzler


Andreas Greinacher, Institut für Immunologie und Transfusionsmedizin, Ernst-Moritz-Arndt-Universität Greifswald, Sauerbruchstraße, D-17489 Greifswald, Germany.
Tel.: +49 3834 865482; fax: +49 3834 865489.


See also Warkentin TE, Pai M, Sheppard JI, Schulman S, Spyropoulos AC, Eikelboom JW. Fondaparinux treatment of acute heparininduced thrombocytopenia confirmed by the serotonin-release assay: a 30-month, 16-patient case series. This issue, pp 2389–96; Goldfarb MJ, Blostein MD. Fondaparinux in acute heparin-induced thrombocytopenia: a case series. This issue, pp 2501–3.

Fondaparinux, a synthetic pentasaccharide that catalyzes the inactivaton of factor (F)Xa by antithrombin, has been shown to induce anti-platelet factor 4 (PF4)/heparin antibodies as frequently as low-molecular-weight heparin [1]; consistent with its immunogenic capacity, presumably antigenic multimolecular complexes of PF4 and fondaparinux have been visualized directly at a molecular level [2]; and, most important, several cases of fondaparinux-associated heparin-induced thrombocytopenia (HIT) have been reported [3].

In spite of these experiences, in this issue of Journal of Thrombosis and Haemostasis, two Canadian groups report on the use of fondaparinux for the anticoagulation of patients with acute HIT. Goldfarb et al. [4] observed eight patients and Warkentin et al. [5] 16; these 24 patients are very well characterized and they probably all had true HIT. This makes these two cohorts different from those previously published [6–9] in which only 1/47 patients also had a documented positive platelet activation test for HIT antibodies. Up until now, it was somewhat unclear how many HIT patients have been treated with fondaparinux.

Combining all 71 patients reported in these cohorts, no new thrombotic events occurred after initiating treatment with fondaparinux [95% confidence interval (CI), 0–5.1%], which looks promising that fondaparinux can provide effective anticoagulation in patients with HIT.

However, beside new thrombosis, the biggest risk for patients treated with alternative anticoagulants is major bleeding. During the introduction of lepirudin into the treatment of HIT, the strong dependence of hirudin pharmacokinetics on renal function and the uncertainties in handling a new drug in a niche indication resulted in major bleeding in more than 10% of patients [10]. Also, with argatroban, a relevant proportion of patients suffered from major bleeding complications [11]. Consequently, for both lepirudin and argatroban, dosing had to be decreased after approval of the drug. As treatment durations varied considerably in the different cohort studies of the various alternative anticoagulants, the bleeding risk per treatment day provides a more meaningful number than the total incidence of major bleedings. Table 1 shows that fondaparinux also seems to compare very favourably with lepirudin, argatroban, and even danaparoid in this regard.

Table 1.   Major bleeding with lepirudin, argatroban, danaparoid and fondaparinux reported in different cohort studies
 Major bleedingTreatment duration daysRisk per treatment day (%)
Lepirudin [11]17.6%151.17
Argatroban [11]5.96%60.99
Danaparoid therapeutic dose [11]5.3%60.89
Fondaparinux therapeutic dose [4,5]1/16

A composite of characteristics makes an anticoagulant effective and reasonably safe in the treatment of HIT. First of all, the drug must be effective, which is the case for all the alternative anticoagulants for HIT. But the drug must also be available: danaparoid has had major supply problems during the past few years, and it has also not been available in the US since 2002. When a drug is needed for an emergency indication such as acute HIT, insecure supply is not acceptable. Dependence of drug pharmacokinetics on renal or hepatic function is another important consideration. The hirudins (lepirudin and desirudin) in particular are excreted > 99% by the kidneys and even small changes in renal function can cause major drug accumulation. The pharmacokinetics of fondaparinux also depends strongly on renal function, and the major bleeding observed with the drug in HIT patients unexpectedly occurred in a patient with renal impairment. This makes fondaparinux a problematic drug in this subset of patients with renal impairment.

Monitoring of the drug levels may help to avoid under- or overdosing. In this regard, monitoring of danaparoid and fondaparinux using anti-FXa assays is much more reliable than monitoring of the direct thrombin inhibitors by the activated partial thromboplastin time (aPTT) or the ecarin clotting time in critically ill patients. The anti-FXa assays do not depend on individual patient factors. In contrast, the aPTT and the ecarin clotting times are dependent on the prothrombin levels of the patient and especially in severely affected HIT patients, prothrombin levels are often low. This results in falsely long aPTTs and, consequently, in inappropriate dose interruptions and under dosing of direct thrombin inhibitors [12]. Although the chromogenic ecarin test overcomes this limitation, this assay is still not automated and also not available in North America. Finally and probably most importantly, HIT is relatively rare and most physicians have rather limited experience handling drugs used exclusively for this niche indication. In contrast, fondaparinux is used on a daily basis in many centers. In this regard, however, one has to consider that the two groups reporting their experience with fondaparinux are located at centers with major experience in the treatment of thrombosis and where availability of anti-FXa testing exists.

The perspective to use fondaparinux for the treatment of HIT and to monitor the drug in these patients prompts several ‘frequently asked questions’:

Why should I monitor fondaparinux at all? There are fixed dose regimens approved for acute thrombosis

Fondaparinux should be monitored in acute HIT until a steady state is reached to avoid under dosage of fondaparinux. HIT is one of the most prothrombotic diseases and prevention of a new thrombosis requires rather aggressive anticoagulation. At the same time, patients in whom the drug accumulates to levels with an increased risk of bleeding must be recognized.

When should I monitor fondaparinux in HIT and which is the target range?

Fondaparinux should be monitored 4 h after the first and second injection with a target peak level of approximately 1.5 aFXa U mL−1; and before the next injection with an aimed trough level of approximately 0.7 aFXa U mL−1. If the peak level is not reached, a second dose of the drug should be given (depending on the aFXa-level 1.5 or 2.5 mg), if the trough level is too high, the dose given at the next day should be reduced. Probably, the dosing scheme used at McMaster with a time difference of 12–14 h instead of 24 h between the first and second injection of fondaparinux also contributed to the effective prevention of new a thrombosis during the most acute phase of HIT.

Is there a difference between fondaparinux and danaparoid, as both are antithrombin-dependent FXa inhibitors?

There is a fundamental difference between the two drugs. The activity of danaparoid is about 30 anti-FXa U mg−1, whereas it is 700 anti-FXa U mg−1 for fondaparinux. Therefore, much more danaparoid must be given on a weight basis than fondaparinux to achieve similar anti-FXa levels. By interacting with PF4, danaparoid detaches the PF4/heparin complexes from the platelet surface and thus (besides its potent anticoagulant properties) additionally inhibits the pathogenesis of HIT antibody induced platelet activation [13]. Moreover, danaparoid has some anti-thrombin activities; in contrast, fondaparinux has no anti-thrombin action. Together with its lesser dependence on renal function, we prefer danaparoid over fondaparinux in patients with acute HIT in our clinical practice.

Fondaparinux has been shown to induce anti-PF4/heparin antibodies. How can I use such a drug in HIT?

Fondaparinux binds to PF4 and forms multimolecular complexes with PF4 [2]. However, it forms relatively few of these complexes. These few complexes are sufficient to induce an immune response but only a small number of these multimolecular complexes is generated, which is not sufficient to activate many platelets thereby promoting thrombin generation. To date, none of the patients reported in the case series have developed apparent worsening of their HIT when treated with fondaparinux.

Fondaparinux has been reported to cause HIT. How does this fit with the above concepts?

The anti-PF4/heparin immune response varies considerably among different patients. With regard to clinical significance and the risk of new a thrombosis they show a certain hierarchy. Some patients form antibodies which react with optimally presented PF4/heparin antibodies in an ELISA but do not cause any clinical problems. Other patients form antibodies which bind to PF4/heparin complexes and activate platelets in the presence of heparin; here, HIT can occur when heparin is being given. A third group of patients forms antibodies that, in addition, can activate platelets in the presence of buffer alone (i.e. in the complete absence of heparin) during the acute phase of HIT [14,15], these patients usually have developed HIT. The last group, representing the most dangerous antibodies, are induced by heparin but become progressively clinically relevant after cessation of heparin, so-called ‘delayed-onset’ HIT antibodies [16]. These antibodies, which bind PF4 and cause strong platelet activation even in the absence of heparin, behave like platelet-activating autoantibodies. It is likely that the few cases of fondaparinux-induced HIT resemble this situation, for example the immune response has been triggered by the few PF4/fondaparinux complexes but then switch to produce strongly reactive autoantibodies independently of fondaparinux. These relatively rare patients require high dose anticoagulation and should be closely monitored to avoid any further thrombin bursts until the platelet counts have recovered considerably, which may take months.

Is fondaparinux dependent HIT really excluded?

No, as fondaparinux binds to PF4 there is a theoretical chance that it is inducing an antigenic conformational change recognized by fondaparinux-dependent antibodies, but based on the approval studies which enrolled several thousand patients, the risk is very low.

In conclusion, fondaparinux emerges as an effective, reasonably safe and cost-effective treatment option for patients with acute HIT.

Disclosure of conflict of interest

The author received honoraria from GSK for lectures.