Is there a need for an alternative in the era of novel anticoagulants?

Abstract

With the development of the new direct oral anticoagulants, many of the unmet needs of the vitamin K antagonists were fulfilled, such as the absence of dietary interactions, few drug interactions, predictable effects and no need for monitoring. However, growing experience with the direct oral anticoagulants indicates there is still room for improvement. Developing antithrombotic agents that optimize their antithrombotic effect while producing little or no risk of bleeding is a long sought after goal. This and other enhanced attributes of candidate drugs are on the horizon.

Keywords:

• anticoagulants
• direct oral anticoagulants
• DOACs
• NOACs
• novel oral anticoagulants

With the advent of the first oral direct thrombin inhibitor, ximelagatran, coming on line in the early 2000s, the medical profession was touting the benefits of a new class of oral anticoagulants to replace the vitamin K antagonists (VKAs) and their many shortcomings. These new anticoagulants would have a number of attributes to overcome the problems associated with the VKAs, including a wide therapeutic window, no need for monitoring and few drug or dietary interactions [1]. Although ximelagatran was promptly withdrawn from the European market because of hepatic toxicity and is no longer in development, there are now four new drugs on the market with similar attributes, one an oral direct thrombin inhibitor and three oral direct factor Xa inhibitors. These drugs posses many of the favorable characteristics noted above. They have proven efficacy and a safety profile better than the VKAs [2]. Clinicians, however, are beginning to experience unanticipated management issues such as managing drug interactions, measuring anticoagulant activity when needed and reversing anticoagulant activity; the question again is asked, ‘do we need new alternative anticoagulants’.

The answer to this question is an unequivocal ‘yes’. We should strive to better our antithrombotic armamentarium until we develop the ideal oral anticoagulant, if that is achievable. What are some of the attributes we would like to see in the quest for better anticoagulant therapy? Most of the characteristics of the currently available new direct oral anticoagulants are already improvements over the VKAs (Table 1). In addition, the following attributes would be highly desirable:

Table 1. Major desirable characteristics of new alternative anticoagulants compared to vitamin K antagonists and new direct oral anticoagulants.

Characteristic	Vitamin K antagonists^†	Direct oral anticoagulants	New alternative anticoagulants
Predictable effect	No	Yes	Yes
Monitoring necessary	Yes	No	No
Drug interactions	Many	Few	Even less
Dietary interactions	Many	None	None
Effects on coagulation easily reversed	No	No	Yes (or no effects)
Assay readily available to measure effect on coagulation	Yes	No	Yes (if present)
Efficacy	Good	Good	Better
Safety (bleeding)	Poor	Better	Best (less or no bleeding)
Efficacy/safety	Moderate	Better	Best
Cost	Moderate (with management costs)	High	Lower

^†Vitamin K antagonists are considered the standard on which other newer agents are compared.

1. A drug that needs no monitoring, but when one needs to know drug concentration or effect (impairment of coagulation), there are widely available, simple assays to inform the clinician. (A practical enhancement)

2. A drug that maximizes antithrombotic efficacy while minimizing bleeding risk, sometimes referred to as the antithrombotic to anticoagulant ratio (i.e., efficacy to safety ratio or one that prevents thrombosis without increasing the risk of bleeding). (A therapeutic enhancement)

3. A drug that provides high value for its price. (A societal enhancement)

Measuring anticoagulant-induced coagulation impairment

Today’s new oral anticoagulants offer great advantages over the VKAs with their predictable effect, the absence of numerous drug or dietary interactions and the limited number of dose modifications, all leading to the ability to treat without the need for monitoring. But there are times when knowledge of impaired coagulation or drug effect is essential, such as in the setting of major bleeding, the need for emergent surgery, trauma or managing drug interactions [3]. For today’s new drugs there are no laboratory assays that meet the needs of simplicity, wide availability, rapid turnaround and linearity. For dabigatran, the traditional thrombin time, because of its high sensitivity, is mainly helpful when normal, indicating little if any drug on board [4]. The dilute thrombin time is a good measure of dabigatran effect, but is not available in the US [5]. The activated partial thromboplastin time has poor linearity and is highly reagent specific and the prothrombin time has very limited sensitivity. The situation is worse for the oral factor Xa inhibitors with the only useful assay being the chromogenic anti-Xa assay [6], but this assay has extremely limited availability and is not practical in urgent settings where it is likely to be needed. If a drug was developed that did not impair coagulation but retained an antithrombotic effect (attribute # 2 above), then such an assay might not be necessary. Until that time, however, it would be advantageous to be able to rapidly measure drug effect in times of need.

Improving the benefit/risk of anticoagulant drugs

There is growing evidence that the antithrombotic potential of a drug is not necessarily linked to its ability to impair coagulation (i.e., anticoagulant effect), thus, improving antithrombotic efficacy while reducing anticoagulant-related bleeding has been a long sought after goal. In some respects, the new oral anticoagulants compared to the VKAs have furthered this prospect with their reduced bleeding profile [2]. Another approach is to have highly effective, readily available and rapidly acting reversal agents. Such reversal agents are on the horizon for the currently available new agents [7]. Other approaches have been to develop anticoagulants with what might be considered built-in reversal strategies. One such creative approach has been pursued by taking advantage of the biotin–avidin binding affinity [8] and attaching biotin to an anticoagulant that can subsequently be removed by administering avidin. Biotin is a member of the vitamin B family and avidin a protein with very low antigenicity derived from the white of hen’s eggs. The viability of this concept was proven with the development of biotinylated idraparinux (idrabiotaparinux) [9,10], idraparinux being a parenteral Xa inhibitor based on the heparin–antithrombin binding moiety that has a very long half-life. This concept has been studied in a clinical trial of patients with pulmonary embolism showing non-inferiority to standard therapy with significantly less bleeding [11].

Another approach has been the development of nucleic acid aptamers [12]. These DNA or RNA single-stranded nucleic acids inhibit, in a high affinity fashion, a selected protein’s function. Although this approach does not necessarily lead to a better risk/benefit ratio, its anticoagulant activity can be rapidly and easily controlled by developing a complementary, anti-sense oligonucleotide to neutralize the aptamer. On a clinical basis, this approach has been most fully studied with the factor IX aptamer known as the REG 1 or 2 Anticoagulation System (REG2; Regado Biosciences, Inc., Basking Ridge, NJ) [13,14]. Studies have shown that factor IXa inhibition provides a clinically useful antithrombotic signal [15]. In the REG system, pegnivacogin, a nuclease-stabilized RNA aptamer, when given subcutaneously, binds to and inhibits factor IXa with high affinity and specificity and is reversed by its complementary oligonucleotide, anivamersen. Initial studies with intravenous pegnivacogin (REG 1) showed its ability to anticoagulate patients and be reversed by anivamersen in a Phase 2 trial in patients with acute coronary syndromes [13]. A recent first-in-human study of REG 2 (subcutaneous pegnivacogin) demonstrated its ability to neutralize thrombin generation in a dose response manner with reversal by anivamersen without evidence of rebound effect [14].

These approaches have employed agents that require parenteral administration. For long-term outpatient anticoagulant therapy, one would ideally need to develop an oral variation of one of these approaches, although long-term, outpatient subcutaneous therapy may be a viable alternative.

Developing an agent that has good antithrombotic efficacy without creating a risk of bleeding is, perhaps, the ultimate goal. As early as the 1950s it was shown that for warfarin, factor II reduction was mostly responsible for its antithrombotic potential which was separate from its ability to impair coagulation [16]. This was subsequently supported by other reports [17]. A number of efforts are underway to identify candidate drugs with this characteristic.

Much attention has been focused on factors in the intrinsic pathway of coagulation above where factor IX exerts its effect [18]. It is well known that factor XI deficiency is not associated with a major bleeding phenotype [19], but clinical evidence also supports the finding that factor XI deficiency is associated with a reduced risk of stroke and deep vein thrombosis [20]. Based on these clinical insights, investigators have studied the effect of inhibiting [21] or lowering factor XI levels [22]. The latter effort is achieved by specific antisense Factor XI oligonucleotides that selectively inhibit factor XI mRNA expression [22,23]. This concept was recently taken to the clinic where a clinical trial of venous thromboembolism prevention in elective knee arthroplasty of two doses of the antisense oligonucleotide showed superiority over standard therapy with enoxaparin and was associated with a >50% reduction in bleeding compared to enoxaparin [24]. To be effective, however, the oligonucleotide needed to be administered at least 1 month before the planned surgery to allow time for reduction in factor XI levels, raising the question of how this approach might be used in the treatment of acute VTE.

Other novel approaches are under investigation as new discoveries in the pathophysiology of thrombosis are made. Protein disulfide isomerase (PDI) is a thiol isomerase that facilitates disulfide bond modification during protein synthesis and folding. It is found in a number of cells in the endoplasmic reticulum, in the dense tubular system of platelets and on cell surfaces [25]. It was shown that PDI plays a role in platelet function, is required for platelet accumulation during thrombus formation and also affects fibrin formation [25,26]. Animal studies have shown that either blockade or absence of PDI can reduce thrombosis in arterial injury model in animals. The quercetin flavonoids have been found to be natural inhibitors of PDI. Synthetic PDI inhibitors have also been developed with an eye towards developing therapeutic antithrombotic agents [26], and trials are ongoing to assess the antithrombotic potential of these inhibitors. If effective as an antithrombotic, it is yet unclear whether they would also impose a risk of bleeding.

Polyphosphate is an anionic, linear polymer of inorganic phosphates that is secreted from activated platelets and acts as a procoagulant stimulus at several points in the coagulation cascade as well as having several other functions in mammalian systems [27]. Polyphosphate has been shown to be a potent activator of the contact pathway, to enhance the rate of factor V activation by factor Xa, thrombin and factor XIa, to down modulate the inhibitory function of tissue factor pathway inhibitor and to enhance fibrin clot structure by means of incorporating itself into fibrin clots and enhancing the activation of thrombin-activatable fibrinolysis inhibitor. If polyphosphate modulates and/or potentiates thrombosis, investigators have considered whether inhibition of polyphosphate might be a useful therapeutic approach. Initial studies looking at non-toxic polyphosphate inhibitors in animal models of thrombosis have shown that several compounds are effective in reducing thrombosis without increasing the risk of bleeding [28].

Of interest is the recent chance discovery of an IgA paraprotein from a patient with markedly impaired ex vivo clot formation, but without evidence of a bleeding disorder [29]. The paraprotein interacted with exosite 1, the site on thrombin that cleaves fibrinogen into fibrin. Investigators theorized, ‘…exosite 1 is protected from the antibody (when a thrombin molecule sits) on a cell or clot surface, so hemostasis is unaffected, but [it blocks] thrombosis [that] occurs in the luminal space, where exosite 1 is exposed and available to the antibody’ [29]. This might allow for the antibody to inhibit thrombosis where thrombosis is not wanted but allow thrombosis to occur when and where it is needed. Investigators have synthesized an IgG4 antibody that mimics the binding site on thrombin of the patients’ paraprotein and plans are underway for clinical trials.

Optimizing anticoagulant value

Finally, drug acquisition cost for the patient is not a minor issue as we have seen with the new anticoagulants and especially for the elderly who make up much of the population requiring anticoagulation. Therapeutic value in healthcare is usually defined as outcomes/price [30]. The goal is to maximize therapeutic outcomes at the lowest reasonable price. Outcomes for anticoagulants must take into account both their efficacy and safety leading to net clinical benefit. It is ironic that this goal, the one most under human control, may be the least likely to be achieved.

Financial & competing interests disclosure

J Ansell has acted as a consultant for and received honoraria from Bristol Myers Squibb, Pfizer, Daiichi Sankyo, Boehringer Ingelheim, Roche Diagnostics, Alere, Perosphere and Instrumentation Laboratories. J Ansell also has equity in Perosphere. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.