Direct oral anticoagulants:a new challenge for pharmacovigilance

More than two thirds of people undergoing long-term anticoagulant therapy with the vitamin-K antagonists (VKAs) warfarin and acenocoumarol have non–valvular atrial fibrillation (AF). This cardiac arrhythmia affects as many as 8.8 million people in the European Union, 700.000 in Italy. Because AF prevalence rises from 0.5 to 1% at the age of 60 up to as high as 10% by age 80 years, AF cases are predicted to double to 1.5 millions by the year 2050, due to general population aging in Italy. AF patients have a 5-fold higher risk of embolic stroke, which is often fatal or disabling. Thromboprophylaxis with VKAs reduces this risk by 64% ¹, but this positive effect is partially offset by an annualized rate of major bleeding complications of 1.2 to 1.5% in the frame of clinical trials, but up to 6.5% in real world registries ². Major bleeding is usually defined as fatal bleeding, or reduction in hemoglobin by at least 2 g/dL, or transfusion of at least two units of blood, or symptomatic bleeds in such critical areas and organs as the central nervous system, the gastrointestinal tract and the retroperitoneal space ³. In spite of the fact that the net clinical benefit is definitely in favor of VKAs, the fear of bleeding complications is the main reason for the underuse of these drugs in AF, as confirmed by a recent study carried out in 70 Italian internal medicine wards ⁴. In particular, advanced age is an important hindrance to the prescription of anticoagulants, despite solid evidence of their effectiveness even in the oldest old (80 years or more), who carries the highest risk of thromboembolism ⁵.
This unsatisfactory background pertaining to the safety of VKAs and their related underprescription has prompted the novel clinical development of oral anticoagulants that, at variance with VKAs, inhibit directly such key enzymes in the coagulation cascade as thrombin and activated factor X (FXa). The direct anticoagulants (DOAs) currently licensed in Italy are the thrombin inhibitor dabigatran etexilate (Pradaxa) and the FXa inhibitors rivaroxaban (Xarelto) and apixaban (Eliquis). Another anti-Xa (edoxaban) is in the advanced development phase but unlikely to be licensed before 2015. At variance with VKAs, all these DOAs have rapid onset of action, short half-life in plasma, limited drug - drug interactions and no need of routine coagulation monitoring for dose adjustment, because their anticoagulant effect is sufficiently constant at fixed oral doses. The main disadvantages of DOAs compared with VKAs are the need of twice daily administration (at least for dabigatran and apixaban, but not for rivaroxaban and edoxaban); the much higher cost despite no laboratory monitoring and, most importantly, the fact that all these drugs are mainly cleared through the kidney, an organ that is very frequently impaired in the oldest old. At the moment there are no antidotes to neutralize the DOAs, but the rapid offset of the anticoagulant effect due to their short half-life makes enough to stop these drugs in most instances, except in situations of life-threatening bleeding. Each DOA has been compared with INR -adjusted warfarin in huge randomized clinical trials carried out worldwide in nearly 80.000 patients with chronic AF ⁶. The primary goal of these trials was to evaluate non-inferiority versus warfarin in terms of thromboprophylaxis, so that any finding of superiority must be considered a secondary end-point owing to the study design. Even though there are some differences between the different DOAs pertaining to their clinical efficacy and effectiveness, some common results broadly emerge ⁶. All drugs were non-inferior to VKAs, with a comparable or lower risk of stroke and systemic embolism and with a trend for all drugs of lower all-cause mortality, that for apixaban reached statistical significance ⁶. What about a most important primary outcome such as bleeding? The results of randomized clinical trials indicate that although in general major bleeding was less frequent than or as frequent as for warfarin-treated patients, the hope of no adverse hemorrhagic event did not materialize and abnormal bleeding remained a fact and deed even with the DOAs. However, the good news are that all the DOAs reduced by at least 50% intracerebral bleeds, the most feared complication of VKA therapy ⁶. It is likely that the observed lower all-cause mortality is largely due to less intracerebral bleeds. Does this important advantage of DOAs makes them the treatment of choice in the increasing number of elderly patients with AF? There are several caveats.
Due to the highly selected patients enrolled in the randomized clinical trials, the risk of all-cause bleeding may be underestimated, and can be truly evaluated only in a real world scenario, such as for instance in the oldest patients with multiple diseases taking multiple drugs, frail and at risk of falls, i.e., the most frequent type of patients with AF ⁵. Preliminary data on real world situations are only available for dabigatran, the first DOA licensed in 2011 in Europe and in 2010 the USA (but in Italy this and other DOAs were licensed with a delay of almost two years, in 2013). Soon after licensing there were case-reports of dramatic and ever fatal bleeding, usually occurring in elderly patients treated with dabigatran despite evidence of some degree of renal impairment. Another concern was the observation, stemming from the randomized clinical trials carried out in patients with AF and in those with other indications such as venous thromboembolism, that the use of this DOA was associated with a higher rate of myocardial infarction. Furthermore, an excess of gastrointestinal bleeding over warfarin was observed in the AF RE-LY trial.
These concerns have been partially offset by the Mini-Sentinel analysis conducted in the USA on behalf of the Food and Drug Administration ⁷, as well as by two observational studies carried out in Denmark ^8,9, that took advantage of the comprehensive registries of deaths, hospitalizations and drug prescriptions available in that country. All in all, these post-marketing, real world data obtained in nearly 130.000 patients with AF treated with warfarin or dabigatran showed that there was no excess gastrointestinal bleeding and myocardial infarction with this DOA, while confirming that intracranial bleeding and all-cause mortality were less frequent than for warfarin-treated patients. Rivaroxaban and apixaban were licensed much after dabigatran, so that similar data for these DOAs are not yet available. Which learning points can be made, in Italy and elsewhere, from the available experience with the new DOAs? First, it is hoped that the solid findings of less intracerebral bleeds and of no need for routine out-patient visits for laboratory monitoring may convince both reluctant physicians and their many untreated patients with AF at high risk of stroke to initiate with these drugs a badly needed and efficacious thromboprophylaxis. Second, those AF patients who take VKAs and are very unstable within the recommended INR therapeutic range (2.0 to 3.0) should consider to switch to a DOA. Indeed, a word of caution on switching stems from one of the Danish studies ⁸, because patients switched from VKAs to dabigatran (but not those naïve to the drug) had an excess of bleeding complications, perhaps because switching did occur more frequently in patients at high risk of bleeding and treated outside the current recommendations. Another crucial indication is to avoid dabigatran when glomerular filtration rates are below 50-60 ml/min ⁵. Even though for the other DOAs the degree of renal elimination is smaller (80% for dabigatran contrasts with 60% for rivaroxaban, 50% for edoxaban and 25% for apixaban), much caution is necessary even for these factor Xa inhibitors.
Finally, a strong recommendations to physicians who are adopting DOAs, either in patients naïve or in those experienced with VKA therapy, is to accurately collect all the bleeding episodes and their circumstances and to communicate them to agencies responsible for pharmacovigilance. A healthcare professional, when observing a suspected adverse drug reaction, should notify it via a post-marketing report to the person responsible for hospital pharmacovigilance, if he/she works in a hospital setting, or to the pharmacovigilance head of local health authorities, if he/she operates in the frame of territorial health service. In Italy the reporting form and a list of national pharmacovigilance heads are obtainable from the Italian Medicines Agency’s website (http://www.agenziafarmaco.gov.it/it). It is of utmost importance that communications pertain not only to the DOAs, but also to warfarin. Signaling only the adverse events occurring with the new anticoagulants would lead to the so called Weber effect, i.e., the over representation of adverse effects that typically occurs when a new drug appear in the market ¹⁰. Perhaps the excess of bleeding observed in the early days of dabigatran licensing may be due not only to the inappropriate use of this drug in high risk patients, but also to over reporting associated with the Weber effect.

Conflicts of interest. The REPOSI registry (see ref. 4) in Italy is partially supported by unrestricted grants from Boehringer, Bayer and Pfizer to the Italian Society of Internal Medicine and the Mario Negri Institute. The author has no personal conflict of interest.