Antithrombotic therapy after bioprosthetic aortic valve replacement

Aortic valve replacement with a tissue valve alone or in association with coronary artery bypass grafting is being practised more and more often in the elderly, accepting the definition of elderly for people aged 65 years or more. Despite the age, many of these patients do not have associated risk factors for thromboembolism and could benefit from initially avoiding anticoagulation after valve replacement. The need for short-term anticoagulation was based on some papers showing a high incidence of thromboembolism in the first 3 months after surgery [1]. The lack of endothelium in the leaflet tissue and the presence of rough surfaces like Dacron rings and knots justified some antithrombotic treatment for 3 months until complete ‘healing’ occurs. Heras et al. [1] estimated the risk to be about 10% but these data were obtained from retrospective studies on populations operated on more than two decades ago. Present studies show that the risk is much lower, about 2–4%. These are rough percentages, for linearized rates do not apply to short periods of time like 3 months (they exaggerate four times the true incidence). A hazard-function type of analysis is more appropriate providing there is enough number of events [2]. This makes comparison difficult. In the last 5 years, several papers studying the usefulness of antiplatelets after tissue valve replacement were published reflecting the growing interest on this field among the surgical community. Most of them were published in this journal [3–5]. Of all, only three were prospective [2,6,7] and only one was randomized (the TRAC study) [7,8]. One common characteristic to most studies is the small sample size: usually about 100 patients in each arm of the trial. This is the main limitation of studies based on surgical population. As Nowell and Jahangiri [9] point out, given an event rate of 2% and a risk ratio of 2.0, a population of 1500 patients would be required in order to obtain a significant difference between the groups. Therefore, there is a place for a large randomized study that can help provide robust evidence. At the present time, there is relevant evidence showing that antiplatelet treatment is as useful as anticoagulation in preventing thromboembolism. It seems obvious that if both treatments are equally useful in preventing thromboembolism the safety profile of antiplatelets and the avoidance of repeated blood tests and dosage adjustments may result in better quality of life for the patients and better treatment compliance. The American Heart Association’s 2006 guidelines recommend the use of antiplatelets after valve replacement with a tissue valve (class I, level of evidence C). Although the recommendation is made for aspirin, it can be extrapolated to other antiplatelets like ticlopidine [2] or clopidogrel (in case of allergy to aspirin) and also to triflusal [8,10]. Triflusal is structurally related to aspirin and has a similar antithrombotic effect. Although less known and used as the other antiplatelet agents, it offers a more favorable safety profile due to the lesser degree of platelet cyclooxygenase inhibition, resulting in a lower risk of bleeding, which is especially useful in the elderly population [10], in the countries where it is available.

In this issue, Colli et al. [11] present the results of a survey on the use of antithrombotic treatment after aortic valve replacement with a tissue valve among the 49 centers participating in the so-called ACTION registry. The ACTION registry was not formally published as a ‘protocol design’ or ‘work-in-progress’ paper, but was mentioned in a review on this matter published in this journal in March 2007 [4]. As these authors stated in that review, the study was initially designed as a randomized trial comparing aspirin to oral anticoagulation. This would be the ‘ACTION paradox’, that is a drug trial sponsored by a medical device company. Instead, they finally launched an international, however mostly European (plus Canada, Israel and India) registry recruiting patients from 49 centers. The same company has launched a similar study, the ANSWER Registry in the USA recruiting 2000 patients. Since there is no detailed description of the study protocol some questions remain in the air: do both registries complement each other? Is this just a marketing strategy? In the methods section of the manuscript they state that it is an open registry where every investigator prescribes the antithrombotic therapy that is deemed appropriate. The only common condition is the implantation of SJM Epic™ or SJM Epic™ Supra porcine bioprosthesis. We miss a full detailed description of the study protocol: definition of events, recruiting period, primary and secondary end points, data validation, etc. In the present status, the study suffers from serious limitations such as the open character and the lack of randomization. We are afraid that some bias is going to be unavoidable. On the other hand being a registry offers some advantages: it shows what happens in the real life without the rigid limitations imposed by a randomized study. In that sense, this preliminary survey gives us some interesting data. There is a large variability among centers even in the same country and in one center, two participating surgeons had different policies, revealing a possible lack of consensus. Interestingly enough, one third of the patients received a combined therapy of oral anticoagulants and aspirin. Given the proven efficacy of both therapies alone, this group will probably not show a significant difference in the rate of thromboembolism and may contaminate the issue of which therapy, antiplatelet or oral anticoagulation, prevents better from thrombotic events. This group, though, may have an impact on the rate of bleeding affecting safety.

A small group of 4% of patients will not receive any antithrombotic therapy. Brueck et al. [5] could not find any difference in efficacy and safety between patients having aspirin or no treatment. The sample size was small and so was the event rate. The ACTION registry has a large population at study and, therefore, may give us relevant information about whether not giving any antithrombotic therapy is an option.

The authors conclude that results form this registry may be the basis for an international randomized study supported by professional organizations. Taking into account the difficulties, above all logistic and financial, in recruiting a large sample of surgical patients, we consider that if this registry can provide relevant information on the use of antiplatelets after aortic valve replacement with a tissue valve, results may lead to a level of evidence C+, thus avoiding the need of a subsequent randomized trial. Otherwise, this will be another missed opportunity.

© European Association for Cardio-Thoracic Surgery 2008