The dabigatran case: the importance of transparency

"Can we rely on RE-LY?" was the title of an editorial in the New England Journal of Medicine¹ in 2009, commenting on the RE-LY study² which compared dabigatran (110 and 150 mg), the first member of the new class of oral anticoagulants, and warfarin in patients with non-valvular atrial fibrillation at risk of stroke. This study showed that both doses of dabigatran were not inferior to warfarin in terms of efficacy (stroke and systemic embolism); and furthermore, at a dose of 150 mg, that dabigatran was more effective than the comparison group, whilst at a dose of 110 mg, it was superior in terms of safety (serious bleeding).

A reliable study?

Despite some reservations, according to the editors, the study can be considered reliable. However, at the time of the request authorization of the drug, the FDA raised doubts about the design and quality of the data; the approval of the drug, which took place in 2010 in the US and in 2011 in Europe, was then delayed due to a request to the manufacturing company Boehringer for additional data.
Five years after the New England Journal of Medicine editorial, the British Medical Journal published an investigation into the lack of transparency of the data provided to regulatory agencies by Boehringer.³ Use of the new fixed-dose anticoagulant, without the need for monitoring blood, was the main message with which the company launched dabigatran on the market and was therefore able to benefit from the advantages granted by regulatory agencies to promote development of innovative medicines. This allowed them to register the medication, presenting the results of a single study, the RE-LY precisely, instead of two studies, as required by US law for drugs without the innovative requirement. The data presented in this single study, and the evidence on which they are based form the main guidelines in the recommendation for using dabigatran, and now seem incomplete. Family members of some of the patients who were enrolled in the registered trial and who died of a haemorrhage during treatment have filed a lawsuit against Boehringer, who recorded these cases as "cardiovascular events", without specifying the presence of haemorrhagic event. The company also omitted important information about monitoring blood concentrations of dabigatran and of adjusting the dose of medication with reductions, compared to warfarin, of the risk of bleeding estimated at around 30-40%, with no effect on the risk of stroke.

No need to monitor plasma concentrations?

The question of the necessity of monitoring plasma concentrations of dabigatran had already been raised by the FDA during the registration and approval process, as well as by the EMA. EMA documents from 2010 showed that during the registration phase the manufacturing company had communicated to the Agency the maximum plasma concentration of dabigatran (200 ng/ml) above which an increased risk of bleeding could be assumed.
The question that is also posed in the British Medical Journal is: "Had Boehringer provided this information to the FDA as well? And if the EMA was in possession of these data, why hadn't they shared it with prescribing doctors?". Boehringer has since declined to say that it never indicated a threshold plasma concentration. In addition, the FDA had faced the problem of considerable variability linked to the pharmacokinetic profile of dabigatran, so at the same dose, the plasma concentration of the drug could vary by up to 5 times. Such concerns had not affected the conditional approval of the drug. Confidential internal company documents made public in the discussion of the case, have brought to light the draft of an article, written in 2011 but only published by Reilly⁴ in 2014, in which he spoke clearly about monitoring plasma concentrations of dabigatran, to improve the risk-benefit ratio of the drug. These results were omitted in the published version of the article, which simply suggested monitoring patient's renal function before and during treatment with dabigatran, in order to reduce the dosage in case of impaired renal function.
In 2012, dabigatran attained blockbuster status, with sales worth $ 1 billion dollars. This result, however, went hand in hand with the high number of spontaneous reports collected by the FDA for dabigatran in 2011: 542 cases of death and 2,367 of haemorrhages. During the same period, 72 deaths were reported for warfarin.³ These data should be reflected on. It is clear there is a need for systematic and independent studies on the correct use of dabigatran (and comparison with the two other drugs in this new category: apixaban and rivaroxaban), but the importance of transparency of data obtained from clinical studies, which until January this year the manufacturers had no legal obligation to share with regulatory agencies, should be equally evident.

The new EMA transparency procedures

After almost two years of discussions and comparisons on the age-old debate concerning the concept of transparency when conducting clinical trials, new European procedures were finally made available from 1 January 2015⁵ which will make possible the publication of and access to data that is too frequently not disclosed by pharmaceutical companies. This information will only be available for medicinal products authorized via the centralized procedure. Before this new policy, the EMA permitted the release of reports about clinical trials on request, according to a policy of access to regulatory documents that had been in place since late 2010.⁶ The stated intent of the EMA is on one hand to increase public trust and understanding of how regulatory decisions are made, and on the other to give university or independent researchers the opportunity to re-evaluate data in Clinical Study Reports (CSR).
The long road that led to the drafting of this new policy started in November 2012; the draft was then made available from June of the following year for the launch of the first public consultation, with the aim of satisfying the needs of the various stakeholders. The EMA received over a thousand comments from health professionals, industries, universities, regulatory agencies and patient organizations.
However, in May 2014, the EMA presented a review of the draft, which showed several points of dispute, including the ability to merely view the documents, without the ability to save, download or simply print them. In addition, restrictions were placed on the terms of use of the data, which required the individual user to reveal themselves personally, in case legal actions were brought by the pharmaceutical company. Another somewhat questionable detail concerned the possibility of redacting data that was considered to be confidential from a commercial point of view by the pharmaceutical company, conveniently foreshadowing the risk of complaints.
Fortunately, and thanks to the hype generated by the clamour against these unexpected changes,⁷ in June 2014 the EMA took a step back, and on 2 October adopted the final version of the procedure: the access to the CSR is now permitted through two types of profile: one that only provides a display of the data to any user, and the second, linked to a more restrictive registration process, which is designed for academic staff, and allows you to download, save and print the required information. The data cannot be used for commercial purposes or for research supporting a marketing authorization application for a medicinal product, or for extensions or changes to an authorization that has already been granted.

Only a small step forward

This still represents only a small step towards completing the first phase of the laborious and slow process to make clinical data transparent, and does not provide for a second, whose aim should be to make data relating to individual patients involved in the trials public.
Several concerns remain, most notably renunciation of the EMA request for companies to send them all original data from clinical studies in a format that allows a reanalysis by the same agency, as they announced in 2012. In addition, the new procedure does not apply to requests for authorization submitted to the EMA before 1 January 2015, or to data related to requests for a line extension, or extensions of indications for drugs already approved before 1 July 2015.
Of great significance is the ability of a pharmaceutical company to obscure information that could undermine their company's economic or competitive position, despite the fact data from clinical trials are not considered to be confidential business information. The EMA has nevertheless recognized that, in limited circumstances,⁵ the redaction of some data before publication can be conceded. A company may also request the redaction of other sections, in addition to those provided, but must be able to provide an adequate justification. This leaves a rather large grey area, in which companies could move around and carry out real censorship, behind which they could conceal information that might affect assessment of the risk-benefit ratio of a drug. In this way, the excellent motives which induced the EMA, and others, to finally make a step towards transparency would be affected.
The story of dabigatran, as well as others that have arisen previously (remember the case of oseltamivir⁸) reminds us that the mere moral obligation of the pharmaceutical industry to make data relevant to the entire scientific community public is not sufficient to ensure knowledge is shared, and to protect the health of patients.