Rosiglitazone and the need for a new drug safety agency

The recent news about the suspension of rosiglitazone, a blockbuster antidiabetes drug, is not reassuring.1 Once again a widely prescribed drug shows a degree of toxicity that should not have allowed it to remain on the market for so long. This is not an isolated case, because it follows in a short time similar concerns over cerivastatin, rofecoxib, rimonabant, and sibutramine, not forgetting the selective serotonin reuptake inhibitors. In addition, a further case is now on the table, concerning the relation between the use of bisphosphonates and cancer.2

At least in the case of sibutramine and rosiglitazone, awareness of the risk of adverse events had been shown in the European Medicines Agency’s request for large randomised trials to investigate possible cardiotoxicity. A trial of sibutramine was committed to in 2002 and the trial begun in January 2003, completed in November 2009, and published in September 2010,3 well after the drug had been suspended, in January 2010. In the case of rosiglitazone the commitment was made in July 2000, the study published in June 2009,4 and the drug suspended in September 2010.

Safety questions can be answered only by large, long term studies. Manufacturers have every interest in prolonging the time between beginning a trial and the final result, during which they carry on earning income. Certainly in the case of rofecoxib a meta-analysis showed the risk of cardiotoxicity four years before its withdrawal.5 No data are available for sibutramine, but in the case of rosiglitazone the first meta-analysis showing an increase in the risk of heart failure was available in 2007.6 In the meantime many patients developed heart disease while seeing no significant improvement in their diabetes.

What can be done to improve the situation? Surely we must raise the bar, by adopting more stringent requirements before new drugs are allowed onto the market. However, obviously it will be hard for a regulatory agency that has authorised a drug to admit that it hadn’t predicted its potential risk (or had underestimated or overlooked it). The usual regulatory reaction is to insert a warning or a contraindication in the summary of product characteristics or to issue a “Dear Doctor” letter. A regulatory agency seldom withdraws a drug from the market; it is more likely to be the manufacturer that takes the initiative.

Our proposal is to separate drug approval from post-marketing pharmacovigilance. Having both functions under a single body, as happens today, implies a sort of conflict of interest, quite apart from the enormous workload. Furthermore, while establishing the benefits of drugs is relatively straightforward, evaluating their safety is a much more complex and time consuming process that can hardly be expected to have resulted in any firm conclusions at the time of approval.

The new safety evaluation agency could have several tasks, including auditing drug companies’ commitments to and conduct of trials, promoting proactive pharmacovigilance, and making the appropriate regulatory decisions. It would first of all follow up drug companies’ commitments to carry out trials required by the agency that had granted the marketing authorisation. It should verify that whatever is needed to fulfil the requirements is done properly, in the shortest possible time. The US Food and Drug Administration has admitted that only 30% of such requests for trials are fully adhered to by companies. (The EMA does not provide any such figures.)

The second task would be to act as a gathering point for all information on safety collected by national agencies, sustaining a true Europe-wide network such as the current European Centres of Pharmacoepidemiology and Pharmacovigilance promoted by the EMA itself. To promote proactive pharmacovigilance, the safety agency could commission and fund independent studies on critical issues of safety. In addition, it would maintain rigorous scrutiny of the literature, including meta-analyses, systematic reviews, and prospective studies on drug safety. Another useful task would be to promote a flow of information to all interested parties well ahead of publication.

The third task for the agency would be to make decisions on the basis of its assessment of the information collected, with the priority aim of ensuring public health. When the risks of a drug exceed the benefits, the agency must decide on its temporary suspension or definitive withdrawal from the market. The agency should work swiftly and without secrecy and be open to external evaluation and input.

The continual problems of toxicity of drugs that have often been approved on the basis of surrogate end points need to be dealt with urgently so that the public’s confidence can be regained and patients’ safety assured.