One of the problems that has concerned me over the years is the relatively restricted nature of interventions available to Drug Regulatory Authorities in the event that a licensed medicine turns out to have unsuspected toxicity. These are suspension, revocation or modification to the summary of product characteristics. In these situations the perceived need to take action in relation to the risk from the product is often greater than the apparent risk itself warrants. There is an understandable tendency to emphasise risk and forget about benefit. An example of this would be the manner in which the on-going controversy about the risks of the third generation oral contraceptive pills and thromboembolic disease. As well as being of great intrinsic interest, this example emphasises that not all risks relate to new or nearly new drugs. Pharmacovigilance needs to remain alert to the potential problems of drugs at all stages in their development. Another example of a relatively old drug running into problems in the past was nomifensine, an effective antidepressant in which evidence came to light about the risks of acute haemolysis under unusual circumstances of use. Given the relatively long standing nature of the medicine itself, the company involved found it easier to withdraw it from sale than risk litigation by continuing use with adequate warnings. Was this the right decision? What happened to the long-term recipients who were receiving benefits from this drug? Did they transfer to an older antidepressant, or to a newer one (for which we had no comparable information), or discontinue treatment? What were the outcomes in relation to recurrence of depression, suicides and adverse effects to replacement therapies? Unfortunately, we do not know the answers to these questions. Clearly, the company involved in manufacturing nomifensine was not going to fund such a study. The issue seems to me to be a public health one and thus to require public funding. Perhaps with the development of multipurpose databases we will be better at managing such an event in future. I hope so, but the question remains unanswered at present.
I believe that the need for post-withdrawal surveillance studies in large databases is clear and is clearly a duty that devolves on the public purse. I hope the need will occur infrequently!
Was this article helpful?