Reflections on using non-inferiority randomised placebo controlled trials in assessing cardiovascular safety of new agents for treatment of type 2 diabetes.

The 2008 Food and Drug Administration (FDA) guidance to industry requires experimental evidence that new agents to treat type 2 diabetes do not have an unacceptable increase in cardiovascular risk. They specify this unacceptable increase to be a risk ratio of 1.3 in non-inferiority trials which may use placebo control. Clinically, this means that if a new agent achieves this threshold of not being 30% worse than placebo it is declared 'non-inferior'. This guidance was in response to safety concerns raised about medications approved on their basis of reducing glycated haemoglobin alone. There was concern that this FDA guidance would stifle new drugs coming to market. On the contrary, there have been a number of exciting new classes of agents approved with improved confidence that they reduce glycated haemoglobin, and that they also do not excessively increase cardiovascular risk. Cardiovascular safety trials have been conducted for a number of novel medications using a non-inferiority approach. However, clinicians need to recognise that the results of non-inferiority trials are not as credible as superiority trials. It is important to closely review the trials before accepting claims of 'non-inferiority' or 'cardiac neutrality' especially when these studies are often compared with placebo, and may be accepting estimates of effect which span potentially clinically meaningful harm. There are compelling reasons to further investigate agents showing promise in non-inferiority trials with superiority trials, which include prespecified subgroups, and with sufficient power and duration to provide robust estimates of harms and benefits to inform clinical decision-making.