Main

In the fast-evolving world of drug discovery, there is often an implicit assumption that newer is better. The launch of innovative medicines is always a reason for celebration, and 2001 had its share of triumphs, with the launch of Novartis' Gleevec being a key highlight. But amid all the excitement over the latest breakthroughs and technologies, perhaps too little time is spent looking back and comparing the present state of events with past situations. How sure are we that all of this 'progress' is actually delivering the goods? Are we really making better medicines?

The structure of the regulatory framework in which new drugs are tested and licensed is not designed for looking back. The regulatory procedures that are used throughout most of the world are geared towards delivering evidence that a new drug is efficacious, not to telling us how well these drugs perform relative to those that are already in existence. For example, although the US Food and Drug Administration (FDA) generally requires data to be submitted from two large-scale placebo-controlled trials before a new drug is approved, the regulations only require applicants to show that the drug works, not that it works any better than existing therapies.

Of particular interest, therefore, was a comparison of results from a number of antipsychotic clinical trials that was recently published in the American Journal of Psychiatry1. The authors made use of the US Freedom of Information Act to access clinical trial data from the FDA database for trials conducted over a six-year period on new and established antipsychotics. The four new drugs tested, which were all approved between 1987 and 1997, were clozapine, risperidone, olanzapine and quetiapine fumarate. Comparing the level of symptom reduction, measured on a standard Brief Psychiatric Rating Scale (BPRS), in patients that were using these new drugs with that for patients being treated with haloperidol, a drug that was developed by Janssen Pharmaceutica in the 1960s, gave rise to a remarkable result. The average symptom reduction among those schizophrenics treated with the new drugs, at 16.6%, was no better than it was for those treated with haloperidol, at 17.3%.

BPRS scores do not describe the whole picture, and clozapine, in particular, is much heralded as giving rise to greater patient compliance than found with established antipsychotics. But the point is nevertheless a valid one: how many new drugs are improvements on the old? That companies are not falling over themselves to answer this question is hardly surprising, as nobody wants to discover that their new product is actually no better than the existing market leaders. But the amount of information that can be gained by performing meta-analyses on past regulatory data is limited, and what is needed are fully randomized assessments that honestly compare available treatments. In order to detect small differences, such trials need to be large and, it follows, expensive.

It may, however, be a drive to save money that gives the impetus to conduct such trials. The increasing trend, among Health Maintenance Organizations (HMOs) and governments that are subsidizing treatments, to analyse all available therapeutic options is leading them to consider sponsoring competitive analyses of those drugs already on the market. One example of this new breed of trial is the PD MED study, which is being conducted by the University of Birmingham Clinical Trials Unit to determine the cost- effectiveness of four classes of Parkinson's disease drugs in patients with early and late-stage Parkinson's disease. Trials such as this might not only allow resources to be directed more efficiently in future, but might also highlight cases in which the development of new medicines has done comparatively little to change patient outcomes. Which might be a welcome wake-up call for us all.