Therapeutic approaches to multiple sclerosis: an update on failed, interrupted, or inconclusive trials of immunomodulatory treatment strategies.

Multiple sclerosis (MS) continues to be a therapeutic challenge, and much effort is being made to develop new and more effective immune therapies. Particularly in the past decade, neuroimmunologic research has delivered new and highly effective therapeutic options, as seen in the growing number of immunotherapeutic agents and biologics in development. However, numerous promising clinical trials have failed to show efficacy or have had to be halted prematurely because of unexpected adverse events. Some others have shown results that are of unknown significance with regard to a reliable assessment of true efficacy versus safety. For example, studies of the highly innovative monoclonal antibodies that selectively target immunologic effector molecules have not only revealed the impressive efficacy of such treatments, they have also raised serious concerns about the safety profiles of these antibodies. These results add a new dimension to the estimation of risk-benefit ratios regarding acute or long-term adverse effects. Therapeutic approaches that have previously failed in MS have indicated that there are discrepancies between theoretical expectations and practical outcomes of different compounds. Learning from these defeats helps to optimize future study designs and to reduce the risks to patients. This review summarizes trials on MS treatments since 2001 that failed or were interrupted, attempts to analyze the underlying reasons for failure, and discusses the implications for our current view of MS pathogenesis, clinical practice, and design of future studies. In order to maintain clarity, this review focuses on anti-inflammatory therapies and does not include studies on already approved and effective disease-modifying therapies, albeit used in distinct administration routes or under different paradigms. Neuroprotective and alternative treatment strategies are presented elsewhere.