Future directions in IAT: better studies, better selection, better timing and better techniques

Despite years of clinical experience, we know very little about who we should be treating with intra-arterial stroke therapy (IAT) and how to deliver this resource-intensive treatment modality. This sobering fact was highlighted by the recent randomized controlled trials (RCTs) which failed to demonstrate improved outcomes with IAT over medical management alone.1–3 Although it is very possible that these results would have been different had the current generation stent retrievers been used as revascularization tools, we should acknowledge that our confidence in the efficacy of an unproven therapy has been misplaced before (eg, the SAMMPRIS trial4). A major lesson from these trials is that, at least in the context of current IAT practice, the signal of benefit from this treatment is smaller than assumed. Furthermore, because higher rates of early revascularization observed with IAT have not translated into proportionally better outcomes, the importance of revascularization, without consideration of other contributing factors, may have been overestimated. Among the factors emerging as critically important in determining clinical outcomes are speed of reperfusion and preprocedure infarct size. Although new trials are necessary, they will not remedy the gaps in our knowledge. Phase III RCTs are difficult to perform due to limited funding, lack of treatment equipoise and a small eligible patient population. The time investment is often so large that, due to concomitant technological advances, the results may not be relevant to current practice. As such, high-quality data from single-center studies or multicenter registries that include natural history data are needed to answer the numerous clinical questions that cannot be addressed in an RCT setting, and to provide data to refine trial selection criteria. This commentary highlights areas of critical need in IAT research and process improvement (box 1). Building on this discussion, this special issue of JNIS is devoted to …