December 24th, 2014
An Examination of MR CLEAN
CardioExchange welcomes this guest post from Dr. Rory Spiegel, who is a third-year and chief Emergency Medicine resident at Newark Beth Israel Medical Center. This piece originally appeared on his blog, EM Nerd.
Published in the NEJM on December 17th 2014, ushered in with the inflated fanfare only the medical industry is capable of, MR CLEAN marks the first successful trial of interventional therapy for acute ischemic stroke. In direct contrast to IMS-3, SYNTHESIS, and MR RESCUE, MR CLEAN is a significantly positive trial. The authors demonstrated success in their primary outcome — “improved neurological outcomes at 90 days” — with an adjusted odds ratio of 1.67 (95% confidence interval [CI], 1.21 – 2.30). Why MR CLEAN was positive when the three trials that came before were negative is still unclear. As discussed in my previous post (as well as far more elegant posts on emlitofnote.com and stemlynsblog.org) it may be due to better equipment, faster symptom onset to recanulization times, and the incorporation of CT angiography to identify a cohort of patients who would truly benefit from these invasive interventional strategies. Conversely it may simply be due to the traditional therapy group performing so poorly.
A closer look at the results from MR CLEAN reveal that though the interventional group outperformed the placebo group by a significant amount (an absolute increase of the number of patients alive and independent by 13.5%) compared to its peers, its performance was far from exceptional. In the IMS-3 trial, 40.2% patients in the control arm (tPA alone) were alive and independent at 90 days compared to only 32.6% of the patients in the intervention arm of the MR CLEAN trial. Even the placebo groups in NINDS and ECASS-3 who received no reperfusion therapies had better outcomes than the patients receiving interventional therapies in the MR CLEAN trial. Twenty-six percent and 45.5% of the control patients in the NINDS and ECASS-3 trials, respectively had a mRS of 0 or 1 at 90 days. Compared to only 11.6% of the patients in the interventional arm of MR CLEAN.
Though it is difficult and not completely appropriate to compare groups from different trials, it does call into question the reasons for MR CLEAN’s staggering success. It may very well have been that the patients in the MR CLEAN cohort were far sicker than the earlier stroke trials, (though both their presenting NIHSS and 90-day mortality rates seem quite similar). It may be that the utilization of CT angiography to select patients for recruitment excluded the majority of the stroke mimics who were included in these earlier trials, and will universally have good outcomes (this seems to be the answer given by the authors when queried by Dr. Ryan Radecki). The subgroup analysis, which indicated only the patients with a NIHSS of greater than 20 demonstrated a statistically significant benefit from endovascular therapy, seems to support this supposition. The authors point to a meta-analysis of the six trials examining endovascular therapy for acute ischemic stroke as additional proof. In this analysis by Fargen et al, published in the J NeuroIntervent Surg, the authors examine the subgroup of patients with radiographically confirmed large vessel occlusion (LVO). Similar to MR CLEAN, patients receiving endovascular therapy had better outcomes at 90 days (a mRS of 0-2 38.3% vs 25.8% respectively ). Even in this combined cohort with radiographically confirmed LVO, outcomes were not as dire as those observed in the MR CLEAN trial.
MR CLEAN’s success may be attributed to the advancements in both procedural proficiency and technological prowess. But it is equally likely that the whimsy of random chance was responsible for these impressive results. On a final note, it is important to remember that all of the trials examining endovascular therapy in acute ischemic stroke were compared to a control group that included the administration of IV tPA, an intervention whose own efficacy is very much in doubt. Although the rate of adverse events in the intervention arm of MR CLEAN did not differ significantly from those given IV tPA, this is only because alteplase comes with its own terrifying set of unpleasantries. When compared to a true placebo group, I am certain the rate of symptomatic intracranial hemorrhage and new ischemic stroke (7.7% and 5.6%, respectively) would appear far more concerning.
Given the universal failure of the previous three trials examining the very same question, surely more confirmatory evidence is required before investing the unimaginable resources required to support the vast infrastructure needed to make interventional therapy a reality. Since MR CLEAN’s success was announced at the 9th annual World Stroke Conference in Istanbul held in October 2014, two trials examining endovascular therapy in acute ischemic stroke, ESCAPE and EXTEND IA, have halted enrollment early for benefit. It will be interesting to see if these premature stoppages were because of preplanned interim analyses or if MR CLEAN’s success influenced their early termination. I hope we invest the time and resources required to answer these questions with the methodological rigor they deserve. It would be frustratingly tragic to once again be forced to practice with continual doubt because we halted all further investigations out of the fear of discovering that reality is not as promising as the false-truth gained from interpreting only the data that pleases us.