How exactly do trialists proceed when deciding upon the appropriate acronyms for their soon-to-be blockbuster trial? Is the proper etiquette to follow a traditional prospective process, utilizing the first letter of each word in a trial’s longer title? Or is the selection of an acronym based on its ability to inspire and only then, retrospectively conceive of a full title? In the case of the DWI or CTP Assessment with Clinical Mismatch in the Triage of Wake-Up and Late Presenting Strokes Undergoing Neurointervention with Trevo, or DAWN Trial, the answer I think is clear.
Published in the NEJM, Nogueira et al enrolled patients presenting to the Emergency Department with an acute ischemic stroke between 6-24 hours after symptom onset, evidence of occlusion of the intracranial internal carotid artery, the first segment of the middle cerebral artery, or both on computed tomographic (CT) angiography or magnetic resonance angiography, and a mismatch between the severity of the clinical deficit and the infarct volume on neuroimaging. Patients were randomized to either standard medical management alone or the addition of endovascular thrombectomy (1).
From September 2014 through February 2017, the authors randomized 206 patients presenting to 26 centers in the United States, Canada, Europe, and Australia. The authors reported extraordinary success. 49% of patients randomized to the thrombectomy arm achieved functional independence (mRS score of 0,1 or 2) at 90 days compared to only 14% of the patients in the control arm. 48% of the patients in the thrombectomy arm experienced a significant early improvement in symptoms (defined as a decrease of NIHSS of 10 points or greater or an absolute score of 0 or 1 at days 5,6 or 7) compared to only 19% of the patients in the control group. At 24 hours, reperfusion was achieved in 77% of the patients in the thrombectomy group with only 36% of the patients in the control group. Neurologic deterioration occurred in 14% of the patients in the thrombectomy group compared to 26% of the controls.
And while these results are fairly robust and likely represent a true utility of endovascular therapy in this subset of stroke patients, this trial contains a number of eccentricities that may have influenced their findings. First, the authors primary endpoint was the mean score for disability on the utility-weighted modified Rankin scale at 90 days. The authors attempted to attribute a weighted value to each mRS score. Essentially, the numerical mRS score is intended to represent the patient’s underlying functional status. And while the mRS functions as an ordinal scale, the integral distances between values are not equal. Shifting from a score of 2 to a score of 0 is not equivalent to shifting from a score of 6 to a score of 4. The authors argue that because of this, a simple ordinal analysis that treats all shifts equally is not appropriate. Rather creating a weighted system, based on patient and clinician centered judgments, which assigns a score to each mRS value better represents the potential benefits of any therapeutic intervention. In this case mRS scores of 0,1,2,3,4,5,6 were assigned values of 10, 9.1, 7.6, 6.5, 3.3, 0, and 0. When taken on face value, this seems like a reasonable outcome measure. In reality, its implementation may be somewhat problematic. We have discussed in the past, the unreliability of the mRS score, and how two clinicians when assessing the same patient will often disagree by as much as 2 points on the mRS. These discrepancies are more prevalent in the intermediate scores (mRS values of 2,3,4) and patients with values on either end of the scale are evaluated more reliably (2,3,4). Even in a simple ordinal analysis these inconsistencies can be problematic, but in a weighted analysis, these inconsistencies will be potentiated as each incremental error causes a shift in the weighted scale rather than just the simple mRS. So to imagine that an unblinded cohort, in whom approximately 25% were evaluated via a phone interview, possessed the granularity to perform such an analysis is hard to believe. The most inviting motivation for such an analysis is the same reason trialists abandoned the traditional dichotomous outcome measure in favor of the standard ordinal analysis, increased statistical power.
The second issue, like the majority of trials examining the utility of endovascular therapy in CVA, the DAWN trial was stopped early for benefit. Using a Bayesian trial design, the authors go to impressive measures in justifying their early stoppage, and given their results, it was likely appropriate. But despite giving themselves apriori statistical permission for such a termination, as with any early stoppage, the potential for falsely optimistic results is high.
Finally, the authors failed to inform us how many patients were screened in the process of recruiting this cohort of 207 patients, and so it is unclear how to place this trial in its appropriate context. I suspect like the remainder of the endovascular trials, this was a highly selected cohort plucked from a far greater population of stroke patients.
Despite these issues, the results of the results of the DAWN trial are far from novel findings. In fact, the trial is fairly consistent with the remainder of the literature examining endovascular therapy for the treatment of acute ischemic stroke. When these therapies are deployed in a population of stroke patients with accessible lesions and salvageable tissue, patients fare notably better than controls. The results of the DAWN cohort highlight the flaws in a selection process that utilizes an empiric time threshold to identify a cohort who will benefit from acute reperfusion therapy. The utilization of empiric time cutoffs is a remnant of an older age, before we had wide access to perfusion imaging. The early studies examining the utility of perfusion imaging were consistently negative. But it is important to understand the context in which they failed to demonstrate a benefit.
In 1995, the NINDS trial group published their findings describing the benefits of the use of tPA in patients presenting in under 3-hours. Shortly after the FDA approved the use of tPA in this subset of stroke patients, further attempts to validate its use in this subgroup of patients ceased. The early studies examining the use of perfusion imaging were not intended to further refine which patients would truly benefit from effective reperfusion therapy, but rather were utilized by industry as a ploy to potential expand the window for the administration of thrombolytics, an agent of questionable therapeutic benefit. If your reperfusion therapy lacks the ability to reperfuse, does it matter if you identify patients with salvageable brain tissue? What the endovascular trials have consistently demonstrated, is that when perfusion imaging is paired with an effective therapeutic modality it is possible to identify a cohort of patients who will benefit from revascularization independent of time from symptom onset.
Finally, given the success of the DAWN trial, it is time to re-examine the utility of thrombolytics when given in concert with endovascular therapies. What added benefit do they provide if any? Given the current industry-driven trial climate and the fact that such a scientific inquiry can only harm Big Pharma’s bottom line, it is doubtful such an endeavor will ever be carried out.
- Nogueira RG, Jadhav AP, Haussen DC, et al. Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct. N Engl J Med. DOI: 10.1056/NEJMoa1706442
- van Swieten JC, Koudstaal PJ, Visser MC, Schouten HJ, van Gijn J. Inter-observer agreement for the assessment of handicap in stroke patients. Stroke. 1988;19(5):604–7.
- Wilson JT, Hareendran A, Grant M, Baird T, Schulz UG, Muir KW, et al. Improving the assessment of outcomes in stroke: use of a structured interview to assign grades on the modified Rankin scale. Stroke. 2002;33(9):2243–6.
- Wilson JT, Hareendran A, Hendry A, Potter J, Bone I, Muir KW. Reliability of the modified Rankin scale across multiple raters: benefits of a structured interview. Stroke. 2005;36(4):777–81.