What Is a Meaningful Difference When Using Infarct Volume as the Primary Outcome?: Results From the HERMES Database
Leon A. Rinkel, Johanna M. Ospel, Scott B. Brown, Bruce C.V. Campbell, Diederik W.J. Dippel, Andrew M. Demchuk, Charles B.L.M. Majoie, Peter J. Mitchell, Serge Bracard, Francis Guillemin, Tudor G. Jovin, Keith W. Muir, Philip White, Jeffrey L. Saver, Michael D. Hill, Mayank Goyal,- Advanced and Specialized Nursing
- Cardiology and Cardiovascular Medicine
- Neurology (clinical)
BACKGROUND:
Ischemic stroke lesion volume at follow-up is an important surrogate outcome for acute stroke trials. We aimed to assess which differences in 48-hour lesion volume translate into meaningful clinical differences.
METHODS:
We used pooled data from 7 trials investigating the efficacy of endovascular treatment for anterior circulation large vessel occlusion in acute ischemic stroke. We assessed 48-hour lesion volume follow-up computed tomography or magnetic resonance imaging. The primary outcome was a good functional outcome, defined as modified Rankin Scale (mRS) scores of 0 to 2. We performed multivariable logistic regression to predict the probability of achieving mRS scores of 0 to 2 and determined the differences in 48-hour lesion volume that correspond to a change of 1%, 5%, and 10% in the adjusted probability of achieving mRS scores of 0 to 2.
RESULTS:
In total, 1665/1766 (94.2%) patients (median age, 68 [interquartile range, 57–76] years, 781 [46.9%] female) had information on follow-up ischemic lesion volume. Computed tomography was used for follow-up imaging in 83% of patients. The median 48-hour lesion volume was 41 (interquartile range, 14–120) mL. We observed a linear relationship between 48-hour lesion volume and mRS scores of 0 to 2 for adjusted probabilities between 65% and 20%/volumes <80 mL, although the curve sloped off for lower mRS scores of 0–2 probabilities/higher volumes. The median differences in 48-hour lesion volume associated with a 1%, 5%, and 10% increase in the probability of mRS scores of 0 to 2 for volumes <80 mL were 2 (interquartile range, 2–3), 10 (9–11), and 20 (18–23) mL, respectively. We found comparable associations when assessing computed tomography and magnetic resonance imaging separately.
CONCLUSIONS:
A difference of 2, 10, and 20 mL in 48-hour lesion volume, respectively, is associated with a 1%, 5%, and 10% absolute increase in the probability of achieving good functional outcome. These results can inform the design of future stroke trials that use 48-hour lesion volume as the primary outcome.