Strength of Association between Infarct Volume and Clinical Outcome Depends on the Magnitude of Infarct Size: Results from the ESCAPE-NA1 Trial

Probability of mRS 0–2 at 90 days adjusted for infarct volume and key baseline variables (age, sex, NIHSS, ASPECTS, occlusion location, alteplase treatment, nerinetide treatment). A, Lines of best fit for volume groups 3 and 4, modelled within each volume group using a LOWESS function (red line). Lines of best fit are not shown for groups 1 and 2 because the volumes are so compressed by the x-axis scale. B, Volumes on a log scale better show the relationship at low infarct volumes (dashed blue line). Dashed vertical gray lines indicate curve shoulders that have been used to define infarct volume groups. C–F, Volume grouping according to visually-defined groupings at the curve shoulders. The graphs illustrate that at low volumes (IQR = 0–15 mL), the infarct volume has no relationship with outcome (C); other factors are deterministic of outcome. In volume groups 2 and 3 with volumes from 15 to 200 mL (D and E), there is progressive importance of volume as a predictor of outcome. At volumes of ≥200 mL (F), a large infarct volume is the determinant factor in predicting outcomes, which are generally poor. RR indicates risk ratio; 90d, ninety days; P, probability.