Response to endovascular reperfusion is not time-dependent in patients with salvageable tissue

Abstract

Objective: To evaluate whether time to treatment modifies the effect of endovascular reperfusion in stroke patients with evidence of salvageable tissue on MRI.

Methods: Patients from the Diffusion and Perfusion Imaging Evaluation for Understanding Stroke Evolution 2 (DEFUSE 2) cohort study with a perfusion-diffusion target mismatch were included. Reperfusion was defined as a decrease in the perfusion lesion volume of at least 50% between baseline and early follow-up. Good functional outcome was defined as a modified Rankin Scale score ≤2 at day 90. Lesion growth was defined as the difference between the baseline and the early follow-up diffusion-weighted imaging lesion volumes.

Results: Among 78 patients with the target mismatch profile (mean age 66 ± 16 years, 54% women), reperfusion was associated with increased odds of good functional outcome (adjusted odds ratio 3.7, 95% confidence interval 1.2-12, p = 0.03) and attenuation of lesion growth (p = 0.02). Time to treatment did not modify these effects (p value for the time × reperfusion interaction is 0.6 for good functional outcome and 0.3 for lesion growth). Similarly, in the subgroup of patients with reperfusion (n = 46), time to treatment was not associated with good functional outcome (p = 0.2).

Conclusion: The association between endovascular reperfusion and improved functional and radiologic outcomes is not time-dependent in patients with a perfusion-diffusion mismatch. Proof that patients with mismatch benefit from endovascular therapy in the late time window should come from a randomized placebo-controlled trial.

Figure 1. Effect of time to treatment on the association between reperfusion and good functional outcome

The adjusted odds ratio (OR) for good functional outcome with reperfusion in patients with the target mismatch (n = 78) is displayed as a function of time from stroke onset to the start of the endovascular procedure. The 95% confidence interval (CI) is indicated by dashed lines. The association between reperfusion and good functional outcome is not modified by time to treatment in patients with target mismatch (adjusted p value for the interaction between reperfusion and time to treatment = 0.6). Values are derived from a multivariate logistic regression model adjusted for age and baseline diffusion-weighted imaging lesion volume.

Figure 2. Day 90 modified Rankin Scale scores among patients with the target mismatch stratified by treatment time and reperfusion status

The distribution of the 90-day modified Rankin Scale (mRS) scores is shown for target mismatch patients stratified by treatment time (≤6 hours vs >6 hours) and reperfusion status. Based on multivariate logistic regression, reperfusion is associated with an increased chance of good functional outcome (odds ratio adjusted for age and baseline diffusion-weighted imaging lesion volume 4.0, 95% confidence interval 1.3–13, p = 0.02), whereas time to treatment (p = 0.6) and the interaction between reperfusion and time to treatment are not (p = 0.6).

Figure 3. Effect of time to treatment on the probability of good functional outcome with reperfusion

The probability of good functional outcome in patients with target mismatch who achieved reperfusion (n = 46) is displayed as a function of time from stroke onset to the start of the endovascular procedure. The 2 curved light blue lines indicate the 95% confidence interval (CI). Short horizontal black lines indicate the mean observed probability of good functional outcome in patients grouped (n = 7 or 8) according to their consecutive time from symptom onset to start of procedure. Black ovals indicate the mean time from symptom onset to start of procedure for each of the 6 subgroups. Blue circles indicate the observed outcome for each individual patient in the cohort, where 0 corresponds to a modified Rankin Scale (mRS) score >2 and 1 to an mRS score ≤2 at day 90. “p for trend” is the adjusted p value for the association between time to treatment and the probability of good functional outcome (p = 0.2). Values are derived from a logistic regression model with time from stroke onset to start of the endovascular procedure as the predictor variable. Probabilities are adjusted for age and baseline diffusion-weighted imaging lesion volume.