Abstract 213: The Probability of Ischemic Stroke Therapeutic Targets at Extended Time Windows
BACKGROUND. A contributing factor in the decreased efficacy of thrombolysis over time may be a decrease in the probability of the therapeutic target, but this has not been well studied. Detection of therapeutic targets beyond 4.5 hr is the premise for reperfusion trials at extended time windows. We quantified the relationship of time from onset to the probability of detection of therapeutic targets.
METHODS. We analyzed a consecutive series of 1103 ischemic stroke patients with NIHSS > 3 and an MRI within 24 hr from last known well. Potential therapeutic targets were occlusion on MRA (ARTERY), M1 or M2 occlusion (MCAO), focal ischemia on perfusion MRI (PERFUSION), and diffusion-perfusion mismatch (MISMATCH). Logistic regression models evaluated hours from onset, NIHSS, age, and sex as predictors of target.
RESULTS. ARTERY, PERFUSION, and MISMATCH probabilities decreased over time (p < 0.0001) at approximately 1% per hour. More severe strokes increased the probability of detection by about 1% per NIHSS point. Logistic regression model found significant relationships of MISMATCH to hours, NIHSS, MCAO, and the interaction of NIHSS X MCAO (Table). MCAO patients had higher probabilities of MISMATCH, which decreased more rapidly over time (Graph). The probability of MISMATCH can be estimated by the logistic regression model: e.g., MCAO and NIHSS 12 at 9 hr = 78%; MCAO and NIHSS 14 at 20 hr = 47%.
CONCLUSION. The probability of potential targets of reperfusion therapy declines over the first 24 hours from stroke onset. A substantial proportion of patients have a target beyond the proven time window for thrombolytic therapy, and the proportion can be estimated by the logistic regression model.
- © 2012 by American Heart Association, Inc.