The steep, time-dependent loss of benefit from reperfusion in clinical trials is consistent with loss of penumbra over the early hours of ischemia, as observed in animal models. Human imaging studies, however, show persistent penumbra for up to 48 h. We investigated core and penumbra volumes and collateral status in relation to time after stroke onset within the first 6 h.

Using data from three multimodal computer tomography-based studies in acute ischemic stroke patients <6 h after onset, we measured core and penumbra volumes, collateral status, and target mismatch (defined as core volume < 50 ml, perfusion lesion volume > 15 ml, mismatch ratio > 1.8). Patients were grouped by onset to imaging time (<3, 3–4.5, 4.5–6 h). We explored correlates of penumbra proportion by multivariable linear regression.

Analysis included 144 subjects. Across time epochs, neither proportions of penumbra (59%, 64%, 75% at <3, 3–4.5, >4. 5 h, respectively, p = 0.4) nor poor collaterals (15/56 (27%), 14/47 (30%), 4/15 (27%) at <3, 3–4.5, >4.5 h, p = 0.9) differed significantly. Penumbra proportion was not clearly related to time to imaging (R²= 0.003; p = 0.5) but a trend for divergent effects by collateral status was seen (slight increase in penumbra over time with good collaterals versus reduced with poor, interaction = 0.08). The proportion of patients with target mismatch did not vary by time (56%, 74%, and 67% at <3, 3–4.5, >4.5 h, p = 0.09).

In a cross-sectional sample imaged within 6 h, neither the proportions of penumbral tissue nor “target mismatch” varied by time from onset. A trend for reducing penumbra proportion only among those with poor collaterals may have pathophysiological and therapeutic importance.