RESUMEN
BACKGROUND: Factors contributing to cerebral edema in the post-hyperacute period of ischemic stroke (first 24-72 hours) are poorly understood. Blood-brain barrier (BBB) disruption and postischemic hyperperfusion reflect microvascular dysfunction and are associated with hemorrhagic transformation. We investigated the relationships between BBB integrity, cerebral blood flow, and space-occupying cerebral edema in patients who received acute reperfusion therapy. METHODS: We performed a pooled analysis of patients treated for anterior circulation large vessel occlusion in the EXTEND-IA TNK and EXTEND-IA TNK part 2 trials who had MRI with dynamic susceptibility contrast-enhanced perfusion-weighted imaging 24 hours after treatment. We investigated the associations between BBB disruption and cerebral blood flow within the infarct with cerebral edema assessed using 2 metrics: first midline shift (MLS) trichotomized as an ordinal scale of negligible (<1 mm), mild (≥1 to <5 mm), or severe (≥5 mm), and second relative hemispheric volume (rHV), defined as the ratio of the 3-dimensional volume of the ischemic hemisphere relative to the contralateral hemisphere. RESULTS: Of 238 patients analyzed, 133 (55.9%) had negligible, 93 (39.1%) mild, and 12 (5.0%) severe MLS at 24 hours. The associated median rHV was 1.01 (IQR, 1.00-1.028), 1.03 (IQR, 1.01-1.077), and 1.15 (IQR, 1.08-1.22), respectively. MLS and rHV were associated with poor functional outcome at 90 days (P<0.002). Increased BBB permeability was independently associated with more edema after adjusting for age, occlusion location, reperfusion, parenchymal hematoma, and thrombolytic agent used (MLS cOR, 1.12 [95% CI, 1.03-1.20], P=0.005; rHV ß, 0.39 [95% CI, 0.24-0.55], P<0.0001), as was reduced cerebral blood flow (MLS cOR, 0.25 [95% CI, 0.10-0.58], P=0.001; rHV ß, -2.95 [95% CI, -4.61 to -11.29], P=0.0006). In subgroup analysis of patients with successful reperfusion (extended Treatment in Cerebral Ischemia 2b-3, n=200), reduced cerebral blood flow remained significantly associated with edema (MLS cOR, 0.37 [95% CI, 0.14-0.98], P=0.045; rHV ß, -2.59 [95% CI, -4.32 to -0.86], P=0.004). CONCLUSIONS: BBB disruption and persistent hypoperfusion in the infarct after reperfusion treatment is associated with space-occupying cerebral edema. Further studies evaluating microvascular dysfunction during the post-hyperacute period as biomarkers of poststroke edema and potential therapeutic targets are warranted.