Reduced brain oxygen response to spreading depolarization predicts worse outcome in ischaemic stroke.

Abstract

Spreading depolarization (SD) describes a propagating neuronal mass depolarization within the cerebral cortex that represents a mediator of infarct development and strongly stimulates the metabolic rate of O2 consumption. Here, we investigated the influence of spreading depolarization on brain tissue partial pressure of O2 (ptiO2) within the peri-infarct tissue of patients suffering malignant hemispheric stroke. This prospective observational trial included 25 patients with malignant hemispheric stroke that underwent decompressive hemicraniectomy followed by subdural placement of electrodes for electrocorticography (ECoG) and neighbouring implantation of a ptiO2 probe within the peri-infarcted cortex. Continuous side-by-side ECoG + ptiO2 recordings were obtained for 3-6 days postoperatively and analysed for the occurrence of SD-independent and SD-coupled ptiO2 changes, radiological findings, as well as their association with clinical outcome at 6 months. During the combined ECoG + ptiO2 monitoring period of 2604 h and among 1022 SDs, 483 (47%) SD-coupled ptiO2 variations were identified as biphasic (59%), hypoxic (36%) or hyperoxic (5%) ptiO2 responses that differed significantly (P < 0.0001). Among the remaining 538/1022 (53%) SDs, no SD-coupled ptiO2 response was detected, which we categorized as 'No response'. The overall infarct progression was 1.7% (interquartile range -2.5-10.9). SD characteristics regarding type, duration and frequency, as well as SD-independent baseline ptiO2 had no association with outcome. In contrast, a high occurrence rate and amplitude of SD-coupled variations in ptiO2 were associated with improved outcome at 6 months (occurrence: r = -0.62, P = 0.035; amplitude: r = -0.57, P = 0.024; Spearman correlation). In conclusion, an absent or reduced ptiO2 response to SD could indicate tissue-at-risk and help direct targeted treatment strategies in ischaemic stroke, which is further evidence that not all SDs are the same but tissue responses coupled to SD such as ptiO2 contain prognostic information. In particular, a lack of SD-coupled ptiO2 variations appears to be a predictor of worse outcome in large hemispheric stroke.