Background Endovascular therapy (EVT) is standard of care for stroke caused by large vessel occlusion. Whether EVT should be performed under general anesthesia (GA) or conscious sedation (CS) is controversial. While a meta-analysis of randomized trials showed better outcome for EVT under GA, observational studies suggested the opposite. A proposed advantage of GA is better reperfusion achieved via more successful handling of the immobile patient. The aim of this study was to investigate if the good outcome seen in patients treated under GA was mediated by better reperfusion.Methods The meta-analysis included 368 individual patients from three randomized controlled trials, of whom 185 patients were randomized to CS. A mediator analysis was performed to examine if the better outcome in the GA arm was driven by higher reperfusion rate.Results The total effect showed a risk difference (RD) of 0.15 (95.04 to 0.25), associating GA with a beneficial outcome. The direct effect of GA constituted a large portion, with an RD of 0.12 (95.01 to 0.22), while only a small portion was mediated through the degree of reperfusion, with an RD of 0.03 (95.02 to 0.04).Conclusion The better outcome after EVT in the GA arm was mainly a direct effecttextemdashthat is, an effect that was not explained by better reperfusion. We also found a better outcome in the GA arm when reperfusion was not achieved. Whether this is an effect of the stable condition and blood pressure under GA or a neuroprotective effect will need to be investigated in future research.Data are available upon reasonable request.