Spreading depolarizations are waves of mass neuronal and glial depolarization that propagate across the injured human cortex. They can occur with depression of neuronal activity as spreading depressions or isoelectric spreading depolarizations on a background of absent or minimal electroencephalogram activity. Spreading depolarizations are characterized by the loss of neuronal ion homeostasis and are believed to damage functional neurons, leading to neuronal necrosis or neurological degeneration and poor outcome. Analgesics and sedatives influence activity-dependent neuronal ion homeostasis and therefore represent potential modulators of spreading depolarizations. In this exploratory retrospective international multicentre analysis, we investigated the influence of midazolam, propofol, fentanyl, sufentanil, ketamine and morphine on the occurrence of spreading depolarizations in 115 brain-injured patients. A surface electrode strip was placed on the cortex, and continuous electrocorticographical recordings were obtained. We used multivariable binary logistic regression to quantify associations between the investigated drugs and the hours of electrocorticographical recordings with and without spreading depolarizations or clusters of spreading depolarizations. We found that administration of ketamine was associated with a reduction of spreading depolarizations and spreading depolarization clusters (P < 0.05). Midazolam anaesthesia, in contrast, was associated with an increased number of spreading depolarization clusters (P < 0.05). By using a univariate odds ratio analysis, we also found a significant association between ketamine administration and reduced occurrence of isoelectric spreading depolarizations in patients suffering from traumatic brain injury, subarachnoid haemorrhage and malignant hemispheric stroke (P < 0.05). Our findings suggest that ketamine-or another N-methyl-d-aspartate receptor antagonist-may represent a viable treatment for patients at risk for spreading depolarizations. This hypothesis will be tested in a prospective study.