Pilot studies of hypothermia for stroke suggest a potential benefit in humans. We sought to test whether hypothermia decreases post-ischemic edema using CT scans from a pilot trial of endovascular hypothermia for stroke. Eighteen patients with acute ischemic stroke underwent therapeutic hypothermia (target = 33 degrees C) for 12 or 24 h followed by a 12-h controlled re-warm using an endovascular system. CT scans obtained at baseline, 36-48 h (right after cooling and re-warming) and 30 days were digitized, intracranial compartment volumes measured using a validated stereological technique, and the calculated change in CSF volume between the three time-points were used as an estimate of edema formation in each patient. Patients were grouped retrospectively for analysis based on whether they cooled effectively (i.e., to a temperature nadir of less than 34.5 degrees C within 8 h) or not. Eleven patients were cooled partially or not at all, and seven were effectively cooled. Baseline demographics and compartment volumes and densities were similar in both groups. At 36-48 h, the total CSF volume had significantly decreased in the not-cooled group compared to the cooled group (P < 0.05), with no significant difference in mean volume of ischemia between them (73 +/- 73 ml vs. 54 +/- 59 ml, respectively), suggesting an ameliorative effect of hypothermia on acute edema formation. At 30 days, the difference in CSF volumes had resolved, and infarct volumes (73 +/- 71 ml vs. 84 +/- 102 ml, respectively) and functional outcomes were comparable. Endovascular hypothermia decreases acute post-ischemic cerebral edema. A larger trial is warranted to determine if it affects final infarct volume and outcome in stroke.