BACKGROUND Therapeutic hypothermia is a promising neuroprotective therapy with multiple mechanisms of action. We demonstrated the feasibility of thrombolysis combined with endovascular hypothermia, but not all patients achieved effective cooling. We sought to identify the factors that determined effective cooling. METHODS In 26 patients who underwent endovascular hypothermia, we computed four measures of effective cooling: time to reach target; Area-Under-the-Curve (AUC) 34 ratio; AUC-34; and AUC-35. By multivariate regression, we examined the effects of age, weight, starting temperature, body mass index, body surface area (BSA), gender, shivering, and total meperidine dose on the four outcome measures. RESULTS In univariate analyses, all four outcome measures were significantly influenced by BSA (p < 0.01 in all univariate analyses). Time to reach target temperature was quicker in older patients (p < 0.01). Shivering and meperidine dose were highly intercorrelated (r = 0.6, p < 0.01) and both marginally influenced all four outcome measures. In multivariate analysis, AUC ratio and time to reach target temperature were significantly influenced by BSA (p < 0.01) and meperidine (p < 0.05); AUC-34 was influenced only by BSA (p < 0.01). The AUC-35 was influenced by BSA (p < 0.01), shivering, and total meperidine dose (p < 0.05). CONCLUSIONS The most important determinant of effective cooling during endovascular hypothermia is BSA; larger patients are more difficult to cool and maintain in therapeutic range. Older patients cool more quickly. Shivering was well controlled by the combination of meperidine, buspirone, and surface counter-warming and only minimally influenced cooling effectiveness. Future trials of therapeutic hypothermia may include added measures to cool larger patients more effectively.