Cerebral swelling often occurs during craniotomy for cerebral tumors. The primary aim in this study was to determine risk factors (intracranial pressure [ICP], patient characteristics, histopathological features, neuroimaging characteristics, anesthetic regimen, and perioperative physiological data) predictive of brain swelling through the dural opening. As a secondary aim the authors attempted to define subdural ICP thresholds associated with brain swelling. The study population consisted of 692 patients (mean age 50+/-15 years) scheduled for elective craniotomy for supratentorial brain tumors. Brain swelling through the dural opening was estimated according to a four-point scale. The patients were dichotomized as those without cerebral swelling (that is, brain below the dura mater [59 patients] or brain at the level of the dura mater [386 patients]) and those with cerebral swelling (that is, moderate brain swelling [205 patients] or pronounced brain swelling [42 patients]). Logistic regression analysis was used to identify subdural ICP (odds ratio [OR] 1.9, 95% confidence interval [CI] 1.72-2.1, p < 0.0001), midline shift (OR 1.06, 95% CI 1.02-1.11, p = 0.008), a diagnosis of glioblastoma multiforme (OR 2.1, 95% CI 1.01-4.3, p = 0.047), and metastasis (OR 2.9, 95% CI 1.3-6.9, p = 0.01) as independent risk factors of intraoperative brain swelling. Thresholds for ICP associated with brain swelling were defined as follows: at an ICP less than 5 mm Hg, brain swelling rarely occurred (5% probability); at an ICP greater than 13 mm Hg, brain swelling occurred with 95% probability; and at an ICP greater than 26 mm Hg, severe brain swelling occurred with 95% probability. Subdural ICP is the strongest predictor of intraoperative brain swelling. It is possible to define thresholds of cerebral swelling and the authors recommend subdural ICP measurement as a tool to initiate preventive measures to reduce ICP before opening the dura mater.