BACKGROUND: The excitotoxic amino acid glutamate is known to aggravate pre-existing neuropathology. Since volatile anesthetics increase plasma amino acid levels, we investigated if the anesthetics isoflurane and propofol increase plasma and cerebrospinal fluid (CSF) glutamate in neurosurgical patients. METHODS: In discectomized patients (n = 15), plasma glutamate was determined at 30 minute intervals before and during isoflurane anesthesia. In craniotomized patients (n = 66), plasma glutamate was assessed during and up to 24 hours after routine isoflurane or propofol anesthesia. CSF samples were withdrawn upon opening of the dura, before surgical manipulations. FINDINGS: During isoflurane anesthesia, plasma glutamate was significantly and reversibly increased in discectomized and craniotomized patients compared to healthy controls (56+/-6 microM; p<0.05), which was mostly sustained in male patients (males: 126+/-12 vs. females: 96+/-6 microM; p<0.05). With propofol, plasma glutamate was increased equally in men and women but to a lesser extent than with isoflurane (mean: 72+/-7 microM). CSF glutamate was significantly increased during isoflurane and propofol anesthesia compared to control lumbar CSF (1.2+/-0.1 microM; p<0.0001), being more prominent in patients with pre-existing brain edema receiving isoflurane (76+/-8 vs. propofol: 40+/-6 microM; p<0.05). CONCLUSIONS: The significant increases in plasma and CSF glutamate which were mostly sustained during isoflurane compared to propofol anesthesia should prompt the identification of anesthetic agents which do not impose a possible burden of glutamate-mediated excitotoxicity in patients with underlying compromised cerebral homeostasis. Detailed neuropsychological investigations following different anesthesia regimen are important to determine if transient elevations in CSF and plasma glutamate levels are of clinical relevance.