The purpose of the present study was to determine the effect of surgery on the time length and quality of survival in patients with recurrent glioblastoma multiforme. Two groups were compared; the first included 18 patients who underwent surgery at the time of tumour recurrence. The second group included 36 patients who did not undergo surgery at the time of tumour recurrence. Both groups were matched according to the following criteria: gender, age, Karnofsky Performance Scale (KPS) score, at the time of initial surgery and of tumour recurrence, extent of initial surgery, interval between initial surgery and tumour reccurence. Both groups received conventional treatment after initial surgery. There are no statistically significant differences between the two groups as regards to the previously mentioned criteria. After tumour recurrence, the median survival time was 5 months in the group of patients undergoing a second resection and 2 months in the group of patients not undergoing repeat surgery. The difference was statistically significant on univariate analysis. Moreover, the median length of time spent in an acceptable condition (KPS >/=60) from the time of tumour recurrence was found to be significantly longer in patient who underwent a second resection (4 months) compared with patients who did not undergo repeat surgery (1 month). Even in a relatively favorable subgroup of reoperated patients, the survival benefit although significant was only 3 months. It was impossible to completely match the two groups of patients suggesting that the difference might have been even less. Although symptomatic improvement is modestly achieved by repeat surgery, its transient nature necessitates clear discussion with patient and family on an individual basis.