The prognostic significance of midline shift at presentation on survival in patients with glioblastoma multiforme

Purpose: While patients with glioblastoma multiforme (GBM) who present with midline shift have a presumably worse prognosis, there is little literatur e evaluating the prognostic significance of this presentation in multivaria te analysis in the context of other known prognostic factors. Methods and Materials: From March 1981 to September 1993, 219 patients unde rwent irradiation for intracranial glioma at our institution. One hundred f ourteen patients with a diagnosis of a primary GBM were analyzed for the in fluence of the presence of midline shift at diagnosis on survival with resp ect to other known prognostic factors, including age, Karnofsky performance status (KPS), and extent of surgery. Eighty-five patients (74%) presented with midline shift. Surgical treatment consisted of subtotal/total resectio n in 86 patients (75%). Among patients presenting with midline shift, 68 (8 0%) underwent subtotal/ total resection before irradiation. Results: Multivariate analysis of the entire cohort of patients found none of the potential prognostic factors analyzed to significantly influence sur vival. The overall median survival was 6 months. However, when multivariate analysis was limited to patients with a KPS of greater than or equal to 70 , only the presence of midline shift and age were found to significantly in fluence survival. Patients with a KPS greater than or equal to 70 and with midline shift present at diagnosis had a median survival of 8 months, as co mpared to 14 months for those not having midline shift at presentation (p = 0.04). Patients with a KPS greater than or equal to 70 and age > 50 years had a median survival of 5 months as compared to 11 months for those less t han or equal to 50 (p = 0.02). Conclusion: In this series, where 80% of patients who presented with a midl ine shift underwent decompressive resection of GBM before irradiation, the presence of midline shift at diagnosis remained an independent prognostic f actor influencing survival among good performance status patients. While th e role of decompressive surgery in this setting is likely of some benefit, the extent of this benefit remains to be defined. (C) 2000 Elsevier Science Inc.