Stereotactic biopsy is often performed for diagnostic purposes before treating patients whose imaging studies highly suggest glioma. Indications cited for biopsy include diagnosis and/or the "inoperability" of the tumor. This study questions the routine use of stereotactic biopsy in the initial management of gliomas. At The University of Texas M. D. Anderson Cancer Center, we retrospectively reviewed a consecutive series of 81 patients whose imaging studies suggested glioma and who underwent stereotactic biopsy followed by craniotomy/resection (within 60 days) between 1993 and 1998. All relevant clinical and imaging information was reviewed, including computerized volumetric analysis of the tumors based on pre- and postoperative MRI. Stereotactic biopsy was performed at institutions other than M. D. Anderson in 78 (96%) of 81 patients. The majority of tumors were located either in eloquent brain (36 of 81 = 44%) or near-eloquent brain (41 of 81 = 51%), and this frequently was the rationale cited for performing stereotactic biopsy. Gross total resection (>95%) was achieved in 46 (57%) of 81 patients, with a median extent of resection of 96% for this series. Diagnoses based on biopsy or resection in the same patient differed in 40 (49%) of 82 cases. This discrepancy was reduced to 30 (38%) of 80 cases when the biopsy slides were reviewed preoperatively by each of three neuropathologists at M. D. Anderson. Major neurologic complications occurred in 10 (12.3%) of 81 surgical patients and 3 (3.7%) of 81 patients undergoing biopsy. Surgical morbidity was probably higher in our series than it would be for glioma patients in general because our patients represent a highly selected subset of glioma patients whose tumors present a technical challenge to remove. Stereotactic biopsy is frequently inaccurate in providing a correct diagnosis and is associated with additional risk and cost. If stereotactic biopsy is performed, expert neuropathology consultation should be sought.