Eighty-five "well-differentiated" astrocytomas in adults (age, greater than or equal to 18 years), operated on between 1950 and 1982, were retrospectively reviewed. The pilocytic variant was not included. Twenty-four clinical and 8 histological factors were analyzed to investigate their importance in predicting length of survival. Multivariate analysis showed that the following variables were correlated with survival time (P less than 0.01): extent of surgical removal, altered consciousness during preoperative examination, focal deficit as presenting symptom, performance status (Karnofsky rating) after surgery, and vessel size in the surgical specimen. Total removal of the tumor was related to a higher 5-year survival rate (51%) than subtotal removal (23.5%), and none of the patients with partial removal survived more than 5 years. Postoperative radiotherapy (40-55 Gy) improved only the 1- and 3-year survival rates. Based on the significant factors provided by multivariate analysis, a score was developed to detect subgroups with different prognoses. Median survival time ranged from 383 days for patients with a score greater than or equal to 2.5 to 1,533 days for those with a score less than 0.5; no patient with a score greater than or equal to 1.5 survived more than 10 years. The percentage of recurring astrocytomas that showed anaplastic areas in the second biopsy specimen was 79%. Total surgical removal is the most important factor in the management of well-differentiated astrocytomas, whereas the efficacy of postoperative radiotherapy still needs to be confirmed by prospective and randomized studies. The rationale for treating incompletely resected astrocytomas with radiation therapy could lie in the high incidence of malignant transformation.