Tumor growth depends on cell division and cell death. To investigate the role of apoptosis in tumor cell death, we examined 83 cases of glial tumors using in situ nonradioactive tailing of DNA breaks. In addition, since p53 protein may participate in the regulation of apoptosis in glioblastoma, we compared the apoptosis ratio (AR) with the labeling index (LI) of p53 protein immunopositivity. The AR in glial tumor parenchyma ranged from 0 to 1.4%: mean AR +/- standard deviation was 0.4 +/- 0.4% (range, 0-1.4) for glioblastoma, 0.3 +/- 0.3% (range, 0.01-0.83) for anaplastic astrocytoma, 0.1 +/- 0.1% (range, 0-0.41) for low-grade astrocytoma, 0.006 +/- 0.008% (range, 0-0.02) for pilocytic astrocytoma, 0.2 +/- 0.2% (range, 0-0.62) for oligodendroglioma and 0.003 +/- 0.004% (range, 0-0.01) for ependymoma. ARs were significantly higher in higher-grade astrocytic tumors than in lower-grade tumors (Mann-Whitney U test: P = 0.0003), although wide variability in each group resulted in overlapping between the groups. p53 protein immunopositivity (more than 25% of nuclei) was found in 15 of 32 glioblastoma cases, while in the remaining 17 none or only a low percentage (up to 6%) of the nuclei were positive. In p53 protein-positive cases mean AR (0.51 +/- 0.47%) was not significantly higher than that in p53 protein-negative cases (0.22 +/- 0.23%; P = 0.1681).