OBJECTIVE Elderly oncology patients are not enrolled in early-phase trials in proportion to the numbers of geriatric patients with cancer. There may be concern that elderly patients will not tolerate investigational agents as well as younger patients, resulting in a disproportionate number of dose-limiting toxicities (DLT). Recent single-institution studies provide conflicting data on the relationship between age and DLT. METHODS We retrospectively reviewed data about patients treated on single-agent, dose-escalation, phase I clinical trials sponsored by the Cancer Therapy Evaluation Program (CTEP) of the National Cancer Institute. Patients' dose levels were described as a percentage of maximum tolerated dose, the highest dose level at which <33% of patients had a DLT, or recommended phase II dose (RP2D). Mixed-effect logistic regression models were used to analyze relationships between the probability of a DLT and age and other explanatory variables. RESULTS Increasing dose, increasing age, and worsening performance status (PS) were significantly related to an increased probability of a DLT in this model (P < 0.05). There was no association between dose level administered and age (P = 0.57). CONCLUSIONS This analysis of phase I dose-escalation trials, involving more than 500 patients older than 70 years of age, is the largest reported. As age and dose level increased and PS worsened, the probability of a DLT increased. Although increasing age was associated with occurrence of DLT, this risk remained within accepted thresholds of risk for phase I trials. There was no evidence of age bias on enrollment of patients on low or high dose levels.