Objective: The 2006 American Society of Clinical Oncology guideline recommends primary prophylaxis (PP) with colony-stimulating factor (CSF) for elderly patients with diffuse aggressive lymphoma receiving chemotherapy, based on the assumption of equal survival and studies showing that CSF saved costs by reducing hospitalization from febrile neutropenia (FN). These analyses examined only one cycle of chemotherapy, and did not consider the costs of CSF in subsequent cycles, the strategy of secondary prophylaxis (SP) or patients' preferences. This study examined the cost-effectiveness of PP with SP. Methods: We conducted a cost-utility analysis to compare PP with CSF to SP with CSF for diffuse aggressive lymphoma. We used a Markov cohort model with a time horizon of 8 cycles of chemotherapy (i.e. 24 weeks), using a payer's perspective (Ontario Ministry of Health). Ontario's 2006 health economic data was used. The cost of hospitalization for FN was obtained from Ontario Case Costing Initiative. Data for efficacies of CSF, probabilities and utilities were obtained from published literature. Monte Carlo simulation was conducted. Results: The ICER of PP to SP was $739,999/QALY. One-way sensitivity analyses (willingness-to-pay threshold =$100,000) showed that if PP were to be cost-effective, the cost of hospitalization for FN had to be >$31,138 (2.5 times > base case), the cost of CSF per cycle