Our aim was to evaluate the cost-effectiveness of docetaxel versus weekly paclitaxel regimen in patients with metastatic breast cancer previously treated with anthracycline from the Spanish National Health Service (NHS) perspective. A Markov model with a 21-day cycle duration was developed to estimate total treatment-related costs and clinical benefits over 5 years of docetaxel (100 mg/m(2)) and weekly paclitaxel (80 mg/m(2)). Patient data were obtained from the Randomized Phase III Study of Docetaxel Compared with Paclitaxel in Metastatic Breast Cancer (TAX-311) and Anglo-Celtic IV trials. Utilities were obtained from literature, and unitary costs (a,not sign2009) from a Spanish health-cost database and the Catalogue of Medicines. Cost and benefits [life-years gained (LYG) and quality-adjusted life years (QALY)] were discounted at 3%. Sensitivity analyses were performed. Docetaxel yields higher health benefits (1.83 LYG; 1.08 QALY) than paclitaxel (1.46 LYG; 0.84 QALY). Global costs (treatment, concomitant medication, adverse events management, progression, best supportive care, and end of life phase) per patient were a,not sign20,052 and a,not sign9,982 with docetaxel and paclitaxel, respectively. Incremental cost-effectiveness ratio (ICER) of docetaxel versus paclitaxel was a,not sign190/LYG and a,not sign295/QALY. Based on a a,not sign30,000/QALY threshold, docetaxel has 99% probability of being cost-effective. ICER was mostly sensitive to hazard ratio (HR) (when varied from 1.46 to 1.09; a,not sign3,517/ QALY), discount over the ex-lab price of TaxolA (R) (75%; a,not sign6,396/QALY) and granulocyte colony-stimulating factor (G-CSF) prophylactic treatment (when administered in 60% of cycles instead of 100%; cost saving). Variations in other inputs, such as time horizon (3-10 years), discount rate (0-5%), or adverse event cost (+/- 25%) were shown not to have relevant influence on the results. Compared to weekly paclitaxel, docetaxel therapy is cost effective for treating metastatic breast cancer patients.