Background Abiraterione is an expensive drug indicated for the treatment of metastasic castration resistant prostate cancer. In order to optimise its use, abiraterone is authorised for use under certain criteria in our hospital.

Purpose To analyse compliance with detailed criteria, response to and safety of abiraterone in clinical practice in a tertiary hospital.

Material and methods Retrospective observational study of all patients who received abiraterone for 4 years (January 2011–December 2014) through clinical history. Use criteria: performance status: ECOG ≤2 and Gleason < 8, serum transaminase levels <2.5 upper limit of normal and presence of bone and/or node but no visceral metastases.

Variables collected from the medical records were: age, performance status, stage of the disease, existence and location of metastases, pretreatments, treatment duration, causes of interruption, prostatic specific antigen (PSA), tolerance and safety.

Results 54 patients were included. Median age was 76 (57–85) years. 41% of patients were consistent with all established use criteria and in 42% of patients it was not possible to know if they were consistent with these criteria due to the absence of information in the clinical history. 17% of patients were not consistent with the criteria. At the beginning of treatment, liver function tests were normal in all patients.

Tolerance of abiraterone was appropriate in 87% of patients and 13% of patients showed moderate adverse events, such as gastrointestinal disorders and asthenia. Two patients had a large increase in transaminase levels, which forced discontinuation of the treatment.

16 patients continue in treatment at time of completion of the study and 38 patients had stopped the treatment during the study period. Median time for finished treatments was 6.1 months (1–31). Discontinuation was due to: 79% lack of efficacy, 5% death, 8% adverse events or intolerance, and 8% other causes.

Conclusion Efficacy and safety results were similar to other studies; a pharmacoeconomic analysis could help in the decision making process. Most patients with the required information available were consistent with the use criteria. The absent data from the clinical history shows that new tools to register and consult clinical data are needed.

No conflict of interest.