RESUMO
STUDY OBJECTIVE: To evaluate nationwide chronic myeloid leukemia (CML) treatment practices over an extended period and across multiple lines of tyrosine kinase inhibitor (TKI) therapy with imatinib, dasatinib, and nilotinib. DESIGN: Retrospective cohort study. DATA SOURCE: Veterans Health Administration (VHA) national database. PATIENTS: A total of 2873 VHA beneficiaries aged 18-89 years who had at least one encounter at any of the ~150 VHA hospitals and 800 VHA clinics, had a diagnosis code for CML, and filled at least one prescription for imatinib, nilotinib, or dasatinib between October 1, 2001, and September 30, 2010. MEASUREMENT AND MAIN RESULTS: The VHA database was used for the time period of October 1, 2000, to September 30, 2012, allowing for a 1-year observation period to identify CML treatments prior to study enrollment and a minimum of a 2-year follow-up period to assess study end points. Primary study end points included change in TKI treatment, gaps in TKI treatment, TKI treatment persistence, and patient survival. Persistence for each distinct line of treatment was defined as the time of continuous therapy, quantified by the number of days covered by the drug from treatment initiation until a 60-day gap in treatment was identified or a switch in treatment occurred. A Kaplan-Meier model was used to evaluate persistence and survival. Of the 2873 patients receiving first-line TKI treatment, 586 (20.4%) switched to a different TKI, constituting second-line treatment. Overall, 245 patients (8.5%) were switched again to third-line treatment. Only 4.4% of patients receiving first-line treatment experienced a gap in therapy of 60 or more days. First-line treatment persistence rates were 75%, 65%, and 55% for the first, second, and third years of treatment, respectively. Five-year survival with first-line treatment was 62%. CONCLUSION: In this national cohort of VHA patients, 1-year persistence of first-line TKI treatment was similar to that in prior studies. Five-year survival was comparable with that in other observational studies but was lower than that in prospective clinical trials. Persistence rates declined after the introduction of the new TKIs.
