Abstract
Background/Aims Tamoxifen is a widely prescribed endocrine therapy for prevention and treatment of breast cancer in pre- and post-menopausal women and is commonly recommended for five years duration. While most people tolerate tamoxifen reasonably well, thromboembolic events such as deep vein thrombosis, pulmonary embolism, myocardial infarction or cerebrovascular accident can be serious complications. Study aims include evaluation of epidemiology of tamoxifen-related thromboembolic event in a large community setting, determination of rate and temporal distribution of thromboembolic events, description of patient clinical characteristics, analysis of time from tamoxifen initiation to first thromboembolic event, description of incidence of events among subgroups of patients related to antiplatelet, anticoagulant and statin medication use.
Methods Retrospective electronic extraction of tamoxifen-use patient data from Marshfield Clinic/ St. Joseph’s Cancer Registry and the Marshfield Clinic Electronic Medical Record (EMR) from 1/1/1992 – 6/31/2009, with manual data validation. Medication use was identified by MedsManager and by Freepharma™. Thromboemoblic events and comorbidity diagnoses were identified by ICD9 codes, and cancer diagnosis by ICD9 code or inclusion in the Registry. Analysis included descriptive statistics, Kaplan-Meier estimates of event-free survival over time and estimates of event rates and hazard ratios. Risk factors were assessed in univariate and multivariate analyses.
Results Primary analysis was limited to 3283 women with breast cancer. Mean age at first tamoxifen use was 62.1 years (range 22 – 97). Tamoxifen adherence rates (censored for death or 18 months after last visit) showed 88.6% of women on tamoxifen at year one, 78.1% at year 2, 67.4% at year 3 and 56.9% at year 4. Thromboembolic events (venous and arterial) occurred in 5% of women under age 45, 7% of women age 45–54, 14% in women age 55–64, 19% in women age 65–74 and 27% in women older than 75 years. Thromboembolism continued to occur over time, unlike other studies that reported clustering of thromboembolism in the first year.
Conclusions Tamoxifen use should be recommended with caution, and patients monitored closely over the entire period of tamoxifen use for development or prevention of thromboembolic events. Further investigation into tamoxifen discontinuation rates is warranted.
