Cost effectiveness of trastuzumab in the adjuvant treatment of early breast cancer: a lifetime model

Abstract

Background: Recent randomised trials have demonstrated a statistically significant effect of trastuzumab on disease-free survival when used as adjuvant therapy for human epidermal growth factor receptor 2 protein (HER2)-positive resectable early stage breast cancer, with a treatment course lasting either 9 or 52 weeks. However, the cost effectiveness of adjuvant trastuzumab with respect to mortality remains uncertain, especially in an Australian setting.

Objective: To estimate the cost effectiveness of trastuzumab in Australia, in a cohort of 50-year-old patients with HER2-positive breast cancer over a lifetime, using (i) disease-free survival and cardiotoxicity data from recent randomised trials; (ii) information on long-term survival of patients with treated primary breast cancer; and (iii) costs of treating local and distant relapses and disease from causes other than breast cancer.

Methods: A Markov model consisting of four health states (remission, loco-regional recurrence, metastatic disease and death) was developed. Transition probabilities corresponded to patterns of relapse and metastases seen in recent trials. The model was run until age 100 years to allow calculation of average survival. Outcome measures were life-years and QALYs (calculated using utility weights reported in the literature). The model was calibrated to reflect literature evidence that the risk of breast cancer recurrence following primary treatment diminishes progressively to zero after about 20 years. It was assumed that the morbidity benefit of trastuzumab observed in trials would be present for 5 years but would then diminish progressively to zero after 8 years. Costs (year 2005 values) and benefits were discounted at 3% per annum.

Results: For every 1000 patients treated with a 52-week course of trastuzumab, there were 136 fewer breast cancer deaths (relative risk reduction 28%). The incremental cost-effectiveness ratios (ICERs) were Australian dollars ($ A)13 730 per year of life saved (YOLS) and $ A22 793 per QALY. The net incremental cost was $ A56.3 million ($ A414 012/cancer death avoided). Cost effectiveness declined (ICER = $ A27 734/QALY) in older patients (age 65 years at treatment initiation). The ICER was driven mainly by the drug acquisition costs, the assumption of the duration of benefit and the discount rate. Cost offsets from reduced costs of treating recurrent or metastatic breast cancer during follow-up were present but these factors were of less importance according to sensitivity analyses. The 9-week treatment schedule approached economic dominance (ICER = $ A1700/QALY) because of decreased costs, improved relative risk for prevention of metastases and more cancer deaths avoided (196).

Conclusion: The results suggest that trastuzumab as adjuvant therapy for early breast cancer may be cost effective when given over either 52 or 9 weeks at current acquisition costs in Australia. However, the overall budget impact of the 52-week course is significant, and the 9-week course appears economically attractive.