Economic evaluation of hormonal therapies for postmenopausal women with estrogen receptor–positive early breast cancer in Canada

Original Article

Medical Economics

Economic evaluation of hormonal therapies for postmenopausal women with estrogen receptor–positive early breast cancer in Canada


S. Djalalov , PhD , * , J. Beca , MSc * ,, E. Amir , MD PhD § , M. Krahn , MD , § # , M.E. Trudeau , MD § ** , J.S. Hoch , PhD * ,,, §

* Centre for Excellence in Economic Analysis Research, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, Toronto, ON.
Pharmacoeconomics Research Unit, Cancer Care Ontario, Toronto, ON.
Canadian Centre for Applied Research in Cancer Control, Cancer Care Ontario, Toronto, ON.
§ University of Toronto, Toronto, ON.
Princess Margaret Hospital, Toronto, ON.
# The Toronto Health Economics and Technology Assessment Collaborative, Toronto, ON.
** Sunnybrook Health Sciences Centre, Toronto, ON.


doi: http://dx.doi.org/10.3747/co.22.2120


ABSTRACT

Background

Aromatase inhibitor ( ai ) therapy has been subjected to numerous cost-effectiveness analyses. However, with most ai s having reached the end of patent protection and with maturation of the clinical trials data, a re-analysis of ai cost-effectiveness and a consideration of ai use as part of sequential therapy is desirable. Our objective was to assess the cost-effectiveness of the 5-year upfront and sequential tamoxifen ( tam ) and ai hormonal strategies currently used for treating patients with estrogen receptor ( er )–positive early breast cancer.

Methods

The cost-effectiveness analysis used a Markov model that took a Canadian health system perspective with a lifetime time horizon. The base case involved 65-year-old women with er -positive early breast cancer. Probabilistic sensitivity analyses were used to incorporate parameter uncertainties. An expected-value-of-perfect-information test was performed to identify future research directions. Outcomes were quality-adjusted life-years ( qaly s) and costs.

Results

The sequential tam ai strategy was less costly than the other strategies, but less effective than upfront ai and more effective than upfront tam . Upfront ai was more effective and less costly than upfront tam because of less breast cancer recurrence and differences in adverse events. In an exploratory analysis that included a sequential ai tam strategy, ai tam dominated based on small numerical differences unlikely to be clinically significant; that strategy was thus not used in the base-case analysis.

Conclusions

In postmenopausal women with er -positive early breast cancer, strategies using ai s appear to provide more benefit than strategies using tam alone. Among the ai -containing strategies, sequential strategies using tam and an ai appear to provide benefits similar to those provided by upfront ai , but at a lower cost.

KEYWORDS: Breast cancer , cost-effectiveness , tamoxifen , aromatase inhibitors

1.  INTRODUCTION

About 75% of all incident breast cancers occur in post-menopausal women, and approximately 80% of those breast cancers express the estrogen receptor (that is, they are er -positive)1. Recurrence and death from er -positive breast cancer can be effectively reduced through estrogen suppression or antagonism. It is estimated that each year in Canada, 7320 er -positive breast cancers will occur in women 65 years of age or older2.

A number of large randomized controlled trials have shown that, in postmenopausal women with er -positive breast cancer, the use of aromatase inhibitors ( ai ) either upfront or after initial treatment with tamoxifen is associated with a statistically significant reduction in the risk of recurrence and, in some cases, an improvement in overall survival3,4. A recently published meta-analysis5 of breast cancer outcomes in adjuvant trials of ai versus tamoxifen ( tam ) considered two cohorts: specifically, trials comparing 5 years of an ai with 5 years of tam , both starting soon after surgery4,6; or trials comparing 5 years of upfront tam with tam therapy switched after 2–3 years to 2–3 years of an ai (for a total of 5 years of hormonal therapy in both groups)79. The first cohort comprised 9856 patients with a mean 5.8 years of follow-up. At 5 years, ai therapy was associated with an absolute 2.9% [standard error ( se ): 0.7%] decrease in recurrence (9.6% for ai vs. 12.6% for tam , 2 p < 0.001) and a nonsignificant absolute 1.1% ( se : 0.5%) decrease in breast cancer mortality (4.8% for ai vs. 5.9% for tam , 2 p = 0.1). The second cohort comprised 9015 patients with a mean 3.9 years of follow-up. At 3 years from treatment divergence (that is, approximately 5 years after starting hormonal treatment), sequential therapy with an ai was associated with an absolute 3.1% ( se : 0.6%) decrease in recurrence (5.0% for ai vs. 8.1% for tam from divergence, 2 p < 0.001) and an absolute 0.7% ( se : 0.3%) decrease in breast cancer mortality (1.7% for ai vs. 2.4% for tam from divergence, 2 p = 0.02). The meta-analysis concluded that recurrence rates are significantly lower with ai s than with tam , either as initial upfront therapy or after 2–3 years of tam 5. Consequently, practice guidelines recommend that all such women receive ai at some point in their adjuvant therapy10.

Therapy with ai has already been subjected to numerous cost-effectiveness analyses from the Canadian perspective1114. However, many of the analyses used assumptions that the matured clinical data do not support, including the assumption that differences in disease-free survival will translate into improvements in overall survival. Furthermore, most ai s have now reached the end of patent protection and are available from generic manufacturers at substantially reduced cost. It is therefore desirable to repeat the cost-effectiveness analyses to identify the optimal strategy for postmenopausal women with er -positive early breast cancer.

The latest guidelines include recommendations for extended treatment with tam beyond 5 years, but no data for ai s beyond 5 years (either upfront or as part of sequential strategies) are currently available; trials are ongoing10. Here, we report the results of a cost-effectiveness analysis, from the Canadian health care perspective, of several 5-year hormonal therapy strategies (upfront and sequential tam and ai therapies) currently being used in patients with er -positive early breast cancer.

2.  METHODS

2.1  Treatment Strategies

We compared three treatment strategies: two upfront therapies (5 years of tam , 5 years of ai ) and a sequential treatment [ tam for 2 years followed by ai for 3 years ( tam ai )]. In an exploratory analysis, we also compared sequential ai for 2 years followed by tam for 3 years ( ai tam ).

2.2  Model

We constructed a state-transition model to simulate lifetime health profiles and to compare treatment strategies in a cohort of 65-year-old postmenopausal women with er -positive early breast cancer. During each yearly cycle, patients faced the possibility of progression to more advanced stages, of adverse events ( ae s), or of death from other causes (Figure 1). Early breast cancer patients might develop a contralateral tumour, locoregional recurrence, or distant metastasis.

 


 

Figure 1 Markov model used in the analysis. a Includes contralateral tumour. b Adverse events of two types, with separate health states, are considered in the model: short-term events (thromboembolism, fractures, arthralgia) and chronic events (stroke, cardiac events, endometrial cancer).

Hormonal therapies increase the risk of ae s, including stroke, cardiac events, endometrial cancer, thromboembolism, fractures, and arthralgia. The effects of each ae were considered separately. Transient ae s—such as thromboembolism, fractures, and arthralgia, which have a short-term cost–utility and mortality impact—were tracked in individual health states for 1 year and subsequently grouped into a post-short-term ae health state. Chronic ae s—such as stroke, cardiac events, and endometrial cancer—were tracked using separate health states to incorporate their long-term management costs, disutilities, and elevated risks of death over and above the breast cancer risks for recurrence or death. We used data from the medical literature (Table i) to assign cost and utility estimates (“ qaly weights”) to each health state over a lifetime time horizon. Costs and health effects were discounted at a common rate of 5%.

TABLE I   Mortality rates, costs, and utilities

 

2.3  Assumptions

We made a few simplifying assumptions when building the Markov model. First, to make the sequential ai tam strategy from its single trial comparable with the meta-analysis data, we assumed that the sequential ai tam and upfront ai strategies have equal risks of recurrence. That assumption was based on results from the Breast International Group ( big ) 1–98 trial3, which showed no difference in recurrence between the two strategies after 5 years. Second, we combined the contralateral tumour and locoregional recurrence health states into one health state. Third, we treated the ai drug class as a group without reference to a specific drug. The latter assumption took account of the fact that adjuvant and neoadjuvant studies have both failed to show any differences between letrozole, anastrozole, and exemestane43,44. Fourth, we assumed that patients could experience non–breast cancer death from any health state, but that breast cancer–related death occurs only after distant metastasis.

2.4  Risk of Death

The annual risk of death among women with distant metastasis treated with hormonal therapy was obtained from a large phase iii randomized trial, Protocol 025 (P025), which compared ai with tam as first-line therapy in metastatic breast cancer15 (Table i). We estimated mortality rates for breast cancer patients without distant metastasis by eliminating deaths from breast cancer, endometrial cancer, any cardiac event, thromboembolic events, and stroke from the age-specific annual mortality rates for Canadian women. Because long-term survival data for patients on sequential therapies beyond 5 years were unavailable, we assumed that the mortality rate from distant recurrence was similar for all four strategies over the expected life of each patient.

Risks of death from hormonal therapy–related ae s were obtained from various studies. The baseline mortality from stroke (13%) and any cardiac event (9%) were obtained from a systematic review that included twenty studies of long-term cancer survivors18,19. The base-case mortality from thromboembolic events (20%) was obtained from Canadian patients with solid tumours and acute venous thromboembolism participating in an open-label randomized controlled trial16,17. The risk of death from fracture in the 1 year after the fracture was estimated to be 22% based on a systematic review of prospective and retrospective studies of 64,316 patients after hip fracture surgery2022. An endometrial cancer mortality rate of 3% was obtained from a 10-year prospective trial of patients with stages i iii endometrial carcinoma23,24. No excess mortality risk accrues to arthralgia or myalgia; age-specific mortality rates for Canadian females were therefore used for those ae s.

2.5  Transition Probabilities

Annual probabilities for first breast cancer events in years 1–5 for patients treated with upfront tam and ai and sequential tam ai therapy were based on a meta-analysis5 updated with the most recently published results from two large studies: big 1–983 and atac 25 (Table ii). The atac trial data with 10-year follow-up were used to calculate the annual probabilities for 6 years onward, and we assumed that the probabilities of recurrence after 6 years were the same for the sequential strategies as for upfront ai . The annual probabilities of recurrence after contralateral tumour and locoregional recurrence were obtained from a previous cost-effectiveness study of early breast cancer treatment11, which were calculated by pooling Kaplan–Meier estimates of disease progression in patients with locoregional recurrence from eight studies; we assumed that those probabilities were the same for all strategies. Probabilities of developing ae s (endometrial cancer, cardiac events, thromboembolic events, fractures, arthralgia, and stroke) while on treatment with each of the strategies were based on the number of grades 3–5 ae s reported in the big 1–98 study3.

TABLE II  Transition probabilities, by treatment strategya

 

2.6  Costs

The cost of generic letrozole was used for ai . Costs for tam and letrozole were based on Ontario Drug Benefit Formulary costs, including a 5% dispensing fee27 (Table i). Costs for breast cancer events were based on a study that integrated data from Canada’s Population Health Model and multiple other sources; it determined costs for breast cancer at diagnosis and for the management of local and distant recurrence26. Costs for locoregional recurrence, including contra-lateral tumour, were based on the average costs of local recurrence, including local treatment followed by systemic therapy. Costs for distant metastasis were based on the weighted average of metastatic sites and included diagnosis and treatment. Follow-up treatment for local recurrence included physician assessments, clinic costs, hematology, biochemistry, a bone scan, chest radiography, liver ultrasonography, and annual mammography, and were based on the average cost of locoregional recurrence (stage ii ). Breast cancer death was allocated costs for palliative care. Costs for treatment of stroke, thromboembolic events, any cardiac event (myocardial infarction, angina), fracture, and endometrial cancer were obtained from published Canadian studies2834,36,45. The treatment cost for arthralgia was based on the annual cost of anti-inflammatory medication (celecoxib 100 mg daily)35. When necessary, all costs were inflated to 2011 Canadian dollars using the medical component of the Canadian Consumer Price Index46.

2.7  Utilities

Utility weights for breast cancer health states were obtained from a meta-analysis of 476 utility values from forty-nine studies39 (Table i). Data for metastatic and early breast cancer states were synthesized by meta-regression. Of the three models presented in the meta-analysis, we used the numbers from model 1, which were weighted by the inverse of the standard deviation for base-case parameters of locoregional recurrence (including contralateral tumour) and distant metastasis. To identify upper and lower bounds for sensitivity analyses, we used the model 2 values for “treatment type” and “response to treatment,” which were weighted by sample size. The utility estimate for individuals without symptomatic breast cancer (“disease-free” health state) was based on the Medical Expenditure Panel Survey conducted in the United States during 2000–200237,38. Preference-based scores were determined using the EQ-5D health questionnaire (EuroQoL, Rotterdam, Netherlands) which was administered within a group of 20,980 women with chronic conditions, of whom about 17% were more than 65 years of age. Utility estimates for thromboembolic events ( n = 126) were obtained from the same study. Utilities for any cardiac events—including myocardial infarction, ischemic heart disease and heart failure ( n = 456), and stroke ( n = 69)—were also derived using the EQ-5D from U.K. patients who had previously experienced cardiac adverse events40. Because the latter study was conducted in patients with type 2 diabetes, we used the percentage utility decrement resulting from the event and subtracted it from the breast cancer disease-free health state utility. Utility values for fractures were obtained from a systematic review of twenty-seven studies for osteoporosis-related conditions41. We used EQ-5D utility values from hip fracture, because hip fractures represent the largest financial burden and the highest incidence34. Utility weights for endometrial cancer were based on representative data from the U.S. National Health Interview Survey and the Healthy People 2000 survey conducted in members of the U.S. population with chronic conditions42.

2.8  Sensitivity Analyses

Plausible ranges of high and low values for the model variables were used to conduct deterministic and probabilistic sensitivity analyses examining the influence and uncertainty of all parameters. To identify parameters with influence on the optimal strategy, we used net benefit to conduct threshold analyses. Several scenarios explored the effects of assumptions about carryover benefits. Monte Carlo simulation was used to perform probabilistic sensitivity analyses. We used gamma distributions to represent uncertainty about cost parameters, because cost data are skewed and cannot be negative. We used beta distributions for the probabilities and utilities, because those estimates are confined to a 0–1 range.

All parameters were randomly sampled from their assigned distributions, and one thousand simulations were performed. We estimated the likelihood of each treatment strategy being optimal across a range of willingness-to-pay ( wtp ) thresholds and illustrated uncertainty in our results with a cost-effectiveness acceptability curve. To assess the value of additional information, we calculated the expected value of perfect information ( evpi ) with a 10-year lifespan of the testing technology, and the partial evpi for the input parameters at various wtp thresholds. The evpi analyses inform decision-makers about the expected value of conducting more research to support a decision. From the approximately 75% population prevalence estimates for er -positive tumours in Canada47, we estimated the number of patients with er -positive breast cancers who were 65 years of age and older in Canada to be 14,160. Based on that population estimate, we estimated the population evpi at various wtp thresholds.

Tables i and ii list selected variables in the model.

3.  RESULTS

3.1   Base-Case Results

Base-case results were based three strategies (up-front tam , upfront ai , and sequential tam ai ), with a calibrated sequential ai tam strategy included in an exploratory analysis. Total survival time was 2.4 months more per patient with the upfront ai and sequential tam ai strategies than with upfront tam (Table iii). Total cost per patient was $175 lower with upfront ai and $581 lower with sequential tam ai than with upfront tam . Overall, compared with upfront tam , ai -containing strategies were more effective and less costly as a result of less recurrence and differences in adverse events. Among the ai- containing strategies, sequential therapy provided similar benefits, but was less costly, mainly because of lower drug costs (sequential therapy strategies use some less costly tam without a large impact on recurrence risk: ai tam , tam ai , and upfront ai involved 2, 3, and 5 years of ai respectively). In the exploratory analysis, the sequential ai tam strategy was the least costly and most effective, but only by very small, statistically nonsignificant differences. Per custom, we did not report incremental cost-effectiveness ratios for the dominant interventions. Figure 2 illustrates the costs and outcome trade-offs for all treatment strategies, including the exploratory sequential ai tam strategy.

TABLE III  Base case resultsa

 

 


 

FIGURE 2 Base-case results on the cost-effectiveness plane for the breast cancer model. Includes the exploratory sequential strategy of aromatase inhibitor ( ai ) followed by tamoxifen ( tam ). Inset illustrates the costs and effects of the four strategies in detail. qaly = quality-adjusted life-years.

3.2  Sensitivity Analyses

Table iv presents threshold values from one-way sensitivity analyses, listing the conditions under which a strategy other than sequential tam ai is optimal. The most sensitive parameters in the model were the probabilities of distant recurrence during treatment. If the probability of distance recurrence during 5 years of treatment were to be 3% lower with upfront ai or 4% higher with sequential tam ai , then upfront ai would become the optimal strategy. Upfront ai would also become optimal if the probability of locoregional recurrence during 5 years of treatment were to be 18% lower with upfront ai , or if the probability of stroke or any cardiac event were to be 18% lower with upfront ai . All of the foregoing cases are within the plausible range. The analysis was less sensitive to costing parameters, utilities, mortalities, and probabilities of ae s except for stroke, any cardiac event, and fractures (for the ai and tam ai strategies).

TABLE IV  Threshold valuesa for the conditions under which base case strategies other than tamoxifen ( tam ) followed by aromatase inhibitor ( ai ) are optimal

 

Scenario analyses to address uncertainty in the estimates of carryover benefit were also conducted. In a more conservative scenario (assuming no carryover benefit), if recurrence after 5 years for locoregional and distant metastasis were equal for all therapies, upfront ai monotherapy would become more expensive compared with the other strategies, upfront tam would become the least costly, and compared with upfront tam , all strategies would have very favourable incremental cost-effectiveness ratios. In a scenario in which the risks of recurrence in the first 5 years are maintained indefinitely for years 6 and beyond, a pattern of results similar to those for the base case are found, with slightly larger differences between the strategies. If the risks of recurrence for years 6 and beyond are used only for years 6–10, and no further differences are assumed for the remainder of the model, upfront ai would be more costly than upfront tam , but the benefit of sequential therapies would remain similar to that of upfront ai , and sequential therapies would be less costly than either monotherapy. Finally, if the annual cost of treatment with ai were to be the same as the cost of tam , up-front tam becomes the most expensive strategy, and upfront ai becomes the least expensive.

3.3   Probabilistic Sensitivity Analysis

Multiple cost-effectiveness acceptability curves were created to illustrate the likelihood of each of the strategies being cost-effective at different wtp thresholds (Figure 3). At a threshold of CA$20,000 per qaly gained, the probabilities of being cost-effective were 41% for a sequential ai tam strategy, 33% for a sequential tam ai strategy, and 22% for an upfront ai strategy. Those probabilities remained stable at higher wtp thresholds.

 


 

FIGURE 3 Cost-effectiveness acceptability curve for six treatment strategies in estrogen receptor–positive early breast cancer. Includes the exploratory sequential strategy of aromatase inhibitor ( ai ) followed by tamoxifen ( tam ). qaly = quality-adjusted life-years.

3.4  EVPI

The overall evpi s per patient for the strategies were, respectively, CA$1,877, CA$4,018, and CA$7,717 at wtp thresholds of CA$20,000, $50,000, and $100,000 per qaly gained. The population evpi is lowest at CA$103 million if the wtp threshold is CA$1,000 per qaly gained, and it steadily increases as shown in Figure 4. The least decision uncertainty occurs at very low wtp thresholds; it increases as the wtp increases. The partial evpi for the parameters varied with the wtp threshold (Figure 5). At the wtp threshold of CA$50,000 per qaly gained, additional information about the breast cancer event and the ae probabilities would be valuable. Should the corresponding partial evpi s be quite low, additional research on breast cancer mortality, utilities, and treatment cost would be of little value.

 


 

FIGURE 4 Plot of the expected value of perfect information ( evpi ) for the analyzed population.

 


 

FIGURE 5 Plot of the partial and total expected value of perfect information ( evpi ) per patient, by quality-adjusted life-year ( qaly ) cost threshold.

4.  DISCUSSION

In our base-case model, sequential tam ai was the most effective and least costly of the three main strategies. Upfront ai was more effective and less costly than upfront tam because of a reduction in breast cancer recurrence and differences in ae s. Among the ai -containing strategies, sequential therapy appears to provide benefits similar to those with upfront ai at a lower cost. When the exploratory sequential ai tam strategy was included, it was numerically the least costly and most effective, but only by a very small margin that might not be clinically meaningful. Among the three ai -containing strategies, sequential ai tam was the least costly, involving the fewest years of ai therapy (2 years, compared with 3 and 5 years for tam ai and upfront ai respectively). The sequential ai tam strategy was included for exploratory purposes only. It assumed risks of recurrence equal to those with the upfront ai strategy because only one study ( big 1–98) had considered the strategy, finding statistically nonsignificant numeric differences from upfront ai . The other strategies were based on a meta-analysis of eight trials, and thus in the estimates for the sequential ai tam strategy entail additional uncertainty. Overall, the findings suggest very small differences between the strategies, particularly between the three strategies using ai either upfront or sequentially. The results of the probabilistic sensitivity analyses confirm that finding, in that none of the four strategies (including the exploratory sequential ai tam strategy) is optimal in more than approximately 40% of the simulations across the wtp thresholds commonly accepted in Canada.

Earlier studies used results from the big 1–9848, ma .1743, and atac 49 clinical trials to analyze the cost-effectiveness of upfront ai and tam 11,12,14 from the Canadian health care perspective. Another study13 determined the cost-effectiveness of sequential tam ai compared with upfront ai and tam , and found trends in costs and effects that are similar to those from our analysis. However, no studies have compared all four potential 5-year strategies in a single cost-effectiveness analysis. In addition, we used clinical data to derive breast cancer event probabilities from a meta-analysis of breast cancer outcomes in adjuvant trials of ai and tam that included 18,871 patients5. The results have higher statistical power and are potentially less biased and more generalizable than results from individual studies. Finally, our study was based on longitudinal data, with longer patient follow-up and safety profiles.

Our sensitivity analyses suggest several conditions under which the sequential tam ai strategy would no longer be optimal. If the risks of distant recurrence were to be lowered by 3% with upfront ai or increased by 4% with sequential tam ai , upfront ai would instead be optimal. Similarly, if disease-free survival were to be improved with upfront ai , that strategy would instead be optimal. Overall, the results are highly sensitive to the comparative effectiveness of sequential tam ai and the upfront ai strategies.

Our study has several limitations. First, although the results are potentially helpful for decision-makers in comparing a number of treatment strategies at one time, the use of heterogeneous patient characteristics from different studies in the model poses inherent difficulties, and inconsistencies in specific baseline estimates can arise. For instance, the reported mortality rate after recurrence in the Austrian Breast Cancer Study Group VIII trial (1807 patients) was significantly lower than that in the other six trials5. To keep the model parameters consistent, we excluded that trial from the analysis, and we also used a single consistent mortality rate from distant recurrence. Additionally, ae probabilities used in the model come from a single trial ( big 1–98)3, while the breast cancer event probabilities were derived from the meta-analysis of several studies. Considering that big 1–983 represents about one quarter of the meta-analysis data, uncertainty about side effects might affect the model results. However, any impact is likely to be small: according to big 1–983, no life-threatening ae s occurred in any group of treatment strategies, and the one-way sensitivity analysis indicated that ae probabilities are not the main drivers of the model. Finally, by assuming that the rates of breast cancer events are similar for all ai -containing strategies after 5 years of treatment and that the mortality rate from metastatic breast cancer is similar for all strategies, we could neglect potential differences in benefit and “carryover” effect among the therapies (unavoidable because of a lack of long-term data for upfront ai and both sequential therapies). We explored those assumptions about carryover benefit in scenario analyses and found that different assumptions would not affect the overall conclusions. It can be clinically reasonable to use equal breast cancer mortality rates because no available data suggest a post-progression benefit based on the adjuvant treatment received. In addition, there is currently little evidence to support any differences in overall survival between the treatments. Using the same mortality rates for all strategies (without stratification by therapy) caused differences in breast cancer event mortality between the model results and the meta-analysis (Table v ).

TABLE V  Comparison of modela and meta-analysis results for four treatment strategies—including the exploratory calibrated aromatase inhibitor ( ai ) → tamoxifen ( tam ) sequential strategy—at 5 years

 

To summarize, various adjuvant therapy strategies for breast cancer are currently in use. We examined 5-year strategies, but strategies of extended adjuvant therapy to 8 or 10 years are attracting increasing interest and are included in the latest American Society of Clinical Oncology guidelines10. However, the lack of comparative data for extended adjuvant therapy using either tam or ai after upfront ai or sequencing makes formal comparisons difficult at the present time.

5.  CONCLUSIONS

Our study was prompted by the practical need to identify a cost-effective strategy among the various strategies currently used for hormonal therapy in postmenopausal women with er -positive early breast cancer. We created a model to explore several 5-year hormonal strategies in the Canadian context: 5 years of upfront tam , 5 years of upfront ai , sequential tam (2 years) followed by ai (3 years), and an exploratory strategy with sequential ai (2 years) followed by tam (3 years). The ai -containing strategies were more effective and less costly than upfront tam . Of the three ai -containing strategies, those using sequential therapy were less expensive because of drug costs (that is, sequential therapies used some less costly tam without a large effect on recurrence risk). Those results are sensitive to small changes in the parameters, and no individual strategy is optimal more than approximately 40% of the time. Moreover, the exploratory sequential ai tam strategy has not gained significant clinical support to date, and more evidence is required. Overall, in postmenopausal women with er -positive early breast cancer, benefits accrue to strategies using ai compared with strategies using tam alone. Among the ai -containing strategies, sequential therapy appears to provide benefits similar to those with upfront ai at a lower cost. Future research is required to address uncertainties about long-term breast cancer events and ae probabilities. As more evidence of clinical effectiveness becomes available, our model will address a current gap in the economic evidence.

6. ACKNOWLEDGMENTS

The Pharmacoeconomics Research Unit is funded by Cancer Care Ontario; the Canadian Centre for Applied Research in Cancer Control is funded by the Canadian Cancer Society. The findings do not necessarily represent the views of the funders.

7. CONFLICT OF INTEREST DISCLOSURES

We have read and understood Current Oncology ’s policy on disclosing conflicts of interest, and we declare the following interests: an honorarium was paid to EA’s institution (Princess Margaret Hospital) by AstraZeneca. The remaining authors declare that they have no conflicts of interest.

8. REFERENCES

1. Anderson WF, Chatterjee N, Ershler WB, Brawley OW. Estrogen receptor breast cancer phenotypes in the Surveillance, Epidemiology, and End Results database. Breast Cancer Res Treat 2002;76:27–36.
cross-ref  pubmed  

2. Canadian Cancer Society’s Steering Committee on Cancer Statistics. Canadian Cancer Statistics 2012. Toronto, ON: Canadian Cancer Society; 2012.

3. Mouridsen H, Giobbie–Hurder A, Goldhirsch A, et al. on behalf of the big 1–98 Collaborative Group. Letrozole therapy alone or in sequence with tamoxifen in women with breast cancer. N Engl J Med 2009;361:766–76.
cross-ref  pubmed  pmc  

4. Forbes JF, Cuzick J, Buzdar A, Howell A, Tobias JS, Baum M on behalf of the Arimidex, Tamoxifen, Alone or in Combination (atac) Trialists’ Group. Effect of anastrozole and tamoxifen as adjuvant treatment for early-stage breast cancer: 100-month analysis of the atac trial. Lancet Oncol 2008;9:45–53.
cross-ref  

5. Dowsett M, Cuzick J, Ingle J, et al. Meta-analysis of breast cancer outcomes in adjuvant trials of aromatase inhibitors versus tamoxifen. J Clin Oncol 2010;28:509–18.
cross-ref  

6. Coates AS, Keshaviah A, Thurlimann B, et al. Five years of letrozole compared with tamoxifen as initial adjuvant therapy for postmenopausal women with endocrine-responsive early breast cancer: update of study big 1–98. J Clin Oncol 2007;25:486–92.
cross-ref  pubmed  

7. Jakesz R, Jonat W, Gnant M, et al. on behalf of abcsg and the gabg. Switching of postmenopausal women with endocrine-responsive early breast cancer to anastrozole after 2 years’ adjuvant tamoxifen: combined results of abcsg trial 8 and arno 95 trial. Lancet 2005;366:455–62.
cross-ref  pubmed  

8. Boccardo F, Rubagotti A, Guglielmini P, et al. Switching to anastrozole versus continued tamoxifen treatment of early breast cancer. Updated results of the Italian Tamoxifen Anastrozole (ita) trial. Ann Oncol 2006(suppl 7):vii10–14.

9. Coombes RC, Kilburn LS, Snowdon CF, et al. on behalf of the Intergroup Exemestane Study. Survival and safety of exemestane versus tamoxifen after 2–3 years’ tamoxifen treatment (Intergroup Exemestane Study): a randomised controlled trial. Lancet 2007;369:559–70.
cross-ref  pubmed  

10. Burstein HJ, Temin S, Anderson H, et al. Adjuvant endocrine therapy for women with hormone receptor–positive breast cancer: American Society of Clinical Oncology clinical practice guideline focused update. J Clin Oncol 2014;32:2255–69.
cross-ref  pubmed  

11. Delea TE, El-Ouagari K, Karnon J, Sofrygin O. Cost-effectiveness of letrozole versus tamoxifen as initial adjuvant therapy in postmenopausal women with hormone-receptor positive early breast cancer from a Canadian perspective. Breast Cancer Res Treat 2008;108:375–87.
cross-ref  

12. Rocchi A, Verma S. Anastrozole is cost-effective vs tamoxifen as initial adjuvant therapy in early breast cancer: Canadian perspectives on the atac completed-treatment analysis. Support Care Cancer 2006;14:917–27.
cross-ref  pubmed  

13. Skedgel C, Rayson D, Dewar R, Younis T. Cost-utility of adjuvant hormone therapies for breast cancer in post-menopausal women: sequential tamoxifen–exemestane and upfront anastrozole. Breast Cancer Res Treat 2007;101:325–33.
cross-ref  

14. Skedgel C, Rayson D, Dewar R, Younis T. Cost–utility of adjuvant hormone therapies with aromatase inhibitors in post-menopausal women with breast cancer: upfront anastrozole, sequential tamoxifen–exemestane and extended tamoxifen– letrozole. Breast 2007;16:252–61.
cross-ref  pubmed  

15. Mouridsen HT. Aromatase inhibitors in advanced breast cancer. Semin Oncol 2004;31(suppl 12):3–8.
cross-ref  

16. Lee AY, Levine MN, Baker RI, et al. on behalf of the clot investigators. Low-molecular-weight heparin versus a coumarin for the prevention of recurrent venous thromboembolism in patients with cancer. N Engl J Med 2003;349:146–53.
cross-ref  pubmed  

17. Anderson FA Jr, Wheeler HB, Goldberg RJ, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary embolism. The Worcester dvt Study. Arch Intern Med 1991;151:933–8.
cross-ref  pubmed  

18. Askoxylakis V, Thieke C, Pleger ST, et al. Long-term survival of cancer patients compared to heart failure and stroke: a systematic review. BMC Cancer 2010;10:105.
cross-ref  pubmed  pmc  

19. Parkash R, Wee V, Gardner MJ, et al. The impact of warfarin use on clinical outcomes in atrial fibrillation: a population-based study. Can J Cardiol 2007;23:457–61.
cross-ref  pubmed  pmc  

20. Bentler SE, Liu L, Obrizan M, et al. The aftermath of hip fracture: discharge placement, functional status change, and mortality. Am J Epidemiol 2009;170:1290–9.
cross-ref  pubmed  pmc  

21. Smektala R, Endres HG, Dasch B, et al. The effect of time-to-surgery on outcome in elderly patients with proximal femoral fractures. BMC Musculoskelet Disord 2008;9:171.
cross-ref  pubmed  pmc  

22. Kannegaard PN, van der Mark S, Eiken P, Abrahamsen B. Excess mortality in men compared with women following a hip fracture. National analysis of comedications, comorbidity and survival. Age Ageing 2010;39:203–9.
cross-ref  pubmed  

23. Martinez AA, Weiner S, Podratz K, et al. Improved outcome at 10 years for serous-papillary/clear cell or high-risk endometrial cancer patients treated by adjuvant high-dose whole abdomino-pelvic irradiation. Gynecol Oncol 2003;90:537–46.
cross-ref  pubmed  

24. Susini T, Amunni G, Molino C, et al. Ten-year results of a prospective study on the prognostic role of ploidy in endometrial carcinoma: dna aneuploidy identifies high-risk cases among the so-called “low-risk” patients with well and moderately differentiated tumors. Cancer 2007;109:882–90.
cross-ref  pubmed  

25. Cuzick J, Sestak I, Baum M, et al. on behalf of the atac/latte investigators. Effect of anastrozole and tamoxifen as adjuvant treatment for early-stage breast cancer: 10-year analysis of the atac trial. Lancet Oncol 2010;11:1135–41.
cross-ref  pubmed  

26. Will BP, Berthelot JM, Le Petit C, Tomiak EM, Verma S, Evans WK. Estimates of the lifetime costs of breast cancer treatment in Canada. Eur J Cancer 2000;36:724–35.
cross-ref  pubmed  

27. Ontario Ministry of Health and Long-Term Care (mohltc). Formulary Search [Web resource]. Toronto, ON: mohltc; 2011. [Current version available online at: https://www.healthinfo.moh.gov.on.ca/formulary/; cited February 28, 2012]

28. O’Brien JA, Patrick AR, Caro JJ. Cost of managing complications resulting from type 2 diabetes mellitus in Canada. BMC Health Serv Res 2003;3:7.
cross-ref  

29. Wagner M, Goetghebeur M, Merikle E, Pandya A, Chu P, Taylor DC. Cost-effectiveness of intensive lipid lowering therapy with 80 mg of atorvastatin, versus 10 mg of atorvastatin, for secondary prevention of cardiovascular disease in Canada. Can J Clin Pharmacol 2009;16:e331–45.
pubmed  

30. Goeree R, Blackhouse G, Petrovic R, Salama S. Cost of stroke in Canada: a 1-year prospective study. J Med Econ 2005;8:147–67.
cross-ref  

31. O’Brien B, Levine M, Willan A, et al. Economic evaluation of outpatient treatment with low-molecular-weight heparin for proximal vein thrombosis. Arch Intern Med 1999;159:2298–304.

32. Casciano R, Tarride JE, Breton MC, Stern L, Langer A. A pharmacoeconomic evaluation of the myocardial ischemia reduction with aggressive cholesterol lowering (miracl) study in Canada. Can J Clin Pharmacol 2004;11:e179–90.
pubmed  

33. Dhalla IA, Smith MA, Choudhry NK, Denburg AE. Costs and benefits of free medications after myocardial infarction. Healthc Policy 2009;5:68–86.

34. Leslie WD, Metge CJ, Azimaee M, et al. Direct costs of fractures in Canada and trends 1996–2006: a population-based cost-of-illness analysis. J Bone Miner Res 2011;26:2419–29.
cross-ref  pubmed  

35. Thorne C. Management of arthralgias associated with aromatase inhibitor therapy. Curr Oncol 2007;14(suppl 1):S11–19.
cross-ref  pubmed  pmc  

36. Pinilla J. Cost minimization analysis of high-dose-rate versus low-dose-rate brachytherapy in endometrial cancer. Gynecology Tumor Group. Int J Radiat Oncol Biol Phys 1998;42:87–90.
cross-ref  pubmed  

37. Sullivan PW, Lawrence WF, Ghushchyan V. A national catalog of preference-based scores for chronic conditions in the United States. Med Care 2005;43:736–49.
cross-ref  pubmed  

38. Mittmann N, Trakas K, Risebrough N, Liu BA. Utility scores for chronic conditions in a community-dwelling population. Pharmacoeconomics 1999;15:369–76.
cross-ref  pubmed  

39. Peasgood T, Ward SE, Brazier J. Health-state utility values in breast cancer. Expert Rev Pharmacoecon Outcomes Res 2010;10:553–66.
cross-ref  pubmed  

40. Clarke P, Gray A, Holman R. Estimating utility values for health states of type 2 diabetic patients using the EQ-5D (ukpds 62). Med Decis Making 2002;22:340–9.
cross-ref  pubmed  

41. Peasgood T, Herrmann K, Kanis JA, Brazier JE. An updated systematic review of health state utility values for osteoporosis related conditions. Osteoporos Int 2009;20:853–68.
cross-ref  pubmed  

42. Gold MR, Franks P, McCoy KI, Fryback DG. Toward consistency in cost–utility analyses: using national measures to create condition-specific values. Med Care 1998;36:778–92.
cross-ref  pubmed  

43. Goss PE, Ingle JN, Martino S, et al. A randomized trial of letrozole in postmenopausal women after five years of tamoxifen therapy for early-stage breast cancer. N Engl J Med 2003;349:1793–802.
cross-ref  pubmed  

44. Ellis MJ, Suman VJ, Hoog J, et al. Randomized phase ii neoadjuvant comparison between letrozole, anastrozole, and exemestane for postmenopausal women with estrogen receptor-rich stage 2 to 3 breast cancer: clinical and biomarker outcomes and predictive value of the baseline PAM50-based intrinsic subtype—acosog Z1031. J Clin Oncol 2011;29:2342–9.
cross-ref  pubmed  pmc  

45. Etchells E, McLeod RS, Geerts W, Barton P, Detsky AS. Economic analysis of low-dose heparin vs the low-molecular-weight heparin enoxaparin for prevention of venous thromboembolism after colorectal surgery. Arch Intern Med 1999;159:1221–8.
cross-ref  pubmed  

46. Statistics Canada. Consumer Price Index, health and personal care, by province (Canada) [Web page]. Ottawa, ON: Government of Canada; 2015. [Available online at: http://www.statcan.gc.ca/tables-tableaux/sum-som/l01/cst01/econ161a-eng.htm; cited July 9, 2012]

47. Allred DC, Brown P, Medina D. The origins of estrogen receptor alpha-positive and estrogen receptor alpha-negative human breast cancer. Breast Cancer Res 2004;6:240–5.
cross-ref  pubmed  pmc  

48. Thurlimann B, Keshaviah A, Coates AS, et al. on behalf of the Breast International Group (big) 1–98 Collaborative Group. A comparison of letrozole and tamoxifen in postmenopausal women with early breast cancer. N Engl J Med 2005;353:2747–57. [Erratum in: N Engl J Med 2006;354:2200]
cross-ref  pubmed  

49. Baum M, Buzdar A, Cuzick J, et al. on behalf of the atac (Arimidex, Tamoxifen Alone or in Combination) Trialists’ Group. Anastrozole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early-stage breast cancer: results of the atac (Arimidex, Tamoxifen Alone or in Combination) trial efficacy and safety update analyses. Cancer 2003;98:1802–10.
cross-ref  pubmed  


Correspondence to: Jeffrey S. Hoch, St. Michael’s Hospital, 30 Bond Street, Toronto, Ontario M5B 1W8. E-mail: jeffrey.hoch@utoronto.ca

(Return to Top)



Current Oncology , VOLUME 22 , NUMBER 2 , April 2015








Copyright © 2019 Multimed Inc.
ISSN: 1198-0052 (Print) ISSN: 1718-7729 (Online)