Cost Utility Analysis of First-Line Hormonal Therapy in Advanced Breast Cancer
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Am J Clin Oncol (CCT) 26(3): 289–296, 2003. © 2003 Lippincott Williams & Wilkins, Inc., Philadelphia
Cost Utility Analysis of First-Line Hormonal
Therapy in Advanced Breast Cancer
Comparison of Two Aromatase Inhibitors to Tamoxifen
George Dranitsaris, M.Pharm., Shailendra Verma, M.D., and
Maureen Trudeau, M.D.
Recent randomized clinical trials (RCT) comparing anastrozole For the past 3 decades, tamoxifen has been the stan-
(Arimidex) and letrozole (Femara) to tamoxifen in the first-line dard first-line hormonal agent in patients with breast
treatment of postmenopausal women with advanced hormone- cancer with advanced disease who are positive for estro-
sensitive breast cancer have demonstrated that both agents were gen (ER)/progesterone (PR) receptors.1 There have been
at least as effective as tamoxifen. In addition, one RCT has
several attempts to replace tamoxifen with selective
revealed significant superiority of letrozole to tamoxifen with
regard to tumor response rate and time to progression. Based on aromatase inhibitors as the treatment standard in the
the efficacy and toxicity data, anastrozole or letrozole may first-line setting.2 However, randomized trials comparing
replace tamoxifen. A cost effectiveness analysis was under- tamoxifen to either formestane or fadrozole failed to
taken to determine whether the new agents are economically demonstrate superiority in terms of time to progression
acceptable alternatives to tamoxifen. In the absence of a ran- (TTP) and overall response.3–5
domized three-arm trial, a decision model was developed to Three randomized trials involving the third generation
simulate and compare the most common therapeutic outcomes. selective aromatase inhibitors, anastrozole and letrozole,
The clinical data were obtained from a meta analysis of modern have now been completed. Two large randomized clini-
(i.e., post-1990) randomized trials. Clinical outcomes data from cal trials powered for equivalence comparing anastrozole
the various trials were statistically pooled using a random
to tamoxifen in the first-line hormonal treatment of
effects model to provide point estimates and 95% confidence
intervals. Total hospital resource consumption was collected advanced breast cancer have recently been published.6,7
from the charts of 87 patients with advanced disease who had In the larger European study, 668 postmenopausal pa-
failed tamoxifen therapy. The model suggested a comparable tients with locally advanced or metastatic breast cancer
duration of quality-adjusted progression-free survival between were randomized in a one-to-one ratio to receive tamox-
letrozole and anastrozole, both being superior to tamoxifen ifen 20 mg or anastrozole 1 mg daily in a double-blinded
(179 days vs. 172 days vs. 161 days). Letrozole and anastrozole fashion.6 The trial was designed and powered to demon-
had overall costs of Can$2,883 and $2,847 per patient, respec- strate equivalence with respect to TTP and overall re-
tively, which were marginally higher than tamoxifen at sponse. Patients randomized to the anastrozole treatment
$Can2,258 per patient. When the costs and benefits were had a comparable overall response relative to tamoxifen
combined, the data generated an incremental cost per quality-
(32.9% vs. 32.6%; p ⫽ NS). Similarly, the median TTP
adjusted progression-free year of $12,500 and $19,600 for
letrozole and anastrozole, respectively, relative to tamoxifen.
was similar between the anastrozole and tamoxifen
Letrozole and anastrozole are both economically acceptable group (8.2 vs. 8.3 months; p ⫽ NS). Both agents were
alternatives to tamoxifen in the first-line treatment setting. well tolerated, but thromboembolic events and vaginal
However, when efficacy and cost effectiveness are considered bleeding were reported more frequently in patients
together, letrozole could be preferentially considered. treated with tamoxifen. The smaller U.S. trial that en-
Key Words: Letrozole—Anastrozole—Tamoxifen—Cost rolled 353 patients was also able to demonstrate equiv-
analysis—Breast cancer. alence between anastrozole and tamoxifen.7
A large double-blind superiority study has also been
published by the International Letrozole Breast Cancer
From the Ontario Cancer Institute/Princess Margaret Hospital, Ot-
Group.8 In this trial, 907 patients with advanced breast
tawa Regional Cancer Centre and Toronto Sunnybrook Regional Can- cancer were randomized to receive tamoxifen 20 mg or
cer Centre, Ontario Canada. letrozole 2.5 mg on a daily basis until disease progres-
Supported by Novartis Canada Inc. sion or patient dropout. Patient groups were well bal-
Address correspondence and reprint requests to George Dranitsaris,
M.Pharm., Department of Molecular Biology, Room 9-113, Ontario
anced with respect to age, ER/PR receptor status, and
Cancer Institute/Princess Margaret Hospital, 610 University Avenue, prior exposure to adjuvant tamoxifen. Patients random-
Toronto, Canada, M5G 2M9. Email: gdranit@ca.inter.net ized to letrozole had superior overall tumor response
289290 G. DRANITSARIS ET AL.
FIG. 1. Decision analytic model of first-line hormonal therapy in advanced breast cancer.
(30% vs. 20%; p ⫽ 0.0006) and TTP (9.4 vs. 6.0 months; monal therapy in postmenopausal women, a decision model
p ⫽ 0.001). was developed. The baseline analysis considered postmeno-
Response and progression-free survival are considered pausal women with ER/PR-positive breast cancer who are
meaningful outcomes in patients with advanced breast anthracycline naïve and have not received first-line hormonal
cancer. Given the findings of these three large random- therapy in the advanced setting. The primary clinical outcome
for measuring successful therapy in the current analysis was
ized trials, it appears that anastrozole and letrozole
quality-adjusted progression-free survival benefit. The face and
should now be considered viable alternatives to tamox- content validity of the model was verified by two oncologists
ifen in the first-line setting. However, in many countries, (S.H. and M.T.) involved in the management of patients with
including Canada, anastrozole and letrozole are substan- breast cancer in Ontario.
tially more expensive on an acquisition cost basis than The model began at the decision node (square) where a
tamoxifen. Therefore, the issues that drug formulary choice would have to be made between letrozole, anastrozole,
committees have to consider are twofold: 1) Does anas- and tamoxifen (Fig. 1). Patients would receive letrozole 2.5 mg,
trozole or letrozole provide good economic value relative anastrozole 1 mg, or tamoxifen 20 mg on a daily basis. After a
to tamoxifen when all the clinical and economic factors 3-month treatment period, the probability of response (circle)
are quantified? and 2) Is one agent more cost effective would be assessed for each of the three agents. Patients re-
than another as an alternative to tamoxifen in the first- sponding to first-line hormonal therapy would continue treat-
line hormonal treatment of advanced breast cancer. In ment until disease progression. In contrast, patients with dis-
this study, a cost-utility analysis was conducted to deter- ease progression at the 3-month interval would be offered
mine whether the approved third-generation aromatase standard chemotherapy consisting of a combination of 5-flu-
orouracil, doxorubicin, and cyclophosphamide (FAC). As an
inhibitors are economically attractive alternatives to ta-
alternative to the baseline analysis, for patients who had pre-
moxifen from the perspective of the publicly funded viously received an anthracycline-based regimen in the adju-
Canadian healthcare system. vant setting and had experienced disease progression, single-
agent paclitaxel or docetaxel were considered in the model.
METHODS Following three cycles of chemotherapy, nonresponders
would then be offered palliative therapy while responders
would continue to receive an additional three cycles of FAC
Development of Decision Model (Fig. 1). For patients with non-life-threatening disease, a sec-
To simulate the comparison of the third generation aro- ond hormonal agent may be an option prior to chemotherapy
matase inhibitors to tamoxifen in the setting of first-line hor- (e.g., aromatase inhibitors 3 tamoxifen). Since randomized
Am J Clin Oncol (CCT), Vol. 26, No. 3, 2003COST ANALYSIS OF HORMONE THERAPY FOR BREAST CANCER 291
crossover trials revealed that response rates for these sequences The cost of drugs, personnel, and supplies were obtained
were comparable at approximately 10%,9 second-line crossover from current pharmacy ordering catalogues along with phar-
therapy was not considered because the final results would be macy and nursing workload measurement statistics, Princess
unaffected. Margaret Hospital (PMH), 2001. Laboratory and diagnostic
costs were obtained from the Departments of Biochemistry,
Clinical Data Microbiology and Diagnostic Imaging, PMH. The costs re-
The clinical data required for the model, which consisted of ported in this study are in Canadian dollars ($Can1 ⫽ $US 0.6)
disease response, side effect rates, and progression-free sur- as of April 1, 2003.
vival estimates for each of the three hormonal alternatives,
were obtained from a meta analysis of modern (post-1990) Quality of Life Data
randomized trials. A computer literature search of Medline, In the original study by Dranitsaris et al., anastrozole and
Cancer Lit, and the Cochrane databases was performed from letrozole were evaluated in the second-line setting. Health-
1990 through 2001 for human clinical studies involving letro- related quality of life values were measured in terms of patient
zole, anastrozole, and tamoxifen. Eligibility criteria relative to preferences for alternative treatment outcomes. Quality-ad-
validity of trial design and analysis were used to identify justed progression-free periods were measured as “healthy
potential studies. To be eligible, studies must have been pub- months equivalent” for the time spent in each outcome of the
lished in a peer-reviewed journal; used a randomized design decision model using the Time Trade-Off technique.12 The
with letrozole 2.5 mg daily, anastrozole 1 mg daily, or tamox- scores in months were then converted to utility measures
ifen 20 mg daily in one of the treatment arms; and subjects between 0 and 1, where 0 represented death and 1 was a state
must have been postmenopausal women with either positive or of optimal quality of life. As previously described,11 the utility
unknown ER/PR status and had not received tamoxifen as a assessments were obtained from a random 25 Canadian women
first-line treatment in the advanced setting. A similar set of living in Ontario, Canada.
study inclusion criteria was used to identify randomized trials The health states evaluated consisted of letrozole and anas-
for first-line FAC and taxane chemotherapy after anthracycline trozole disease responses and failures that were then followed
failures. During the literature review, effort was given to avoid by chemotherapy. Health state utilities for tamoxifen were not
the inclusion of duplicate publications. evaluated in the original study. Since the large randomized
Studies were selected on the basis of the inclusion criteria. trials reported that the aromatase inhibitors and tamoxifen were
The following data were abstracted from accepted studies: comparably well tolerated,6 – 8 the assumption made was that
drugs, dose, frequency of administration, definition of re- health state utilities between these two classes of drugs would
sponse, study design (e.g., blinded vs. non-blinded), eligibility be identical.
criteria, study populations, incidence of drug-related toxicity,
number of withdrawals caused by adverse drug reactions, and
all clinical outcomes. Response and adverse effect rates from Cost Utility Analysis
the different studies were combined using a random effects The clinical, economic, and respondent preference data were
model to calculate a point estimate and 95% confidence inter- then combined in a cost-utility analysis comparing the two
vals for each outcome.10 In the case of the reported estimates aromatase inhibitors to tamoxifen. The primary outcome deter-
for median progression-free survival, equal weighting was used mined in the study was the incremental cost per quality-
with studies of similar sample sizes. adjusted progression-free year gained, which was calculated by
dividing the difference in cost relative to tamoxifen (numera-
Estimation of Treatment Costs tor) by the difference in quality-adjusted progression-free sur-
The analytic time period for this investigation was from the vival benefit (denominator). The stability of the baseline results
start of first-line hormonal therapy until disease progression, was then tested by a comprehensive sensitivity analysis. This
and a Canadian healthcare system perspective was taken. In a procedure included substituting paclitaxel and docetaxel in
previous but recent pharmacoeconomic evaluation of anastro- place of FAC chemotherapy, and reanalyzing the data using the
zole and letrozole in the second-line setting, Dranitsaris et al.11 upper and lower 95% confidence limits of response and che-
conducted a retrospective chart review to measure chemother- motherapy costs.
apy treatment costs in patients who failed second-line hormonal
therapy. The chemotherapy protocols evaluated consisted of
FAC, paclitaxel and docetaxel. Since chemotherapy costs RESULTS
would be the same regardless of when treatment was initiated
(i.e., after a first-line or second-line failure to hormonal thera- Meta Analysis of Randomized Trials
py), it was considered acceptable to use these costs in the Seven randomized trials with first-line anastrozole,
current decision analysis. letrozole, or tamoxifen in one of the treatment arms met
Patient data obtained from the charts included demographic the study inclusion criteria. These trials provided a total
information, hormone receptor status, radiation history, and of nine treatment arms for statistical pooling (anastrozole ⫽
previous anticancer therapy. Hospital resource consumption 2, letrozole ⫽ 1, and tamoxifen ⫽ 6). There were six trials
associated with anticancer therapy was then collected. This published in peer-reviewed journals (Table 1).13–28 The six
consisted of costs for hospitalization, outpatient clinic visits, tamoxifen trials had used 20 mg daily. Similarly, there were
antiemetics (e.g., ondansetron), chemotherapy (including prep-
aration and administration), laboratory tests, patient monitor-
seven comparative studies with first-line FAC chemother-
ing, adverse effect management, and all related physicians fees. apy: four with docetaxel and three trials with paclitaxel
All patient costs were captured from the first cycle up to 3 (Table 1).
weeks after the last cycle of chemotherapy. Resource utilization The outcomes of the statistical pooling of clinical data
during the progression-free survival interval included costs for with first-line hormonal therapy generated an overall
physician visits and monitoring. objective response of 29.0% (95% CI: 25% to 33%) for
Am J Clin Oncol (CCT), Vol. 26, No. 3, 2003292 G. DRANITSARIS ET AL.
TABLE 1. Clinical outcomes derived from metaanalysis of published randomized trials
No. No. patients Pooled response rate Med. TTP
Treatment trials evaluated (95% CI) (days) Refs.
6, 7, 30
Anastrozole 2 511 29.0% (25.0–33.0)* 255
8
Letrozole 1 453 30.0% (26.0–35.0) 287
4–8, 13, 14
Tamoxifen 6 1,352 24.2% (18.6–29.7) 214
15–21
FAC 7 907 52.1% (46.7–57.5) 236.5
22–25
Doxetaxel 4 647 37.2% (28.6–45.8) 139
26–28
Paclitaxel 3 73 22.5% (15.5–29.5) 105
FAC, fluorouracil, doxorubicin, cyclophosphamide; med. TTP, time to progression.
* Pooled results of both studies were available in abstract form.31
anastrozole, 30.0% (95% CI: 26% to 35%) for letrozole, Quality of Life Data for the Model
and 24.2% (95% CI: 18.6% to 29.7%) for tamoxifen, Utility values associated with the use of aromatase
respectively (Fig. 2). The associated progression-free inhibitors from public female volunteers are presented in
survival was approximately 255 days for anastrozole, Table 4. In both groups of respondents, utility scores
287 days for letrozole, and 214 days for tamoxifen were lowest under the scenarios where patients failed
(Table 1). These data were subsequently used in the hormonal therapy and subsequent chemotherapy. The
model. For FAC chemotherapy, the pooled response was health state with the highest utility score (i.e., highest
52.1% (95% CI: 46.7% to 57.5%), whereas rates for quality of life) was when a response to hormonal therapy
paclitaxel and docetaxel were 22.5% (95% CI: 15.5% to was achieved. Both groups considered this health state to
29.5%) and 37.2% (95% CI: 28.6% to 45.8%), respec- be equivalent to approximately 0.80 on a quality of life
tively (Table 1). scale between 0 and 1 (Table 4).
With these data and the meta analysis results, the
Economic Data from the Chart Review model was used to estimate the quality-adjusted progres-
In the original study,11 87 patients who received che- sion-free survival for each strategy. With the utilities
motherapy between 1997 to mid-1999 met the inclusion from public volunteers, the aromatase inhibitors pro-
criteria for chart review. Twenty-four patients received vided a modest benefit over tamoxifen, but the differ-
first-line FAC chemotherapy. For second-line taxane ences between letrozole and anastrozole were within 7
administration, 34 received paclitaxel and 29 docetaxel days of each other (Table 5). Therefore, the differences
(Table 2). Most patient’s tumors were ER/PR positive, between the aromatase inhibitors were too small to con-
with a small proportion having an unknown receptor clude than one agent provided more quality-adjusted
status. All patients had disease progression during ta- benefit than another.
moxifen prior to receiving chemotherapy. Patients re-
ceived a median of five cycles of FAC and four cycles of
Cost Utilities Analysis
paclitaxel or docetaxel at standard doses (Table 2). The
The utility values were then combined with the clini-
total cost of FAC was approximately $627 per cycle,
cal and economic data for the cost-utility analysis. The
whereas second-line paclitaxel and docetaxel were
average cost per patient for anastrozole and letrozole was
$1,680 and $2,653 per cycle (Table 3).
$2,847 and $2,883 compared to $2,258 for tamoxifen,
respectively (Table 6). Despite their substantially higher
drug acquisition cost, letrozole and anastrozole had over-
all costs that were marginally higher than tamoxifen
while providing greater quality-adjusted survival bene-
fits. These findings were primarily due to the higher
overall tumor response rate and progression-free survival
benefit with the aromatase inhibitors (Table 1). Hence,
the main economic driver behind this outcome was that
a smaller number of patients treated with letrozole and
anastrozole would require chemotherapy over the time
periods evaluated by the model.
The next exercise was to estimate the average cost per
quality-adjusted progression-free year. This was accom-
plished by dividing the average overall cost of each agent
by the model estimated quality-adjusted progression-free
FIG. 2. Pooled response rate of first-line hormonal survival benefit (Table 6). The results indicated comparable
agents. cost effectiveness ratios between tamoxifen ($5,100) and
Am J Clin Oncol (CCT), Vol. 26, No. 3, 2003COST ANALYSIS OF HORMONE THERAPY FOR BREAST CANCER 293
TABLE 2. Demographic and chemotherapy administration data derived from the chart audit
Paclitaxel Doxetaxel
Parameters FAC (n ⫽ 24) (n ⫽ 34) (n ⫽ 29)
At the start of chemotherapy (range)
Median age 63 (43–75) 51 (31–69) 54 (31–69)
Median ECOG score* 1 (0–3) 1 (1–3) 1 (0–2)
Median number of metastatic sites 2 (1–4) 2 (1–4) 2 (1–5)
Median disease free period (months)† 47 (1–174) 37 (1–164) 31 (0–137)
Receptor status
ER positive 23 (96%) 20 (59%) 18 (62%)
PR positive 19 (79%) 17 (50%) 16 (55%)
Unknown 0 (0%) 5 (15%) 2 (7%)
Radiation history
Radiation to primary site 19 (79%) 25 (74%) 20 (69%)
Radiation to metastatic site(s) 14 (58%) 22 (65%) 13 (45%)
Adjuvant therapy
None 6 (25%) 10 (29%) 9 (31%)
Hormonal 12 (50%) 13 (38%) 13 (45%)
Chemotherapy 6 (25%) 16 (47%) 11 (38%)
Prior treatment of advanced disease
Hormonal 24 (100%) 34 (100%) 29 (100%)
Chemotherapy 0 (0%) 30 (88%) 25 (86%)
Current treatment for metastatic disease
Median dose (mg/m2) 153 (93–215) 101 (73–109)
Median number of cycle (range) 5 (1–10) 4 (1–10) 4 (1–12)
Median number treatment days (range) 248 (19–1520) 84 (21–277) 105 (21–252)
ECOG, Eastern Cooperative Oncology Group; ER, estrogen receptor; FAC, fluorouracil, doxorubicin, cyclophosphamide; PR,
progesterone receptor.
* A measure of patient performance status on a scale from 0 to 4 where 0 represents no symptoms.
†
From initial diagnosis to first metastases.
letrozole ($5,900) with anastrozole ($6,100) being the least Sensitivity Analysis
economically favorable of the three agents. A series of one-way sensitivity analyses were then
The final analysis was to estimate the incremental cost conducted, characterized by substituting the cost of pac-
per quality-adjusted progression-free year gained when litaxel and docetaxel in place of FAC and using the 95%
using letrozole or anastrozole as alternatives to tamox- CI limits for response rates and FAC chemotherapy
ifen. This was accomplished by dividing the difference in costs. All the comparisons were relative to tamoxifen
overall cost relative to tamoxifen by the difference in only since no “head-to-head” comparative trial between
utility values. Letrozole had a slight economic advantage the aromatase inhibitors has been reported to date. Given
over anastrozole ($12,500 vs. $19,600), suggesting better the higher cost of taxane chemotherapy (Table 3), there
value for limited healthcare resources. Notwithstanding, was an improvement in the cost effectiveness of letrozole
both of these estimates are below the $20,000 threshold, and anastrozole, with letrozole providing the greatest
which has been suggested as a marker for “good eco- improvement in economic value (Table 7). In all the
nomic value” for new medical technologies.29 other scenarios evaluated, letrozole remained the pre-
TABLE 3. Cost of chemotherapy in patients who progressed on hormonal treatment
FAC Paclitaxel Docetaxel
Resource item (n ⫽ 127 cycles) (n ⫽ 139 cycles) (n ⫽ 138 cycles)
Overall drug cost* 118 1,075 1,737
Patient monitoring† 112 101 150
Ambulatory day care visit‡ 112 112 41
Side effects management§ 212 304 634
Physician fees 88 91 73
Total cost per cycle (95% CI) $627 (512–742) $1,680 (1,574–1,976) $2,653 (2,363–3,053)
FAC, fluorouracil, doxorubicin, cyclophosphamide.
* Consists of costs for drug acquisition, preparation, administration, and supplies.
†
Consists of cost for laboratory tests and all related radiological examinations.
‡
Cost of “chair time” to receive chemotherapy.
§
Includes costs for standard premedication and antiemetics.
Am J Clin Oncol (CCT), Vol. 26, No. 3, 2003294 G. DRANITSARIS ET AL.
TABLE 4. Health state utilities derived using the time which include letrozole, anastrozole, and exemestane,
trade-off technique are more effective than megestrol acetate in patients
whose disease progressed on tamoxifen.2,30 The success
Utilities from public volunteers of these trials inspired additional studies against tamox-
Health state scenarios* (n ⫽ 25)†
ifen, the current first-line standard in the advanced dis-
Letrozole ease setting.
No response and progression 0.45 (0.37–0.55) Two equivalence trials comparing anastrozole to ta-
during FAC moxifen were conducted, one in North America (n ⫽
No response to letrozole but 0.67 (0.55–0.79) 353) and a large one (n ⫽ 668) in Europe.7,8 Both studies
response to FAC
Response to letrozole 0.80 (0.70–0.92) demonstrated that anastrozole was clinically equivalent
Anastrozole to tamoxifen in terms of overall clinical response and
No response and progression 0.45 (0.37–0.55) TTP. In the smaller North American study, a retrospec-
during FAC tive analysis determined that patients randomized to the
No response to anastrozole 0.67 (0.55–0.79)
but response to FAC
anastrozole group had a longer median TTP than those
Response to anatrozole 0.80 (0.70–0.92) treated with tamoxifen (11.1 vs. 5.6 months). However,
when the data from both of these equivalence trials were
FAC, fluorouracil, doxorubicin, cyclophosphamide. combined, the superiority analysis of the pooled data
* Given the comparable side effect rates reported in random- failed to show significant differences in overall tumor
ized trials, an assumption was made that tamoxifen would have response and TTP.31 In contrast to the equivalence study
the same health state utilities as the aromatase inhibitors.
†
A quality-of-life estimate for a particular health state between
design adopted for anastrozole, a large (n ⫽ 907) double-
0 and 1, with 0 being death and 1 representing a state of perfect blind randomized superiority trial comparing letrozole to
health. These measures were used to weigh each health state by tamoxifen was recently reported. Letrozole was found to
the quality of life experienced by a patient living through that be superior to tamoxifen in overall response, TTP, and
time period. time to treatment failure.8
Anastrozole and letrozole have now been approved in
ferred alternative to tamoxifen, with modest but consis- Canada and the United States as alternatives to tamox-
tent economic advantages over anastrozole (Table 7). ifen in the first-line treatment of hormone-sensitive
breast cancer. All three agents have advantages and
disadvantages in terms of disease response, TTP, toler-
DISCUSSION ability, and cost. To quantify these differences and to
In the past few years, there have been several impor- provide evidence for informed formulary decision mak-
tant advances in hormonal therapy directed at the man- ing, a cost utility analysis was performed. Within the
agement of postmenopausal women with advanced analytic time frame evaluated, the findings revealed that
breast cancer. The new selective aromatase inhibitors, both anastrozole and letrozole provide additional quality
adjusted progression-free survival relative to tamoxifen,
and this benefit is available at an acceptable cost to the
TABLE 5. Quality adjusted progression-free survival publicly funded Canadian healthcare system.
benefit The analysis also revealed modest but consistent eco-
nomic advantages with letrozole over anastrozole, re-
Public volunteers gardless of the type of chemotherapy offered at disease
Strategy (n ⫽ 25) (days) progression and with variations in tumor response rates
and cost. The primary driver behind this finding was the
Anastrozole 3 chemotherapy* 172
Letrozole 3 chemotherapy* 179 slightly longer TTP with letrozole relative to tamoxifen,
Tamoxifen 3 chemotherapy* 161 which translated to fewer patients requiring chemother-
apy for progressive disease. The economic benefits be-
* FAC, fluorouracil, doxorubicin, cyclophosphamide. came more evident if single-agent chemotherapy with
TABLE 6. Baseline results of second-line hormonal therapy
Outcome Tamoxifen Letrozole Anastrozole
Average cost/patient $2,258 $2,883 $2,847
Quality-adjusted progression-free benefit (years)* 0.44 0.49 0.47
Average cost-effectiveness ratio† $5,100 $5,900 $6,100
Cost per quality-adjusted progression-free year gained‡ — $12,500 $19,600
All estimates were rounded to the nearest hundred.
* Converted from days in Table 5 to years.
†
Average cost divided by the quality-adjusted progression-free benefit.
‡
Difference in cost relative to tamoxifen divided by difference in quality adjusted progression-free benefit.
Am J Clin Oncol (CCT), Vol. 26, No. 3, 2003COST ANALYSIS OF HORMONE THERAPY FOR BREAST CANCER 295
TABLE 7. One-way sensitivity analysis of primary findings
Sensitivity scenario Letrozole Anastrozole
Base case* $12,500 $19,600
Failure followed by paclitaxel chemotherapy $7,200 $10,800
Failure followed by docetaxel chemotherapy $2,600 $5,100
95% CI of response letrozole (26.0%–35.0%) $17,400–$8,900 —
95% CI of response anastrozole (25.0%–33.0%) — $33,400–$12,800
95% CI cost/cycle FAC ($512–$742) $13,100–$11,900 $20,500–$18,800
95% CI cost/cycle paclitaxel ($1,574–$1,976) $7,600–$6,200 $11,200–$9,400
95% CI cost/cycle docetaxel ($2,363–$3,035) $3,800–$1,100 $6,500–$3,200
FAC, fluorouracil, doxorubicin, cyclophosphamide.
All estimates were rounded to the nearest hundred.
* Outcome presented as the incremental cost per quality-adjusted progression-free year secondary to using the aromatase inhibitors
in place of tamoxifen.
expensive drugs such as paclitaxel or docetaxel were the findings of the current study, letrozole would be the
offered to nonresponders. Costly chemotherapy would be preferred choice.
delayed in a slightly higher number of patients treated (ref. 10 not cited]
with letrozole than with anastrozole.
Given the findings of the current study and the ran-
domized superiority trial, letrozole would be considered REFERENCES
the hormonal treatment of choice in postmenopausal 1. Forbes JF. The control of breast cancer: the role of tamoxifen.
women with advanced breast cancer. These conclusions Semin Oncol 1997;24(suppl 1):5–19.
may have important implications with respect to health- 2. Harvey HA. Emerging role of aromatase inhibitors in the treatment
care policy in the management of patients. Both of the of breast cancer. Oncology 1998;12(suppl 5):32–5.
aromatase inhibitors provide additional quality-adjusted 3. Perez Carrion R, Alberola CV, Calabresi F, et al. Comparison of
the selective aromatase inhibitor formestane with tamoxifen as first
progression-free survival benefits relative to tamoxifen. line hormonal therapy in postmenopausal women with advanced
In Canada, anastrozole and letrozole have the same drug breast cancer. Ann Oncol 1996;5(suppl 7):S19 –24.
acquisition cost, both eightfold higher than tamoxifen. 4. Falkson CI, Falkson HC. A randomized study of CGS 16949A
However, as was suggested by the cost-utility analysis, (fadrozole) versus tamoxifen in previously untreated patients with
metastatic breast cancer. Ann Oncol 1996;7:465–9.
letrozole as a formulary alternative to tamoxifen would 5. Thurliman B, Beretta K, Bacchi M, et al. First-line fadrozole
result in a modest increase in overall treatment costs for versus tamoxifen in postmenopausal women with advanced breast
the associated progression-free survival benefit. A simi- cancer: prospective randomized trial of the Swiss Group for Clin-
lar conclusion can also be drawn with anastrozole. ical Cancer Res SAKK 20/88. Ann Oncol 1996;5:471–9.
There are a number of limitations in the current anal- 6. Bonnetere J, Thurlimann JFR, Robertson M, et al. Anastrozole
versus tamoxifen as first-line therapy for advanced breast cancer in
ysis that have to be addressed. Even though the clinical 668 postmenopausal women: results of the tamoxifen or Arimidex
data for the decision model was abstracted from random- randomized group efficacy study. J Clin Oncol 2000:18:3748 –57.
ized trials, none of the studies directly compared anastro- 7. Nabholtz JM, Buzdar A, Pollak W, et al. Anastrozole is superior to
zole to letrozole. Therefore, the comparisons should only be tamoxifen as first-line therapy for advanced breast cancer in
postmenopausal women: results of a North American multicenter
made against tamoxifen. Another drawback to the decision randomized trial. J Clin Oncol 2000;18:3758 – 67.
model was the failure to consider second-line hormonal 8. Mouridsen H, Gershanovich, Sun Y, et al. Superior efficacy of
therapy for nonresponding patients without life-threatening letrozole versus tamoxifen as first line therapy for postmenopausal
metastases. However, limited crossover data suggest that women with advanced breast cancer: results of a phase III study of
response rates between second-line tamoxifen and the aro- the International Letrozole Breast Cancer Group. J Clin Oncol
2001;19:2596 – 06.
matase inhibitors are comparable.9 Hence, the inclusion of 9. Thurlimann B, Robertson JFR, Bonneterre J, et al. Efficacy of
a second-line hormonal sequence in the model would not tamoxifen following anastrozole as first-line treatment for ad-
have altered the findings of the analysis. The final limitation vanced breast cancer in postmenopausal women. Breast Cancer
that should be acknowledged was that the cost of managing Res Treat 2000;64(abstract 162):51.
10. DerSimonian R, Laird N. Meta-analysis in clinical trials. Con-
drug-related side effects such as thromboembolic events trolled Clin Trials 1986;7:177– 88.
was not included in the analysis. This was a bias against the 11. Dranitsaris G, Leung PP, Mather J, et al. Cost utility analysis of
aromatase inhibitors, since the incidence of thromboem- second line hormonal therapy in advanced breast cancer: a com-
bolic events was higher with tamoxifen.6 – 8 parison of two aromatase inhibitors to megestrol acetate. Anti-
In conclusion, the cost-utility analysis revealed that Cancer Drugs 2000;11:591– 601.
12. Torrance GW. Utility approach to measuring health-related quality
letrozole and anastrozole are both economically accept- of life. J Chron Dis 1987;40:593– 600.
able alternatives to tamoxifen in the first-line treatment 13. Hayes DF, Van Zyl JA, Hacking A, et al. Randomized comparison
setting. However, given the available clinical data and of tamoxifen and two separate doses of toremifene in postmeno-
Am J Clin Oncol (CCT), Vol. 26, No. 3, 2003296 G. DRANITSARIS ET AL.
pausal patients with metastatic breast cancer. J Clin Oncol 1995; with sequential methotrexate and 5-FU in patients with advanced
13:2556 – 66. breast cancer after anthracycline failure: a randomized phase III
14. Stuart NSA, Warwick J, Blackledge GRP, et al. A randomized study with crossover on progression by the Scandinavian Breast
phase III cross-over study of tamoxifen versus megestrol acetate in Group. Eur J Cancer 1999;35:1194 –201.
advanced and recurrent breast cancer. Eur J Cancer 1996;32A: 24. Bonneterre J, Roche H, Monnier A, et al. Taxotere versus 5-FU ⫹
1888 –92. Navelbine as second line chemotherapy in patients with metastatic
15. Alonso MC, Tabernero JM, Ojeda B, et al. A phase III randomized breast cancer [Abstract 564]. Proc Am Soc Clin Oncol 1997;16:
trial of CNF versus CAF in patients with metastatic breast cancer. 162a.
Breast Cancer Treat Res 1995;34:15–24. 25. O’Shaughnessy J. Results of a large phase III trial of Xeloda/
16. Bennett JM, Muss HB, Doroshow JH, et al. A randomized multi- Taxotere combination therapy vs. Taxotere monotherapy in meta-
center trial comparing CMF with CAF in the therapy of metastatic static breast cancer patients [Abstract 62]. Breast Cancer Treat Res
breast carcinoma. J Clin Oncol 1988;6:1611–20. 2000;64:155.
17. French epirubicin study group. A prospective randomized phase III 26. Nabholtz JM, Gelman K, Bontenbal M, et al. Multicenter, random-
trial comparing combination chemotherapy with cyclophospha- ized comparative study of two doses of paclitaxel in patients with
mide, fluorouracil, and either doxorubicin or epirubicin. J Clin metastatic breast cancer. J Clin Oncol 1996;14:1858 – 67.
Oncol 1988;6:679 – 88.
27. Dieras V, Marty M, Tubiana N, et al. Phase II randomized study of
18. Italian Multicenter Breast Study with Epirubicin. Phase III ran-
paclitaxel versus mitomycin in advanced breast cancer. Semin
domized study of fluorouracil, epirubicin, and cyclophosphamide
Oncol 1995;22:33–9.
in advanced breast cancer: an Italian multicentre trial. J Clin Oncol
1988;6:976 – 82. 28. O’Reilly SM, Moiseyenko V, Talbot DC, et al. A randomized
19. Lopez M, Papaldo P, Di Lauro L, et al. FAC vs. FEC in metastatic phase II study of Xeloda vs. paclitaxel in breast cancer patients
breast cancer. Oncology 1989;46:1–5. failing previous anthracycline therapy [Abstract 627]. Proc Am Soc
20. Smalley RV, Lefante J, Bartolucci A, et al. A comparison of CAF Clin Oncol 1998;17:163a.
and CMFVP in patients with advanced breast cancer. Breast 29. Laupacis A, Feeny D, Detsky AS, et al. How attractive does a new
Cancer Treat Res 1983;3:209 –20. technology have to be to warrant adoption and utilization? Can
21. Stewart DJ, Evans WK, Shepherd FA, et al. Cyclophosphamide Med Assoc J 1992;146:473– 81.
and fluorouracil combined with mitoxantrone versus doxorubicin 30. Kaufmann M, Bajetta E, Dirix LY, et al. Exemestane is superior to
for breast cancer: superiority of doxorubicin. J Clin Oncol 1997; megestrol acetate after tamoxifen failure in postmenopausal
15:1897–905. women with advanced breast cancer. J Clin Oncol 2000;18:1399 –
22. Nabholtz JM, Senn HJ, Bezwada WR, et al. Prospective random- 411.
ized trial of docetaxel versus mitomycin plus vinblastine in patients 31. Robertson J, Buzdar A, Nabholtz J, et al. Anastrozole (Arimi-
with metastatic breast cancer progressing despite previous anthra- dexTM) versus tamoxifen as first-line therapy for advanced breast
cycline-containing chemotherapy. J Clin Oncol 1999;17:1413–24. cancer in post-menopausal women—prospective combined analy-
23. Sjostrom J, Blomqvist C, Mouridsen H, et al. Docetaxel compared sis from two international trials. Eur J Cancer 2000;36:S88.
Am J Clin Oncol (CCT), Vol. 26, No. 3, 2003You can also read
