These slides were released by the speaker for internal use by Novartis

1. These slides were released by the speaker for internal use by Novartis

2. Risk/benefit analysis: tamoxifen vs aromatase inhibitors • Edith A Perez • (Mayo Clinic, Jacksonville, FL; Mayo Foundation, Rochester, MN, USA)

3. Tamoxifen Hot flushes Genitourinary symptoms Endometrial cancer Thromboembolism Favorable effects on lipids? Reduction in baseline CVD? Protective effect on bone AIs Hot flushes Arthralgia/arthritis Myalgia Bone loss/fractures vs tamoxifen Increase in CVD & hypercholesterolemia? Fewer gynecological symptoms Fewer endometrial cancers Fewer thromboembolic and DVT events (including grade 3–5) Adverse event profiles

4. Hot flushes reported in adjuvant AI trials • Significant reduction vs tamoxifen with upfront AI (anastrozole, letrozole) • No significant reduction vs tamoxifen with sequential tamoxifen-AI therapy (exemestane, anastrozole) • Significant increase vs placebo with extended adjuvant letrozole ATAC Trialists’ Group Lancet 2005;365:60–2; Thürlimann et al.N Engl J Med 2005;353:2747–57; Coombes et al. N Engl J Med 2004;350:1081–92; Jakesz et al.Lancet 2005;366:455–62; Goss et al. J Natl Cancer Inst 2005;97:1262–71

5. Gynecological symptoms/endometrial cancer reported in adjuvant AI trials • Compared with tamoxifen • Significant reduction in vaginal bleeding, vaginal discharge, endometrial biopsies • Non-significant reduction in endometrial cancer • Compared with placebo • Significant reduction in vaginal bleeding ATAC Trialists’ Group Lancet 2005;365:60–2; Thürlimann et al.N Engl J Med 2005;353:2747–57; Coombes et al. N Engl J Med 2004;350:1081–92; Jakesz et al.Lancet 2005;366:455–62; Goss et al. J Natl Cancer Inst 2005;97:1262–71

6. Thromboembolic (TE) events reported in adjuvant AI trials • Compared with tamoxifen • Significant reduction in TE events including venous TE, deep venous TE and SAEs • Compared with placebo • No difference in incidences of TE events or stroke/transient ischemic attack ATAC Trialists’ Group Lancet 2005;365:60–2; Thürlimann et al.N Engl J Med 2005;353:2747–57; Coombes et al. N Engl J Med 2004;350:1081–92; Jakesz et al.Lancet 2005;366:455–62; Goss et al. J Natl Cancer Inst 2005;97:1262–71

7. Arthralgia reported in adjuvant AI trials • In upfront and switching trials • All AIs associated with significant increase in arthralgia vs tamoxifen • AIs 5.4–36% vs tamoxifen 3.6–29% • In extended adjuvant setting • Significant increase in arthralgia vs placebo ATAC Trialists’ Group Lancet 2005;365:60–2; Thürlimann et al.N Engl J Med 2005;353:2747–57; Coombes et al. N Engl J Med 2004;350:1081–92; Jakesz et al.Lancet 2005;366:455–62; Goss et al. J Natl Cancer Inst 2005;97:1262–71

8. CV disease reported in adjuvant AI trials Within the limitations of trial methodology and cross-trial comparisons, data suggest • Non-significant increase in CV events: AI vs tamoxifen • ATAC • BIG 1-98 • IES • ABCSG-8/ARNO • No difference in CV events: AI vs placebo • MA.17 core analysis • post-unblinding analysis Letrozole is not licensed in all European countries for use in the early adjuvant setting ATAC Trialists’ Group Lancet 2005;365:60–2; Thürlimann et al.N Engl J Med 2005;353:2747–57; Coombes et al. N Engl J Med 2004;350:1081–92; Jakesz et al.Lancet 2005;366:455–62; Goss et al. J Natl Cancer Inst 2005;97:1262–71

9. Cardioprotective effect of tamoxifen • EBCTCG experience at 15 years1 • Subset of ~15,000 women treated with 5 years of tamoxifen vs control • 189 vs 169 vascular deaths (tamoxifen vs control, NS) • 32 trials comparing tamoxifen with control group: metastatic, adjuvant, and prevention settings2 • > 52,000 patients • Relative risk ratio for fatal MIs, tamoxifen vs control: 0.62 (95% CI: 0.41–0.93) 1. EBCTCG Lancet 2005;365:1687–717; 2. Braithwaite et al. J Gen Intern Med 2003;18:937–47

10. Hypercholesterolemia reported in adjuvant AI trials Data from adjuvant AI trials suggest • Non-significant increase in hypercholesterolemia: AI vs tamoxifen • ATAC: 9% vs 3.5% (p = NR) • BIG 1-98: All grades: 43.5% vs 19.1% (p = NR) Grade 1: 35.1% vs 17.3% (p = NR) • ITA: 9.3% vs 4.0% (p = 0.04) • IES: NR • ABCSG-8/ARNO: NR • No difference in hypercholesterolemia: AI vs placebo (fasting) • MA.17: 16% vs 16% (p = 0.79) NR: not reported Letrozole is not licensed in all European countries for use in the early adjuvant setting Arimidex PI www.arimidex.com Sept 2005; Thürlimann et al.N Engl J Med 2005;353:2747–57; Coombes et al. N Engl J Med 2004;350:1081–92; Jakesz et al.Lancet 2005;366:455–62; Goss et al. J Natl Cancer Inst 2005;97:1262–71

11. Differences in data capture In BIG 1-98 cholesterol data were collected every 6 months Simpler data collection in the ATAC trial – non-specific request to report AE

12. BIG 1-98 targeted AE: cholesterol • All grades: 43% vs 19% • Measurements: 90% non-fasting, variable time of day • ≥ grade 1 hypercholesterolemia: any cholesterol measurement above ULN at 6-month visit • No central laboratory measurements

13. 15 Letrozole Tamoxifen 10 Total cholesterol (mmol/L*) 5 0 18 60 0 6 12 42 48 54 30 36 24 Months Let 3209 Tam 3226 2606 2599 2660 2671 2546 2588 2264 2255 1716 1679 1253 1226 863 831 698 695 527 519 6 25 Effect of letrozole vs tamoxifen on total serum cholesterol† Cholesterol values remained stable in the letrozole arm and decreased in the tamoxifen arm by ~12% *To convert mmol/L to mg/dL, multiply by 39 Perez E. Personal communication

14. Lipid lowering effect of tamoxifen? • Yes • Tamoxifen studies • 10 trials, 6 vs placebo; 657 patients • Decrease in cholesterol seen in all studies • Median decrease: 12.5% (range 3–17) • Decrease due to LDL cholesterol Herrington & Klein Womens Health Issues 2001;11:95102

15. Effect of AI on lipid levels vs placebo • ATENA1 and MA.17 lipid substudy2 • Increase in serum cholesterol observed 6 months after tamoxifen discontinuation • No relevant difference between two treatment groups for any lipid parameters • LEAP study3 (early results) • Anastrozole (n = 29), exemestane (n = 32) and letrozole (n = 29) have similar, non-detrimental effects on serum lipids in postmenopausal women • Conclusion • AIs lack the lipid-lowering effects of tamoxifen 1. Markopoulos et al. Anticancer Drugs 2005;16:879–83; 2. Wasan et al. Ann Oncol 2005;16:707–15 3. McCloskey et al. Breast Cancer Res Treat 2005;94:(abstract 2052)

16. Osteoporosis/fractures reported in adjuvant AI trials • Compared with tamoxifen • Upfront trials: significant increase in osteoporosis and/or fractures • Switching trials: significant increase in osteoporosisSignificant increase or trend for increase in fractures • Compared with placebo • Significant increase in new patient-reported osteoporosis • No significant increase in fracture rate ATAC Trialists’ Group Lancet 2005;365:60–2; Thürlimann et al.N Engl J Med 2005;353:2747–57;Coombes et al.N Engl J Med 2004;350:1081–92; Jakesz et al.Lancet 2005;366:455–62; Boccardo et al. J Clin Oncol 2005;23:5138–47;Goss et al. J Natl Cancer Inst 2005;97:1262–71; Mincey & Perez ASCO Edu Book 2005;27–34

17. Management of AEs Bone loss, osteoporosis and fractures • Tamoxifen protects against bone loss in postmenopausal women • HRT – contraindicated in patients with BC • Routine monitoring, calcium/vitamin D supplements • Bisphosphonates prevent bone loss • Efficacy differs between agents • Early generation oral bisphosphonates often poorly tolerated • CTIBL more rapid than ‘natural’ postmenopausal bone loss – may need more potent IV agents CTIBL, cancer treatment-induced bone loss HRT, hormone replacement therapy

18. Letrozole + upfront ZOL Letrozole + delayed ZOL Overall difference p value Mean D in BMD(total hip) +1.40% -2.10% 3.50% < 0.001 Z-FAST, ZO-FAST and E-ZO-FAST RANDOMIZE Upfront zoledronic acid 4 mg q 6 months + Letrozole (2.5 mg/day) x 5 years Delayed* zoledronic acid 4 mg q 6 months + Letrozole (2.5 mg/day) x 5 years BMD at 12 months1 Accrual completed: ZO-FAST, n = 1066; Z-FAST, n = 602; E-ZO-FAST: n = 500 *Initiation of zoledronic acid determined by post-baseline BMD 1. Brufsky et al. J Clin Oncol 2005;23:12s(abstract 533)

19. NCCTG N03CC Upfront or delayed bisphosphonate for women receiving letrozole after tamoxifen Inclusion criteria • Postmenopausal • Prior tamoxifen • T-score ≥ -2.0 SD (n = 550) RANDOMIZE Upfront zoledronic acid 4 mg q 6 months Delayed* zoledronic acid 4 mg q 6 months Letrozole x 5 years Calcium, vitamin D (all patients) *Initiation of zoledronic acid determined by post-baseline BMD T-score < -2.0 SD 10/05: enrollment complete; PI: BA Mincey

20. ASCO bone health guidelines • BMD screening in breast cancer • All women aged > 65 years • All women aged 60–64 years with risk factors • Postmenopausal women of any age receiving AIs • Premenopausal women with medical/surgical premature menopause • Repeat BMD annually after initial exam • Treatment • T-Score > -1.5: lifestyle, reassurance • T-Score -1.5 to -2.5: calcium, vitamin D • T-Score <-2.5: bisphosphonates Hillner et al.J Clin Oncol 2003;21:4042–57

21. Effect of AIs on QoL • Censoring at recurrence may underestimate QoL advantage of AIs • No significant adverse effects of AIs on QoL reported in adjuvant trials to date • AI vs tamoxifen: anastrozole (ATAC), exemestane (IES) • AI vs placebo: letrozole (MA.17) Fallowfield et al.J Clin Oncol 2004;22:4261–71; Fallowfield et al.J Clin Oncol 2006;24:910–17; Whelan et al.J Clin Oncol 2005;23:6931–40

22. Adverse event profilesConclusions • Tamoxifen associated with thromboembolic and gynecological events • AIs generally well tolerated but associated with some bone loss and arthralgia • Tamoxifen protects against postmenopausal bone loss • Routine monitoring, calcium/vitamin D supplements can minimize fracture risk in most patients • Bisphosphonates can prevent AI-associated bone loss • Further investigation of effects on cardiovascular system and lipid metabolism • Beneficial effects of tamoxifen • No increase in cardiovascular disease or hypercholesterolemia, letrozole vs placebo (MA.17) • AEs associated with AIs generally more preventable or manageable than AEs associated with tamoxifen