1. Postmenopausal Hormone Therapy And The Risk of Breast Cancer A Contrary Thought Leon Speroff, M.D.

2. The Cover of The Lancet July 9-15, 2005 “If everything has to be double-blinded, randomised, and evidence-based, where does that leave new ideas?” Speroff

3.                                                                                 Most Important Unanswered Question Postmenopausal Hormone Therapy and the Risk of Breast Cancer: Do hormones initiate new tumor growth or promote the growth of pre-existing tumors? Speroff

4. WHI: E/P Updated Breast Cancer Report E/P Placebo Ratio Invasive breast ca Year 1 12 19 cases 0.62 (0.29-1.23) Year 2 26 32 0.77 (0.46-1.30) Year 3 29 22 1.26 (0.73-2.20) Year 4 44 27 1.54 (0.95-2.49) Year 5 43 21 1.99 (1.18-3.35) Year 6 + 45 29 1.35 (0.85-2.16) Noninvasive 47 37 (NS) Deaths 4 4 JAMA 2003;289:3243

5. Reanalysis of World’s Breast Cancer Data >>>>>>>>>>>>>>>>>>>>>>>>>>>>> Findings: Current users 5+ years: RR = 1.35 (1.21-1.49) No effect of family history Lancet 350:1047, 1997 Speroff

6. Reanalysis of World’s Breast Cancer Data Lancet 350:1047, 1997 Current and recent users had no metastatic disease. Decreased risk of fatal breast cancer in users. Speroff

7. AN APPARENT PARADOX The observational studies that find: At the same time, indicate: Increased risk Decreased mortality Speroff

8. BETTER PROGNOSIS FOR ESTROGEN USERS BETTER PROGNOSIS FOR ESTROGEN USERS Detection/surveillance Bias: Hormone users have more mammograms. Different biology, corrected for mammography: Fewer large tumors, More grade 1 tumors. Bonnier, et al, Obstet Gynecol 85:11, 1995 Manjer, et al, Int J Cancer 92:919, 2001 Gertig, et al, Br Ca Res Treat 80:267, 2003 Pappo, et al, Ann Surg Oncol 11:52, 2004 Speroff

9. An Answer to the Apparent Paradox Detection/surveillance bias = Earlier diagnosis Hormonal effects on a pre-existing tumor = Less aggressive stage PLUS Speroff

10. Review of Oregon Experience Long-term hormone users had: More tumors detected by mammography More ductal ca-in-situ More stage I, node negative tumors Better survival rates (100% after 12 yrs) in tumors detected by mammography) No differences in histology or ER status Long-term hormone users had: More tumors detected by mammography More ductal ca-in-situ More stage I, node negative tumors Better survival rates (100% after 12 yrs) in tumors detected by mammography) No differences in histology or ER status Am J Surg 2008;196:505

11. The Hormonal Effect On Pre-Existing Tumors Differentiation of tumor cells (or inhibition of de-differentiation) allowing time for the stromal reaction that leads to earlier detection. Speroff

13. Causation or Early Detection Similar results with: hormone therapy, oral contraceptives, and pregnancy. Observations that favor early detection: Increase observed very fast. ER+ lower grade and stage disease. Return to baseline after therapy. Better survival rates. Speroff

14. Ontogeny of Breast Cancer Cancer Starts Here Stem Cells Transition Ductal Cells Lobular Cells Hormone Effects

15. Speroff WHI: Updated Breast Cancer Report E/P Placebo Ratio Invasive breast ca Year 1 12 19 cases 0.62 (0.29-1.23) Year 2 26 32 0.77 (0.46-1.30) Year 3 29 22 1.26 (0.73-2.20) Year 4 44 27 1.54 (0.95-2.49) Year 5 43 21 1.99 (1.18-3.35) Year 6 + 45 29 1.35 (0.85-2.16) Noninvasive 47 37 (NS) After adjustments 1.20 (0.94-1.53) !! JAMA 2003;289:3243 2010;304:1684

16. Speroff WHI: Updated Breast Cancer Report Problem with tumor size and localized disease: Tumor size of 1.5-1.8 25-28% positive nodes in literature and SEER 15.8% in WHI placebo group WHI: No nodes examined-9.9/9.1%; missing info-4.0/4.7% Tumors less than 1 cm with no node information were classified as localized disease!!

17. Maturitas 2006;55:103 0 1 2 3 4 5 6 0.8 0.6 0.4 0.2 0 DETECTION (%) E+P PLACEBO

18. SEER Data, 1983-1987 OHSU HRT WHI HRT

19. SEER Data, 1983-1987 OHSU No HRT WHI No HRT

20. Speroff WHI: Updated Breast Cancer Report E/P Placebo Ratio Invasive breast ca Year 1 12 19 cases 0.62 (0.29-1.23) Year 2 26 32 0.77 (0.46-1.30) Year 3 29 22 1.26 (0.73-2.20) Year 4 44 27 1.54 (0.95-2.49) Year 5 43 21 1.99 (1.18-3.35) Year 6 + 45 29 1.35 (0.85-2.16) After adjustments 1.20 (0.94-1.53) !! Risk decreased with time!! JAMA 2003;289:3243 2010:304:1684 Maturitas 55:103, 2006

21. Speroff WHI Comparison: Trial & Observ. Data “Both yield same conclusions when adjusted for time from menopause to treatment.” Problem: Trial- more BSO, parity differences, less mammography, less prior use, fewer risk factors, older, heavier. THE TWO POPULATIONS DIFFER IN RISK PROFILE!! Am J Epidemiol 2008;167:1207 JNCI 2013;105:526 \

22. Speroff WHI: Updated E-Only Breast Cancer Report Overall: HR=0.80; CI=0.62-1.04 Adherent Pts: HR=0.67 CI=0.47-0.97 No effect on in-situ disease. Only ductal cancers and in women with no prior hormone therapy. More follow-up mammograms/biopsies/aspirations. JAMA 2006;295:1647

23. Speroff WHI: Differences Between E-P and E Arms 1.Cardiovascular E-only: more obese, more hypertension & diabetes, less activity. 2.Breast Cancer E-only: – more early and less late births. – 21% more previous and 17% more with longer duration of hormone use. TWO DIFFERENT POPULATIONS! Ann Epidemiol 2003;13:S78

24. Int J Cancer 2005;114:448 Breast Cancer Res Treat 2008;107:103 Int J Cancer 2011;128:144 French E3N Cohort Study 133,744 women; 8.6 years follow-up 55% gels; 45% patches E alone RR = 1.29 (1.02-1.65) E/Progesterone RR = 1.00 (0.83-1.22) E/Progestins RR = 1.77 (1.40-2.24) SPEROFF

25. Int J Cancer 2005;114:448 Breast Cancer Res Treat 2008:107:103 French E3N Cohort Study Nonoral E/Progestins <2yrs: 1.37 (1.07-1.72) <1yr: 1.7 (1.3-2.3) Problems: Users & nonusers not comparable Very fast detection! ? Bioequivalent doses ? E/Progestins: More potent differentiation SPEROFF

26. Cancer 101:1490, 2004 Nurses Health Study: Risk of Invasive Breast Cancer ER+/PR+ <5 yrs 5+ yrs E alone 46 1.02 (0.77-1.38 73 1.37 (1.06-1.78) E/P 112 1.74 (1.40-2.17) 99 2.05 (1.64-2.57) E/P users: younger, lower stage & grade, increase only in ER+/PR+ & greater in lean women. Speroff

27. Is E/P Better? Very large prospective study, 374,465 screened women in 6 U.S. mammography registries: <5 yrs 5+ yrs E alone 0.86 (0.71-1.03) 0.92 (0.84-1.00) E/P 0.85 (0.73-0.98) 1.49 (1.36-1.63) After E/P for 5+ yrs: lower grade & stage, more ER+ J Clin Oncol 21:431, 2003 Speroff

28. 1,081 E only; 1,399 E-P; 4,956 nonusers: Breast Ca Case All Causes Mortality Stage I: E only1.04 (0.77-1.42)1.23 (0.72-2.10) E-P0.69 (0.48-0.99)0.52 (0.26-1.04) Stage II: E only0.86 (0.65-1.14)1.01 (0.72-1.41) E-P0.53 (0.39-0.73)0.69 (0.48-0.98) Br J Cancer 93:392, 2005

29. Breast Cancer Mortality Cancer Epidem Biomark Prev 17:864, 2008 Collaborative Breast Cancer Study Cohort: 12,269 women in Wisc., Mass., NH; followed 1980 to 2006 Tx at DxAdj. Rate Ratio Former E 0.86 (0.71-1.05) Current E 0.91 (0.77-1.09) Former E-P 0.96 (0.62-1.50) Currrent E-P 0.69 (0.55-0.88) E-P for 5 or more years 0.60 (0.43-0.84)

30. 1.56 Speroff U.S. Breast Cancer Prevalence NEJM 356:1670, 2007 Rate decreased 2.5% in 2002, 7% in 2003, level in 2004. Mostly ER+ tumors in women ages 50-69, BUT SAME DECREASE IN WOMEN 70+ (low use of hormones). Two possible reasons: 1. Use of mammography declined 2000 through 2005. 2. This decrease occurred within two years of initial WHI reports: WILL PRE-EXISTING TUMORS REGRESS OR SHOW UP LATER?

31. 1.56 Speroff Geneva Prevalence Statistics BMC Cancer 2006;6:78 Beginning in 1997, peak of breast cancer in Geneva: Increased in younger women, peak at age 60-64. Increase only in Stage I & II disease, ER+ tumors. Increase only in hormone users.

32. 1.56 Speroff E-P Favorably Influences Gene Expression BMC Medicine 2006;4:16 In ER positive tumors, E-P therapy was associated with better survival, altering the regulation of 276 genes involved in DNA repair and cell-cycle regulation.

33. 1.56 Speroff Progestins & PR-A, PR-B Molec Endocr 19:574, 2005 Br Ca Res Treat 79:233, 2003 1.Genes up-reg. by E are down-reg. by progestins. 2.PR-A excess: aggressive, poorly diff. tumors. 4. PR-A dominant in absence of progestins. 4. Progestins decrease breast tissue levels of PR-A, producing benefical change in PR-A:PR-B ratio.

34. 1.56 Speroff Benefits of Progesterone Receptor Molec Endocrinol 2008;22:1812 1. PR functions with and without ligand. 2. Antagonizes inflammatory response. 3. Blocks expression of oncogenic growth factors. 4. Inhibits induction of aromatase enzyme activity. 5. Decreases expression of COX-2, mediator of aromatase and HER-2/neu.

35. 1.56 Speroff Evidence for Beneficial Effect of Progestins 1. E/P increases receptor-postiive tumors quickly. 2. E/P down regulates estrogen-regulated genes. 3. E/P actives repair and normal function genes. 4. E/P alters the PR-A:PR-B ratio. 5. E/P associated with lower grade/stage tumors and reduces breast cancer mortality.

36. The Message for Clinicians Effect greater with E/P, more rapid, and lower grade/stage, better survival rates: JAMA 289:3243, 2003 JAMA 289:3254, 2003 Cancer 97:1387, 2003 Cancer 100:2328, 2004 Cancer Causes Control 17:695, 2006 Speroff

37. The Message for Clinicians Effect in ER+/PR+, lobular cancers, only in current users: JAMA 289:3254, 2003 Br J Cancer 91:644, 2004 Cancer 100:2328, 2004 Cancer 101:1490, 2004 Cancer Causes Control 17:695, 2006 Arch Intern Med 166:1027, 2006 Speroff

38. The Message for Clinicians There is either a small increase in the risk of breast cancer with E/P or the data reflect an impact on pre-existing tumors. It’s possible that E/P causes greater differentiation and earlier detection of pre- existing tumors resulting in better outcomes. Speroff

39. The Message for Patients The Risk of Breast Cancer: The evidence does not support a major increase in risk. Positive family history not a contraindication. Speroff

40. The Message for Patients The Risk of Breast Cancer: 1. A contrasting example. 2. An alternative explanation. Speroff