Postmenopausal Estrogen- Progestin Therapy and Breast Cancer
abstract & commentary
Ross and colleagues from the university of southern California reported the results of a population-based, case-control study of breast cancer risk associated with postmenopausal hormone therapy, both estrogen alone and combined estrogen-progestin in sequential and daily, continuous regimens. Combined estrogen-progestin therapy was associated with a higher odds ratio per five years of use (OR = 1.24; CI = 1.07-1.45) compared with estrogen alone (OR = 1.06;CI = 0.97-1.15). Sequential estrogen-progestin regimens were associated with a higher odds ratio per five years of use (OR = 1.38; CI = 1.13-1.68) compared with a daily, continuous combined regimen (OR = 1.09; CI = 0.88-1.35). The difference between sequential and continuous regimens was not statistically significant. Ross et al conclude that the addition of a progestin to an estrogen postmenopausal regimen increases the risk of breast cancer compared with the use of estrogen alone. (Ross RK, et al. J Natl Cancer Inst 2000;92:328-332.)
COMMENT BY LEON SPEROFF, MD
This report followed by approximately two weeks the article in JAMA reviewed in last month’s issue of OB/GYN Clinical Alert. Both reports are excellent examples of over interpretation of relatively weak epidemiological results.
In the introduction to this case-control study, Ross et al, with some obvious pride, claim that this study provides the most definitive and detailed data yet available between estrogen-progestin use and breast cancer risk. If that is the case, then their conclusion that this study "provides strong evidence that the addition of a progestin enhances markedly the risk of breast cancer relative to estrogen use alone" hardly warrants the modifying adjective and adverb so carefully chosen by Ross et al.
Let me point out some of the weaknesses that were interpreted as strengths. The difference between sequential and daily estrogen-progestin regimens was not statistically significant, but Ross et al were obviously unencumbered by this fact in their emphasis upon a difference between the regimens, especially and unfortunately, in the press release originating from the University of Southern California. The increased relative risks (RR) associated with sequential and daily estrogen-progestin regimens for more than 10 years were based on 27 cases/14 controls and 13 cases/20 controls, respectively; hardly the robust numbers claimed by Ross et al in their introduction. Indeed, where the numbers of cases were substantial, the estrogen-only users, the analysis of estrogen alone regimens indicated no statistically significant increased risk of breast cancer, even with increasing duration of use up to more than 15 years (OR = 1.06; CI = 0.97-1.15). When analyzed by stage of disease, combined estrogen-progestin regimens were associated with a barely significant increase in localized disease and no significant increase in in situ or advanced disease; sequential estrogen-progestin regimens with a significant increase in localized disease and no significant increase in in situ or advanced disease; and daily estrogen-progestin regimens with no significant increase in any of the categories. This variation and the strength of the associations (RRs that range from 0.98 to 1.44) do not provide evidence of a major effect.
The emphasis and interpretation of the current reports examining the effect of estrogen and progestin could have been of a totally different nature. The numbers indicated no significant increased risk with estrogen therapy, even of long duration, a conclusion based on more cases compared with the number of cases in estrogen-progestin users. If one placed the emphasis where the greater numbers are, the message would be a reassuring one.
|Table 1-Relative Risks of Breast Cancer Associated with Postmenopausal Estrogen-Progestin Treatment (Statistically Significant)|
|Reference||Relative Risk||(Confidence Interval)|
|Colditz GA, et al, 1995.1||1.41||(1.15-1.74)|
|Persson I, et al, 1999.2||1.7||(1.1-2.6) > 6 yrs|
|Magnusson C, et al, 1999.3||1.68||(1.39-2.03)|
|Schairer C, et al, 2000.4||1.40||(1.1-1.8)|
|Table 2-Relative Risks of Breast Cancer Associated with Postmenopausal Estrogen-Progestin Treatment (Not Statistically Significant)|
|Reference||Relative Risk||(Confidence Interval)|
|Kaufman DW, et al, 1984.5||1.7||(0.9-3.3)|
|Ewertz M, et al, 1988.6||1.36||(0.98-1.87)|
|Bergvist L, et al, 1989.7||4.4||(0.9-22.4)|
|Yang CP, et al, 1992.8||1.2||(0.6-2.2)|
|Stanford JL, et al, 1995.9||0.9||(0.6-1.2)|
|Newcomb PA, et al, 1995.10||0.75||(0.49-1.15) < 5 yrs|
|1.12 (0.72-1.76)||> 5 yrs|
|LaVecchia C, et al, 1995.11||1.6||(0.4-6.3)|
|World reanalysis, 1997.12||1.53||(0.80-2.92)|
|Brinton LA, et al, 1998.13||0.99||(0.7-1.3)|
|Persson I, et al, 1999.2||1.4 (0.9-2.3)||1-6 yrs|
The available epidemiologic evidence summarized in tables 1 and 2 on the effect of combined estrogen-progestin treatment on the risk of breast cancer indicates a mixed story, not a uniform and consistent result (10 negative studies and 4 positive studies). Those who implicate progestins have given great weight to the observation that proliferation and mitotic activity peak in the luteal phase. Recent studies, however, indicate that prolonged exposure to a constant level of progestin (unlike pregnancy or a menstrual cycle) provides an inhibiting influence, a possible advantage for the postmenopausal regimen of the daily, continuous administration of estrogen-progestin.
There is another aspect of this controversial issue that deserves more publicity. Most of the studies that have examined the breast cancer mortality rates of women who had used postmenopausal hormone therapy have documented improved survival rates.14-23 Even studies that detect an increased risk of breast cancer in hormone users indicate a paradoxical better outcome. This undoubtedly partly reflects earlier diagnosis in users because the greater survival rate in current users is associated with a lower frequency of late stage disease.12,15, 19,22,24-27 There is also evidence to suggest that estrogen users develop smaller, better-differentiated (lower grade) tumors, and that surveillance/detection bias is not the only explanation for better survival.25,27-31 These biologic differences imply that hormone treatment promotes the growth of a malignant locus already in place, and it presents clinically with a more favorable biology. This conclusion is consistent with the fact that virtually all the positive studies find that any increase in risk disappears within five years of discontinuing hormone therapy, and tumors occur at an earlier stage and a younger age in women using hormone therapy. Lower grade tumors are present even when there is no difference in the prevalence of mammography comparing hormone users and nonusers, or when the data are adjusted for the method of detection.21,23,31 Thus, an important effect is on grade of disease, tumor differentiation, and aneuploidy. An excess of grade I tumors has been documented equally in users of estrogen alone and in users of combined estrogen and progestin.32
It is, in my view, appropriate to emphasize the benefits of postmenopausal hormone therapy, point out the continuing concern regarding the relationship between estrogen use and breast cancer (particularly long-term use), and to emphasize the absence of definitive evidence linking such therapy to an increased risk of breast cancer, as well as the uniform data indicating better outcomes in hormone users who develop breast cancer. Additional case-control and cohort studies will only confirm the variability and inconsistency of the findings. Thus, a definitive answer must await the results of the ongoing randomized trials.
1. Colditz GA, et al. N Engl J Med 1995;332:1589.
2. Persson I, et al. Cancer Causes Control 1999;10: 253-260.
3. Magnusson C, et al. Int J Cancer 1999;81:339-344.
4. Schairer C, et al. JAMA 2000;283:485-491.
5. Kaufman DW, et al. JAMA 1984;252:63-67.
6. Ewertz M. Int J Cancer 1988;42:832-838.
7. Bergkvist L, et al. N Engl J Med 1989;321:293.
8. Yang CP, et al. Cancer Causes Control 1992;3:475.
9. Stanford JL, et al. JAMA 1995;274:137.
10. Newcomb PA, et al. Am J Epidemiol 1995;142: 788-795.
11. La Vecchia C, et al. Br J Cancer 1995;72:244-248.
12. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet 1997;350:1047-1059.
13. Brinton LA, et al. Menopause 1998;5:145-151.
14. Bergkvist L, et al. Am J Epidemiol 1989;130:221-227.
15. Hunt K, et al. Br J Obstet Gynaecol 1990;97:1080.
16. Henderson BE. Arch Intern Med 1991;151:75-78.
17. Persson I. Int J Cancer 1996;67:327-332.
18. Willis DB. Cancer Causes Control 1996;7:449-457.
19. Grodstein F, et al. N Engl J Med 1997;336: 1769-1775.
20. Sellers TA, et al. Ann Intern Med 1997;127:973-980.
21. Schairer C, et al. J Natl Cancer Inst 1999;91:264-270.
22. Fowble B. J Clin Oncol 1999;17:1680-1688.
23. Jernström H, et al. Br J Cancer 1999;80:1453-1458.
24. Strickland DM, et al. Obstet Gynecol 1992;80:400.
25. Bonnier P, et al. Obstet Gynecol 1995;85:11.
26. Bonnier P, et al. Int J Cancer 1998;79:278-282.
27. Salmon RJ, et al. Oncol Rep 1999;6:699-703.
28. Magnusson C, et al. Breast Cancer Res Treat 1996; 38:325-334.
29. Holli K, et al. J Clin Oncol 1998;16:3115-3120.
30. O’Connor IF, et al. J Clin Pathol 1998;51:935-938.
31. Bilimoria MM, et al. Ann Surg Oncol 1999;6:200-207.
32. Harding C, et al. BMJ 1996;312:1646-1647.
The following statements regarding the risk of breast cancer associated with postmenopausal hormone therapy are true except:
a. The data available from case-control and cohort studies do not provide uniform, consistent results.
b. Women who use postmenopausal hormone therapy and develop breast cancer while using hormone therapy have a reduced risk of dying of breast cancer.
c. Postmenopausal hormone users have more breast exams and mammograms.
d. The relative risks of breast cancer associated with postmenopausal hormone therapy are in the range recognized as strong associations (> 2.0).