The effect of hormonal contraception in the progression or occurrence of breast cancer remains controversial. One of the frequently asked questions concerning levonorgestrel-releasing intrauterine system is whether it affects the risk of breast cancer. The present study is a continuation of the previously published postmarketing epidemiologic study which was performed to establish the performance, user satisfaction, and continuation rate of the levonorgestrel-releasing intrauterine system in a large population (n = 17,360) of women regularly using the levonorgestrel system for contraception.1 The purpose of that study was to collect information on the incidence of complications requiring hospital treatment in levonorgestrel system users, either during or after levonorgestrel system use. Due to the possibility of hormonal contraception as a risk factor for hormone-related cancers, breast cancer was chosen among the target diagnoses. Because of the rarity of other hormone-related cancers, such as endometrial cancer, among women of reproductive age, only breast cancer was chosen as a target diagnosis in the present study.
MATERIALS AND METHODS
This epidemiologic study was based on a questionnaire sent to 23,885 women in Finland who had had a levonorgestrel-releasing intrauterine system (Mirena, Schering Oy, Turku, Finland) inserted between January 1990 and December 1993. The women were asked to sign and return the insertion form included in the levonorgestrel system sales package. A questionnaire form was sent in April 1996 to the women who returned the insertion form and whose identification information was reliable, allowing the use of the information provided for scientific research and for comparison with other health documents. These signatures represent the women's informed consent. The objective of the study was to collect information on complications occurring in levonorgestrel system users (either during or after the levonorgestrel system use) which required hospital treatment. In the present study, a potential association between the use of the levonorgestrel system and the occurrence of breast cancer was assessed. The study was based on a reanalysis of the data obtained in an earlier postmarketing study1–4 and the data on breast cancer diagnoses from Finnish Cancer Registry (between January 1990 and December 2000 for levonorgestrel system users and from year 1998 for the yearly average incidence for the Finnish female population). The study was conducted in accordance with the ethical principles of the Declaration of Helsinki. The ethics committee approval for the previous postmarketing study was received from the Ethical Committee of the Medical Faculty of the University of Turku. For the present study, the approval for collecting information from the cancer registry was granted by the Finnish National Research and Development Centre for Welfare and Health. This approval includes permission from the Data Protection Ombudsman.
The statistical methods used in this study to compare the breast cancer incidence among the women who were inserted with an levonorgestrel system and the yearly breast cancer incidence in Finland were the calculation of breast cancer incidences and their 95% confidence intervals. The incidence of breast cancer diagnosis per 100,000 woman-years was calculated using the following formula: 105 times the number of subjects with the diagnosis divided by the number of woman-years.
The confidence intervals were based on the χ2 distribution.5 The Fisher exact test was used to compare the breast cancer incidences within each age group. Spearman correlation coefficients were calculated within each levonorgestrel system age group for the relationship between breast cancer incidence and time since levonorgestrel system insertion.
Of the total of 23,885 levonorgestrel system users to whom questionnaires were sent, 18,527 (77.6%) users returned the questionnaire. A total of 1,167 questionnaires had to be left out from further analyses due to lacking or uninterpretable data or lack of signature on the questionnaire, leaving questionnaires from a total of 17,360 levonorgestrel system users available. At the time the women filled out the questionnaire, 66.5% of the levonorgestrel system users were still using the levonorgestrel system. The mean age of the levonorgestrel system users was 35.4 years at baseline (determined as the time of insertion of the levonorgestrel system). The disposition of woman-years and the number of breast cancers diagnosed in each age group are presented in Figure 1. The breast cancer diagnoses (n = 165) that occurred in the study population during the entire study period from 1990 to 2000 were obtained from the national Finnish Cancer Registry. The number of breast cancer cases and gathered woman-years in age groups less than 30 years and more than 54 years, respectively, were not considered large enough for epidemiologic analyses, therefore the incidences were not further analyzed for these age groups.
The incidence of breast cancer with 95% confidence intervals in levonorgestrel system users is presented in Figure 2 by age group per 100,000 woman-years, and is compared with the incidence data derived from the national Finnish Cancer Registry from the year 1998. In 3 age groups (40–44 years, 45–49 years, and 50–54 years), the point estimates for the breast cancer incidence in the average population seems to be higher than among the levonorgestrel system users, and in 2 age groups (30–34 years and 35–39 years, Fig. 3), the point estimate for the levonorgestrel system users seems to be higher than in the average Finnish female population. Based on the 95% confidence intervals for the incidences of breast cancer and Fisher exact test, there is no indication of a difference between the levonorgestrel system users and average Finnish female population in any of the 5-year age groups.
When the incidence of breast cancer among the levonorgestrel system users was observed as a function of time elapsed from the intrauterine system insertion to the occurrence of breast cancers, there was no apparent association between the length of time elapsed from the intrauterine system insertion up to 10 years and the yearly incidence of breast cancer either in any 5-year age category or in the pooled data (Fig. 3). This does not support a causal relationship between the use of the levonorgestrel system and the occurrence of breast cancer. However, a causal relationship cannot be totally excluded, either.
Although numerous epidemiologic studies have tried to find out whether there is an association between the use of oral contraceptives and breast cancer6 (reviewed in Burkman et al7), relatively little has been published relating to nonoral hormonal contraception. Regarding progestin-only contraceptives, a large study on the incidence of breast cancer in users of the subdermal 6-rod levonorgestrel implant (Norplant, Wyeth, Madison, NJ) has been published previously.8,9 This Norplant postmarketing surveillance study, which is a 5-year cohort study with 78,000 woman-years of observation, did not show any statistically significant difference in the occurrence of malignant neoplasms, including breast cancer, between levonorgestrel implant users and controls8,9 (reviewed by Sivin10). Also, a recent case-control study failed to detect any increase in breast cancer risk after the use of levonorgestrel implants or injectable progestin-only contraceptives.11 Because the systemic exposure to levonorgestrel is considerably less with the levonorgestrel system than with the levonorgestrel implant due to lower serum levonorgestrel levels,10,12 the breast cancer risk of levonorgestrel system users is not expected to be higher than that of levonorgestrel implant users. These data are nevertheless important, because there are more than 4 million current users of the levonorgestrel system.
The reliability of the source data regarding the general incidence of breast cancer is high,13 because in Finland almost 100% of all cancer diagnoses are included in the Finnish Cancer Registry, from which both the overall breast cancer incidence as well as the cancer diagnoses of the levonorgestrel system users were derived. Regarding the levonorgestrel system users, final data for analyses was available from 72.7% of women to whom the questionnaire was originally sent, which can be considered a large enough proportion to represent an average levonorgestrel system user. However, a nonresponse bias cannot be totally excluded, either. Inevitably, the levonorgestrel system users are included in the overall Finnish female population from which the Finnish Cancer Registry's incidence rates are derived. Because there is no official registry of all levonorgestrel system users in Finland, it is impossible to subtract the levonorgestrel system users from the overall Finnish population. In case the proportion of levonorgestrel system users to the overall Finnish population would be significant in any age category, this could cause a bias in the results. However, based on sales figures and marketing surveys on the levonorgestrel system use in Finland, it can be estimated that the proportion of levonorgestrel system users among Finnish women aged 15–44 years is below 10%. Also, because the yearly breast cancer incidence of the Finnish female population is from 1 year (1998) and the breast cancers of the levonorgestrel system users are collected over a period of 10 years (1990–2000), the effect of the removal of the cases that occurred in the levonorgestrel system users in 1998 (18 cases) on the population incidences is considered negligible.
Factors that influence breast cancer incidence include factors related to genetic background, reproductive and hormonal factors, factors related to life-style and environmental factors (such as smoking and alcohol use and geographic location) as well as various factors related to socioeconomic status and education.14 These causative factors have been identified also in the Finnish population-based studies (Lillberg K. Psychological stress, personality and risk of breast cancer: Follow-up studies in the Finnish Twin Cohort. Thesis, Department of Public Health, University of Helsinki, 2003). The relative influence of these factors on breast cancer risk varies considerably,14,15 and because of the study setting, it was impossible to control such confounding factors in the present study.
Among the strongest factors that influence the incidence of breast cancer risk are education and social status, with women with a higher socioeconomic status having an elevated risk of breast cancer compared with women with a low socioeconomic status.16 In the present study the socioeconomic status of the breast cancer patients retrieved from the Finnish Cancer Registry was not recorded, but the data represent the average socioeconomic status of breast cancer patients in Finland. As described earlier, the proportions of women in leading (ie, professional) as well as in white-collar positions is higher among the sample of levonorgestrel system users in this study compared with the Finnish women in general.1 Thus, this should increase the incidence of breast cancer among the levonorgestrel system users. However, no increase was seen in the results of the present study (Fig. 3). Furthermore, according to the product information, women are encouraged to undergo an annual gynecologic examination (including an examination of the breasts) after the insertion of the levonorgestrel system. Due to the close surveillance, the incidence of breast cancer could therefore be further accentuated. On the other hand, the levonorgestrel system is in Finland considered as the first line contraceptive only for parous women. It can, therefore, be assumed that the percentage of parous women is higher among levonorgestrel system users than in the general population. If that is the case, the protective effect of higher parity may partially compensate for the increased risk associated with a higher socioeconomic class.
In conclusion, this postmarketing study does not support an association between levonorgestrel-releasing intrauterine system use and the development of breast cancer. On the other hand, an elevated breast cancer risk cannot be totally eliminated either, due to the limited control of confounding factors in the present study. Additional larger studies with a different methodologic approach are needed either to confirm or refute the findings in this study.
1. Backman T, Huhtala S, Blom T, Luoto R, Rauramo I, Koskenvuo M. Length of use and symptoms associated with premature removal of the levonorgestrel intrauterine system: a nation-wide study of 17,360 users. BJOG 2000;107:335–9.
2. Backman T, Huhtala S, Luoto R, Tuominen J, Rauramo I, Koskenvuo M. Advance information improves user satisfaction with the levonorgestrel intrauterine system. Obstet Gynecol 2002;99:608–13.
3. Backman T, Huhtala S, Tuominen J, Luoto R, Erkkola R, Blom T, et al. Sixty thousand woman-years of experience on the levonorgestrel intrauterine system: an epidemiological survey in Finland. Eur J Contracept Reprod Health Care 2001;6(suppl):23–6.
4. Backman T, Rauramo I, Huhtala S, Koskenvuo M. Pregnancy during the use of levonorgestrel intrauterine system. Am J Obstet Gynecol 2004;190:50–4.
5. Armitage P, Berry G. Statistical methods in medical research. 2nd ed. Blackwell Scientific; Oxford, UK 1987, p. 132–4.
6. Breast cancer and hormonal contraceptives: collaborative reanalysis of individual data on 53,297 women with breast cancer and 100,239 women without breast cancer from 54 epidemiological studies. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet 1996;347:1713–27.
7. Burkman R, Schlesselman JJ, Zieman M. Safety concerns and health benefits associated with oral contraception. Am J Obstet Gynecol 2004;190:S5–22.
8. International Collaborative Post-Marketing Surveillance of Norplant. Meirik O, Farley TMM, Diaz S, et al. Post-marketing surveillance of Norplant contraceptive implants: I. Contraceptive efficacy and reproductive health. Contraception 2001;63:167–86.
9. International Collaborative Post-Marketing Surveillance of Norplant. Meirik O, Farley TMM, Collins J, et al. Post-marketing surveillance of Norplant contraceptive implants: II. Non-reproductive health. Contraception 2001;63:187–209.
10. Sivin I. Risks and benefits, advantages and disadvantages of levonorgestrel-releasing contraceptive implants. Drug Saf 2003;26:303–35.
11. Strom BL, Berlin JA, Weber AL, Norman SA, Bernstein L, Burkman RT, et al. Absence of an effect of injectable and implantable progestin-only contraceptives on subsequent risk of breast cancer. Contraception 2004;69:353–60.
12. Luukkainen T, Lahteenmaki P, Toivonen J. Levonorgestrel-releasing intrauterine device. Ann Med 1990;22:85–90.
13. Teppo L, Pukkala E, Lehtonen M. Data quality and quality control of a population-based cancer registry: experience in Finland. Acta Oncol 1994;33:365–9.
14. Key TJ, Verkasalo PK, Banks E. Epidemiology of breast cancer. Lancet Oncol 2001;2:133–40.
15. Clemons M, Goss P. Estrogen and the risk of breast cancer. [published erratum appears in: N Engl J Med 2001;344:1804]. N Engl J Med 2001;344:276–85.
16. Pukkala E. Cancer risk by social class and occupation: a survey of 109,000 cancer cases among Finns of working age. Basel (Switzerland): Karger, 1995.
© 2005 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
17. Fay MP, Feuer EJ. Confidence intervals for directly standardized rates: a method based on the gamma distribution. Stat Med 1997;16:791–801.