1 INSERM, équipe E3N, Institut Gustave Roussy, 39, Rue C.Desmoulins, 94805 Villejuif Cedex, France and 2 INSERM E0357, Hôpital Bichat, ci Bernard, 48, Rue H. Huchart, 75018, Paris, France
3 To whom correspondence should be addressed. Email: clavel{at}igr.fr
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Key words: breast cancer/cohort study/fertility drugs/infertility/risk factors
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The two most common causes of female infertility are ovulation deficiency (32%) and tubal damage (26%) (Thonneau et al., 1991
). Endometriosis is a third cause, whose prevalence is more difficult to evaluate (between 1 and 50%) according to Schweppe (1988)
. Ovulation deficiency may have environmental or psychological causes. It can also be due to a number of pathologies characterized by hormonal disorder or by irregular menstruation and ovulation, all of which require treatment. Essentially, treatment is based on a hormonal therapy that stimulates egg development and release by increasing gonadotrophin (FSH and LH) levels at a specific moment in the woman's cycle. The most commonly prescribed drug, clomiphene citrate (CC), competitively blocks estrogen receptors involved in the negative feedback of estradiol release by the ovaries, which in turn increases the production of FSH and LH.
The pathologies causing infertility and the therapies for these pathologies are associated with major variations in hormonal levels as compared with those of fertile women, and these may be carcinogenic to the breast. Many reproductive factors related to hormone dependence, in particular a high lifetime number of cycles and high levels of endogenous hormones (Kelsey et al., 1993; Clavel-Chapelon, 2002
) have been shown to be associated with breast cancer risk. Likewise, hormonal treatments such as HRT or oral contraceptives have been associated with a slight increase in breast cancer risk (Collaborative Group on Hormonal Factors in Breast Cancer, 1996
, 1997
). Epidemiological studies that have explored the risk of breast cancer associated with infertility give rather conflicting results, as reviewed by Klip et al. (2000)
.
We evaluated the impact of infertility treatment on breast cancer risk, using data from the E3N prospective cohort of 100 000 women.
![]() |
Materials and methods |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Women who reported a history of cancer other than basal cell carcinoma at baseline (n=4567) or for whom no date of diagnosis was available (n=239) were excluded from the initial pool of 98 997 subjects. Those who reported that they had never had sexual intercourse (n=1636) were also excluded from analysis. This left 92 555 women for the main analysis.
Information on infertility was recorded in three questionnaires. In the first two, sent out in 1990 and 1992, women were asked whether they had been treated for infertility and, if so, what treatment(s) they had received: fertility drugs (FDs), IVF, surgery or other complementary alternative medicine. The brand names of six drugs were mentioned: Clomid® (CC), Ondogyne® (cyclofenil), Inductor® and Neopergonal® (both HMG), Humegon® (menotrophin, a purified preparation of gonadotrophin) and GCE® (chorionic gonadotrophin). An additional field was provided for other drugs. In 1995, a third questionnaire was sent out to women who had mentioned in any of the first two that they had been treated with FDs. Start and end dates of use were requested for each drug. A total of 4629 answers were obtained, which enabled us to calculate the overall duration of treatment. In cases of inconsistency between the three questionnaires, the dedicated infertility questionnaire (the last one) was chosen as the source.
Information on potential confounders such as reproductive factors (age at menarche, menopausal status, i.e. premenopausal, post-menopausal with natural or with artificial menopause including partial and total oophrectomy, oral contraceptive use, pregnancies and breast feeding), social and anthropometric characteristics [educational level, body mass index (BMI), marital status], personal history of breast disease, family history of breast cancer and smoking habits were recorded at baseline.
All questionnaires asked participants whether cancer had been diagnosed, requesting the addresses of their physicians and permission to contact them. Deaths in the cohort were detected from reports by family members or by the postal service and by searching the insurance company (MGEN) file, which contains information on vital status. Cause of death information was obtained from the National Service on Causes of Deaths (INSERM). Information on the reimbursement of hospital fees of non-respondents to any questionnaires was obtained from the MGEN file. In this case, the subject's physician was contacted for diagnostic information, making it possible to find additional breast cancer cases. Only 1815 women could not be traced in the MGEN file (names misspelt, names changed after divorce, no longer insured with the MGEN, etc.), and non-respondents in this group were considered lost to follow-up.
For each participant, duration of participation was calculated from the date of return of the first questionnaire up to the date of breast (or other) cancer diagnosis, date of death, date of last questionnaire returned or date of end point (fixed at June 28, 2000).
To investigate the relationship between infertility treatment and breast cancer risk, a proportional hazards regression model (Cox model) was used to estimate the relative risks (RRs) and 95% confidence intervals (CIs) after adjustment for potential confounders. We used age as the time scale to adjust for age properly (Commenges et al., 1998), allowing us to express baseline risk as a function of age instead of a function of the time since inclusion. Age is withdrawn from the covariate list. All factors were entered in the model as categorical variables (see Table I for complete coding). Menopausal status was recoded as menopause yes (whatever the reason) or no. Similarly, oral contraceptive was used as a dichotomic variable ever consumer versus never. The variable on a first-degree familial history of cancer gave the number of affected first-degree relatives (mother, sisters or daughters) and that on a personal history of benign breast disease, the number of past diseases (including mastosis, fibro-cystic disease, cyst and fibro-adenoma). No bivariate analysis was performed due to the large sample size making even slight differences statistically significant. All potential confounding factors were included. Variables not significantly associated with breast cancer risk were subsequently removed. We checked that this did not modify the estimation of other parameters.
|
After exploring the overall relationship between a history of treatment for infertility (pharmacological and non-pharmacological) and breast cancer, we focused on the relationship with FD ever use (including women who declared IVF, since such treatments are generally accompanied by FDs), duration of use and age at first use. We then explored differential associations between breast cancer and potential confusing factors in women untreated and treated by FDs. We finally investigated the group of women who received any of the thee major fertility drugs: Clomid®, GCE® and Humegon®.
The SAS® v8.2 program was used for all statistical analysis.
![]() |
Results |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
A total of 2571 invasive breast cancer cases with no previous history of cancer were recorded since entry in the study. Of these, 2510 (97.6%) were confirmed by a pathology report. Cases that were only self-reported were also included, as self-reporting proved to be extremely accurate (1.6% false positive).
The distribution of the covariates used as confounders is presented in Table I by history of infertility treatment for parous and nulliparous women. Women treated for infertility problems more frequently had reported a history of benign breast disease than untreated women. They were less corpulent and more often post-menopausal at inclusion. Logically they also had fewer children and their first full-term pregnancy occurred at a later age. Age at menarche and family history of breast cancer were similar across subgroups defined by infertility treatment. We also stratified the population of treated women by parity. The distribution of the main characteristics was comparable between the latter two subgroups, though treated nulliparous women were more frequently premenopausal at inclusion and more often had had a personal history of benign breast disease than treated parous women.
Table II gives the overall association between treatment (both pharmacological and non-pharmacological) and breast cancer. In the whole study population, a history of treatment for infertility was not associated with any change in breast cancer risk (RR = 0.95, CI 0.821.11). A similar result was found with use of fertility drugs (RR = 0.94, CI 0.781.12). No modification of the breast cancer risk was found associated with long duration of use of FDs and early age at first use.
|
|
|
![]() |
Discussion |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Few studies have assessed the risk of breast cancer associated with infertility itself. Klip et al. (2000) reviewed 10 cohorts of subfertile women and four casecontrol studies. Although some subanalyses in these studies concluded that there was a significant association between certain types of subfertility (anovulation, progesterone deficiency, amenorrhoea, tubal or other ovarian disorders) and breast cancer risk (Moseson et al., 1993
; Garland et al., 1998
), most found no overall relationship (Gammon and Thompson, 1990
; Venn et al., 1995
, 1999
).
However, the lack of power reflected in the very large CIs and the small number of cases limits the statistical inferences of the study of Klip et al. (2000). The absence of complete adjustment for known risk factors for both breast cancer and infertility, such as nulliparity or BMI (Grodstein et al., 1994
), casts doubts on their conclusions. The use of FDs was not taken into account, except in a subanalysis in Venn's study (Venn et al., 1995
). Furthermore, a difficulty underlined by Healy and Venn (2003)
which is common to all such studies concerns the reliability of information about the causes of infertility (ovulation disorders or other disorders).
Use of FDs as a risk factor for breast cancer was studied in a few cohort studies. Inconsistent results from six cohort studies with relatively short follow-up times (Ron et al., 1987; Venn et al., 1995
, 1999
; Rossing et al., 1996
; Modan et al., 1998
; Potashnik et al., 1999
) have shown associations ranging from 0.5 to 2.6. No clear pattern of risk was apparent (Ricci et al., 1999
) whatever the type of treatment and the dose or duration (Burkman et al., 2003
). The latest published cohort study (Doyle et al., 2002
) did not support any association between ovarian stimulation and increased breast cancer risk (SIR = 116, 84156). Two casecontrol studies (Braga et al., 1996
; Grabrick et al., 2002
) have analysed a possible interaction with family history of breast cancer: they found a non-significant increase, which is compatible with our own result. This increase could suggest that such women are more sensitive to hormonal factors than the general population, as hypothesized by Andrieu et al. (1995)
to explain the possible interaction between abortions and family history of breast cancer.
Our study offered the opportunity to consider a variety of adjustment factors over a large number of person-years and breast cancer cases. However, in our study, infertility was inferred from the reported use of infertility treatment instead of being based on the usual definition, inability to conceive after 1 year of unprotected regular sexual intercourse. No questions in the E3N questionnaire enquired about delay of conception or desire to conceive. We probably targeted a subset of the whole infertile population, and an unknown proportion of the nulliparous women in our cohort is composed of women who were unaware that they were infertile or who refused to get adequate treatment. Therefore, it was impossible to identify a group of infertile non-treated women that would have served to make an adjustment on infertility. This might also explain the lower prevalence of infertility (percentage of treated women) that we found (7.1%) compared with the figures in the literature (between 8 and 18%; Thonneau et al., 1991). The difference is even greater in the generation born between 1925 and 1934, with a prevalence estimated at 4.4%. The most common FDs came on the market in the late 1960s (Clomid® in 1968 and Humegon® in 1967); on average, such drugs were available during 80% of their reproductive life, as defined by the delay between age 18 and menopause. We therefore also performed subanalyses on women born between 1935 and 1950, for whom such drugs were available during >95% of their reproductive life. We noted no appreciable difference in results for this population.
Another explanation for the lower prevalence in the E3N cohort is the difficulty in accurately remembering drugs taken in past decades. The response rate for the infertility questionnaire was 78%, lower than the usual rate for the follow-up questionnaires but identical to that for the dietary questionnaire, which was similar in terms of complexity. Beyond the fact that infertility has for a long time not been considered a medical issue, the treated women may not all have been fully aware of exactly which treatments were specifically targeting infertility. This underlines both the difficulty of obtaining full information and the importance of prospective studies, which reduce the risk of recall bias. Incomplete data are problematic, especially with regard to duration of treatment. However, a subanalysis of subjects with complete data led to RRs similar to those estimated in the whole population. There was no statistical power for examination of cancer risk by dose of drug, due to limited sample sizes. The majority of studies on cancer risk in fertile women have faced similar limitations.
The population of infertile women who have never received treatment is difficult to identify through self-administered questionnaires. An accurate assessment of infertility requires medical expertise beyond the scope of large cohort studies. We thought it best to define them as women who received treatment for infertility and assumed that our control of confounding factors would counterbalance the presence of false positives to some extent. As >45% of women had reached menopause when they replied to the second questionnaire (90% were >43 years old), we assumed that very few women started fertility treatments afterwards and were not identified for this particular reason. A limitation of the present study is its inability to determine the cause of infertility. Infertility and treatment for infertility may be two independent risk factors of breast cancer, but it is extremely hard to disentangle the effect of infertility on breast cancer risk from that of its treatment. When studying the impact of the treatment, infertility is a perfect example of a confounding factor.
We found no effect of the use of infertility treatment. However, we cannot definitively exclude the possibility that use of fertility treatment increases breast cancer risk in some subgroups and that infertility and its treatment have differential effects on breast cancer susceptibility.
![]() |
Acknowledgements |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Braga C, Negri E, La Vecchia C, Parazzini F, Dal Maso L and Franceschi S (1996) Fertility treatment and risk of breast cancer. Hum Reprod 11, 300303.[Abstract]
Burkman RT, Tang MT, Malone KE, Marchbanks PA, McDonald JA, Folger SG, Norman SA, Strom BL, Berstrein L, Ursin G, Weiss LK, Daling JR, Simon MS and Spirtas R (2003) Infertility drugs and the risk of breast cancer: findings from the National Institute of Child Health and Human Development Women's Contraceptive and Reproductive Experience Study. Fertil Steril 79, 844851.[CrossRef][ISI][Medline]
Clavel-Chapelon F (2002) Cumulative number of menstrual cycles and breast cancer risk: results from the E3N cohort study of French women. Cancer Causes Control 13, 831838.[CrossRef][ISI][Medline]
Collaborative Group on Hormonal Factors in Breast Cancer (1996) 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. Lancet 347, 17131727.[ISI][Medline]
Collaborative Group on Hormonal Factors in Breast Cancer (1997) Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Lancet 350, 10471059.[CrossRef][ISI][Medline]
Commenges D, Letenneur L, Joly P, Alioum A and Dartigues JF (1998) Modelling age-specific risk: application to dementia. Stat Med 17, 19731988.[CrossRef][ISI][Medline]
Doyle P, Maconochie N, Beral V, Swerdlow AJ and Tan SL (2002) Cancer incidence following treatment for infertility at a clinic in the UK. Hum Reprod 17, 22092213.
Gammon MD and Thompson DW (1990) Infertility and breast cancer: a population-based case-control study. Am J Epidemiol 132, 708716.[Abstract]
Garland M, Hunter DJ, Colditz GA, Manson JE, Stampfer MJ, Spiegelman D, Speizer F and Willett WC (1998) Menstrual cycle characteristics and history of ovulatory infertility in relation to breast cancer risk in a large cohort of US women. Am J Epidemiol 147, 636643.[Abstract]
Glud E, Kjaer SK, Troisi R and Brinton LA (1998) Fertility drugs and ovarian cancer. Epidemiol Rev 20, 237257.[ISI][Medline]
Grabrick DM, Viernkant RA, Anderson KE, Cerhan JR, Anderson VE and Seller TA (2002) Association of correlates of endogenous hormonal exposure with breast cancer risk in 426 families (United States). Cancer Causes Control 16, 333341.
Grodstein F, Goldman MB and Cramer DW (1994) Body mass index and ovulatory infertility. Epidemiology 5, 247250.[ISI][Medline]
Healy DL and Venn A (2003) Infertility medication and the risk of breast cancer. Fertil Steril 79, 852854.[CrossRef][ISI][Medline]
Kelsey JL, Gammon MD and John EM (1993) Reproductive factors and breast cancer. Epidemiol Rev 15, 3647.[ISI][Medline]
Klip H, Burger CW, Kenemans P and van Leeuwen FE (2000) Cancer risk associated with subfertility and ovulation induction: a review. Cancer Causes Control 11, 319344.[CrossRef][ISI][Medline]
Modan B, Ron E, Lerner-Geva L, Blumstein T, Menczer J, Rabinovici J, Oelsner G, Freedman L, Mashiach S and Lunenfeld B (1998) Cancer incidence in a cohort of infertile women. Am. J Epidemiol 147, 10381042.[Abstract]
Moseson M, Koenig KL, Shore RE and Pasternack BS (1993) The influence of medical conditions associated with hormones on the risk of breast cancer. Int J Epidemiol 22, 10001009.[Abstract]
Potashnik G, Lerner-Geva L, Genkin L, Chetrit A, Lunenfeld E and Porath A (1999) Fertility drugs and the risk of breast and ovarian cancers: results of a long-term follow-up study. Fertil Steril 71, 853859.[CrossRef][ISI][Medline]
Ricci E, Parazzini F, Negri E, Marsico S and La Vecchia C (1999) Fertility drugs and the risk of breast cancer. Hum Reprod 14, 16531655.
Ron E, Lunenfeld B, Menczer J, Blumstein T, Katz L, Oelsner G and Serr D (1987) Cancer incidence in a cohort of infertile women. Am J Epidemiol 125, 780790.[Abstract]
Rossing MA, Daling JR, Weiss NS, Moore DE and Self SG (1996) Risk of breast cancer in a cohort in infertile women. Gynecol Oncol 60, 37.[CrossRef][ISI][Medline]
Schweppe KW (1988) In Rock JA and Schweppe KW (eds) Recent Advances in Management of Endometriosis. Parthenon Publishing Group, Lancashire, UK, pp. 1330.
Thonneau P, Marchand S, Tallec A, Ferial ML, Ducot B, Lansac J, Lopes P, Tabaste JM and Spira A (1991) Incidence and main causes of infertility in a resident population (1,850,000) of three French regions (19881989). Hum Reprod 6, 811816.[Abstract]
Venn A, Watson L, Lumley J, Giles G, King C and Healy D (1995) Breast and ovarian cancer incidence after infertility and in vitro fertilisation. Lancet 346, 9951000.[ISI][Medline]
Venn A, Watson L, Bruinsma F, Giles G and Healy D (1999) Risk of cancer after use of fertility drugs with in-vitro fertilisation. Lancet 354, 15861590.[CrossRef][ISI][Medline]
Submitted on December 19, 2003; resubmitted on May 7, 2004; accepted on June 24, 2004.