Hormonal contraception has a protective effect over ovarian and endometrial cancer development. Relative risk of ovarian cancer decreases by ∼20% for each 5 years of use; it is ∼50% for 15 years of use and decreasing with further use. The protective effect gained declines as time passes from its last use, but a significant effect remains a long time after ceasing. The effect is independent from the type of formulation used. Hormonal contraceptives do not protect from mucinous types of ovarian tumours. Relative risk reduction of endometrial cancer is even higher; the estimated relative risk decrease is ∼50% with 4 years of use, ∼70% with 12 years of use and decreasing with further use. After ceasing oral contraception, the risk begins to rise from its reduced levels but it is still ∼50% even after >20 years after its last use. Hormonal contraception could be used for primary protection from ovarian and endometrial cancer development.
Hormonal contraception has been introduced in the early 1960s and they have become a dominant form of female contraception in most developed countries. It is estimated that more than 300 million of women have used them. In United Kingdom, ∼25% of women aged 16–49 years and in USA ∼20% of women aged 15–44 years take oral contraceptives. The percentages are even higher in younger population, reaching in UK ∼60% of women aged 16–24 years. Often, oral-contraceptive users do it for prolonged periods and at a time of their life of good health while a great proportion of them are pushed by their doctors to do so for the non-contraceptive benefits of them. Apart from protection from an undesired pregnancy and the risks of pregnancy termination, the relation of these drugs with cancer is extremely important.
Ovarian cancer represents a common and rather aggressive type of gynaecological cancer, being responsible for most of the deaths from gynaecological malignancies. Hormonal contraception is associated with alterations in the biological activity of the ovaries, causing reversible anovulation, as well as by changes in pituitary function by reducing the levels of gonadotrophins and, consequently, of ovarian sex steroids. This might reduce the risk of ovarian cancer development. It is true that from past several years a beneficial effect on ovarian cancer risk has been supposed. This supposed protective effect of hormonal contraception over ovarian cancer might have serious public health implications: since advances in screening and early diagnosis of ovarian cancer are limited, primary prevention resulting from hormonal contraception use would be a major progress.
Endometrial cancer has not the unfavourable prognosis of ovarian cancer. However, it has a higher prevalence than that of ovarian cancer, thus having greater public health significance. Hormonal contraception seems to be associated with alterations in the mitotic activity of the endometrium and, as a result, it might alter the risk of endometrial cancer development.
Nevertheless, after so many years in use, the association of ovarian and endometrial cancers with oral contraception remains an extremely challenging issue to be answered. Malignancies become common in women aged above 40 years while oral contraceptives are used mainly by women aged between 20 and 30 years: this elapsed time makes it difficult to design proper epidemiological studies with well-studied population. Furthermore, women may use different formulations of oral contraceptives for various periods in their lifetime and the time between cessation of hormonal contraceptive use and cancer development might differ widely as well. Moreover, ovarian cancer is a histologically heterogeneous group of tumours with varying invasiveness. These parameters may impact on the effect of hormonal contraceptives on ovarian and endometrial cancer risk.
This review article aims to estimate the effect of hormonal contraception use on ovarian and endometrial cancer risk as well as to clarify the role of several parameters that might affect this effect.
Ovarian cancer
Ovarian cancer risk in oral-contraceptive users and the relation with the duration of use
Nowadays, there is a substantial body of epidemiological data examining the relationship between hormonal contraception and ovarian cancer. Beral, on behalf of the Collaborative Group on Epidemiological Studies of Ovarian Cancer, recently has published a collaborative re-analysis of data from 45 epidemiological cohort and case-control studies including 23 257 women with ovarian cancer and 87 303 controls from 21 countries. In total, 31% (7308) of the women with ovarian cancer and 37% (32717) of the controls had used oral contraceptives and the average duration of use was 4.4 and 5.0 years, respectively. The overall relative risk (RR; stratified by study, age, parity and hysterectomy) for ever users versus never users was found to be 0.73 (95% confidence interval (CI): 0.70–0.76, p < 0.0001) ( Table 1 ). Furthermore, the longer the duration of use the lower the risk for ovarian cancer development. The overall relative risk decreased by 20% (95% CI: 18–23%, p < 0.0001) for each 5 years of use; in women who had taken oral contraceptives for ∼15 years, the risk for ovarian cancer was almost halved and decreasing with further use. It is important to note that the gradual decrease in ovarian cancer risk observed with increasing duration of use was statistically significant ( p < 0.00001) and seems to start after at least 1 year of treatment; less than 1-year users versus non-users (mean 0.4 years) had an RR of 1.0 (95% CI: 0.91–1.10), users for 1–4 years (mean 2.4 years) had an RR of 0.78 (95% CI: 0.73–0.83), users for 5–9 years (mean 6.8 years) had an RR of 0.64 (95% CI: 0.59–0.69), users for 10–14 years (mean 11.6 years) had a relative risk of 0.56 (95% CI: 0.50–0.62) and, finally, users for 15 years or more (mean 18.3 years) had a relative risk of 0.42 (95% CI: 0.36–0.49).
| Study | Study design | Population | Follow-up | Mean duration of use | RR (95% CI) |
|---|---|---|---|---|---|
| Collaborative Group on Ovarian Cancer 21 Countries | Re-analysis of 45 epidemiological cohort and case-control studies in case-control design | Cases: 23257 Controls: 87303 | – | Cases: 4.4 years Controls: 5.0 years | 0.73 (0.70–0.76) |
| Hannaford et al UK | Cohort Study Recruitment: 1968-1969 | Ever users: 23377 Never users: 23796 | Ever users: 744000 women-years Never users: 339000 women-years | 44 months | 0.51(0.33–0.78) |
| Vessey and Painter UK | Cohort Study Recruitment: 1968-1974 | Ever and never users: 17032 | Ever users: 353000 women-years Never users: 187000 women-years | NR | 0.5 (0.3–0.7) |
There are also two more recent cohort studies whose data are not included in the aforementioned collaborative re-analysis. The Oxford-Family Planning Association (Oxford-FPA) contraceptive study includes 17 032 women recruited at 17 family planning clinics in England and Scotland at ages 25–39 years between 1968 and 1974. Some early data on the ovarian cancer incidence in relation to oral contraceptives have already been published, and they have been included in the re-analysis of the Collaborative Group on Epidemiological Studies of Ovarian Cancer. However, the latest findings till the end of 2004 based on much larger numbers of cases and longer follow-up have recently been published, including a total of 540 000 woman-years of observation and 58 ovarian cancer cases in the control group and 48 cases in the oral contraceptives group. The overall ovarian cancer relative risk for ever users versus never users in this study was found to be 0.5 (95% CI: 0.3–0.7) ( Table 1 ). Furthermore, this group of investigators found that ovarian cancer risk is significantly lower in women on oral contraceptives for more than 4 years, suggesting that a minimum period of use may be necessary to obtain their protective effect.
The Royal College of General Practitioners’ oral contraceptive study started in 1968 and, over a period of 14 months, 23 377 women who were using oral contraceptives and 23 796 women who had never used them were recruited. The latest results including a total of 339 000 woman-years of observation for never users and 744 000 woman-years for ever users have recently been published. The mean duration of oral contraceptive use in the study was 44 months. A statistically significant reduced ovarian cancer risk (relative risk (RR) 0.51, 95% CI: 0.33–0.78) was found in ever users of oral contraceptives as compared with never users ( Table 1 ). Furthermore, a statistically significant gradual decrease ( p = 0.0015) with increasing duration of oral contraceptives use was observed: for oral contraceptive users of <48 months the RR was 0.58 (95% CI: 0.33–1.04), for users of 49–96 months it was 0.57 (95% CI: 0.30–1.07) and for users of ≥97 months the RR was 0.38 (95% CI: 0.16–0.88).
It seems, therefore, that the use of hormonal contraception has a protective effect over ovarian cancer development. It is also important to note that a beneficial effect seems to exist in women with a family history or a genetic predisposition to ovarian cancer. Narod et al. conducted a study on 207 women with a genetic predisposition to ovarian cancer (179 carriers of BRCA1 and 28 carriers of BRCA2) and 161 sister controls. They found an overall RR (adjusted for geographic area of residence, age, parity and age at first birth) of 0.5 (95% CI: 0.3–0.8) for ever users versus never users. The RR was even lower with increasing duration of use: for >6 years’ users it was 0.4 (95% CI: 0.2–0.7). McGuire et al. have studied ovarian cancer risk in a case control study of 36 BRCA1 carriers with ovarian cancer, 381 non-carriers cases and 568 population controls. A statistically significantly reduced RR (adjusted for age, ethnicity and parity) was found for ever users versus never users among non-carriers (RR 0.55, 95% CI: 0.41–0.75) and a reduced, although not reaching statistical significance, the RR among carriers (RR 0.54, 95% CI: 0.26–1.13). However, the RR for ovarian cancer was statistically significantly reduced for use ≥7 years both in BRCA1 carriers (RR 0.22, 95% CI: 0.07–0.71) and in non-carriers (RR 0.43, 95% CI: 0.30–0.63). A trend for reduced ovarian cancer development was also found in a case-control study conducted by Whittemore et al., including 147 ovarian cancer cases and 304 controls with BRCA1 and MRCA2 mutations. The RR was 0.85 for ever users versus never users (95% CI: 0.53–1.40) and that for ≥6 years of use 0.62 (95% CI: 0.35–1.10). However, Modan et al., in a case-control study including 240 ovarian cancer cases carriers of BRCA1 and BRCA2 592 ovarian non-carriers cases and 2257 controls, failed to find any protective effect of oral contraceptives on BRCA carriers. The RR for ever users versus never users, as expected, was 0.53 (95% CI: 0.34–0.54) in non-carriers but it was 1.07 (95% CI: 0.63–1.83) in carriers. A limitation of the study is thought to be the absence of any information on carrier status in controls.
Protective effect according to the age of first use of oral contraceptives and the interval from last use
Another interesting question is whether the protective effect gained with the use of oral contraceptives remains unchanged through woman’s life or reduces as time passes from the last use of oral contraceptives. In the re-analysis of the Collaborative Group on Epidemiological Studies of Ovarian Cancer, it was found that the more recent the use of oral contraceptives the lower the RR of ovarian cancer. The proportional decline in RR per 5 years of oral contraceptive use was found to be 29% for those who had stopped use <10 years previously, 19% for those who had ceased use 10–19 years previously and 15% for those who had ceased use 20–29 years previously ( p for trend = 0.004). For >30 years elapsed time, there were not sufficient data for permanent conclusions since these patients took oral contraception for short periods in their life (early periods of oral contraception), but it seems that even in these patients a reduced risk might exist ( Table 2 ). It is also important to note that the longer the duration of use the higher the protection irrespective of the time elapsed from ceasing ( Table 2 ). Furthermore, despite the initial perception that the protective effect lasts only for a certain period of a woman’s life after use of oral contraceptives, it seems that a significant effect remains a long time after ceasing.
| Ever-users | Duration of oral contraceptives’ use | Decline in risk % for every 5 years of use (95% CI) | |||
|---|---|---|---|---|---|
| <5 years | 5-9 years | ≥10 years | |||
| Current or use <10 years previously | |||||
| Relative risk (95% CI) | 0.57 (0.50-0.64) | 0.88 (0.75-1.04) | 0.52 (0.43-0.64) | 0.39 (0.33-0.47) | 28.9 (23.0-34.3) |
| Mean duration of use | 5.8 years | 1.7 years | 7.2 years | 14.7 years | |
| Last use 10–19 years previously | |||||
| Relative risk (95% CI) | 0.67 (0.62-0.73) | 0.85 (0.75-0.97) | 0.62 (0.53-0.73) | 0.51 (0.44-0.59) | 19.4 (14.2-24.2) |
| Mean duration of use | 5.6 years | 1.6 years | 6.9 years | 13.8 years | |
| Last use 20-29 years previously | |||||
| Relative risk (95% CI) | 0.76 (0.71–0.97) | 0.81 (0.74–0.89) | 0.69 (0.60–0.78) | 0.60 (0.51–0.72) | 15.1 (8.5–21.2) |
| Mean duration of use | 4.6 years | 1.8 years | 6.7 years | 11.8 years | |
| Last use ≥30 years previously | |||||
| Relative risk (95% CI) | 0.86 (0.76–0.97) | 0.83 (0.73–0.95) | Insufficient data | Insufficient data | Insufficient data |
| Mean duration of use | 2.5 years | 1.5 years | |||
Data analysis of the Royal College of General Practitioners’ Study by time since last use of oral contraception has similar findings; there was a statistically significant ( p for trend = 0.041) decline in the risk of ovarian cancer with longer time since last use. However, the protective effect was found for at least 15 years after stopping, with reduced (although statistically non-significant) RR observed still after longer time intervals. It is impressive that the findings of the Oxford-Family Planning Association (Oxford-FPA) contraceptive study were in the same direction; there was a trend towards increasing ovarian cancer risk with time elapsed since oral contraception discontinuation but, a statistically significant in this study, protective effect still exists even in the longest interval group (>20 years).
Another parameter that was studied is whether the age of onset of oral contraception plays a role in ovarian cancer risk protection. When duration of use and time since last use were taken into account in the collaborative re-analysis of the epidemiological studies of ovarian cancer, the age of onset of oral contraception seems to play no role on the protection from ovarian cancer risk, not even the age of last use.
Protective effect according to the formulations of oral contraceptives
During the past decades, new formulations of oral contraceptives containing fewer amounts of oestrogen and progestins as well as new progestins come on the market. It is obvious that most of the epidemiological studies involved women who primarily have used older formulations containing higher doses of oestrogen (≥50 μg) and progestins. This is due to the fact that ovarian cancer usually develops in the age of 50s after oral contraception use. Although, low-dose pills effectively suppress ovulation, they might have a different effect in suppression of gonadotropin levels and, thus, their protective effect could be different from the older higher-dose preparations.
An answer to this question was attempted in the collaborative re-analysis of the epidemiological studies of ovarian cancer by dividing the women according to the mid-year of use in 1960s, 1970s and 1980s; oral contraceptives typically used in 1980s contained less than one half of the dose of those typically used in the 1960s with those used in the 1970s being in the middle. The investigators found that for a given time since last use, the calendar year of use had no effect on the ovarian cancer risk. Thus, they concluded that the lower-dose oral contraceptives have a similar protective effect on ovarian cancer risk.
Similar finding were reported by Ness et al. in their case-control study, who found that ovarian cancer risk reduction was similar in women who started oral contraception before 1972, between 1972 and 1980 and after 1980. However, this group of investigators attempted a direct answer by further dividing cases and controls in those taken high oestrogen/high progestin and those taken low oestrogen/low progestin formulations; they found an identical RR in the low-dose users (RR = 0.5, 95% CI: 0.3–0.6) and in the high-dose ones (RR = 0.5, 95% CI: 0.3–0.7).
Protective effect according to the tumour histology
As already mentioned, ovarian cancer represent a heterogeneous group of tumours with varying invasiveness. An interesting question is whether hormonal contraception has the same protective effect irrespective of tumour type. In the collaborative re-analysis of the epidemiological studies of ovarian cancer, a sub-analysis of the ovarian cancer cases according to the histological subtype has been made; data for histological subtype were available in 33 out of 45 epidemiological studies including 17 099 out of 23 257 cases. It should be noted that, among these women, ovarian cancer risk declined by 21% for each 5 years of hormonal contraception use, which is similar to the 20% observed in all women, thus indicating that the results were representative of the whole population.
The protective effect for every 5 years’ use of hormonal contraception was broadly similar for epithelial and non-epithelial tumours ( Table 3 ). However, among the epithelial tumours, there was heterogeneity across histological types (test for heterogeneity p = 0.0007). This was due to the fact that hormonal contraception seems to have little or no effect on mucinous types of ovarian tumours whereas the protective effect seems to be almost similar in the other types (serous, endometroid and clear cells) of epithelial tumours. Furthermore, the findings did not differ when mucinous and serous ovarian cancer cases were divided according to their invasiveness in malignant and borderline ones ( Table 3 ). It seems, therefore, that the proportional risk reduction is similar between the different histological types irrespective of their invasiveness, with the exception of mucinous tumours whose incidence seems to be affected very marginally.
