Oral contraceptive and IUD use and endometrial cancer: A population-based case–control study in Shanghai, China

Authors


Abstract

Oral contraceptive (OC) and intrauterine device (IUD) use have been shown to be protective factors for endometrial cancer in several epidemiological studies; however, few studies have been conducted in Chinese populations. We evaluated the association between OC and IUD use and endometrial cancer risk in a population-based case–control study among Chinese women in Shanghai, China. The study included 1,204 newly diagnosed endometrial cancer cases and 1,212 age frequency-matched healthy controls. Logistic regression models were used to estimate adjusted odds ratios (OR) and their 95% confidence intervals (95% CI). In our study population, 18.5% cases and 24.9% controls reported having ever used OCs with an OR of 0.75 (95% CI, 0.60–0.93), after adjusting for known risk or protective factors for endometrial cancer. The risk of endometrial cancer decreased with long-term use of OCs with the OR for more than 72 months of use being 0.50 (95% CI, 0.30–0.85). The effect of OC use remained 25 or more years after cessation of use; the associated OR was 0.57 (95% CI = 0.42–0.78) as compared to nonusers. Similarly, fewer cases than controls had ever used IUD, with the multivariable adjusted OR being 0.53 (95% CI = 0.43–0.65). A reduction in risk was observed regardless the duration of use or age at first and last use. These results suggest that OC and IUD use may confer long-term protection against endometrial cancer. © 2006 Wiley-Liss, Inc.

The association of oral contraceptives (OCs) and risk of endometrial cancer has been investigated in a number of epidemiological studies conducted in Western countries.1, 2, 3, 4, 5, 6, 7, 8, 9 Use of OCs, especially combined oral contraceptives (COC), has generally been seen to reduce the risk of endometrial cancer, although the evidence is not entirely consistent.10 The relationship between OC use and endometrial cancer risk has only been evaluated in one small scale Chinese study.11 No conclusions could be drawn due to the low prevalence of OC use in the study population.11

Owing to the practice of limiting family size in China, there has been widespread use of contraceptives since the mid-1970s; 98.6% of married women used contraception, and 29.3% used OCs in 1982.12 Many previous OC users have now entered the high risk age of endometrial cancer (passed the age of menopause, aged 50–69), offering an opportunity to evaluate the effect of OCs on the risk of endometrial cancer in the Chinese population. Intrauterine devices (IUD), a birth control method used by approximately 100 million women worldwide,13 are the most popular contraceptive method (29.5%) in China.12 Few epidemiological studies have examined the relationship between IUD use and endometrial cancer risk.11, 14, 15, 16, 17, 18, 19 A small-scale study we conducted in China in 1993 found no significant association.11 In this report, we describe the association of OC and IUD use with endometrial cancer in a recently completed population-based case–control study conducted in Shanghai, China.

Subjects and methods

The Shanghai Endometrial Cancer Study is a population-based case–control study. Cases consisted of permanent female residents of urban Shanghai, who were newly diagnosed with endometrial cancer between January 1997 and December 2003 and who were between 30 and 69 years of age at the time of diagnosis. Through the population based Shanghai Cancer Registry, 1,452 eligible endometrial cancer cases were identified and in-person interviews were completed for 1,204 (82.8%) of them. The major reasons for nonparticipation included refusal (135 cases, 9.3%), death before interview (66 cases, 4.5%), inability to locate the subject (23 cases, 1.6%), absence during the study period (12 cases, 0.8%) and other miscellaneous reasons (12 cases, 0.8%).

The Shanghai Resident Registry, which registers all permanent residents of urban Shanghai, was used to randomly select female controls that were frequency-matched to cases by age (5-year intervals). The number of controls in each age-specific stratum was determined in advance according to the age distribution of incident endometrial cancer cases in 1996. Women who had had a hysterectomy were not eligible (n = 63). Interviews were completed for 1,212 (74.4%) of the 1,629 eligible controls. Reasons for nonparticipation of controls included refusal (340 controls, 20.9%) and absence during the study period (61 controls, 3.7%).

All study participants were interviewed in person by trained interviewers and were measured for weight, circumferences of the waist and hips, and sitting and standing heights. The standardized, structured questionnaire used for the study covered demographic factors, menstrual and reproductive history, hormone use, usual dietary habits, prior disease history, regular exercise, tobacco and alcohol use, weight history and family history of cancer. Subjects were asked if they had ever used OCs, injectable contraception, an intrauterine device (IUD), hormone replacement therapy and other female hormones to treat endometriosis, help in becoming pregnant or treat acne. Subjects who had ever taken any OCs were queried for age at first use, duration of use, age at last use and the name of the contraceptive used for each episode. Information on age at first use, duration of use, and age at last use was collected for IUD users.

Menopausal status was defined as no menstruation during the past 12 months, excluding those lapses caused by pregnancy or breastfeeding. For premenopausal women, years of menstruation was calculated by taking age at diagnosis for cases or age at interview for controls and subtracting age at menarche. For postmenopausal women, years of menstruation was calculated by taking age at menopause and subtracting age at menarche. For both groups, years of oral contraceptive and/or injected contraceptive use and years of pregnancy were subtracted from the total years of menstruation. Exercise was measured as whether the participant regularly took part in any kind of exercise/sports activity during last 10 years. Passive smoking was defined as either exposure to smoking at home or smoking in the workplace.

Statistical analyses were conducted using SAS Version 9.1 (SAS Institute, Cary, NC). Odds ratios (OR) and their corresponding 95% confidence intervals (95% CI) were used to estimate the association of endometrial cancer risk with contraceptive use. Unconditional logistic regression was used to obtain maximum likelihood estimates of ORs and their 95% CIs, with adjustment for potential confounding variables. Age and BMI (weight/height2) were included as a continuous variable throughout the data analyses, and categorical variables were treated as dummy variables in the model. We found that age did not modify the effect of OC use and adjusting it categorically produced similar results. Trend tests were conducted by treating categorical variables as the ordinal values in the models. All statistical tests were based on 2-sided probability and a significance level of p ≤ 0.05.

Results

Table I shows comparisons of cases and controls on demographic factors and the known risk factors for endometrial cancer. Compared to controls, cases had an earlier age at menarche, later age at menopause and longer duration of menstruation. Cases were also more likely to have a higher educational level, a family history of colorectal, endometrial, or breast cancer among first-degree relatives, a higher BMI or WHR, a higher frequency of nulliparity and use of female hormones other than HRT, and were less likely to be engaged in regular exercise. All of the above-mentioned variables were considered to be potential confounders and were adjusted for in subsequent analyses. Cases and controls did not differ significantly regarding family income, hormone replacement therapy, smoking status, or passive smoking status.

Table I. Comparison of Cases and Controls on Demographics and Selected Endometrial Cancer Risk Factors1
 Cases (n = 1,204)Controls (n = 1,212)p-Value
  • 1

    Subjects with missing values were excluded from the analysis.

  • 2

    Family history of colorectal, endometrial, and breast cancer among first-degree relatives.

  • 3

    Median (25th, 75th percentile) are presented.

  • *

    Exogenous hormones for other reasons, such as treatment of endometriosis, helping to become pregnant, and treatment of acne.

  • §

    Among parous women.

Age (mean ± SD)54.9 ± 8.555.1 ± 8.50.75
Education (%)
 No formal education7.911.0 
 Elementary education14.112.9 
 Middle and high school62.963.3 
 Professional, college and above15.112.80.03
Marital status (%)
 Unmarried1.51.1 
 Married or cohabiting87.687.6 
 Separated/divorced/widowed10.911.30.63
Family income in previous year (yuan) (%)   
 <10,00011.611.2 
 10,000–14,99921.721.3 
 15,000–19,99919.918.5 
 20,000–29,99921.925.6 
 ≥30,00024.923.40.33
Number of pregnancies (%)
 Never7.43.6 
 116.913.8 
 225.226.0 
 322.324.9 
 416.917.6 
 ≥511.314.1<0.01
Cancer among first-degree relative (%)28.54.4<0.01
Natural menopause (%)97.099.1<0.01
Hormone replacement therapy (%)4.44.00.66
Exogenous hormone use for other reasons (%)*8.53.3<0.01
Ever smoking (%)3.43.50.94
Passive smoking (%)63.966.10.25
Regular exercise (%)28.733.50.01
Age at menarche314.0 (13.0, 16.0)15.0 (13.0, 16.0)<0.01
Age at menopause(among postmenopausal women)350.3 (48.6, 52.5)49.6 (47.3, 51.1)<0.01
Years of menstruation333.3 (30.1, 36.1)31.5 (28.2, 34.4)<0.01
Age at last live birth3§28.0 (26.0, 30.0)28.0 (26.0, 31.0)0.02
Body mass index (BMI)325.2 (22.9, 28.1)23.5 (21.4, 26.0)<0.01
Waist-to-hip circumference ratio (WHR)30.84 (0.81, 0.87)0.81 (0.78, 0.85)<0.01

Among all study subjects, 223 cases (18.5%) and 302 controls (24.9%) reported having ever used OCs in the past. The multivariable-adjusted ORs of endometrial cancer associated with OC use among all subjects are presented in Table II. After adjusting for age at menarche, years of menstruation, menopausal status, BMI and other potential confounders, the OR for any use of OCs compared to never use was 0.75 (95% CI 0.60–0.93). The OR and 95% CI for more than 72 months of use versus never use were 0.50 (0.30–0.85). Compared to never users, the ORs varied by age at first use (OR = 0.65, 95% CI, 0.49–0.85; OR = 0.90, 95% CI, 0.67–1.23, respectively, for age at first use of OCs of less than 30 years, and greater or equal to 30 years). Analysis based on discontinuing use of OCs showed that the reduced risk of endometrial cancer among OCs users was more pronounced in women who had discontinued use for more than 25 years (OR = 0.57, 95% CI = 0.42–0.78).

Table II. Association of Contraceptive Methods and the Risk of Endometrial Cancer
 CasesControlsOR195% CIOR295% CI
  • 1

    Logistic regression adjusted for age.

  • 2

    Logistic regression adjusted for age, education level, body mass index (BMI), family history of colorectal, endometrial, and breast cancer among first-degree relatives, age at menarche, years of menstruation, menopausal status, number of pregnancies, age at last live birth, physical activity, exogenous hormone use (hormone replacement therapy, exogenous hormone use for other reasons).

  • 3

    Trend test based on exposed participants.

OC use
 Never9819101.0 1.0 
 Ever2233020.690.56–0.830.750.60–0.93
Duration of OC use (months)
 Never9819101.0 1.0 
 <673591.150.81–1.640.940.64–1.38
 6–2358700.770.54–1.100.740.50–1.09
 24–7265930.650.47–0.900.750.52–1.07
 ≥7227800.310.20–0.490.500.30–0.85
     ptrend3 = 0.14
Age at first OC use (year)
 Never9819101.0 1.0 
 <301181660.660.51–0.850.650.49–0.85
 ≥301051360.720.55–0.940.900.67–1.23
Years since last OC use (years)
 Never9819101.0 1.0 
 1–241281580.750.59–0.970.960.72–1.27
 ≥25951440.610.46–0.810.570.42–0.78
IUD use
 Never6745391.0 1.0 
 Ever4856370.530.44–0.640.530.43–0.65
Duration of IUD use (months)
 Never6745391.0 1.0 
 <601051540.510.39–0.670.530.39–0.72
 60–1681391570.620.47–0.810.740.55–1.00
 168–2161061420.520.39–0.700.480.35–0.67
 ≥2161251760.530.41–0.690.450.34–0.60
     ptrend3 = 0.34
Age at first IUD use (year)
 Never6745391.0 1.0 
 <301972890.580.46–0.740.580.55–0.76
 30–342172820.550.44–0.690.540.42–0.69
 ≥35701160.450.33–0.630.470.33–0.67
     ptrend3 = 0.18
Age at last IUD use (year)
 Never6745391.0 1.0 
 <401571420.810.62–1.050.910.68–1.22
 40–44771220.430.31–0.590.510.35–0.72
 45–491291730.520.40–0.680.500.37–0.68
 ≥501121880.460.35–0.590.400.30–0.53
     ptrend3 = 0.89
Years since last IUD use (years)
 Never6745391.0 1.0 
 <52313560.420.33–0.520.420.32–0.54
 ≥52542810.670.54–0.820.650.52–0.82
Injectable contraception use
 Never117111771.0 1.0 
 Ever22350.950.59–1.541.070.63–1.81

We also evaluated the association between other contraceptive methods and endometrial cancer risk. Use of injectable contraception was similar between cases (2.7%) and controls (2.9%), and was not associated with risk. IUD, the most common contraceptive method in the study population, was associated with a decreased risk of endometrial cancer (OR = 0.53, 95% CI = 0.43–0.65). The risk reduction did not vary significantly by duration of IUD use or age at first and last use (Table II). Compared to women who never used any contraceptive method, reductions in endometrial cancer risk were observed for subjects who had used only OCs (OR = 0.64, 95% CI = 0.46–0.88); only IUD (OR = 0.48, 95% CI = 0.38–0.60); and both methods (OR = 0.48, 95% CI = 0.35–0.65) during their reproductive years (data not shown). Among the 116 cases and 160 controls who had used both OCs and IUD, the vast majority (70.2% cases and 85.6% controls) took OCs prior to use of IUD.

The association of OC use and endometrial cancer was further evaluated stratifying by menopausal status (Table III). The patterns of OC use and endometrial cancer were similar for both pre- and postmenopausal women, although the point estimates of the duration of OC use were only significant for postmenopausal women, probably due to the small number of study participants in the premenopausal group.

Table III. Effects of OC Use on the Risk of Endometrial Cancer Stratified on Menopausal Status
 Premenopausal womenPostmenopausal women
CasesControlsOR195% CICasesControlsOR195% CI
  • 1

    Logistic regression adjusted for age, education level, body mass index (BMI), family history of colorectal, endometrial, and breast cancer among first-degree relatives, age at menarche, years of menstruation, number of pregnancies, age at last live birth, physical activity, exogenous hormone use (hormone replacement therapy, exogenous hormone use for other reasons).

  • 2

    Trend test based on exposed participants.

  • 3

    Cut point is 25 years since last OC use among postmenopausal women.

OC use
 Never4153791.0 5665311.0 
 Ever87680.780.50–1.221362340.700.53–0.92
Duration of OC use (months)
 Never4153791.0 5665311.0 
 <2447390.830.50–1.3884900.820.58–1.16
 ≥2440290.660.31–1.42521440.580.39–0.86
   ptrend2 = 0.17  ptrend2 = 0.11
Age at first OC use (year)
 Never4153791.0 5665311.0 
 <3047440.750.41–1.37711220.610.43–0.86
 ≥3040241.140.61–2.13651120.830.57–1.21
Years since last OC use
 Never4153791.0 5665311.0 
 1–19 (24)48381.130.61–2.1052951.080.70–1.67
 ≥20 (25)336300.700.37–1.26821330.590.43–0.82

Additional analyses were conducted stratifying by BMI and the number of pregnancies, respectively (Tables IV and V). Among both leaner and overweight women, OC users were at lower risk of endometrial cancer than never users. However, the inverse association between OC use and endometrial cancer risk was slightly stronger (OR = 0.76, 95% CI = 0.57–1.01) among leaner women. The reduced risk associated with OC use was more evident among women with more than 2 pregnancies than among women with 2 or fewer pregnancies. However, none of the multiplicative interaction tests were significant. Passive smoking was associated with slightly reduced risk of endometrial cancer in our data (OR = 0.92, 95% CI = 0.76–1.12). We found no interaction between OC use and passive smoking on endometrial cancer risk (data not shown). Owing to the small number of active smokers, nulliparous women, and women who used hormone replacement therapy after menopause in our study, we were hampered in our efforts to evaluate the association of OC use with endometrial cancer risk in these specific subgroups.

Table IV. Effects of OC Use on the Risk of Endometrial Cancer Stratified on BMI
 BMI <25BMI ≥25
CasesControlsOR195% CICasesControlsOR195% CI
  • 1

    Logistic regression adjusted for age, education level, family history of colorectal, endometrial, and breast cancer among first-degree relatives, age at menarche, years of menstruation, menopause status, number of pregnancies, age at last live birth, physical activity, exogenous hormone use (hormone replacement therapy, exogenous hormone use for other reasons).

  • 2

    Trend test based on exposed participants.

OC use
 Never4786041.0 5033061.0 
 Ever1051970.760.57–1.011181050.860.62–1.20
Duration of OC use (months)
 Never4786041.0 5033061.0 
 <2460870.830.57–1.1971421.000.65–1.54
 ≥24451100.680.45–1.0347630.720.46–1.15
   ptrend2 = 0.75  ptrend2 = 0.41
Age at first OC use (year)
 Never4786041.0 5033061.0 
 <30511130.590.41–0.8667530.870.57–1.32
 ≥3054841.030.69–1.5351520.860.54–1.36
Years since last OC use
 Never4786041.0 5033061.0 
 1–1942681.130.72–1.7730251.240.65–2.35
 ≥20601240.610.43–0.8786790.770.53–1.11
Table V. Effects of OC Use on the Risk of Endometrial Cancer Stratified on Number of Pregnancies
 1–2>2
CasesControlsOR195% CICasesControlsOR195% CI
  • 1

    Logistic regression adjusted for age, education level, body mass index (BMI), family history of colorectal, endometrial, and breast cancer among first-degree relatives, age at menarche, years of menstruation, menopause status, age at last live birth, physical activity, exogenous hormone use (hormone replacement therapy, exogenous hormone use for other reasons).

  • 2

    Trend test based on exposed participants.

  • 3

    Cut point was 25 years since last OC use among women having more than 2 pregnancies.

OC use
 Never4224021.0 4734651.0 
 Ever84801.010.69–1.491362210.650.49–0.86
Duration of OC use (months)
 Never4224021.0 4734651.0 
 <2450311.270.77–2.1079980.660.46–0.94
 ≥2434490.750.43–1.33571230.630.42–0.94
   ptrend2 = 0.51  ptrend2 = 0.65
Age at first OC use (year)
 Never4224021.0 4734651.0 
 <3039450.750.45–1.24771210.600.42–0.85
 ≥3045351.410.83–2.39591000.720.49–1.07
Years since last OC use
 Never4224021.0 4734651.0 
 1–19 (24)37331.340.76–2.33621000.900.60–1.35
 ≥20 (25)345450.800.49–1.32711170.520.36–0.74

Discussion

This large, population-based, case–control study found that reduced risk of endometrial cancer was associated with ever-use of OCs and long duration (72 months or more) of OC use. This association was independent of other major risk factors, including age at menarche, menopausal status, years of menstruation, BMI, number of pregnancies and so on. We found that the inverse association between OC use and endometrial cancer risk remained 25 or more years after cessation of use. Similarly, IUD use was also associated with a reduced risk of endometrial cancer and the effect appeared to be long lasting.

Our results are consistent with most prior investigations that risk of endometrial cancer is reduced with OC use, particularly among long-term users of OCs.1, 2, 3, 4, 5, 6, 7, 8, 9, 10 To our knowledge, the association of OC use and endometrial cancer risk among Chinese women has only been evaluated in a smaller scale study (268 cases and 268 controls) that we conducted in Shanghai; and a nonsignificant inverse association was reported (OR = 0.8, 95% CI = 0.4–1.8).11 Few previous studies have reported an inverse association between OC use and the risk of endometrial cancer among postmenopausal women, and results have been mixed regarding the effect of duration of OC use.1, 4, 5 In our study, 20.9% of postmenopausal cases and 31.5% of postmenopausal controls reported ever using OCs during their reproductive years. Among postmenopausal women, we found that previous OC use was consistently associated with reduced endometrial cancer risk, and long-term (≥24 months) use conferred about a 42% reduction in risk.

Few studies have evaluated the association between duration since discontinuing OC use and endometrial cancer risk.4, 5 Jick et al. reported that the relative risk of endometrial cancer was 0.62 (95% CI = 0.19–2.07) for women who had stopped use of OCs more than 21 years previously as compared to never users.4 In a multicenter, case–control study, the reduced odds ratios observed in OC users lessened with increasing time after the cessation of use (OR = 0.40, 95% CI = 0.2–0.8 for 15–19 years since last OC use, OR = 0.67, 95% CI = 0.4–1.3 for 20 or more years since last OC use, as compared to nonusers).5 In our data, the inverse association with OC use was found to last for more than 25 years, especially among postmenopausal women.

The inverse association between OC use and endometrial cancer may be due to the progestin component of OCs, which acts to reduce the proliferation of the endometrium. Combined oral hormonal contraceptives (COC) were introduced in China in the 1960s and low-dose oral hormonal contraceptives, such as norethisterone with ethinyl estradiol combined pill and the megestrol combined pill, were introduced in 1967.20 Although specific information regarding dosage of OCs was unavailable, most women (89.1% cases and 81% controls) started to use OCs after 1967; and over 77% of our participants reported using COCs. A single combined OC formulation (OC No. 1), which contains 0.625 mg of norethindrone and 35 μg ethinyl estradiol, was the most commonly used COC in our study population (over 85% of all COC users). Further analyses restricted to subjects using OC no. 1 showed no substantial change in the observed OC use and endometrial cancer association. The composition of OCs may have varied over time,21, 22 but we found no association between calendar year of beginning OC use and endometrial cancer.

Our finding of an inverse association between IUD use and endometrial cancer is, in general, consistent with the literature.14, 15, 16, 17, 18 The reduction in endometrial cancer risk does not appear to depend on duration of use, age at first and last use, or time since last use, which is also consistent with the results of previous studies.14, 15, 16, 17, 18 It has been suggested that IUD use may induce the inflammatory actions that eliminate abnormal and precancerous endometrial epithelial cells;23 decrease hyperplasia of the endometrium, a known risk factor for endometrial cancer;15 and reduce the concentration of estrogen receptors.24 Hormonal IUD, which contains progesterone or levonorgestrel, has now been introduced in many countries. Studies are needed to evaluate its effects on endometrial cancer.25 Use of injectable contraception, with hydroxyprogesterone as the most common formulation, was not very common in our study population and was often used only for short-term periods. We found no association between use of injectable contraception and risk of endometrial cancer, although one previous study did show that use of depot medroxyprogesterone acetate (DMPA) contraceptive was associated with reduction in endometrial cancer risk (OR = 0.21, 95% CI = 0.06–0.79).26

The strengths of our study include the population-based study design, large sample size, and detailed information on a wide range of potential confounders. However, several limitations should be mentioned. Owing to the nature of self-reported OC exposure, misclassification errors are unavoidable, which might have biased the OC use and endometrial cancer association. While nondifferential misclassification is likely to bias the result towards the null, differential misclassification can be biased in either direction. The median (25th–75th percentile) interval between diagnosis and interview for cases was 5.6 (3.4–9.2) months, which minimized the influence of disease diagnosis on recall of usual OC use. With 9.3% of cases and 20.9% of controls refusing to participate in the study, selection bias could be a concern. However, the frequency of OC use among controls in this study was similar to that found in a previous study of breast cancer that we conducted in Shanghai, which had a much higher (92%) response rate.27

In summary, this population-based, case–control study of the risk of endometrial cancer found an inverse association between OC use and endometrial cancer risk. Further studies are needed to investigate the biochemical responsible for this inverse association and to clarify the biological mechanisms involved.

Acknowledgements

We thank Dr. Fan Jin for her contributions in implementing the study in Shanghai and Ms. Bethanie Hull for technical support in the preparation of the manuscript. This study would not have been possible without the support of all of the study participants and research staff of the Shanghai Endometrial Cancer Study.

Ancillary