Abstract
Purpose
Few studies have examined methods of contraception, beyond oral contraceptives (OCs) and tubal ligation, in relation to ovarian cancer risk.
Methods
Nine hundred two cases with incident ovarian/peritoneal/tubal cancer were compared to 1800 population-based controls. Women self-reported all methods of contraception using life calendars.
Results
Each of the contraceptive methods examined reduced the risk of ovarian cancer as compared to use of no artificial contraception. Comparing ever versus never use, after adjustment for potentially confounding factors and all other methods of contraception, the methods of contraception that emerged as protective were OCs (adj OR 0.75, 95% CI 0.61–0.93); tubal ligation (adj OR 0.63, 95% CI 0.51–0.77); intrauterine devices (IUDs) (adj OR 0.75, 95% CI 0.59–0.95); and vasectomy (adj OR 0.77, 95% CI 0.61–0.99). While for oral contraceptives and tubal ligation, the longer the duration of use, the greater the effect, for IUDs the pattern was reversed: significant protection occurred with short duration and progressively greater risk (albeit non-significant) was seen with longer duration of use.
Conclusions
In the largest case-control study to date, a range of effective methods of contraception reduced the risk for ovarian cancer. OCs and tubal ligation reduced ovarian cancer risk with lower odds ratios with longer duration of use, whereas IUDs reduced risk overall, having the greatest impact with short duration of use.
Keywords: contraception, contraceptive methods, oral contraceptives, tubal ligation, IUDs, ovarian cancer
Several forms of contraception have been shown to reduce the risk of developing ovarian cancer. Oral contraceptives reduce risk in a duration-dependent fashion and the effects of oral contraceptives last for at least 20 years after cessation of use (1–4). Tubal ligation has also been shown to consistently reduce risk (5–8). Increasingly, oral contraceptives (OCs) are considered for chemoprophylaxis against ovarian cancer, particularly in high risk women (9–11).
Few studies have examined use of other methods of contraception in relation to ovarian cancer risk. Small case-control studies demonstrated some risk reduction with non-hormonal contraceptive methods but small numbers of women used each method, the findings were not entirely consistent by method, and risk reductions were not significant (1, 12–14). Two larger studies reached somewhat conflicting conclusions. In a large, case-control study by our group, all methods of contraception (intrauterine devices or IUDs, barrier methods, and vasectomy, as well as OCs and tubal ligation) reduced ovarian cancer risk as compared to no contraception use; ever use versus never use also reduced risk in multiparous but not nulliparous women (15). An analysis from the prospective Nurses’ Health Study cohort reported an increased risk associated with IUD use and no association for other contraceptive methods (16).
The finding that multiple contraceptive methods reduce ovary cancer risk must be scrutinized as possibly representing a bias by fertility status. Ovarian cancer rates are higher among infertile women (17, 18). In turn, many infertile women spend long periods of time practicing unprotected intercourse. Contraceptive users may thus appear to be protected from ovarian cancer only because they are less likely to be infertile. Alternatively, the finding that methods of contraception beyond OCs and IUDs are protective, may provide insights into ovarian carcinogenesis.
Here we attempt to re-examine the results from our earlier case-control study in a newly conducted population-based case-control design. Our earlier study was conducted in the Delaware Valley in and around Philadelphia and our current study was conducted in Western Pennsylvania and surrounding regions (15). As we did previously, we attempt to separate parity from contraceptive use and to examine the specificity of contraceptive effects on risk reduction for ovarian cancer. Here we examine OCs, IUDs, any barrier methods, tubal ligation, and vasectomy (in a partner) in relation to ovarian cancer risk.
Subjects and Methods
Subjects were enrolled in a population-based case-control study conducted in a contiguous region comprising Western Pennsylvania, Eastern Ohio, and Western New York State. Cases were residents of this geographic region with histologically confirmed primary epithelial ovarian, fallopian tube, or peritoneal cancer diagnosed between February 2003 and November 2008. Both invasive and borderline tumors were included. Women were referred from hospital tumor registries, clinical practices, or pathology databases and contacted with the permission of their gynecologists. Eligible women were at least 25 years of age and within 9 months of initial diagnosis. Controls consisted of women at least age 25 who lived in telephone exchanges wherein cases resided. Random digit dialing was used to identify age-eligible women, and these were further screened by the study team to ensure that they had not had a previous oophorectomy or diagnosis of ovarian cancer. Eligible women were then invited to participate. Potential controls were frequency matched by 5-year age group and telephone exchange to cases in a ratio of approximately 2:1. Women were interviewed in their homes by trained interviewers. The questionnaire included a reproductive and gynecological history, a contraceptive history, a medical history, a family history, and information on lifestyle practices. All study subjects gave informed consent for participation.
From Pennsylvania and Ohio, we identified 2458 potential cases with histologically confirmed borderline or invasive epithelial ovarian cancer or tubal/peritoneal cancer. After excluding women who were ineligible on the basis of age and time since diagnosis; deceased; residence outside of the counties in which referral hospitals were located; prior diagnosis of ovarian cancer; or inability to speak English, there were 997 who had incident cancer and were eligible for the study. Two hundred thirty one women were untraceable and 115 women refused to participate or their physicians refused on their behalf. Thus, 651 women completed case interviews. From New York, we identified 420 potential cases. After excluding women who were ineligible based on the criteria cited above, there were 273 who had incident cancer and were eligible for the study. Fourteen women were untraceable and eight women refused to participate or their physicians refused on their behalf, resulting in a sample of 251 women who completed interviews.
Overall, 902 women with ovarian, tubal/peritoneal cancer completed an interview and are included in these analyses. For brevity, we subsequently use the term ovarian cancer to describe all cases.
Controls were identified from 90,540 random digit dialing calls. Of these, 46,752 reached non-working numbers; 26,237 were unresolved (never reached a person); 14,899 reached an ineligible or indeterminate household (no woman within the age range or no information given); and 808 refused to participate. Of the 1844 eligible women willing to be interviewed in the initial screening, 1802 controls completed an interview. Two controls had an oophorectomy prior to the interview and were further excluded from our study, and 1800 controls completed an interview.
Cases included 677 women with invasive epithelial ovarian tumors, 97 with borderline epithelial ovarian tumors, 75 with peritoneal tumors, 32 with fallopian tumors, and 21 women with “other” or a missing type. The diagnosis of ovarian cancer was confirmed by local pathology in all cases.
CONTRACEPTIVE USE
Standardized 2-hour-long interviews were conducted by trained interviewers in the homes of participating women. A “life” calendar marked with important events that each participant recalled during her life was used to enhance memory of distant information. Using the calendar, the interviewer led each woman through a recall of her sexual activity, use of various contraceptive methods, pregnancy attempts, and reproductive events, for every month from sexual debut until a reference date. The reference date was calculated as 9 months before the interview (for both cases and controls) to ensure that exposures occurred before ovarian cancer diagnoses in cases and within a similar time frame for cases and controls. All contraceptive use was recorded, including the type of contraception, frequency of use, duration of use, and reason for use. Finally, we asked women about any medical consultation for fertility problems.
COVARIATES
Detailed information on demographic factors, physical characteristics, medical history, life-style, and family history was obtained by interview. These included factors that have been previously associated with ovarian cancer: race, education, family history of ovarian cancer, number of live births and pregnancies, and breast-feeding.
STATISTICAL ANALYSIS
All analyses were restricted to women who had ever had sexual intercourse with a man. Thirty three cases and 23 controls that had never had intercourse were excluded on the basis that they would not have had the opportunity for exposure to contraceptive methods for contraception. Because we did not engage in individual matching of cases and controls, we used unconditional logistic regression analyses. We adjusted the ORs for any residual effect of age and for gravidity (each as continuous variables), race (white/black/other), self-reported infertility (yes/no for diagnosis or use of infertility medications) and history of ovarian cancer in any first-degree relative (yes/no). We included these covariates because they were the strongest covariates related to ovarian cancer in our data. The inclusion of education and breast-feeding did not affect our results. We subsequently adjusted for all other forms of contraception other than the one of interest (eg. for OCs, this analysis included the covariates listed above plus ever use of IUD, tubal ligation and vasectomy). Oral contraceptive use was categorized as use for contraception, for non-contraceptive uses such as to control abnormal bleeding or menstrual pain, or for both contraception and other uses. Barrier methods included condoms, diaphragms, foam, sponges, or cervical caps. The reference group of no contraception included women who reported never using OCs, birth control implants, IUDs, any barrier methods, tubal ligation, or vasectomy (in a partner). These women may have used natural family planning (that is, having intercourse during times when the woman believed she was not ovulating), withdrawal, or nothing. We do not report as a separate category of contraception birth control implants, because the number of women using these methods in our study was small (16 cases and 46 controls).
Results
Study subjects were predominantly white, post-high school graduates, 60 or older, and post-menopausal (Table 1). The commonly-found protections with increasing education, numbers of pregnancies/live births and breast-feeding were observed. Cases were more likely to be African-American than controls, suggesting a selection bias among this small segment of subjects.
Table 1.
Demographic and Reproductive Characteristics of Ovarian Cancer Cases and Controls in the HOPE Study
| Cases (n=869) | Controls (n=1779) | Crude OR (95% CI) | Adjusted OR* (95% CI) | |
|---|---|---|---|---|
| Race | ||||
| White | 823 (95) | 1734 (96) | 1.00 | - |
| Black | 35 (4) | 29 (3) | 2.55 (1.55–4.19) | - |
| Other | 11 (1) | 16 (1) | 1.45 (0.67–3.14) | - |
| Education | ||||
| Not high school graduate | 80 (9) | 82 (5) | 1.00 | 1.00 |
| High school graduate | 294 (34) | 527 (30) | 0.58 (0.41–0.82) | 0.53 (0.37–0.76) |
| Post- high school | 495 (57) | 1170 (66) | 0.45 (0.33–0.63) | 0.40 (0.28–0.56) |
| Age (years) | - | |||
| ≤ 30 | 13 (1) | 24 (1) | 1.00 | - |
| 30–< 40 | 43 (5) | 107 (6) | 0.74 (0.35–1.59) | - |
| 40–< 50 | 155 (18) | 391 (22) | 0.73 (0.36–1.47) | - |
| 50–< 60 | 270 (31) | 563 (32) | 0.88 (0.44–1.76) | - |
| ≥ 60 | 388 (45) | 694 (39) | 1.03 (0.52–2.05) | - |
| Number of Pregnancies | ||||
| 0 | 135 (16) | 144 (8) | 1.00 | - |
| 1 | 114 (13) | 188 (11) | 0.65 (0.46–0.90) | - |
| 2 | 216 (25) | 458 (26) | 0.50 (0.38–0.67) | - |
| 3 | 167 (19) | 426 (24) | 0.42 (0.31–0.56) | - |
| 4 | 113 (13) | 284 (16) | 0.42 (0.30–0.59) | - |
| ≥5 | 124 (14) | 279 (16) | 0.47 (0.34–0.65) | - |
| Number of Live Births | ||||
| 0 | 181 (21) | 207 (12) | 1.00 | 1.00 |
| 1 | 117 (13) | 228 (13) | 0.59 (0.43–0.79) | 0.63 (0.45–0.86) |
| 2 | 263 (31) | 593 (34) | 0.50 (0.39–0.65) | 0.59 (0.43–0.80) |
| 3 | 170 (19) | 418 (23) | 0.47 (0.35–0.61) | 0.56 (0.39–0.82) |
| 4 | 73 (8) | 190 (11) | 0.44 (0.31–0.62) | 0.54 (0.33–0.88) |
| ≥5 | 65 (7) | 143 (8) | 0.52 (0.36–0.74) | 0.72 (0.38–1.37) |
| Body Mass Index (kg/m2) | ||||
| ≤18.5 | 8 (1) | 21 (1) | 1.00 | 1.00 |
| 18.5–<25.0 | 289 (32) | 645 (37) | 1.18 (0.52–2.69) | 1.03 (0.44–2.41) |
| 25–< 30 | 255 (30) | 522 (29) | 1.28 (0.56–2.93) | 1.11 (0.48–2.60) |
| ≥30 | 317 (37) | 591 (33) | 1.41 (0.62–3.22) | 1.20 (0.52–2.81) |
| Age at menarche (%) | ||||
| ≤11 | 174 (20) | 439 (25) | 1.00 | 1.00 |
| 12 | 247 (29) | 458 (26) | 1.36 (1.07–1.71) | 1.33 (1.04–1.67) |
| 13 | 235 (27) | 478 (27) | 1.24 (0.98–1.56) | 1.26 (0.98–1.59) |
| ≥14 | 213 (24) | 404 (23) | 1.33 (1.04–1.69) | 1.28 (1.00–1.64) |
| Menopausal at interview | ||||
| No | 259 (30) | 562 (32) | 1.00 | 1.00 |
| Yes | 590 (68) | 1160 (65) | 1.10 (0.93–1.32) | 0.81 (0.63–1.05) |
| Unknown | 20 (2) | 56 (3) | 0.78 (0.46–1.32) | 0.74 (0.43–1.26) |
| Hysterectomy | ||||
| No | 689 (80) | 1472 (83) | 1.00 | 1.00 |
| Yes | 180 (20) | 306 (17) | 1.26 (1.02–1.53) | 1.24 (0.99–1.53) |
| Tubal ligation | ||||
| No | 666 (77) | 1162 (65) | 1.00 | 1.00 |
| Yes | 201 (23) | 616 (35) | 0.57 (0.47–0.69) | 0.62 (0.51–0.75) |
| Fibroid Tumors | ||||
| No | 673 (77) | 1451 (82) | 1.00 | 1.00 |
| Yes | 196 (23) | 327 (18) | 1.29 (1.06–1.58) | 1.34 (1.02–1.53) |
| Ever breastfed | ||||
| No | 565(65) | 893 (50) | 1.00 | 1.00 |
| Yes | 304 (35) | 886 (50) | 0.54 (0.46–0.64) | 0.62 (0.52–0.74) |
| Oral Contraceptive Use | ||||
| No | 340 (37) | 512 (28) | 1.00 | 1.00 |
| Yes | 529 (63) | 1267 (72) | 0.64 (0.54–0.75) | 0.67 (0.55–0.81) |
| Family history of ovarian cancer | ||||
| No | 833 (96) | 1723 (97) | 1.00 | - |
| Yes | 36 (4) | 56 (3) | 1.33 (0.87–2.03) | - |
| Family history of breast cancer | ||||
| No | 718 (83) | 1522 (86) | 1.00 | 1.00 |
| Yes | 151 (17) | 257 (14) | 1.25 (1.00–1.55) | 1.23 (1.00–1.53) |
| Family history of endometrial cancer | ||||
| No | 843 (97) | 1726 (97) | 1.00 | 1.00 |
| Yes | 26 (3) | 53 (3) | 1.00 (0.62–1.62) | 0.99 (0.61–1.60) |
| Smoking Status | ||||
| Never | 432 (50) | 893 (51) | 1.00 | 1.00 |
| Former | 282 (32) | 543 (30) | 1.07 (0.89–1.29) | 1.09 (0.90–1.31) |
| Current | 155 (18) | 343 (19) | 0.93 (0.75–1.16) | 0.96 (0.77–1.21) |
Adjusted for age, number of pregnancies, race, and family history of ovarian cancer. Adjusted odds ratios were not calculated for age, number of pregnancies, family history of ovarian cancer, and race because these variables were used in the adjustment.
OR, odds ratio; CI, confidence interval
We found a reduction in the risk of ovarian cancer for ever versus never use of each of the methods for contraception analyzed (Table 2). However, after adjustment for age, race, family history of ovarian cancer, infertility, and gravidity, significant protection was seen only with IUDs as well as OCs for contraception and tubal ligation. After further adjustment for all other forms of effective contraception, IUDs, OCs for contraception, and tubal ligation remained significantly protective; now vasectomy also reached a level of significant protection.
Table 2.
Odds Ratios for Ovarian Cancer Comparison of Ever-Use of Contraceptive Methods with Never-Use
| Variable | Cases | Controls | Crude OR | 95% CI | Adjusted OR* | 95% CI | Adjusted OR** | 95% CI |
|---|---|---|---|---|---|---|---|---|
| OC use | ||||||||
| Neither | 340 (39) | 512 (29) | 1.00 | 1.00 | 1.00 | |||
| OCs for contraception | 370 (43) | 880 (49) | 0.63 | 0.53–0.76 | 0.66 | 0.54–0.81 | 0.75 | 0.61–0.93 |
| OCs for non-contraception | 53 (6) | 108 (6) | 0.74 | 0.52–1.06 | 0.73 | 0.51–1.06 | 0.76 | 0.53–1.11 |
| OCs for both# | 106 (12) | 279 (16) | 0.57 | 0.44–0.74 | 0.58 | 0.44–0.78 | 0.63 | 0.47–0.85 |
| IUDs | ||||||||
| Never | 763 (88) | 1479 (83) | 1.00 | 1.00 | 1.00 | |||
| Ever | 104 (12) | 299 (17) | 0.67 | 0.53–0.86 | 0.69 | 0.54–0.89 | 0.75 | 0.59–0.95 |
| Barrier | ||||||||
| Never | 305 (35) | 570 (32) | 1.00 | 1.00 | 1.00 | |||
| Ever | 564 (65) | 1209 (68) | 0.87 | 0.74–1.03 | 0.94 | 0.79–1.13 | 0.95 | 0.79–1.13 |
| Tubal ligation | ||||||||
| Never | 666 (77) | 1162 (65) | 1.00 | 1.00 | 1.00 | |||
| Ever | 201 (23) | 616 (35) | 0.56 | 0.47–0.69 | 0.62 | 0.51–0.75 | 0.63 | 0.51–0.77 |
| Vasectomy | ||||||||
| Never | 746 (86) | 1468 (83) | 1.00 | 1.00 | 1.00 | |||
| Ever | 121 (14) | 309 (17) | 0.77 | 0.61–0.97 | 0.84 | 0.67–1.06 | 0.77 | 0.61–0.99 |
Adjusted for age, number of pregnancies, race, infertility and family history of ovarian cancer
Adjusted for age, number of pregnancies, race, infertility, family history of ovarian cancer, ever use of oral contraceptives, ever use of IUDs, ever use of barriers, tubal ligation, and vasectomy.
Both refers to the use of oral contraceptives for both contraception and non-contraception
Ever barrier use is defined as ever use of condoms, diaphragms, foam, sponges, or cervical caps.
OC, Oral Contraceptive
As ever use of contraceptive methods is complicated by admixing users of other methods, mixed methods, and none of these methods over a lifetime, we also compared users of each method with women who used no artificial contraception, defined as use of only natural family planning, withdrawal, rhythm, or no contraception (Table 3). Each of the methods significantly reduced the risk of ovarian cancer as compared to no artificial contraception.
Table 3.
Odds Ratios for Ovarian Cancer: Comparison of Ever Use of Contraceptive Methods with No Artificial Contraception
| Variable | Cases | Controls | Crude OR | 95% CI | Adjusted OR* | 95% CI |
|---|---|---|---|---|---|---|
| OCs | ||||||
| No contraception† | 111 | 119 | 1.00 | 1.00 | ||
| OCs for contraception | 370 | 880 | 0.45 | 0.34–0.60 | 0.52 | 0.36–0.72 |
| OCs for non-contraception | 106 | 279 | 0.41 | 0.29–0.57 | 0.35 | 0.22–0.56 |
| Intrauterine device | ||||||
| No contraception† | 111 | 119 | 1.00 | 1.00 | ||
| IUD | 104 | 299 | 0.37 | 0.27–0.53 | 0.41 | 0.27–0.62 |
| Barrier | ||||||
| No contraception† | 111 | 119 | 1.00 | 1.00 | ||
| Barrier | 564 | 1209 | 0.50 | 0.38–0.66 | 0.58 | 0.43–0.79 |
| Tubal ligation | ||||||
| No contraception† | 111 | 119 | 1.00 | 1.00 | ||
| Tubal ligation | 201 | 616 | 0.35 | 0.26–0.47 | 0.35 | 0.25–0.51 |
| Vasectomy | ||||||
| No contraception† | 111 | 119 | 1.00 | 1.00 | ||
| Vasectomy | 121 | 309 | 0.42 | 0.30–0.59 | 0.48 | 0.31–0.79 |
Reference category
Adjusted for age, pregnancies, race, infertility and family history of ovarian cancer
No contraception refers to women who reported never using any contraceptive method (OC, IUD, barrier, tubal ligation, vasectomy, patch, shot, or cream); such women might have used family planning, withdrawal, or nothing.
Next, we examined the association between contraception and ovarian cancer among women with zero, one, two or three or more pregnancies (Table 4). Both OCs for contraception and tubal ligation significantly reduced risk in some but not all gravidity categories, without a clear pattern of greater or lesser effects as gravidity increased. IUD use, despite generating protective odds ratios similar to those for OCs, did not produce significant results in any gravidity category, possibly due to small cell sizes. Vasectomy also did not produce significant reductions in risk in any gravidity category.
Table 4.
Adjusted Odds Ratios for Ovarian Cancer: Comparison of Ever-Use of Contraceptive Methods with Never-Use by Gravidity Group
| Gravidity= 0 | Gravidity = 1 | Gravidity = 2 | Gravidity ≥ 3 | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | Case | Control | Adj. OR* | 95% CI | Case | Control | Adj. OR* | 95% CI | Case | Control | Adj. OR* | 95% CI | Case | Control | Adj. OR* | 95% CI |
| OCs | ||||||||||||||||
| Neither | 58 | 43 | 1.00 | - | 38 | 60 | 1.00 | - | 79 | 100 | 1.00 | - | 165 | 308 | 1.00 | - |
| For contraception | 43 | 53 | 0.46 | 0.25–0.86 | 50 | 77 | 0.99 | 0.58–2.02 | 98 | 264 | 0.51 | 0.34–0.77 | 179 | 486 | 0.85 | 0.64–1.14 |
| For non-contraception | 12 | 13 | 0.61 | 0.25–1.52 | 6 | 16 | 0.60 | 0.44–2.23 | 12 | 18 | 0.89 | 0.40–1.99 | 22 | 61 | 0.77 | 0.45–1.32 |
| For both# | 21 | 34 | 0.31 | 0.15–0.67 | 20 | 35 | 0.99 | 0.22–1.69 | 27 | 76 | 0.50 | 0.28–0.88 | 38 | 134 | 0.70 | 0.45–1.09 |
| IUDs | ||||||||||||||||
| Never | 129 | 136 | 1.00 | - | 103 | 161 | 1.00 | - | 194 | 386 | 1.00 | - | 336 | 796 | 1.00 | - |
| Ever | 3 | 7 | 0.32 | 0.07–1.56 | 11 | 27 | 0.67 | 0.32–1.40 | 22 | 72 | 0.59 | 0.36–0.99 | 68 | 193 | 0.84 | 0.62–1.16 |
| Tubal ligation | ||||||||||||||||
| Never | 126 | 128 | 1.00 | - | 102 | 151 | 1.00 | - | 165 | 300 | 1.00 | - | 273 | 583 | 1.00 | - |
| Ever | 6 | 15 | 0.36 | 0.13–1.05 | 12 | 37 | 0.49 | 0.24–1.00 | 51 | 158 | 0.60 | 0.41–0.87 | 131 | 406 | 0.75 | 0.58–0.97 |
| Barrier | ||||||||||||||||
| Never | 52 | 60 | 1.00 | - | 40 | 73 | 1.00 | - | 78 | 135 | 1.00 | - | 134 | 301 | 1.00 | - |
| Ever | 82 | 83 | 1.09 | 0.65–1.82 | 74 | 115 | 1.23 | 0.74–2.03 | 138 | 323 | 0.77 | 0.55–1.10 | 270 | 688 | 0.96 | 0.74–1.23 |
| Vasectomy | ||||||||||||||||
| Never | 120 | 133 | 1.00 | 100 | 165 | 1.00 | 179 | 359 | 1.00 | 346 | 811 | 1.00 | ||||
| Ever | 12 | 10 | 1.38 | 0.56–3.37 | 14 | 22 | 1.14 | 0.55–2.35 | 37 | 99 | 0.81 | 0.53–1.23 | 58 | 178 | 0.89 | 0.64–1.24 |
Adjusted for age, race, family history of ovarian cancer, and infertility
Both refers to the use of oral contraceptives for both contraception and non-contraception
By duration, OCs had a progressively greater impact with 4 or fewer years, 5–9 years and 10 or more years of use (Table 5). Similarly, longer duration of tubal ligation was associated with lower risk. For IUD use, the pattern was reversed: significant protection occurred with short duration (≤ 4 years) use and progressively greater risk was seen with longer duration of use (adj ORs 0.53 for ≤4 years; 1.11 for 5–9 years; 1.40 for ≥10 years). We further explored whether time since last IUD use might drive these observations. Although there was a trend toward reduced risk with increasing time since last use, this was eliminated with adjustment for (i.e. not independent of) IUD duration. We did not have data on duration of vasectomy.
Table 5.
Adjusted Odds Ratios and 95% Confidence Intervals for Ovarian Cancer Comparison of Duration of Use of Contraceptive Methods with Never Use
| Duration |
||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| ≤ 4 years | 5–9 years | ≥ 10 years | ||||||||||
| Cases | Controls | OR* | 95% CI | Cases | Controls | OR* | 95% CI | Cases | Controls | OR* | 95% CI | |
| OCs for contraception | 255 (29) | 555 (31) | 0.91 | 0.75–1.10 | 78 (9) | 194 (11) | 0.78 | 0.59–1.05 | 37 (4) | 131 (7) | 0.52 | 0.35–0.76 |
| OCs for non-contraception | 44 (5) | 93 (5) | 0.93 | 0.64–1.36 | 7 (1) | 8 (0) | 1.60 | 0.58–4.47 | 2 (0) | 7 (0) | 0.53 | 0.11–2.62 |
| OCs for both | 57 (7) | 99 (6) | 1.22 | 0.87–1.73 | 29 (3) | 84 (5) | 0.72 | 0.46–1.12 | 20 (2) | 96 (5) | 0.40 | 0.25–0.67 |
| IUDs # | 58 (7) | 228 (12) | 0.53 | 0.39–0.72 | 20 (2) | 34 (2) | 1.11 | 0.63–1.96 | 25 (3) | 33 (2) | 1.40 | 0.82–2.39 |
Adjusted for age, pregnancies, race, family history of ovarian cancer, and infertility.
Data on duration of use of IUDs was unavailable for 4 women.
IUD, intrauterine device
OCs, Oral Contraceptives
Additional analyses showed our observations to be robust. Contraception use before the first pregnancy or episode of trying was protective (OCs for contraception, OR = 0.87 (0.69–1.11); IUD, OR = 0.81 (0.41–1.60)). Adjustment for breast-feeding in the multivariate analyses had no substantial effect on our results for ever-never use. Analyses including only epithelial ovarian cancer (excluding fallopian and peritoneal cancers) essentially replicated those shown here with the result for vasectomy slightly enhanced in these analyses when adjusted for confounders plus all other contraceptive methods (OR 0.73, 95% CI 0.56–0.94).
Finally, we examined use of concomitant methods of birth control over a lifetime (Table 6). The majority of women used more than one contraceptive method over a lifetime and of these, the most common combination was OC use plus another method. For instance, of the 424 women whose contraceptive use included vasectomy, only 89 (21%) did not also use OCs.
Table 6.
Contraception Methods Used by Cases and Controls
| Contraceptive Method | Total | Cases | Controls | P |
|---|---|---|---|---|
| Never Users | 231 | |||
| Oral contraceptive only | 153 | 56 (6) | 97 (5) | 0.30 |
| IUD only | 26 | 8 (1) | 18 (1) | 0.82 |
| Tubal ligation only | 75 | 24 (3) | 51 (3) | 0.88 |
| Vasectomy only | 20 | 11 (1) | 9 (1) | 0.03 |
| Barrier only | 373 | 153 (18) | 220 (12) | 0.0003 |
| OC and IUD | 24 | 4 (0) | 20 (1) | 0.09 |
| OC and tubal ligation | 156 | 38 (4) | 118 (7) | 0.02 |
| OC and vasectomy | 54 | 17 (2) | 37 (2) | 0.83 |
| OC and barrier | 471 | 166 (19) | 305 (17) | 0.21 |
| IUD and tubal ligation | 24 | 4 (1) | 20 (1) | 0.09 |
| IUD and vasectomy | 3 | 1 (0) | 2 (0) | 0.98 |
| IUD and barrier | 28 | 12 (1) | 16 (1) | 0.25 |
| Tubal ligation and vasectomy | 0 | - | - | - |
| Tubal ligation and barrier | 92 | 24 (3) | 68 (4) | 0.16 |
| Vasectomy and barrier | 56 | 18 (2) | 38 (2) | 0.92 |
| OC, IUD and tubal ligation | 41 | 7 (1) | 34 (2) | 0.03 |
| OC, IUD and vasectomy | 12 | 3 (0) | 9 (1) | 0.56 |
| OC, IUD and barrier | 80 | 28 (3) | 52 (3) | 0.67 |
| OC, tubal ligation and vasectomy | 5 | 1 (0) | 4 (0) | 0.54 |
| OC, tubal ligation and barrier | 288 | 71 (8) | 217 (12) | 0.002 |
| OC, vasectomy and barrier | 198 | 48 (6) | 150 (8) | 0.008 |
| IUD, tubal ligation and vasectomy | 0 | - | -- | - |
| IUD, tubal ligation and barrier | 20 | 4 (0) | 16 (1) | 0.22 |
| IUD, vasectomy and barrier | 10 | 2 (0) | 8 (0) | 0.39 |
| Tubal ligation, vasectomy and barrier | 5 | 1 (0) | 4 (0) | 0.54 |
| OC, IUD, tubal ligation and vasectomy | 1 | - | 1 (0) | 0.48 |
| OC, IUD, tubal ligation and barrier | 86 | 18 (2) | 68 (4) | 0.02 |
| OC, IUD, vasectomy and barrier | 42 | 10 (1) | 32 (2) | 0.21 |
| OC, tubal ligation, vasectomy and barrier | 18 | 6 (1) | 12 (1) | 0.96 |
| IUD, tubal ligation, vasectomy and barrier | 0 | - | - | - |
| OC, IUD, tubal ligation, vasectomy and barrier | 6 | 3 (0) | 3 (0) | 0.37 |
Discussion
We report here the largest case-control study to examine whether effective methods of contraception, beyond OCs and tubal ligation, reduce the risk of ovarian cancer. Consistent with the results of our previous case-control analysis (15), we found that use of a variety of different contraceptive methods generally reduced risk of ovarian cancer as compared to use of no artificial contraception. Such an analysis is almost certainly confounded by fertility in that women who are infertile or sub-fertile would be less likely to use effective methods of contraception and more likely to develop ovarian cancer. In our current, more discriminating analyses of ever versus never use, comparisons to OCs, and use within parity categories, the methods of contraception that emerged as protective were OCs, tubal ligation, IUDs, and vasectomy. Vasectomy is intriguing but we were less informed about this relationship with ovarian cancer since we had no data on duration of use. IUDs were particularly interesting because 1) they are not traditionally thought to reduce the risk of ovarian cancer; 2) duration-response analyses showed a counter-intuitive pattern wherein shorter use reduced risk and longer use (albeit non-significantly) increased risk.
Our results replicated a plethora of earlier studies linking OCs and tubal ligation to reduced ovarian cancer risk (5–7, 19–20). In particular, our results mirror adjusted ORs for ever-use of OCs reported from a meta-analysis (0.7) and a pooled analysis (0.66) (2, 21). We partially, but did not fully, replicate a much more limited literature that has addressed the relation between other forms of contraception (barrier, IUD, or vasectomy) and the risk of ovarian cancer. In these studies, the reported ORs were generally less than 1.0; however, none of those studies had enough women in any contraception category, other than OCs, to show strong effects or to explore more fully comparisons between categories (1, 12–15, 22). In the only prospective study to examine contraception methods beyond OCs and tubal ligation, Tworoger et al. found a significant relative risk of 1.76 associated with IUD use (16). Unfortunately, only 18 IUD users informed the analysis and thus duration and time since last use of IUDs was not reported. Here, we did not show significant risk reductions for barrier methods and vasectomy but we did find that shorter-duration IUD use reduced risk while longer duration IUD use increased risk.
In previous analyses stratifying by parity or gravidity groups, results have been mixed. In our previous study, we found risk reductions to be limited to multigravid women (15). In our current study, we found a patchy set of associations that did not clearly demonstrate a limitation by gravidity category but was not fully consistent between gravidity categories, perhaps on the basis of the sizes of individual stratification cells. All methods were protective before the first pregnancy, a time during which women might not yet know their fertility potential and thus not yet adjust their contraceptive strategy. All of this suggests that confounding by fertility status is an unlikely explanation for our observations.
A variety of biological explanations have been offered to explain the protective effect of OCs on ovarian cancer risk. These include: (1) excessive ovulation promotes risk; (2) elevated steroid hormone levels increaser risk; (3) unopposed estrogen increases risk; and (4) pelvic inflammation increases risk (23–27). Tubal ligation has been posited to have an effect via a reduction in utero-ovarian blood flow resulting in altered local hormonal and growth factor levels, or via its protection against the ascension of inflammants (26–28). Some IUDs contain progestin, which has been proposed to reduce the risk of ovarian cancer (25). However, only a tiny fraction of IUD-users in the current analysis (N=14) reported using progestin-containing IUDs. IUDs, particularly older formulations, such as the Dalkon Shield, increased the risk for pelvic inflammatory disease. The hazard likely occurred because of the particular construction of the multifilament string attached to the Dalkon shield. But it also may have related to insertion through a cervix infected with the bacterial sexually transmitted infections that cause pelvic inflammatory disease, as suggested by the close temporal relationship between insertion and pelvic inflammatory disease, and the relative safety of modern-day use which is confined to women monogamous without cervical infections (29). These facts may explain our counterintuitive finding of a reversed duration-response relationship (longer use associated with increasing risk). IUDs must be replaced every 5–10 years depending on the product – longer use would imply more insertions and thus greater risk of infection and inflammation. Shorter use might actually reduce upper genital tract inflammation by virtue of killing sperm. Indeed, the marginal reduction in risk from vasectomy might suggest some protection from reduced exposure to sperm
Strengths of this study include the population-based ascertainment of cases and controls, the large number of women interviewed, the use of life calendars and emphasis on recall of contraceptive use and reproductive experiences, and the structured interviews to enhance recall. All of these methodological features reduce the potential for selection and information bias. The greatest study limitation was the challenge of separating out the effects of various contraceptive methods since use of more than one method over a lifetime was the norm. We attempted to separately delineate methods by adjusting for all methods in logistic regression analyses and by comparing each method to no effective method. Nonetheless, residual confounding remains a real concern. Other study limitations included: (1) selection against women with short life expectancies post-diagnosis who may have become debilitated or died prior to interview and (2) inaccurate recall of past contraceptive experiences. Women may have incorrectly recalled past events, such as the duration of use of contraceptive methods. It is less likely that women would misremember ever versus never-use of contraceptive methods (30, 31). Previous studies have found that among ever-users of OCs identified by medical records, 80% or more reported OC use; an even larger proportion of IUD users identified by medical records reported IUD use (30–35).
In summary, from this large study of contraceptive methods and ovarian cancer, we confirmed that OCs and tubal ligation reduced risk for ovarian cancer. Short-term IUD use reduced risk but long term IUD use tended toward elevating risk. Since contraceptive methods are modifiable and since ovarian cancer is highly lethal, these findings should be added to other considerations when selecting contraceptive methods.
Acknowledgments
Grant funding: R01CA095023
Abbreviations
- IUD
intrauterine device
- OC
oral contraceptive
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.The Cancer and Steroid Hormone Study Group. The reduction in risk of ovarian cancer associated with oral contraceptives use. N Engl J Med. 1987;316:650–655. doi: 10.1056/NEJM198703123161102. [DOI] [PubMed] [Google Scholar]
- 2.Whittemore AS, Harris R, Itnyre J the Collaborative Ovarian Cancer Group. Characteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control studies. II. Invasive epithelial ovarian cancers in white women. Am J Epidemiol. 1992;136:1184–1203. doi: 10.1093/oxfordjournals.aje.a116427. [DOI] [PubMed] [Google Scholar]
- 3.Purdie D, Green A, Bain C, Siskind V, Ward B, Hacker N, et al. Reproductive and other factors and risk of epithelial ovarian cancer: an Australian case-control study. Survey of Women’s Health Study Group. Int J Cancer. 1995;62:678–684. doi: 10.1002/ijc.2910620606. [DOI] [PubMed] [Google Scholar]
- 4.Ness RB, Grisso JA, Klapper J, Schlesselman JJ, Silberzweig S, Vergona R, et al. the SHARE Study Group. Risk of ovarian cancer in relation to estrogen and progestin dose and use characteristics of oral contraceptives. Am J Epidemiol. 2000;152:233–241. doi: 10.1093/aje/152.3.233. [DOI] [PubMed] [Google Scholar]
- 5.Hankinson SE, Hunter DJ, Colditz GA, Willett WC, Stampfer MJ, Rosner B, et al. Tubal ligation, hysterectomy, and risk of ovarian cancer: a prospective study. JAMA. 1993;270:2813–2818. [PubMed] [Google Scholar]
- 6.Miracle-McMahill HL, Calle EE, Kosinski AS, Rodriguez C, Wingo PA, Thun MJ, et al. Tubal ligation and fatal ovarian cancer in a large prospective cohort study. Am J Epidemiol. 1997;145:349–357. doi: 10.1093/oxfordjournals.aje.a009112. [DOI] [PubMed] [Google Scholar]
- 7.Rosenblatt KA, Thomas DB. Reduced risk of ovarian cancer in women with a tubal ligation or hysterectomy. The World Health Organization Collaborative Study of Neoplasia and Steroid Contraceptives. Cancer Epidemiol Biomarkers Prev. 1996;5:933–935. [PubMed] [Google Scholar]
- 8.Kjaer SK, Mellemkjaer L, Brinton LA, Johansen C, Gridley G, Olsen JH. Tubal sterilization and risk of ovarian, endometrial and cervical cancer. A Danish population-based follow-up study of more than 65 000 sterilized women. Int J Epidemiol. 2004;33:596–602. doi: 10.1093/ije/dyh046. [DOI] [PubMed] [Google Scholar]
- 9.Narod SA, Risch H, Moslehi R, Dorum A, Neuhausen S, Olsson H, et al. Oral contraceptives and the risk of hereditary ovarian cancer. N Engl J Med. 1998;339:424–428. doi: 10.1056/NEJM199808133390702. [DOI] [PubMed] [Google Scholar]
- 10.Antoniou AC, Rookus M, Andrieu N, Brohet R, Chang-Claude J, Peock S, et al. Reproductive and hormonal factors, and ovarian cancer risk for BRCA1 and BRCA2 mutation carriers: results from the International BRCA1/2 Carrier Cohort Study. Cancer Epidemiol Biomarkers Prev. 2009;18:601–610. doi: 10.1158/1055-9965.EPI-08-0546. [DOI] [PubMed] [Google Scholar]
- 11.Modugno F, Moslehi R, Ness RB, Nelson DB, Belle S, Kant JA, et al. Reproductive factors and ovarian cancer risk in Jewish BRCA1 and BRCA2 mutation carriers (United States) Cancer Causes Control. 2003;14:439–446. doi: 10.1023/a:1024932427503. [DOI] [PubMed] [Google Scholar]
- 12.Parazzini F, La Vecchia C, Negri E, Bocciolone L, Fedele L, Franceschi S. Oral contraceptive use and the risk of ovarian cancer: an Italian case-control study. Eur J Cancer. 1991;27:594–598. doi: 10.1016/0277-5379(91)90226-4. [DOI] [PubMed] [Google Scholar]
- 13.Booth M, Beral V, Smith P. Risk factors for ovarian cancer: a case-control study. Br J Cancer. 1989;60:592–598. doi: 10.1038/bjc.1989.320. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Salazar-Martinez E, Lazcano-Ponce EC, Gonzalez Lira-Lira G, Escudero-De los Rios P, Salmeron-Castro J, Hernandez-Avila M. Reproductive factors of ovarian and endometrial cancer risk in high fertility population in Mexico. Cancer Res. 1999;59:3658–3662. [PubMed] [Google Scholar]
- 15.Ness RB, Grisso JA, Vergona R, Klapper J, Morgan M, Wheeler JE for the Study of Health and Reproduction (SHARE) Study Group. Oral contraceptives, other methods of contraception and risk reduction for ovarian cancer. Epidemiology. 2001;12:307–312. doi: 10.1097/00001648-200105000-00010. [DOI] [PubMed] [Google Scholar]
- 16.Tworoger SS, Fairfield KM, Colditz GA, Rosner BA, Hankinson SE. Association of oral contraceptive use, other contraceptive methods, and infertility with ovarian cancer risk. Am J Epidemiol. 2007;166:894–901. doi: 10.1093/aje/kwm157. [DOI] [PubMed] [Google Scholar]
- 17.Glud E, Kjaer SK, Troisi R, Brinton LA. Fertility drugs and ovarian cancer. Epidemiol Rev. 1998;20:237–257. doi: 10.1093/oxfordjournals.epirev.a017983. [DOI] [PubMed] [Google Scholar]
- 18.Ness RB, Cramer DW, Goodman MT, Kjaer SK, Mallin K, Mosgaard BJ, et al. Infertility, fertility drugs, and ovarian cancer: a pooled analysis of case-control studies. Am J Epid. 2002;155:217–224. doi: 10.1093/aje/155.3.217. [DOI] [PubMed] [Google Scholar]
- 19.Hannaford PC, Sivasubramaniam S, Elliott AM, Angus V, Iversen L, Lee AJ. Cancer risk among users of oral contraceptives: cohort data from the Royal College of General Practitioner’s oral contraception study. BMJ. 2007:335–651. doi: 10.1136/bmj.39289.649410.55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Collaborative Group on Epidemiological Studies of Ovarian Cancer. Ovarian cancer and oral contraceptives: collaborative reanalysis of data from 45 epidemiological studies including 23,257 women with ovarian cancer and 87,303 controls. Lancet. 2008;371:303–314. doi: 10.1016/S0140-6736(08)60167-1. [DOI] [PubMed] [Google Scholar]
- 21.Stanford JL. Oral contraceptives and neoplasia of the ovary. Contraception. 1991;43:543–556. doi: 10.1016/0010-7824(91)90003-x. [DOI] [PubMed] [Google Scholar]
- 22.Dorjgochoo T, Shu XO, Li HL, Qian HZ, Yang G, Cai H, Gao YT, Zheng W. Use of oral contraceptives, intrauterine devices and tubal sterilization and cancer risk in a large prospective study, from 1996 to 2006. Int J Cancer. 2009;15;124(10):2442–2449. doi: 10.1002/ijc.24232. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Fathalla MF. Incessant ovulation – a factor in ovarian neoplasia? Lancet. 1971;2:163. doi: 10.1016/s0140-6736(71)92335-x. [DOI] [PubMed] [Google Scholar]
- 24.Cramer DW, Welch WR. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis. J Natl Cancer Inst. 1983;71:717–721. [PubMed] [Google Scholar]
- 25.Risch HA. Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of adrogens and progesterone. J Natl Cancer Inst. 1998;90:1774–1786. doi: 10.1093/jnci/90.23.1774. [DOI] [PubMed] [Google Scholar]
- 26.Ness RB, Cottreau C. Possible role of ovarian epithelial inflammation in ovarian cancer. J Natl Cancer Inst. 1999;91:1459–1467. doi: 10.1093/jnci/91.17.1459. [DOI] [PubMed] [Google Scholar]
- 27.Ness RB, Grisso JA, Cottreau C, Klapper J, Veragona R, Wheeler JE, et al. Factors related to inflammation of the ovarian epithelium and risk of ovarian cancer. Epidemiology. 2000;11:111–117. doi: 10.1097/00001648-200003000-00006. [DOI] [PubMed] [Google Scholar]
- 28.Cramer DW, Xu H. Epidemiologic evidence for uterine growth factors in the pathogenesis of ovarian cancer. Ann Epidemiol. 1995;5:310–314. doi: 10.1016/1047-2797(94)00098-e. [DOI] [PubMed] [Google Scholar]
- 29.MacIsaac L, Espey E. Intrauterine contraception: the pendulum swings back. Obstet Gynecol Clin North Am. 2007;34:91–111. doi: 10.1016/j.ogc.2007.02.004. [DOI] [PubMed] [Google Scholar]
- 30.West SL, Savitz DA, Koch G, Strom BL, Guess HA, Hartzema A. Recall accuracy for prescription medications: self-report compared with database information. Am J Epidemiol. 1995;142:1103–1112. doi: 10.1093/oxfordjournals.aje.a117563. [DOI] [PubMed] [Google Scholar]
- 31.Coulter A, Vessey M, McPherson K. The ability of women to recall their oral contraceptive histories. Contraception. 1986;33:127–137. doi: 10.1016/0010-7824(86)90079-x. [DOI] [PubMed] [Google Scholar]
- 32.Harlow SD, Linet MS. The agreement between questionnaire data and medical records: the evidence of accuracy of recall. Am J Epidemiol. 1989;129:233–248. doi: 10.1093/oxfordjournals.aje.a115129. [DOI] [PubMed] [Google Scholar]
- 33.Maggwa BN, Man JK, Mbugua S, Hunter DJ. Validity of contraceptive histories in a rural community in Kenya. Int J Epidemiol. 1993;22:692–697. doi: 10.1093/ije/22.4.692. [DOI] [PubMed] [Google Scholar]
- 34.Stolley PD, Tonascia JA, Sartwell PE, Tuckman MS, Tonascia S, Rutledge A, et al. Agreement rates between oral contraceptive users and prescribers in relation to drug use histories. Am J Epidemiol. 1978;107:226–235. doi: 10.1093/oxfordjournals.aje.a112529. [DOI] [PubMed] [Google Scholar]
- 35.Wingo PA, Lee NC. Use of life calendar to enhance the quality of exposure and risk factors histories. Am J Epidemiol. 1988;128:921. [Google Scholar]