You are on page 1of 9


Association between contemporary hormonal contraception

BMJ: first published as 10.1136/bmj.k3609 on 26 September 2018. Downloaded from on 4 October 2018 by guest. Protected by copyright.
and ovarian cancer in women of reproductive age in Denmark:
prospective, nationwide cohort study
Lisa Iversen,1 Shona Fielding,2 Øjvind Lidegaard,3 Lina S Mørch,3 Charlotte W Skovlund,3
­Philip C Hannaford1
Academic Primary Care, Abstract Results
Institute of Applied Health Objectives During 21.4 million person years, 1249 incident
Sciences, University of To investigate the association between contemporary ovarian cancers occurred. Among ever users of
Aberdeen, Aberdeen AB25
2ZD, UK combined hormonal contraceptives (including hormonal contraception, 478 ovarian cancers were
Medical Statistics Team, progestogen types in combined preparations and all recorded over 13 344 531 person years. Never users
Institute of Applied Health progestogen-only products) and overall and specific had 771 ovarian cancers during 8 150 250 person
Sciences, University of types of ovarian cancer. years. Compared with never users, reduced risks
Aberdeen, Aberdeen, UK
3 Design of ovarian cancer occurred with current or recent
Rigshospitalet, Juliane
Marie Centre, Department of Prospective, nationwide cohort study. use and former use of any hormonal contraception
Gynaecology, University of (relative risk 0.58 (95% confidence interval 0.49 to
Copenhagen, Denmark Setting
Denmark, 1995-2014. 0.68) and 0.77 (0.66 to 0.91), respectively). Relative
Correspondence to: L Iversen risks among current or recent users decreased Participants with increasing duration (from 0.82 (0.59 to 1.12)
(ORCID 0000-0002-2524-5229)
All women aged 15-49 years during 1995-2014 were
Additional material is published with ≤1 year use to 0.26 (0.16 to 0.43) with >10
eligible. Women were excluded if they immigrated
online only. To view please visit years’ use; P<0.001 for trend). Similar results were
the journal online. after 1995, had cancer (except non-melanoma skin
achieved among women followed up to their first
Cite this as: BMJ 2018;362:k3609 cancer), had venous thrombosis, or were treated for
switch in contraceptive type. Little evidence of infertility before entry (final study population included
major differences in risk estimates by tumour type or
Accepted: 30 July 2018 1 879 227 women). Women were categorised as
progestogen content of combined oral contraceptives
never users (no record of being dispensed hormonal
was seen. Use of progestogen-only products were
contraception), current or recent users (≤1 year after
not associated with ovarian cancer risk. Among ever
stopping use), or former users (>1 year after stopping
users of hormonal contraception, the reduction in the
use) of different hormonal contraceptives.
age standardised absolute rate of ovarian cancer was
Main outcome measures 3.2 per 100 000 person years. Based on the relative
Poisson regression was used to calculate relative risk risk for the never use versus ever use categories
of ovarian cancer among users of any contemporary of hormonal contraception (0.66), the population
combined hormonal contraceptives and by prevented fraction was estimated to be 21%—that
progestogen type in combined preparations and is, use of hormonal contraception prevented 21% of
all progestogen-only products, including non-oral
ovarian cancers in the study population.
preparations. Separate analyses examined women
followed up to their first contraception type switch Conclusions
and those with full contraceptive histories. Duration, Use of contemporary combined hormonal
time since last use, and tumour histology were contraceptives is associated with a reduction in
examined and the population prevented fraction were ovarian cancer risk in women of reproductive age—an
calculated. effect related to duration of use, which diminishes
after stopping use. These data suggest no protective
effect from progestogen-only products.
What is already known about this topic
Previous research relating to the use of older and higher dose preparations of Introduction
oestrogen containing older progestogens has shown a reduced risk of ovarian
Hormonal contraception is popular. At least 100
cancer in users of combined oral hormonal contraceptives
million women worldwide use hormonal contraception
This reduced risk has been shown to persist for many years after stopping use of every day.1 In 2012, an estimated 238 719 women
these combined oral contraceptives around the world were diagnosed with a new ovarian
What this study adds cancer and 151 917 died from the condition.2 Survival
from ovarian cancer is poorer than from many other
In this large, prospective, population based study of women aged 15-49 years in
cancers; with an overall, age standardised survival
Denmark, use of contemporary combined oral contraceptives containing newer
rate at five years of 30-40% worldwide, varying
progestogens was associated with a reduction in the risk of ovarian cancer
according to stage at diagnosis.3 Previous research
Few women in the study were exclusive users of progestogen-only contraceptives;
has shown a reduced risk of ovarian cancer in users of
therefore, evidence regarding progestogen-only contraceptives was limited, with
combined oral contraceptives, an effect that persists
no evidence of beneficial ovarian cancer effects among exclusive users of these
for many years after stopping use.4-9 However, most
of the evidence relates to the use of older and higher

the bmj | BMJ 2018;362:k3609 | doi: 10.1136/bmj.k3609 1

Such changes have occurred in formulations of combined Study population oral contraceptives (notably reductions in oestrogen All women living in Denmark aged 15-49 years from dose and the introduction of newer progestogens such 1 January 1995 to 31 December 2014 were eligible as desogestrel. Women were categorised as • National Register of Medicinal Product Statistics. The study 28 days were filled in prospectively. assigns a unique number to each resident for each model. and an increased use of progestogen. combined oral contraception4-9 as well as for other types of hormonal contraception. using the • National Birth Register.20 gaps between prescriptions of fewer than the risk of cancer or cardiovascular disease.10 Therapeutic Chemical Classification (ATC) system code It is important for users of contemporary combined MG03G in the National Prescription Registry). based on The Danish Sex Hormone Register Study. a benefit is specific to any particular formulation. in use surrounding 95% confidence intervals were calculated since 1968. with never users as the reference 2 doi: 10. stopped use Data linkage (estimated date when prescription finished. contraception that have occurred over time. patterns for the study unless they immigrated after 1995 (total of administration (continuous v monthly cycles in number eligible=1 904 094). We assumed that dataset uses an individual’s personal identification the levonorgestrel intrauterine system was used for four number in the Civil Registration System to link data years unless another hormonal contraceptive product from several national registries: was redeemed or pregnancy occurred before the end • Statistics Denmark. Such information is important for infertility treatment. linked to contemporary hormonal contraception.17 18 follows a national contraception (redemption date of prescription for a cohort of women aged 15-79 years. death. thereby ensuring accurate reflect the impact of substantial changes in hormonal linkage of different databases. (including progestogen-only oral contraceptives and whichever came first.bmj. of ovarian cancer. Anatomical only preparations.362:k3609 | the bmj . Using the methods of Nielsen and to investigate the relation between hormone use and colleagues. During follow-up. These data were updated daily to allow identification of when women started use Methods (date when prescription was redeemed). emigration. prescription for ovarian stimulating drugs.3 (SAS Institute). rate ratios (referred to here as relative risks) and their The personal identification number system. The to experience the same pattern of ovarian cancer women were followed up to first diagnosis of ovarian benefit as users in the older studies. bilateral putting into context the risks of other known cancers oophorectomy. number cohort of young Danish women. and whether cancer (ICD-10 (international classification of diseases. Therefore. for hospital discharge in SAS version 9. defined daily doses.11 Current evidence is insufficient for and for six months after delivery. • Danish Cancer Registry. dose. or switched use of hormonal been described previously. women were the levonorgestrel releasing intrauterine system) censored permanently at the date of diagnosis of non- also should know whether they are at a reduced risk ovarian cancer (except non-melanoma skin cancer). had venous thrombosis. RESEARCH dose preparations of oestrogen containing older in Denmark and is used as a key personal identifier BMJ: first published as 10.k3609 | BMJ 2018. and route of administration. Adjusted incidence diagnoses and surgeries since 1977. or were treated for a withdrawal bleed occurs). 10th revision)19 code C56). of packages. and drospirenone). current or recent users (up to one since January 1994 and considered complete since year after stopping use). Protected by copyright. such Women were censored temporarily during pregnancy as breast cancer. new non-oral routes of infertility before study entry (indicated by a redeemed administration. progestogens.1136/bmj. It was established different product). venous thrombosis. Downloaded from http://www.12-16 We report here Hormonal contraception a large investigation of contemporary hormonal Information about redeemed prescriptions included contraceptive use and ovarian cancer in a national date of redemption. this evidence might not in all national registries. Women were excluded which 21 days of combined hormonal contraception if they had cancer (except non-melanoma skin is followed by seven hormone-free days during which cancer). never users (no record of being dispensed hormonal for hormonal contraceptive prescriptions dispensed contraception). or end of follow-up (31 December 2014). or former users (more than 1 January 1995 one year after stopping use) of different hormonal contraceptives. ATC code. The final oral contraceptives to know whether they are likely study population comprised 1 879 227 women. for all births since 1973 and age distribution of the cohort as standard. gestodene. for education information of the four year period.k3609 on 26 September 2018. for histologically verified cancers since 1943 and family history of Statistical analysis premenopausal (age <50 years) breast or ovarian We calculated age standardised incidence rates of cancer in mothers or sisters ovarian cancers per 100 000 person years. which has number of packs issued).1136/bmj. or second unilateral on 4 October 2018 by guest. Risk of information on smoking status (since 1991) and ovarian cancer among users of the different product body mass index (since 2004) groups was analysed by a Poisson regression model • National Health Register. age Users of other hormonal contraceptive methods 50 years.

During digit 3. analyses stratified the data by duration of use.0) years. former will be shared with the media and disseminated to users.7). were carried out in invited to contribute to the writing or editing of this the subset of parous women with this information.8 (5. asking them to share patient mass index when ascertained antenatally (<18. a higher percentage of parous women were 6 9 To minimise the effects of previous use while using non-oral routes of hormonal contraception ensuring that we had sufficient data for analysis. no). for those aged was related to combined oral contraception (86%. 1249 women aged Studies of Ovarian Cancer.bmj. 35-39. from previous use of oral combined contraception. non-smoker. college/university education.1136/bmj. unknown) (expressed as a percentage) of the ovarian cancers in • Tubal sterilisation (yes. 40- (population prevented fraction=prevalenceexposure(1− 44. no) • Polycystic ovary syndrome (yes.24 When necessary. Downloaded from http://www. reduction International Cancer Control Tumour.k3609 3 . Most hormonal contraceptive use first change in hormonal contraception. university education The population prevented fraction is the proportion with research qualifications. Up to 31 December 2003. Since this was a population smoking status was ascertained antenatally) and body study using existing data.362:k3609 | doi: 10. 15 years on or after 1 January 1995 (that is. we BMJ: first published as 10. RESEARCH group. with values contraception. Metastasis in the age standardised absolute rate of ovarian cancer the bmj | BMJ on 4 October 2018 by guest. unknown). Node. 3. Previous levonorgestrel intrauterine system were older and more studies have indicated a persisting protective effect likely to be parous than women using other products.8 years (standard deviation 6. our or oral progestogen-only products than nulliparous risk estimates for different hormonal preparations women.23 after which the Union for Among ever users of hormonal contraception. The results The main analysis compared ever users. and 20-29 years had a mean follow-up of 16. 45-49 years) relative risk)) associated with ever use of any hormonal • Education (elementary school only. Current or recent users of the and for epithelial ovarian cancers only. • Parity (0. unknown at the time to interpret the results. 756 943 women mucinous (M8470/8480/8490). 20-24. and current or recent users of any hormonal members of the public through appropriate media contraception with never users as reference. further education excluding college/university. We tested for the trends for duration of use and 13 344 531 person years of follow-up. and 369 309 aged 40 years and registration) and ending with the behaviour invasive older had a mean follow-up of 4. staging classification. At study entry.5-25. non-epithelial (M8620/8631/8650/8 10. document for readability or accuracy.8 (4. by time since last current use.4 and used the following 15-49 years had incident ovarian cancer during more ICD-O-3 codes21: epithelial clear cell (M8310/8313).1136/bmj. The adjusted models included the following (TNM) classification was used. Patient and public involvement Patients or members of the public were not invited to Additional models adjusting for smoking status comment on the research question. 379 808 aged 862/8890/8933/8951/8963/9080/9084/9110). table 1). 2. Few women (contributing about 2-3% of total were derived mainly among women followed up to person years) had a family history of early (before age their first change in hormonal contraception during 50 years) ovarian or breast cancer. recent use. >30. table 2).7 (2. the study. and by ovarian tumour histology. Median age at Comparisons between main contraceptive groupings ovarian cancer diagnosis was 44. and epithelial serous younger than 20 years then had a mean follow-up of (M8441/8460).5. use versus ever use of any hormonal contraception. >4) We calculated the population prevented fraction • Age group (15-19. or (current smoker.4 years per woman).9) years. 771 ovarian cancers set to the median duration within each category. 4. 30-34. time varying covariates: converted TNM classification information into the FIGO stage classification. high school  contraception by using the relative risk of never only. epithelial 11. study design.2) years.9-47. Ovarian tumour histology used the same groupings Results as the Collaborative Group on Epidemiological During the 20 year study period. than 21. In never users. relevant outcomes was not possible.4 Generally. Separate channels. malignant tumour not otherwise specified (all other 373 167 aged 30-39 years had a mean follow-up morphology codes supplied alongside the C56 cancer of 13.k3609 on 26 September 2018. >25-30. no) • Family history of breast or ovarian cancer (yes.5 per 100 000 person years. no) the cohort that has been prevented by ever use of any • Hysterectomy (yes. 1. They were not 18. 25-29. tumour staging The age adjusted incidence of ovarian cancer was in the Danish Cancer Registry used the International highest in women who were never users of hormonal Federation of Gynaecology and Obstetrics (FIGO) contraception (7.0). Protected by copyright.4 million person years of observation (about epithelial endometrioid (M8380). 478 ovarian time since last current use by including the duration/ cancers were found among ever users of any hormonal time since variables as an ordinal variable.25 We then examined tumour FIGO • Calendar year stage by age at diagnosis and calculated the frequency • Hormonal contraceptive use of the different tumour types by stage at diagnosis. no) hormonal contraception. those with 7 612  267/8 839 374 person years of current or a fully documented hormonal contraceptive history). • Endometriosis (yes.22 occurred during 8 150 250 person years.4 years (interquartile were also conducted for women followed up to their range 38.

66 to 0. PCOS.2 0.6 (5. *From this year onwards to end of the study.2 2.4 (5.3 9.3 2.4 2.9 24.6 (7. U=university.8 23.5 5.6 1.2   Cyproterone (1995-) 317 961 27.0 (4.7 0.1 0.2 1.3 2.9 0.8 0.6 1.76)).4 3. 15 333 680 person years).2 1.5 0.8 (8.7) 21.7 0.5 2.6 6. Most of the risk 4 doi: 10. ¶Available since 1991 only in parous women (n=512 143.1 2.1 (4. dienogest 14 081 33.4   LNG-IUS (1995-) 708 111 40.5 86. descrip- tive percentages represent the percentage of person time with a given characteristic.2 0.6) 11. including 0.2 0. Overall.0 2.1 (4.5 24.0) 27.6 0.9 2.6 2.0 0.2 0.7 0.3 24.4 24.1 6.6 0.77 (0.1136/bmj.0 86. §Available since 2004 only in parous women (n=322 641.8 79.3 0.7) 32. progestogen-only products.2 (8.5 0.04 1.5 2.5 (5.3 23.9 4.3) 31.9   Levonorgestrel (1995-) 1 016 015 30.1 (5.68)) as well as for former use (0.6 (5.49 to this analysis.2 57.9 (10.8) 14. BMI=body mass index. (%)† Nulliparous PCOS ­sterilisation Hysterectomy Endometriosis history (mean Smoking on market during study*) years mean (SD)) E U (%) (%) (%) (%) (%) (%)‡ (SD))§ (%)¶ Main use groups Never use 8 150 250 35.58 to 0.1 0.1 2.9 24.9 (7.2 (7.5 24.3   Desogestrel (2001-) 123 539 32.5   Drospirenone (2001-) 592 556 26.2 6.8) 9. polycystic ovary syndrome.4 0.1 0.1136/bmj.03 2. the pattern of relative risks seen in the full risk than users of combined oral contraceptives. women followed beyond their first reduced risk of ovarian cancer among current users switch in hormonal contraception) were similar (table was stronger with increasing duration of hormonal 1S).5) 31.8 12.2  (2009-) Non-oral   Patch (2003-) 15 358 27.5 (5.6   Norgestimate (1995-) 721 004 27.1 25.1 0. ever users total period of observation.2 per 100 000 person years.4 0.5 2.7 0.7) 20.4 24.3 24.4 24.1 0.2 (7.7 23.7% of total number) contributing 25% quickly in those with shorter durations of use (table 3). and was non.7 0.4 (8.8) 15.9 Oral.6 0.7 0.5 78.8) 14.0 77.8 78.5 1.8) 18.9 24.66 (95% confidence interval 0.8 24. The reduced ovarian cancer the full cohort and among parous women followed up risk among previous hormonal contraceptive users to their first switch in hormonal contraception when diminished with time since stopping use.4 2.3 2.0 9.5 Current or recent use 8 839 374 29.8 11.3 1.8   Desogestrel (1995-) 1 659 258 27.4 2.7) 24.8 25. 73% unknown body mass index).1) 8.5 0.RESEARCH Table 1 | Characteristics of never. 50 µg ethinylestradiol   Norethisterone (1995-2002) 57 516 32.k3609 | BMJ 2018.1 0.5 0. The cohort (that is.3) 14.9 (8.91)).3 7.8) 22.7 0.2 33.1 0.5 Former use. E=elementary school.2 (4. Type of hormonal No of Education Tubal Family BMI contraception (dates person Age (years.2 2. since last use and duration were examined.1 2.3) 17.0) 18.3 2.8) 28.4 63.0   Estradiol valerate.4) 54.1 0.0) 24.3 2.8 0. When both time not shown).4 86.6 0.3 0.7) 18.6 (9.0 (6.2 0.1 0. Protected by copyright.6 0.4 0.7 68.362:k3609 | the bmj .4 0. evidence Restricting the analysis to women with a complete among former users of any hormonal contraception contraception exposure history (those aged 15 years on indicated greater protection with longer durations of or after 1 January 1995) reduced the cohort to 596 395 use and a suggestion that the protection waned more women (31.9) 13. †Percentage of participants with elementary school education only and percentage with university education.3 79.7 (8.5 2.0 (6.3 (8.6 55.8 6.4 9.6 85.1 0.9 80.6) 19.4 46.5 0.3 0.7 Current or recent use of combined hormonal contraception Oral.2 1.0 (4. ‡Family history of premenopausal breast or ovarian cancer.1 1.9) 43.06 3.3 (7. In use of any hormonal contraception (0.5 (8. SD=standard deviation.1 7. we adjusted for smoking and body mass index (data significant by 10 years after last use.7 0.3 0.7 (5.9 (4.bmj.7 (7.5 68.9 2.2 0.2   Gestodene (1995-) 2 985 909 27.3 2.0) 26. Overall similar risk estimates were found in both contraceptive use.5 0.4 2.7) 18.9 0.6   Levonorgestrel (1995-2009) 82 756 34.0 (7.3) 21. LNG-IUS=levonorgestrel intrauterine system.3 0.8 (8.8 (5.k3609 on 26 September 2018.4) 24.3 24.4 (11.2 4.7 6.5) 46.0 1.7 0.3) 10.3 1.0 2.8 8.5 Non-oral   MPA depot (1995-) 27 832 27.5 24.3) 22.3 1.7 (4.1 0.58 (0.8) 24.3 0.2 24.2 0. Downloaded from http://www.1 74.8 (4.1 0.1) on 4 October 2018 by guest.1 0.1 0.8) 35.0 89.4 0.4 1.7 45.4 26.2 (5.0 0.1 1.6 4.9 (5.0) 9. Reduced different progestogens.1 0.0   Levonorgestrel (1995-2005) 11 544 37.4) 45. For combined seemed to have a smaller reduction in ovarian cancer products.5 72. reflecting the first switch in hormonal contraception (71% of the younger age of this subset of women. MPA=medroxyprogesterone acetate.7 (4.4 2.4 2.5) 18.0 0.1 0.0) 24.1 0. 57% unknown smoking).3   Vaginal ring (2002-) 124 280 28.8 12. of any hormonal contraception had a reduced risk of there was little evidence of important differences ovarian cancer compared with never users (relative risk between combined oral contraceptives containing 0.2) 11. was 3.7 Descriptive statistics calculated as the average person time with a given characteristic divided by the total amount of person time on a specific hormonal contraceptive.6 (8. former. Current or recent users of progestogen-only products did not protect against ovarian cancer. >12 months ago 4 505 157 36.1 0.3 (5. of the total period of observation (5 417 268 person Among women followed in the cohort up to their years) and 6% of all ovarian cancers.5   Implant (1999-) 58 371 26.2 23. and current or recent users of different types of hormonal contraception BMJ: first published as 10. at least for those with sufficient risk estimates were also observed for current or recent usage to produce precise risk estimates (table 4).1) 26.6 11.0 Current or recent use of progestogen-only products Oral   Norethisterone (1995-) 158 073 34.6 (7.8 8. the levonorgestrel containing intrauterine system.9) 28. 20-40 µg ethinylestradiol   Norethisterone (1995-) 165 211 28.

7 0.9 million women. the population prevented effects found. and so few is.001).82 (0.99)   >10 years 665 281 51 3. because of the time varying fraction was estimated to be 21%—that is.80 (0. effect.08) *Adjusted for calendar year.49 to 0. estimates were imprecise although broadly compatible types of progestogens. duration of current use No of ovarian Age standardised incidence Adjusted relative risk Use category No of person years cancer events per 100 000 person years (95% CI)* Never use 8 150 250 771 7. Most of the hormonal contraceptive how contemporary hormonal contraceptives affected use related to combined oral contraceptives. ‡Ptrend=0.53 (0.26 (0. hysterectomy.74)   >10 years 930 943 18 1.9 0.61 to 0.03. If some III or IV compared with 15% (30/204) of mucinous women did not (that is. Discussion The data linkage study design also enabled us to Principal findings adjust for several important confounding variables.3 0. Our data do not suggest a protective was associated with a reduced risk of any epithelial effect from progestogen-only products.77 (0.95) Duration of current or recent use of any hormonal contraception†   ≤1 year 1 265 020 42 5.8 0. age. endometriosis.76) Former use (any hormonal) 4 505 157 244 5.2 0.0 0. in particular endometrioid.59 to 1.001. Based on the relative risk for the when we incorrectly deemed them to be a user).78 (0. 73% of the total ovarian Strengths and limitations of study cancers) was examined among women followed up Strengths of our study included its nationwide to their first switch in hormonal contraception (table coverage of nearly 1.58 (0. parity.5 1. Similar to most studies of hormonal with younger women (table 4S.91) Current or recent use (any hormonal) 8 839 374 234 on 4 October 2018 by guest.0 0. hormonal contraception was seen with nearly all types Current or recent use of any hormonal contraception of ovarian cancer. younger than 50 years) living in Denmark. Ptrend<0. of the redeemed prescription.58 to 0.72 (0. the women would be classified as a non-taker. more than 21 3S). RESEARCH Table 2 | Relative risk of ovarian cancer in users of hormonal contraception according to time since last current use and BMJ: first published as 10. and the bmj | BMJ 2018. mucinous.1136/bmj.8 0. Thereafter.45 to 0. contraception currently available. however.50 to 0. it also periods of current use and persisted for several years meant that the study could not provide information on after stopping use. When the subset of epithelial ovarian cancers (773/1064.16 to 0.62 (0.6 0.k3609 on 26 September 2018. use of variables used in our study.68) Current or recent use (combined oral) 7 751 904 175 3.59 to 1.76 (0. Protected by copyright.66 to 0. In this study of women of reproductive age (that The cohort was younger than 50 years. †Ptrend<0.k3609 5 .66 (0. such never use versus ever use categories of hormonal misclassification would underestimate the protective contraception (0. tubal sterilisation. We were not able to combined oral contraceptives containing different adjust for some factors.64) Current or recent use (progestogen only) 1 087 470 59 4.57 (0.1136/bmj.55 to 0.93)   >5-≤10 years 1 397 257 83 4. education. few ovarian cancer. and family history of breast or ovarian cancer.bmj. an effect that strengthened with longer were examining the effects of oral contraception.3 0.43) Time since last current use of any hormonal contraception‡   >1-≤5 years 2 442 620 110 5. and and serous epithelial cancer (table 5). There was not so we had limited statistical power to detect such an a significant reduction in clear cell epithelial or non.44 to 0. polycystic ovary syndrome. and the contraceptives was similar to that seen for all ovarian examination of the many different forms of hormonal cancers.26 While this age contraception was associated with a reduced risk of restriction meant that we could be confident that we ovarian cancer. However. in whom most was little evidence of important differences between cases of ovarian cancer occur. The reduced risk associated with with the results of the main analyses (table 2S).61 to 0.12)   >1-≤5 years 4 063 111 103 4.66). epithelial ovarian cancer.77)   >5-≤10 years 2 580 300 71 4. such as breastfeeding. the pattern of relative risks for different hormonal million person years of prospective follow-up.4 0. we women will have used hormone replacement therapy found that ever use of any contemporary hormonal known to increase ovarian cancer risk. they remained being a non-user tumours (table 5S). this misclassification of hormonal contraception prevented 21% of ovarian current use would only be present for the time period cancers in the study population.00 Ever use (any hormonal contraception) 13 344 531 478 4. Over 60% contraceptives. we have assumed that women who were (323/502) of serous tumours were diagnosed at stage dispensed a prescription subsequently used it. women were exclusive users of such products.362:k3609 | doi: 10. The linkage of Age at diagnosis was associated with FIGO stage prescribing and cancer registration information with a higher percentage of women in the oldest age avoided recall bias regarding patterns of hormonal group being diagnosed at stage III or IV in comparison contraception use.0 0. Downloaded from http://www. There ovarian cancer risk in older women.

education.41 to 0.61)   Desogestrel.50 to 0.15) (0. and family history of breast or ovarian cancer.84 635 891 29 0. 35 µg ethinylestradiol 375 778 11 0.29) Current or recent use of combined hormonal contraception—oral   Norethisterone.41 40 304 2 0. the no change cohort) No of person years No of cancer events Adjusted relative risk (95% CI)* Never use 8 150 250 771 1.1136/bmj. 20-30 µg ethinylestradiol 988 952 17 0.23 to 1. Protected by copyright. Our on 4 October 2018 by guest. therefore. hysterectomy.48 to 1.59 to 1.98) (0.39 to 0.84) Former use (any hormonal) 2 590 322 164 0.61 (0.85 (0.21) >1-≤5 years 1 038 886 47 0.362:k3609 | the bmj .73 (0. it seems to their first switch of hormonal contraception in the unlikely that there are important differences in these study was intended to minimise such effects.45 to 0.71 to 1.14 to 2.16 to 8.11 to 20.08 (0. information about endometriosis and polycystic ovary We did not have information about hormonal syndrome only related to women admitted to hospital contraceptives prescribed before entry to the study. 6 doi: 10.23)  Desogestrel 12 155 0 — Current or recent use of progestogen-only contraception—non-oral   MPA depot 7321 3 6. 30-35 µg ethinylestradiol 116 090 7 1. we do not believe that our results were contraception meant that it was not possible to seriously affected by this misclassification.84 to 1. a problem that should diminish the adjustment for these variables in the subset of women further a woman was from her last use.91) (0. 20-35 µg ethinylestradiol 1 887 047 42 0.64)  Levonorgestrel 6972 1 1.73)   Gestodene.60 291 123 9 0. risk estimates. while still temporal associations. endometriosis. Some of examine the association between duration of use and the effects attributed to the use of a particular product time since last current use among users of specific could reflect lingering effects from the use of a previous preparations.16 (0.33) *Adjusted for calendar year.58 to 1.30 (0. Because older for whom these data were available did not materially women are further away from their last use.54 to 2. However.71)   Norethisterone.35 (0.56 (0.RESEARCH Table 3 | Relative risk of ovarian cancer in former users of hormonal contraception by time since last current use and duration of use BMJ: first published as 10. Information on have used hormonal contraception previously.14 316 850 27 0. Time since last current use of hormonal contraception 1-<5 years 5-<10 years ≥10 years Duration of use No of person No of ovarian Relative risk No of person No of ovarian Relative risk No of person No of ovarian Relative risk years cancer events (95% CI)* years cancer events (95% CI)* years cancer events (95% CI)* ≤1 year 667 835 37 0.57 (0.64)   Norgestimate.k3609 on 26 September 2018.61)   Levonorgestrel.76)   Levonorgestrel. Such smoking and body mass index was only available in misclassification would underestimate the protective parous women for part of the study period. polycystic ovary syndrome. and family history of breast or ovarian cancer.21 to 0. age.37)   Cyproterone. 30 µg ethinylestradiol 142 147 0 —   Estradiol valerate.84 (0. hysterectomy.62 to 2.57 (0.00 Ever use (any hormonal contraception) 7 183 430 293 0. age.75 (0. for these conditions. 30-35 µg ethinylestradiol 519 113 11 0.79)   Drospirenone.45 (0.35) LNG-IUS=levonorgestrel intrauterine system. tubal sterilisation. providing enough data for reliable risk estimates. 50 µg ethinylestradiol 36 494 4 1. given the little evidence overall product or products.74) Current or recent use (combined OC) 4 316 888 103 0. However.81 (0.81) (0. could Some women who were classified as never users could be subject to residual confounding.13) (0.k3609 | BMJ 2018.56) (0.62 to 1.87 (0.40)  Implant 10 575 0 —  LNG-IUS 172 265 18 0. tubal sterilisation. Our analysis of women followed of major differences between preparations.81 470 243 45 1. polycystic ovary syndrome.53 to 1.33 (0.70) Current or recent use (progestogen only) 276 221 26 0.20) >5 years 735 899 26 0. dienogest 1022 0 — Current or recent use of combined hormonal contraception—non-oral  Patch 2253 0 —   Vaginal ring 11 729 0 — Current or recent use of progestogen-only contraception—oral  Norethisterone 66 934 4 0.54 to 1.56 (2. 50 µg ethinylestradiol 47 335 6 1.27 to 0.02) Current or recent use (any hormonal) 4 593 109 129 0. education.18 to 0.bmj. MPA=medroxyprogesterone acetate. 20-35 µg ethinylestradiol 188 928 5 1.76 (0.63 to 0. parity. and change the risk estimates. endometriosis.44 to 2. Downloaded from http://www.62 (0.50 to 3. parity.66 308 126 22 0.21 (0. Little Table 4 | Relative risk of ovarian cancer in users by different types of hormonal contraception in women followed up to their first switch in hormonal contraception (that is. *Adjusted for calendar year.45 to 0.41 to 1.1136/bmj. because most instances of ovarian cancer occur in The low incidence of ovarian cancer among women older women (71% of our events were in women older followed up to their first switch in type of hormonal than 40 years).

endometriosis. Our slightly stronger with imprecise.06) Mucinous epithelial ovarian cancer — 204 —   Never use 8 150 250 107 1.08) Malignant tumour not otherwise specified No of ovarian cancer events — 274 — Never use 8 150 250 172 1.58 (0.68 (0.32 to 1. Protected by copyright.59 (0. our study 58% of the total period of observation in ever users included women aged 15-49 years.38 to 1.96 ( on 4 October 2018 by guest.64 (0.80) Former use 4 505 157 44 0.4 The overall relative risk between ever Studies4 5 have not found differences in ovarian and never users was 0. Furthermore. most of whom arising from current or recent use of combined oral will have been premenopausal.58 to 0.56 to 0.93) Non-epithelial ovarian cancer No of ovarian cancer events — 73 — Never use 8 150 250 33 1. women who were generally older than our cohort (mean or reduced statistical power to continue to observe a age of diagnosis of ovarian cancer was 56 years.62 to 0. polycystic ovary syndrome.bmj. many years previously.09 (0. demonstrating that how recent a woman unknown previous use could not occur).73 (0. with significant reduction due to the relatively small periods only 18% of tumours diagnosed in women younger of observation (7% of total person years for ever use).55 (0. and family history of breast or ovarian cancer. Downloaded from http://www.4 However. when use. cohort).k3609 on 26 September 2018. because many contraception got stronger with longer durations of use would have stopped using hormonal contraception and persisted for a number of years after stopping use. education.00 Current or recent use 8 839 374 58 0.35 to 0.05) *Adjusted for calendar year.72)   Former use 4 505 157 97 0.76).73 (95% confidence interval cancer risk by oestrogen dose of combined oral 0.1136/bmj.97)   Former use 4 505 157 52 0.41) Endometrioid epithelial ovarian cancer — 135 —   Never use 8 150 250 93 1. tubal sterilisation.1136/bmj.20) Former use 4 505 157 19 1. than 45 years). observation for ever users of hormonal contraceptives Similar to the Collaborative Group4 and other in this age group had a higher proportion of information more recent investigations.34 to 1. we restricted our analysis to women with complete no significant heterogeneity by menopausal status or contraceptive histories (in whom misclassification from age was seen.00   Current or recent use 8 839 374 45 0.00   Current or recent use 8 839 374 80 0.41) Serous epithelial ovarian cancer — 502 —   Never use 8 150 250 325 1. The Collaborative Group on The apparent loss of protection 10 years after stopping Epidemiological Studies of Ovarian Cancer’s reanalysis hormonal contraception could be due to the loss of of oral contraception data from 45 studies included biological effect (assuming a causal association exists).70 to 0. RESEARCH Table 5 | Relative risk of different histological types of ovarian cancer associated with hormonal contraception (full BMJ: first published as 10. evidence of important differences was seen between of ovarian cancer per five years of oral contraceptive the combined oral contraceptives. by categories of hormonal contraception use Histology and use category No of person years No of cancer events Adjusted relative risk (95% CI)* Epithelial ovarian cancer Any epithelial ovarian cancer — 902 —   Never use 8 150 250 566 1. reduced risk for any ovarian cancer among ever users of any hormonal contraception (relative risk 0.76)) was possibly due to By contrast with most previous research.5-9 we found that the risk relating to current or recent use than those for ever reductions among current users of any hormonal users in a study recruiting older women. The found that younger and premenopausal women same studies were unable to investigate associations seemed to have greater percentage reductions in risk with combined pills containing different progestogens.72 (0.57 (0.00   Current or recent use 8 839 374 155 0. Thus.75 (0. albeit than the other two factors.28 to 1.00 Current or recent use 8 839 374 21 0. the bmj | BMJ 2018.27)   Former use 4 505 157 11 0.91) Clear cell epithelial ovarian cancer — 61 —   Never use 8 150 250 41 1.47 to 0. the periods of contraceptives. The Collaborative Group’s analysis contraceptives (when assessed by decade of use).00   Current or recent use 8 839 374 21 0.42 to 0.44 to 0. hysterectomy.65 (0.70)   Former use 4 505 157 181 0.k3609 7 . non-significant risk estimates.96)   Former use 4 505 157 21 0.58 (0.362:k3609 | doi: 10. after accounting for time since last use.00   Current or recent use 8 839 374 9 0. age.50 to 1.57 to 2.41 to 0.64 (0.65 to 1. parity.66 (95% Comparison with other studies confidence interval 0. the pattern of last used hormonal contraceptives was more important results for individual combinations was similar.

4 28 medroxyprogesterone. it is in fact a heterogeneous detect a significant protective effect. In line with current and found little evidence of important differences in understanding.22 The Nurses’ study had insufficient data tumour type) and Collaborative Group (mean age at to provide risk estimates for products containing diagnosis 56 years) studies.29 However. This high incidence in their risk of ovarian cancer. ØL is the principal investigator of the Danish Sex Hormone a decreased risk of serous. although it is not yet known how long for. she is guarantor for the paper. we had good similar among endometrioid. suppress ovulation so protection against neoplastic with the reduced risk estimates seen in the full cohort development is feasible.30 Recent understanding suggests that only products (table 4).60 (95% confidence interval 0. The consortium found that a five first draft and subsequent revisions of the paper with input from year increase in duration of oral contraceptive use and all authors.1136/bmj.12 Based on results from our prospective study. and clear Register study. current of mucinous tumours has been observed previously in or recent use of progestogen-only products among all Denmark30 and elsewhere.28 and it is widely recognised women implied a smaller effect on ovarian cancer risk that improvements in pathology.2 million products. It has been suggested that combined oral types. but the exact mechanisms by analysis due to lingering effects from previous combined which hormonal contraceptives reduce ovarian cancer pill usage. LSM provided expertise on ovarian cancer 8 doi: 10. full cohort analysis could be due to a lingering effect the epidemiological evidence suggests a longlasting from previous combined pill usage. Downloaded from http://www. Our findings suggest that the protective effects of Although we lacked power to examine patterns of risk current or recent use of hormonal contraception is by duration of use and progestogen type. Protected by copyright. 95% confidence interval 1. desogestrel.83. Whatever the biological mechanisms. and provided expertise regarding cell tumours but not mucinous tumours.72.1136/bmj.9 years depending on histological norgestrel).45 to 0.4 It is important to note that our study 2. imaging.bmj.14-16 with findings from those It has been suggested that recent downward trends studies able to investigate exclusive use14 16 suggesting in ovarian cancer mortality rates in North America and a protective effect.16 to was not associated with the incidence of mucinous 2.29 Combined hormonal contraceptives of biological effect from progestogen-only products. imprecise risk estimates. or drospirenone. which supports the notion that these ovarian cancer Soini and colleagues12 13 reported a standardised mortality benefits are likely to continue.3-68. mucinous and serous statistical power for the most often used progestogens types of epithelial ovarian cancer.27 In our study. contemporary combined hormonal contraceptives are and serous ovarian carcinomas. with patterns similar Finnish studies did not adjust for parity or previous to those seen with older combined oral products. and so had limited statistical power to to be one disease. only 3% users of combined preparations containing different of all ovarian cancers are now thought to be mucinous progestogens. Our finding of an increased risk was Europe can be partly attributed to the use of combined based on only three exposed ovarian cancers and a oral contraceptives. SF undertook all the more than 10 years of use were both associated with statistical analyses and wrote the first draft of the methods and results. incidence ratio of 0. as mucinous. of preparations containing the oestrogen mestranol of Ovarian Cancer found that oral contraceptive use (hazard ratio 1.k3609 | BMJ 2018.362:k3609 | the bmj . The Ovarian Cancer Cohort Consortium (OC3) there is insufficient evidence to suggest similar recently examined risk factors for different histological protection among exclusive users of progestogen-only types of ovarian cancer among more than 1. endometrioid. hormonal contraception. Alternatively. compared with users of combined products markers will alter the reporting of the morphological (table 2). Asia. no protection epithelial tumours could have been misclassified against ovarian cancer was seen in users of on 4 October 2018 by guest. In the no change cohort followed up to the subtypes of ovarian cancer. most mucinous tumours in our study overall ovarian cancer risk among current or recent were diagnosed at FIGO stage I. some of the first switch in hormonal contraception. endometrioid.88) and first generation progestogens (1. women from 21 prospective studies in Europe. 1.65) were associated with an increased ovarian cancer investigated women of reproductive age and who were risk.28 Similarly. This risk estimate was based on although ovarian cancer is clinically considered only 26 events. Presently.13 Our findings do still associated with a reduced risk of ovarian cancer not concur with those studies.RESEARCH The Nurses’ Health Study II reported that short term use the Collaborative Group on Epidemiological Studies BMJ: first published as 10. the group of neoplasms with different pathogenesis lack of significance could be because of an absence pathways. resulting in very fraction of 21% with use of hormonal contraception.k3609 on 26 September 2018. advised on analyses. norgestimate. protection against most types of ovarian cancer from Few studies have examined use of depot combined oral contraception. Contributors: LI conducted the literature review and wrote the and North America.29 Therefore. but found no relation with those containing younger than those in both the OC3 (median age at second generation progestogens (levonorgestrel and diagnosis 61. possibly because the in women of reproductive age.11 to tumours. with decreased risks for mucinous. and serum estimates. use of oral contraceptives (shown to have a persisting The reduced risk seems to persist after stopping use.29 indicating a relatively high incidence of contraceptives do not differ from progestogen-only pills mucinous tumours in our cohort. protective effect). suggesting that the protection found in the risk are unclear.76) for ovarian cancer and use of the levonorgestrel Conclusions intrauterine system among women in Finland.31 We found a population prevented very small total observation period.

371 Lifestyle and Health Cohort Study. Yeates D.96:889-92. doi:10.2014.bmj.1136/bmj.1001/jama. Keiding N.bmj. Ovarian cancer study. ovary. doi:10. Lancet 2015. cervix. La Vecchia C. Mellemkjaer L. Mäenpää J.95:1181-9. et al. analysis. JAMA 2009. Thrombotic stroke and myocardial infarction with hormonal The lead author affirms that the manuscript is an honest.2011. 2 GLOBOCAN 2012: Estimated cancer incidence. (2015).1093/annonc/mdw306 College of General Practitioners’ Oral Contraception Study.pdf accessed 05/03/2018.1052 Health Data Board. Reeves G.1038/sj. ovarian cancer and 87. Grénman S. Rodriguez T. Kjaer SK.1038/bjc.bjc. Impact of levonorgestrel-releasing intrauterine system disclosure form at www.aspx. Løkkegaard E.302:298-305.icmje. doi:10. Br J use of hormonal contraceptives: the Norwegian-Swedish Women’s Cancer 2011.1111/ Control 2017. nothing to disclose. J Clin Oncol 2016. Gaitskell K. doi:10. Bandera EV.ch39. provided the original work is properly cited and the use is non.1016/S0140-6736(14)62038-9 doi:10.0/. Oral contraceptive use and cancer: final doi:10. Adams-Campbell LL.90:1386-91. and approved the final manuscript. 1992. Collaborative Group 26 Beral V. Nevertheless. Kelemen LE.1471-0528. Ovary and fallopian tube. 2014 FIGO staging for ovarian. Revised FIGO staging for gynaecological cancer. Palmer JR. Cancer Causes worldwide 2015 (ST/ESA/SER. BJOG 2012.133:401-4. D’Cruz AK. doi:10. in the Pukkala E. diseases for oncology. Grénman S. individual data for 25. Gynecol Oncol 2014. Risk for reproductive factors and risk of ovarian cancer in the European invasive and borderline epithelial ovarian neoplasias following Prospective Investigation into Cancer and Nutrition. tube and peritoneal cancer.2016. Jensen A. Alberg AJ. John Wiley & Sons. doi:10. Schock H. See: http://creativecommons. Adami H-O. Jensen A. contraception.1016/j. RESEARCH epidemiology.28:385-91. doi:10. Prat J.pmed.08.303 controls.0b013e328348a6e7 8 Bethea TN. N Engl J Med 2017. LSM and CWS 15 Wilailak S. CWS wrote the initial programmes Group Steering Committee.bmj. 1010. Trabert B. et al. http://globocan. Negri E. Løkkegaard E. doi:10. Ovarian and tubal study of reproductive factors and exogenous hormone use in cancer in Denmark: an update on incidence and survival. et al.un.1371/journal. Using Creative Commons Attribution Non Commercial (CC BY-NC 4.377:2228-39. The corresponding 11 Mørch LS.008 Cancer Cohort Consortium. Br J Cancer 2004. Hannaford PC.257 women with individual participant meta-analysis of 52 epidemiological studies. Hurskainen R. Moser K. Nordisk Foundation. the work was obtained from the Danish Data Protection Agency and doi:10.1097/PAT. et al.676.pdf and declare: use on the cancer risk of the ovary and fallopian tube. 3rd ed. Int J Cancer 1991. Skovlund CW. Andreasen AH.12948 Control 2017. Menopausal hormone use and ovarian cancer risk: 45 epidemiological studies including 23. Oral contraceptive use and Swedish Women’s Lifestyle and Health Cohort Study. doi:10. Fielding S.1016/j.124:292-9. Mäenpää J. doi:10.1097/AOG.55:1281-4. World Health discrepancies from the study as planned have been explained.ygyno. section 3). Acta support from the Novo Nordisk Foundation for the submitted work. Iversen L.2016. Ovarian Cancer Association Consortium BMJ: first published as 10. Br J Obstet Gynaecol 1989.0) license. 22 Shafrir AL. A prospective 30 Gottschau M. the results.1001182. PCH. Brierley JD. during the conduct 14 Urban M. Depot became employed by the Novo Nordisk Foundation after the work medroxyprogesterone acetate and epithelial ovarian cancer: reported in the manuscript was completed. Rosenberg L. 21 Fritz A.43:420-32. and endometrium United States on hormonal contraception and venous thrombosis for in black South African women: case-control study.1471-0528.k3609 on 26 September 2018. Cancer risk in women using the levonorgestrel-releasing collection. 18 Lidegaard Ø. Trends in contraceptive use in the United States born from 1947 to Jack A.1016/S0140-6736(14)61687-1.02. Reproductive doi:10. Cancer Causes Gynecol Scand 2016.109:djx144. 6 Vessey M. Department of Economic and Social Affairs. N Engl J Med mortality and doi:10. author attests that all listed authors meet authorship criteria and that Lidegaard Ø. Staging and summary_table_site_sel. World Health Organization. International statistical classification of important aspects of the study have been omitted. Allen NE.385:1835-42. J Natl Cancer Inst 2017. Acta Obstet relation to ovarian cancer risk among Black women. annepidem. 7 Moorman PG. Ethical approval is not required Contraceptives. Carioli G.105:1436-42.1989.1016/j. Reeves G. 2000.x Ethical approval: The data were analysed and held within the secure 16 The WHO Collaborative Study of Neoplasia and Steroid data repository at Statistics Denmark. All authors designed the study. et al. Ovarian cancer Collaborative Group On Epidemiological Studies Of Ovarian and oral contraceptives: collaborative reanalysis of data from Cancer. Krüger-Kjaer 24 Owens G.1111/aogs. ajog. Contraception 2013. Poole EM. doi:10.e1-9. et al. Committee law (chapter . Keiding on 4 October 2018 by guest.6601715 report from the Oxford-Family Planning Association contraceptive 28 Wentzensen N. Lidegaard O. prognostic grouping. ØL Oncol 2016.2016. Lund E. Hannibal CG.9:e1001182. doi:10.1111/j. Current gaps in registries. Weir HK. 10th revision. fallopian registries in 67 countries (CONCORD-2). Am J Obstet Gynecol 2017. that no 19 World Health Organization.1175660 reports grants from the Novo Nordisk Foundation.0000000000000356 Competing interests: All authors have completed the ICMJE uniform 13 Soini T. Paavonen J. Ann Lifetime cancer risk and combined oral contraceptives: the Royal Oncol 2016.004 Pathology 2011. prescription registries to define continuous drug use: how to which permits others to distribute. Data sharing: Data sharing is not available. Weiderpass E. Injectable and oral contraceptive of the study. International classification of commercial.27:2017-25. paper.2017. in the writing of the intrauterine system in Finland. et al. UICC manual of clinical oncology. Paavonen J. Hermon C. doi:10. and interpretation of data. Obstet Gynecol 2014.887 patients from 279 population-based 25 Mutch DG.04. NEJMoa1700732 Funding: Supported by a grant (No 11645) from the Novo Nordisk 12 Soini T. Downloaded from http://www. eds.1016/S0140-6736(08)60167-1 27 Kumle M. of oral contraceptive use and ovarian cancer among women Population Division. Norwegian- 5 Tsilidis Subscribe: http://www. a review with emphasis on new developments and pathogenesis. during the conduct of the study.2009. Doll R. Peto R.03298. or in the decision to submit the paper for publication.17:384-8. for register based studies according to Danish Research Ethics 17 Mørch LS.49:191-5. Depot-medroxyprogesterone acetate (DMPA) and risk of epithelial ovarian cancer. PCH advised on analyses and provided expertise on ovarian cancer epidemiology: the need for new population-based hormonal contraception. trendsContraceptiveUse2015Report. Poole EM. This is an Open Access article distributed in accordance with the 20 Nielsen LH.002 Web appendix: Supplemental tables No commercial reuse: See rights and reprints http://www.iarc. LSM reports grants from the Novo Med 2012.tb03341.8178 factors and ovarian cancer risk in African-American 29 McCluggage WG. Lancet 2008. and has been an expert witness in two legal cases in the use and cancers of the breast.1056/ NEJMoa1111840 and transparent account of the study being reported. Vipupinyo C.07.26:654-62.34:2888-98.385:977. Hermon C.66.013 4 Beral V. accurate. doi:10. Ann Epidemiol 2016. approval for Hormone therapy and ovarian cancer. adapt. Epidemiology Working programmes for data management. CONCORD Working Group. build upon this work fill gaps between prescriptions. interpreted research. Contemporary hormonal contraception and the risk of breast cancer. for data management and prepared all the data from the national Members of the EWG SC. In: O’Sullivan B. Hurskainen R. Egger S. Peto R. Sivasubramaniam S. eds. Accessed 08/02/2017. Foundation. doi:10.371:303-14. 9 Iversen L. 2015: Global surveillance of cancer survival 1995-2009: analysis of 483-510. Pharmacoepidemiol Drug non-commercially.88:678-83.1002/9781119013143.1007/s10552-016-0840-4 31 Malvezzi M. 3 Allemani C.216:580. risk factors by histologic subtype: an analysis from the Ovarian contraception. and license their derivative works on different Saf 2008.2012. Lancet 2015. Lee AJ. doi:10. Organization. PLoS the plaintiff in 2011 and 2012. and that any diseases and health related problems.119:672-7.1056/ no others meeting the criteria have been omitted. Banks E. Suraseranivong V.A/349). 9th ed.1549 terms. Hannaford PC. advised on the statistical analysis. in alphabetical order. Løkkegaard E. on Epidemiological Studies of Ovarian Cancer.1080/0284186X. Fielding S.1007/s10552-017-0876-0 development/desa/population/publications/pdf/family/ 23 Shepherd JH. Key TJ. et al. Skovlund CW.1002/pds. Pukkala E. Percy C. Carreira H. doi:10. Andreasen AH.x prevalence worldwide in 2012. Braaten T. LI. Protected by copyright.2013. The funder had no role in the study design. doi:10. Kitchener H. et al. et al. Global trends and predictions in ovarian cancer mortality.28:371-83. Morphological subtypes of ovarian carcinoma: women.1016/j. A prospective cohort study 1 United Nations. and SF have a multicentre case-control study.366:2257-66. remix. doi:10. www. and wrote the initial 10 Doherty JA.