sign in sign up
<<
Home , Nomegestrol, Nomegestrol acetate

BMJ

Confidential: For Review Only Prolonged exposure to high doses of and risk of in women: French cohort study on 253,777 exposed women

Journal: BMJ

Manuscript ID BMJ-2020-054621

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the 24-Jan-2020 Author:

Complete List of Authors: Weill, Alain; French National Health Insurance, Department of Public Health Studies; Scientific Interest Group - Health Product Epidemiology (ANSM-CNAM EPI-PHARE SIG) NGUYEN, Pierre; Scientific Interest Group - Health Product Epidemiology (ANSM-CNAM EPI-PHARE SIG) Labidi, Moujahed; Lariboisière Hospital, AP-HP, Department of Neurosurgery Cadier, Benjamin; French National Health Insurance, Department of Public Health Studies Passeri, Thibault; Lariboisière Hospital, AP-HP, Department of Neurosurgery Duranteau, Lise; Bicêtre Hospital, AP-HP, Unit and Rare Disorders of Sex Development Bernat, Anne-Laure; Lariboisière Hospital, AP-HP, Department of Neurosurgery Yoldjian, Isabelle; French National Agency for the Safety of Medicines and Health Products (ANSM), Endocrinology and gynaecology product manager Fontanel, Sylvie ; Grand Est regional health authority Coste, Joel; French National Health Insurance, Department of Public Health Studies; Cochin Hospital, AP-HP, Biostatistics and Epidemiology Unit

Keywords: , meningioma, hyperandrogenism

https://mc.manuscriptcentral.com/bmj Page 1 of 56 BMJ

1 2 3 Prolonged exposure to high doses of cyproterone acetate and risk of 4 5 6 meningioma in women: French cohort study on 253,777 exposed women 7 8 9 10 11 12 Confidential: For Review Only 13 14 Short title: cyproterone acetate and risk of meningioma study 15 16 17 18 19 20 21 22 1,2 2 3, 1 23 Alain Weill (ORCID 0000-0001-8687-9092), Pierre Nguyen , Moujahed Labidi , Benjamin Cadier , 24 Thibault Passeri 3, Lise Duranteau 4, Anne-Laure Bernat 3, Isabelle Yoldjian 5, Sylvie Fontanel 6, Sébastien 25 26 Froelich 3, Joël Coste 1,7 27 28 29 30 31 1. Department of Public Health Studies, French National Health Insurance, 75986 Paris cedex 20, 32 2 Scientific Interest Group - Health Product Epidemiology (ANSM-CNAM EPI-PHARE SIG), Saint-Denis, France 33 3 Department of Neurosurgery, Lariboisière Hospital, AP-HP, University of Paris, Paris, France 34 4 Gynaecology Unit and Rare Disorders of Sex Development, Bicêtre Hospital, AP-HP, Le Kremlin-Bicêtre, France 35 5 Endocrinology and gynaecology product manager, French National Agency for the Safety of Medicines and Health 36 Products(ANSM), Saint-Denis 37 6 Grand Est regional health authority, Nancy, France 38 7 Biostatistics and Epidemiology Unit – Cochin Hospital, AP-HP, Paris, France 39 40 41 42 Address for correspondence: 43 44 Corresponding author: Alain Weill ([email protected]) 45 46 Groupement d’intérêt scientifique; Epidémiologie des produits de santé GIS EPI-PHARE ANSM-CNAM 47 42 Bd de la Libération 48 93200 Saint Denis 49 Phone +331 558 742 24 50 51 Word count abstract: 410 52 Total main text for BMJ (5,550) 53 54 55 56 57 58 59 60

1 https://mc.manuscriptcentral.com/bmj BMJ Page 2 of 56

1 2 3 Abstract 4 5 6 7 8 Objectives: To assess, in a large-scale, nationwide cohort of women, the risk of meningioma 9 10 associated with exposure to high-dose cyproterone acetate, a progestin indicated for clinical 11 12 Confidential: For Review Only 13 hyperandrogenism. 14 15 Design: Observational cohort study. 16 17 Setting: Data from the French national health data system (SNDS) between 2007 and 2015. 18 19 20 Participants: All women aged 7 to 70 years, living in France, with at least one 21 22 reimbursement for high-dose cyproterone acetate with no history of meningioma or benign 23 24 brain tumour or long-term disease status. 25 26 27 Main outcome measure: The event of interest was surgery (resection or decompression) or 28 29 radiotherapy for intracranial meningioma. Relative and absolute risks of meningioma were 30 31 estimated according to the cumulative dose of cyproterone acetate. 32 33 Results: 253,777 women participated in the main study on women initiating cyproterone 34 35 36 acetate between 2007 and 2014, including 139,222 (289,544 woman-years [WY]) in the 37 38 “exposed” group and 114,555 (439,949 WY) in the “very slightly exposed” control group. 39 40 Sixty-nine in the “exposed” group and 20 meningiomas in the “very slightly 41 42 43 exposed” group were treated by surgery or radiotherapy. The incidence of meningioma in the 44 45 two groups was 23.8 and 4.5 per 100,000 WY, respectively, i.e. a crude relative risk of 5.2 46 47 [3.2-8.6] (HRa = 6.6 [4.0-11.1]). The HRa for a cumulative dose of cyproterone acetate of 48 49 50 more than 60 grams was 21.7 [10.8-43.5]. After discontinuation of exposure for one year, the 51 52 risk of meningioma in exposed women was 1.8-fold higher [1.0-3.2] than in very slightly 53 54 exposed women. We observed 485 cases of meningioma among the 131,485 women already 55 56 exposed to cyproterone acetate in 2006 (risk of 387 per 100,000 WY in the group with the 57 58 59 highest exposure in terms of cumulative dose). Over almost a decade (2007-2015), more than 60

2 https://mc.manuscriptcentral.com/bmj Page 3 of 56 BMJ

1 2 3 500 cases of treated meningioma in women in France were attributable to prolonged exposure 4 5 6 to cyproterone acetate. Meningiomas located in the anterior skull base and middle skull base, 7 8 particularly the medial third of the middle skull base, involving the spheno-orbital region, 9 10 appeared to be very specific to cyproterone acetate. 11 12 Almost Confidential: 30% of women continued orFor resumed Review cyproterone acetateOnly after surgery or 13 14 15 radiotherapy for meningioma, although cyproterone acetate was contraindicated in this setting 16 17 in 2011. 18 19 Conclusions: This study confirms and more clearly defines the magnitude of the risk of 20 21 22 meningioma associated with high-dose cyproterone acetate. A strong dose-effect relationship 23 24 was observed, together with a marked reduction of the risk after discontinuation of treatment 25 26 for at least one year, which constitute two major arguments in favour of a biologically 27 28 29 plausible causal relationship. 30 31 32 33 34 35 Key words: cyproterone acetate, meningioma, hyperandrogenism 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3 https://mc.manuscriptcentral.com/bmj BMJ Page 4 of 56

1 2 3 What is already known on this topic 4 5 6 Cyproterone acetate is a synthetic progestin with potent anti- and anti- 7 8 9 effects. 10 11 12 Since 2007,Confidential: several case series of meningiomas For have Review described patients Only who developed one or 13 14 more meningiomas after using cyproterone acetate at doses of 25 to 100 mg daily for 5 to 30 15 16 years. The first publications reported cases of meningioma occurring in transgender patients 17 18 19 in the process of male-to-female transitioning comprising hormonal therapy, then in women 20 21 treated for hirsutism or various benign gynaecological or dermatological diseases, and finally 22 23 several cases of men treated for prostate cancer or paraphilia. 24 25 26 The presence of receptors on meningiomas supports the biological plausibility 27 28 29 of an association. 30 31 32 Two very weak epidemiological studies reported contradictory results concerning the 33 34 increased risk of meningioma in women exposed to cyproterone acetate 35 36 37 38 39 40 What this paper adds 41 42 43 44 This study showed that women who initiated high-dose cyproterone acetate treatment had a 45 46 sevenfold increased risk of meningioma treated by neurosurgery or radiotherapy. 47 48 49 A strong dose-effect relationship was observed with a greater than 20-fold increased risk 50 51 beyond a cumulative exposure of 60 grams, which corresponds to treatment at a dose of 50 52 53 54 mg per day, 20 days per month for five years. 55 56 57 58 59 60

4 https://mc.manuscriptcentral.com/bmj Page 5 of 56 BMJ

1 2 3 A strong dose-effect relationship was also observed in women already exposed in 2006, at the 4 5 6 beginning of the study, with a risk of 4 per 1,000 woman-years in the group with the highest 7 8 exposure in terms of cumulative dose which corresponds to a 30-fold increased risk. 9 10 11 Meningiomas located in the anterior skull base and middle skull base, particularly the medial 12 Confidential: For Review Only 13 third of the spheno-orbital region, appeared to be very specific to exposure to high-dose 14 15 16 cyproterone acetate. 17 18 19 Regression of this risk after one year of discontinuation of cyproterone acetate constitutes an 20 21 additional argument in favour of a causal relationship, also suggesting spontaneous regression 22 23 of meningiomas that may therefore no longer require any treatment. 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

5 https://mc.manuscriptcentral.com/bmj BMJ Page 6 of 56

1 2 3 Introduction 4 5 6 7 8 Cyproterone acetate (CA) is a synthetic progestin with a potent anti-androgen action. CA has 9 10 been approved in European countries by national marketing authorisation procedures since 11 12 Confidential: For Review Only 13 the 1970s and is available on prescription under various trade names and dose strengths (1, 2, 14 15 10, 50 and 100 mg). The indications for CA also vary considerably from one country to 16 17 another. It is indicated in men (50 or 100 mg) with inoperable prostate cancer and paraphilias. 18 19 20 In women, the 50 mg dose strength is indicated in nineteen European countries including 21 22 France, , , , , The , etc. for various disorders of the 23 24 hirsutism/hyperandrogenism spectrum. In other European countries, 10 mg tablets are 25 26 27 marketed and indicated for signs of androgenization and certain forms of hirsutism 28 29 (supplementary S1: approved indications in ). CA is also used at a low dose of 2 mg 30 31 combined with 35 µg for the treatment of hirsutism and severe related to 32 33 androgen-sensitivity and/or hirsutism in women of reproductive age, and at a dose of 1 mg in 34 35 36 combination with valerate 2 mg as replacement therapy. 37 38 Since 2007, several case series have reported cases of meningioma associated with prolonged 39 40 exposure (5 to 30 years) to high doses of CA (25 to 100 mg daily) in both men and women. 41 42 43 [1–11]. The labelling of CA (10, 50 and 100 mg) was modified in 2009 [12] and, since 2011 44 45 for the brand name drug and 2013 for generics, the labelling now indicates that cases of 46 47 meningioma have been reported during prolonged use of CA [13]. A history or concomitant 48 49 50 presence of meningioma now also constitutes a contraindication to the use of CA [12,13]. 51 52 Only two very low powered epidemiological studies on the link between CA and meningioma 53 54 have been published, with conflicting results. The first study reported 4 cases, 2 men and 2 55 56 women, corresponding to a relative risk of 11.4 (4.3-30.8) [14]. A second study revealed an 57 58 59 60

6 https://mc.manuscriptcentral.com/bmj Page 7 of 56 BMJ

1 2 3 excess risk in men (4 cases) with an odds ratio of 3.3 (1.0-10.6), but concluded that CA was 4 5 6 not associated with an increased risk of meningioma in women [15]. 7 8 The primary objective of this study was to evaluate the real-life impact of prolonged exposure 9 10 to high-dose CA in women on the risk of meningioma. This study had several secondary 11 12 objectives:Confidential: 1) to evaluate the dose-effect For relationship; Review 2) to define Onlythe course of the risk of 13 14 15 meningioma after discontinuation of CA; 3) to identify specific features of cyproterone 16 17 acetate-related meningiomas; 4) to measure the effective CA discontinuation rate after 18 19 treatment of a meningioma; 5) to estimate the number of operated or radiated cases of 20 21 22 meningioma attributable to exposure to CA in France. 23 24 25 26 27 Methods 28 29 30 31 The study design was that of an “exposed/unexposed” cohort study conducted on French 32 33 34 national health data system (SNDS, formerly SNIIRAM, established in 2006) data. This study 35 36 compared the incidence of treated meningioma (event of interest) in females between the ages 37 38 of 7 and 70 years exposed to high-dose CA and in those who prematurely discontinued CA 39 40 41 and who were therefore only very slightly exposed. 42 43 The main study was conducted on a cohort of women who initiated CA between 2007 and 44 45 2014 with a follow-up until the end of 2015. The cohort was used to test, under optimal 46 47 epidemiological conditions, the hypothesis of an association between exposure and risk of 48 49 50 meningioma and to evaluate a possible dose-effect relationship. A complementary analysis 51 52 was conducted on a population of women already exposed to CA in 2006 and followed until 53 54 the end of 2015. This population was used to define, based on the hypothesis of a causal 55 56 57 association, the characteristics of cyproterone acetate-related meningiomas and the number of 58 59 operated or irradiated meningiomas attributable to CA exposure. 60

7 https://mc.manuscriptcentral.com/bmj BMJ Page 8 of 56

1 2 3 Data source 4 5 6 SNDS covers the entire population, i.e. 67 million inhabitants in France. Each person is 7 8 identified by a unique, anonymous number. Since 2006, SNDS contains information on all 9 10 outpatient (drugs, imaging, laboratory tests, etc.) and inpatient (diagnoses, procedures 11 12 performedConfidential: coded according to the common For classification Review of medical Only procedures [CCAM], 13 14 15 etc.) care. The health expenditure of patients with long-term diseases (LTDs), such as cancer, 16 17 diabetes, etc., is fully reimbursed, and their diagnosis is recorded according to ICD-10 codes. 18 19 SNDS has been extensively used in France to conduct real-life studies, especially concerning 20 21 22 the use, safety and efficacy of drugs [16–18]. Many pharmacoepidemiological studies have 23 24 been based on the use of this database [19–25]. The SNDS also contains sociodemographic 25 26 data and, when applicable, the date of death. 27 28 29 Study populations 30 31 Women satisfying the following criteria were included in the main study, conducted on the 32 33 cohort of women who initiated CA: 1) covered by the national health insurance general 34 35 36 scheme, including local mutualist sections (i.e. 87% of the population living in France); 2) 37 38 between the ages of 7 and 70 years at the time of initiation of CA; 3 ) to whom CA was 39 40 dispensed between 2007 and 2014 in the absence of any dispensing of CA in 2006. 41 42 Women satisfying criteria 1) and 2) but to whom CA was dispensed in 2006 were included in 43 44 45 a complementary cohort (“exposed in 2006”). 46 47 People administratively assigned as male and undergoing male-to-female transitioning were 48 49 not included in the study. 50 51 52 Exclusion criteria 53 54 The following women were excluded from the cohorts studied: 55 56  Women with a long-term disease [LTD] (including lupus, cancer and all other chronic 57 58 59 diseases with 100% national health insurance coverage) with LTD status starting, at 60

8 https://mc.manuscriptcentral.com/bmj Page 9 of 56 BMJ

1 2 3 the latest, one month before initiation of CA. This exclusion was justified by the fact 4 5 6 that continuous treatment with CA 50 mg daily can be proposed as a method of 7 8 contraception in women with systemic lupus erythematosus or other chronic diseases 9 10 (rheumatoid arthritis, etc.). In general, we excluded women with a serious illness 11 12 (cancer,Confidential: neurological disease, etc.) For likely toReview interfere with the Only risk of meningioma and 13 14 15 especially the therapeutic management of meningioma. 16 17  Women with a known history of “benign neoplasm of meninges”, “benign neoplasm 18 19 of brain and other parts of central nervous system” or “neoplasm of uncertain or 20 21 22 unknown behaviour of meninges” (ICD-10 3-digit codes D32, D33, D42 as principal 23 24 diagnosis [DP], related diagnosis [DR] or associated diagnosis [DAS] during a 25 26 hospital stay and/or as LTD) between 1st January 2006 and the date of start of follow- 27 28 29 up. 30 31 32 Definition of exposure and follow-up 33 34 Women in the main study cohort were considered to be “exposed” when they had received a 35 36 first dispensing of CA between 1st January 2007 and 31st December 2014 and had received a 37 38 39 cumulative dose greater than or equal to 3 grams, i.e. at least three standard packs of twenty 40 41 50 mg tablets, during the first six months following this first dispensing. The study start date 42 43 corresponded to the date of first dispensing of CA (index date). Follow-up began six months 44 45 after the first dispensing of CA. This group of “exposed” women was compared to a group of 46 47 st 48 “very slightly exposed” women, to whom CA was dispensed for the first time between 1 49 50 January 2007 and 31st December 2014 and who rapidly discontinued treatment, after having 51 52 received a cumulative dose strictly less than 3 grams, i.e. one or two standard packs dispensed 53 54 55 during the first six months following this first prescription. 56 57 58 59 60

9 https://mc.manuscriptcentral.com/bmj BMJ Page 10 of 56

1 2 3 Endpoints 4 5 6 The event of interest was neurosurgical resection, decompression or radiotherapy for one or 7 8 more intracranial meningiomas. The event of interest was identified by the combination of a 9 10 first hospitalisation for meningioma coded as the principal diagnosis (DP) or related diagnosis 11 12 Confidential: For Review Only 13 (DR) according ICD-10 as D32 “Benign neoplasm of meninges” with at least one surgical 14 15 procedure for tumour resection or a list of codes likely to be used in meningioma surgery or a 16 17 stereotactic radiosurgery procedure or a fractionated radiotherapy procedure (supplementary 18 19 material S2). 20 21 22 Follow-up 23 24 All women were followed until 31st December 2015 (end of study date) or until the first of the 25 26 following events, when it occurred before the end of the study: event of interest; “lost to 27 28 29 follow-up”, defined by a period of more than 24 months with no national health insurance 30 31 reimbursement; for the “very slightly exposed” group, date of resumption of CA; for the 32 33 “exposed” group, censorship after one year without dispensing of CA; death. 34 35 36 A sensitivity analysis was performed, in which patients could change treatment group, 37 38 therefore corresponding to time-dependent exposure. Women in the “very slightly exposed” 39 40 group for whom a new prescription of CA was identified after the start of follow-up were 41 42 43 switched to the ”exposed” group. Women in the “exposed” group who discontinued treatment 44 45 for one year were switched to the “discontinuation” group. 46 47 Covariables and confounding factors 48 49 50 The following characteristics of women and their care pathways were also considered: 51 52  Sociodemographic characteristics: age, complementary universal health insurance 53 54 (CMUc: social welfare marker of deprivation, corresponding to free complementary 55 56 57 health insurance. It is attributed under certain conditions of residence and resources 58 59 when the annual income in 2015 was less than €8,645 for a single person and €12,967 60

10 https://mc.manuscriptcentral.com/bmj Page 11 of 56 BMJ

1 2 3 for a couple. These income levels are below the poverty threshold), region of 4 5 6 residence; 7 8 9  The initial prescriber’s specialty (first prescriber identified in the database): 10 11 gynaecologist, dermatologist, endocrinologist, general practitioner or other; 12 Confidential: For Review Only 13 14  Coprescription of oestrogens (supplementary S3); 15 16 17  Presumed context of prescription: 18 19 20 21 o Reimbursed laser hair removal procedure, treatment for hirsutism, 22 23 hospitalisation with a diagnosis of hirsutism, laboratory assay corresponding to 24 25 the diagnosis of hirsutism (LH, FSH, , DHA), pelvic ultrasound; 26 27 28 o History of specific treatment for acne (drugs with ATC codes D10AB, 29 30 31 D10AD01, D10AD02, D10AD03, D10AE01, D10AF02, D10AF52, D10AX, 32 33 D10AX30, D10BA01, D10BX); 34 35 36 o Gynaecologist, dermatologist, endocrinologist consultations; 37 38 39 o The context of prescription of CA was evaluated as a function of health care 40 41 utilization. The algorithms used are presented in supplementary S4. 42 43 44 Schematically, possible prescription for hirsutism was adopted on the basis of 45 46 hospitalisation and/or LTD for hirsutism, or laser hair removal reimbursed by 47 48 national health insurance (indicated for hirsutism), or concomitant or previous 49 50 51 drug treatment for hirsutism (, , etc.) (supplementary 52 53 S5), or a laboratory work-up (FSH/LH, testosterone, DHA) and 54 55 abdominopelvic ultrasound and follow-up compatible with hirsutism. 56 57 58 59 60

11 https://mc.manuscriptcentral.com/bmj BMJ Page 12 of 56

1 2 3 o Possible prescription for acne was adopted or considered, in the absence of 4 5 6 hirsutism, on the basis of a drug treatment for acne (isotretinoin, topical 7 8 erythromycin… [codes listed above]. 9 10 11 Statistical analyses 12 Confidential: For Review Only 13 Cox proportional hazards models were used to compare the incidence of events between the 14 15 16 various groups: “exposed”, “very slightly exposed” for the main analysis and “exposed”, 17 18 “very slightly exposed” and “discontinuation” for the sensitivity analysis. 19 20 We performed adjustment for baseline characteristics (age, expressed as a continuous 21 22 23 variable, CMUc, initial prescriber’s specialty, context of indication, coprescription of 24 25 oestrogens) by only maintaining those variables significantly associated with the development 26 27 of meningioma and related to the dose of CA. 28 29 Cumulative dose and age were considered to be time-dependent variables in the analysis. 30 31 32 Results according to the anatomical location of meningioma are presented according to the 33 34 initial CCAM classification (French common classification of medical procedures), grouped 35 36 into 6 groups (anterior skull base; middle skull base; posterior skull base; convexity, not 37 38 39 involving the dural venous sinuses; convexity, involving the dural venous sinuses; falx cerebri 40 41 and tentorium) and a second more precise classification into 16 groups presented in 42 43 supplementary S2. 44 45 46 Data were analysed with SAS Enterprise Guide version 4.3 software. 47 48 49 50 Regulatory and ethical aspects 51 52 53 The SNDS (formerly SNIIRAM) is a strictly anonymous database, comprising all 54 55 reimbursement data derived from mandatory health insurance, particularly data derived from 56 57 processing of electronic or paper health care reimbursement forms and data derived from 58 59 60 health care institutions via the PMSI. This study was conducted in the context of the specific

12 https://mc.manuscriptcentral.com/bmj Page 13 of 56 BMJ

1 2 3 CNIL framework agreement: CNIL decision DE-2011-078 dated 5 September 2011 4 5 6 authorising the CNAM to perform personal health data processing in order to conduct 7 8 retrospective cohort studies designed to describe possible statistical associations between the 9 10 use of a pharmaceutical product and the development of a disease or death. 11 12 Confidential: For Review Only 13 Extraction and analysis of SNDS data were performed by CNAM or ANSM employees, 14 15 personally authorised to access the SNDS databases. 16 17 18 19 20 Patient and public involvement 21 22 No patients were involved in setting the research question or the outcome measures, nor were 23 24 they involved in developing plans for recruitment, design, or implementation of the study. No 25 26 27 patients were asked to advise on interpretation or writing up of results. 28 29 The results were presented to patient association representatives. Patients who had used 30 31 cyproterone acetate 50/100 mg over the previous two years (and their general practitioners) 32 33 34 received an information letter at home providing detailed information on what to do 35 36 (supplementary S6). 37 38 39 40 41 42 43 Results 44 45 46 47 48 A total of 253,777 women were included in the main study cohort, 139,222 (54.9%) in the 49 50 “exposed” group and 114,555 (45.1%) in the “very slightly exposed” control group (Figure 51 52 1). 53 54 55 The characteristics of the women at cohort entry are shown in Table 1. These women had a 56 57 mean age of 29.4 years and more than 40% of them were younger than 25. The initial 58 59 prescriber was a gynaecologist in more than one-half of cases (56.7%). About 30% of women 60

13 https://mc.manuscriptcentral.com/bmj BMJ Page 14 of 56

1 2 3 had undergone a laboratory work-up compatible with investigation of hirsutism and less than 4 5 6 40% had undergone pelvic ultrasound, which is essential for the investigation of hirsutism. 7 8 The presumed context of CA prescription therefore remained difficult to define: 31.6% of 9 10 prescriptions could correspond to the treatment of acne without hirsutism and 13.1% of 11 12 prescriptionsConfidential: were compatible with management For Review of hirsutism, according Only to the algorithms 13 14 15 used in this study. More than 55% of prescriptions were not associated with any markers of 16 17 acne treatment, or treatment or investigation of hirsutism. Overall, on inclusion, the 18 19 “exposed” population was very similar to the “very slightly exposed” population. In both 20 21 22 groups, treatment was predominantly initiated by gynaecologists (55.9% versus 57.8%). 23 24 Prescriptions for possible hirsutism were slightly more frequent in the “exposed” group, as 25 26 laboratory screening for possible hirsutism was performed in 36.0% and 23.2% of women, 27 28 29 respectively, and follow-up compatible with hirsutism was identified in 15.8% versus 9.8% of 30 31 women, respectively. The mean duration of follow-up was 2.1 years for “exposed” women 32 33 and 3.8 years for “very slightly exposed” women. 34 35 Age at the time of initiation of treatment was strongly associated with the risk of meningioma 36 37 38 with a hazard ratio [HR] compared to the reference 25-34 years age-group of 0.2 (95%CI: 0.0- 39 40 1.2) for women under the age of 25 years; 10.4 (4.8-22.5) for women aged 45 to 54 years and 41 42 42.3 (15.9-112.7) for women 65 years and older. In view of the major impact of age on the 43 44 45 risk of meningioma, the other patient characteristics were all tested after adjustment for age. 46 47 CMUc, the context of prescription and the prescriber’s specialty were not associated with the 48 49 risk of meningioma. In contrast, coprescription of oestrogens was significantly associated 50 51 52 with the risk of meningioma with an HR of 1.6 (1.1-2.4) (supplementary S7). 53 54 55 56 Risk of meningioma associated with CA 57 58 59 60

14 https://mc.manuscriptcentral.com/bmj Page 15 of 56 BMJ

1 2 3 Sixty-nine meningiomas in the “exposed” group and 20 meningiomas in the “very slightly 4 5 6 exposed” group were treated by surgery or radiotherapy. The incidence of meningioma in the 7 8 two groups was 23.8 per 100,000 woman-years and 4.5 per 100,000 woman-years, 9 10 respectively, i.e. a crude relative risk of 5.2 [3.2-8.6] and an adjusted hazard ratio (HR) of 11 12 6.6 [4.0-11.1].Confidential: Analysis according to the For cumulative Review dose of CA demonstrated Only a dose-effect 13 14 15 relationship with a higher risk associated with a higher cumulative dose. The HR, which was 16 17 not significantly different from 1 for exposure to less than 12 grams of CA, rapidly increased 18 19 for higher cumulative doses: HR: 11.3 [5.8-22.2] for 36 to 60 grams and 21.7 [10.8-43.5] for a 20 21 22 cumulative dose of CA of more than 60 grams (Table 2). 23 24 The risk of meningioma after discontinuation to CA for one year was 1.8-fold higher [1.0-3.2] 25 26 than the risk related to very slight or limited exposure. The risk was 4.2-fold higher [2.2-8.0] 27 28 29 when the cumulative dose of CA before discontinuation was 12 grams or higher 30 31 (supplementary S8). 32 33 34 35 In the complementary cohort of women already exposed to CA in 2006 (N=131,485), the 36 37 38 mean age was 32.1 years. These women with long-term exposure were also more often taking 39 40 oestrogens (55.5% versus 31.9%). The incidence of meningioma in the exposed population 41 42 was 141 per 100,000 woman-years (447 cases), i.e. a greater than 20-fold excess risk 43 44 45 (adjusted HR: 21.2 [12.8-35.1]). A high dose-effect relationship was also observed. Beyond a 46 47 cumulative dose of 60 grams after 2006, the incidence was 387 per 100,000 WY. 48 49 (supplementary S9). 50 51 52 53 54 Meningioma in women exposed to CA 55 56 Among the 516 women (69 from the main cohort and 447 from the complementary cohort) 57 58 exposed to CA, hospitalised and treated aggressively for meningioma, 96.0% underwent a 59 60

15 https://mc.manuscriptcentral.com/bmj BMJ Page 16 of 56

1 2 3 neurosurgical procedure and 4.0% were treated by radiotherapy (Table 3). The mean length of 4 5 6 the initial hospital stay was 10 days. The locations of operated meningiomas in these exposed 7 8 women were strikingly different from those observed in "very slightly exposed" women. The 9 10 two most common anatomical locations of meningioma among women exposed to CA 11 12 accordingConfidential: to the classification into 6 groups For were: Review 1) anterior skull Only base (n=190), and, 2) 13 14 15 middle skull base (n=130). The absolute and relative risks of meningioma in the various 16 17 anatomical regions of the brain (16 groups) differed considerably according to the level of 18 19 exposure (Supplementary Material S10). In particular, meningiomas of the anterior skull base 20 21 22 (with a 32-fold increased risk) and meningiomas of the convexity of the brain not involving 23 24 the dural venous sinuses were almost specific for prolonged exposure to CA. Inversely, the 25 26 risk of meningioma was increased 3- to 5-fold for other anatomical regions, such as the falx 27 28 29 cerebri and the tentorial incisura. 30 31 Antiepileptic drug treatment was continued for one to two years after discharge from hospital 32 33 in 29.5% of patients. All-cause 30-day and one-year mortality rates after the initial stay were 34 35 1.2% (6 cases) and 1.8% (8 cases), respectively. 36 37 38 Among the 516 cases of treated meningioma in women exposed to CA, 152 (29.5%) women 39 40 resumed CA during the year following neurosurgery or radiotherapy. Resumption of CA after 41 42 neurosurgery or radiotherapy decreased significantly over time (47.7% resumption rate in 43 44 45 2007 versus 15.0% in 2014 and 19.0% in 2015). The CA resumption rate was therefore 37% 46 47 during the period before modification of the labelling (2007-2011) versus 22% after 48 49 modification of the labelling for the brand name drug (2012-2015). 50 51 52 The estimated number of cases attributable to CA was more than 500 in France for the period 53 54 2007 to 2015 (62/year) (Supplementary Material S9). This estimate is based on a conservative 55 56 calculation not taking into account cases diagnosed more than one year after discontinuation 57 58 of CA, cases diagnosed after an initial classification of "very slightly exposed" and 59 60

16 https://mc.manuscriptcentral.com/bmj Page 17 of 56 BMJ

1 2 3 subsequent resumption of CA and meningiomas managed by medical follow-up without 4 5 6 surgery or radiotherapy (Supplementary Material S11). 7 8 9 10 11 12 Confidential: For Review Only 13 Discussion 14 15 16 Main findings of the study 17 18 This study demonstrates a strong, dose-dependent association between exposure to CA and 19 20 the risk of meningioma requiring invasive treatment. The most heavily exposed women, 21 22 23 already treated in 2006, with a cumulative dose of more than 60 grams of CA between 2007 24 25 and 2015, presented the greatest risk of meningioma, with 4 cases of meningioma per 1,000 26 27 woman-years. The risk of meningioma decreased markedly after stopping treatment for one 28 29 30 year. 31 32 Meningiomas occurring during CA exposure were located in particular sites, with an almost 33 34 32-fold excess risk of anterior skull base meningioma. The very great majority of women with 35 36 37 meningioma had taken CA for off-label indications and about 30% of them had even 38 39 continued or resumed CA after treatment of the meningioma, which had become formally 40 41 contraindicated since 2011. 42 43 44 45 46 Comparison with published data 47 48 49 50 Risk and causality 51 52 53 Only two epidemiological studies have assessed the link between CA and meningioma. The 54 55 first study, by Gil M et al., 2011, conducted on a retrospective Spanish cohort [14], identified 56 57 4 cases (2 men and 2 women) of meningioma in the group exposed to high-dose CA. The 58 59 60 absolute risk was 60 per 100,000 person-years and the relative risk (regardless of gender) was

17 https://mc.manuscriptcentral.com/bmj BMJ Page 18 of 56

1 2 3 11.4 (95%CI 4.3-30.8), corresponding to an intermediate risk between those observed in our 4 5 6 cohort and women already exposed in 2006. Due to its low power, this study did not allow 7 8 any subgroup analysis, description and adjustment for other risk factors, or a study of the 9 10 dose-effect relationship. It nevertheless confirmed, as early as 2011, an already reported 11 12 signal andConfidential: provided a very important contribution For Review to the identification Only of a probable causal 13 14 15 relationship. The second study, by Cea-Soriano L et al., 2012 [15], a case-control study 16 17 nested in a cohort of the population, was more generally designed to study 18 19 the link between hormonal therapy and meningioma. This study found an excess risk in men 20 21 22 associated with cyproterone acetate (4 cases) with an OR of 3.3 (1.0-10.6), but surprisingly 23 24 failed to show any association between cyproterone acetate and the risk of meningioma in 25 26 women. This finding was highlighted in the conclusion of the abstract. 27 28 29 This study raises a number of questions: firstly, women received a relatively low daily dose of 30 31 2 mg of CA in combination with an oestrogen. Under these conditions, the observed risk [OR 32 33 = 1.51 (0.33-6.86)] cannot be compared to that observed in our study, based on the use of 34 35 high doses (50 mg tablets, possibly scored). Moreover, CA 50 mg tablets are not used in 36 37 38 women in the United Kingdom (Table 1) [26]. The study was sponsored by a drug company, 39 40 the initial marketing authorisation holder of CA in Europe. Furthermore, two of the four 41 42 authors were salaried employees of this company and two others received unrestricted grants 43 44 45 from the company. Experience has shown that conflicts of interest with a drug company can 46 47 influence the choice of methods, study populations and finally the results, particularly in a 48 49 study of the relationships between and severe adverse effects [27]. 50 51 52 53 54 The three main identified risk factors for meningioma, apart from age and female gender, are 55 56 genetic predispositions, especially attributed to hereditary mutations of the neurofibromatosis 57 58 type 2 gene [28–31], environmental or medical exposure to high-dose ionising radiation [32– 59 60

18 https://mc.manuscriptcentral.com/bmj Page 19 of 56 BMJ

1 2 3 42] and sex hormones. Arguments in support of an association between sex hormones and 4 5 6 meningiomas include: the increased incidence of meningioma after puberty in women 7 8 compared to men (2:1), with the highest ratio (3.15:1) observed in women of childbearing 9 10 age, observations that meningiomas increase in size during the luteal phase of the menstrual 11 12 cycle andConfidential: during or regress after For delivery, Review the presence of Only oestrogen, progesterone 13 14 15 and androgen receptors in certain meningiomas, [43–50], and a moderately elevated 16 17 association between and meningiomas [51,52]. 18 19 Numerous studies have also investigated whether the use of exogenous sex hormones, such as 20 21 22 oral contraceptives and/or hormone replacement therapy (HRT), is associated with an 23 24 increased risk of meningioma. There is no evidence, at the present time, of an excess risk of 25 26 meningioma in women using oral contraceptives. No association has been established 27 28 29 between past or present use of oral contraceptives and the risk of meningioma [45,47,53]. 30 31 More consistent associations have been demonstrated with hormone replacement therapy 32 33 (HRT), although the increased risk appears to remain limited. The meta-analysis by Benson et 34 35 al., [54] concluded on a significantly increased risk of meningioma with HRT with a relative 36 37 38 risk of 1.35 [1.21-1.49]. The main published studies have reported excess risks of 39 40 meningioma ranging between 1.2 and 2.2. These various studies appear to show that the use 41 42 of progestin is an independent risk factor for the development and growth of meningiomas. In 43 44 45 any case, the magnitude of the risk observed with HRT is much lower than that observed with 46 47 high-dose CA in our study. Wielmels, in 2010 [28], called for larger-scale studies on the use 48 49 of exogenous hormones in women with meningioma, with particular attention to stratification 50 51 52 according to hormone composition (i.e. oestrogen and/or progesterone), duration and age of 53 54 use, as well as subgroups of meningiomas defined by tumour receptor expression. 55 56 In addition, the presence of progesterone receptors in arachnoid tissue, from which 57 58 meningiomas arise, has been known for four decades [55–59]. Several authors have reported a 59 60

19 https://mc.manuscriptcentral.com/bmj BMJ Page 20 of 56

1 2 3 heterogeneous distribution of progesterone receptors according to the site of meningiomas. A 4 5 6 rostrocaudal gradient of progesterone receptors, with a higher density of receptors in the skull 7 8 base, has often been reported, and was confirmed by several publications [60,61]. Our finding 9 10 of a predominance of anterior skull base meningiomas related to exposure to CA, a powerful 11 12 progestin,Confidential: is perfectly consistent with these For biological Review findings. Only 13 14 15 Peyre et al. recently compared 40 patients operated for meningioma after prolonged treatment 16 17 with progestins with 530 women with meningiomas arising in the absence of any hormone 18 19 therapy [62]. Progestin-related meningiomas, were more frequently multiple meningiomas 20 21 22 situated in the skull base (a predominant location also observed in our study). These authors 23 24 showed a higher incidence of PIK3CA and TRAF7 mutations and a lower incidence of NF2 25 26 mutations in progestin-related meningiomas compared to non-progestin-related meningiomas, 27 28 29 suggesting a hormone-induced mutational shift [62]. 30 31 Biologically, an association between exposure to high-dose CA and the development of 32 33 meningioma is therefore highly plausible. This argument in favour of a causal relationship 34 35 provides further support to the other four arguments provided by our study: the strength of the 36 37 38 association, the dose-effect relationship, the reduction of the risk after stopping treatment 39 40 (also see below) and the specificity of certain tumour locations, especially the anterior skull 41 42 base. The use of CA determines a new form of meningiomatosis, corresponding to hormone- 43 44 45 sensitive meningiomas. 46 47 48 49 Use of high-dose CA in women 50 51 52 Like the study by Bernat et al., 2015 [7], which reported a series of 12 women, only one of 53 54 whom was treated for hirsutism, with one or more meningiomas following treatment with 55 56 high-dose CA, the present study reports extensive off-label use of CA. The other reported 57 58 indications for CA [7,63,64] were acne, hyperandrogenism, contraception, dysmenorrhoea, 59 60

20 https://mc.manuscriptcentral.com/bmj Page 21 of 56 BMJ

1 2 3 polycystic ovary syndrome, alopecia and menopause. Most of the clinical cases reported in 4 5 6 the international scientific literature were concentrated in 4 countries (France, Spain, , 7 8 Belgium), reflecting very different indications and uses compared to those observed in Asian 9 10 and Anglo-Saxon countries, where the use of high-dose CA for dermatological and 11 12 gynaecologicalConfidential: indications such as contraception For Review is not approved. Only High-dose CA is not 13 14 15 currently marketed in the , due to the availability of alternative treatments. 16 17 France accounts for about 60% of all CA 50 mg sales [65] (women and men users) among 18 19 European countries with similar populations (Germany, Spain, United Kingdom, Italy). In 20 21 22 some European countries (Belgium, Cyprus, Germany, Greece, Luxembourg, The 23 24 Netherlands, and ) CA is marketed at a lower dose strength of 10 mg, with or without 25 26 the 50 mg dose strength (supplementary S1), which probably limited the risk of CA-related 27 28 29 meningioma in these countries. CA, marketed under various trade names, is extensively used 30 31 in other Latin American countries, especially and . 32 33 34 35 Clinical management of CA-related meningiomas and regression after discontinuation of CA 36 37 38 Several case series of CA-related meningiomas have reported regression or stabilization of 39 40 meningiomas after discontinuation of CA, suggesting that surgery could possibly be avoided 41 42 in these patients [2–4,6–8,11,66]. The results of the present study, which revealed a marked 43 44 45 reduction of the risk of invasive treatment for meningioma after one year of discontinuation of 46 47 CA, are consistent with the results of previous studies comprising magnetic resonance 48 49 imaging (MRI) follow-up. Discontinuation of cyproterone acetate in the presence of 50 51 52 meningioma should be considered to be a valid treatment option prior to surgery provided 53 54 close clinical and MRI surveillance with comparative measurement of tumour volumes is 55 56 ensured. [7,8,11,66] 57 58 59 60

21 https://mc.manuscriptcentral.com/bmj BMJ Page 22 of 56

1 2 3 Surgical complications of meningioma, despite its “benign” histology, are not uncommon: 4 5 6 antiepileptic drug treatment was continued for more than one year after surgery in more than 7 8 30% of patients in our study. van der Vossen S et al. [67] showed that 40% of patients 9 10 operated for meningioma experienced cognitive or emotional disorders an average of 2.5 11 12 years afterConfidential: surgery. Furthermore, surgery For for skull Review base meningioma, Only one of the predominant 13 14 15 locations of CA-related meningioma, is one of the most challenging forms of surgery and is 16 17 associated with a much higher risk than surgery of meningiomas of the convexity. 18 19 20 21 22 Strengths and limitations of the study 23 24 25 26 Strengths 27 28 29 The large cohort (1,300,000 woman-years exposed) studied here provides new information 30 31 about the dose-effect relationship, the anatomical locations of operated meningiomas, the 32 33 effect of discontinuation of treatment, and the attributable number of cases, which have not 34 35 been previously documented. 36 37 38 This study, based on the SNDS national databases, was conducted on a large unselected 39 40 population, in which exposure to CA was measured prospectively over time by pharmacists 41 42 dispensing drugs, recorded by bar code for the purposes of financial monitoring, and 43 44 45 independently of the event of interest with no recall bias, unlike older studies that were 46 47 generally based on retrospective exposure data derived from meningioma registries. The 48 49 extensive use of CA in twenty European countries makes this study even more important. 50 51 52 Another strength of this study is that it used an event of interest linking a diagnosis – 53 54 meningioma - and interventional management - neurosurgery (in 96% of cases) or 55 56 radiotherapy. This definition of the event ensures a high specificity and consequently 57 58 minimizes the effect of misclassification error on risk estimation. This is a particularly 59 60

22 https://mc.manuscriptcentral.com/bmj Page 23 of 56 BMJ

1 2 3 important point, as a small asymptomatic meningioma (not causing any neurological deficit) 4 5 6 may be discovered incidentally on imaging, which could induce differential errors when 7 8 imaging is performed more frequently in more closely monitored exposed women. The 9 10 requirement of an associated therapeutic procedure [68] restricts the study to tumours with 11 12 present orConfidential: imminent neurological consequences, For Review less subject to a Only discovery bias, and also 13 14 15 confirmed by histological examination, making the event defined in this way highly specific. 16 17 Such a definition of cases, used in the context of a pharmacoepidemiological approach, also 18 19 avoided the limitations of a pharmacovigilance approach based on heterogeneous spontaneous 20 21 22 reporting by healthcare professionals. Only a pharmacoepidemiological approach can provide 23 24 estimates of absolute risk, relative risk and attributable number of cases of an adverse event 25 26 [69]. 27 28 29 30 31 Limitations 32 33 Like all real-life observational studies, this study presents a number of limitations. First of all, 34 35 the indications for high-dose CA were poorly defined, although this does not affect the level 36 37 38 of risk, as no association was observed in our study between the context of prescription and 39 40 the risk of meningioma. However, the poorly defined indications constitute a serious 41 42 limitation in terms of determination of the benefit-risk balance of the drug, which is essential 43 44 45 in view of the very extensive off-label use of CA in France. The "Meningioma and CA" 46 47 temporary specialized scientific committee (CSST) formed by the French Medicines Agency 48 49 formulated its conclusions at the meeting held on 1st October 2018, considering, in particular, 50 51 52 that the use of CA for off-label indications such as acne, seborrhoea and moderate hirsutism 53 54 must now be formally banned [70]. 55 56 Another limitation of this study was the failure to take into account the combined oral 57 58 contraceptive pill containing 2 mg of CA, which is not reimbursed and therefore not present 59 60

23 https://mc.manuscriptcentral.com/bmj BMJ Page 24 of 56

1 2 3 in the SNDS. However, exposure to this low dose of CA, 25-fold lower than the 50 mg/day 4 5 ® 6 dosage studied here, and sometimes coprescribed with Diane , would have had only a 7 8 marginal impact on calculation of the cumulative dose of CA. 9 10 According to several published case series, the majority of meningiomas occur after 10 to 30 11 12 years of Confidential:exposure to CA. However, at theFor time ofReview this study, the SNDS Only database comprised 13 14 15 only 10 years of follow-up, which required us to constitute an "incident" cohort of women 16 17 who initiated CA between 2007 and 2014 for the main analysis and to consider women 18 19 already taking CA in 2006, and initiating treatment at an unknown date, only in the 20 21 22 complementary analyses. This approach makes presentation of the results of the study slightly 23 24 more complicated, but enhances the validity of the conclusions. 25 26 Several limitations also concern the clinical and histological characteristics of meningioma: 27 28 29 first of all, the SNDS does not directly comprise any data concerning MRI results, histology 30 31 results or the single or multiple nature of meningiomas, except in the case of operations 32 33 involving distinct anatomical regions. 34 35 Finally, the circumstances of discovery of meningiomas are not specified, but the 36 37 38 neurosurgical management, most likely according to current European guidelines [68], 39 40 reflects the symptomatic nature of the tumour related to its volume and/or location. Fewer 41 42 than 10% of treated patients had undergone MRI between 6 and 18 months before the 43 44 45 operation. The MRI results (number and size of meningiomas) are not available, but only 46 47 reimbursement of the procedure can be identified in the database. It is therefore very likely 48 49 that this study underestimates the overall incidence of intracranial meningiomas, as the 50 51 52 majority of these tumours, small or asymptomatic tumours, are simply followed by MRI. 53 54 Nevertheless, the endpoint used in this study (neurosurgical resection or radiotherapy) 55 56 corresponds to a pragmatic and clinically significant event. 57 58 59 60

24 https://mc.manuscriptcentral.com/bmj Page 25 of 56 BMJ

1 2 3 The risk of meningioma associated with other potent progestins must also be evaluated 4 5 6 according to a similar methodology, especially those molecules administered in the context of 7 8 perimenopausal hormone replacement therapy, such as acetate [71,72] and 9 10 acetate [73]. In the current state of knowledge, there is no evidence to suggest 11 12 that the riskConfidential: is of the same order of magnitude For as thatReview associated with Only cyproterone acetate. 13 14 15 16 17 Conclusion 18 19 More than 400,000 women in France used high-dose CA between 2006 and 2015, 20 21 22 predominantly for off-label indications. A strong and dose-dependent association between 23 24 exposure to CA and meningioma treated by surgery or radiotherapy was observed in these 25 26 women. Based on the available follow-up, the risk of meningioma decreased markedly after 27 28 29 stopping treatment. In the light of these results, prescription of cyproterone acetate, especially 30 31 for off-label indications, must be called into question, with more detailed assessment of the 32 33 risk/benefit balance in approved indications and more thorough screening and monitoring of 34 35 meningioma by brain imaging (MRI) when prescription of CA appears to be necessary. 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

25 https://mc.manuscriptcentral.com/bmj BMJ Page 26 of 56

1 2 3 References 4 5 6 7 1 Gazzeri R, Galarza M, Gazzeri G. Growth of a Meningioma in a Transsexual Patient after 8 –Progestin Therapy. N Engl J Med 2007;357:2411–2. 9 doi:10.1056/NEJMc071938 10 11 2 Froelich S, Dali-Youcef N, Boyer P, et al. Does cyproterone acetate promote multiple 12 meningiomas?Confidential: Endocr Abstr Proc 10th For Eur Congr Review Endocrinol Berl Only Ger 2008;16:158. 13 14 3 Gonçalves AMG, Page P, Domigo V, et al. Abrupt Regression of a Meningioma after 15 16 Discontinuation of Cyproterone Treatment: Fig 1. Am J Neuroradiol 2010;31:1504–5. 17 doi:10.3174/ajnr.A1978 18 19 4 Cebula H, Pham TQ, Boyer P, et al. Regression of meningiomas after discontinuation of 20 cyproterone acetate in a transsexual patient. Acta Neurochir (Wien) 2010;152:1955–6. 21 doi:10.1007/s00701-010-0787-2 22 23 5 Bergoglio MT, Gómez-Balaguer M, Almonacid Folch E, et al. Symptomatic meningioma 24 25 induced by cross- treatment in a male-to-female transsexual. Endocrinol 26 Nutr 2013;60:264–7. doi:10.1016/j.endonu.2012.07.004 27 28 6 Sys C, Kestelyn P. Unilateral proptosis and blindness caused by meningioma in a patient 29 treated with cyproterone acetate. GMS Ophthalmol Cases 2015;5. doi:10.3205/oc000027 30 31 7 Bernat AL, Oyama K, Hamdi S, et al. Growth stabilization and regression of 32 meningiomas after discontinuation of cyproterone acetate: a case series of 12 patients. 33 Acta Neurochir (Wien) 2015;157:1741–6. doi:10.1007/s00701-015-2532-3 34 35 36 8 Botella C, Coll G, Lemaire J-J, et al. [Intra cranial meningiomas and long term use of 37 cyproterone acetate with a conventional dose in women. A report of two cases of tumor 38 decrease after treatment withdrawal]. Neurochirurgie 2015;61:339–42. 39 doi:10.1016/j.neuchi.2015.05.002 40 41 9 Alderman CP. Probable Drug-Related Meningioma Detected During the Course of 42 Medication Review Services. Consult Pharm 2016;31:500–4. 43 doi:10.4140/TCP.n.2016.500 44 45 46 10 Borghei-Razavi H, Fragoza-Padilla V, Hargus G, et al. Meningioma: The Unusual 47 Growth in a Transsexual Patient after Estrogen-Progesterone Therapy. SOJ Neurol 48 2014;1:1–3. 49 50 11 Zairi F, Aboukais R, LE Rhun E, et al. Close follow-up after discontinuation of 51 cyproterone acetate: a possible option to defer surgery in patients with voluminous 52 intracranial meningioma. J Neurosurg Sci 2017;61:98–101. doi:10.23736/S0390- 53 54 5616.16.03243-4 55 56 12 HMA 2009. PhVWP Report on Cyproterone acetate and the risk of meningiomas 57 Provided to the CMD(h) on behalf of National Competent Authorities. Changes apply to 58 cyproterone-containing drugs with dosages higher than 2mg . 59 2009.https://www.hma.eu/fileadmin/dateien/Human_Medicines/CMD_h_/Product_Infor 60

26 https://mc.manuscriptcentral.com/bmj Page 27 of 56 BMJ

1 2 3 mation/PhVWP_Recommendations/Cyproterone/CMDh-PhVWP-012- 4 Rev0_2009_11.pdf 5 6 7 13 ANSM. Résumé des Caractéristiques du Produit ; ANSM - Mis à jour le : 20/11/2012 ; 8 ANDROCUR 50 mg, comprimé sécable. http://agence- 9 prd.ansm.sante.fr/php/ecodex/rcp/R0221731.htm 10 11 14 Gil M, Oliva B, Timoner J, et al. Risk of meningioma among users of high doses of 12 cyproteroneConfidential: acetate as compared with For the general Review population: evidence Only from a 13 population-based cohort study: Risk of meningioma among users of high doses of 14 cyproterone acetate. Br J Clin Pharmacol 2011;72:965–8. doi:10.1111/j.1365- 15 16 2125.2011.04031.x 17 18 15 Cea-Soriano L, Blenk T, Wallander M-A, et al. Hormonal therapies and meningioma: Is 19 there a link? Cancer Epidemiol 2012;36:198–205. doi:10.1016/j.canep.2011.08.003 20 21 16 Tuppin P, Rudant J, Constantinou P, et al. Value of a national administrative database to 22 guide public decisions: From the système national d’information interrégimes de 23 l’Assurance Maladie (SNIIRAM) to the système national des données de santé (SNDS) in 24 France. Rev Epidemiol Sante Publique 2017;65 Suppl 4:S149–67. 25 26 doi:10.1016/j.respe.2017.05.004 27 28 17 Bezin J, Duong M, Lassalle R, et al. The national healthcare system claims databases in 29 France, SNIIRAM and EGB: Powerful tools for pharmacoepidemiology. 30 Pharmacoepidemiol Drug Saf 2017;26:954–62. doi:10.1002/pds.4233 31 32 18 Moulis G, Lapeyre-Mestre M, Palmaro A, et al. French health insurance databases: What 33 interest for medical research? Rev Med Interne 2015;36:411–7. 34 35 doi:10.1016/j.revmed.2014.11.009 36 37 19 Maura G, Blotière P-O, Bouillon K, et al. Comparison of the short-term risk of bleeding 38 and arterial thromboembolic events in nonvalvular atrial fibrillation patients newly 39 treated with dabigatran or rivaroxaban versus vitamin K antagonists: a French nationwide 40 propensity-matched cohort study. Circulation 2015;132:1252–60. 41 doi:10.1161/CIRCULATIONAHA.115.015710 42 43 20 Weill A, Dalichampt M, Raguideau F, et al. Low dose oestrogen combined oral 44 45 contraception and risk of pulmonary embolism, stroke, and myocardial infarction in five 46 million French women: cohort study. BMJ 2016;353:i2002. doi:10.1136/bmj.i2002 47 48 21 Basson M, Mezzarobba M, Weill A, et al. Severe intestinal malabsorption associated 49 with olmesartan: a French nationwide observational cohort study. Gut 2016;65:1664–9. 50 doi:10.1136/gutjnl-2015-309690 51 52 22 Lemaitre M, Kirchgesner J, Rudnichi A, et al. Association Between Use of Thiopurines 53 or Tumor Necrosis Factor Antagonists Alone or in Combination and Risk of Lymphoma 54 55 in Patients With Inflammatory Bowel Disease. JAMA 2017;318:1679–86. 56 doi:10.1001/jama.2017.16071 57 58 23 Blotière P-O, Raguideau F, Weill A, et al. Risks of 23 specific malformations associated 59 with prenatal exposure to 10 antiepileptic drugs. Neurology Published Online First: 12 60 June 2019. doi:10.1212/WNL.0000000000007696

27 https://mc.manuscriptcentral.com/bmj BMJ Page 28 of 56

1 2 3 24 Meyer A, Rudant J, Drouin J, et al. Effectiveness and Safety of Reference Infliximab and 4 Biosimilar in Crohn Disease: A French Equivalence Study. Ann Intern Med 5 6 2019;170:99–107. doi:10.7326/M18-1512 7 8 25 Giral P, Neumann A, Weill A, et al. Cardiovascular effect of discontinuing statins for 9 primary prevention at the age of 75 years: a nationwide population-based cohort study in 10 France. Eur Heart J Published Online First: 30 July 2019. doi:10.1093/eurheartj/ehz458 11 12 26 MedicinesConfidential: and Healthcare products RegulatoryFor Review Agency. High-dose Only cyproterone acetate: 13 potential risk of (multiple) meningiomas. Patients with existing meningioma or a history 14 of meningioma must not be prescribed high-dose (≥25 mg per day) cyproterone acetate. 15 16 2009.https://www.gov.uk/drug-safety-update/high-dose-cyproterone-acetate-potential- 17 risk-of-multiple-meningiomas 18 19 27 Stegeman BH, de Bastos M, Rosendaal FR, et al. Different combined oral contraceptives 20 and the risk of venous thrombosis: systematic review and network meta-analysis. BMJ 21 2013;347:f5298. doi:10.1136/bmj.f5298 22 23 28 Wiemels J, Wrensch M, Claus EB. Epidemiology and etiology of meningioma. J 24 Neurooncol 2010;99:307–14. doi:10.1007/s11060-010-0386-3 25 26 27 29 Malmer B, Henriksson R, Grönberg H. Familial brain tumours-genetics or environment? 28 A nationwide cohort study of cancer risk in spouses and first-degree relatives of brain 29 tumour patients. Int J Cancer 2003;106:260–3. doi:10.1002/ijc.11213 30 31 30 Hemminki K, Tretli S, Sundquist J, et al. Familial risks in nervous-system tumours: a 32 histology-specific analysis from and . Lancet Oncol 2009;10:481–8. 33 doi:10.1016/S1470-2045(09)70076-2 34 35 36 31 Hansson CM, Buckley PG, Grigelioniene G, et al. Comprehensive genetic and epigenetic 37 analysis of sporadic meningioma for macro-mutations on 22q and micro-mutations within 38 the NF2 locus. BMC Genomics 2007;8:16. doi:10.1186/1471-2164-8-16 39 40 32 Hijiya N, Hudson MM, Lensing S, et al. Cumulative incidence of secondary neoplasms 41 as a first event after childhood acute lymphoblastic leukemia. JAMA 2007;297:1207–15. 42 doi:10.1001/jama.297.11.1207 43 44 33 Preston DL, Ron E, Yonehara S, et al. Tumors of the nervous system and pituitary gland 45 46 associated with atomic bomb radiation exposure. J Natl Cancer Inst 2002;94:1555–63. 47 48 34 Ron E, Modan B, Boice JD, et al. Tumors of the brain and nervous system after 49 radiotherapy in childhood. N Engl J Med 1988;319:1033–9. 50 doi:10.1056/NEJM198810203191601 51 52 35 Sadetzki S, Flint-Richter P, Ben-Tal T, et al. Radiation-induced meningioma: a 53 descriptive study of 253 cases. J Neurosurg 2002;97:1078–82. 54 doi:10.3171/jns.2002.97.5.1078 55 56 57 36 Sadetzki S, Flint-Richter P, Starinsky S, et al. Genotyping of patients with sporadic and 58 radiation-associated meningiomas. Cancer Epidemiol Biomarkers Prev 2005;14:969–76. 59 doi:10.1158/1055-9965.EPI-04-0366 60

28 https://mc.manuscriptcentral.com/bmj Page 29 of 56 BMJ

1 2 3 37 Preston-Martin S. Descriptive epidemiology of primary tumors of the spinal cord and 4 spinal meninges in Los Angeles County, 1972-1985. Neuroepidemiology 1990;9:106–11. 5 6 doi:10.1159/000110757 7 8 38 Ryan P, Lee MW, North B, et al. Amalgam fillings, diagnostic dental x-rays and tumours 9 of the brain and meninges. Eur J Cancer B Oral Oncol 1992;28B:91–5. 10 11 39 Swerdlow AJ, Cooke R, Beckers D, et al. Risk of meningioma in European patients 12 treatedConfidential: with growth hormone in childhood: For results Review from the SAGhE Only cohort. J Clin 13 Endocrinol Metab Published Online First: 17 August 2018. doi:10.1210/jc.2018-01133 14 15 40 Braganza MZ, Kitahara CM, Berrington de Gonzalez A, et al. Ionizing radiation and the 16 17 risk of brain and central nervous system tumors: a systematic review. Neuro-Oncol 18 2012;14:1316–24. doi:10.1093/neuonc/nos208 19 20 41 Taylor AJ, Little MP, Winter DL, et al. Population-based risks of CNS tumors in 21 survivors of childhood cancer: the British Childhood Cancer Survivor Study. J Clin 22 Oncol 2010;28:5287–93. doi:10.1200/JCO.2009.27.0090 23 24 42 Neglia JP, Robison LL, Stovall M, et al. New primary neoplasms of the central nervous 25 system in survivors of childhood cancer: a report from the Childhood Cancer Survivor 26 27 Study. J Natl Cancer Inst 2006;98:1528–37. doi:10.1093/jnci/djj411 28 29 43 Gurcay AG, Bozkurt I, Senturk S, et al. Diagnosis, Treatment, and Management Strategy 30 of Meningioma during Pregnancy. Asian J Neurosurg 2018;13:86–9. doi:10.4103/1793- 31 5482.181115 32 33 44 Blitshteyn S, Crook JE, Jaeckle KA. Is there an association between meningioma and 34 hormone replacement therapy? J Clin Oncol 2008;26:279–82. 35 36 doi:10.1200/JCO.2007.14.2133 37 38 45 Lee E, Grutsch J, Persky V, et al. Association of meningioma with reproductive factors. 39 Int J Cancer 2006;119:1152–7. doi:10.1002/ijc.21950 40 41 46 Wigertz A, Lönn S, Mathiesen T, et al. Risk of brain tumors associated with exposure to 42 exogenous female sex hormones. Am J Epidemiol 2006;164:629–36. 43 doi:10.1093/aje/kwj254 44 45 47 Wigertz A, Lönn S, Hall P, et al. Reproductive factors and risk of meningioma and 46 47 glioma. Cancer Epidemiol Biomarkers Prev 2008;17:2663–70. doi:10.1158/1055- 48 9965.EPI-08-0406 49 50 48 Chakravarthy V, Kaplan B, Gospodarev V, et al. Houdini Tumor: Case Report and 51 Literature Review of Pregnancy-Associated Meningioma. World Neurosurg 52 2018;114:e1261–5. doi:10.1016/j.wneu.2018.03.187 53 54 49 Hortobágyi T, Bencze J, Murnyák B, et al. Pathophysiology of Meningioma Growth in 55 Pregnancy. Open Med Wars Pol 2017;12:195–200. doi:10.1515/med-2017-0029 56 57 58 50 Kerschbaumer J, Freyschlag CF, Stockhammer G, et al. Hormone-dependent shrinkage 59 of a sphenoid wing meningioma after pregnancy: case report. J Neurosurg 60 2016;124:137–40. doi:10.3171/2014.12.JNS142112

29 https://mc.manuscriptcentral.com/bmj BMJ Page 30 of 56

1 2 3 51 Kubo M, Fukutomi T, Akashi-Tanaka S, et al. Association of breast cancer with 4 meningioma: report of a case and review of the literature. Jpn J Clin Oncol 2001;31:510– 5 6 3. doi:10.1093/jjco/hye109 7 8 52 Custer BS, Koepsell TD, Mueller BA. The association between breast carcinoma and 9 meningioma in women. Cancer 2002;94:1626–35. doi:10.1002/cncr.10410 10 11 53 Benson VS, Pirie K, Green J, et al. Lifestyle factors and primary glioma and meningioma 12 tumoursConfidential: in the Million Women Study For cohort. ReviewBr J Cancer 2008; 99Only:185–90. 13 doi:10.1038/sj.bjc.6604445 14 15 54 Benson VS, Kirichek O, Beral V, et al. Menopausal and central nervous 16 17 system tumor risk: large UK prospective study and meta-analysis. Int J Cancer 18 2015;136:2369–77. doi:10.1002/ijc.29274 19 20 55 Donnell MS, Meyer GA, Donegan WL. Estrogen-receptor protein in intracranial 21 meningiomas. J Neurosurg 1979;50:499–502. doi:10.3171/jns.1979.50.4.0499 22 23 56 Poisson M, Magdelenat H, Foncin JF, et al. [Estrogen and progestin receptors in 24 meningiomas: a study in 22 cases (author’s transl)]. Rev Neurol (Paris) 1980;136:193– 25 203. 26 27 28 57 Maiuri F, Montagnani S, Gallicchio B. Estrogen and progesterone receptors in 29 meningiomas. Surg Neurol 1986;26:435–40. 30 31 58 Goffin J. Estrogen- and progesterone-receptors in meningiomas. Review article. Clin 32 Neurol Neurosurg 1986;88:169–75. 33 34 59 Omulecka A, Papierz W, Nawrocka-Kunecka A, et al. Immunohistochemical expression 35 of progesterone and estrogen receptors in meningiomas. Folia Neuropathol 2006;44:111– 36 37 5. 38 39 60 Bouillot P, Pellissier JF, Devictor B, et al. Quantitative imaging of estrogen and 40 progesterone receptors, estrogen-regulated protein, and growth fraction: 41 immunocytochemical assays in 52 meningiomas. Correlation with clinical and 42 morphological data. J Neurosurg 1994;81:765–73. doi:10.3171/jns.1994.81.5.0765 43 44 61 Ülgen E, Bektaşoğlu PK, Sav MA, et al. Meningiomas Display a Specific 45 Immunoexpression Pattern in a Rostrocaudal Gradient: An Analysis of 366 Patients. 46 47 World Neurosurg 2019;123:e520–35. doi:10.1016/j.wneu.2018.11.201 48 49 62 Peyre M, Gaillard S, de Marcellus C, et al. Progestin-associated shift of meningioma 50 mutational landscape. Ann Oncol 2018;29:681–6. doi:10.1093/annonc/mdx763 51 52 63 HAS. COMMISSION DE LA TRANSPARENCE Avis 19 décembre 2012 Examen du 53 dossier des spécialités inscrites pour une durée de 5 ans à compter du 31 décembre 2005 54 (JO du 3 août 2007) ANDROCUR 50 mg, comprimé sécable B/ 20 (CIP : 323 510-0) 55 ANDROCUR 100 mg, comprimé sécable B/ 60 (CIP : 340 417-5) Laboratoires BAYER 56 57 SANTE. 2012. https://www.has-sante.fr/portail/upload/docs/evamed/CT- 58 10007_ANDROCUR_Avis2_CT10007_RI_PIS.pdf 59 60

30 https://mc.manuscriptcentral.com/bmj Page 31 of 56 BMJ

1 2 3 64 HAS. COMMISSION DE LA TRANSPARENCE Avis 22 juin 2016 acétate de 4 cyprotérone ANDROCUR 50 mg, comprimé sécable B/ 20 (CIP : 34009 323 510 0 8) 5 6 ANDROCUR 100 mg, comprimé sécable B/ 60 (CIP : 34009 340 417 5 4). 2016. 7 https://www.has-sante.fr/portail/upload/docs/evamed/CT- 8 14651_ANDROCUR_PIS_RI_Avis2_CT14651.pdf 9 10 65 Agence nationale de sécurité du médicament et des produits de santé (ANSM. Source 11 OCTAVE. 2019. 12 Confidential: For Review Only 13 66 Bernat AL, Bonnin S, Labidi M, et al. Regression of Giant Olfactory Groove 14 Meningioma and Complete Visual Acuity Recovery after Discontinuation of Cyproterone 15 16 Acetate. J Ophthalmic Vis Res 2018;13:355–8. doi:10.4103/jovr.jovr_21_17 17 18 67 van der Vossen S, Schepers VPM, Berkelbach van der Sprenkel JW, et al. Cognitive and 19 emotional problems in patients after cerebral meningioma surgery. J Rehabil Med 20 2014;46:430–7. doi:10.2340/16501977-1795 21 22 68 Goldbrunner R, Minniti G, Preusser M, et al. EANO guidelines for the diagnosis and 23 treatment of meningiomas. Lancet Oncol 2016;17:e383–91. doi:10.1016/S1470- 24 2045(16)30321-7 25 26 27 69 Coste J. Diverging approaches of pharmacovigilance and pharmacoepidemiology to 28 assessing drug safety: epistemological and ethical implications. Pharmacoepidemiol 29 Drug Saf 2017;26:600–2. doi:10.1002/pds.4190 30 31 70 ANSM. Meningioma and cyproterone acetate temporary specialized scientific committee 32 (CSST) report to the Agence nationale de sécurité du médicament et des produits de santé 33 (French National Agency for Medicines and Health Products Safety). 34 35 2018.https://ansm.sante.fr/S-informer/Points-d-information-Points-d- 36 information/Androcur-et-generiques-acetate-de-cyproterone-50-mg-et-100-mg-et-risque- 37 de-meningiome-l-ANSM-publie-des-recommandations-pour-la-prise-en-charge-des- 38 patients-Point-d-information 39 40 71 Champagne P-O, Passeri T, Froelich S. Combined hormonal influence of cyproterone 41 acetate and on meningioma: a case report. Acta Neurochir (Wien) 42 2019;161:589–92. doi:10.1007/s00701-018-03782-4 43 44 45 72 Passeri T, Champagne P-O, Bernat A-L, et al. Spontaneous regression of meningiomas 46 after interruption of nomegestrol acetate: a series of three patients. Acta Neurochir (Wien) 47 2019;161:761–5. doi:10.1007/s00701-019-03848-x 48 49 73 Shimizu J, Matsumoto M, Yamazaki E, et al. Spontaneous regression of an asymptomatic 50 meningioma associated with discontinuation of progesterone administration. 51 Neurol Med Chir (Tokyo) 2008;48:227–30. 52 53 54 55 56 57 58 59 60

31 https://mc.manuscriptcentral.com/bmj BMJ Page 32 of 56

1 2 3 4 Acknowledgements 5 6 7 The authors would like to thank Dr Anthony Saul for English revision of the manuscript, Prof. 8 9 Emilie Sbidian (dermatologist) for his expertise, Prof. O Lyon-Caen (neurologist), Dr R Dray- 10 11 Spira (epidemiologist), Prof. M Zureik (epidemiologist), Dr E Richer (obstetrician- 12 Confidential: For Review Only 13 14 gynaecologist), Dr M Laanani (epidemiologist), Dr A Merlo (general practitioner), L 15 16 Hoisnard (public health resident) and Prof. H Dufour (neurosurgeon) for providing us with 17 18 their comments and suggestions. 19 20 21 Footnotes 22 23 24 Contributors: SyFo, SeFr, AW, JC, and had the idea for the study, AW and JC conceived and 25 26 27 planned the study. AW drafted the manuscript. BC performed data management and BC, PN 28 29 performed statistical analyses. AW and JC ensured project and study management. All 30 31 authors (AW, PN, ML, BC, TP, LD, ALB, YL, SyFo, SeFr and JC) contributed to 32 33 34 interpretation of the data and revised the manuscript. All authors approved the final 35 36 manuscript. 37 38 39 AW and JC are the guarantors. 40 41 42 Funding: This research was funded by the French National Health Insurance Fund 43 44 (CNAMTS) and the Health Product Epidemiology Scientific Interest Group (ANSM-CNAM 45 46 47 EPI-PHARE SIG). 48 49 50 AW, JC, BC are employees of French National Health Insurance Fund, PN is an employee of 51 52 the French National Agency for Medicines and Health Products Safety. The present paper 53 54 55 represents the opinions of the authors and does not necessarily reflect the position of their 56 57 employers. 58 59 60

32 https://mc.manuscriptcentral.com/bmj Page 33 of 56 BMJ

1 2 3 Competing interests: All authors have completed the ICMJE uniform disclosure form at 4 5 6 www.icmje.org/coi_disclosure.pdf and declare: no support from any organisation for the 7 8 submitted work; no financial relationships with any organisations that might have an interest 9 10 in the submitted work in the previous three years; no other relationships or activities that 11 12 could appearConfidential: to have influenced the submitted For work. Review Only 13 14 15 16 Ethical approval: This research was authorised by the French Data Protection Agency 17 18 (CNIL). 19 20 21 Data sharing: No additional data available by author (French law to access SNDS 22 23 https://www.indsante.fr). 24 25 26 Transparency: The lead author (AW) affirms that this manuscript is an honest, accurate and 27 28 29 transparent account of the study being reported; that no important aspects of the study have 30 31 been omitted; and that any discrepancies from the study as planned (and, if relevant, 32 33 registered) have been explained. 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

33 https://mc.manuscriptcentral.com/bmj BMJ Page 34 of 56

1 2 3 Figure and tables 4 5 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Figure 1. Flow chart of the cyproterone acetate and risk of meningioma study population 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

34 https://mc.manuscriptcentral.com/bmj Page 35 of 56 BMJ

1 2 3 4 5 Table 1: Characteristics, follow-up, and modalities of exclusion from the exposed and very slightly 6 exposed women of the 2007-2014 cyproterone acetate initiation cohort: main analysis 7 8 9 Total Exposed Very slightly exposed p 10 N (%) 253,777 139,222 (55%) 114,555 (45%) 11 Age (years) at initiation 12 Confidential: For Review Only mean age (SD) 29.4 (10.9) 29.1 (10.9%) 29.9 (10.9%) <.0001 13 median age [IQR] 27 [21-37] 27 [20-36] 28 [21-37] 14 7-24 101,655 (40.1%) 58,430 (42.0%) 43,225 (37.7%) 15 25-34 76,056 (30.0%) 40,367 (29.0%) 35,689 (31.2%) 16 35-44 49,732 (19.6 26,681 (19.2%) 23,051 (20.1%) 17 45-54 20,349 (8.0%) 10,587 (7.6%) 9,762 (8.5%) 18 55 et + 5,985 (2.4%) 3,157 (2.3%) 2,828 (2.5%) <.0001 19 Year of initiation of CA 20 2007 39,481 (15.9%) 21,361 (15.3%) 18,120 (15.8%) 21 2008 36,403 (14.3%) 19,543 (14.0%) 16,860 (14.7%) 22 2009 33,427 (13.2%) 18,273 (13.1%) 15,154 (13.2%) 23 2010 33,752 (13.3%) 18,383 (13.2%) 15,369 (13.4%) 24 2011 30,120 (11.9%) 16,390 (11.8%) 13,730 (12.0%) 25 2012 26,539 (10.5%) 14,313 (10.3%) 12,226 (10.7%) 26 2013 31,327 (12.3%) 17,905 (12.9%) 13,422 (11.7%) 27 2014 22,728 (9.0%) 13,054 (9.4%) 9,674 (8.4%) <.0001 28 29 CMUc beneficiary (a) 30 Yes 18,499 (7.3%) 9,093 (6.5%) 9,406 (8.2%) 31 No 235,278 (92.7%) 130,129 (93.5%) 105,149 (91.8%) <.0001 32 Initial prescriber's specialty 33 Gynaecologist 144,018 (56.7%) 77,816 (55.9%) 66,202 (57.8%) 34 Dermatologist 29,289 (11.5%) 12,944 (9.3%) 16,345 (14.3%) 35 Endocrinologist 25,557 (10.1%) 18,171 (13.1%) 7,386 (6.4%) 36 General practitioner 46,412 (18.3%) 25,440 (18.3%) 20,972 (18.3%) 37 Other 8,499 (3.3%) 4,850 (3.5%) 3,649 (3.2%) 38 Missing data 2 (0.0%) 1 (0.0%) 1 (0.0%) <.0001 39 Associated procedures/care (b) 40 Hosp. with diagnosis of hirsutism 3,911 (1.5%) 3,017 (2.2%) 894 (0.8%) <.0001 41 Reimbursement for laser hair removal 2,194 (0.9%) 1,641 (1.2%) 553 (0.5%) <.0001 42 Lab tests (FSH/LH, testosterone, DHA) 76,729 (30.3%) 50,104 (36.0%) 26,625 (23.2%) <.0001 43 Pelvic ultrasound 98,206 (38.8%) 56,278 (40.4%) 41,928 (36.6%) <.0001 44 Anti-acne medication 94,379 (37,2%) 50,449 (36,2%) 43,930 (38,4%) <.0001 45 Duration of follow-up (years) 46 mean duration (SD) 2.9 (2.3) 2.1 (1.6) 3.8 (2.6) 47 <.0001 median duration [IQR] 2.0 [1.0-4.4] 1.4 [1.0-2.5] 3.7 [1.5-6.1] 48 <.0001 < 2 years 128,428 (50.6) 92,024 (66.1%) 36,404 (31.8%) 49 [2 years; 5 years[ 73,064 (28.8) 36,820 (26.4%) 36,244 (31.6%) 50 ≥ 5 years 52,285 (20.6) 10,378 (7.5%) 41,907 (36.6%) 51 <.0001 woman-years (WY) 729,493 289,544 439,949 52 53 Modality of exclusion from the cohort 54 at 31/12/2015 117,788 (46.4%) 32,143 (23.1%) 85,645 (74.8%) <.0001 55 death 242 (0.1%) 77 (0.1%) 165 (0.1%) <.0001 56 event of interest 89 (0.0%) 69 (0.0%) 20 (0.0%) <.0001 57 discontinuation of exposure for the exposed group 106,933 (76.8%) - 58 lost to follow-up - - 3,429 (3.0%) 59 new dispensing for the very slightly exposed group - - 25,296 (22.1%) 60

35 https://mc.manuscriptcentral.com/bmj BMJ Page 36 of 56

1 2 3 4 5 Table 2: Incidence, relative risk and adjusted hazard ratio of meningioma according to exposure to 6 cyproterone acetate 7 8 Incidence 9 WY Cases per 100,000 RR [95%CI] HRa [95%CI] (a) 10 WY 11 Very slightly exposed (< 3 g) 439,949 20 4.5 Ref. Ref. 12 Confidential: For Review Only 13 Exposed (>= 3 g) 289,544 69 23.8 5.2 [3.2 - 8.6] 6.6 [4.0 - 11.1] 14 Cumulative dose 15 16 [3 g; 6 g [ 53,744 2 3.7 0.8 [0.2 - 3.5] 1.1 [0.3 - 4.9] 17 [6 g; 12 g [ 79,202 6 7.6 1.7 [0.7 - 4.1] 2.2 [0.9 - 5.6] 18 19 [12 g; 36 g [ 115,594 30 26.0 5.7 [3.2 -10.1] 6.4 [3.6 - 11.5] 20 21 [36 g; 60 g [ 29,390 16 54.4 12.0 [6.2 - 23.1] 11.3 [5.8 - 22.2] 22 60 g and higher 11,615 15 129.1 28.4 [14.5 - 55.5] 21.7 [10.8 - 43.5] 23 24 a Adjustment for age as time-dependent variable and oestrogens on inclusion 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

36 https://mc.manuscriptcentral.com/bmj Page 37 of 56 BMJ

1 2 3 Table 3: Global description of meningioma in the main cohort (2007-2014) and complementary 4 analysis with women already exposed to cyproterone acetate in 2006 5 6 7 8 Very slightly Exposed P 9 exposed 10 11 N (%) 516 (100%) 36 (100%) 12 Initial treatmentConfidential: For Review Only 13 neurosurgery 495 (95.9%) 32 (88.9%) 14 radiotherapy 21 (4.1%) 4 (11.1%) >=.05 15 Age at diagnosis 16 mean age (SD) 48.1 (9.5) 50.5 (13.1) >=.05 17 median age (IQR) 47.1 [42.1-53.6] 50.4 [40.2-58.6] 18 7-24 1 (0.2%) 0 (0.0%) 19 25-34 33 (6.4%) 4 (11.1%) 20 35-44 170 (32.9%) 9 (25.0%) 21 45-54 204 (39.5%) 11 (30.6%) 22 55 -64 78 (15.1%) 6 (16.7%) 23 65 and older 30 (5.8%) 6 (16.7%) >=.05 24 Cumulative dose 25 26 < 3 g 0 (0.0%) 36 (100.0%) 27 [3 g; 6 g [ 2 (0.4%) 0 (0.0%) 28 [6 g; 12 g [ 14 (2.7%) 0 (0.0%) 29 [12 g; 36 g [ 149 (28.9%) 0 (0.0%) 30 [36 g; 60 g [ 141 (27.3%) 0 (0.0%) 31 60 g and higher 210 (40.7%) 0 (0.0%) <.0001 32 Anatomical location of meningioma 33 Anatomical locations in 6 groups according to CCAM 34 Group 1: Anterior skull base 190 (36.8%) 7 (19.4%) 35 Group 2: Middle of skull base 130 (25.2%) 9 (25.0%) 36 Group 3: Posterior skull base 20 (3.9%) 3 (8.3%) 37 Group 4: Convexity, not involving dural venous sinuses 107 (20.7%) 4 (11.1%) 38 Group 5: Convexity, involving dural venous sinuses 19 (3.7%) 2 (5.6%) 39 Group 6: Falx and tentorium 29 (5.6%) 7 (19.4%) 40 Missing (no surgery) 21 (4.1%) 4 (11.1%) <.01 41 42 Anatomical location in 16 groups according to CCAM 43 44 Anterior skull base (a) 120 (23.3%) 3 (8.3%) 45 Optic chiasma and/or hypothalamic region 70 (13.6%) 4 (11.1%) 46 Middle skull base 47 (9.1%) 3 (8.3%) 47 Medial third of middle skull base: spheno-orbital region 81 (15.7%) 4 (11.1%) 48 Sella turcica 2 (0.4%) 2 (5.6%) 49 Cerebellopontine angle and/or internal auditory meatus 6 (1.2%) 2 (5.6%) 50 Brain convexity, invading a dural venous sinus 6 (1.2%) 0 (0.0%) 51 Clivus 4 (0.8%) 0 (0.0%) 52 Jugular foramen 2 (0.4%) 0 (0.0%) 53 Foramen magnum 2 (0.4%) 0 (0.0%) 54 Petroclival region 0 (0.0%) 1 (2.8%) 55 Brain convexity, not involving dural venous sinuses 107 (20.7%) 4 (11.1%) 56 Brain convexity, invading a dural venous sinus 19 (3.7%) 2 (5.6%) 57 Falx cerebri 19 (3.7%) 5 (13.9%) 58 Tentorium cerebelli 4 (0.8%) 0 (0.0%) 59 Tentorial incisure 6 (1.2%) 2 (5.6%) NC 60 Missing (no surgery) 21 (4.1%) 4 (11.1%)

37 https://mc.manuscriptcentral.com/bmj BMJ Page 38 of 56

1 2 3 MRI 4 between 6 and 18 months before the event 52 (10.1%) 8 (22.2%) <.05 5 one year after the event 477 (96.3%) 31 (86.1%) >=.05 6 Length of hospital stay for initial treatment 7 mean duration (SD) in days 10.3 9.1 >=.05 8 median duration [IQR] in days 8 [7-12] 7 [6-10] 9 less than 7 days 128 (24.8%) 11 (30.6%) 10 7 to 9 days 193 (37.4%) 15 (41.7%) 11 10 days and longer 195 (37.8%) 10 (27.8%) >=.05 12 Confidential: For Review Only Death 13 14 death 0 to 30 days 6 (1.2%) 0 (0.0%) >=.05 15 death 0 days to 1 year 8 (1.8%) 0 (0.0%) >=.05 16 Antiepileptic drug treatment 17 date of discharge to 1 year 297 (57.6%) 26 (72.2%) >=.05 18 1 year to 2 years after discharge from hospital 152 (29.5%) 13 (36.1%) >=.05 19 Hospitalisation for seizures 20 date of discharge to 1 year 9 (1.7%) 0 (0.0%) >=.05 21 1 year to 2 years after initial discharge from hospital 9 (1.7%) 0 (0.0%) >=.05 22 Neurosurgery reoperation 23 date of discharge to 1 year 11 (2.1%) 2 (5.6%) >=.05 24 1 year to 2 years after the initial operation 10 (1.9%) 0 (0.0%) >=.05 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

38 https://mc.manuscriptcentral.com/bmj Page 39 of 56 BMJ

1 2 3 Supplementary Material 4 5 6 7 8 9 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1 https://mc.manuscriptcentral.com/bmj BMJ Page 40 of 56

1 2 3 Supplementary Material S1 4 5 Table 1A: List of approved indications in women and marketing of high-dose cyproterone acetate (10 6 7 mg and 50 mg by country in Europe 8 9 10 11 Country Approved indications 12 Confidential: For Review Only Cyproterone acetate 10 mg tablet 13 14 Moderately severe signs of androgenization in the woman, e.g. moderately severe hirsutism; moderately severe androgenetic alopecia; severe and moderately severe forms of acne and seborrhoea. In 15 , Belgium, Cyprus, moderately severe forms of acne and seborrhoea supplementary administration of Androcur tablets is 16 Greece, Luxembourg, Portugal indicated in cases where the clinical picture is refractory to other treatments and no satisfactory results 17 have been achieved with Diane-35 alone. 18 19 Distinct androgenic manifestations in women requiring hormone treatment: severe forms of acne, if 20 associated with inflammation and nodules (acne papulopustulosa, acne nodulocystica) or there is a risk 21 of scarring, moderate to severe forms of hirsutism, moderate to severe forms of androgenetic alopecia. 22 Germany If CPA at lower doses than part of a low-dose cyproterone acetate-estrogen combination 2 mg / 0.030 mg or 2 mg / 0.035 mg (combined oral contraceptive, pill) has not been shown to be medically effective 23 or if another antiandrogenic therapy is used was not effective. When treating acne, hormone treatment 24 should be balanced against systemic antibiotic treatment. 25 26 Idiopathic hirsutism in women when other therapies have failed. Treatment of severe forms of acne in Netherlands 27 women if a treatment with low doses of cyproterone acetate has not lead to improvement. 28 29 Cyproterone acetate 50 mg Tablet 30 31 Severe signs of androgenization, e.g. very severe hirsutism, severe androgenetic alopecia, often attended Austria by severe forms of acne and/or seborrhoea, as second line therapy if low dosed cyproterone acetate 32 (e.g. Diane 35) has no benefit. 33 34 Belgium, Bulgaria, , 35 Czech Republic, , 36 Germany, , , Severe signs of androgenization, e.g. very severe hirsutism, severe androgenetic alopecia, often attended Luxembourg, Netherlands, by severe forms of acne and/or seborrhoea. 37 Poland, , , 38 Spain 39 40 Estonia Severe symptoms of androgenization for women in fertile age. 41 42 Severe non-neoplastic hirsutism in women (idiopathic, polycystic ovary syndrome), with a major impact France 43 on emotional and social life 44 Severe signs of androgenization, e.g. very severe hirsutism with severe androgenetic alopecia, often 45 Slovenia 46 attended by severe forms of acne and/or seborrhoea. 47 48 Sweden Pronounced hirsutism in women of childbearing age 49 50 51 52 53 54 55 56 57 58 59 60

2 https://mc.manuscriptcentral.com/bmj Page 41 of 56 BMJ

1 2 3 4 Table 1 B: List of approved indications in men and marketing of high-dose cyproterone acetate (50 mg and 100 5 mg) by country in Europe 6 7 8 Country Approved indications 9 Cyproterone acetate 50 mg tablet 10 11 Belgium, Bulgaria, Croatia, 12 Cyprus, CzechConfidential: Republic, Estonia, For Review Only Finland, Germany, Greece, Italy, treatment in inoperable carcinoma of the prostate. 13 Latvia, Lithuania, Luxembourg, 14 Poland, Romania, Slovakia, 15 Slovenia 16 17 France Palliative anti-androgenic treatment of prostate cancer 18 Palliative treatment of metastasized or locally progressed inoperable carcinoma of the prostate, when treatment 19 with GnRH-analogues or surgical intervention has proven to be ineffective, is contra-indicated or when oral therapy 20 Netherlands is preferable. Initially to reduce "flare", caused by initial increased levels of serum testosterone at the onset of GnRH-agonist treatment. For the treatment of "hot flushes" occurring during treatment with GnRH- or after 21 orchidectomy. 22 Treatment of advanced hormone dependent carcinoma of the prostate. Reduction of androgen levels at the 23 Spain beginning of treatment with GnRH agonists. Treatment of hot flushes caused by decreased levels of . 24 25 Austria, Belgium, Bulgaria, 26 Croatia, Cyprus, Czech Republic, Estonia, Finland, Germany, Italy, 27 Reduction of sex drive in sexual deviations in men. 28 Latvia, Lithuania, Luxembourg, Netherlands, Norway, Poland, 29 Romania, Slovakia, Slovenia 30 31 Greece Reduction of sex drive in sexual deviations in adult men. 32 33 , United Kingdom Control of libido in severe hypersexuality and/or sexual deviation in the adult male. 34 35 Spain Reduction of increased sex drive (hypersexuality). 36 Sweden Reduction of sex drive in sexual deviations in men and/or hypersexuality in men 37 38 Cyproterone acetate 100 mg tablet 39 Czech Republic, , 40 Antiandrogen treatment in inoperable carcinoma of the prostate. , Italy, Portugal, Slovenia 41 42 Management of patients with prostatic cancer (1) to suppress "flare" with initial LHRH analogue therapy, (2) in long- term palliative treatment where LHRH analogues or surgery are contraindicated, not tolerated, or where oral 43 United Kingdom therapy is preferred, and (3) in the treatment of hot flushes in patients under treatment with LHRH analogues or 44 who have had orchidectomy. 45 Palliative antiandrogen treatment in inoperable carcinoma of the prostate, including the reduction of hot flushes Austria 46 during the therapy with Gn-RH agonists or after an orchiectomy. 47 France Palliative anti-androgenic treatment of prostate cancer 48 49 Ireland Reduction of drive in sexual deviations 50 51 France Reduction of sexual drive in paraphilias in association with a psychotherapeutic management 52 53 54 55 56 57 58 59 60

3 https://mc.manuscriptcentral.com/bmj BMJ Page 42 of 56

1 2 3 4 Supplementary Material S2 (table A, B, and C) 5 6 Table A Classification of anatomical regions according to the site of resection algorithm 7 8 9 Classification of anatomical regions according to site of resection Six groups Sixteen groups

10 CCAM procedure Code Code 6 Code 16 11 12 Confidential: For Review Only Resection of anterior skull base tumours via bilateral frontal ACFA015 1 11 13 craniotomy 14 15 Resection of anterior skull base tumours via bilateral frontal 16 ACFA026 1 11 17 craniotomy and an ethmoidal approach 18 19 Resection of anterior skull base tumours via unilateral frontal ACFA001 1 11 20 craniotomy 21 22 Resection of tumours of the optic chiasma and/or hypothalamic 23 ACFA022 1 12 24 region via craniotomy 25 26 27 Resection of middle skull base tumours via craniotomy ACFA011 2 21 28 29 Resection of tumours of the medial third of the middle skull base 30 ACFA013 2 22 involving the spheno-orbital angle via craniotomy 31 32 33 Resection of lesions of the sella turcica via transsphenoidal KAFA001 2 23 34 approach 35 36 37 Resection of lesions of the sella turcica via craniotomy KAFA002 2 23 38 39 Resection of tumours of the cerebellopontine angle and/or internal 40 acoustic meatus [internal auditory canal] via a retrosigmoid ACFA010 3 31 41 infraoccipital approach 42 43 Resection of tumours of the cerebellopontine angle and/or internal 44 acoustic meatus [internal auditory canal] via a suprapetrosal ACFA012 3 31 approach 45 46 Resection of tumours of the cerebellopontine angle and/or internal 47 acoustic meatus [internal auditory canal] via a translabyrinthine ACFA007 3 31 48 approach 49 50 Resection of extraparenchymal tumours of the cerebellar convexity ACFA018 3 32 51 with dural venous sinus invasion via craniotomy 52 53 Resection of extraparenchymal tumours of the cerebellar convexity ACFA008 3 32 54 without dural venous sinus invasion via craniotomy 55 56 Resection of tumours of the clivus via a transoral or nasosphenoidal 57 ACFA020 3 33 58 approach 59 60

4 https://mc.manuscriptcentral.com/bmj Page 43 of 56 BMJ

1 2 3 4 Resection of tumours of the clivus via craniotomy ACFA004 3 33 5 6 7 8 Resection of tumours of the jugular foramen via craniotomy ACFA023 3 34 9 10 Resection of tumours of the foramen magnum without vertebral 11 ACFA019 3 35 artery rerouting via craniotomy 12 Confidential: For Review Only 13 14 Resection of tumours of the petroclival region without facial nerve ACFA016 3 36 15 rerouting via a transpetrosal approach 16 17 Resection of extraparenchymal tumours of the brain convexity ACFA002 4 41 18 without dural venous sinus invasion via craniotomy 19 20 Resection of extraparenchymal tumours of the brain convexity with 21 ACFA028 5 51 22 dural venous sinus invasion via craniotomy 23 24 25 Resection of tumours of the falx cerebri via craniotomy ABFA010 6 61 26 27 Resection of tumours of the tentorium cerebelli via infratentorial 28 ABFA008 6 62 29 craniotomy 30 31 Resection of tumours of the tentorial incisure via supratentorial ABFA009 6 63 32 craniotomy 33 34 35 Classification in 6 groups 36 Group 1: Anterior skull base 37 Group 2: Middle skull base 38 Group 3: Posterior skull base 39 Group 4: Convexity without dural sinus invasion 40 Group 5: Convexity with dural sinus invasion 41 Group 6: Falx and tentorium 42 43 44 Classification in 16 groups 45 Anterior skull base (code 11) 46 Optic chiasma and/or hypothalamic region (12) 47 Middle skull base (21) 48 Medial third of the middle skull base involving the spheno-orbital angle (22) 49 Sella turcica (23) 50 Cerebellopontine angle and/or internal acoustic meatus (31) 51 Cerebellar convexity with or without dural venous sinus invasion (32) 52 53 Clivus (33) 54 Jugular foramen (34) 55 Foramen magnum (35) 56 Petroclival region (36) 57 Brain convexity without dural venous sinus invasion (41) 58 Brain convexity with dural venous sinus invasion (51) 59 Falx cerebri (61) 60 Tentorium cerebelli (62)

5 https://mc.manuscriptcentral.com/bmj BMJ Page 44 of 56

1 2 3 Tentorial incisure (63) 4 5 Table B List of other procedures likely to be used for meningioma surgery 6 7 CCAM 8 code Description of the procedure 9 10 ADPA001 Optic nerve decompression via craniotomy 11 ADPA020 Optic nerve decompression via orbital approach 12 ADPA023Confidential:Optic nerve decompression via transsinus For [transethmoidal] Review approach Only 13 ADPA016 Infraorbital nerve decompression via a direct approach 14 15 ADPA008 Facial nerve decompression via a transmastoid approach 16 ADPA011 Facial nerve decompression via a suprapetrosal approach 17 ADPA021 Facial nerve decompression via a transmastoid approach and suprapetrosal approach 18 BKFA001 Resection of orbital lesions, via lateral approach 19 20 BKFA002 Resection of orbital lesions via a conjunctival-palpebral approach 21 BKFA003 Resection of orbital lesions via a coronal approach 22 KAFA001 Resection of lesions of the sella turcica via a transsphenoidal approach 23 KAFA002 Resection of lesions of the sella turcica via craniotomy 24 25 KAFE900 Resection of lesions of the sella turcica via video-assisted transsphenoidal surgery 26 27 28 29 30 Table C : Radiotherapy procedures 31 32 CCAM Code Description of the procedure 33 Preparation for external beam radiotherapy with three-dimensional dosimetry and DVH after MRI 34 scanning, virtual simulation using the beam's eye view (BEV) function and three-dimensional ZZMP010 35 reconstruction, and creation of a personalized compensatory filter or personalized focal mask 36 and/or multi-leaf collimator configuration 37 38 Preparation for external beam radiotherapy with three-dimensional dosimetry and DVH after MRI 39 ZZMP900 scanning, virtual simulation using the beam's eye view (BEV) function and three-dimensional 40 reconstruction, and multi-leaf collimator configuration for dynamic use 41 ZZMP011 Preparation for fractionated-dose stereotactic brain irradiation 42 Preparation for single-dose stereotactic intracranial irradiation with placement of an invasive head ZZMP012 43 frame 44 45 Preparation for stereotactic external beam radiotherapy without respiratory gating with three- ZZMP016 dimensional dosimetry and DVH after multimodal numerical fusion scanning and virtual 46 simulation using the beam's eye view (BEV) function and three-dimensional reconstruction 47 48 AANL001 Single-dose stereotactic brain irradiation with invasive head frame 49 AANL002 Fractionated-dose stereotactic brain irradiation session with noninvasive head frame 50 ZANL001 Single-dose stereotactic intracranial external beam radiotherapy with invasive head frame 51 52 Single-dose, image-guided stereotactic external beam radiotherapy without respiratory gating using ZZNL049 53 a photon-producing machine 54 Single-dose, image-guided stereotactic external beam radiotherapy without respiratory gating using 55 ZZNL055 a dedicated photon-producing machine 56 57 Image-guided stereotactic external beam radiotherapy session without respiratory gating using a ZZNL058 58 photon-producing machine 59 60

6 https://mc.manuscriptcentral.com/bmj Page 45 of 56 BMJ

1 2 3 Image-guided stereotactic external beam radiotherapy session without respiratory gating using a ZZNL059 4 dedicated photon-producing machine 5 6 ZZNL045 Proton beam therapy session 7 External beam radiotherapy session using a photon-producing machine with an energy less than 5 ZZNL020 8 MV by 1 or 2 beams, without the use of a beam modifier 9 External beam radiotherapy session using a photon-producing machine with an energy less than 5 10 ZZNL021 11 MV by 1 or 2 beams, with the use of a beam modifier 12 Confidential:External beam radiotherapy session For using a photon-producingReview machine Only with an energy less than 5 ZZNL023 13 MV by 3 or 4 beams, without the use of a beam modifier 14 External beam radiotherapy session using a photon-producing machine with an energy less than 5 15 ZZNL024 16 MV by 3 or 4 beams, with the use of a beam modifier 17 External beam radiotherapy session using a photon-producing machine with an energy less than 5 ZZNL025 18 MV by 5 or more beams, without the use of a beam modifier 19 External beam radiotherapy session using a photon-producing machine with an energy less than 5 20 ZZNL026 21 MV by 5 or more beams, with the use of a beam modifier 22 External beam radiotherapy session using a photon-producing machine with an energy greater than ZZNL027 23 or equal to 5 MV and less than 15 MV by 1 or 2 beams, without the use of a beam modifier 24 External beam radiotherapy session using a photon-producing machine with an energy greater than 25 ZZNL028 26 or equal to 5 MV and less than 15 MV by 1 or 2 beams, with the use of a beam modifier 27 External beam radiotherapy session using a photon-producing machine with an energy greater than ZZNL030 28 or equal to 5 MV and less than 15 MV by 3 or 4 beams, without the use of a beam modifier 29 External beam radiotherapy session using a photon-producing machine with an energy greater than 30 ZZNL031 31 or equal to 5 MV and less than 15 MV by 3 or 4 beams, with the use of a beam modifier 32 External beam radiotherapy session using a photon-producing machine with an energy greater than ZZNL033 33 or equal to 5 MV and less than 15 MV by 5 or more beams, without the use of a beam modifier 34 External beam radiotherapy session using a photon-producing machine with an energy greater than 35 ZZNL034 36 or equal to 5 MV and less than 15 MV by 5 or more beams, with the use of a beam modifier 37 External beam radiotherapy session using a photon-producing machine with an energy greater than ZZNL036 38 or equal to 15 MV by 1 or 2 beams, without the use of a beam modifier 39 External beam radiotherapy session using a photon-producing machine with an energy greater than 40 ZZNL037 41 or equal to 15 MV by 1 or 2 beams, with the use of a beam modifier 42 External beam radiotherapy session using a photon-producing machine with an energy greater than ZZNL039 43 or equal to 15 MV by 3 or 4 beams, without the use of a beam modifier 44 External beam radiotherapy session using a photon-producing machine with an energy greater than 45 ZZNL040 46 or equal to 15 MV by 3 or 4 beams, with the use of a beam modifier 47 External beam radiotherapy session using a photon-producing machine with an energy greater than ZZNL042 48 or equal to 15 MV by 5 or more beams, without the use of a beam modifier 49 External beam radiotherapy session using a photon-producing machine with an energy greater than 50 ZZNL043 51 or equal to 15 MV by 5 or more beams, with the use of a beam modifier 52 Image-guided stereotactic external beam radiotherapy session using a dedicated photon-producing ZZNL059 53 machine, without respiratory gating 54 External beam radiotherapy session using a photon-producing machine with an energy greater than 55 ZZNL900 or equal to 5 MV and less than 15 MV by 1 or 2 beams, with dynamic use of a multi-leaf 56 collimator [intensity modulation] 57 External beam radiotherapy session using a photon-producing machine with an energy greater than 58 ZZNL902 or equal to 5 MV and less than 15 MV by 5 or more beams, with dynamic use of a multi-leaf 59 collimator [intensity modulation] 60

7 https://mc.manuscriptcentral.com/bmj BMJ Page 46 of 56

1 2 3 External beam radiotherapy session using a photon-producing machine with an energy greater than 4 ZZNL903 or equal to 15 MV by 5 or more beams, with dynamic use of a multi-leaf collimator [intensity 5 modulation] 6 External beam radiotherapy session using a photon-producing machine with an energy greater than 7 ZZNL904 or equal to 15 MV by 1 or 2 beams, with dynamic use of a multi-leaf collimator [intensity 8 modulation] 9 External beam radiotherapy session using a photon-producing machine with an energy greater than 10 ZZNL905 or equal to 5 MV and less than 15 MV by 3 or 4 beams, with dynamic use of a multi-leaf 11 collimator [intensity modulation] 12 Confidential:External beam radiotherapy session For using a photon-producingReview machine Only with an energy greater than 13 ZZNL906 or equal to 15 MV by 3 or 4 beams, with dynamic use of a multi-leaf collimator [intensity 14 modulation] 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

8 https://mc.manuscriptcentral.com/bmj Page 47 of 56 BMJ

1 2 3 Supplementary Material S3 4 5 List of ATC classes of oestrogens 6 7 8 ATC code Description of the ATC class 9 G03AA05 AND OESTROGEN 10 G03AA06 AND OESTROGEN 11 G03AA07 AND OESTROGEN 12 Confidential: For Review Only G03AA09 AND OESTROGEN 13 14 G03AA10 AND OESTROGEN 15 G03AB03 LEVONORGESTREL AND OESTROGEN 16 G03AB04 NORETHISTERONE AND OESTROGEN 17 G03CA03 OESTRADIOL 18 19 G03CA04 ESTRIOL 20 G03CA09 PROMESTRIENE 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

9 https://mc.manuscriptcentral.com/bmj BMJ Page 48 of 56

1 2 3 Supplementary Material S4 4 5 Algorithm used to attribute a specific disease to patients 6 7 Context of prescription for hirsutism algorithm 8 Laser hair removal reimbursed by French national health insurance 9 Laser hair removal procedures reimbursed and CCAM code corresponding to the treatment of hirsutism 10 CCAM procedure code Description of the procedure 11 QZNP027 Electric skin hair removal session 12 QZNP030Confidential: Skin hair removal over more than For 150 cm 2,Review with laser or flash lamp Only 13 QZNP029 Skin hair removal over less than 50 cm2, with laser or flash lamp 14 QZNP028 Skin hair removal over 50 cm2 to 150 cm2, with laser or flash lamp 15 16 or 17 Drug treatment for hirsutism (Supplementary Material S5) 18 or 19 Hospitalisation for hirsutism DP/DR or DAS codes ICD-10: L68 or E28 20 21 or 22 [at least one pelvic ultrasound 23 and at least one specific laboratory assay 24 and exclusively a gynaecology or endocrinology follow-up visit 25 and gynaecologist or endocrinologist as initial prescriber] 26 27 28 29 Context of prescription for acne algorithm 30 [Treatment of acne 31 and 32 dermatology follow-up visit 33 and no indication for hirsutism] 34 or 35 [Treatment of acne 36 and 37 a gynaecology or endocrinology follow-up visit AND a dermatology follow-up visit 38 and (gynaecology + endocrinology) / dermatology ratio less than 0.5 39 40 and 41 no indication for hirsutism] 42 or 43 Treatment of acne 44 and 45 a gynaecology or endocrinology follow-up visit AND a dermatology follow-up visit 46 and 47 dermatologist as initial prescriber 48 and 49 no indication for hirsutism] 50 or 51 [Treatment of acne 52 and follow-up visit with none of these specialists 53 and dermatologist, gynaecologist and endocrinologist as initial prescriber 54 no indication for hirsutism 55 56 57 58 59 60

10 https://mc.manuscriptcentral.com/bmj Page 49 of 56 BMJ

1 2 3 Supplementary Material S5 4 5 6 Concomitant or previous drug treatment for hirsutism 7 8 9 ATC CODE NAME 10 C03DA01 SPIRONOLACTONE 11 C03EA04 ALTIZIDE AND POTASSIUM-SPARING AGENTS 12 Confidential: For Review Only 13 G04CB01 FINASTERIDE 14 L02AE01 BUSERELIN 15 L02AE02 16 L02AE03 GOSERELIN 17 18 L02AE04 TRIPTORELIN 19 L02BB01 20 L02BB02 21 L02BB03 22 23 L02BB04 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

11 https://mc.manuscriptcentral.com/bmj BMJ Page 50 of 56

1 2 3 Supplementary Material S6: 4 5 Personalized letter sent to women to whom cyproterone acetate was prescribed and reimbursed during the 6 7 previous 24 months 8 French version of the patient letter available at: 9 https://ansm.sante.fr/content/download/162911/2130915/version/1/file/Androcur_modele_courrier_cnam_a 10 nsm_patient_190702.pdf 11 French version of the physician letter available at: 12 Confidential: For Review Only https://ansm.sante.fr/content/download/162915/2130957/version/1/file/Androcur_modele_courrier_cnam_a 13 14 nsm_PS_190702.pdf 15 16 English translation of the patient letter 17 18 Dear Madam / Dear Sir, 19 Your doctor has prescribed a drug containing cyproterone acetate in the form of 50 or 100 mg tablets 20 ® 21 (Androcur or one of its generics). Cyproterone acetate is an effective drug for the treatment of certain 22 diseases, but it can predispose to the development of meningioma. Meningioma is a tumour that arises from a 23 membrane surrounding the brain: the meninges. These tumours are usually benign. However, they can be 24 responsible for serious disorders, possibly requiring a major surgical operation. 25 26 The risk of meningioma increases as a function of the dose and duration of treatment, as a study conducted by 27 28 French national health insurance (CNAM)* showed that the risk of developing meningioma for all patients is 29 sevenfold higher after 6 months of treatment at a mean dose greater than or equal to 25 mg daily and is 20- 30 fold higher beyond a cumulative dose of 60 g, i.e. about 5 years of treatment at a dosage of 50 mg daily, or 10 31 years at a dosage of 25 mg daily. 32 33 It is therefore recommended to regularly review the justification for prescribing cyproterone acetate. When 34 35 prolonged treatment is considered to be essential, regular monitoring by brain imaging (MRI) must be ensured. 36 37 Your doctor must now give you two documents: a document providing information about this drug and 38 another document declaring that he/she has informed you about these risks. This information declaration must 39 be established and co-signed each year by your doctor and yourself; this declaration is essential in order to 40 41 obtain your drug from the pharmacy. 42 43 You are encouraged to consult your doctor if you are currently taking this drug or if you have taken it in the 44 past. Your doctor will assess whether or not brain imaging needs to be performed. If you are still taking 45 treatment, your doctor will assess with you whether or not this prescription should be continued. In the 46 meantime, you are advised not to stop or modify your treatment unless advised by your doctor. 47 48 49 The doctor who prescribed this drug to you has been informed about this situation and is going to contact you 50 for a new visit, if he/she has not already explained the situation to you. You can also contact your doctor 51 directly. 52 53 Yours sincerely 54 55 56 Professor Olivier Lyon-Caen Doctor Dominique Martin 57 Conseil National de l'Assurance Maladie Physician ANSM Director-General 58 59 For more information: www.ansm.sante.fr 60 - "Enter "Androcur and generics" in the search engine and download the following documents:

12 https://mc.manuscriptcentral.com/bmj Page 51 of 56 BMJ

1 2 3 Risk of meningioma: "Annual patient information declaration", Treatment by cyproterone acetate (Androcur 4 and generics, 50 or 100 mg tablets) and "Androcur and generics: information for patients". 5 6 7 ------8 * CNAM study: "Exposure to high doses of cyproterone acetate and risk of meningioma in women: a cohort study in France 9 from 2006 to 2015." 10 11 12 Confidential: For Review Only 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

13 https://mc.manuscriptcentral.com/bmj BMJ Page 52 of 56

1 2 3 4 Supplementary Material S7: 5 Incidence and hazard ratio of meningioma according to age, CMUc, indication, initial prescriber’s specialty and 6 coprescription of oestrogens 7 8 9 Incidence per 10 Number of Woman-years 100,000 woman- HR 95%CI events 11 years 12 Confidential: For Review Only 13 14 Age 15 7-24 193,627 1 1 0.2 0.0 1.2 16 25-34 242,606 9 4 Ref. 17 35-44 172,300 33 19 5.8 2.7 12.6 18 19 45-54 90,552 31 34 10.4 4.8 22.5 20 55-64 24,643 7 28 8.7 3.1 23.9 21 65 and older 5,765 8 139 42.3 15.9 112.7 22

23 24 CMUc 25 Yes 53,922 6 11 1.0* 0.5 2.1 26 No 675,193 83 12 Ref. 27 28 Context of prescription 29 Acne 228,382 19 8 0.9* 0.5 1.5 30 Hirsutism 85,595 12 14 1.1* 0.6 2.1 31 Not determined 415,327 58 14 Ref. 32 33 Initial prescriber’s specialty 34 Gynaecologist 419,312 59 14 1.5* 0.9 2.5 35 Dermatologist 87,813 3 3 0.7* 0.3 1.7 36 37 Endocrinologist 66,382 8 12 1.5* 0.7 3.2 38 General practitioner 130,852 17 13 Ref. 39 Other (Hospital/other specialty) 24,936 2 8 1.0* 0.3 3.5 40 41 Coprescription of oestrogens 42 Yes 297,205 48 16 1.6* 1.1 2.4 43 No 432,099 41 9 Ref. 44 * Age-adjusted HR 45 CMUc: social welfare marker of deprivation, corresponding to free complementary health insurance. It is attributed under certain conditions 46 of residence and resources when the annual income in 2015 was less than €8,645 for a single person and €12,967 for a couple. These income 47 levels are below the poverty threshold 48 49 50 51 52 53 54 55 56 57 58 59 60

14 https://mc.manuscriptcentral.com/bmj Page 53 of 56 BMJ

1 2 3 Supplementary Material S8: 4 Table: Incidence, relative risk and adjusted hazard ratio for meningioma after discontinuation of cyproterone 5 acetate (2007-2014 incident cohort) 6 7 8 Incidence per HRa age and oestrogen WY Cases RR [95%CI] 9 100,000 WY [95%CI] 10 11 Very slightly exposed (<3 g) 440 20 4.5 Ref. Ref. 12 Confidential: For Review Only Exposed (>= 3 g) 13 336 73* 21.7 4.8 [2.9-7.8] 5.6 [3.4-9.3] 14 Exposure discontinued 15 366 34 9.3 2.0 [1.2-3.6] 1.8 [1.0-3.2] 16 Exposure discontinued; according to cumulative dose until discontinuation 17 [3 g; 12 g[ 18 276 16 6 1.3 [0.7-2.5] 1.3 [0.6-2.4] 19 12 g and higher 20 90 18 20 4.4 [2.3-8.3] 4.2 [2.2-8.0] 21 22 *The difference in the number of cases, 73 cases (in this Table) in exposed subjects (three-group sensitivity analysis) and 69 cases (Table 2) 23 in exposed subjects (two-group main analysis) can be explained by management of exposure as a time-dependent variable in the three- group analysis, as 4 women initially classified in the "very slightly exposed" group subsequently resumed cyproterone acetate and 24 developed a meningioma treated by surgery or radiotherapy, resulting in a total of 73 cases (follow-up of these 4 women was censored at 25 the time of resumption of cyproterone acetate in the main analysis). 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

15 https://mc.manuscriptcentral.com/bmj BMJ Page 54 of 56

1 2 3 Supplementary Material S9: 4 Table: Incidence, relative risk and adjusted hazard ratio for meningioma according to exposure to cyproterone 5 6 acetate: women already exposed in 2006 7 8 9 Incidence per WY Cases RR [95%CI] HRa [95%CI]a 10 100,000 WY 11 12 Confidential: For Review Only Very slightly exposed (<3 g) 221,299 16 7.2 Ref. Ref. 13 14 Exposed (≥ 3 g) 317,084 447 140.9 19.5 [11.8-32.1] 21.2 [12.8-35.1] 15 16 according to the cumulative dose after 2006 17 [3 g; 6 g [ 26,026 0 0 - - 18 19 [6 g; 12 g [ 55,398 8 14.4 2.0 [0.9-4.7] 5.0 [2.0-12.5] 20 [12 g; 36 g [ 125,083 119 95.1 13.2 [7.8-22.2] 12.7 [7.4-22.0] 21 22 [36 g; 60 g [ 59,638 125 209.6 29.0 [17.2-48.8] 19.5 [11.5-33.1] 23 60 g and higher 50,94 195 382.8 53.0 [31.8-88.2] 31.1 [18.4-52.0] 24 a adjustment for oestrogen at the time of initiation and age as a time-dependent variable. 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

16 https://mc.manuscriptcentral.com/bmj Page 55 of 56 BMJ

1 2 3 Supplementary Material S10: 4 Table: Description of the anatomical locations of meningioma (≥ 6 cases) treated by neurosurgical resection 5 according to exposure to cyproterone acetate (cohort and women already exposed in 2006) 6 7 8 Very slightly Exposed group 9 exposed group Relative risk 10 Anatomical location of meningioma treated exposed vs. very HRa 95%CI by resection 11 Inc. per Inc. per slightly exposed Number Number 12 Confidential: Formillion Reviewmillion Only of cases of cases 13 WY WY 14 15 Anterior skull base 120 198 3 5 43.6 [13.9-137.1] 32.5 [11.9- 89.1] 16 17 Brain convexity, not involving dural venous 107 176 4 6 29.6 [10.7-79.1] 32.5 [11.9-89.1] 18 sinuses 19 20 Medial third of middle skull base, involving 81 134 4 6 22.1 [8.1-60.2] 25.1 [9.1-69.3] 21 the spheno-orbital region 22 23 24 Optic chiasma and/or hypothalamic region 70 115 4 6 19.1 [7.0-52.2] 17.6 [6.3-48.8] 25 26 Middle skull base 47 77 3 5 17.1 [5.3-54.9] 22.8 [7.0-74.5] 27 Brain convexity, invading a dural venous 28 19 31 2 3 10.4 [2.4-44.5] 9.5 [2.1-42.4] 29 sinus 30 31 Falx cerebri 19 31 5 8 4.1 [1.5-11.1] 3.8 [1.4-10.7] 32 33 Tentorial incisure 6 10 2 3 3.3 [0.7-16.2] 5.2 [1.0-26.9] 34 Cerebellopontine angle and/or internal 6 10 2 3 3.3 [0.7-16.2] 4.7 [0.9-25.0] 35 auditory meatus 36 37 Cerebellar convexity, invading or not 6 10 0 0 Infinity 38 invading the dural venous sinuses 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

17 https://mc.manuscriptcentral.com/bmj BMJ Page 56 of 56

1 2 3 Supplementary Material S11 4 Estimation of the number of cases of meningioma attributable to exposure to cyproterone acetate in France 5 6 between 2007 and 2015 7 8 9 By assuming a causal relationship between exposure to cyproterone acetate and meningioma, the 10 11 attributable fraction (AF) and the attributable number of cases (ANC) of meningioma among women 12 Confidential: For Review Only exposed to cyproterone acetate can be estimated by means of the classical formula: AF = (RR-1)/RR. 13 14 The AF for the incident cohort is (6.6-1)/6.6), i.e. 0.848. 15 16 The AF for the prevalent cohort is (21.2-1)/21.2), i.e. 0.953. 17 The study population only concerned general scheme and local mutualist section beneficiaries, i.e. 18 19 87% of the population living in France. 20 The attributable number of cases can be estimated as follows: ANC = ((69x0.848) + 21 22 (447x0.953))/0.87) = 556.9, i.e. more than 500 attributable cases for the period 2007 to 2015 (62/year) 23 24 based on a conservative calculation not taking into account cases diagnosed more than one year after 25 discontinuation of cyproterone acetate, cases diagnosed after an initial classification of "very slightly 26 27 exposed" and subsequent resumption of cyproterone acetate and meningiomas managed by medical 28 follow-up without neurosurgery or radiotherapy. 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

18 https://mc.manuscriptcentral.com/bmj