IJC International Journal of Cancer

Transcription

1 IJC International Journal of Cancer Postmenopausal hormone therapy and the risk of breast cancer in Norway Marta Roman 1,2, Solveig Sakshaug 3, Sidsel Graff-Iversen 4, Siri Vangen 2, Elisabete Weiderpass 1,5,6,7, Giske Ursin 1,8 and Solveig Hofvind 1,9 1 Cancer Registry of Norway, Oslo, Norway 2 Women and Children s Division, National Advisory Unit for Women s Health, Oslo University Hospital, Oslo, Norway 3 Department of Pharmacoepidemiology, Norwegian Institute of Public Health, Oslo, Norway 4 Department of Chronic Diseases, Norwegian Institute of Public Health, Oslo, Norway 5 Department of Community Medicine, Faculty of Health Sciences, University of Tromsø, the Arctic University of Norway, Tromsø, Norway 6 Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden 7 Department of Genetic Epidemiology, Folkh alsan Research Center, Helsinki, Finland 8 Department of Preventive Medicine, University of Southern California, Los Angeles, CA 9 Oslo and Akershus University College of Applied Sciences, Faculty of Health Science, Oslo, Norway There is convincing evidence that combined estrogen progestin therapy (EPT) increases the risk of breast cancer. However, the effect of different formulations, preparations and routes of administration is largely unknown. Estrogen only-therapy (ET) is, in general, not associated or weakly associated with breast cancer risk. We investigated the effect of hormone therapy (HT) with ET, EPT, and tibolone on risk of invasive breast cancer. Information on HT use was obtained from the Norwegian Prescription Database, and breast cancer incidence from the Cancer Registry of Norway. Poisson regression was used to estimate the incidence rate ratios (RR). We analyzed data from 686,614 Norwegian women, aged years in January 2004, followed until December 2008, of whom 178,383 (26%) were prescribed HT. During the average 4.8 years of follow-up, 7,910 invasive breast cancers were registered. Compared with nonusers, current users of estradiol norethisterone acetate (NETA)(EPT) had a RR of 2.74 (95% CI: ). Users of the high dose estradiol NETA formulation Kliogest VR had a RR of 3.26 (95% CI: ), while users of the low dose Activelle VR had a RR of 2.76 (95% CI: ). Current users of tibolone had a RR of 1.91 (95% CI: ). Current users of ET with oral or transdermal estradiol had a RR of 1.40 (95% CI: ), and 1.40 (95% CI: ), respectively. The increased incidence rates approximates one extra invasive breast cancer case diagnosed for every 259 women using estradiol NETA for one year, and one extra case for every 475 women using tibolone. In conclusion, use of estradiol NETA and tibolone preparations is associated with an increased breast cancer risk. There have been a number of epidemiological and experimental studies of hormone therapy (HT) and breast cancer risk There is now convincing evidence that combined estrogen progestin therapy (EPT) is associated with a substantially higher risk of breast cancer than preparations containing estrogen alone (ET). 4,5,12 14 The effect of tibolone is less investigated. 10,15,16 In 2012, the International Agency for Research on Cancer classified the evidence for an increased risk of breast cancer as sufficient for EPT and limited for ET. 17,18 Key words: breast neoplasms, hormonal therapy, estrogen, progestin, tibolone, female DOI: /ijc History: Received 13 Jan 2015; Accepted 30 July 2015; Online 19 Aug 2015 Correspondence to: Solveig Hofvind, Cancer Registry of Norway, Oslo, Norway or Oslo and Akershus University College of Applied Sciences, Faculty of Health Science, Majorstua 0403, Oslo, Norway, Tel: 1[ ], Solveig.hofvind@kreftregisteret.no There is less evidence as to whether risk differs between continuous and sequential EPT regimens, 3,10,13,15,19 21, or whether dose or route of administration matters. 10,21 Further, there is limited documentation on whether type of estrogen or progestin preparation matter, specifically whether EPT combinations containing androgenic progestin are associated with higher risk than other EPTs. In addition, there is limited evidence as to whether estrogen-like compounds, such as tibolone, increase the risk. 17,18 For women considering HT treatment of menopausal symptoms, there is a strong need for more knowledge to clarify the risk of breast cancer associated with different HT preparations. We used information from the Norwegian population-based registries to assess some of the remaining questions on use of HT and breast cancer risk, such as the impact of different regimens, types of estrogen, strength, and route of administration. Methods Study population All Norwegian citizens are assigned an 11-digit unique personal identification number at birth or immigration. The Int. J. Cancer: 138, (2016) VC 2015 UICC

2 Roman et al. 585 What s new? Women who are considering treatment for menopausal symptoms must weigh the risks and benefits carefully, as combined estrogen progestin therapy can significantly elevate the risk of breast cancer. But, as this study shows, not all types of estrogen progestin therapy are created equally when it comes to breast cancer risk. Data from the Norwegian Prescription Database and the Cancer Registry of Norway shows that among women aged years, breast cancer risk increases threefold in association with use of estradiol norethisterone and increases almost twofold with the use of the synthetic steroid tibolone. Oral and transdermal estrogen therapies showed only moderate risk increases. personal identification number allows univocal linkage between national registries. To explore the influence of HT on the risk of breast cancer, we linked information from national registries with information about date of birth, immigration and emigration status, death, cause of death, education level and actual residence (Statistics Norway and the Population Registry), redeemed prescriptions (the Norwegian Prescription Database) and cancer cases (the Cancer Registry of Norway). Cancer reporting has been mandatory by law in Norway since The overall completeness of solid cancers was estimated at 98.8% for the registration period , and 99.3% of the breast cancer cases were histologically proven. 23 The study was approved by the regional ethics committee in the South East region of Norway, and concession to data linkage was granted by the Norwegian Data Protection Authority. All women born between 1925 and 1959, alive and residing in Norway as of January 1, 2004 (aged years) were included in the study group. Among the 895,281 initially identified women, we excluded 94,333 who were born outside Norway (Fig. 1). The exclusion was done to ensure reliable information on previous history of cancer and number of childbirths. Further, we excluded women registered with a cancer diagnosis before January 2004 (n 5 52,074), and those with a breast cancer diagnosis during the first 3 months of the study period (January to March 2004; n 5 531), to ensure a minimum latency time in case of HT use. Furthermore, we also excluded women prescribed sex hormones other than ET, EPT or tibolone during the study period (n 5 33,299), to avoid a possible misclassification. Lastly, we excluded women who only redeemed one prescription dated after June 2004 (n 5 28,430), because we assumed the short duration of use to not affect breast cancer risk. This left 686,614 for analyses. We followed the women for incident breast cancer and use of HT until December 31, vaginal or injection). We retrieved data on use of any sex hormones (ATC group G03) in the period Use of HT was defined from redeemed prescriptions for ET (ATC group G03C) and EPT (ATC group G03F). Postmenopausal hormone use HT in Norway includes components of natural estrogens (estradiol and estriol) and tibolone (estrogenic, progestogenic and weak androgenic activity). To prevent risk of endometrial cancer in women with an intact uterus, progestins are added to the estrogen therapy. Progestins can be added on a continuous or sequential regimen. All continuous and sequential combined regimens of estrogen progestin available in The Norwegian prescription database The Norwegian Prescription Database includes detailed information about all redeemed prescriptions in Norway on an individual level since January The registry includes information on the date of the redeemed prescription, specific Anatomical Therapeutic Chemical (ATC) code, number of packages redeemed, package size, brand name, strength of the dosage and route of administration (oral, transdermal, Figure 1. Flow chart of study participants.

3 586 Hormone therapy and risk of breast cancer Figure 2. Follow-up definition of study participants. Norway contain estradiol and norethisterone acetate (Activelle VR, Kliogest VR, Trisekvens VR, Trisekvens forte VR, Novofem VR and Sequidot VR ). Use of other progestin types, such as medroxyprogesterone acetate or dienogest, is almost nonexistent in Norway. 24 ET products for treatment of vaginal atrophy symptoms contain lower amounts of estrogens compared with the oral formulations used for other purposes. Vaginal preparations contain estradiol (Vagifem VR ) and estriol (Ovesterin VR ). Ovesterin VR oral tablets are also used for this purpose. Duration of HT use was estimated for each different type of drug as number of total treatment days, calculated from the package size multiplied by the number of packages prescribed regarding the dosing intervals recommended. The estimated duration of HT use was extended by 4 months to account for prolonged HT use beyond the treatment days prescribed. Gaps between prescriptions were assumed to represent continuous use, unless these gaps were longer than 4 months, when they were considered to represent a stop in use 25 with eventual possible re-uptake. Women were classified as mixed users if they redeemed at least two contemporary prescriptions of different hormone types during the estimated duration of HT use. The women were included in the various type of HT preparation categories based on the specific product dispensed. All women redeeming a prescription for any HT were considered current users. Women were defined as baseline users if they redeemed their first prescription during the first 6 months of the study period (January to June 2004) and as new current users if they redeemed their first prescription after June All women in the study population contributed personyears at risk as a nonuser, current user and/or previous HT user. Person-years at risk were calculated from start of the study period, January 2004, until censoring or end of followup. Women were censored at death, emigration, breast cancer diagnosis (invasive or ductal carcinoma in situ), other tumour diagnosis or end of follow-up (December 31, 2008), whichever date came first (Fig. 2). Nonusers contributed person-years at risk from January 2004 until the date of the first redeemed prescription, if any, or end of follow-up. Women contributed person-years at risk as current users according to the accumulated duration of treatment for the type of HT dispensed. If there were gaps of more than 4 months between prescriptions, women contributed person-years at risk as a previous user from the date that the estimated duration of HT use ended, until the next redeemed prescription date if any, or end of the study period. Because it is believed that HT probably acts as a promoter of breast cancer carcinogenesis with some latency time, cases occurring during the first 3 months after start of use (current user) or after cessation of use (past user) were allocated to the former HT status category. These women contributed person-years to the former HT status from the estimated duration of use until the date of breast cancer diagnosis. Current use of HT was categorized according to user type (baseline user or new user), type of HT component (estradiol,

4 Roman et al. 587 Table 1. Characteristics of women in the study population according to hormone therapy use assessed at the end of the follow-up period ( ) Hormone therapy use Nonusers 1 (n 5 508,231) Estradiol-only users 2 (n 5 64,023) Estriol-only users 2 (n 5 14,405) Tibolone-only users 2 (n 5 9,420) Estradiol NETA only users 2 (n 5 48,144) Mixed users 2 (n 5 42,391) Breast cancer, n (&) 5,602 (11.0) 513 (8.0) 148 (10.3) 165 (17.5) 1,070 (22.2) 412 (9.7) Age at baseline, mean (SD) 60.0 (10.1) 60.0 (7.9) 68.6 (8.0) 56.9 (5.9) 57.0 (6.8) 57.8(7.8) Age at baseline, 5 years, n (%) ,859 (19.1) 5,695 (8.9) 232 (1.6) 921 (9.8) 7,067 (14.7) 5,954 (14.0) ,187 (16.4) 12,781 (20.0) 603 (4.2) 2,478 (26.3) 11,918 (24.8) 11,912 (28.1) ,315 (16.8) 16,429 (25.7) 1,351 (9.4) 3,370 (35.8) 14,283 (29.7) 10,765 (25.4) ,547 (13.7) 12,215 (19.1) 1,828 (12.7) 1,674 (17.8) 8,271 (17.2) 6,124 (14.4) ,066 (11.4) 8,066 (12.6) 2,658 (18.5) 674 (7.2) 3,869 (8.0) 3,603 (8.5) ,133 (11.2) 5,395 (8.4) 3,378 (23.5) 245 (2.6) 1,840 (3.8) 2,382 (5.6) ,124 (11.4) 3,442 (5.4) 4,355 (30.2) 58 (0.6) 896 (1.9) 1,651 (3.9) Age at 1st birth Nulliparous 68,972 (13.6) 4,939 (7.7) 1,470 (10.2) 811 (8.6) 4,656 (9.7) 3,469 (8.2) <25 275,864 (54.3) 38,967 (60.9) 7,716 (53.6) 5,654 (60.0) 30,058 (62.4) 25,779 (60.8) ,879 (22.4) 14,775 (23.1) 3,804 (26.4) 2,224 (23.6) 9,674 (20.1) 9,405 (22.2) 30 49,516 (9.7) 5,342 (8.3) 1,415 (9.8) 731 (7.8) 3,756 (7.8) 3,738 (8.8) No of births Nulliparous 68,972 (13.6) 4,939 (7.7) 1,470 (10.2) 811 (8.6) 4,656 (9.7) 3,469 (8.2) 1 64,360 (12.7) 7,543 (11.8) 1,728 (12.0) 1,249 (13.3) 6,514 (13.5) 5,627 (13.3) 2 180,413 (35.5) 26,880 (42.0) 4,624 (32.1) 4,414 (46.9) 20,922 (43.5) 18,765 (44.3) 3 124,910 (24.6) 16,986 (26.5) 3,820 (26.5) 2,283 (24.2) 11,853 (24.6) 10,411 (24.6) 4 69,576 (13.7) 7,675 (12.0) 2,763 (19.2) 663 (7.0) 4,199 (8.7) 4,119 (9.7) 1 Women not prescribed for ET or EPT during the study period. 2 Includes current and previous users.

5 588 Hormone therapy and risk of breast cancer Figure 3. Risk of invasive breast cancer associated with hormone therapy use by hormone therapy status and user type. CI, confidence interval; RR, relative risk. Adjusted for age (5-year), number of births, age at 1st birth and time (offset). estriol, tibolone, estradiol NETA and mixed), type of combined estrogen progestin regimen (continuous use, sequential use and other types of HT), route of administration (oral estradiol, transdermal estradiol, vaginal estradiol, oral estriol, vaginal estriol, oral estradiol NETA, transdermal estradiol NETA and mixed patterns) and preparation type of the various brands and strengths of oral formulations (for estradiol: Progynova VR 1 mg, Progynova VR 2 mg; for estriol: Ovesterin VR 1 mg, Ovesterin VR 2 mg; for tibolone: Livial VR 2.5 mg; for estradiol NETA: Activelle VR 1 mg daily estradiol and 14 mg NETA per month, Kliogest VR 2 mg daily estradiol and 28 mg NETA per month, Novofem VR 1 mg daily estradiol and 12 mg NETA per month, Trisekvens VR 2 mg daily estradiol and 10 mg NETA per month, Trisekvens forte VR 4 mg daily estradiol and 10 mg NETA per month; and other products and mixed patterns). Statistical analysis The outcome of interest was incident invasive breast cancer diagnosed during the study period. Incidence rate ratios (RR) with 95% confidence intervals (95% CI) of invasive breast cancer associated with different categorizations of HT use were estimated by Poisson regression. The number of incident breast cancer was analyzed as a log-linear function of exposure time (t), HT use (h), and adjusting covariates (x). The model was expressed as ln(m hx ) 5 a 1 ln(t hx ) 1 b h h 1 b x x; where a is the intercept, ln(t hx ) is the time offset of the Poisson regression model given by the log of exposure time and b is the slope of the regression line. Exposure time was measured as person-years at risk. We adjusted for age in 5-year intervals (45 49, 50 54, 55 59, 60 64, 65 69, and years), number of births (nulliparous, 1, 2, 3 and 4), and age at first live birth (nulliparous, <25, and 30). HT use was categorized as previously described (i.e., by user type, type of HT component, type of combined regimen, route of administration and preparation type). The reference group were nonusers of HT. Test for differences between levels of the study variables were done by forward difference coding for comparison of adjacent levels of categorical variables in the regression models. All tests were twosided with a 5% significance level. Statistical analyses were performed using SAS 9.2 (SAS Institute, Cary, NC). Three sensitivity analyses were performed. In the first analysis, baseline users and new users were defined as first prescription redeemed at 3 and 9 months from the beginning of the study, instead of 6 months. In the second, we restricted the analyses to women with information on education level. In the third sensitivity analysis, we excluded women below age 55 years to perform analyses on postmenopausal women only. Results We followed 686,614 women born in Norway with no previous history of cancer from 2004 to The average duration of follow-up was 4.8 years. A total of 178,383 (26%) women had redeemed a HT prescription during the study period. The number of incident breast cancer cases was 7,910. The proportion of breast cancer cases was 11.0& for nonusers, 8.0& for women who used estradiol only, 10.3& for estriol users, 17.5& for tibolone users and 22.2& for estradiol NETA users (Table 1). Estriol users were older (t-test for difference in mean age p < 0.001), whereas tibolone and estradiol NETA users were younger (p < and p < 0.001, respectively) compared with nonusers. The proportion of nulliparous women was lower among all types of HT users compared with nonusers (v 2 d test for difference in proportions, all p-values <0.001; Table 1). Compared with nonusers, there was an increased risk of breast cancer among current users of HT (RR ; 95% CI: ), whereas previous use was not associated with increased risk (RR ; 95% CI: ; Fig. 3). The risk of breast cancer was doubled in women using HT at baseline (RR ; 95% CI: ), whereas RR for new users was 1.27 (95% CI: ; Fig. 3). The RR for current users of estradiol and estriol did not reach statistically significant levels. Tibolone and estradiol NETA users had a RR of 1.91 (95% CI: ) and 2.74 (95% CI: ), respectively, compared with nonusers (Fig. 4). Previous use of estradiol, estriol, tibolone or estradiol NETA were not associated with an increased risk when compared with nonusers of HT (results not shown). Compared with nonusers, the RR of breast cancer was 1.18 (95% CI: ) for current baseline users of estriol, 1.30 (95% CI: ) for baseline users of estradiol, 2.12

6 Roman et al. 589 Figure 4. Risk of invasive breast cancer by current use of different types of hormone therapy. CI, confidence interval; RR, relative risk; NETA, norethisterone. Adjusted for age (5-year), number of births, age at 1st birth and time (offset). Others: comprises the other hormone therapy types not relevant for the specific hormone of interest. Figure 5. Risk of invasive breast cancer by current use of different oral preparations of hormone therapy. CI, confidence interval; RR, relative risk; NETA, norethisterone. Adjusted for age (5-year), number of births, age at 1st birth and time (offset). (95% CI: ) for baseline users of tibolone and 3.00 (95% CI: ) for baseline estradiol NETA users. Among new users, the RR was 0.94 (95% CI: ) for estradiol, 0.89 (95% CI: ) for estriol, 1.47 (95% CI: ) for tibolone and 2.17 (95% CI: ) for estradiol NETA users (results not shown). Current users of continuous EPT preparations of estradiol NETA had a RR of 2.80 (95% CI: ), and those using sequential EPT preparations had a RR of 2.31 (95% CI: ) compared with nonusers (Fig. 4). The RRs did not significantly differ between continuous and sequential EPT preparations (p ). Users of oral and transdermal formulations of estradiol both had a RR of 1.40 (Fig. 4). Among current users of estradiol NETA, users of oral formulations were associated with an increased risk of breast cancer (RR ; 95% CI: ), whereas users of transdermal formulations had a nonsignificant increase in risk (RR ; 95% CI: ). Compared with nonusers, current use of 1 mg estradiol and 2 mg estradiol (Progynova VR ) were associated with a RR of 1.52 (95% CI: ) and 1.68 (95% CI: ), respectively (Fig. 5). No differences in the risk of breast cancer were found between the 1 mg and 2 mg estradiol formulations (p ). Women using the highest NETA dose continuous preparation Kliogest VR (2 mg daily of estradiol and 28 mg NETA per month) showed a RR of 3.26 (95% CI: ), whereas women using the lowest NETA dose sequential preparation Trisekvens VR (2 mg daily estradiol and

7 590 Hormone therapy and risk of breast cancer 10 mg NETA per month) had a RR of 2.70 (95% CI: ). No significant differences were found between Kliogest VR and Trisekvens VR (p ). Current users of Trisekvens forte VR (4 mg daily estradiol and 10 mg NETA per month) were at higher risk of breast cancer (RR ; 95% CI: ). The risk was not statistically significant different between current users of Trisekvens VR and Trisekvens forte VR (p ). Compared with nonusers, women using the low-dose NETA preparation Activelle VR (1 mg daily of estradiol and 14 mg NETA per month) had an increased RR of breast cancer of 2.76 (95% CI: ), whereas current users of Novofem VR (1 mg daily of estradiol and 13 mg NETA per month) had a nonsignificant RR (Fig. 5). The risk was not statistically significant different between current users of the low NETA dose preparations Activelle VR and Trisekvens VR (p ), but significant differences were found between current users of Activelle VR and the high NETA dose Kliogest VR (p ). Crude incidence rates of breast cancer per 1000 personyears was 2.2 in nonusers, 4.3 in current tibolone users and 6.0 in current estradiol NETA users. This translates to an absolute risk difference of 2.1 for tibolone and 3.8 for estradiol NETA users as compared with nonusers per 1000 person-years. Assuming that the difference in risk between nonusers and tibolone and estradiol NETA users is due to HT, these results imply that use of tibolone resulted in about one extra case of breast cancer for roughly every 475 women taking tibolone for 1 year, and the use of estradiol NETA resulted in about one extra case of breast cancer for every 259 women taking estradiol NETA for 1 year. Further, applying the absolute risk difference to the hormone use in Norway from 2004 to 2008 (number of person years: 30,699 for tibolone and 144,792 for estradiol NETA), the use of tibolone and estradiol NETA is estimated to have resulted in 64 and 559 additional cases of invasive breast cancer, over the average follow-up of 4.8 years. These numbers correspond to 7.9% of all invasive breast cancer cases diagnosed among women aged years during the study period (0.8% due to tibolone use and 7.1% due to estradiol NETA use). No significant differences were found when baseline users were defined at 3 and 9 months from the beginning of the study as compared with the 6 months definition used. Nor did the results change when we adjusted for education in the restricted analyses in women with known education level (n 5 665,909; 96.9% of women). The results excluding women below age 55 years showed a moderate increased risk of breast cancer for users of EPT (RR ; 95% CI: ) and tibolone (RR ; 95% CI: ) compared with EPT and tibolone users in the full database analyses. Discussion In this study including information about prescriptions from 2004, we found that users of estradiol NETA had the highest risk of breast cancer compared with nonusers, of which Kliogest VR, Activelle VR and Trisekvens VR users had about threefold elevated risk. Furthermore, users of tibolone had a nearly doubled risk. We found moderate elevated risks of breast cancer in users of oral and transdermal ET preparations of estradiol. Our results are in agreement with previous studies that found that EPT use is associated with an increased risk of breast cancer. 7,10,13,19,26,27 The risk estimates associated with EPT use are similar to those found in prospective European studies, 10,15,19 but higher than in U.S. observational studies 6,13 and in the Women s Health Initiative randomized controlled trial. 7,26 Preparations containing estradiol NETA are rarely used in the United States, but are common in Norway and other Nordic countries, 20 whereas medroxyprogesterone is the most common form of EPT in the United States. However, previous studies suggest that neither the type of estrogen nor the type of progestin influence the risk of breast cancer. 10,21,28 Thus, differences between the European and U.S. studies may be explained by differences in the monthly doses of progestin. In our cohort, 62% of HT users were current users at baseline. Baseline HT users had a higher risk of invasive breast cancer than women who redeemed their first prescription 6 months after start of follow-up (new users). Baseline users are likely to have been HT users prior to the study start in January The increased risk in baseline users supports the previous findings that the risk increases with duration of HT use. 1,5,10,13,20 Among new users of estradiol NETA, who were possibly more likely to have used these products exclusively, the RR was 2.17 (95% CI: ). Sensitivity analyses on the definition of baseline users at 3 and 9 months from the beginning of the study did not show statistically significant differences when compared with the 6 months definition used. Tibolone use was associated with a nearly twofold higher breast cancer risk than in nonusers. Tibolone is a tissuespecific regulator with estrogenic and androgenic properties. Use of tibolone reduces the level of active estrogen in breast tissue, whereas it has limited effect on cell proliferation. 29 Hence, tibolone has been promoted as less risky for the breast than therapies with natural estrogens. 30,31 However, our study supports the results from previous studies that found an increased risk of breast cancer among tibolone users, 10,15,21 although one case control study found no association. 32 Tibolone was introduced in Norway in 2000, around the same time as several major studies emerged documenting an increased risk of EPT on breast cancer risk. 4 6 Thus, we cannot exclude the possibility that previous EPT users had, by 2004, started to use tibolone instead, and that our risk estimates partly reflect previous EPT use. Nevertheless, we found that new users of tibolone had an increased risk of breast cancer, and the risk did not differ significantly in magnitude compared with baseline users of tibolone (data not shown). Overall, use of ET preparations was not associated with breast cancer risk. This is in agreement with evidence from some studies, 2 4,11,14,33 35 but not with others that found a moderate increase in breast cancer risk with ET

8 Roman et al. 591 preparations. 5,10,13,21,36 39 We found a moderate elevated risk in users of oral and transdermal estradiol, which is in agreement with results from Lyytinen et al. 40 in Finland that found a similar increased risk of breast cancer for oral and transdermal users of estradiol. Use of vaginal estradiol is relatively common compared with oral and transdermal use. In our study, 67% of the contribution in person-years at risk of breast cancer due to estradiol use was related to vaginal use. Vaginal preparations of ET, which are used for the treatment of symptoms of vaginal atrophy, have lower amounts of estrogens than oral preparations of the same compounds. In our study, oral or vaginal use of estriol, usually prescribed for treatment of vaginal atrophy in elderly women, was not associated with increased risk of invasive breast cancer, which corroborates the findings in the study by Lyytinen et al. 40 Oral use of EPT with estradiol NETA was associated with an increased risk of breast cancer, whereas transdermal use was not associated with any increased risk. However, the risk estimate for transdermal use was based on a small sample of exposed women (1.6%), which limits our statistical power and encompasses a wide confidence interval, making the risk estimates inconclusive. A study by Lyytinen et al. 41 reported that oral and transdermal use of estradiol NETA preparations of EPT were associated with a comparable increased risk of breast cancer. A study from France reporting only oral preparations of estradiol NETA found also an increased risk. 39 Our findings of an increased risk of breast cancer for both continuous and sequential combined preparations of EPT are in agreement with previous studies. 3,10,13,15,19 21 Astudyshowed an increased risk of breast cancer for continuous compared with sequential preparations of estradiol NETA. 27 We found a higher RR in continuous than in sequential preparations, but the differences were not significant. Because use of EPT with progestin types other than NETA was almost nonexistent in Norway after 2004, 24 we could not compare effect estimates of various progestin types. An increased risk of breast cancer with use of EPT with progestin types of NETA is in line with previous European studies. 10,20 In our study, women who redeemed oral formulations of Kliogest VR, Activelle VR, Trisekvens VR and Trisekvens forte VR had an increased risk of invasive breast cancer, ranging from a RR of 2.70 for the lowest monthly progestin dose Trisekvens VR preparations with 2 mg daily estradiol and 10 mg NETA per month to 3.26 for the highest monthly progestin dose Kliogest VR with 2 mg daily estradiol and 28 mg NETA per month. The high-dose preparation Kliogest VR has twice the amount of estrogens and progestin per month than Activelle VR and showed a significantly increased risk of breast cancer when both products were compared. However, no differences in the risk of breast cancer were found between the high NETA dose Kliogest VR and the lower NETA dose Trisekvens VR. A previous study carried out in Finland found no significant differences between the low-dose Activelle VR and high-dose Kliogest VR regardless of the duration of exposure. 41 We found no significant differences in magnitude in the risk of breast cancer between users of the continuous preparation Activelle VR and the sequential preparation Trisekvens VR, which have a close monthly cumulative dosage of norethisterone. The estradiol dose in the oral formulations of estradiol NETA varied between 1.0 and 4.0 mg, but an increasing daily dose of estrogen was not associated with an increased risk of breast cancer in our study. Preparations of Trisekvens VR and Trisekvens forte VR with estrogen amounts of 2 and 4 mg daily estradiol, respectively, but same dose of 10 mg NETA per month, had a similar increased breast cancer risk when tested for differences. An absence or a minimal evidence of the impact of estrogen doses of this magnitude in the risk of breast cancer have been previously reported. 6,10,40,42 Our statistical power to test some of the differences between preparations was limited and should be considered carefully. The high-dose preparation Kliogest VR and the low progestin dose Trisekvens forte VR were withdrawn from the market in Norway in 2008 and 2006, respectively, and Activelle VR is the most used estradiol NETA preparation. Overall, the increased risk of breast cancer associated with estradiol NETA preparations, including the lower dose preparations, is substantially higher than expected despite the short follow-up time in our study. Some women may have been prescribed ET alone and progestin alone equivalently to a combined EPT prescription. These types of combined regimens were commonly used in the United States. 13 However, only 1% of women in our study population had combined prescriptions of progestin alone plus ET or EPT. Because this is not a typical prescription in Norway, and because of concern that this may or may not have been used concurrently, we excluded women prescribed progestin alone (ATC group G03D) from the analysis (see Fig. 1). Strengths and limitations Our population-based study included nearly 700,000 women aged years. A major strength is the registry-based information on exposure to hormones throughout the study period, rather than information based on self-reported HT use. Another strength is the linkage to other populationbased registries, including the Cancer Registry of Norway, which has a high coverage and a completeness higher than 98%. 23 Our study has several limitations. On one hand, we excluded some women based on what happened during follow-up (women who only redeemed one prescription dated after June 2004 (3.2%), women prescribed sex hormones other than ET, tibolone or EPT during the study period (3.7%)), which may lead to some time-related biases. 43 However, the bias introduced by these exclusions is expected to be small. On the other hand, we do not know whether women had used HT before study start. Nevertheless, previous use does not seem to influence the risk of breast cancer, which minimizes the possible bias. 1,10,44,45 Further, we did not have information on menopausal status. Premenopausal

9 592 Hormone therapy and risk of breast cancer women are at a higher risk of breast cancer compared with menopausal women of the same age, and HT users are more likely to be postmenopausal, which would underestimate the risk among HT users. Nonetheless, the sensitivity analyses excluding women below age 55 years showed a moderate increased risk of breast cancer for HT users 55 years or older (EPT and tibolone) compared with HT users in the full database analyses, which evidences a possible underestimation of the effect of HT use in our results by the lack of information on menopausal status. Moreover, we had no information on use of opportunistic or organized mammographic screening. Women prescribed HT are likely to be more closely followed by their general practitioner or gynaecologist and be recommended to have regular screening mammograms. Also, we lack information on hysterectomy. Recommendation for ET use is usually limited to women with a previous hysterectomy, whereas EPT is recommended in women with an intact uterus. Furthermore, low-dose vaginal formulations of estriol are available without prescription in Norway. Users of vaginal estriol would be under-registered and included in the reference population of nonusers, which could lead to underestimation of the risk conferred with their use. Thus, the lack of association between use of vaginal estriol and breast cancer risk in our study should be considered with caution. Also, the maximum follow-up time of our study was 5 years, which may be considered relatively short. However, the magnitude of the risk estimates we found was greater than that initially expected based on other studies. REFERENCES Absolute risk with HT Use of tibolone and estradiol NETA may have resulted in roughly 623 extra cases of breast cancer in Norway over the mean follow-up of 4.8 years, i.e., 7.9% of the invasive breast cancers in this study from 2004 through 2009, or 130 new cases per year among Norwegian women aged years. Bakken et al. did a similar estimate based on data from 10 years earlier, i.e., self-reported HT use during At that time, they estimated that approximately 300 new cases of breast cancer per year were due to current use of HT among Norwegian women aged years, 19 substantially higher than our estimate. However, Bakken et al. s assessment was prior to the publication of the Women s Health Initiative trial results in 2002, which changed medical practice also in Norway by decreasing use of EPT preparations markedly. 7 Conclusion In this registry-based study of HT use in Norwegian women aged years, the highest risk of breast cancer was found among users of EPT with estradiol NETA, associated with about threefold elevated risk compared with nonusers. Furthermore, tibolone conferred a nearly doubled risk, and also oral and transdermal ET preparations of estradiol seemed to confer an elevated risk. The increased incidence approximates one extra invasive breast cancer for every 259 women using estradiol NETA for 1 year and one for every 475 women using tibolone. 1. Collaborative Group on Hormonal Factors in Breast Cancer. Breast cancer and hormone replacement therapy: collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Lancet 1997; 350: Persson I, Weiderpass E, Bergkvist L, et al. Risks of breast and endometrial cancer after estrogen and estrogen-progestin replacement. Cancer Causes Control 1999; 10: Magnusson C, Baron JA, Correia N, et al. Breastcancer risk following long-term oestrogen- and oestrogen-progestin-replacement therapy. Int J Cancer 1999; 81: Ross RK, Paganini-Hill A, Wan PC, et al. Effect of hormone replacement therapy on breast cancer risk: estrogen versus estrogen plus progestin. J Natl Cancer Inst 2000; 92: Schairer C, Lubin J, Troisi R, et al. Menopausal estrogen and estrogen-progestin replacement therapy and breast cancer risk. JAMA 2000; 283: Colditz GA, Rosner B. Cumulative risk of breast cancer to age 70 years according to risk factor status: data from the Nurses Health Study. Am J Epidemiol 2000; 152: Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women s Health Initiative randomized controlled trial. JAMA 2002; 288: Beral V, Banks E, Reeves G. Evidence from randomised trials on the long-term effects of hormone replacement therapy. Lancet 2002; 360: Chlebowski RT, Hendrix SL, Langer RD, et al. Influence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: The Women s Health Initiative randomized trial. JAMA 2003; 289: Beral V. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet 2003; 362: Olsson HL, Ingvar C, Bladstrom A. Hormone replacement therapy containing progestins and given continuously increases breast carcinoma risk in Sweden. Cancer 2003; 97: Colditz GA. Estrogen, estrogen plus progestin therapy, and risk of breast cancer. Clin Cancer Res 2005; 11:909s 17s. 13. Saxena T, Lee E, Henderson KD, et al. Menopausal hormone therapy and subsequent risk of specific invasive breast cancer subtypes in the California Teachers Study. Cancer Epidemiol Biomarkers Prev 2010; 19: Anderson GL, Limacher M, Assaf AR, et al. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: The Women s Health Initiative randomized controlled trial. JAMA 2004; 291: Stahlberg C, Pedersen AT, Lynge E, et al. Increased risk of breast cancer following different regimens of hormone replacement therapy frequently used in Europe. Int J Cancer 2004; 109: Formoso G, Perrone E, Maltoni S, et al. Short and long term effects of tibolone in postmenopausal women. Cochrane Database Syst Rev 2012; 2:CD Grosse Y, Baan R, Straif K, et al. A review of human carcinogens-part a: Pharmaceuticals. Lancet Oncol 2009; 10: Pharmaceuticals. Volume 100 A. A review of human carcinogens. IARC Monogr Eval Carcinog Risks Hum 2012;100: Bakken K, Alsaker E, Eggen AE, et al. Hormone replacement therapy and incidence of hormonedependent cancers in the Norwegian Women and Cancer Study. Int J Cancer 2004; 112: Lee SA, Ross RK, Pike MC. An overview of menopausal oestrogen-progestin hormone therapy and breast cancer risk. Br J Cancer 2005; 92: Bakken K, Fournier A, Lund E, et al. Menopausal hormone therapy and breast cancer risk: Impact of different treatments. The European Prospective Investigation into Cancer and Nutrition. Int J Cancer 2011; 128: Cancer Registry of Norway. Regulation on the collection and processing of personal health data in the Cancer Registry of Norway (cancer registry regulations). Oslo, Norway: Cancer Registry of Norway, Larsen IK, Smastuen M, Johannesen TB, Langmark F, Parkin DM, Bray F, Moller B. Data quality at the Cancer Registry of Norway: an overview of comparability, completeness, validity and timeliness. Eur J Cancer 2009; 45: Hofvind S, Sakshaug S, Ursin G, et al. Breast cancer incidence trends in Norway explained by

10 Roman et al. 593 hormone therapy or mammographic screening? Int J Cancer 2012; 130: Nielsen LH, Lokkegaard E, Andreasen AH, et al. Using prescription registries to define continuous drug use: how to fill gaps between prescriptions. Pharmacoepidemiol Drug Saf 2008; 17: Chlebowski RT, Anderson GL, Gass M, et al. Estrogen plus progestin and breast cancer incidence and mortality in postmenopausal women. JAMA 2010; 304: Lyytinen H, Pukkala E, Ylikorkala O. Breast cancer risk in postmenopausal women using estradiol-progestogen therapy. Obstet Gynecol 2009; 113: Flesch-Janys D, Slanger T, Mutschelknauss E, et al. Risk of different histological types of postmenopausal breast cancer by type and regimen of menopausal hormone therapy. Int J Cancer 2008; 123: Notelovitz M. Postmenopausal tibolone therapy: biologic principles and applied clinical practice. MedGenMed 2007; 9: Archer DF, Hendrix S, Gallagher JC, et al. Endometrial effects of tibolone. J Clin Endocrinol Metab 2007; 92: Cummings SR, Ettinger B, Delmas PD, et al. The effects of tibolone in older postmenopausal women. N Engl J Med 2008; 359: Opatrny L, Dell Aniello S, Assouline S, et al. Hormone replacement therapy use and variations in the risk of breast cancer. BJOG 2008; 115:169 75; discussion Stanford JL, Weiss NS, Voigt LF, et al. Combined estrogen and progestin hormone replacement therapy in relation to risk of breast cancer in middle-aged women. JAMA 1995; 274: Manson JE, Chlebowski RT, Stefanick ML, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women s Health Initiative randomized trials. JAMA 2013; 310: Fournier A, Berrino F, Riboli E, et al. Breast cancer risk in relation to different types of hormone replacement therapy in the E3N-EPIC cohort. Int J Cancer 2005; 114: Colditz GA, Hankinson SE, Hunter DJ, et al. The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. N Engl JMed1995; 332: Chen CL, Weiss NS, Newcomb P, et al. Hormone replacement therapy in relation to breast cancer. JAMA 2002; 287: Li R, Gilliland FD, Baumgartner K, et al. Hormone replacement therapy and breast carcinoma risk in Hispanic and non-hispanic women. Cancer 2002; 95: Fournier A, Berrino F, Clavel-Chapelon F. Unequal risks for breast cancer associated with different hormone replacement therapies: results from the E3N cohort study. Breast Cancer Res Treat 2008; 107: Lyytinen H, Pukkala E, Ylikorkala O. Breast cancer risk in postmenopausal women using estrogen-only therapy. Obstet Gynecol 2006; 108: Lyytinen H, Dyba T, Pukkala E, et al. Do the dose or route of administration of norethisterone acetate as a part of hormone therapy play a role in risk of breast cancer: National-wide case-control study from Finland. Int J Cancer 2010; 127: Chen WY, Manson JE, Hankinson SE, et al. Unopposed estrogen therapy and the risk of invasive breast cancer. Arch Intern Med 2006; 166: Suissa S, Azoulay L. Metformin and the risk of cancer: time-related biases in observational studies. Diabetes Care 2012; 35: Chlebowski RT, Kuller LH, Prentice RL, et al. Breast cancer after use of estrogen plus progestin in postmenopausal women. N Engl J Med 2009; 360: Fournier A, Mesrine S, Dossus L, et al. Risk of breast cancer after stopping menopausal hormone therapy in the E3N cohort. Breast Cancer Res Treat 2014; 145: