One effect and two causes: Growth acceleration and breast cancer risk after hormone replacement therapy

Dieter Hölzel, Anne Schlesinger-Raab, Kathrin Halfter
doi: https://doi.org/10.1101/2024.02.16.24302914

Abstract

Background An increased risk for breast cancer (BC) following hormone replacement therapy (HT) with estrogen and progesterone in women has been reported in several studies. However, HTs are associated with two distinct effects, a BC risk (RF) and an acceleration of BC growth (GAF). The interaction of both effects is analyzed.

Methods Using data from the U.S. population on BC incidence and life expectancy specific cohorts and their disease trajectories are modelled. First, age-specific BCs are randomly generated for the age interval 50-80 years based on public data. Second, this simulated cohort subsequently receives a HT over 4 years that accelerates the growth of prevalent BCs. In a third cohort additional BCs are simulated caused by HT under treatment. The cumulative incidence of BCs is modelled for up to 30 years using different assumptions on BC growth duration, GAF and RF, as well as the duration of HT. The Women’s Health Initiative Study (WHI-S) is also simulated in a fourth cohort assuming a GAF 1.4 and RF 2.0.

Results Studies modelling the risk of BC after HT imply two main findings: First, the growth of prevalent BCs is accelerated in parallel to the start of HT. This results in an increased BC incidence where the relative risk is equivalent to the GAF. The duration of HT defines the turning point of this increase. The second finding demonstrates that during the 15 years of BC growth an inherent RF for BC through HTs becomes observable only after a comparable delay. The combined effect results in an overlap of age-specific BCs, which develop at the same time, both growing faster under HT.

Varying parameters can explain different results in the WHI-S. According to the WHI almost 6 million women decided to discontinue or not to start HT. This offers a valid explanation (assuming a GAF 2) for the 10% decline in incidence around 2002. Estimates which report one million additional BCs associated with HT in Western countries since 1990 can thus, for the most part, be attributed to more rapidly growing prevalent BCs.

Conclusion HT accelerates the growth of prevalent BCs and can also cause BCs which become symptomatic after 10 or more years. In combination these effects increase BC incidence. Differentiated information on these inherent risks and benefits should support shared decision-making for HT.

Introduction

Hormone replacement therapies (HT) have been used in different forms for more than 70 years, at the same time the discussion on their risks and benefits has not abated. 1,2 Reasons for the continued controversy surrounding HT include side effects such as increased risks of breast cancer (BC), endometrial cancer, dementia, and stroke, while also reducing the risk of diabetes, fractures or colorectal cancer. 3 The medical indication of HT for menopausal symptoms remains relevant to women’s health care. Data on the increased risk of breast cancer (BC) due to combined HTs with estrogen and progesterone were presented in 2002 by the Women’s Health Initiative Studies (WHI-S). The publication and subsequent change in HT prescription patterns resulted in a marked decrease in BC incidence in the USA. 4,5 Further data on HT and BC risk are available from the Million Women Study (MWS) 6 and from meta-analyses (Meta-A). 7,8 Meta-A estimated that 1 million BCs were caused by HT in the first 5 years after onset in Western countries since 1990, but with the following condition that: “…if these associations are largely causal…”. We hypothesize that HTs do result in additional BCs caused by a combination of two effects, an acceleration of growth of prevalent BCs and increased BC risk. In this modelling study we examine the interaction of these two effects and estimate the risk for HT users.

Methodology

Four cohorts of female patients are randomly generated in order to model the effects of HT according to normal life expectancy and the incidence from SEER Cancer Statistics Review, 1975-2002 9 :

  1. BCs are randomly generated within a virtual cohort of 100,000 50 year old women over an observation period of 30 years taking life expectancy into account. 9,10 This cohort serves as the control group.

  2. The control cohort receives HT. The HT effect is estimated using the duration of BC growth (GD), the duration of HT (HTD), and a growth acceleration factor (GAF).

  3. A risk factor (RF) takes effect, which becomes relevant in the initiation of additional BCs through HT. Both effects, the faster growth in the second cohort and an increased BC risk are combined in this third cohort, starting and ending with HT.

  4. Endocrine prevention is modelled in a separate cohort for additional comparison.

For hormone receptor positive (HR+) BCs, initial modelling is performed with the parameters GD of 15 years, HTD of 4 years and the factor 2 as GAF and RF. This RF of 2 is taken from the Gail model, which was used to recruit participants for chemoprevention studies. 11 Briefly, these studies show that even 15 years after the end of prevention therapy, the incidence was cut in half and thus concluded that HR+ BCs grow for at least 15 years 12,13

The growth of BCs can be described using volume doubling times (VDT). VDTs of 72 and 170 days are used for HR- and HR+ BCs, respectively. These values are the mean VDT from two studies. 14,15 This means, for example, that 10.5 volume doublings (VD) or 2.1/4.9 years are required for the growth of HR-/HR+ BC from 2.5mm (pT1a) to 28.0mm (pT2). After 33 VD a 20.0mm BC with about 8,6 billion cells has developed from one cancerous cell, which takes 15.4 years in HR+ BC. This time span is comparable to the GD introduced above. A GAF of 2 can also be considered biologically plausible, as shown by the 2.4-fold faster growth of HR- versus HR+ BCs. These values given above are then used to generate growth trajectories and adjusted BC diagnosis timelines (see Fig. 1).

Figure 1:
Figure 1:
Figure 1:

Tumor growth and growth acceleration by HT with GD 15 yrs, HTD 4 yrs and factor 2 for GAF and RF.

A: HT begins at the age of 50 at time 0. The straight lines represent the number of cells with exponential growth, the right axis shows the “age of development” of the prevalent BCs. If a BC is initiated 15 yrs earlier, it is excluded due to positive mammography (red dot, a prevalent 1 mm BC which has grown over 9.3 yrs from initiation and further 6 yrs until detection by mammography). BCs which start between - 15 and -11 yrs grow a maximum of 2 yrs faster under HT (orange trajectory) and are detected during HT. BCs which are initiated before the start of HT (green) grow faster over 4 yrs at different tumor ages, their growth slows again after the end of HT. Their detection is brought forward in time. The age-specific BCs developing under HT (red trajectories) and additional BCs (blue) developing if there is a BC risk by HT grow faster under HT. The first BCs initiated by HT would appear after about 13 yrs, the last after 19 yrs.

B: Transformation of the time of diagnosis of BC by HT. HT promotes tumor growth. Four periods can be distinguished: P1 – Incidence increase with a RR of HTD/GAF. P2 – Same incidence until GD- HTD*((GAF-1)/GAF) (2-13 years). P3a – Incidence decreases because diagnosis was brought forward due to accelerated growth (green). P3b – HT induces new BCs (blue). P4 – End of HT effect. The transformations and relative risks can be easily calculated if an annual incidence of e.g. 100 is assumed.

Comparison and modelling of the WHI-S uses published data such as age distribution and hazard ratios of cumulative incidences to estimate GAF and RF. Statistical analyses and modelling using data matrices of 3 million elements for e.g. 100,000 test persons and their events over a period of 30 years were done using the R-System version 3.1.3 16.

Results

Simulation of BC incidence with plausible model parameters for BC risk and growth acceleration due to HT

The time-dependent GAF effect of HT on disease trajectories is described in Fig. 1. The interaction with RF and the resulting BC incidence estimate is shown in Fig. 2. In the example with 100,000 virtual test persons 119+263+281+248 BCs were generated in the first 4 years of the control group (SD ± 11-17). For the next 30 years, there are 9,961 BCs and 27,612 deaths without BC to expect. Two years of endocrine prevention would eradicate an estimated 2,731 or 27.3% BCs, assuming 50% treatment efficacy.

Figure 2:
Figure 2:

Simulation of incidence for 100,000 women aged 50 years with the model parameters GD 15 years, HTD 4 years, the factor 2 for GAF and RF and life expectancy. Overlapping lines were shifted manually.

A: Cumulative incidences generated according to the SEER data from 2000 (red) with 9,961 BCs or 11% over 30 years, with faster growth through 4 years of HT (green) with the same number and with an additional BC risk (n=911) through HT (blue). With 2 years of endocrine prevention and 50% BC reduction, 2,731 or 27.4% of prevalent BCs could be avoided in 17 years (black). The 4 periods of Fig. 1 are marked.

B: The first 7 years of the cumulative incidences in Fig. 2A. Due to the inclusion criterion “negative mammography”, almost no BCs are diagnosed in the first six months without HT, but the first faster-growing BCs are diagnosed after 3 months with HT.

C: Contours of the bar charts of the annual BCs: Without HT (red), with HT and GAF 2 (green), with HT, GAF 2 and RF 2 (blue). The beginning and end of HT results in 4 periods: At the beginning by growth acceleration of all prevalent BCs and with a time delay for the BCs occurring age-specifically under HT and those initiated by HT, both of which are accelerated for 4 years and occur in P3 over 6 years. The annual fluctuations are due to random generation (Poisson distributed: SD 19). The mammography effect is evident in the first year. The absolute number of annual BCs decreases with age despite increasing incidence because 27,612 women do not reach the age of 80 due to competing risk.

D: Time dependence of the relative risks with and without additional BC risk of HT and 95% CI. The risks differ from the 13th year onwards. The inclusion criterion negative mammography and GAF 2 result in the increased initial risk.

The cumulative incidences show 3 inflection points (Fig.2): Period 1 ends after HTD/GAF years and gives the best estimate for the relative risk (RR), which is equivalent to GAF. The 911 BCs diagnosed in the control group within the first 4 years become detectable a maximum 2 years earlier through HT. The incidence is thereby effectively doubled during this period (cohort 2). The RR decreases during Period 2. At the beginning of Period 3 (GD-HTD*((GAF-1)/GAF) the incidence in the HT group decreases up to Period 4 (GD+HTD). In the risk group (cohort 3), the incidence increases in Period 3 until the end of the BC risk enhancing HT. Both prevalent and new BC types are exposed to HT between 0 and 4 years. When the GAF and RF factors are equal, the cumulative incidence shows no effect. In Period 4, after 19 years, the HT effect on BC risk returns to the “naturally” expected value. Fig.2C-D show the inflection point’s dependence of the RR in a year-on-year comparison.

Women Health Initiative Studies

The observational WHI-S reported a RR of 2 for BC in the first years. The same risk was modelled with GAF 2 in Fig. 2. 17 The WHI-S with 16,121 women showed a cumulative incidence congruent to the simulation illustrated in Fig. 2 of this article. 18 In a separate randomized study the WHI presented different results. Herein 8,506 women were recruited for combination HT, their age distribution, study cancellation after 5.6 years, life expectancy and the inclusion criterion of a negative mammogram are available data from the study. The cumulative hazard rates after 5.6/24 years were 1.26/1.28. 4,5,19. Using this data from the randomized study the two factors can be estimated and under the specification of GAF 1.4 and RF 2.0 a study replicate model was generated (Fig.3).

Figure 3:
Figure 3:
Figure 3:

Simulation of the randomized WHI study

The model parameters are: 8,506 women (n=659/659/811/418 BCs in the four cohorts, see Fig.2), age distribution according to the WHI-S (MW: 63.6 yrs), GD 15 yrs (3,313 total yrs of tumor growth), HTD 5.6 yrs (accelerated tumors n=379, total years of earlier detection =-682 yrs), negative mammogram, incidence according to SEER data from 2000, life expectancy considered as competing risk. They result in the estimates GAF 1.4 and RF 2.0 (+152 additional BCs). For a better overview overlapping lines were shifted manually.

A: Cumulative incidences of the four groups. Control group of the WHI-S (red), and the HT groups without (green) and with BC risks (blue) show the combined effect for 30 years of follow-up. The turning points are at 4, 13.4 and 20.6 years.

B: Cumulative incidences of the first 7 years. The effect of the negative mammogram is illustrated, whereby no BCs occur in the control group in the first 6 months. In the control group 202 BCs were generated in 7 years, in the first year only 13 from day 186, in the HT group 249 and 13 appear before day 186 due to GAF 1.4.

C: Contours of the bar charts of the annual BCs. Due to the higher age of 63.6 years, the number of patients at risk and thus the annual BCs decrease significantly compared to Fig.2C.

D: Time dependence of the relative risks with and without additional BC risk by HT and 95% CI (see Fig. 2)

The age distribution of the study model resulted in a mean of 63.5 years (WHI-S 63.4 years). As in the most recent publication the median age after 24 years of follow-up was 82.0 years, and 3,182 of the 8,506 patients in the HT group were still at risk.

After 8 years, 296/262 (RR 1.22 CI: 1.02-1.45) and after 24 years 761/609 (RR 1.27 CI: 1.14-1.41) BCs were generated in the two groups with/without HT. This can be compared to data of the WHI-S which are 277/207 after 8 years and 584/447 after 24 years. The most striking difference between the original and the modelled study was found in the first 4 years (Fig.3B), in which the randomized WHI-S even shows a protective effect of HT. 5 In the first year 13/27 BCs were generated (Fig.3C) in the modelled study, which also shows the time dependence of the RR with large confidence intervals due to small case numbers (Fig.3D). Thirty years of complete follow-up also explain the higher number of cases. Overall, the WHI-S can be reproduced with the two factors GAF 1.4 and RF 2 except for the first few years.

The Million women study and Meta analyses

The results of the updated Meta-A summarize data from 58 prospective and retrospective studies with a total of 143,887 BCs and 424,972 controls. 7,20 Data on BC risk by histological type in the MWS was presented in 2006 and showed an RR of 1.71 for ductal BC with a treatment duration of 3.4 years for combination HT. 6 Fig.3C-D also shows large RRs at the beginning due to the high risk after a negative mammogram. The Meta-A results show an increase in RR, which reaches 2.26/2.51 in women taking HT with treatment durations of 10-14/≥ 15 years. However, there is no continuous increase in risk in any simulation with longer HT. It is therefore remarkable that there appears to be a decrease in the RR to 1.2-1.3 for women who had undergone HT in the first 5 years after the end of long HTs. The findings correspond to those seen Fig.3D. The changes in RR are not available over the course of the years, not even for the excellent MWS with 43,022 BCs and 169,041 controls. Only the increased BC mortality after 20 years is available 21, but no details on the prognostic factors of the BCs that occurred during the course. This data are important to better understand the circumstances of the estimated one million additional BCs caused by HT worldwide. 7 But detailed data on various drugs, the onset and duration of HT and long-term follow-up are needed in order to provide reliable estimates for GAF and RF in simulations.

Discussion

The effect of HT with estrogen and progesterone is modelled based on currently known facts on tumor growth. The median time from tumor initiation to METs in HR-/HR+ BCs is 5/10 years. 22 Since BCs grow more slowly, the 1.5-fold growth time for BC is plausible. A BC growth over 15 years is also shown in prevention studies for HR+ BCs. 12 Therefore, there are 2 effects caused by HT: A growth acceleration and a BC risk, represented by GAF and RF. Available data can be simulated, modelled, and reinterpreted considering these two factors.

Growth acceleration GAF

An accelerated growth causes the increase in incidence at the beginning of HT (Fig. 1). Other published findings on HT and breast biology support these findings, for example HTs can increase breast density and cell activity in the short term and BCs diagnosed during HT are more likely luminal A-like 23. In addition, elevated Ki67 levels indicate increased proliferation even in healthy breast tissue due to HT. 24 Furthermore, there is no evidence that dormant BCs exist and can be reactivated. Therefore, the estimated one million BCs are likely to be prevalent, not yet detectable, only faster-growing BCs. This can also be corroborated by the MWS with a reported RR 1.45 after one year of HT use. 7,20 As the modelling in Fig.2 shows, a constant RR is to be expected from the beginning of HT, which is equivalent to the GAF. This phase lasts for the time limit of HTD/GAF.

A closer look at the initial time period can also be revealing: If negative mammograms in the first few months after the onset of a study suggest lower incidences, a GAF should result in an extremely high RR in the first few months, even with small numbers of patients in the studies (Fig. 2-3). The time dependence can be described overall by reporting the annual BCs, the RR, or at best continuously with the cumulative incidence in the study and control arms of a study. The observational WHI study shows this initial effect with an RR of 2 (see also Fig. 2).

If HT causes faster growth of HR+ BCs, this would mean that in HT users larger BCs and BCs with an increased number of positive LNs are diagnosed. During screening more interval cancers were reported and the BCs spread earlier, which would increase mortality. 19,20 However, the increase in mortality in the WHI-S is not significant even after 20 years. This finding is due to the number of cases, but also to a follow-up period which is too short because initiated BCs mostly occur after 15 years and METs occur a further 15 to 20 years later. 18 An average tumor diameter of 17/15.0mm is reported for BCs with/without HT 19, which results in a mortality rate of 84.6/87.4% after 15 years. 25

Wash-out phase

As HT users were also recruited for the studies, the wash-out time must be taken into account.19 Together with other inclusion and exclusion criteria, this improves the homogeneity of the study groups. However, a GAF and RF effect of HT is certainly not reversible. That is why there is no safe wash-out interval. 26 An interruption of HT results in intermittent growth, after restarting HT the accelerated growth and the associated BC risk returns. As for the BC risk, its relationship to the duration of HT has to be considered as well. If HT is used for the first time at the onset of the study, age-specific BCs are bound to occur at the same time as the BCs caused by HT. The growth of both BCs is accelerated by HT and they occur in the Period 3 interval. This reasoning appears plausible because neither prevention studies nor adjuvant therapies reveal an efficacy window: They affect all BCs, large and small prevalent BCs. Since BCs are not eliminated by the wash-out time, the durations before and after the start of the study must be added together for past users. With HT durations of more than 10 years and a GAF of 2 the first HT-induced BCs will occur during HT; the two effects overlap. This explains the increasing RR in the Meta-A study for HT users of 10-14 and >15 years. So far, there is no biological explanation, no memory for the duration of HT and a RF that increases with HTD.

The decline in incidence

With the end of HT its effects also stop. This is shown by the marked drop in BC incidence in the U.S. reported for 2002 following the publication of the WHI-S. 9,27 In 2002 203,500 BCs were expected in the U.S., in 2008 only 182,460 28,29. In the age interval between 50 and 70 years, the incidence was 356 per 100 000. A decrease in the incidence of 21,040 (10.3%) BCs would be explained by 5.9 million women foregoing HT under the assumption of a GAF of 2. This decline is explained by the simultaneous renunciation of HTs and cancellation of ongoing HTs. Approximately 20 million fewer prescriptions were made in 2003 compared to 2002, or a decrease of approximately 5 million HTs from 2001 to 2003 reported for the whole U.S. population 30 ,31 A further explanation, such as a regression of prevalent BCs through withdrawal of millions of HTs, is not required for this rapid decline. In addition, this decline confirms that switching growth acceleration on and off responds quickly and that BC risk at all does not play a role in this short-term decline. As the two effects GAF and RF cannot be reversed by stopping HT, the stop still has an effect corresponding to the GD for 15 years. Until 2018, the incidence shows no clear trend, which speaks in favor of compensation and thus equal effects of GAF and RF (Fig. 2C). 27

The HT effect can also explain the long-term incidence trend in the USA. The temporary increase due to screening in the 1980s continued with the increasing utilization of HT until 2001. In 2006, the incidence was the same as in 1986. 9 Following the publication of the missing prevention of heart disease and side effects of HT study 32, incidence and HT utilization began to decline. This was historically unique reaction in a population due to the 2002 WHI-S publication. This suggests that the incidence peak between 1986 and 2002 was an artefact merely bringing prevalent BCs forward through screening and HT.

Open questions

HTs cause BC growth acceleration and have an additional BC risk. Even with short HTs, the effects overlap in the range of the growth duration of BCs at around 15 years. The effects such as those on cell activity and breast density change in the short term. Therefore, the time dependence of the effects must be analyzed continuously with up to 30 years of follow-up. Even if this allows reliable conclusions to be drawn for patients, many questions regarding the quantification of the effects remain unanswered.

How does the growth of BCs depend on the estrogen and progesterone receptors and molecular subtype? Does it depend on the duration of HT or on age? Can a GAF in HR+ BCs even approach the 2.5-fold growth of triple negative BCs and how does this correlate with Ki67 in BC and healthy breast tissue? 24 Comparable questions arise regarding BC risk. Are there biomarkers or even genes 33 which can recognize HT as cause of BCs? Can temporary growth due to HT be recognized in the spatial spread of a BC? Today, molecular biological differences in the BCs of HT users and non-users could improve our understanding of the effect. This may also answer the question whether stopping HT can stop or even eradicate an autonomously growing occult BC? 2,34 Comparisons with studies on estrogen treatment alone and the HT different type and prescription and BC risk are obvious, especially as MWS and WHI are contradictory in important aspects. 5,7 Hypothesis-based follow-up studies investigating BC risk from the onset of HT could provide clarification. Particular attention should be paid to the first months after screening-based recruitment and after the end of HT.

The view for patients

This reinterpretation of HT and BC risk may be relevant for patients. The decline in incidence in 2002 shows that millions of women have stopped or not started HTs, predominantly from fear of BCs. Women for whom HTs were indicated due to menopausal symptoms have accepted a reduced quality of life 3,35. The fear from BC needs to be put into perspective with a modified understanding of the interrelation of its benefits and risks. Firstly, BCs which occur under HT are not caused by HT. At 5 years of HT, about 2 prevalent BCs are brought forward for every 100 HT users. These BCs grow somewhat faster under HT. This may worsen the prognosis. Therefore, annual screening for HT users would be beneficial. Secondly, there is also a BC risk due to HT. A RF of 2 would also lead to about 2 additional BCs per 100 HT users, which, however, are not expected until 14 to 20 years later if short HTs are started at the age of 50. Early detection and regular mammography screening can improve the prognosis of these BCs. Risk reduction also includes the indication of endocrine prevention therapies, as currently approved in postmenopausal women in the UK, in whom the number of prevalent BCs can be almost halved, today perhaps with treatment of 2 years only (Fig.2A). 12

Limitations

As with all statistical modelling, the better the data, the more meaningful the results. This is related to the specification of the turning points at the end of the three Periods 1-3, which could not be derived from aggregating Forrest plots data, but from the time dependence of the cumulative incidence with a 30-year follow-up as shown in Fig. 2-3A. This would allow the growth duration of HR+ BCs and the range of HT-initiated BCs to be assigned more precisely.

The simulations show that the effects of HT can be modelled with the GAF and RF. A study “twin” with the same age distribution and the same long-term results can be generated for each of the non-consistent WHI cohorts. Incidence effects can also be explained with data on HT use in a population. Other modifying factors such as body mass index or a time dependency, e.g. on the gap between menopause and the start of HT, also require better data with correlating clinical findings.

Conclusion

Modelling of correlations and hypotheses testing can be done by generating study twins using available data. This is demonstrated with a parameter constellation of the WHI-S. HTs promote and terminate the growth of prevalent BCs. Both, the renunciation of HT and its discontinuation caused the 10.3% decline in incidence in the USA in 2002. There is a BC risk, but it occurs with HT over a few years about 15 years after the onset of HT, probably with a factor of 2. The 1 million BCs attributed to HT are not newly caused, but prevalent BCs which have been shifted forward. The small additional time-delayed risk increases the benefit of short indicated HT.

Data Availability

All data produced are available online at: http://seer.cancer.gov/ https://seer.cancer.gov/statistics-network/ http://www.tumorregister-muenchen.de/en/facts/specific_analysis.php

http://seer.cancer.gov/

https://seer.cancer.gov/statistics-network/

http://www.tumorregister-muenchen.de/en/facts/specific_analysis.php

Conflict of Interest Disclosures

The authors declare no conflict of interest.

Author Contributions

Concept and design: Hölzel

Acquisition, analysis, interpretation of data: All authors

Drafting of the manuscript: All authors

Critical revision of the manuscript for intellectual content: Halfter, Schlesinger-Raab

Statistical analysis: All authors

Administration, technical, material support: Halfter

Supervision: Hölzel, Schlesinger-Raab

Acknowledgment

We would like to thank all former MCR employees for their contributions to the excellent data quality. We would like to thank Doris Mayr, Sylvia Heywang-Köbrunner, Christian Thaler and Michael Lauseker for their valuable scientific and technical advice. We thank the heads of the breast cancer project group of the Comprehensive Cancer Centre of the Ludwig-Maximilians-University and the Technical University of Munich for repeated presentations and enriching discussions of results.

Funding: None

Footnotes

  • hoe{at}ibe.med.uni-muenchen.de, schlesi{at}ibe.med.uni-muenchen.de

Abbreviations

BC
Breast cancer
GD
Growth duration of BCs
GAF
Growth acceleration factor
HT
Hormone replacement therapy
HTD
Hormone replacement therapy duration
HR+/-
Hormone receptor positive/negative
MET
Metastasis, metastases
MWS
Million Women Stud
RF
BC risk factor
RR
relative risk
VDT
Volume doubling time
WHI-S
Women’s Health Initiative Studies

References

  1. 1.↵
    Lobo RA. Hormone-replacement therapy: current thinking. Nature reviews Endocrinology 2017; 13(4): 220–31.
    OpenUrl
  2. 2.↵
    Narod SA. Hormone replacement therapy and the risk of breast cancer. Nature reviews Clinical oncology 2011; 8(11): 669–76.
    OpenUrl
  3. 3.↵
    Marjoribanks J, Farquhar C, Roberts H, Lethaby A, Lee J. Long-term hormone therapy for perimenopausal and postmenopausal women. The Cochrane database of systematic reviews 2017; 1: Cd004143.
    OpenUrl
  4. 4.↵
    Writing Group for WHI Investigators. 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(3): 321–33.
    OpenUrlCrossRefPubMedWeb of Science
  5. 5.↵
    Chlebowski RT, Anderson GL, Aragaki AK, et al. Association of Menopausal Hormone Therapy With Breast Cancer Incidence and Mortality During Long-term Follow-up of the Women’s Health Initiative Randomized Clinical Trials. Jama 2020; 324(4): 369–80.
    OpenUrlCrossRefPubMed
  6. 6.↵
    Reeves GK, Beral V, Green J, Gathani T, Bull D. Hormonal therapy for menopause and breast-cancer risk by histological type: A cohort study and meta-analysis. The Lancet Oncology 2006; 7(11): 910–8.
    OpenUrlCrossRefPubMedWeb of Science
  7. 7.↵
    Collaborative Group on Hormonal Factors in Breast Cancer. Type and timing of menopausal hormone therapy and breast cancer risk: individual participant meta-analysis of the worldwide epidemiological evidence. Lancet 2019; 394: 1159–68.
    OpenUrlCrossRefPubMed
  8. 8.↵
    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. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet 1997; 350(9084): 1047–59.
    OpenUrl
  9. 9.↵
    Noone A, Howlader N, Krapcho M, Miller D, Brest Ae. SEER Cancer Statistics Review, 1975-2017 National Cancer Institute. Bethesda, MD,: http://seer.cancer.gov/ (accessed Feb. 5, 2024).
  10. 10.↵
    Munich Cancer Registry. http://www.tumorregister-muenchen.de/en/facts/specific_analysis.php (accessed Feb. 5, 2024).
  11. 11.↵
    Gail MH, Brinton LA, Byar DP, et al. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. Journal of the National Cancer Institute 1989; 81(24): 1879–86.
    OpenUrlCrossRefPubMedWeb of Science
  12. 12.↵
    Cuzick J, Sestak I, Forbes JF, et al. Use of anastrozole for breast cancer prevention (IBIS-II): Long-term results of a randomised controlled trial. Lancet 2020; 395: 117–22.
    OpenUrlCrossRefPubMed
  13. 13.↵
    Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for the prevention of breast cancer: current status of the National Surgical Adjuvant Breast and Bowel Project P-1 study. Journal of the National Cancer Institute 2005; 97(22): 1652–62.
    OpenUrlCrossRefPubMedWeb of Science
  14. 14.↵
    Weedon-Fekjaer H, Lindqvist BH, Vatten LJ, Aalen OO, Tretli S. Breast cancer tumor growth estimated through mammography screening data. Breast cancer research : BCR 2008; 10(3): R41.
    OpenUrlCrossRef
  15. 15.↵
    Ryu EB, Chang JM, Seo M, Kim SA, Lim JH, Moon WK. Tumour volume doubling time of molecular breast cancer subtypes assessed by serial breast ultrasound. European radiology 2014; 24(9): 2227–35.
    OpenUrlCrossRefPubMed
  16. 16.↵
    R Core Team. R: A language and environment for statistical computing. R: Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/. 2015.
  17. 17.↵
    Chlebowski RT, Kuller LH, Prentice RL, et al. Breast cancer after use of estrogen plus progestin in postmenopausal women. The New England journal of medicine 2009; 360(6): 573–87.
    OpenUrlCrossRefPubMedWeb of Science
  18. 18.↵
    Chlebowski RT, Manson JE, Anderson GL, et al. Estrogen plus progestin and breast cancer incidence and mortality in the Women’s Health Initiative Observational Study. Journal of the National Cancer Institute 2013; 105(8): 526–35.
    OpenUrlCrossRefPubMedWeb of Science
  19. 19.↵
    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(24): 3243–53.
    OpenUrlCrossRefPubMedWeb of Science
  20. 20.↵
    Beral V. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet 2003; 362(9382): 419–27.
    OpenUrlCrossRefPubMedWeb of Science
  21. 21.↵
    Beral V, Peto R, Pirie K, Reeves G. Menopausal hormone therapy and 20-year breast cancer mortality. Lancet 2019; 394(10204): 1139.
    OpenUrl
  22. 22.↵
    Engel J, Eckel R, Halfter K, Schubert-Fritschle G, Hölzel D. Breast cancer: emerging principles of metastasis, adjuvant and neoadjuvant treatment from cancer registry data. Journal of cancer research and clinical oncology 2022.
  23. 23.↵
    Wunderle M, Pretscher J, Brucker SY, et al. Association between breast cancer risk factors and molecular type in postmenopausal patients with hormone receptor-positive early breast cancer. Breast cancer research and treatment 2019; 174(2): 453–61.
    OpenUrl
  24. 24.↵
    Hofseth LJ, Raafat AM, Osuch JR, Pathak DR, Slomski CA, Haslam SZ. Hormone replacement therapy with estrogen or estrogen plus medroxyprogesterone acetate is associated with increased epithelial proliferation in the normal postmenopausal breast. The Journal of clinical endocrinology and metabolism 1999; 84(12): 4559–65.
    OpenUrlCrossRefPubMedWeb of Science
  25. 25.↵
    Engel J, Weichert W, Jung A, Emeny R, Hölzel D. Lymph node infiltration, parallel metastasis and treatment success in breast cancer. Breast (Edinburgh, Scotland) 2019; 48: 1–6.
    OpenUrl
  26. 26.↵
    Anderson GL, Chlebowski RT, Rossouw JE, et al. Prior hormone therapy and breast cancer risk in the Women’s Health Initiative randomized trial of estrogen plus progestin. Maturitas 2006; 55(2): 103–15.
    OpenUrlCrossRefPubMedWeb of Science
  27. 27.↵
    _SEER NCI. Cancer Statistics Explorer Network. https://seer.cancer.gov/statistics-network/ (accessed Feb. 5, 2024.
  28. 28.↵
    Jemal A, Thomas A, Murray T, Thun M. Cancer statistics, 2002. CA: a cancer journal for clinicians 2002; 52(1): 23-47.
    OpenUrlCrossRefPubMedWeb of Science
  29. 29.↵
    Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA: a cancer journal for clinicians 2008; 58(2): 71–96.
    OpenUrlCrossRefPubMedWeb of Science
  30. 30.↵
    Ravdin PM, Cronin KA, Howlader N, et al. The decrease in breast-cancer incidence in 2003 in the United States. The New England journal of medicine 2007; 356(16): 1670–4.
    OpenUrlCrossRefPubMedWeb of Science
  31. 31.↵
    Hersh AL, Stefanick ML, Stafford RS. National use of postmenopausal hormone therapy: annual trends and response to recent evidence. Jama 2004; 291(1): 47–53.
    OpenUrlCrossRefPubMedWeb of Science
  32. 32.↵
    Hulley S, Grady D, Bush T, et al. Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. Heart and Estrogen/progestin Replacement Study (HERS) Research Group. Jama 1998; 280(7): 605–13.
    OpenUrlCrossRefPubMedWeb of Science
  33. 33.↵
    Narod SA. Which Genes for Hereditary Breast Cancer? The New England journal of medicine 2021; 384(5): 471–3.
    OpenUrlCrossRef
  34. 34.↵
    Shapiro S, Farmer RD, Seaman H, Stevenson JC, Mueck AO. Does hormone replacement therapy cause breast cancer? An application of causal principles to three studies: Part 1. The Collaborative Reanalysis. The journal of family planning and reproductive health care 2011; 37(2): 103–9.
    OpenUrl
  35. 35.↵
    Pinkerton JV. Hormone Therapy for Postmenopausal Women. The New England journal of medicine 2020; 382(5): 446–55.
    OpenUrlCrossRefPubMed
Back to top
Posted February 17, 2024.
Download PDF
Data/Code
Email
One effect and two causes: Growth acceleration and breast cancer risk after hormone replacement therapy
Dieter Hölzel, Anne Schlesinger-Raab, Kathrin Halfter
medRxiv 2024.02.16.24302914; doi: https://doi.org/10.1101/2024.02.16.24302914
Twitter logo Facebook logo LinkedIn logo Mendeley logo
Citation Tools

Subject Area