Identifying socioeconomic and ethnic disparities in key health outcomes by hormone replacement therapy use in women aged 40-80

There are known disparities in hormone replacement therapy prescribing and uptake in UK primary care.(1) The research team was awarded funding from NIHR RfPB(2) to identify groups of women who are receiving HRT prescriptions and which women are experiencing or likely to be experiencing the most severe menopause symptoms which is in progress, scheduled to complete in March 2025. That work will tell us about disparities in HRT use by ethnicity, socioeconomic status and region of the UK using routinely collected healthcare data. However, we know from speaking with patient and public representatives, that women from ethnic minority backgrounds are not confident that the benefits or risks they may experience from using HRT, would be the same as the general population.

In this proposal, we plan to investigate disparities in health outcomes with a particular focus on ethnic and socioeconomic differences according to HRT use. This can be done using datasets of routinely-collected healthcare data which include data on self-defined ethnic group and the Townsend deprivation index calculated using individual postcodes. This has the advantage over Index of Multiple Deprivation (IMDS) in that it does not include a measure of health which would confound the analyses. It is also evaluated at output area which is more granular than the IMDS and hence potentially more accurate for analyses.

Some health outcomes are of particular importance in relation to HRT use. These include cancers, comprising breast, ovarian and endometrial;(3) cardiovascular disease generally, and thrombosis in particular,(4) dementia(5) and osteoporosis, which may be reduced with HRT use.(6) These risks will also differ depending on the type of HRT used (transdermal vs oral), the duration of use and whether oestrogen alone, combined HRT or type of progesterone.(7)

************************************************
Importance in terms of improving the health and/or wellbeing of the public and/or to patients and health care services

Some women have concerns about the risks from taking HRT,(8) but those from non-white ethnic backgrounds may have different background risks of adverse health outcomes and carry a different pattern of comorbidities from white populations. This adds additional uncertainties to the risks of developing serious health conditions for which there is currently very little evidence. It is important for all women, regardless of ethnicity or comorbidities, to understand whether there may be additional benefits or harms from taking HRT so they can make fully informed decisions regarding any treatment.

Little is known about the associations between ethnicity, socioeconomic status and comorbid conditions with health outcomes if women choose to use HRT. Some conditions or comorbidities may be more common or less well-controlled in some ethnic groups or in people living with higher levels of deprivation such as haemoglobinopathies (sickle cell disease or thalassemia),(9) diabetes,(10) obesity(10) or hypertension.(11) People living with some of these conditions may be excluded from trials assessing the efficacy or safety of medications or treatments such as HRT. This means that the only way to evaluate the impact of taking medications in some populations is to look at routinely collected data on health outcomes once the medications are being widely used.

Women living with some conditions may already have a higher risk of poor health outcomes, yet it is uncertain how HRT may change this risk.(12) For example, we already know that some health conditions, such as diabetes, are more prevalent both in areas of high social deprivation(10) and in some ethnic minority groups in the UK.(13) Additionally, obesity is also known to be prevalent in areas of high social deprivation, and is reported to be associated with more severe menopause symptoms(14). Both obesity and diabetes are risk factors for cardiovascular disease (CVD), and the effect of HRT on CVD risk remains uncertain.(15)

There therefore remains a question as to whether findings from previous studies on the risks of health outcomes with HRT use can be extended to women from different ethnic backgrounds, and whether risks are similar in different socioeconomic groups or people living with other health conditions. Furthermore, the 2023 NICE Draft menopause guidelines(7) have highlighted the need for research on the impact of HRT on health outcomes in both ethnic minority groups and those with early menopause. The health outcomes specified by NICE include cardiovascular disease, breast, endometrial and ovarian cancer, dementia and all-cause mortality.

*************************************
Review of existing evidence – How does the existing literature support this proposal?

There are known disparities in breast cancer diagnosis in the United States, with black women having both the lowest odds of being diagnosed (35% lower compared with white women) and yet have the poorest survival.(16) Additionally, there are large differences in survival by socioeconomic status.(17) These inequalities may be because of the inequalities in access to healthcare in the United States. Incidence of breast cancer in the UK has been reported to be lower in South Asian and Black women compared with white women,(18) but age at diagnosis is lower in Black and Asian women compared to white women and Black women are more likely to be diagnosed with higher grade aggressive tumours.(19)

The current evidence on whether HRT use should be avoided for women with risk factors can be conflicting and confusing. For breast cancer, there is a small increase in risk with HRT which increases with the age of the woman, whether it is combined vs oestrogen alone and duration of treatment,(20) but whether and how these risks differ in different ethnic groups and social groups who may be living with different burdens of comorbidities(21) is not known.

There is an increased risk of endometrial cancer for sequential combined, but not for continuous combined HRT.(7) There is an increased risk in ovarian cancer with oestrogen only or combined HRT use for five years or more which increases with duration of use.(7)

Reports on the effect of HRT use on risk of endometrial and ovarian cancer by ethnic group are sparse,(22) though incidence of and death from ovarian cancer may be lower in some ethnic groups compared to white populations.(23, 24) A recent systematic review concluded that there were a lack of high quality studies on HRT use and endometrial cancer and it remains unclear how to advise women with a history of endometriosis regarding the management of menopausal symptoms(25) and whether advice should differ by ethnic group. In the USA, there is a high burden of morbidity due to osteoporosis and associated fractures. There is some evidence that prevalence of osteoporosis may be lower in black women compared to white women(26) but mortality following fracture may be higher in black women(27) and Asian populations may have a higher risk of osteoporosis.(26) Furthermore, deprivation is a risk factor for osteoporosis(28) and there are racial disparities in the management of osteoporosis.(29) Thalassaemia and sickle cell are known to adversely affect bone density,(30) yet little is known about menopause in women living with sickle cell(31) and how living with this condition may affect health outcomes.

There are ethnic disparities in cardiovascular disease, where Black women in the US have a lower life expectancy compared with women from other ethnicities.(32) In the UK, it has been found that women of Pakistani origin were at highest risk and those of Chinese origin at lowest risk of poor cardiovascular outcomes.(33) HRT is associated with an increased risk of venous thromboembolism,(34) and risks are greater with oral HRT compared with transdermal.(7) There is some evidence that some HRT formulations may increase risk of stroke and these risks may be higher in Black women compared to white women.(7) However, the impact of HRT on CVD risk generally is uncertain and there is very little evidence on how risks differ by ethnicity, meaning it is difficult for clinicians to advise women from different ethnic groups on the best treatment options.

There are also uncertainties about the association between HRT use and dementia.(5) An analysis of Danish registry data found that combined HRT is associated with all-cause and late dementia, particularly with long-term use.(5) The 2023 NICE draft menopause guideline(7) reports that combined HRT may increase the risk of dementia if started over the age of 65. They found no evidence of an increased risk of dementia with oestrogen only HRT.

The evidence for each of these outcomes for women from ethnic minority group and the 2023 NICE Draft menopause guidelines(7) have highlighted research needs on evaluating the impact of HRT across a range of health outcomes, both in those from ethnic minority groups and in those with early menopause.

This proposed work will address these research recommendations. It will provide an opportunity to increase knowledge using UK data and provide information to groups of women who may be under-represented in clinical trials. This can be achieved by linking primary care records with cancer registry, hospital admissions and Office for National Statistics mortality data.

General Practice data

This project will examine health outcomes in women who were exposed to HRT.

• demographics - age, sex, ethnicity, deprivation, geographical region, gp practice,
• HRT uses
• cancer diagnoses, thrombosis, dementia, osteoporosis and all-cause mortality
• co-morbidities
• prescribed medication

Hospital Episode Statistics - Admitted Patients

Hospital Episode Statistics - Outpatients

Hospital Episode Statistics - Emergency Care

Hospital Episode Statistics - Critical Care

Cancer Registry

Dr Jennifer Hirst - University of Oxford

Prof Julia Hippisley-Cox - University of Oxford

Prof Carol Coupland - University of Oxford

Dr Sarah Hillman - University of Warwick