20 November 2019

HRT: The art of dealing less in numbers and more in people

Clinical Endocrinology Women

In shades of the notorious, Women’s Health Initiative (WHI) study 1 back in 2002 that saw a generation of doctors abandon menopausal hormone therapy prescribing, a recent meta-analysis in The Lancet 2 has reignited concerns among Australian GPs.

In a survey of more than 200 GPs, more half admitted the recent study had them worried about the risk of breast cancer associated with the menopausal treatment.

In addition, doctors reported that a significant proportion (40%) of their menopausal patients also had concerns about the breast cancer risk, courtesy of The Lancet paper, according to the survey conducted by the education provider, HealthEd.

While still significant, the reaction is nowhere near the backlash that occurred after the WHI publication in JAMA back in in 2002. However, as we now know, much of the angst at that time was the result of the research findings being misinterpreted by the global media.

In a nutshell, absolute risk was confused with relative risk. The 26% increase in relative risk of breast cancer associated with HRT (an extra eight cases per 10,000 women) was incorrectly interpreted as one in four women would develop the malignancy if they took the hormonal treatment.  Even though the errors were soon found and the corrections issued, much of the damage had already been done.

Fast forward to 2019. As soon as the recent The Lancet paper was published, experts on the clinical side of the fence were suggesting that the paper was a flawed exercise in numbers pushing an epidemiologist agenda without proper consideration of the patient context.

The Lancet paper was a meta-analysis of 58 observational studies involving more than 600,000 postmenopausal women, of whom 108,000 had developed breast cancer between 1992 and 2018.

Among their conclusions, research authors suggested that after five to 14 years on combined estrogen progestogen therapy, the risk of breast cancer doubled compared with never users. The authors also said that if these estimates were correct it could translate to one million extra cases of breast cancer being attributable to menopause hormone therapy over the past 30 years.

Fairly chilling statements.

However, within a day of its publication, critics were pointing out flaws in the research. Firstly, there is an inherent risk in pooling data from observational studies – observational studies cannot determine cause and effect. Also the study had a median diagnosis year of 1999, so did not take into account significant new therapy changes currently in practice.

In addition, in the context of the high rates of breast cancer in the community overall, the additional risk due to HRT is very small. Critics also suggested the researchers selectively reported the data in parts of the study.

Behind the almost immediate and highly critical response, are concerns among women’s health experts that the dramatic conclusions might again deter GPs and patients from  menopausal hormonal therapy unless these findings are quickly and definitively put into perspective.

Once again we might see symptomatic menopausal women lose access to a very effective treatment option on the basis of incorrectly exaggerated concerns.

Four of Australia’s leading clinical experts were so concerned about the potential for over-reaction to the paper that they recently participated in a webinar panel discussion to address the issue.

The webinar, conducted by HealthEd, attracted one of the largest audiences in the  education provider’s history.

The panel included Dr Terri Foran, sexual health physician at the Royal Hospital for Women in Sydney; Dr Sonia Davidson, endocrinologist and president of the Australian Menopause Society; Dr Bronwyn Stuckey, a clinical endocrinologist at Sir Charles Gairdner Hospital, Perth; and gynaecologist Dr Elizabeth Farrell, head of the Menopause Unit at Monash Medical Centre in Melbourne.

Dr Davidson said that the statements made in the media would be “terrifying” for the typical patient and highly concerning for most GPs, especially given that almost all women whom she sees already have a high level of concern around breast cancer and HRT.

“This paper took us a huge step backwards in being able to discuss breast cancer with women,” she said.

All of the experts commented on the failure of the study authors to adequately communicate the absolute risk as opposed to the relative risk.

“We’re women and we get breast cancer because we’re female,” said Dr Farrell. “And obviously, if we have hormone exposure, we are exposed to a slightly greater risk of breast cancer.”

The paper presented big numbers, but that’s the prevalence of breast cancer. The incremental increase related to hormone therapy has not been explained in plain language that would make sense to a typical menopause sufferer.

Several members of the panel pointed to the study being a collage of largely observational studies and that simply was not good enough to justify the media statements about the findings.

“Epidemiology has its place because you can’t do randomised control for everything,” said Dr Stuckey. But she said the role of epidemiology was usually to highlight which were the questions that needed answering with more rigorous research using more scientifically robust methods such as randomised controlled studies.

The issue of age of the studies included in the meta-analysis again came up in the panel discussion. In the period in which the data was collected, menopausal treatments were different to what is used today. Current treatments have been shown to confer less of a breast cancer risk than previous therapeutic options.

Dr Davidson said, these days, the much safer “body-identical” micronised progesterone was being increasingly used.

“All the international societies have said, if you’re looking at progesterone, this is probably your safest option, not only for breast cancer, but for VTE and for other issues and other risks,” she said.

However such therapy was not available, let alone in use, when the studies included in the meta-analysis were conducted.

Despite all the criticisms of The Lancet study and the unwarranted claims made by the study authors, several of the panelists made the point that epidemiology had made an important contribution in this area of medicine.

According to Dr Stuckey, epidemiological studies have played an important role in our understanding of menopause and menopausal hormonal treatments over time.

“We’ve learned something from WHI,” she said, particularly around the risks of progestogen and breast cancer.

“[But it’s] a matter of how we present risks and benefits in the context of the quality of life of women.”

In the end, the key issue is one of weighing up the risks versus the benefits for the individual female patient sitting in front of you in your clinical practice.

Even aside from all the flaws in The Lancet meta-analysis highlighted above, there was no recognition within the paper of the importance of the individual circumstances of the patient.

“Epidemiologists don’t sit in front of a patient who’s having terrible symptoms,” Dr Stuckey said. “I would say that epidemiologists deal in numbers, we deal in people”.

All the experts agreed, we need to ensure the research is accurately reported and the risk of breast cancer associated with menopausal treatments stays in perspective especially for those women whose quality of life may depend on being able to take it.


JAMA. 2002 Dec 11;288(22):2819

Lancet 394 (10294) 1159-1168, Sept 28, 2019

COVID-19 live update