roundups & rapid reactions

Scientists comment on a press release from the Menopause Society about testosterone therapy and cardiovascular disease in postmenopausal women, about a presentation to be given at the 2026 Annual Meeting of The Menopause Society in San Diego.

Dr Adam Brentnall, Reader in Biostatistics, Queen Mary University of London (QMUL), said:

“This is an interesting finding, and assuming the data are robust and the study was well conducted, the authors would be right to describe their results as a ‘potentially important’ safety signal that needs further investigation.

“However, additional caution is also needed because there is no accompanying report, and insufficient detail in the press release to assess the scientific value of the study. One way the analysis could go wrong is if the reason for the difference between the groups is not testosterone use, but an unmeasured factor, such as cholesterol, or something else that was not available for the analysis.”

Prof Stephen Burgess, Professor in Biostatistics, University of Cambridge, said:

“This is a press release rather than a peer-reviewed study, so we are limited as to what information we have. The press release describes the analyses performed and gives the headline results, but we do not have the same level of detail to judge the findings as we would with a peer-reviewed study. Additionally, this is an observational comparison: it tells us that women who take testosterone have higher risk of cardiovascular disease compared to women who do not take testosterone. It cannot tell us whether testosterone is responsible for that difference or not. However, the study authors report that their analysis compared individuals who were matched based on many other potential cardiovascular risk factors, which helps narrow down the range of possible causes.

“This study shows a substantial difference between cardiovascular disease risk in women who take testosterone and those who do not – an eight-fold difference in risk. Cardiovascular disease is relatively common, so an eight-fold difference is concerning. While this study cannot prove that testosterone treatment is the cause of this difference in risk, it does raise concerns, particularly as the authors have accounted for several other potential causes. It is well-known across societies and cultures that men have shorter average life expectancy than women, and testosterone is a potential reason for this. Previous studies have raised similar concerns about the cardiovascular risk of testosterone supplementation in men. This study provides suggestive evidence that testosterone therapy may increase cardiovascular disease risk in post-menopausal women, but further research is needed to strengthen our confidence in this finding.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, Open University, said:

“To comment on a conference press release like this is pretty frustrating for me. The SMC ask me for comments because I’m an applied statistician, and not (in this case) an expert on the clinical and physiological aspects of menopause studies and hormone treatment. As a statistician, my expertise is in understanding how far the procedures of the study allow it to come to the conclusions that it reaches. Does what they did allow the researchers to come to the conclusions they claim? Are there important limitations of the work?

“But here I have only a conference press release. Sometimes a full research paper turns out to be available as well, but if there is one in this case, I couldn’t find it. There are some indications in the press release that the research used appropriate methods, to some extent anyway. But often the devil is in the detail. I don’t yet have access to the detail, and I don’t know whether any of the work has been subject to peer review.

“The study is observational. That is, women were not allocated at random to have testosterone therapy or not to have it, as would occur in a randomised controlled trial (RCT). Instead, the researchers found women who had had testosterone therapy, and other women who had not, and compared the chance of a new diagnosis of cardiovascular disease (CVD) in the two groups over a follow-up period averaging about 3 years. They found that, though the chance of a new CVD diagnosis was higher in women who had had testosterone therapy than in women who hadn’t, neither group had a particularly high risk of having a CVD episode during the relatively short follow-up time.

“Because it’s an observational study, it’s not possible to be sure that the observed difference in CVD risk was actually caused by the testosterone therapy. It could have been caused instead, in whole or in part, by some other factor or factors, that just happen to be associated with whether the women had the testosterone therapy or not.

“The researchers were aware of this, and used a method called propensity score matching to allow for a long list of other factors that might affect CVD risk. (The factors are listed in the press release.) They also carried out alternative statistical analyses, called sensitivity analyses, to check various aspects.

“It’s admirable that they did this, but without access to more details of exactly what was done, I can’t assess how convincing these approaches were. It’s possible that they were done as well as they could have been – but in any case it’s never possible to be sure that absolutely every relevant factor has been taken into account, so as with other observational studies, one just can’t be certain that the observed difference in CVD risk is caused by the testosterone therapy. It might be, but there remain other possibilities.

“It’s presumably because of these remaining doubts that the press release says only that testosterone therapy may increase CVD risk in postmenopausal women, and the lead researcher is quoted as asking for more research, rather than making a specific clinical recommendation. There may be other limitations of the study as well, but the press release doesn’t mention them. (They would be listed in a full research paper.)

“Further research on this topic would, however, not necessarily be straightforward. Observational studies, however carefully done, tend to leave important doubts about cause and effect. RCTs are expensive, particularly if most of the women involved are not going to have a new CVD, and either they have to have a long follow-up (to catch enough CVD cases), or they have to use indicators associated with CVD rather than actual CVD diagnoses, which makes it harder to draw conclusions about the risks of actual CVD.”

Press release title: ‘Testosterone Therapy May Increase Risk of Cardiovascular Disease in Postmenopausal Women’.

There is no paper.

Declared interests

Dr Adam Brentnall: “No conflicts for this.”

Prof Stephen Burgess: “I have no relevant conflict of interest to declare.”

Prof Kevin McConway: “I have no conflicts of interest to declare.”