Specific HRT tablets linked to increased risk of heart disease and blood clots

[LoremIpsumCici]

Certain hormone replacement therapy (HRT) tablets containing both estrogen and progestogen are associated with a higher risk of heart disease and rare but serious blood clots known as venous thromboembolism (VTE) in women around the age of menopause, finds a study from Sweden published by The British Medical Journal.

https://www.bmj.com/content/387/bmj-2023-078784

Thanks for the link @LoremIpsumCici
I have read most of it and one thing that stands out and is worth reading is the 'limitations of the study'.

First, it's observational not a controlled study.

The women they studied were not categorised/excluded/measured for obesity or smoking. (Two of the biggest risks for CVD.)

There was no guarantee that they took HRT- they are referred to as 'intention to treat' (from medical records) meaning it was prescribed but no record of whether they took it or for how long.

It doesn't include body-identical progesterone.

It's been known for a long time that oral HRT has a higher risk of blood clots. Be interesting to see if any drs in the UK comment on the study.

@JinglingSpringbells
Yes, limitations are always worth reading, every study has them!
But the ones on this study have been impressively mitigated.

First, it's observational not a controlled study.

The discrepancy between previous observational studies and randomised controlled trials has been attributed to population differences and methodological biases. Analysing observational data by using a design that emulates a target trial has been suggested to mitigate such pitfalls. Other than the randomisation, which is not possible in an observational study, emulating a target trial is advocated as being more accurate for causal inference than a traditional observational study.

This study was designed to emulate target trials (that is, hypothetical pragmatic trials that would have answered the causal question of interest) of the effect of menopausal hormone therapy compared with no menopausal hormone therapy on the risk of cardiovascular disease outcomes. Table 1 shows the protocol of the target trial and its observational emulation. 138 trials were emulated, one trial each month during the study period, with each trial having a one month enrolment period. This ensures that all eligible initiators and events are included in the analysis. This approach increases statistical power compared with selecting only one of those instances as time zero and accounts for the fact that individuals can meet eligibility criteria multiple times during the study period; figure 1 illustrates the study design. The emulation of a series of trials, such that each individual may participate in multiple trials, has been successfully applied in previous studies when comparing treatment versus non-treatment.

By clearly defining the target trial protocol and its observational emulation, we ensured eligibility and treatment alignment from the start, preventing selection bias by excluding prevalent users at baseline. We aimed to estimate the average treatment effect for ischaemic heart disease, cerebral infarction, myocardial infarction, and venous thromboembolism with the use of different types of contemporary menopausal hormone therapy.

The women they studied were not categorised/excluded/measured for obesity or smoking. (Two of the biggest risks for CVD.)

True, but they did exclude from both the population that did not take HRT and the population the did take HRT women who, before the start of follow-up, had a previous history of ischaemic heart disease, stroke, peripheral vascular disease, peripheral arterial disease, or cancer. They also excluded women who appeared for the first time or with a new personal identification number in the registers after 2005. This group represents either people who immigrated to Sweden or women who had received gender affirming care. This was due to missing pre-immigration medical data and potential confounding factors from gender transition procedures.

It is likely the rate of smoking and obesity were not too terribly different between the two populations, so having that data isn’t likely to have had a big impact on the differences observed.

There was no guarantee that they took HRT- they are referred to as 'intention to treat' (from medical records) meaning it was prescribed but no record of whether they took it or for how long.

True there is no guarantee they took the HRT, but it wasn’t just prescribed, but also the prescriptions were regularly redeemed (filled and collected) by the women.
”We retrieved the redeemed prescriptions for the study period from the prescribed drug register, which categorises information according to Anatomical Therapeutical Chemical codes. The relevant Anatomical Therapeutical Chemical codes for systemic menopausal hormone therapy products are listed in supplementary table S4. Additional information on name, trade name, dose, number of packages, defined daily doses per package, tablets per package, route of administration, and date of redemption was available.”

Women who did not collect their prescriptions were not included in the study.

It doesn't include body-identical progesterone.
Correct, that is recommended for future study

The women they studied were not categorised/excluded/measured for obesity or smoking. (Two of the biggest risks for CVD.)

I missed this as well
”To account for differences in predisposing diseases and disorders between initiators and non-initiators, we adjusted for hypertension, heart disease, and diabetes (defined by the use or non-use of drugs for these conditions).”

On the guarantee they took HRT, if a woman started redeeming prescriptions she was dropped from the nonHRT population. Similarly, if a woman stopped redeeming prescriptions, she was dropped from the HRT population.

Thus, in the per protocol analyses, all initiators were continuous users and all non-initiators were never users.

Looking at the stats, it seems incredible that something showing such a huge increase in risk/events hasn't been picked up by the media.

It really needs someone far better at stats than me to interpret it.

It was picked up by the media. I just decided to find and post the study the articles were based on as often journalists tend sensationalise results and scaremonger:

Telegraph
https://www.telegraph.co.uk/news/2024/11/28/hrt-linked-to-risk-of-heart-disease-and-blood-clots/

The Sun
https://www.thesun.co.uk/health/31991064/specific-types-hrt-raise-risk-heart-disease-blood-clots/

MSN Medical News
https://www.msn.com/en-gb/health/other/specific-hrt-tablets-linked-to-increased-risk-of-heart-disease-and-blood-clots/ar-AA1uTwQP

The Mirror
https://www.mirror.co.uk/news/health/women-taking-certain-hrt-drugs-34201010

The Irish Times
https://www.irishtimes.com/health/2024/11/28/certain-hormone-replacement-therapy-tablets-associated-with-higher-risk-of-heart-disease/

Medical Dialogues
https://medicaldialogues.in/mdtv/cardiology/videos/some-hrt-tablets-may-be-linked-to-increased-risk-of-heart-attack-and-blood-clot-study-138901

Australian Broadcasting Company
https://www.abc.net.au/listen/programs/healthreport/hrt-tablets-heart-disease-clot/104660116

BMJGroup
https://bmjgroup.com/certain-hrt-tablets-linked-to-increased-heart-disease-and-blood-clot-risk/

Thanks

Most of the links are behind paywalls.

The BMJ journo summary says that blood clots could be found in 7:1000 women in a year. That's 0.7% added risk.

I'm not for a minute undermining that but given the limitations which the authors themselves list, I'm just wondering if this is telling us anything new? (ie that oral HRT has always been known to have a small clot risk)?

Be interesting to see if the BMS adds comment.

@LoremIpsumCici I found this on the BMS site and thought it was worth sharing. https://thebms.org.uk/2024/11/bms-response-to-new-study-published-in-the-bmj-certain-hrt-tablets-linked-to-increased-heart-disease-and-blood-clot-risk/