Venous Thromboembolism, Myocardial Infarction and Ischaemic Stroke in Feminising Hormone Therapy

Evidence Note

What is the question?

The clinical question is narrow: does oestrogen-based gender-affirming therapy raise the risk of venous thromboembolism — a clot in the leg or lung — or of myocardial infarction (heart attack) or ischaemic stroke? And if so, by how much, caused by what, and what can be done about it?

The full review grades each outcome on four axes — did the treatment cause it, how big is the effect, would it matter to a patient, and was it measured in trans people at all — and keeps them apart, because in this field they are constantly run together.

What does the evidence say?

The venous clot risk is real, and roughly a doubling. Two independent modern cohorts — Kaiser Permanente in the United States and a national clinic in Amsterdam — put the relative risk near 1.8 to 2.0 times that of cisgender comparators. In absolute terms that is on the order of 14 to 17 extra clots per 1000 people over eight years. The events fall disproportionately in older and more medically complex patients, and they accumulate over years, not in the first few weeks.

Its direction is better established than its size. No study compares trans women on oestrogen with trans women off it. So how much of the risk is caused by modern prescribed oestradiol — rather than by the older high-dose regimens some patients once took, by other medications, or by differences between the groups being compared — is genuinely uncertain. “Roughly twofold” is a good estimate, not a fixed biological constant.

The frightening historical numbers do not describe modern therapy. The very high clot risks reported decades ago came from high-dose ethinylestradiol, a synthetic oestrogen no longer prescribed for this purpose. Contemporary 17β-oestradiol has a materially smaller effect on clotting. A “45-fold” figure still circulates in some summaries; it belongs to an obsolete regimen, could not be verified against its original source, and is deliberately kept out of this review’s graded findings.

Arterial risk is not the venous story in reverse. On the most recent data, heart-attack and stroke rates in trans women are not higher than in comparable cisgender men — and heart attacks are actually recorded less often.

But that lower number is not proof of protection. It is an observed comparison against general-population rates, not a demonstrated effect of the hormones. It does not show that feminising therapy prevents heart attacks — and it is held to exactly the same scrutiny as the clot findings. A reassuring number gets no free pass.

What actually reduces the risk?

Avoid ethinylestradiol. This is the clearest and best-supported move — the international standards of care recommend against it. Its relevance now is legacy risk and unsupervised, self-sourced use, not modern prescribing.

Transdermal (patch or gel) oestradiol for higher-risk patients. A defensible preference, and WPATH SOC-8 suggests it — a conditional recommendation — for people over 45 or with a previous clot. But this rests on menopausal (cisgender) data and on how the liver handles oral oestrogen, not on trials in trans people. No study in trans populations has directly tested whether route changes clot risk.

Prefer 17β-oestradiol to conjugated equine oestrogens. Reasonable, but the evidence is indirect. And an accompanying progestogen should not be assumed to be clot-neutral.

Use the lowest effective anti-androgen dose. For cyproterone, the main established concern is not clots but meningioma — a usually-benign tumour of the brain lining whose risk rises with cumulative dose. The European Medicines Agency restricted higher doses for this reason, so the lowest effective dose is the sensible rule.

What is not needed. Routine blood-thinning prophylaxis is not indicated for hormone therapy alone, and screening everyone for inherited clotting disorders before starting is not recommended. After a clot, continuing therapy alongside anticoagulation may be reasonable on individual assessment — though that evidence is borrowed from cisgender populations, and the largest study explicitly excluded its trans patients.

How strong is the evidence?

The venous direction is well supported; almost everything downstream of it is extrapolated. Two genuinely independent cohorts agree on the near-twofold clot signal. But the choice of route, the choice of molecule, and how to manage therapy after a clot are largely inferred from cisgender menopausal data, because no trial in trans people isolates any of them. No cohort has an untreated trans comparison group. So the review is confident about which way the risks point, and much more cautious about their exact size and cause.

Where are the uncertainties?

Several of the questions that matter most for an individual decision have not been answered directly. How much of the clot association is actually caused by modern oestradiol, rather than by era and confounding. Whether patch versus tablet changes clot risk in trans people specifically — never directly tested. Injectable oestradiol and clots — no comparative data, in either direction. The independent contribution of cyproterone or of an added progestogen to clot risk — not quantified. And long-term arterial risk as patients age and stay on therapy for decades.

What does this mean for practice?

For clinicians. The venous risk is real, modest and concentrated in older, comorbid patients. Manage the conventional risk factors, prefer 17β-oestradiol, consider transdermal for those at higher risk, and use the lowest effective cyproterone dose. Do not read the lower heart-attack rate as cardioprotection. And do not withhold desired therapy on clot grounds without weighing it against the benefit.

For patients and families. There is a real, modest increase in clot risk — larger with age and other risk factors, small when you are young and otherwise healthy. The alarming historical numbers are from a drug no longer used this way. The therapy has not been shown to protect your heart, whatever a reassuring statistic might seem to say. These are things to weigh with a clinician who knows your history — not reasons to start or stop treatment on your own.

The benefit belongs in the same decision. Feminising therapy is desired and clinically important to many of the people who seek it, and is associated with improvements in gender dysphoria and psychological well-being. The review grades that benefit at the same standard as the harms — neither inflated because it is welcome, nor dismissed. Withholding or interrupting desired therapy carries its own costs, and those belong in the decision alongside the clot risk.

For anyone reading the headlines. Two opposite errors are common: treating a real but modest clot risk as a catastrophe, and treating a lower heart-attack count as proof of safety. They are the same mistake in different directions — reading which way the risk points as though it settled how large it is, and reading a comparison as though it proved a cause.