TRT and Venous Thrombosis Risk: What We Know After the TRAVERSE Trial

Venous thromboembolism (VTE), comprising deep vein thrombosis (DVT) and pulmonary embolism (PE), is the third most common acute cardiovascular disease. A clot forms most often in the veins of the legs, and when it breaks off and reaches the lungs it can be life-threatening. The question of whether testosterone therapy raises this risk has returned regularly for years — from early warnings by regulatory agencies, through reports from individual centers, to the largest randomized trial to date.

An honest answer requires distinguishing three levels of evidence: randomized trials (the strongest, but often too small to detect rare events), observational studies (large, but prone to confounding) and biological mechanisms (plausible, but not clinically proven). In this article we go through them in turn, quoting numbers only from the source publications.

The best-documented mechanism is the effect of testosterone on blood formation. Androgens stimulate red blood cell production: they increase erythropoietin output and lower hepcidin, which raises iron availability for the bone marrow. The result can be a rise in hematocrit — the proportion of blood volume occupied by red cells. The higher the hematocrit, the greater the blood viscosity and, in theory, the greater the tendency to form clots. That is why monitoring hematocrit is a permanent part of TRT follow-up, and we describe the details in our entry on hematocrit on TRT.

The second possible mechanism concerns clotting factors, platelets and inflammation. Testosterone may affect the concentrations of some coagulation and fibrinolysis proteins, and some studies suggest changes in leukocyte and platelet counts during treatment. The third is conversion to estradiol, which in women on hormone therapy is linked to a higher risk of VTE — but for men on TRT this is a hypothesis, not a proven phenomenon. Finally, it cannot be ruled out that some thrombotic events result from overlapping risk factors: surgery, immobilization, cancer or untreated thrombophilia.

How should this result be read? First, TRAVERSE is the largest and best-designed cardiovascular safety trial of testosterone: it answers the question about heart attack, stroke and cardiovascular death, and in that respect it showed no adverse effect. Second, pulmonary embolism, atrial fibrillation and acute kidney injury were secondary events, and their higher incidence in the testosterone group was noted by the authors as a safety signal, not a settled causal effect — the trial was not designed and statistically powered to settle these endpoints.

Third, the study population was older men with heart disease or a high risk of it; results should not be transferred directly to younger men without such burdens, although there is no reason to think the situation is worse for them. Fourth, a single preparation (a transdermal gel) was used; results for long-acting injections may differ because the concentration profile is different. We describe cardiovascular follow-up more broadly in our entry on heart attack risk on TRT.

Before TRAVERSE was published, answers were sought in pooled analyses of smaller trials. One of them included only randomized trials.

These data must be read together with TRAVERSE. Venous thrombosis is a relatively rare event, so trials of a hundred to a few hundred people lasting a few months have very little statistical power to detect a difference. A result of 'no significant association' in such an analysis therefore means 'not detected', not 'excluded'. That is why the signal from TRAVERSE — the only trial that was both large and long — should be taken seriously, with caution about its size.

Since the most plausible mechanism runs through red blood cells, the natural question is whether the risk of thrombosis depends on whether hematocrit actually rose. A large observational analysis from a US clinical database addresses it.

The practical sense is this: risk does not seem to be spread evenly across everyone treated but concentrated in those whose hematocrit rises above a threshold (52% in this study). It is not clear whether hematocrit itself is the cause or a marker of other features (such as sleep apnea, smoking, dehydration or an overshoot dose) — but under both interpretations regular monitoring and acting on high hematocrit are justified. Specific decision thresholds and ways of proceeding are covered in the entry on the safe hematocrit value.

Not every man starting TRT has the same baseline risk of thrombosis. The greatest risk factor for another event is considered to be a previous venous thrombosis or pulmonary embolism. Also relevant are inherited thrombophilias (e.g. the factor V Leiden mutation) and acquired ones (e.g. the presence of lupus anticoagulant), cancer, recent surgery, prolonged immobilization or travel, obesity, smoking, dehydration, and concurrent use of other agents that increase red cell counts (e.g. erythropoietin, anabolic agents outside medical supervision).

Case reports and clinical series from a single center specializing in clotting disorders suggested that men who developed thrombosis after starting TRT more often turned out to have an unrecognized thrombophilia, and the authors of those papers recommended screening before treatment. These are, however, data of low reliability (no control group representative of all those treated, a referral center), and routine thrombophilia testing is not a standard in TRT guidelines. A more sensible approach is a careful personal and family history — whether anyone in the family had thrombosis at a young age or recurrent clots — and only then a decision about tests.

The opposite myth is just as wrong: that testosterone 'causes clots' and is therefore dangerous for everyone. The meta-analysis of randomized trials found no association, and in the observational study of hypogonadal men without erythrocytosis the risk was not raised compared with untreated men. The picture is therefore this: the risk is probably small in the whole population, may be important in selected subgroups, and therefore calls for individual assessment, not panic or dismissal.

It is also worth remembering the limits of the available data. TRAVERSE tested one preparation in older men with cardiovascular burdens; the meta-analyses include short, small trials; observational analyses cannot fully eliminate the influence of confounders. None of these sources allows a single reliable figure to be given for 'by how much the risk of thrombosis increases'.

A sound approach to thrombosis on TRT is neither alarmism nor carelessness. The best available data do not point to a clear increase in risk for those treated overall, but the signal from TRAVERSE and the higher risk in people with erythrocytosis or thrombophilia justify a careful history before treatment, regular hematocrit checks and knowing the warning symptoms. For a man with a previous thrombosis, the decision about TRT should be made together with a doctor who knows his history.