Stop using testosterone levels to explain gynaecomastia on enzalutamide

Design

Prospective analysis of testosterone kinetics and adverse events within a trial of intermittent ARPI monotherapy

Population

38 men with biochemically recurrent prostate cancer, median age 64.6 years, median PSA 4.38 ng/mL, no androgen deprivation

Primary outcome

Relationship between serum testosterone and incidence of gynaecomastia

Effect

Median testosterone 316 to a peak of 832.5 ng/dL (63% median rise); gynaecomastia in 16/38 (42%) with no correlation to peak level or duration

Androgen receptor pathway inhibitor monotherapy without androgen deprivation is attractive in biochemically recurrent prostate cancer precisely because it leaves testosterone alone — and the price is a rise in serum testosterone and the breast symptoms that go with it. The working assumption has been that the gynaecomastia tracks the testosterone: higher peak, longer exposure, worse breasts. This prospective analysis tested the assumption in 38 men given three months of intermittent enzalutamide with no androgen deprivation.

The testosterone rise was substantial. Median baseline was 316 ng/dL (range 167 to 723), and the median peak was 832.5 ng/dL — a median increase of 63%. Gynaecomastia symptoms were reported by 16 of 38 men, 42%. And neither peak testosterone nor the duration of raised testosterone correlated with whether a man got it. More striking still, gynaecomastia during the first course did not reliably predict gynaecomastia in the second, in the same patient.

Two things follow for clinic. First, do not order or interpret a testosterone level as a way of anticipating, explaining or managing breast symptoms on ARPI monotherapy — it will not tell you. Second, counsel on the rate, not on the mechanism: roughly four in ten men will get symptoms, it is not predictable from their numbers, and a man who escaped it on one course is not thereby safe on the next. That changes when prophylaxis is discussed. If breast symptoms are a reason a patient would stop treatment, the conversation about prophylactic tamoxifen or breast-bud radiotherapy belongs before the first course rather than after symptoms appear, because there is no level to watch that would warn you.

• Do not use testosterone monitoring to predict or explain gynaecomastia on ARPI monotherapy.
• Quote a symptom rate of about 42% when consenting for enzalutamide monotherapy.
• Raise prophylaxis before the first course if breast symptoms would threaten adherence.
• Re-ask at every course — an uneventful first course does not predict the second.
• Examine the breasts and ask directly; men frequently do not volunteer this.

Why it matters

It removes a test clinicians have been using to reason about a common toxicity, and shifts the decision to before treatment starts.

Don't overread it

With 38 patients, absence of correlation is not proof that testosterone plays no part.

The statistics, in plain English

Thirty-eight patients is small, and an absent correlation in a small sample is weaker evidence than a positive one — the study could have missed a modest relationship. But the size of the gap between expectation and observation is what carries this: if gynaecomastia were driven by the peak testosterone, a 63% median rise across the cohort should have produced a visible gradient, and it did not. The within-patient finding is the most persuasive part, because it holds the patient's own biology constant: the same man, the same drug, a different course, a different outcome.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.