Doxycycline for CD138-positive chronic endometritis did not improve live births

Design

Multicentre, adaptive, double-blind, placebo-controlled randomised trial with Bayesian analysis (27 UK sites)

Population

438 women with at least two first-trimester miscarriages and CD138-positive chronic endometritis

Primary outcome

Cumulative live birth or ongoing pregnancy at 24 weeks or more

Effect

52% vs 50%; adjusted RR 1.02 (credible interval 0.85 to 1.21)

This UK trial randomised 438 women aged 18 to under 42 with at least two consecutive first-trimester miscarriages and chronic endometritis, defined as CD138-positive cells on luteal-phase biopsy. They received doxycycline 100 mg twice daily for 14 days or identical placebo before conception. Women and clinicians were blinded.

Live birth or ongoing pregnancy occurred in 113 of 219 (52%) on doxycycline and 109 of 219 (50%) on placebo (adjusted RR 1.02, credible interval 0.85 to 1.21). Repeat biopsies in 98 placebo and 104 doxycycline women showed the same variation in CD138 staining, suggesting that biopsy findings fluctuate between cycles. A separate group who screened negative had a lower rate (30%) than the placebo arm (41%).

Some guidelines advise testing for chronic endometritis and treating it with antibiotics, and others do not. This trial offers no support for treating on CD138 staining alone. The trial stopped at an interim analysis under a Bayesian design.

• Consider not offering endometrial biopsy for CD138 staining solely to treat chronic endometritis in recurrent miscarriage.
• Discuss that a 14-day doxycycline course did not change live birth rates in this trial.
• Keep standard recurrent miscarriage work-up, which this trial does not change.
• Treat a clinically evident infection on its own indications.
• Check the local guideline position, which has not yet changed in response to this trial.

Why it matters

It removes the placebo-controlled gap behind a practice that guidelines currently treat inconsistently.

Don't overread it

One trial of one regimen; it does not change any guideline yet, and it does not address other antibiotic regimens or other definitions of chronic endometritis.

The statistics, in plain English

A credible interval of 0.85 to 1.21 runs on both sides of 1.0, so the data are compatible with a small benefit or a small harm but give no sign of the large benefit that testing would need. The lower rate in screen-negative women is a non-randomised comparison and should not be read as an effect.

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.