438 Women with Recurrent Miscarriage and Confirmed Chronic Endometritis Randomized… Head-to-Head Comparison of Doxycycline vs. Placebo
Live Birth or Pregnancy Continuation Beyond 24 Weeks at 52% vs. 50%… No Antibiotic Benefit Confirmed
CD138 Levels Changed Similarly in Both Antibiotic and Placebo Groups… Calling the ‘CD138-Positive → Antibiotics’ Formula into Question
There is a diagnostic test that women who have experienced recurrent miscarriage or recurrent implantation failure often encounter: the evaluation for Chronic Endometritis (CE).
When a biopsy returns a “CD138-positive” result, treatment often follows: taking an antibiotic such as doxycycline, followed by a repeat endometrial biopsy to verify whether CD138 has cleared.
Yet an essential question remained unanswered: “If a woman with recurrent miscarriage tests positive for CD138 and takes antibiotics, does it truly lower the risk of another miscarriage and increase her chances of having a baby in the next pregnancy?”
The results of a large-scale randomized clinical trial directly comparing doxycycline against a placebo to answer this exact question have now been published.
To state the conclusion first: no significant difference was identified in live birth or sustained pregnancy outcomes between women who took doxycycline and those who received a placebo.
First, What Is ‘CD138’?
“CD” stands for Cluster of Differentiation, referring to surface markers used to identify cell types and their differentiation states. Among them, CD138 is used as an immunohistochemical marker to detect plasma cells.
Chronic endometritis is a state of persistent, low-grade inflammation within the endometrium, often presenting without noticeable symptoms. Histologically, plasma cell infiltration into the endometrial stroma has long been regarded as its hallmark, and CD138 immunostaining is utilized to visualize these cells more reliably.
Simply put, CD138 itself is neither an inflammatory substance nor a bacterium. It is best understood as a cellular “signpost” used to identify plasma cells and related cell types within endometrial tissue.
Consequently, testing CD138-positive does not inherently mean “a specific pathogenic bacteria is present in the uterus.”
While dysbiosis of the endometrial microbiome or specific microorganisms are frequently cited as potential causes of chronic endometritis, the definitive etiology remains unclear. Furthermore, diagnostic thresholds—specifically, how many CD138-positive cells per area are required to diagnose chronic endometritis—vary substantially across studies.
Nonetheless, as associations between chronic endometritis, recurrent miscarriage, and recurrent implantation failure were reported, administering antibiotics to women testing positive for CD138 became widely adopted in clinical practice. Some studies reported improved pregnancy outcomes in women whose chronic endometritis cleared post-treatment. However, a significant portion of these were observational studies, and well-designed, placebo-controlled randomized clinical trials remained scarce.
This is precisely why this new clinical trial commands attention.
2,178 Women with Recurrent Miscarriage Screened… 438 CD138-Positive Women Randomized
A research team from the University of Warwick and the University of Oxford, among other institutions in the UK, recruited women aged 18 to 41 who had experienced two or more consecutive early pregnancy losses.
A total of 2,178 women were screened across 27 UK medical centers, of whom 505 were confirmed to be CD138-positive. Ultimately, 438 participants were randomized into the trial.
The investigators performed an endometrial biopsy during the luteal phase and evaluated chronic endometritis via CD138 immunohistochemistry.
In this study, having 5 or more CD138-positive cells per 10 mm² of endometrial tissue was classified as the presence of chronic endometritis.
However, unified international consensus on how many CD138 cells define chronic endometritis is still lacking. This trial broadly included CD138-positive stromal cells rather than strictly counting only cells with classic plasma cell morphology. The researchers noted that these diagnostic threshold differences should be considered when interpreting the findings.
Among the 438 participants, 219 received doxycycline 100 mg twice daily for 14 days. The remaining 219 received an identical-appearing placebo on the exact same schedule.
Crucially, the medication was not administered during pregnancy; it was started on the first day of the menstrual cycle prior to attempting pregnancy.
Furthermore, neither the patients, the treating physicians, nor the outcome assessors knew who was taking doxycycline and who was receiving the placebo. Conducted across multiple centers as a double-blind, randomized, placebo-controlled trial, the study design provided high methodological rigor.
Doxycycline 52% vs. Placebo 50%… Virtually No Difference
What the researchers prioritized as the primary outcome was not simply whether participants achieved pregnancy. They evaluated whether participants achieved a live birth or sustained pregnancy beyond 24 weeks gestation by the end of the study period.
The results ran counter to common clinical expectations.
In the doxycycline group, 113 of 219 women (52%) achieved a live birth or maintained pregnancy beyond 24 weeks, compared to 109 of 219 women (50%) in the placebo group. While the doxycycline group showed a 2 percentage point numerical increase, this difference was not statistically significant (adjusted relative risk [aRR] 1.02; 95% Bayesian credible interval [CrI] 0.85–1.21).
During the study period, 299 of the 438 participants (68%) conceived on their first attempt. Including subsequent pregnancies, a total of 374 pregnancies occurred, of which 316 (84%) were spontaneous conceptions. Comparing first-pregnancy outcomes or time-to-pregnancy separately yielded identical conclusions: women who took doxycycline demonstrated no clear clinical advantage.
A More Intriguing Finding… CD138 Changed Similarly in the Placebo Group
Alongside pregnancy outcomes, another finding warrants close attention: the question of whether antibiotics actually clear CD138.
The investigators conducted a follow-up endometrial biopsy after treatment in 202 participants. Surprisingly, the quantity and distribution of CD138-positive cells changed similarly not only in the doxycycline group, but also in the placebo group that received no antibiotics at all.
This means that observing a reduction in CD138 on a subsequent test following antibiotic treatment cannot be assumed to be a direct therapeutic effect of the drug. The researchers suggested the likelihood of inter-cycle variation, where CD138 levels fluctuate naturally across menstrual cycles.
This observation is significant. Historically, when CD138 turned negative after treatment, clinicians and patients readily concluded that “the antibiotics resolved the chronic endometritis.” Yet this trial demonstrates that comparable clearance patterns occurred in the untreated placebo cohort as well.
‘CD138-Positive → Bacterial Infection → Antibiotics’ Are Not Synonymous
Another issue raised by this study is how a CD138-positive result should be interpreted clinically.
A CD138 assay does not detect or isolate bacteria. It is a histological examination that identifies plasma cells associated with inflammation within the endometrium. Consequently, being “CD138-positive” does not equate to having an active “bacterial infection.”
Diagnostic standards also remain unstandardized. Testing methodologies, tissue processing, and cell count criteria differ between studies, and consensus is divided over what threshold constitutes true chronic endometritis.
In this study, re-analyzing the data using more stringent, higher CD138 cutoffs did not alter the primary outcome: no evidence was found that doxycycline improved pregnancy or birth rates.
Ultimately, reflexively linking “CD138-positive → chronic endometritis → bacterial infection → antibiotics” oversimplifies a biological process that still requires substantial clarification.
Why Did Earlier Studies Report Antibiotics Were Beneficial?
Prior literature has suggested that treating chronic endometritis could improve pregnancy outcomes.
In particular, clinical reports indicating higher live birth rates in women whose follow-up biopsies confirmed resolution of chronic endometritis led to the widespread adoption of antibiotic regimens, including doxycycline.
The American Society for Reproductive Medicine (ASRM) 2026 Committee Opinion on Recurrent Pregnancy Loss noted that a meta-analysis pooling 12 earlier studies observed improved live birth rates when follow-up biopsies confirmed clearance of chronic endometritis after treatment.
However, the ASRM simultaneously pointed out that prior evidence was constrained by heterogeneous patient populations, divergent definitions of chronic endometritis, and a notable absence of robust randomized clinical trials. The committee also acknowledged that in this 438-patient randomized trial, doxycycline failed to significantly improve miscarriage or live birth outcomes.
This highlights why observational studies must be distinguished from randomized, placebo-controlled trials.
Even if women whose CD138 cleared after antibiotics experienced higher birth rates, that correlation does not prove that the antibiotics caused the successful delivery. Multiple confounding factors could be involved—such as the possibility that women with a more favorable underlying baseline prognosis naturally cleared inflammation, or that CD138 levels simply fluctuated across cycles.
This clinical trial randomized patients between doxycycline and placebo specifically to isolate and minimize these confounding influences.
This Does Not Mean ‘Antibiotics Are Never Needed for Chronic Endometritis’
Conversely, these findings should not be overextended.
This study does not demonstrate that antibiotics are ineffective for all presentations of chronic endometritis.
More precisely, it shows that: in women with recurrent miscarriage categorized as having chronic endometritis based on CD138-positive stromal cells, administering pre-conceptional doxycycline 100 mg twice daily for 14 days did not increase the likelihood of live birth or pregnancy maintenance beyond 24 weeks compared to placebo.
These conclusions cannot be directly applied to cases involving clinically confirmed, specific bacterial pathogens, pelvic inflammatory disease, or other distinct intracavitary pathologies.
The trial also had methodological boundaries. Following an adaptive Bayesian design, the trial was stopped early at an interim analysis, and the observed prevalence of CD138 positivity was higher than the investigators initially projected.
Furthermore, because the post-treatment repeat biopsy was conducted within the same treatment cycle, this study alone cannot determine whether CD138 changes persisted into subsequent menstrual cycles.
What Matters Is Not the ‘Biomarker Value’, but Whether a Baby Was Delivered
The clinical takeaway for women undergoing treatment for recurrent miscarriage or subfertility is straightforward:
A laboratory biomarker improvement does not necessarily equal an improved pregnancy and live birth outcome.
Converting CD138 from positive to negative is a surrogate endpoint. In recurrent miscarriage care, the meaningful clinical outcome for patients is whether the intervention reduces pregnancy loss, sustains gestation, and increases the probability of delivering a healthy infant.
In this trial, doxycycline did not outperform placebo in achieving that primary clinical objective.
Rather than summarizing the findings as “antibiotics are useless,” it is more accurate to state that evidence does not support routinely prescribing doxycycline to women with recurrent miscarriage based solely on a CD138-positive test result to improve live birth rates.
This trial demonstrates that the familiar clinical sequence of “positive test → treatment → negative re-test” does not automatically translate into “treatment → increased probability of a live birth.”
※ Study Source: Published online August 2026 in the international journal Human Reproduction. Conducted by researchers from Warwick Medical School, Warwick Clinical Trials Unit, and the Nuffield Department of Women’s & Reproductive Health at the University of Oxford. Title: “A double-blind multi-centre randomised placebo-controlled trial of doxycycline in women with recurrent miscarriage and chronic endometritis defined by CD138 positive stromal cells.” DOI: 10.1093/humrep/deag133, PMID: 42618515.
※ The contents of this article do not replace individualized medical diagnosis or care. Diagnostic criteria for chronic endometritis and the clinical relevance of CD138 test results can vary depending on a patient’s clinical history. Testing positive for CD138 is not synonymous with an active bacterial infection, nor does this imply that antibiotics are unnecessary when specific clinical infections or other therapeutic indications exist. Decisions regarding diagnostic testing and antibiotic therapy must be made in consultation with an attending physician, taking into account other etiologies of recurrent pregnancy loss, histological findings, infectious evaluations, and individual clinical circumstances.
※ The images used in this article were generated using artificial intelligence (ChatGPT, OpenAI) as illustrative reference materials and do not depict real individuals.
※ Copyright © 2026 THE FERTILITY NOTES. All rights reserved. Unauthorized reproduction, duplication, distribution, scraping, AI training, or data utilization of all content, including articles, photos, and images, is strictly prohibited and subject to legal liability.