- Design
- Systematic review and random-effects meta-analysis of observational studies
- Population
- Seven studies measuring maternal serum or plasma cytokines in pregnancy
- Primary outcome
- Association of maternal cytokines with offspring autism
- Effect
- CXCL10 OR 1.62 (1.24-2.11); IL-10 OR 1.34 (1.06-1.69); IL-17A OR 0.81 (0.70-0.94); no robust single marker
A systematic review and meta-analysis (6 October 2026) pooled seven observational studies that measured cytokines or chemokines in maternal blood during pregnancy and compared mothers of children later diagnosed with autism with controls.
Several markers showed nominal associations: higher odds of autism with IL-10 (OR 1.34, 95% CI 1.06-1.69), CXCL10 (OR 1.62, 1.24-2.11), IL-1α (OR 1.42, 1.23-1.65) and IL-5 (OR 1.49, 1.18-1.89), and lower odds with IL-17A (OR 0.81, 0.70-0.94). Heterogeneity was substantial for some markers, including IL-10. The authors concluded that maternal immune profiles may differ but that no single cytokine is a robust predictor.
This supports maternal immune activation as a line of research. It does not support testing cytokines in pregnancy, and it should not be used to imply that a maternal infection or inflammatory condition caused a child's autism.
- Do not order maternal cytokine panels to predict autism; no marker is validated.
- If parents ask whether an infection in pregnancy caused their child's autism, explain the evidence is associative and inconsistent.
- Continue routine infection prevention in pregnancy, including recommended vaccination, on its own merits.
- Expect larger prospective cohorts before any clinical use is considered.
Why it matters
Parents increasingly ask about immune causes of autism, and this gives an honest answer: some differences, nothing predictive.
Don't overread it
Seven observational studies reporting nominal associations cannot show that maternal inflammation causes autism.
The statistics, in plain English
Many cytokines were tested, so a few will reach significance by chance. 'Nominal' means before any correction for that, which is why the authors do not claim a firm biomarker.
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.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for obstetrics & gynaecology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free