- Design
- Multi-site observational study with Bayesian latent class modelling and environmental sampling
- Population
- 2,724 people with suspected mpox at four DR Congo sites, 2024–2026
- Primary outcome
- Proportion of PCR positives consistent with environmental MPXV DNA
- Effect
- 35% (95% CrI 31 to 39) at Ct <40; 18% at Ct <37; 5% at Ct <34
A multi-site observational study in DR Congo took PCR cycle threshold (Ct) values from 2,724 people with suspected mpox, sampled clinic and laboratory surfaces at two sites, and used a Bayesian model to separate true infections from positives consistent with environmental viral DNA. Serology in a subset tested the model.
An estimated 35% of positives with Ct below 40 were likely false positives. Of 35 people classified as environmental positives, 88.6% had no serological evidence of orthopoxvirus infection. Lowering the cut-off to Ct below 37 reduced the false positive rate to 18%, and below 34 to 5%, with little loss of sensitivity.
In a high-transmission clinic, surfaces carry enough viral DNA to contaminate swabs. The lesson travels to any setting testing large numbers for a DNA virus: a high-Ct positive in a patient with an atypical picture needs clinical correlation and possibly a repeat, not automatic isolation and contact tracing.
- Treat a high-Ct (late) mpox positive with caution when the clinical picture does not fit
- Repeat the swab from a fresh lesion when in doubt
- Clean sampling areas between patients
- Ask the laboratory for the Ct value, not only positive or negative
Why it matters
Outbreak counts and individual isolation decisions may both be inflated by contamination, not infection.
Don't overread it
The false positive proportion is a model estimate from high-transmission African clinics and will be much lower where mpox is rare.
The statistics, in plain English
The 35% figure comes from a latent class model, which infers which results are real without a perfect reference test. The credible interval (31% to 39%) is narrow, but it reflects model assumptions as well as data. The serology check was in only 65 people, so it supports the model without proving it.
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 infectious diseases, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free