- Design
- pragmatic, prospective, single-centre, non-blinded randomised controlled trial
- Population
- 277 adult intensive care patients with suspected community-acquired pneumonia
- Primary outcome
- vancomycin-free hours alive within 7 days of enrolment
- Effect
- 109.7 versus 105.7 hours, adjusted difference 4 hours (95% CI -9.5 to 18.2), p=0.458; MRSA PCR negative predictive value 98.9%; no difference in 30-day mortality
MRSA nasal PCR has an excellent negative predictive value in pneumonia, and stewardship programmes have promoted it as a way to stop empiric vancomycin early. STOP-Vanc tested whether it actually does that, randomising 277 adult intensive care patients with suspected community-acquired pneumonia to usual care with or without MRSA nares PCR after admission.
The test performed as advertised: negative predictive value 98.9% in the intervention arm. The behaviour did not follow. Vancomycin-free hours alive over seven days were 109.7 in the intervention arm against 105.7 in the control, an adjusted difference of 4 hours (95% CI -9.5 to 18.2, p=0.458). Thirty-day mortality did not differ either.
This is the most useful kind of negative trial, because it separates two things that get conflated. The diagnostic property is not in doubt; what failed was the assumption that giving a clinician a reassuring result leads them to stop an antibiotic. Fear of missing MRSA in a sick patient is not overcome by a nasal swab. The implication for practice is that a de-escalation intervention needs the de-escalation built into it — a stewardship review, a protocolised stop, a prompt — rather than a result dropped into the record. Single-centre and unblinded, with 277 patients, it cannot exclude a small effect; it does show that the test alone is not the intervention.
- A negative MRSA nares PCR supports stopping vancomycin — but someone has to actually stop it
- Pair the swab with a stewardship prompt or a protocolised review, or expect no change in prescribing
- The 98.9% negative predictive value applies to this population and prevalence, not universally
- Do not use a positive nasal swab to justify continuing vancomycin: colonisation is not pneumonia
- Where MRSA prevalence is lower, as in much of Indian community-acquired pneumonia, empiric vancomycin is harder to justify in the first place
The statistics, in plain English
An adjusted difference of 4 vancomycin-free hours with an interval from -9.5 to 18.2 includes zero comfortably, so no effect was detected; with 277 patients the trial could not have detected a difference of a few hours anyway. A negative predictive value of 98.9% depends on how common MRSA is in the population tested — in a setting with higher prevalence the same test would miss proportionally more. The outcome was modelled using a longitudinal state transition approach adjusted for baseline covariates, which handles death and drug use together rather than counting only survivors.
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 emergency & critical care, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free