The hospital antibiogram is the wrong number for bone and joint infection

Design

retrospective cohort study using a national administrative database linked to hospital antibiograms

Population

15841 hospitalisations for acute haematogenous osteomyelitis or acute bacterial arthritis at 46 US children's hospitals, September 2015 to December 2024

Primary outcome

frequency and appropriateness of empiric MRSA antimicrobial therapy

Effect

17% of pathogen-coded cases were MRSA; 58% of children without identified MRSA received empiric MRSA therapy; disease-specific MRSA rate half the hospital-wide rate (ratio 0.5, range 0.2 to 1.17)

Guidelines advise empiric MRSA cover for acute haematogenous osteomyelitis and acute bacterial arthritis where the hospital's overall MRSA rate exceeds 10% to 20%. This study asked whether that rule is producing sensible prescribing. It used the Pediatric Health Information System database to identify 15841 hospitalisations for these two infections at 46 US children's hospitals from September 2015 to December 2024, and collected each hospital's own antibiogram.

Of the 48% of children who had a pathogen-specific discharge code, 17% had MRSA. Among children in whom MRSA was not identified, 58% had still received empiric MRSA therapy. Most participating hospitals had a hospital-wide MRSA rate above 20% - which is what triggers the guideline - but the MRSA rate specifically among osteoarticular infections was consistently lower, at about half the hospital-wide figure (ratio 0.5, range 0.2 to 1.17).

So the threshold is being applied to the wrong denominator. A hospital-wide rate is dominated by skin and soft tissue and by device-associated infection; bone and joint infection in children behaves differently, and using the aggregate number ratchets up empiric vancomycin or clindamycin for a majority of children who do not need it. The fix the authors propose is unglamorous and achievable: build a syndrome-specific antibiogram for osteoarticular infection, and set empiric policy from that. Most units have the isolates to do it.

• Ask your microbiology laboratory for an osteoarticular-specific antibiogram rather than using the hospital-wide MRSA rate.
• Take a culture before the first dose wherever the child's condition allows - the 52% with no pathogen code is the deeper problem here.
• Set a fixed review point at 48 to 72 hours to narrow therapy once cultures are back.
• Remember that clindamycin resistance patterns are local; an empiric choice that works in one city may not in the next.
• In India, where community-associated MRSA prevalence varies widely between centres, the argument for a local syndrome-specific antibiogram is stronger, not weaker.

The statistics, in plain English

The headline weakness is that only 48% of these hospitalisations carried a pathogen-specific discharge code, so the 17% MRSA figure describes the children in whom an organism was identified and coded, not all of them. If MRSA were systematically more likely to be cultured and coded - plausible, since it is more likely to be severe - the true proportion would be lower still, which strengthens rather than weakens the argument. The ratio of 0.5 comparing disease-specific to hospital-wide MRSA rates has a range across hospitals of 0.2 to 1.17, so at least one hospital saw no advantage from a specific antibiogram; this is a reason to compute your own rather than to adopt the average.

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.