MRSA in the journal Science - spread, outbreaks and an argument for active surveillance

I have a story tonight at CIDRAP about a paper published this evening in the journal Science. To respect fair use and make sure my colleagues get clicks, I just quote the story here — but then I want to talk about why I think it's such an important study.

   A multi-national team of researchers has applied a new genomic tool to a 50-year-old bacterial foe, using minute mutations to track the spread of drug-resistant staph both across continents and within a single hospital.
   On a global scale, their sleuthing tracked the movement of one clone of methicillin-resistant Staphylococcus aureus (MRSA) back and forth across the planet, pinpointing when individual cases transported infections across national borders to spark new outbreaks. Separately, their method demonstrated that what appeared to be a hospital epidemic of MRSA was not a single outbreak, but rather a mixed event of patient-to-patient transmission of one strain that was accompanied by multiple importations from outside the hospital of similar but unrelated strains. ...
   In a briefing yesterday for the press, the authors emphasized the latter finding, pointing out that the traditional infection control measures usually applied to hospital outbreaks would not curb the spread of infections that were carried in undetected from outside. Their method, they said, provides a proof of concept for using cutting-edge genomics to uncover the precise pathways by which MRSA spreads within hospitals—not only tracing its path from patient to patient, but also identifying the bug in patients whose undetected bacterial carriage could spark outbreaks but have not yet.

 If you'd like more, here's a very good story at Scientific American, one at BBC Health and one by the Associated Press; and Science Daily's version.

Now, the details. This team (which has 15 members from almost as many institutions) secured two collections of MRSA isolates: 43 collected from all over the globe between 1982 and 2003, and 20 from a single hospital in Thailand, collected between October 2006 and November 2007. All of the isolates were ST239, which is a hospital-acquired strain that is particularly prevalent in Asia. They analyzed them using high-throughput sequencing, with a particular analyzer (Illumina) that could produce whole genomes of up to 96 isolates very quickly (an extraordinary advance from the weeks and months it used to take to achieve a single whole genome). Then they compared the genomes, looking for single-letter changes in the genetic code (single-nucleotide polymorphisms, SNPs or "snips," and also insertions and deletions of nucleotides). They used those findings to construct a "family tree" of 239 that tracks very nicely with the known history of MRSA's emergence and initial spread, and that pinpoints rare but intriguing importations of clones from certain areas into other parts of the world.

But it's what they found in the Thai hospital isolates that is especially interesting. (Most of this is not explicit in the paper, but was related in the press briefing that Science conducted on Wednesday). The differences that can be seen in the whole-genome analysis can't be discerned by earlier identification methods, so the isolates collected at the hospital appeared to be the same. However, they weren't the same. Some of them were very closely related, and formed what seems to have been a chain of person-to-person transmission — a true hospital-acquired outbreak. But others of them were not so closely related, either to the outbreak or to each other. What they were, instead, were individual importations into the hospital of a hospital strain that had been acquired outside the hospital, and were carried in by staff, patients, visitors.

You can see where this is going, right? If all the cases in the hospital had represented patient to patient transmission within a known outbreak, excellent infection control might have corralled them. But some of them were not part of that outbreak, so infection control measures aimed at that outbreak would not have kept those other cases from spreading. What would have stopped them from spreading, as the authors pointed out, is detecting them at some other point in their entry into the hospital:

..."That implies you have to have a different perspective on where you are going to apply your infection-control procedures and strategies," co-author Dr. Sharon Peacock of the University of Cambridge said during the briefing.

What that sounds like — and the authors acknowledged as much — is an argument for active detection and isolation/active surveillance and testing/search and destroy, the process of screening some percentage of patients coming into a hospital for MRSA carriage so that the bug can be detected and dealt with long before its presence triggers an outbreak. It is probably not a coincidence that the majority of the authors (including Peacock) are British, and search and destroy has recently become widely accepted in the UK; in fact, the National Health Service recently made it mandatory.

But search and destroy remains remarkably controversial here in the US, despite strong proof of concept demonstrations in healthcare institutions such as Evanston-Northwestern Healthcare, and adoption throughout the VA system. I'll be interested to see whether this paper makes a dent in the overall resistance to search and destroy, and if not, to hear why not.

The cite is: Harris SR, Feil EJ, Holden MTG, et al. Evolution of MRSA during hospital transmission and intercontinental spread. Science 2010 Jan 22;327(5964):469-74