John Bartlett's Game Changers in Infectious Disease: 2011

John G. Bartlett, MD


October 26, 2011

In This Article

Tackling MRSA

In Veterans Affairs (VA) hospitals, Jain and colleagues[15] studied morbidity and mortality associated with methicillin-resistant Staphylococcus aureus (MRSA) infections. In 2007, the CDC reported that the burden of healthcare-associated infections involving MRSA in the United States was 94,360 per year, with a yearly mortality of 18,650. This finding contributed to increasing concern about healthcare-associated infections.

The goal of the VA study was to implement a "MRSA bundle" that included universal MRSA surveillance, contact precautions, hand hygiene, and what was described as a "culture change" based on the principle of "positive deviance." Infection control and prevention would become the responsibility of everyone involved in the care of patients and not just an infection control team. The study was sequential from 2007-2010 and data were collected for 1,934,598 patients. The sequential design included a control period from 2005 through mid-2007 followed by implementation of the bundle.

At more than 2 years of follow-up, compared with baseline, the frequency of healthcare-associated MRSA infections declined by 48%, with a 62% reduction in ICU patients and a 45% reduction in non-ICU patients. This impressive record of success was attributed to an intensive effort by the VA system to better control the epidemic of nosocomial infections involving MRSA.

The report by Huskins and coworkers,[16] a similar study addressing the same issue, funded by the National Institutes of Health (NIH), was carried out in 18 ICUs at academic centers. Findings were reported in the same issue of The New England Journal of Medicine.[16] In this study, the ICUs used a cluster randomization protocol for standard contact barrier precautions vs gloves only when taking care of ICU patients whose surveillance cultures of nose or stool indicated vancomycin-resistant enterococcus or MRSA. Results based on intervention in 5434 patients vs 3705 control patients showed no benefit in reducing infections involving either of these pathogens (hazard ratio 1.05; 95% confidence interval 0.8-1.4).

Thus, the VA study and the NIH-funded ICU trial addressed similar issues using surveillance cultures and barrier precautions to control MRSA infections in the hospital, and, although neither study was perfect, they came to opposite conclusions. Criticism of the VA study includes the problem of sequential design because many other factors could change rates of MRSA infections, including the previously noted data for CLABSI.[6] The most serious concern about the NIH-funded ICU trial is that the cultures were processed in Bethesda, Maryland, resulting in a prolonged delay (averaging about 5 days) before the results were known to the investigators.

Why Is This a Game Changer?

The more important conclusion of these 2 studies is that there is no conclusion. We had a similar controversy in 2008 when, using standardized infection control methods, a study in Chicago showed a 45% reduction in MRSA infections,[17] and a study in surgical patients in Finland showed a 10% increase in S aureus infections.[18] Everyone in hospital practice is aware of the enormous concern for MRSA and the repeated emphasis on the ritualistic methods of infection control.

What Does This Mean to the Practitioner?

  • MRSA is an incredibly important pathogen in nosocomial infections.

  • Practitioners need to be aware that these reports[15,16] came to diametrically opposed conclusions. Practice justified by one study needs to be balanced by consideration of the other.

  • Infection control of MRSA and other resistant pathogens is critically important to clinicians, health systems, payers, and patients, as nicely pointed out by Platt[19] in his editorial accompanying these conflicting studies in The New England Journal of Medicine.

  • We need to be cautious about the "bundle" that looks good in terms of scientific methods. Either study, considered on its own, could drive decisions and have a massive impact on practice, with tremendous allocation of resources and no clinical benefit. This emphasizes the need for fastidious attention to study design and, possibly, the need for 2 multicenter trials.