Surveillance and management of all types of Staphylococcus aureus bacteraemia

In the United Kingdom reporting of bacteraemia due to methicillin resistant Staphylococcus aureus (MRSA) infections is mandatory, and reduction in bacteraemia rates is a performance target for NHS trusts. Rates of S aureus bacteraemia remain high around the world, so we need forms of surveillance that will allow better understanding of its causes.

In this week's BMJ Wyllie and colleagues describe the use of linked data in Oxfordshire hospitals to investigate secular trends in bacteraemia caused by S aureus.¹ Using anonymised data on hospital admissions of patients and linking them to information on isolates of S aureus, Wyllie and colleagues found that about a third of patients with S aureus bacteraemia died within 30 days. The risk of death was similar for methicillin sensitive and methicillin resistant S aureus infections.

Between 1997 and 2003, rates of MRSA in these Oxfordshire hospitals increased while rates for methicillin sensitive S aureus (MSSA) strains remained constant. In other words, methicillin resistant strains did not displace methicillin susceptible strains: indeed they added considerably to the burden of disease. This paper serves as a reminder that health services must concentrate efforts on preventing all kinds of S aureus bacteraemia, to appreciate the importance of both methicillin resistant and methicillin sensitive strains, and to look critically at the successes and failures of control measures. Furthermore, these findings will reflect the experience of many readers and pose important questions.

Is the increase in rates of bacteraemia related to novel strains of bacteria, to changes in the population of patients, to changes in healthcare practices, or to a combination of factors? And how do these trends relate to the work of infection control teams? Many infection control teams share a common experience: they recognise cases of MRSA; initiate general “search and destroy” procedures such as screening and isolation of patients; then after a period of apparent success they abandon these measures as rates of colonisation increase and services are reconfigured to concentrate on high risk patients.

A mathematical model published two years ago describes how such loss of infection control can occur by stealth: measures for screening and isolation may seem effective for years but, as increasing numbers of colonised patients are discharged and readmitted, infection rates reach a threshold where suddenly resources become overwhelmed.² Loss of control at one hospital has knock-on effects at units that share the same pool of colonised patients. Such is the experience in the UK.

Around a third of humans are colonised with S aureus. Conservative estimates of the number of MRSA carriers worldwide range from 2 million to 53 million, and this pool is growing.³ The Netherlands is one of the few countries where this rising tide has been held back. A model developed using Dutch data suggests that one factor necessary for control is attempted eradication of carriage on discharge from hospital.⁴ Optimistically, this Dutch model suggests that, even when MRSA becomes endemic, it may be possible to reverse the situation by a coordinated reinstatement of search and destroy measures (including eradication on discharge). To do this properly would require a huge investment in facilities, however, and might take a decade or so to bear fruit.

For practical purposes we may be already past the point of no return. Given that the patients studied by Wyllie and colleagues were general medical and surgical patients and were not selected from high risk groups, it may be more pragmatic to concentrate on measures that prevent all forms of S aureus bacteraemia (such as better management of vascular devices) and to optimise treatment of bacteraemia.⁵,⁶

For example, some doubt remains about the optimal duration of antibiotic treatment for S aureus bacteraemia and carefully planned multicentre prospective comparative trials in selected patient groups are needed to evaluate antibiotics, including several recently licensed agents, for the treatment of MRSA bacteraemia. For the longer term we need an objective, evidence based debate about the desirability and feasibility of controlling transmission and colonisation.

Eradication of MRSA alone will not solve the problem of invasive S aureus infection, not least because strains of S aureus that are sensitive to methicillin still account for many infections. S aureus is a genetically diverse species,⁷ and the acronym MRSA includes a bundle of successful clones which have acquired the mec gene that confers resistance.⁸ Sometimes one cannot see the species for the gene. Measures that focus on detecting carriage draw attention away from the real problem of invasive disease and shake the foundation of reasoned intellectual debate on staphylococcal infection.

As Wyllie and colleagues suggest, collecting patient centred data over long periods at representative centres would allow more detailed surveillance and could inform prospective intervention studies on the prevention and treatment of bacteraemia.¹ Along with greater understanding of the evolutionary biology of these strains of bacteria,⁷,⁸ better management of community acquired MRSA, and more rational use of antibiotics (antibiotic stewardship), such surveillance could greatly improve the management of invasive staphylococcal infection and save lives.