Control of endemic MRSA—what is the evidence? A personal view

Abstract

Although there is extensive literature on the control of MRSA, when that concerning epidemics is excluded, only a limited amount remains regarding the control of endemic MRSA. Several guidelines have been recently published recommending stringent control measures, which are often suggested based on their success in controlling MRSA outbreaks in hospitals with few MRSA or in containing MRSA cases introduced into a hospital with no MRSA. In these settings, multiple measures are usually introduced with apparently successful results. However, results may not be generalizable to other settings and we do not know the minimum effective measures required for MRSA containment. This paper aims critically to review the literature to determine whether evidence exists for the value of the infection control measures that are widely recommended in the endemic setting. Much of this literature is based on observational studies, with few randomized, controlled trials having been conducted. More well-designed studies are required before many of the principles on which we build infection control programmes can be regarded as evidence based.

Introduction

Methicillin resistance in Staphylococcus aureus (MRSA) was first reported in 1961.¹ Since then, MRSA has gone from being regarded as ‘not only rare but of doubtful clinical significance’² to a major nosocomial pathogen throughout the world.3., 4., 5., 6. Initial outbreaks were managed with a ‘seek and destroy’ attitude; later it was accepted that eradication of MRSA was almost impossible.⁷ Some have even suggested that endemic MRSA ‘may not need any control’⁸ and that ‘trying to control MRSA causes more problems than it solves’,⁹ but most believe that MRSA containment is warranted.¹⁰ Perhaps the most pressing reason for control is the emergence of vancomycin-intermediate and -resistant S. aureus (VISA and VRSA).11., 12., 13. A recent case–control study has demonstrated that previous vancomycin use and MRSA infection are independent risk factors for infection with S. aureus with reduced susceptibility to vancomycin.¹⁴ Because vancomycin usage is largely driven by rates of MRSA and higher rates of MRSA contribute to the emergence of VISA, containment is desirable on this basis alone.

Because of the nature of outbreaks and the difficulty randomizing many infection control interventions, much of the published material on MRSA control consists of retrospective observational studies and opinion, with a few prospective studies and fewer well-designed interventional studies. Importantly, one cannot assume a fall in MRSA infection rates after an intervention is a direct result of that intervention, as rates of endemic MRSA tend to rise and fall with time.¹⁵ Interpretation of the literature is hampered by differing methods of data reporting, absence of denominators, lack of statistical analysis and power calculations, and it is often difficult to draw valid conclusions, let alone compare studies.

Recently published guidelines for MRSA control¹⁶ have highlighted a rarely addressed issue: can control measures in one setting be generalized to other settings? This question applies to both epidemic and endemic MRSA and also specific settings, such as intensive care, acute and non-acute wards, where MRSA may have widely variable transmission dynamics. Many studies cited as models for control of endemic MRSA have taken place during an outbreak or when cases of MRSA have been introduced into wards, hospitals or geographic areas with few or no MRSA, or in hospitals with an increase in low baseline levels of MRSA. In these settings, various measures have been introduced simultaneously, often with apparently successful results. However, in the endemic context, control measures must be efficacious, ongoing and sustainable. We contend that the ability to respond to, and the measures required to control, MRSA in these settings may not necessarily be the same required to control MRSA in institutions where MRSA in highly endemic.

Whilst accepting that many of the recommended measures are likely to be successful, we currently lack an understanding of the minimum effective measures that are necessary and feasible to control MRSA in the endemic setting. This is important, as we believe that the ability of an institution to respond to MRSA may depend upon the context of the problem. For example, the ability to respond to a sudden increase in MRSA cases in an institution with a low level of MRSA or no MRSA with multiple simultaneous control measures that are expensive and resource-intensive may be different from the ability of an institution with high endemic MRSA levels to maintain these measures on an ongoing and widespread basis. Thus, if we knew that one or two measures were crucial, then emphasis could be placed on these measures for the long-term control of endemic MRSA. For example, if we knew that hand hygiene was the most important factor, resources could be directed towards this rather than trying to accommodate all recommendations, which may be unsustainable in the long-term. Because of the likely contribution of multiple factors to the epidemiology and transmission of MRSA, there is unlikely to be a single solution for all institutions, but knowledge of the relative importance of the numerous recommendations would be helpful to prioritize resources. Because most studies report institution of multiple interventions, regardless of the setting, it is impossible to determine what the most important components were, and therefore, it is impossible to determine whether control of endemic MRSA requires the same measures as control of epidemic MRSA.

The aim of this review is critically to analyse the literature on MRSA control in the endemic setting. A major obstacle to achieving this is the difficulty in excluding or including studies based on the authors' determination of whether the setting was epidemic or endemic, as this is not usually defined in the publication. A prolonged epidemic in one study may be considered endemic in another.¹⁷ Wenzel et al.¹⁸ have defined an increased case rate in several different ways, but comment that the exact numbers required to define an outbreak in an individual institution depends on the type of patient and the baseline MRSA levels. Thus we have not excluded studies based on these labels (endemic or epidemic), but rather included those adding to the overall evidence base. Our review is based on a Medline search using the terms MRSA and methicillin-resistant Staphylococcus aureus, with careful back searching of references. We have not tried to include all references on MRSA control because the numbers are too great, but have tried to include illustrative examples with evidence for and against certain measures.