Effect of Colonization with Methicillin-Resistant Staphylococcus aureus on Subsequent Infection

Sir—We are indebted to Davis et al. [1] for their valuable study, which strengthens the role of active surveillance. The followup period of 12 months is laudable, because infections detected after a patient’s discharge from the hospital could also be observed. One important and very alarming finding in the study [1] is that 2.7%–4.4% of patients in the studied population were infected with methicillin-resistant Staphylococcus aureus (MRSA) during their hospitalization, regardless of their colonization status. Also alarming is the higher risk of MRSA infection for noncolonized patients, compared with colonized patients, in the intensive care units of surgical and trauma departments. Only one-third of MRSA infections occurred in patients who were previously colonized with MRSA. Two-thirds of MRSA infections are apparently acquired via transmission. This is a strong argument for the roles of active surveillance, contact isolation precautions, and hand hygiene (the utility of the latter has recently and repeatedly been questioned). A systematic review has shown moderate evidence for a preventive effect of isolation precautions [2], but we postulate that concerted use of these precautions could have prevented, at least partially, the morbidity and mortality associated with 19 MRSA infections reported in Davis et al. [1]. We would like to obtain more information on the use of barrier precautions and hand hygiene in the institution discussed in Davis et al. [1]. There are 4 additional comments that we wish to make. First, the rate of MRSA prevalence is generally defined in the literature as the proportion of all S. aureus isolates that are methicillin resistant. Thus, the actual rate of MRSA in the study population would be 15.9% of isolates (26 MRSA isolates out of 163 S. aureus isolates), rather than the 3.4% of isolates reported by Davis et al. [1]. Second, the question of whether MRSA is more virulent than methicillin-susceptible S. aureus (MSSA) cannot be answered by comparing the incidence rates of MRSA infection in patients who are colonized with either MSSA or MRSA. Ideally, the rate of MSSA infection in patients who are colonized with MSSA would have been compared with the rate of MRSA infection in those who are colonized with MRSA. The latter figure is reported as 19%, but how many patients with MSSA colonization acquire MSSA infection? Wertheim et al. [3] have recently reported a relative risk of 3.0 for MSSA bacteremia in MSSA carriers. A case-case-control study design, as shown by Kaye et al. [4], would answer this question, because risk factors for infection might be different in the 3 groups (MSSA, MRSA, or no colonization). Third, the use of susceptibility patterns to determine whether MRSA strains are nosocomial or community-acquired strains seems somewhat outdated. Modern molecular techniques for making this determination are available, including PFGE for strain typing and epidemiological analysis [5] and PCR for determination of the mecA genetic element and virulence factors associated with community-acquired strains [6]. It is not clear whether the infections in noncolonized patients originate from patients with MRSA colonization or from another source. Finally, Davis et al. [1] used only swabs of the nares to determine colonization status. Adding cultures of the throat, as is common in epidemiological studies, would have increased the sensitivity of testing [7] and thus changed the statistical significance of the findings. In conclusion, the findings of Davis et al. [1] are plausible and valuable from a pathogenetic and epidemiologic point of view. Colonization with MRSA clearly leads to infection, but not only in patients who are colonized. Reduction of transmission is the key, and intervention studies are needed. As stated by Cooper et al. [2] in their systematic review evaluating the evidence for barrier precautions: “lack of evidence ... should not be mistaken for evidence of lack of effect” [2, p. 538].