JSTOR

Methods

From January 1999 through December 2003, nasal swab specimens were collected routinely from all patients hospitalized in the 13‐bed general ICU of Hospital São Lucas (Porto Alegre, Brazil), a 600‐bed hospital. The nasal swab specimens were obtained when patients were admitted to the ICU and weekly thereafter. Patients with nasal carriage of MRSA received 2% mupirocin intranasally 3 times daily for 5 days and received a chlorhexidine bath once daily for 3 days. The patients did not routinely receive any other intervention, and decolonization procedures were adopted only in the ICU. Trained ICU nursing staff members applied mupirocin and chlorhexidine baths. Patients were followed up weekly with nasal swab cultures to verify eradication of nasal carriage of MRSA. Eradication was defined as 1 MRSA‐negative nasal swab culture.

Nosocomial infections were defined according to the criteria of the US Centers for Diseases Control and Prevention.⁶ S. aureus was identified by conventional microbiology methods, and resistance to oxacillin was determined by Kirby‐Bauer disk‐diffusion method, according to NCCLS guidelines.⁷ Resistance to mupirocin was assessed during the final year of the study, according the aforementioned methods and guidelines.

The incidence of nosocomial infection due to all S. aureus strains and to MRSA, the percentage of S. aureus–positive nasal swab cultures, and the percentage of MRSA strains were determined each year (1999 through 2003), and the results were compared by use of the χ² test. was considered to be statistically significant.

Results

A total of 2,200 patients admitted to the ICU during 1999‐2003 were screened for nasal carriage of S. aureus (424 patients in 1999, 458 patients in 2000, 445 patients in 2001, 415 patients in 2002, and 458 patients in 2003). Of these patients, 511 (23.2%) were found to have nasal carriage of S. aureus, and 364 (16.5%) were carriers of MRSA (71% of all S. aureus strains). Rates of eradication of nasal carriage of MRSA ranged from 69.7% to 75% of patients (mean, 72.3% of patients).

During this period, there were 139 nosocomial S. aureus infections, and 111 of them (79.8%) were due to MRSA. The most common site of S. aureus infection was the inferior respiratory tract (88 cases [63.3%]), followed by the bloodstream (29 cases [20.9%]), and skin and soft tissue (12 cases [9%]). Other less common infections accounted for 7% of the cases.

The overall cumulative incidence of nosocomial MRSA infection was 35 (8.2%) of 424 cases in 1999, 29 (6.3%) of 458 cases in 2000, 19 (4.3%) of 445 cases in 2001, 15 (3.6%) of 415 cases in 2002, and 13 (2.8%) of 458 cases in 2003. There was a trend toward lower incidences of nosocomial MRSA infection during the study period (Figure), and there was a significant decrease in the incidence of such infections in 2003 ( ).

(69 KB)

Figure.  Incidence of nosocomial infection with methicillin‐resistant Staphylococcus aureus (MRSA) and percentage of S. aureus–positive nasal swab cultures, from 1999 through 2003. The incidence of nosocomial MRSA infection (diamonds) was significantly lower in 2003 than in previous years ( ). The percentages of swab cultures positive for all strains of S. aureus (gray bars) and for MRSA (white bars) decreased significantly during the study period ( and , respectively). The median length of stay in an intensive care unit was 14 days in 1999, 15 days in 2000, 18 days in 2001, 19 days in 2002, and 16 days in 2003.

Open New Window

The incidence of nosocomial infection due to all S. aureus strains also decreased during the study period: 42 (9.9%) of 424 cases in 1999, 41 (9.0%) of 458 cases in 2000, 24 (5.4%) of 445 cases in 2001, 16 (3.9%) of 415 cases in 2002, and 16 (3.4%) of 458 cases in 2003. The cumulative incidence of S. aureus infection was significantly lower in 2002 and 2003 ( ). As shown in the Figure, the percentage of nasal swab cultures positive for MRSA also decreased significantly during those years. Resistance to mupirocin was found in 0.05% of MRSA isolates collected in 2003.

Discussion

Despite the fact that mupirocin nasal ointment has been demonstrated to be effective in eradicating nasal carriage of S. aureus and in decreasing the incidence of S. aureus infection among patients undergoing cardiothoracic surgery and hemodialysis,^4,5 there are no data demonstrating the same benefit in other populations.^8,9 Wertheim et al.⁹ have recently published the first randomized controlled trial involving nonsurgical patients. However, their study showed that treatment of nasal carriers of S. aureus with mupirocin ointment did not decrease the incidence of nosocomial S. aureus infection among treated patients.

Our study presents the findings from 5 years of active follow‐up surveillance for nosocomial MRSA infections in the ICU of a tertiary‐care teaching hospital after the introduction of routine treatment of all nasal carriers of MRSA. The present study differs from previous ones in many aspects. First, our study population included only patients in the ICU, where nosocomial S. aureus infections occur more frequently. Second, we applied mupirocin ointment and chlorhexidine baths only to nasal carriers of MRSA. Third, although previous studies have tried to demonstrate a decrease in the cumulative incidence of nosocomial S. aureus infection in mupirocin‐treated groups, compared with placebo groups, we aimed to investigate whether the application of systematic MRSA decolonization procedures to patients in a specific ICU decreases the cumulative incidence of nosocomial infection due to MRSA. Therefore, we had a distinct end point.

The findings of the present study show a significant decrease in the cumulative incidence of nosocomial MRSA infection among patients in the ICU during the 5‐year follow‐up period. The incidence of nosocomial infection due to all S. aureus strains also decreased, despite the fact that decolonization procedures were applied only to nasal carriers of MRSA; this decrease may have occurred because ∼80% of nosocomial S. aureus infections were caused by MRSA. This result is corroborated by the finding that the incidence of nosocomial methicillin‐susceptible S. aureus infection did not decrease during the study period, although it was significantly lower in 2002 ( ; data not shown). Lower annual incidences of nosocomial infection due to MRSA were accompanied by a significant decrease in the prevalence of nasal carriage, as determined by the percentage of positive nasal swabs. It is possible that lower incidences of nosocomial MRSA infection were due to the lower prevalence of MRSA carriage during the study period and that decolonization procedures led to this lower prevalence of MRSA nasal carriage.

The reduction in the prevalence of nasal carriage of all S. aureus strains observed in our study may be explain by the reduction in the prevalence of MRSA, because 71% of S. aureus strains were MRSA. Moreover, the prevalence of nasal carriage of methicillin‐susceptible S. aureus did not changed during the study period ( ; data not shown). Resistance to mupirocin was not a concern for the treatment of nasal carriers in our population, because such resistance was not prevalent among MRSA isolates in 2003, despite the fact that mupirocin was used for 4 years.

Our study has some limitations. It was not possible to control for many confounding variables. Both the prevalence of nasal carriage and the incidence of nosocomial infection could have been reduced by many other factors not included in a multivariate analysis. However, some of the factors, such as the use of vancomycin and changes in patients' clinical profiles, did not change during those years at our institution (data not shown). Another limitation was that we did not perform a molecular analysis of the strains. Thus, we could not determine the role of patient‐to‐patient transmission of such strains. However, an outbreak of vancomycin‐resistant enterococcus infection was observed in the ICU during the study period. Patients infected with vancomycin‐resistant enterococcus were placed in contact isolation, and it is possible that some MRSA‐infected patients were also colonized by vancomycin‐resistant enterococcus. The effect of this potential bias is probably of little importance, because the number of patients infected with vancomycin‐resistant enterococcus was smaller than the number of patients infected with MRSA.

We were not fully able to conclude, by use of our study methodology, whether decolonization procedures are the only or the main cause of the reduction in the incidence of MRSA infection. However, it is important to point out that treatment with mupirocin nasal ointment and chlorhexidine baths for nasal carriers of MRSA is a possible determinant of the decreasing prevalence of nasal carriage of MRSA, and it is probable that such procedures play a role in the reduction of the incidence of MRSA infection.

In contrast to previous studies, we proposed a distinct end point: the incidence of nosocomial MRSA infection in a specific ICU, in which decolonization procedures were applied to nasal carriers of MRSA. By use of such a design, we demonstrated a significant reduction in the incidence of nosocomial MRSA infection. Considering methodological limitations, we outline the potential role of mupirocin and chlorhexidine decolonization procedures in determining such a decrease. Further studies with similar end points are necessary to reinforce evidence of the effectiveness of these interventions in nonsurgical patients.