From October to November 2002, a cluster of skin and soft tissue infections due to MRSA involving pediatric and maternity patients occurred at a New York City hospital. The hospital has a labor and delivery unit and 2 units that house both healthy newborns and maternity patients. Healthcare workers on these units typically care for patients on all the units. After the outbreak was recognized, the following interventions were implemented: 1) nursing and medical personnel from the involved areas were informed of the outbreak and potential modes of transmission of staphylococci, 2) contact precautions were emphasized for all patients with suspected or proven skin infections, 3) alcohol-based hand sanitizers were placed in involved areas, 4) healthcare workers from involved units were screened for nasal MRSA colonization, and 5) environmental surfaces (including cord clamps, antitheft transponders, and temperature sensors of baby warmers) were tested for MRSA contamination. Healthcare workers colonized with MRSA were treated with intranasal mupirocin and furloughed until repeat cultures were negative. To identify any other potential case-patients, letters concerning the outbreak were sent to pediatricians who cared for newborns discharged from the affected units during the outbreak period. Cases were defined as MRSA infections in patients who stayed on the labor and delivery, nursery, or maternity units at any time from October 2002 to December 2002. The medical records of the patients were reviewed for information regarding prior healthcare exposures, receipt of antimicrobial agents, underlying medical conditions, treatment, and clinical outcome.

Cultures related to the outbreak were grown on tryptic soy agar plates supplemented with 3% sheep blood; colonies consistent with S. aureus were identified according to standard techniques. All isolates underwent susceptibility testing with the Etest method (AB Biodisk, Solna, Sweden). Ribotyping was performed with the Riboprinter Microbial Characterization System (Qualicon, Wilmington, DE, USA), as previously noted (13). In addition, isolates of MRSA collected during the outbreak were fingerprinted by pulsed-field gel electrophoresis (PFGE), as previously described (13). PFGE results were interpreted according to known criteria (14).

SCCmec typing was performed by using multiplex polymerase chain reaction (PCR), under conditions described by Oliveira et al. (15). Primers to detect the mecA gene were included as an internal positive control (15). Multilocus sequence typing (MLST) was performed on selected isolates as described by Enright et al. (16). Bidirectional DNA sequencing of 7 amplified housekeeping genes was performed with an automated fluorescent dye-terminator sequencing system (Applied Biosystems, Foster City, CA, USA). Allelic types were assigned by using the MLST database (available from www.mlst.net).

The presence of genetic sequences encoding several staphylococcal toxins was also investigated for the outbreak isolates. Based on the previously reported distribution of enterotoxins in CA-MRSA from the United States (7), the following toxins were selected for investigation: staphylococcal enterotoxin A (SEA), B (SEB), C (SEC), H (SEH), and K (SEK). In addition, strains were screened for PVL and toxic shock syndrome toxin-1 (TSST-1). Previously published primers and conditions were used to detect sequences encoding for SEA, SEB, SEC, SEH, PVL, and TSST-1 (17–19). Genes encoding for SEK were detected with the following primers: SEK forward: 5´-TGGATCAATGGAAATCACAAAA-3´ and reverse: 5´-TTTGGTAGCCCATCATCTCC-3´ (predicted product size 287 bp). The specificity of amplification was verified by bidirectional sequencing of the product.

The identification of MW2 in the outbreak of the neonatal-maternity unit prompted a retrospective investigation to determine the regional prevalence of MRSA resembling this strain. In 1999, 2001, and 2003, surveillance studies were performed in Brooklyn, New York. Each surveillance study involved collecting all single-patient isolates of S. aureus from clinical microbiology laboratories during a 3-month interval. Each study included 11–15 hospitals. Susceptibility testing was performed in the central research laboratory by using the agar dilution method according to NCCLS methodology (20). All MRSA isolates were then screened for a phenotype of susceptibility to clindamycin and ciprofloxacin (typical for MW2). Isolates possessing this susceptibility pattern underwent ribotyping and SCCmec typing. The study was approved by the Institutional Review Board at the State University of New York (SUNY) Health Science Center and Maimonides Medical Center.

From October 18 to November 28, 2002, a total of 8 patients with skin and soft tissue infections due to MRSA were identified. During this period, 3.5 cases of MRSA infection occurred each month in the nursery and maternity units. In contrast, no MRSA infections had been reported from the involved units in the 10 months before the outbreak. Two patients were mothers, and 6 were neonates; in no instance were both the mother and her child infected. All had been hospitalized on an involved unit at some point from October 16 to November 6, 2002. Review of medical records showed that none of the patients had prior hospital exposure, underlying chronic medical conditions, or recent antibiotic therapy.

Clinical manifestations of the infections are included in Table 1. None of the patients had evidence of infection upon admission to the hospital. The timing of hospitalization and onset of clinical symptoms are shown in Figure 1. Patients stayed on the unit for an average of 5 days (range 2–12 days). Clinical infection developed in 4 of the newborns and 1 mother while in the hospital. Symptoms developed in 2 newborns and 1 mother 2, 10, and 24 days, respectively, after discharge. β-Lactam antimicrobial agents were initially administered for 6 patients. Definitive therapy generally consisted of topical or systemic antimicrobial agents active against MRSA; 1 patient required surgical drainage. All patients had clinical resolution of infection.

Two additional suspected cases were reported by pediatricians to the Infection Control Department. The first was in an infant, born in November 2002, who was seen as an outpatient for pustulosis; however, the site was not cultured. The second case involved another infant, also born in November 2002, who was readmitted to the hospital 4 days later for treatment of omphalitis. Multiple cultures yielded no growth. No additional cases were reported from December10, 2002, to December 31, 2003.

Susceptibility testing showed that all 8 isolates were susceptible to clindamycin, ciprofloxacin, trimethoprim-sulfamethoxazole, rifampin, doxycycline, linezolid, and vancomycin. Of the 8 clinical isolates, 7 (isolates P1-P7) belonged to 1 ribotype that was identical to the prototypical MW2 strain (Figure 2). PFGE confirmed that the 7 isolates were identical and closely related to MW2 (Figure 2). All 7 contained SCCmec type IV. Since the 7 isolates appeared identical, MLST was performed on one of the isolates and showed sequence type 1. The PFGE and MLST pattern are the same as CA-MRSA clone USA 400, which also includes MW2 (21). Among these 7 isolates, all contained SEK and PVL, 6 contained SEC and SEH, and 5 contained SEA. None was found to have genes encoding SEB or TSST-1. The eighth clinical isolate, from a catheter-site infection, was distinct from the outbreak strain by ribotyping and PFGE (Figure 2). For this isolate, SCCmec was nontypable, and MLST typing confirmed a distinct allelic profile. None of the genes encoding toxins was detected.

A total of 189 healthcare workers worked on the involved units during the outbreak period. Screening cultures of the anterior nares were performed in 176 of the workers in November 2002. Three of the cultures were positive for MRSA, including 2 from the nursing staff and 1 from a pediatrician. The 3 MRSA strains possessed a susceptibility pattern typical for the multidrug-resistant hospital strains, with resistance to clindamycin and ciprofloxacin. They belonged to ribotypes distinct from the outbreak clone, and PFGE confirmed these isolates were unrelated to MW2 (Figure 2). For the 3 isolates, SCCmec was nontypable with the multiplex PCR method. None of the 27 environmental samples collected in November 2002 yielded positive cultures for MRSA.

A total of 4,345 isolates of S. aureus were collected in the 3 surveillance studies conducted in 1999, 2001, and 2003; susceptibility data for these isolates are given in Table 2. A total of 1,913 (44%) isolates were methicillin-resistant. Of the 1,913 MRSA isolates, 118 (6%) possessed the screened phenotype (susceptible to both clindamycin and fluoroquinolones). Among the 118 isolates, 40 different ribotypes were identified. A total of 11 isolates possessed the same ribotype pattern as the outbreak clone, MW2. Of the 11 isolates, 4 were known to come from children. One HIV-infected adult died of overwhelming sepsis within 24 hours of hospitalization. Sources of the cultures included skin and soft-tissue in 7 patients, blood/sterile body fluid in 3 patients, and the genital tract in 1 patient. Nine of the 11 isolates had SCCmecIV. The number of isolates resembling MW2 remained relatively constant during the 3 surveillances (4 in 1999, 3 in 2001, and 4 in 2003).