Study population, demographics, and sample selection.

As part of a community study on S. aureus transmission in Northern Manhattan, we implemented a cluster network analysis of ST398 transmission between September 2010 and March 2013. Individuals who had previously tested positive for ST398 colonization or infection were eligible to enroll (28, 42). These index participants were asked to complete a structured questionnaire and provide information on their daily contacts and the nature and frequency of these interactions. In addition, swabs were collected from the anterior nares, throat, and groin from consenting participants (Becton, Dickinson Culturette Systems). A standardized set of environmental items was also sampled with sterile premoistened swabs in all households: doorknobs, television remote control, living room light switch, toy, couch or bed, computer or radio, house phone or index cellular phone, bathroom sink, kitchen, and appliance handle (33).

A snowball enrollment strategy was used to identify contacts of index participants. Named household and outside-household contacts of the index were approached and after providing informed consent underwent the same study procedures. Contacts were further defined by their type of relationship (e.g., family, work, school, sports, or other social life contacts). Recruitment occurred in waves, such that contacts of study participants were asked to participate in the study, as were the contacts of these newly recruited contacts. Follow-up household visits were carried out twice: between 3 and 6 or 6 to 12 months after enrollment, respectively. We enrolled individuals from 15 networks that consisted of 958 named members. Of these, 348 were enrolled and 273 individuals agreed to complete the questionnaire. We compared demographic variables and risk factors for S. aureus infections between individuals colonized with ST398 and those colonized with other S. aureus isolates or not colonized. Data were analyzed using SAS 9.1 software (SAS Institute, Cary, NC).

Culture swabs were incubated overnight at 37°C in high-salt 6.5% broth and plated onto mannitol salt agar (Becton, Dickinson) for 48 h at 35°C. Positive mannitol-fermenting yellow colonies were isolated onto 5% sheep blood agar plates (Becton, Dickinson), and single colonies were selected for further analysis. S. aureus was identified by a coagulase and protein A detection kit (Murex Staphaurex). S. aureus-positive isolates were genotyped by spa sequencing using Ridom-staph software (24, 25). Of 2,590 swabs, we identified 690 S. aureus isolates. Based on spa typing and Spa clustering, 175 isolates belonged to Spa-CCt571, consistent with CC398.

Of these, we selected 153 isolates, including all human-colonizing isolates (n = 96) and a subset of environmental isolates (57/79) for sequencing. We also selected an additional 135 ST398 isolates from previous studies on S. aureus infection and colonization in Northern Manhattan and the Bronx as follows: (i) MSSA infections between 2007 and 2011 (n = 69) (24, 28), (ii) infectious isolates from the Dominican Republic and Martinique (n = 14) (22), (iii) community studies on colonization between 2003 and 2010 (n = 38) (23, 33, 42), and (iv) colonizing isolates from pigs in Minnesota and swine veterinarians in Minnesota, Illinois, and Indiana (n = 14) (26). In total, we obtained 288 novel sequences. Previously published sequences from international studies were also included in phylogenetic analyses (31, 32).