Experimental Design

This study was carried out in a large academic ophthalmology setting. A standardized method (described below) was used to sample 12 randomly selected examination rooms and shared imaging rooms in 2 ophthalmology clinic buildings belonging to the same system (1 freestanding outpatient clinic, clinic 1; and 1 hospital-adjoining clinic, clinic 2). To be able to cover as many locations as possible from the clinic, the same surface in 3 different examination rooms was sampled as a pool. For example, all tonometer tips from each set of 3 rooms were pooled into 1 sample. Each group of rooms was denoted group A, group B, etc. ( Table 1 ). Items sampled in each examination room included patient-contact surfaces, such as slit-lamp, headrests and chinrests and tonometer tips; employee-contact surfaces such as computer keyboards and hand-sanitizer dispensers; and general contact surfaces such as doorknobs. Excluding shared imaging areas, 12 examination rooms of 43 (28%) were sampled ( Table 1 ).

Shared imaging equipment and imaging room contact surfaces (doorknobs, keyboards, head/chinrests, A- and B-scan ultrasound probes) in both clinic buildings were also tested. In these areas, a total of 3 surfaces were not always available, so the number of samples included in each pool can be found in Table 1 . The sampled surfaces were identical in all subsequent samplings. Overall, a total of 28 pooled samples from both examination and imaging rooms were collected on each date, representing a total of 76 individual surfaces.

Sampling was performed quarterly for 1 year. For the initial survey, rooms were selected randomly, and the same cadre of rooms and surfaces was screened for each subsequent survey. In each case, sampling was conducted at the end of the workday, after the clinics had closed and before daily cleaning and disinfection procedures were performed by the housekeeping staff.

Sample Collection and Processing

Samples were systematically collected from commonly touched surfaces by either cotton-tipped swabs (for smaller surface areas) that were premoistened to enhance sensitivity or by 10.4 × 8.0 inch (26.5 × 20.3 cm) dry electrostatic cloths (for larger surface areas). The electrostatic nature of the cloth enhances sensitivity of bacterial collection. For doorknob samples, the area immediately surrounding the doorknob was also sampled, up to about 1 foot from the knob and on both sides of the door. Slit-lamp headrests and chinrests were sampled at both forehead points and chin points of contact. Sampling was conducted by trained personnel wearing clean clothing covers and gloves, which were changed between each sample collected. After each sample was collected, swabs were immediately placed in tubes containing sterile trypticase soy broth (TSB, BD BBL; Becton, Dickinson) with 2.5% NaCl, while cloths were placed in sterile empty transportation bags (Nasco Whirl-Pak; Fort Atkinson, WI, USA). All samples were taken immediately to the Diagnostic and Research Laboratory on Infectious Diseases and processed as previously described, with a few modifications. Briefly, a pre-enriched medium (TSB with 2.5% NaCl) was added to each bag containing the dry cloths. Tubes with swabs as well as bags with cloths were incubated for 24 hours at 35°C, then grown on mannitol salt agar plates (BD BBL mannitol salt agar; Becton, Dickinson). Finally, 3 colonies with typical mannitol reactions were plated on 5% sheep blood plates (Remel; blood agar [trypticase soy agar, TSA, with 5% sheep blood]; Thermo Fisher Scientific, Lenexa, KS) to be further characterized. Typical morphologic colonies of S. aureus were confirmed by standard biological testing. The MRSA phenotype was confirmed by growth on oxacillin screen agar plates (OSA, BD BBL Becton-Dickinson) containing 6 μg/mL of oxacillin supplemented with NaCl following the Clinical Laboratory Standards Institute protocols.

Phenotyping and Genotyping

Antimicrobial resistance of all S. aureus isolates was established by the Kirby Bauer method using an array of 15 antibiotics (ampicillin 10 μg, amoxicillin/clavulanate 30 μg, oxacillin 1 μg, cephalothin 30 μg, cefpodoxime 10 μg, ciprofloxacin 2 μg, erythromycin 15 μg, gentamicin 1 μg, amikacin 30 μg, clindamycin 2 μg, tetracycline 30 μg, doxycycline 30 μg, trimethoprim/sulfamethoxazole 25 μg, moxifloxacin 5 μg, and chloramphenicol 30 μg). Resistance to vancomycin was tested using vancomycin screen agar plates (BD BBL Vancomycin Screen Agar; Becton-Dickinson). For quality control purposes, the following strains were included: S. aureus (ATCC 43300); S. aureus (ATCC 29213); S. aureus (ATCC 25923); Enterococcus faecalis (ATCC 23212); Escherichia coli (ATCC 25922); and Pseudomonas aeruginosa (ATCC 27853).

Molecular characterization of S. aureus isolates was performed as previously described. Briefly, pulsed-field gel electrophoresis was performed in a CHEF mapper system (Bio-Rad Laboratories, Nazareth, Belgium) to separated restriction fragments obtained from the digestion of chromosomal DNA with SmaI. Salmonella serotype Branderup strain H9812 was digested with XbaI and used as a molecular size marker. Relatedness between isolates was evaluated by dendrogram analysis of the resulting pulsed-field gel electrophoresis band patterns using commercial software (BioNumerics, v 6.6, Applied Maths, Ghent, Belgium). The dendrogram was generated by using the Dice coefficient and the unweighted pair group method with arithmetic averages, with a 1% band position tolerance. Based on percentage of similarity, band patterns were classified as clones or the same pulsotype (≥98%) and grouped in clusters (≥80%). To establish US types, all S. aureus isolates obtained in this study were compared against a Centers for Disease Control and Prevention (CDC) database containing 100 S. aureus strains with the most typical band patterns for each US type, using ≥80% similarity as the cutoff point. For MRSA isolates, mecA gene confirmation and SCCmec typing was performed as well.