Impact of strain typing methods on assessment of relationship between paired nares and wound isolates of methicillin-resistant Staphylococcus aureus

Abstract

The anterior nares are the site of choice for the Veterans Administration methicillin-resistant Staphylococcus aureus (MRSA) surveillance program; however, a correlation between nares colonization and concomitant wound infections has not been well established. The purpose of this study was 3-fold: to determine the relatedness of MRSA isolates from 40 paired wound and nares specimens by four different strain typing methods, to determine concordance of typing methods, and to establish a baseline of MRSA types at this medical center. Isolates were typed by repetitive PCR (rep-PCR) (DiversiLab System; DL) and SpectraCell Raman analysis (SCRA) (commercially available methods that can be performed within a clinical lab), pulsed-field gel electrophoresis (PFGE), and an antibiotic susceptibility profile (AB). Whole-genome optical mapping (WGM) (OpGen, Inc.) was performed on selected isolates. All methods agreed that 26 pairs were indistinguishable and four pairs were different. Discrepant results were as follows: 4 where only SCRA was discordant, 3 where only AB was discordant, 2 where both DL and AB were discordant, and 1 where both DL and SCRA were discordant. All WGM agreed with PFGE. After discrepancy resolution, 80% of the pairs were indistinguishable and 20% were different. A total of 56% of nares results were nonpredictive if negative nares and positive wound cultures are included. Methods agreed 85 to 93% of the time; however, congruence of isolates to a clade was lower. Baseline analysis of types showed that 15 pairs were unique to single patients (30 strains, 38%; 47% of the matching pairs). Twenty-five strains (30%) represented a single clade identical by PFGE, SCRA, and DL, decreasing specificity. Typing method and institutional type frequency are important in assessing MRSA strain relatedness.

Fig 1

Visual comparison of PFGE gels from selected pairs. Analysis of selected pairs with the discrepant results. The final assessment was that the strains were different for patients 10 and 18 and the same for patients 1, 14, and 29. Strains from patient 27 and 15 were not discrepant and are included as controls to illustrate similarity of strains (patient 27) and differences (patient 15); all testing methods agreed that both strains from patient 27 were the same and both strains from patient 15 were different.

Fig 2

Whole-genome map comparison of the naris and wound strains from patient 10. Comparison of the high-resolution, ordered restriction maps of two isolates from the same patient demonstrates significant genomic differences, indicated by white spaces and noted as the number of kilobases.

Fig 3

Greater discrimination by PFGE. Each isolate was designated group G by DL. However, the strains of patients 17 (strains 35 and 44), 14 (strains 41 and 47), 37 (strains 57 and 60), 29 (strains 64 and 65), and 27 (strains 55 and 76) have distinct patterns by PFGE when evaluated for a single band difference. Other methods of evaluating gel patterns can allow up to 3 band differences and still consider the strains to be in the same group.