Whole-Genome Sequencing Reveals the Contribution of Long-Term Carriers in Staphylococcus aureus Outbreak Investigation

Abstract

Whole-genome sequencing (WGS) makes it possible to determine the relatedness of bacterial isolates at a high resolution, thereby helping to characterize outbreaks. However, for Staphylococcus aureus, the accumulation of within-host diversity during carriage might limit the interpretation of sequencing data. In this study, we hypothesized the converse, namely, that within-host diversity can in fact be exploited to reveal the involvement of long-term carriers (LTCs) in outbreaks. We analyzed WGS data from 20 historical outbreaks and applied phylogenetic methods to assess genetic relatedness and to estimate the time to most recent common ancestor (TMRCA). The findings were compared with the routine investigation results and epidemiological evidence. Outbreaks with epidemiological evidence for an LTC source had a mean estimated TMRCA (adjusted for outbreak duration) of 243 days (95% highest posterior density interval [HPD], 143 to 343 days) compared with 55 days (95% HPD, 28 to 81 days) for outbreaks lacking epidemiological evidence for an LTC (P = 0.004). A threshold of 156 days predicted LTC involvement with a sensitivity of 0.875 and a specificity of 1. We also found 6/20 outbreaks included isolates with differing antimicrobial susceptibility profiles; however, these had only modestly increased pairwise diversity (mean 17.5 single nucleotide variants [SNVs] [95% confidence interval {CI}, 17.3 to 17.8]) compared with isolates with identical antibiograms (12.7 SNVs [95% CI, 12.5 to 12.8]) (P < 0.0001). Additionally, for 2 outbreaks, WGS identified 1 or more isolates that were genetically distinct despite having the outbreak pulsed-field gel electrophoresis (PFGE) pulsotype. The duration-adjusted TMRCA allowed the involvement of LTCs in outbreaks to be identified and could be used to decide whether screening for long-term carriage (e.g., in health care workers) is warranted. Requiring identical antibiograms to trigger investigation could miss important contributors to outbreaks.

Keywords: MRSA; Staphylococcus aureus; outbreaks; whole-genome sequencing.

FIG 1

All pairwise differences between early (<2 months since acquisition) and late (>12 months since acquisition) nasal swab samples from 7 patients with previously negative nasal swabs. Included for comparison are samples from patients positive at entry to the study (time of acquisition unknown).

FIG 2

Duration-adjusted TMRCA for outbreaks with (i) no evidence of a long-term carrier (direct contacts between all cases); (ii) likely LTC (indirect ward contacts or preoutbreak LTC); or (iii) LTC unclear/possible (evidence of a postoutbreak LTC).

FIG 3

Pairwise SNV differences for all pairs within an outbreak, where isolates had differing antibiograms (a) or differing PFGE pulsotypes (b).