Determination of cgMLST Thresholds for MRSA, MDR E. coli, and VRE by correlating molecular typing data with epidemiological information

Question

In outbreak investigations, two primary Whole Genome Sequencing (WGS)-based methods, namely cgMLST and SNP, quantify typing differences through differing alleles or nucleotides, respectively. Currently, there is no universally accepted threshold for these measures to assess the likelihood of transmission. Thresholds in the literature are frequently determined through retrospective analysis of outbreaks within their epidemiological context. This retrospective approach results in the adoption of diverse thresholds, undermining the potential standardization achievable with the cgMLST method. Our aim is to establish universally applicable threshold values for MRSA, MDR E. coli, and VRE through a novel correlation method.

Methods

We conducted routine WGS-based cgMLST typing on inpatient-derived multidrug-resistant organisms (MDROs). Allelic profiles of eligible MDROs were exported, and cgMLST distances were computed. For each observed cgMLST distance, we calculated patient pairs with a successful identification of an epidemiological link. Considering different levels of spatial and temporal interactions, various epidemiological cut points were examined. The threshold was defined as the cgMLST distance beyond which the percentage of epidemiological contacts found no longer changed significantly.

Results

Threshold values were determined as 4 for MRSA and 3 for both E. coli and VRE. A notable finding was that even with identically typed isolates, no epidemiological link could be identified in half of the cases. For VRE, the value was even below 40%.

Discussion

We present our correlation method as an innovative approach for establishing universally applicable threshold values in molecular typing. However, it's important to note that the thresholds presented were derived exclusively from our dataset and may not hold general validity. We encourage other university hospitals to employ this method with their datasets to validate and confirm the findings.