An assessment of Listeria typing data reported by national public health reference laboratories has shown positive results.
The findings come from the fifth external quality assessment (EQA-5) scheme for typing of Listeria monocytogenes.
The test was organized for laboratories providing data to the Food- and Waterborne Diseases and Zoonoses Network (FWD-Net) managed by the European Centre for Disease Control and Prevention (ECDC).
This program is responsible for EU-wide surveillance of listeriosis and detecting and investigating foodborne outbreaks. Objectives are to assess the quality and comparability of the typing data reported by participating public health national reference labs.
A change was made from including quality assessment of pulsed-field gel electrophoresis (PFGE) in EQA-4 to including a molecular typing-based cluster analysis using PFGE and/or whole genome sequencing (WGS) derived data during the fifth external quality assessment project.
Two separate sets of 11 test isolates were selected for serotyping and molecular typing-based cluster analysis.
Twenty-two labs signed up and 20 completed the exercise, representing a decrease in participation of 13 percent from the previous assessment. The majority, 65 percent, of labs completed the full EQA process. In total, 18 did the serotyping part and 15 did the molecular typing-based cluster analysis.
In the serotyping part, the 11 Listeria monocytogenes isolates were tested to assess the participants’ ability to obtain the correct serotype. Conventional serotyping results were provided by six participants and molecular serotyping by 17 of them. Five performed both serotyping methods.
For the six who used conventional serotyping, performance was high, with four out of 11 isolates correctly serotyped. Fifteen of the 17 labs that used molecular serotyping were able to correctly serotype all 11 isolates.
“The performance of the conventional serotyping results was acceptable but a decrease was observed compared with EQA-4 (the fourth assessment), where the same number of errors were reported, but more labs participated. Again this year, the main problem was an uncommon serotype 3a isolate that two labs reported as 1/2a. One of the labs repeated the mistake from EQA-3, reporting 1/2a instead of 3a. The performance of the PCR-based molecular serotyping was high … The two errors were from two different laboratories reported in two different isolates,” the ECDC reported.
“In general, a trend toward substituting conventional serotyping with molecular was observed through the five EQAs, reflecting a decrease in participation in conventional serotyping from 63 percent to 33 percent and an increase in molecular serotyping from 44 percent to 94 percent from EQA 1 to 5.”
Most participants, 12 of 15, reported cluster analysis using WGS-derived data, while three reported only using PFGE data. Four submitted cluster data based on both PFGE and WGS.
Seven participants performed cluster analysis using PFGE-derived data. Performance was high, with six correctly identifying the cluster of closely related isolates defined by a pre-categorization from the EQA provider among the 11 cluster test isolates.
Twelve labs performed cluster analysis using WGS-derived data and nine correctly identified the cluster of closely related isolates among the test isolates. Two labs only analyzed WGS data from 10 and seven isolates respectively due to data quality not meeting their own qualitative control standards. But, they identified the correct cluster among the remaining isolates.
Only one lab used external assistance for sequencing. Different sequencing platforms were used including MiniSeq, MiSeq, HiSeq, NextSeq and Ion Torrent. All reported using commercial kits for library preparation. Out of the 12 participants, eight used Illumina’s Nextera kit.
Out of the 12 labs, an allele-based method was preferred since eight used core genome multilocus sequence type (cgMLST) and four did single nucleotide polymorphism (SNP) analysis. The one lab not identifying the correct cluster had used SNP analysis.
Three labs reported SNP distances that were several orders of magnitude higher when compared to those using allele-based methods.
“This is problematic in terms of inter-laboratory comparability and cluster definitions and makes the use of SNP distances obtained from non-standardized SNP analyses less suitable for communicating about genetic clusters when investigating international outbreaks,” said ECDC.
“In general, the reported cgMLST results were at a comparable level of allelic difference (0-7) within the cluster isolates despite analysis with different schemes. This highlights the advantages of cgMLST as a method for inter-laboratory comparability and communication about cluster definitions.”