ECDC: Molecular surveillance system relies on labs to produce comparable results

By Joseph James Whitworth

- Last updated on GMT

©iStock/Sebastian Kauliltzki
©iStock/Sebastian Kauliltzki

Related tags Type i and type ii errors

A number of technical issues reducing the quality of typing results have been identified in EU labs, according to the European Centre for Disease Prevention and Control (ECDC).

The seventh round of the external quality assessment (EQA) scheme for typing of Shiga toxin/verocytotoxin-producing E. coli (STEC/VTEC) was organised for 30 labs in the Food- and Waterborne Diseases and Zoonoses network (FWD-Net).

Issues identified could be improved by optimising lab procedures, training and capacity building.

It included O:H serotyping, genotyping of virulence genes (eae, ehxA, vtx1 and vtx2), genes for Enteroaggregative E. coli (EAEC), subtyping of vtx genes, and phenotypic detection of β-glucuronidase, enterohaemolysin, ESBL production, sorbitol fermentation and VT production.

More than half (17/30) of the laboratories participated in the four parts of the scheme (PFGE, sero-, geno- and phenotyping) and three labs completed all methods in all parts.

In total, 19 labs did the PFGE part and almost all (29/30) in a selection of the other three.

ECDC said the gel quality varied considerably resulting in a highly variable quality of profiles for the individual test strains.

However, fourteen labs were able to produce a gel of sufficient quality to enable profile detection and inter-laboratory comparison.

Most (13/19) in the PFGE part completed the lab (gel) and the analysing BioNumerics (BN). BN is software initially developed for PFGE gel analysis.

All 30 submitted results, however, deviation between the methods registered and those performed were identified for half (15/30) of the participants.

Poor performance of O-grouping

In the serotyping, the 27 participants performed O grouping but only 17/27 did H typing.

Only two were able to type all ten test strains correctly for O-grouping. Twelve reported an incorrect O-group for one or more strains.

The general performance for H-typing was better with the majority (11/17) correctly H-typing all ten test strains.

In the genotyping part, the 28 participants submitted results for eae and vtx genes, while 18 labs did for the ehxA gene.

Twenty-five labs gave subtyping results, and 21 reported for EAEC; aaiC (17/28) and aggR (21/28).

Phenotyping problems

In the phenotyping part, the participation rate was highest for sorbitol fermentation (26/28) and lowest for VT detection (8/28).

A total of 18 labs performed detection of β-glucuronidase, enterohaemolysin or ESBL production.

VTEC in Europe

In 2014, confirmed cases of VTEC infections were 5,955, a slight decrease from 2013, and seven deaths were reported.

As in previous years, the most commonly reported VTEC O group was O157 (46.3% of cases with known serogroup) followed by O26, O103, O145, O91, O146 and O111.

Three of the four labs reported incorrect eae results, giving a false negative result whereas the other incorrect eae results were for different strains.

In all, eae was misidentified seven times; five false negatives and three false positives and ehxA was misidentified six times, two false negatives and four false positives.

One laboratory missed the presence of both eae and ehxA in one strain and reported incorrect haemolysin and vtx1 subtype.

A total of 25 labs subtyped vtx1 and vtx2, the majority (22/25) subtyping vtx1 correctly but only half (13/25) reporting the correct vtx2 subtype for all ten test strains.

Incorrect vtx2 subtypes were reported 24 times, the majority (14/24) due to reporting two different subtypes for a strain positive for only one vtx2 subtype.

Performance for detection of enterohaemolysin production was fairly poor, with nine of 14 reporting correct results. Most (19/23) of the incorrect findings were false negatives.

Performance for detection of VT production was also fairly poor with five of eight getting 100%.

ECDC said inconsistency in the number of tests performed per strain and per lab have been a ‘recurrent problem’ throughout the EQAs.

“The participants have not explained why a specific test was not performed on all 10-test strains. This was particularly evident for O grouping and H typing where laboratories submitted multiple entries for not done.

“These inconsistencies reduce comparability between the tests and the laboratories and complicate the analyses. Furthermore, in all EQAs there have been deviations between the methods registered and the methods performed.”

Related topics Food Safety & Quality

Related news

Show more

Follow us


View more