Expanding use of whole genome sequencing pathogen analysis globally will improve trace back investigations during foodborne outbreaks, according to researchers.
A study published in the journal Eurosurveillance reports experiences of using whole genome sequencing (WGS) data to investigate outbreaks of Shiga toxin-producing E. coli (STEC) O157:H7 in England between 2013 and 2017.
The research scientists said more WGS use would ensure rapid implementation of interventions to protect public health, inform risk assessment, and help manage national and international foodborne outbreaks of the potentially deadly E. coli O157:H7.
In the 1990s and early 2000s, pulsed field gel electrophoresis (PFGE) and multilocus variable number tandem repeat analysis (MLVA,) respectively, were used reactively in outbreaks already identified by epidemiological links. Public Health England (PHE) implemented WGS as the molecular typing method for all isolates of E. coli O157:H7 in June 2015.
“Analysis of WGS data delivered an unprecedented level of strain discrimination when compared with MLVA. The robustness of the WGS method ensured confidence in the microbiological identification of linked cases, even when epidemiological links were obscured,” said researchers.
Forensic-level microbiological typing provided by WGS can generate a fine-tuned definition to identify outbreak patients.
There is often poor patient recall of food consumed before onset of symptoms in epidemiological investigations, particularly when the product is a side dish such as salad leaves or raw vegetables, or something such as herbs and spices in foods with multiple ingredients.
In a 2014 STEC outbreak associated with raw milk, four additional patients, identified by WGS, had consumed the milk but initially failed to recall an accurate food history or were unaware the milk was unpasteurized.
Analysis of the STEC O157:H7 dataset at PHE showed that by exploring the context of the deeper phylogenetic relationship between isolates, the source of infection could be linked to specific geographical regions of the United Kingdom.
Analysis of WGS data from a 2016 outbreak linked to contaminated mixed leaf salad revealed the outbreak strain belonged to an uncommon “clade” in the PHE database. The clade included a high proportion of patients reporting recent travel to Mediterranean countries, compared with other clades in the PHE database. Imported red Batavia lettuce leaves were suspected as the vehicle of infection, although there was no microbiological evidence.
“As more countries implement standardized, open access WGS data for routine surveillance of STEC, cross border exchange of WGS data will have a major impact on the ability to investigate national and international outbreaks of foodborne disease,” said researchers.
Change in serotypes for HUS cases
Meanwhile, another study in the same journal found no major changes to hemolytic uremic syndrome (HUS) incidence rates in France between 2007 and 2016, but there was a shift in serogroups. A severe and sometimes fatal type of kidney failure, HUS can be a complication of E. coli infection.
In France, STEC surveillance relies on voluntary HUS surveillance in children younger than 15 years. The French pediatric HUS surveillance system is coordinated by Santé publique France, the national public health agency. Most labs do not test STEC in stool routinely in patients with diarrhea unless clinicians request it. Specific tests are necessary for diagnosis of E. coli infections. The infections can be misdiagnosed because the pathogen causes symptoms similar to common gastronomic viruses, which are often referred to as “stomach flu.”
From 2007 to 2016, a median of 116 pediatric HUS cases was reported each year with a low of 74 and high of 162. During the study period, 1,215 cases were notified, including 32 who were part of six outbreaks. Four of these were foodborne; two in 2011, one in 2012 and one in 2013.
HUS was preceded by diarrhea in 1,099 patients, of whom 545 had bloody diarrhea. The median delay between diarrhea onset and HUS diagnosis was six days. Half of the patients required blood transfusions, and 389 of 1,215 required dialysis. Eleven deaths were reported.
The most frequent STEC serogroups identified were:
- O157 with 212 cases, 23 percent of 942 cases with a stool analysis;
- O26 with 106 cases; and
- O80 with 73 cases).
Other frequent serogroups were O111 with 18 cases and O145 and O55 with 14 cases each.
An important shift occurred regarding STEC serogroups in pediatric HUS cases between 2007 and 2016. STEC O157 has decreased since 2011, while the serogroup O26 has increased since 2010 and O80 emerged in 2012.
Researchers said the surveillance system is evolving, more widespread stool testing of gastroenteritis cases for STEC and routine use of WGS will improve ability to further describe STEC strain characteristics and capacity for outbreak detection.
STEC typing assessment
Finally, 25 labs have participated in the eighth external quality assessment scheme for typing of STEC.
It is organized for labs providing data to the Food- and Waterborne Diseases and Zoonoses Network (FWD-Net), which is managed by the European Centre for Disease Prevention and Control (ECDC). The aim is to assess the quality and comparability of typing data reported by public health national reference laboratories in the network.
Participation fell by 17 percent from the previous assessment, possibly because of adding molecular typing-based cluster analysis (using PFGE and/or WGS without a standard protocol) or removing two independent methods in previous assessments: quality assessments of PFGE and phenotypic analysis.
Twenty labs performed the serotyping part and 57 percent reported results still using phenotypic serotyping.
Full O:H serotyping was performed by 60 percent of 25 labs, with an average score of 86 percent. The more common European serotypes generated the highest scores, e.g. 100 percent for both O157:H7 isolates, while the less frequent O187:H28 had an average score of only 53 percent. Not all labs demonstrated the capacity to determine all O groups and H types. Participation in H typing was low at only 15 of 25.
Eighteen labs performed molecular typing-based cluster analysis using any method. Eleven did cluster analysis using WGS-derived data. Performance was high, with 10 correctly identifying the cluster of closely related isolates.
An allele-based method was preferred for the WGS-based cluster identification since 72 percent used core genome or whole genome multilocus sequence type (cgMLST/wgMLST) compared to 27 percent using single nucleotide polymorphism (SNP) for the main reported cluster analysis.
Seven participants in cluster analysis only used PFGE and three did not identify the correct cluster.
(To sign up for a free subscription to Food Safety News, click here.)