Microbiological

All 12 cases in the outbreak were identified as STEC O157:H7 PT2 harbouring stx subtype, stx2a. WGS confirmed that isolates from nine cases were identical and the remaining three isolates differed 1–2 SNPs from the outbreak profile.

The outbreak strain was unique and had not been detected in the preceding 3 years of routine WGS for STEC isolates in England. The outbreak strain fell within the same clade as STEC O157:H7 isolated from UK livestock indicating that the source was of UK domestic origin.

Epidemiological investigations

Cases had onset dates between 28 September 2017 and 23 November 2017 (Fig. 1).

Fig. 1.

Epidemic curve for outbreak cases and timeline of key events in the investigation.

Eleven cases were resident in England and were distributed nationally and one case was resident in Scotland. Nine cases were male and three were female. Cases ranged in age from 1 to 65 years with a median of 16 years. Information on clinical symptoms was available for the 11 cases resident in England. All reported bloody diarrhoea, 10 with abdominal pain, seven with vomiting and five with fever. Eight of 12 cases required hospitalisation and four developed HUS. No deaths were reported.

Review of the ESQs of 11 cases yielded no obvious common exposure between cases. None reported foreign travel in the 7 days prior to onset. Three cases reported direct or indirect contact with ruminants but the wide geographical dispersal of cases suggested a nationally distributed foodborne source of infection was likely.

The case-case study identified that Retailer A was reported by 90.9% outbreak cases (10/11) compared to 19.8% of the control group (106/536, P < 0.001). No other supermarkets were reported more often amongst outbreak cases than the control group (Table 1). Among food exposures, 13 food categories were reported by more than half of outbreak cases with an OR >1 (Supplementary file S1). Free text analysis identified 20 items which were reported significantly more often amongst outbreak cases, among which consumption of beef from Retailer A was most frequently reported (Table 2).

Table 1.

Frequency of shopping at national UK supermarket chains^a reported by outbreak cases compared to STEC- control cases by univariable analysis

^a Reporting of supermarket chains amongst cases was determined from parsing free text fields used to collect data on where specific food items were purchased into a new categorical variable.

^b Control group comprised of all primary STEC O157:H7 cases, not associated with travel abroad or any known outbreaks. The controls were matched to cases on gender and age group and controls were only selected for the months September to December in 2017 and the preceding years (2009–2016).

Table 2.

Frequency of food exposures^a among STEC outbreak cases compared to STEC-control cases by univariable analysis, NESSS data

^a Exposures not collected via standard binary questions but coded through parsing free text fields into coded variables.

Seven cases were interviewed using a trawling questionnaire and all reconfirmed shopping at Retailer A. Consumption or handling of raw beef intended for cooking was reported by all seven and included one or more products of burgers (n = 7), minced beef (n = 3), beef pies (n = 2), steak (n = 1) and roast beef (n = 1). Six of seven cases reported purchasing the beef products from Retailer A.

Ten cases were re-interviewed between 8 December and 11 December 2017 using a questionnaire focussed on burger and mince products. Nine cases reported eating cooked frozen burgers from Retailer A and one from Retailer B. Left over raw product remained in three households and was retrieved and tested. Six cases provided loyalty card data.

Retailer A reviewed the purchase history for the six cases with loyalty card data. This confirmed that 4/6 had purchased the retailer's own branded frozen burgers between August and November 2017. Online shopping history from a further case (without a loyalty card) indicated that they had ordered a different type of Retailer A's own branded fresh burgers. From loyalty card data and patient interviews, specific product details were available for 10/12 cases. In total, nine reported burgers from Retailer A, seven a specific own-brand frozen burger product and two a specific own-brand fresh burger product.

Food chain investigations

STEC O157:H7 was isolated from two leftover burgers sampled from homes of two different cases and reported as STEC O157:H7 PT2 Stx2 on 21 December 2017. On 29 December 2017, WGS confirmed that the two isolates were identical to the outbreak strain. A sample of burger retained at the meat preparation premises was tested and confirmed to have the outbreak strain by WGS.

Product trace-back investigations indicated the burgers were produced at a single meat preparation premise, who produced frozen burgers for all major UK supermarkets. Batch code and production information stamped on the product packages of the two samples obtained from cases freezers stated that they were produced on 5 September 2017 at 07:56 h. The third sample from the meat preparation premises from which the outbreak strain was confirmed by PHE was produced at 07:47 h on the same day.

All three burgers were likely to have been produced from the same batch of material. Fifteen different blocks of raw material from one supplier were used to produce the implicated product. A small quantity of the raw material from one of the batches was also used to produce burgers sold by Retailer B. (No outbreak cases were linked to burgers sold by Retailer B.) Four formal final burger products produced from the same batch of raw ingredient as the implicated batch but at different times were tested and no STEC was detected (Table 3). This indicated that the contamination may have been limited to one mix prepared between 07:00 and 08:00 h on 5 September 2017. The implicated burgers had a ‘best before’ date of September 2018. It was confirmed that the products were distributed exclusively across the UK and not exported abroad and that 30 252 packs had been produced from that batch.

Table 3.

Microbiological sampling and STEC testing results of product samples taken during the outbreak investigation

The meat preparation premises were inspected on 4 January 2018. A review of company records on cleaning, environmental swabbing and food product sampling results was undertaken. Fourteen hygiene swabs were taken at several points along the food processing areas at the premises. Two swabs showed Enterobacteriaceae counts of >10² and <10⁴ cfu per swab but were negative for β-glucuronidase-positive E. coli and coagulase-positive Staphylococci. All 14 swabs were negative for STEC O157. Inspection of the premises confirmed that there were no breaches in compliance against food safety requirements.

The burger packs carried the advice ‘Check food is piping hot’ twice. The label had a safety statement with a ‘call out’ explanation mark on the side of the pack giving advice to cook ‘until no raw meat remains’. The cooking instructions had been verified by an independent laboratory who cooked the product under the grill using the cooking instructions as per the pack and concluded that the recommended temperature of 70 °C or above for 2 min would be achieved.

Retailer A undertook additional testing of the cooking instructions on 18 December 2017 (for grilling, the most common cooking method) which confirmed that post-cook, the burgers (including the core) remained at over 70 °C for at least 2 min. The FSA concluded that if the cooking instructions were adequately followed, the necessary temperature to achieve a 6-log reduction in E. coli in the burgers would be achieved and that the cooking guidance had an adequate safety margin, regardless of the method of cooking burgers (grilled or shallow fried).

Following confirmation of the vehicle of the outbreak, seven food handlers who prepared the burgers for eight outbreak cases were interviewed about storage and cooking of the burgers at home. All reported good storage and cooking practices at home, although this was self-reported. Four cases reported freezer temperatures between −20 and −25 °C, lower than the recommended freezer temperatures of around −18 °C.

Outbreak control measures

Upon confirmation of STEC O157:H7 in two burger samples on 20 December 2017, Retailer A voluntarily removed the implicated products from shop shelves pending further investigations. Following confirmation of the STEC strain as PT2 on 21 December 2017, investigators considered all relevant factors including the severity of illness caused by the outbreak strain, the long shelf life (at least 12 months) of frozen burgers and the likelihood of customers having frozen burgers in their homes and recommended a voluntary recall of the implicated product. The FSA and Retailer A issued a product recall notice on the 22 December 2017 advising those who might have purchased any of the affected batches not to consume the product, and to return it to Retailer A for a full refund. Recall notices were displayed in store, media coverage publicised the recall and Retailer A directly contacted 57 000 customers who had purchased the products to notify them of the recall. A number of customers returned the product and approximately 19 000 of the 30 252 units were destroyed. Retailer B also removed frozen burgers made on two production runs from the same raw material on 22 December 2017.

No further cases linked to the outbreak have been reported to date, indicating that the intervention was successful.