At each stage, the E. coli sneaked through. It came in with the feces caked on the hide of at least one cow, a so-called "super-shedder" of bacteria, and persevered. The E. coli wasn't caught on the kill floor, survived cleaning and clung on during dehiding, in which a cow's skin is peeled away.
It reached the cutting table – a bacteria watershed, where the cow is cut into different types of beef, including "trim," the odds and ends that become hamburger. The E. coli went undetected in the 325 grams of beef trim tested from this particular 2,000-pound batch, so it moved through. When alarms sounded, it was in stores.
As Ottawa eyes changes to food safety programs, the Canadian Food Inspection Agency is seeking the cause of an E. coli outbreak at XL Foods in Brooks, Alta. There wasn't one massive failure, but tiny ones that added up. Officials say the system worked, more or less, but couldn't stop it.
"That super-shedder that we had way upstream got through all those hoops and hurdles, and looked clean, [but] had just enough E. coli to basically get there [undetected]," CFIA meat programs director Richard Arsenault says.
The CFIA says the company failed in "trend analysis," or connecting the dots of tests that showed E. coli in discarded batches, but questions remain. Testing may need improvement. Cross contamination, particularly at an Edmonton Costco store, made things worse. The union that represents workers at the facility has said the production line moves too fast and the hot water for washing surfaces once failed. Messages left for company executives and spokesmen were not returned on Monday.
Whatever caused it, the recall is one of the largest in Canadian history, at more than 1.5-million pounds of beef from every province and territory. Nine people have fallen ill, four of them tied by genetic testing to tainted meat. It all raises questions about the system itself – if everything worked, more or less, how can super-shedders be caught in the future? (Agriculture Minister Gerry Ritz, whose file includes the CFIA, has declined interviews.)
The XL Foods plant is one of the three largest slaughterhouses in Canada; many smaller producers buy killed, skinned and cleaned carcasses from there. It was at another plant – the CFIA won't say which – that testing on beef trim came back positive for E. coli on Sept. 4. The plant had got the carcass from XL Foods. A day earlier, U.S. officials found E. coli in an XL shipment at the border. CFIA officials swarmed the Brooks plant trying to figure out what went wrong. But they still let it operate, thinking it was a one-off. People were already getting sick. The recall began Sept. 16, and the shutdown was announced Sept. 27.
Workers say the plant had been speeding up. The morning "A-shift" is being asked to handle as many as 2,400 of the day's 4,000 cows, said Jean Mulimbi, a union steward with the United Food and Commercial Workers Canada local 401, representing plant workers. The afternoon "B shift" then handles fewer cows and works fewer hours, he said. It leaves less time to clean equipment.
"All this is coming from the line speed," Mr. Mulimbi said. "...We tried to fix it, but nobody paid attention."
The production line's speed has been a bargaining issue at the plant. "One of the most glaring examples of [the company's] lack of concern is shown by the company forcing high line speeds on workers, even though most of the production areas are operating short-handed," the union said in a spring update on plant conditions.
At the plant, cows are first examined by XL Foods and CFIA staff – some cows have 30 pounds of mud and feces on their hides from feed lots, trucks and holding pens – but an E. coli-carrying super-shedder looks the same. They're then stunned and taken to the kill floor, where they're eviscerated and de-hided. This is the messy part. The carcasses are then chilled, and held for a day or more before being cut up. That's why animals slaughtered on one day (Aug. 23 in this case) take up several production days (five).
Canada ranks poorly on ability to track contaminated food, says Sylvain Charlebois, one of the authors of a 2010 food security study by the University of Regina and University of Saskatchewan. In it, Canada ranked fourth overall, but 14th in traceability. The CFIA has faced questions about its response – 12 days passed from the first positive test to a public notice and recall. But Dr. Charlebois (a CFIA adviser) says it functioned well. "I would argue two to three weeks would actually be an adequate time to respond," he said, noting two months passed between the first illness and first recall during Maple Leaf's listeria outbreak several years ago.
The plant could reopen as soon as this week, Dr. Arsenault said, calling the recall an extreme case in an otherwise improving system. "It's like a combination of a bunch of things that all line up," he said. "And you get this multiple hurdle thing, where the hurdle isn't as high as it's perhaps designed to be all the time, but it was pretty darn close. Just enough got through."