A growing reliance on new, cheaper and faster testing for infectious diseases has experts questioning the accuracy of a reported decline of E. coli O157 cases in the U.S., challenging one of the nation’s few food safety success stories.
Rather than identifying the specific strain of E. coli, such as O157 or O104, faster tests only recognize a poisonous toxin known as shiga, casting doubt on the accuracy of official numbers reported by the Centers for Disease Control and Prevention, according to foodborne disease epidemiologists interviewed by News21.
Most E. coli strains are harmless, but a few hundred harbor shiga toxin, which causes bloody diarrhea and can lead to kidney failure and death, usually in small children. E. coli O157 and other shiga toxin strains are typically transmitted by consuming undercooked ground beef, raw produce and unpasteurized milk.
“If (physicians) know there’s shiga toxin in there, that’s all they need to know for treating the patient,” said Dr. Tim Jones, Tennessee’s state epidemiologist and a leading researcher of foodborne disease surveillance.
But knowing what strain of E. coli a patient has is crucial to uncovering foodborne outbreaks and improving public health surveillance. The interests of public health and patient care do not always align, however, especially when cost is an issue.
“All of the sudden, it’s cheaper for clinicians to do toxin tests,” said Dr. Laurene Mascola, chief of communicable disease control for the Los Angeles County Department of Health. “So that’s sort of a scary trend for the future.”
Since 1996, CDC has been assessing national trends in foodborne diseases based on surveillance data through a system called FoodNet. Of the foodborne diseases monitored during the last 15 years, only E. coli O157 declined enough – by 50 percent – to meet the government’s 10-year national health objective set in 2000, according to the agency’s annual report released in June 2011.
“It’s definitely true that the reported rates have gone down, but it’s not possible to tell from that if the true rates have gone down,” said Dr. Bela Matyas, a foodborne disease epidemiologist and chief health officer for Solano County in Northern California.
For decades, physicians had the tools to test for only E. coli O157 by taking a patient’s stool sample and growing a culture, which could take two to three days.
During the last 10 years, new rapid tests have allowed doctors to test broadly for the shiga toxin without identifying the exact strain. These tests allow for diagnosis within a few hours instead of a few days.
But as the Center for Infectious Disease Research & Policy at the University of Minnesota reported in July 2011, “Clinical laboratories are increasingly using (rapid) tests to detect (E.coli), either along with or instead of the traditional method of growing the bacterium in a culture.”
When labs forgo culture testing, the specific strain of E. coli is not identified. It is not known how many of these unidentified strains are O157, and they are not counted in CDC reports.
CDC officials acknowledge that they can’t account for rapid testing in their surveillance. The agency does not know how prevalent the toxin tests for E. coli are becoming and is uncertain how surveillance results are affected, said Dr. Olga Henao, who leads coordination of FoodNet.
Part of Mascola’s job in Los Angeles involves urging physicians to perform culture tests, but arm-twisting only goes so far, she said, when hospitals don’t consider culture tests cost effective. Her experience highlights the growing divide between individual patient care and public health.
“They really don’t care from a personal point of view whether you have salmonella or shigella; they just want you to get better,” Mascola said. “Whereas from a public health point of view, we want to know about every possible foodborne organism to see whether or not it’s part of a bigger picture.”
Without a culture test, labs lack the puzzle pieces to link cases in a way that identifies an outbreak. This makes it almost impossible to trace an outbreak to its food source, putting the public’s health at greater risk.
Additionally, research demonstrates that rapid testing is a problem when used for other foodborne diseases. A 2010 Minnesota Department of Health study found traditional non-rapid testing identified cryptosporidiosis, which causes diarrhea and vomiting, as a correct illness 97 percent of the time. Rapid testing in clinical laboratories led to a correct cryptosporidiosis diagnosis 56 percent of the time.
Similar tests are coming out for campylobacter, which can cause bloody diarrhea, according to Dr. Kirk Smith, a senior foodborne epidemiologist in Minnesota and co-author of the study. Campylobacter sickens more than 2.4 million Americans each year.
Since 2009, the CDC has recommended that labs perform both a rapid test and a culture test for E. coli, a suggestion Tennessee’s Jones said is increasingly unrealistic because of the costs and labor involved. More labs, he said, are beginning to use only rapid tests.
A survey Jones conducted found that only 16 percent of labs in his state performed rapid tests for E. coli in 2007. By 2010, that rate had increased to 52 percent. During the same period, a third of Tennessee’s E. coli cases were tested only for shiga toxin, meaning a specific strain was never identified.
“CDC is saying, ‘We don’t know how good those tests are, we’re not going to count them,’ … and it’s going to look like (E. coli) is going away,” Jones said. “When that’s absolutely not going to be the case.”
To combat the increased use of rapid tests, some states are requiring that clinical labs send samples to state public health labs for a culture tests after shiga toxin is found.
Although this process leads to a delay in matching specimens and places an added burden on state labs operating on taxpayer dollars, it is better than the alternative, Jones said. He added that rapid tests are an unavoidable reality of modern public health surveillance for a growing number of infectious diseases.
“Next it’s going to be salmonella, and who knows what else after that,” he said.