Changes in E coli testing pose surveillance challenges

Jul 21, 2011 (CIDRAP News) – The big Escherichia coli O104:H4 epidemic in Germany comes at a time when methods of testing for pathogenic E coli are changing, posing some challenges for public health officials intent on detecting E coli outbreaks quickly and monitoring trends as accurately as possible.

Clinical laboratories are increasingly using enzyme immunoassay (EIA) tests to detect Shiga toxin–producing (pathogenic) E coli (STEC), either along with or instead of the traditional method of growing the bacterium in a culture. EIA tests detect the Shiga toxin (ST) rather than the bacterium itself and are faster than culturing.

The use of ST tests has pluses and minuses, according to experts. The big concern is that if an ST test is used alone, the information yielded is too limited. The tests detect any strain of STEC, but they don't tell the serotype, which is important for identifying and investigating outbreaks.

In contrast, culturing detects E coli O157:H7, the most common STEC, but it doesn't sniff out any of the other STEC serotypes, such as O26, O45, O111, or the O104 strain that has caused Germany's big E coli crisis.

Because of these limitations, the US Centers for Disease Control and Prevention (CDC) recommends that, when testing a sample for STEC, labs should run ST tests and cultures simultaneously. If the toxin test is positive, then the culture will tell if the strain is O157 or not. If the strain is something other than O157, more specialized testing is necessary to identify it.

Thus, some state health departments are requesting or requiring clinical labs, when they get a positive ST test, to send a specimen to the state lab for further testing so the strain can be identified and analyzed.

As an example of the uncertainties related to testing, increases in the reported numbers of non-O157 E coli strains have generated considerable attention lately, but the increases probably reflect only the increased use of ST tests, not a real rise in incidence of cases, according to experts.

CDC surveillance shows that detected E coli O157 cases have been falling, while detected non-O157 cases have been rising in recent years. The FoodNet surveillance system, which gathers foodborne disease reports in 10 states, showed that in 2010, for the first time, non-O157 E coli cases were slightly more common than O157 cases: 1.0 versus .0.9 case per 100,000 population.

The FoodNet report points out that changing lab practices, including the use of culture-independent tests for STEC, can affect the reported incidence of infection and that the increasing use of such tests may have a negative effect on current surveillance strategies.

Positives and negatives
Robert Tauxe, MD, deputy director of the CDC's Division of Foodborne, Bacterial and Mycotic Diseases, sees positives and negatives in the growing use of ST tests for E coli.

"We think of them as screening tests," he said in a recent interview. "The basic issue is that the screening tests can provide evidence of Shiga toxin–producing organisms fast, but it won't be known what sort of organism it is."

"The key is that when a test is positive for Shiga toxin, then the lab needs to send either isolates or E coli from that specimen or the broth that was incubated overnight to the state public health lab so they can see what organism was producing the toxin," Tauxe said. "We don't really know how many labs are actually following that, but we think at this point a lot of them are."

"If the lab doesn't forward the materials to the state public health lab, then we don't know if it's an O157 or some other STEC, and there's nothing to enter into PulseNet [the CDC database of DNA fingerprints for pathogens], and our surveillance will take a major step backward," he added.

Tim F. Jones, MD, Tennessee's state epidemiologist, commented, "One of the responses of state health departments to help get around this problem has been to require that any Shiga toxin–positive test be reported, and require the lab to send in the broth. They can just do the STEC test, but send in the sample, and then we at the state health department will absorb the cost and time and do the culture. It seems to be working. We don't get 100%, but we're getting some."

Jones worries about O157 cases going undetected if ST testing is used alone. "Now some O157 cases that we were recognizing with culture are getting lost in that unknown group," he said.

But he added, "We're also detecting non-O157 outbreaks, which we were completely missing before. And that part's good, as long as we can get isolates in and figure out what they are."

Tauxe agrees, saying an advantage of the use of Shiga toxin tests is that more and more non-O175 STEC infections are being identified and reported to FoodNet,. "So O157 has been decreasing, we think because of measures that have been taken in the beef industry, but non-O157 has been increasing, we think because of these tests," he said.

Others agree that the rising numbers of non-O157 reports reflect the increasing use of ST tests and a resulting increase in detection rather than a real increase in cases. This week Connecticut public health researchers published a report in Clinical Infectious Diseases (CID) on O157 and non-O157 cases identified in the state from 2000 through 2009. They reported that the incidence of non-O157 infections has been increasing, but when they adjusted their data for non-O157 cases that would have been detected if ST testing instead of culture for O157 had been used initially, they concluded that both types decreased in the course of the decade.

How widely are ST tests used?
Data are scarce on the numbers of clinical labs using ST tests versus culture or both. But Katie Miller, MPH, CPH, and colleagues at the Washington State Department of Health recently conducted a statewide survey and came up with some.

"For diagnosis, as the CDC recommends, it's important for all clinical labs to do both culture and nonculture Shiga toxin tests," Miller said. She said Washington requires clinical labs to submit specimens to the state lab if they detect ST or E coli O157.

She and her colleagues conducted a survey on E coli testing practices at all microbiology labs in the state, except those that send samples to reference labs, and got a 100% response, Miller said in an interview. The survey showed that:

  • About 25% of labs, handling about 40% of all samples, are both culturing and using ST assays. "We expect that number to rise," Miller said.
  • About 65% of labs, accounting for about half of all specimens, are only culturing for O157—so non-O157 strains would not be detected initially.
  • Four labs that collectively handle about 13% of all specimens use only ST tests.
  • Overall, about half of all specimens are being screened with an ST test, either alone or in combination with culturing.

The survey also revealed that most labs that are using ST testing started doing so in 2009 and 2010, when Washington had its biggest increases in reported non-O157 cases, Miller said. "We do feel that the increased Shiga toxin testing very likely contributed to the [apparent] increased incidence of non-O157 STEC in Washington," she commented.

Miller sees problems with exclusive initial use of either type of test. "Those labs that use only Shiga toxin assay are delaying the picking up of O157," she said. "It's important to start those tests simultaneously and immediately, for prompt diagnosis and to detect outbreaks." (Outbreak requires an isolate of the organism so investigators can determine if a series of cases is caused by the same strain.)

At the same time, since many labs are culturing only for O157, "that's potentially missing a great deal of non-O157," she said.

Use of ST testing alone can also pose problems for clinical decision making, according to some. "Use of the Stx [Shiga toxin] EIA without simultaneous or subsequent culture for O157 STEC delays or prevents further characterization of isolates and does not provide complete information for clinical decision making, as there is substantial variability in clinical outcome among the different STEC serogroups," a team of experts from several state health departments said in a report in the April issue of Foodborne Pathogens and Disease.

Good cooperation from labs
Carlota Medus, PhD, MPH, an epidemiologist in foodborne, vector-borne and zoonotic diseases at the Minnesota Department of Health (MDH), shares the concerns about the trend in testing for STEC, but says it has not caused any big problems in Minnesota as yet.

"We built our systems of outbreak detection and investigation assuming we'll have an isolate, that we'll actually get the E coli here," she said. "Having the actual bacteria, we can do serotyping and PFGE [pulsed-field gel electrophoresis, or DNA fingerprinting]. Now the emergence or acceptance of nonculture methods in clinical labs is a challenge we're facing."

So far it hasn't been much of a problem in Minnesota, Medus said. The MDH now asks labs, when they have a positive test result, to send a stool sample or the enrichment broth so the state lab can culture and analyze it.

"Since our labs have started using these tests, clinical labs have been excellent in sending the broths," she said. Also, most labs in the state have followed the CDC recommendation about running a culture for O157 if they use an ST test. "Instead of replacing culture, they added a nonculture test. So now they're detecting pathogenic strains of E coli that before you couldn't detect," she said.

However, "In the US as a whole that's not necessarily the case," Medus added. "A lot of labs have started using nonculture methods instead."

A burden for state labs
One effect of the growing use of ST tests is that it is increasing the workload for state health labs, experts say.

"This is a lot of additional work in our state public health labs, because all these E coli strains themselves or the broths are being sent to state labs, who are now doing more work than they used to because they're doing the isolation of the organism," said Tauxe of the CDC. "It does mean more of a workload in state labs in a time of decreasing budgets."

Washington's Miller agreed: "It is adding an increased burden to public health labs, because it comes here and we culture and we retest with Shiga toxin assay."

Shari Shea, director of food safety for the Association of Public Health Laboratories (APHL), said she does not have hard data but thinks that most clinical labs send at least some of their ST-positive specimens to their state public health lab. Anecdotal reports are that some states "are hit very hard by the need to subculture from broth and do not have the resources to keep up; other states are able to handle the testing," she commented by e-mail.

"We are running out of resources to do this work, which means the public could be at risk for an undetected outbreak of non-O157 STEC," Shea said.

Implications for prevention
Another potential drawback of ST testing alone is that it can hamper efforts to prevent STEC infections, according to Tauxe.

"If we can't tell the difference between O157 and non-O157, it's harder to know which target to attack with prevention," he said.

"There's a lot we still don't know about the non-O157," he added. "O157 is mostly [found in] ground beef and other beef and some produce; it's not clear that that's the case for the non-O157. We wouldn't want to just be doing everything for all STEC. Our prevention measures are more effective when they're more targeted."

The 10-year Connecticut study that was published this week provided support for Tauxe's point. It showed that, compared with O157 cases, non-O157 cases as a group were somewhat less likely to be associated with ground beef (17.8% vs 25.5% of cases) and more likely to be associated with drinking untreated surface water (13.9% vs 7.5%) and with international travel (15.3% vs 2.5%.)

See also:

CID abstract on E coli cases in Connecticut from 2000 to 2009

October 2009 CDC recommendations on E coli testing by clinical labs

2010 CDC FoodNet report

January 2011 Emerging Infectious Diseases report with estimates of foodborne illness cases

April 2011 Foodborne Pathogens and Disease report on lab practices for STEC identification

This week's top reads