Skip to main content
Premium Trial:

Request an Annual Quote

Study Results Suggest Overdiagnosis of C. Difficile With Molecular Testing


NEW YORK (GenomeWeb) – After tracking test results and outcomes for hospitalized adults over several years, researchers from the University of California Davis have collected new evidence that molecular tests may not be the best tool to determine whether patients should be treated for Clostridium difficile infection.

Their study, published last week in JAMA Internal Medicine, found that though PCR testing may accurately detect the presence of C. diff, a significant proportion of patients appear to test positive for the bacterium regardless of whether it is actually causing their clinical symptoms.

As a result, testing for the presence of C. diff toxins, rather than the bacterium itself, appears to be the more accurate way to identify patients whose symptoms actually are linked to this organism and who need to be treated accordingly, the researchers found.

Christopher Polage, the study's first author, and a member of the UC Davis pathology faculty, told GenomeWeb that he and his team set out to answer a lingering question for clinical C. diff diagnosis — whether patients who are positive for C. diff infection based on molecular testing but negative for disease-associated toxins actually need C. diff-targeted treatment.

"At the time we originally planned the study there was very little data in the literature that was recent that looked at clinical outcomes in toxin-negative patients," he said.

"There was some very old data —from the 90s — where they looked at people who were positive by culture, but negative by toxin assays, and their conclusion back then was that these individuals were indistinguishable from people who didn't have C. diff."

But for some reason, Polage said, this data was superseded in recent years by a push for the adoption of molecular tests to define clinical C. diff infection and guide clinical decision-making.

"There were some prominent people calling very loudly, saying that if you are not doing a DNA detection test you are missing patients and harming people. It was close to that stark," Polage said.

"Initially, we did a retrospective outcomes study and we couldn’t find any evidence of any negative outcomes in toxin-negative patients despite what everyone was saying in meetings and on listserves. … Then we planned the prospective study to really provide the outcomes data that was lacking," he explained.

In their study, Polage and his UC Davis colleagues collected a systematic set of data comparing toxin assay results with PCR test results for a total of 1,416 adults who were hospitalized and tested 72 hours or longer after their admission between December 2010 and October 2012.

The group then divided this cohort, based on their results, into three groups: toxin-positive/PCR-positive (double-positive), toxin-negative/PCR-positive, and toxin-negative/PCR-negative (double negative).

They then tracked a set of outcomes, including duration of diarrhea during up to 14 days of treatment, the rate of CDI-related complications such as colectomy, megacolon, and ICU admission, as well as CDI-related death within 30 days.

According to the authors, 21 percent of the overall cohort tested positive for C. diff using PCR. However, more than half of those patients were negative for the presence of C. diff toxins based on UC Davis' standard immunoassay-based clinical toxin testing.

This discordance alone does not necessarily indicate that the PCR testing results in overdiagnosis. But when the researchers then looked at how clinical symptoms and outcomes stratified among the three groups, it was clear that PCR testing alone did not seem to be appropriately selecting patients with a true clinical C. diff infection rather than a symptomless colonization.

Toxin-negative/PCR-positive subjects in the study had lower C diff. bacterial load, and less fecal inflammation and diarrhea than the double-positive group. They also had shorter duration of diarrhea, and suffered no CDI-related complications compared to 10 complications recorded for the double positive patients.

Not only did the toxin-negative/PCR-positive subset look much different clinically than the double-positive group, they also appeared very similar to the PCR-negative patients, tracking closely in terms of every symptom and outcome measure the researchers looked at.

According to Polage and his coauthors, the results "strongly suggest that most patients with negative toxin test results and C. difficile detected by PCR do not need treatment for CDI."

This does not mean that PCR testing or other molecular strategies do not have usefulness, however, Polage stressed.

The high negative predictive value of PCR means these technologies offer a unique contribution to infection prevention, due to the potential of toxin-negative/PCR-positive patients to still spread CDI within hospitals healthcare facilities.

But in terms of clinical decision making, determining whether a patient should be treated for CDI or not, the study results demonstrate clearly that the defining measure should be toxin testing, Polage said.

While Polage and his colleagues were still working on their study, a large effort in the UK came to a similar conclusion.

"What the people in Britain proposed was basically a reversal of the approach people have been using here in the US," Polage said. "They were in a position over in the UK to use their results to reverse practice … and redefine guidelines to drive practice back toward toxin testing. People in theUS, despite that, have remained much more skeptical."

"The message I want people to take home with them [from our new data] is not that there is no role for molecular testing," he said. "But what we need to get across to physicians is that most of their treatment decisions should be based on toxin presence or absence, and that they should view these other molecular positives as most likely not needing treatment, whereas the reflex up to now would have been that they all need treatment."

Though the implications for toxin-negative patients are clear in both the UK study and the UC Davis team's work, Polage said that there are still some other unanswered questions that may warrant future study, for instance, how to classify and how to best treat patients who current toxin immunoassay methods deem C. diff negative, but who might test positive for lower levels of toxin using more sensitive methods.

"People continue to have anxiety that the toxin immunoassays are not quite as sensitive as we might like them to be," he said.

Several groups are working on developing methods for measuring C. diff toxins at a much lower concentration of limit of detection, including one working with Quanterix's Simoa technology, but according to Polage, there isn't necessarily definitive evidence that the additional patients such technology would identify would actually require treatment.

Polage and his colleagues addressed this question briefly in their analysis, looking at a subset of their cohort that was PCR-positive and toxin-negative by immunoassay, but would have been considered toxin-positive using an older culture-based technology, called the cell cytotoxin assay. Among these subjects, outcomes were similar to those of the regular toxin-negative/PCR-positive group.

"In our data we didn't see evidence that many of those people need treatment … but, there might be value in trying to do a larger trial, specifically asking the question about that sub-cohort, which for us was about 15 to 20 percent of the patients with positive PCR results," he said.