Hospitals that introduced polymerase chain reaction (PCR) testing for Clostridium difficile toxin genes in stool were suddenly confronted with an apparent 40% to 50% increase in detection of the organism not a good thing in a world of public reporting and value-based purchasing. This did not represent an increased infection rate, but, rather, the detection of colonization by the organism, something that is found in as many as 8% of patients without diarrhea admitted to the hospital. This is not just a bad thing for the reputation and finances of the hospital, it is a bad thing for the patients who receive unnecessary treatment.

In 2013, Planche and colleagues reported the results of a large study evaluating the ability of various diagnostic methods to predict outcomes in those with positive results.1 They found that the detection of C. difficile by PCR for toxin gene or by a glutamate dehydrogenase assay was not predictive of increased mortality in the absence of a positive test by either enzyme immunoassay (EIA) or cytotoxigenic culture positivity, with the latter being the most highly correlated with fatality. In contrast, Rao et al, using a two-step system that detects C. difficile glutaraldehyde dehydrogenase and toxin, recently concluded that toxin detection was not predictive of severe disease or mortality,2 although their results have been questioned by Planche and colleagues.3

Thus, we have a problem and a controversy, one that now has been addressed by another group. Polage and colleagues set out “to determine the natural history and need for treatment of patients who are toxin immunoassay negative and polymerase chain reaction (PCR) positive (Toxin−/PCR+) for CDI.” They performed a series of tests on unformed PCR of stool samples submitted to their laboratory for C. difficile testing that began with PCR for toxin gene (the result of which was not reported to the clinician) and toxin immunoassay.

Of the 1416 hospitalized adults whose stools were tested, 131 (9.3%) were Toxin+/PCR+, 162 (11.4%) Toxin−/PCR+, and 1123 (79.3%) were Toxin−/PCR−. Patients who were Toxin+/PCR+ had had greater prior antibiotic exposure, more frequently had leukocytosis, had more diarrhea on day 1, and were more likely to have elevated fecal lactoferrin than those in the other groups — who largely did not differ from each other clinically. They also had a longer duration of diarrhea than Toxin−/PCR+ patients and Toxin−/PCR− patients (P < 0.001), and had a greater risk of diarrhea during the follow-up. In multivariate analysis, Toxin+/PCR+ status had the strongest effect on the duration of diarrhea. Toxin−/PCR+ status and pretest exposure to metronidazole or oral vancomycin were not significant predictors in the multivariable model. One hundred percent of Toxin+/PCR+ patients were treated for a median duration of 14 days with vancomycin or metronidazole. Among the Toxin−/PCR+ patients, 40.7% were treated, but only for a median duration of 6 days; 32.1% of those Toxin−/PCR− received one of these antibiotics for a median of 5 days.

Complications, such as megacolon, need for colectomy or intensive care unit care, and death, occurred in 10 (7.6%) of 131 Toxin+/PCR+ patients, compared to 0 of 162 who were Toxin−/PCR+ and 3 (0.3%) of 1123 (P < 0.001) Toxin−/PCR− patients. There was no significant difference between the latter two groups. In addition, 11 (8.4%) Toxin+/PCR+ patients died compared to 1 (0.6%) and none (P < 0.001) who were Toxin−/PCR+ and Toxin−/PCR−, respectively. The single death in the Toxin−/PCR+ group occurred in a patient with severe comorbidities who had uncomplicated recurrent diarrhea that had resolved prior to withdrawal of care.

These data provide a strong argument that the diagnosis of disease caused by Clostridium difficile should be made on the basis of the detection of toxin, not toxin gene (or glutamate dehydrogenase). This would appear to reduce unnecessary treatment, as well as heartburn, for Infection Prevention Programs and hospital administrators.

The data also provide an argument for antimicrobial stewardship — and not just regarding apparent unnecessary anti-C. difficile treatment in Toxin−/PCR+ patients. Why was a median duration of 6 days of such treatment administered to approximately one-third of patients who were Toxin−/PCR−? 


  1. Planche TD, Davies KA, Coen PG, et al. Differences in outcome according to Clostridium difficile testing method: A prospective multicentre diagnostic validation study of C. difficile infection. Lancet Infect Dis 2013;13:936–945.
  2. Rao K, Micic D, Natarajan M, et al. Clostridium difficile ribotype 027: Relationship to age, detectability of toxins A or B in stool with rapid testing, severe infection, and mortality. Clin Infect Dis 2015;61:233–241.
  3. Planche T, Wilcox M, Walker AS. Fecal-free toxin detection remains the best way to detect Clostridium difficile infection. Clin Infect Dis 2015;61:1210-1211.