Overdiagnosis of C. difficile infection (CDI) and the importance of toxin detection

Clostridium difficile is an anaerobic toxin-producing intestinal pathogen that causes hospital and community-acquired diarrhea and colitis. In the U.S., estimates run as high as 500,000 cases of C. difficile infection (CDI) annually, with close to 29,000 deaths from the infection.^1,2 In Europe, the numbers probably are similar.³ CDI is only now beginning to be recognized in Asia. This bacterium has taken advantage of our overutilization of antibiotics, which represents by far the primary predisposing factor for the initial infection and onset of CDI. There are other predisposing factors (e.g., advanced age and length of hospitalization); however, the overuse of antibiotics is the most common cause. Toxin detection may help identify patients who would benefit from antibiotic treatment and could lower the unnecessary use of antibiotics.

Today’s antibiotics are effective at killing not only “targeted” pathogens but also our own endogenous protective intestinal flora. There is an urgency to decrease the overuse of antibiotics, not only to reduce the occurrence of CDI, but to prevent the development of antibiotic-resistant bacteria. C. difficile has evolved several antibiotic resistance mechanisms. Probably the most important is the single-site mutation that has occurred in the gyrase gene of some ribotypes, making them resistant to fluoroquinolone antibiotics.⁴ Fluoroquinolones are common antibiotics given for respiratory conditions, so overtreatment of pneumonia can lead to a subsequent case of C. difficile.

There are more than 30 in vitro diagnostic tests for C. difficile. These tests detect glutamate dehydrogenase (GDH), toxins A and B, or the genes for toxins A and B in fecal specimens. Patients can have C. difficile and these specific markers in their feces and still be completely asymptomatic. Patients who have diarrhea also can have these markers in their feces without C. difficile being the cause of the diarrhea. Therefore, all of these tests should be used in conjunction with patient clinical histories to establish an accurate diagnosis of CDI.

“Potential C. difficile excretors”

Over the past few years, several articles have been published that more accurately define the types of tests best suited for CDI. In the largest study to date on C. difficile testing methods, Planche et al⁵ performed a study involving four hospital diagnostic labs in the United Kingdom. Samples were from hospital and community patients and were submitted for routine testing for C. difficile. All assays were performed at the individual sites, and testing consisted of toxigenic culture and tissue culture assay. Tests that detected GDH, toxins A/B, or toxin genes were included, and all sites underwent a training period provided by assay manufacturers. There were 12,420 samples tested from 10,186 patients. The optimum algorithm compared with tissue culture was either GDH or nucleic acid amplification testing (NAAT) as the initial screen, as both had similar sensitivities. These assays were combined with immunological toxin tests. Immunological toxin positives were not seen to have a high enough sensitivity to be the primary screen, but when positive after testing with GDH or NAAT, were considered confirmatory for C. difficile.

Patients who were positive for toxin had a higher case-fatality rate than those who were either toxigenic culture-positive but toxin-negative, or those who were negative by both methods, leading the authors to conclude that a positive toxin result exhibited greater accuracy for CDI, and identified patients who needed treatment. Patients who were culture-positive but cytotoxin-negative had good outcomes even without treatment. The authors suggested the term “potential C. difficile excretor” for patients who were toxigenic culture-positive but toxin-negative, indicating that this group of patients may still be infectious. The authors noted that management of patients who were excretors was unclear.

Algorithm testing with a GDH-toxin or a toxin-NAAT combination stratified patients into three categories: those who had CDI, those who were potential excretors, and those who were negative for C. difficile. In this large study, GDH and NAAT tests exhibited a high NPV, accurately ruling out CDI. Reflex testing with a toxin assay as the second portion of the algorithm reduced the number of patients with discordant results to a more manageable number.

More insights from studies

A study in the United Kingdom performed by Baker et al⁶ at about the same time similarly focused on the clinical characteristics of patients who had a positive toxin test compared to those with a negative toxin test, but had a positive PCR result. Although the number of samples and patients was smaller, the results showed a mortality rate of 38 percent in patients who were positive for toxin compared to 11 percent in those who were PCR-positive but toxin-negative. Patients negative by both assays had a mortality rate of 10 percent, which was similar to the group that was toxin-negative but PCR-positive. Longer bouts of diarrhea were observed in the toxin-positive patients, and there was a high resolution of symptoms without treatment at 14 days in toxin-negative patients. The use of Ct values for quantitating target DNA was not supported.

In a study performed at the University of California Davis Medical Center by Polage et al,⁷ 1,416 hospitalized patients were analyzed. Patients who were positive for toxin and by PCR had higher numbers of C. difficile, higher duration of diarrhea, a greater risk of diarrhea, and higher frequency of hypervirulent 027 strains than other groups. In addition, they had higher levels of fecal lactoferrin (i.e., more extensive intestinal inflammation). This group of patients also had more CDI-related complications (megacolon, fulminant colitis, ICU care related to CDI) and deaths. The authors noted that their findings were consistent with the view that CDI is a toxin-mediated inflammatory disease, and that in the absence of toxin, patients had less antibiotic exposure, lower C. difficile counts, less inflammation, and milder symptoms despite minimal or no treatment. Further, the findings in this study corroborated the results of Planche et al, leading the authors to conclude that most patients with negative toxin test results may not need treatment for CDI, even if C. difficile is detected by NAAT or toxigenic bacterial culture. The results also indicated that patients who had severe or complicated CDI missed by toxin tests were rare. Based on these findings, the authors suggested that as many as half of patients positive by PCR are overdiagnosed and exposed to unnecessary treatment.

The results of several smaller recent studies further support the importance of toxin detection. Beaulieu et al⁸ found that patients who were positive only by PCR had lower white blood cell counts and fewer complications, shorter length of stay in the hospital, and in general, were clinically less severe. Koo et al⁹ noted a doubling in incidence following the implementation of PCR, and found that patients who had toxin in their stool were sicker and hospitalized for longer periods. Patel et al¹⁰ looked at community-acquired CDI and found that white blood cell counts were significantly higher in toxin-positive patients. Their results similarly showed that toxin-positive patients were clinically more severe. Yuhashi et al¹¹ found that the duration of symptoms was significantly shorter in toxin-negative patients and comparable to patients who were negative for the organism. In an extension of the work by Planche et al, Kumar et al¹² found that there were no patients who were toxin-positive but GDH-negative, and that a positive toxin result was associated with higher white blood cell counts and C-reactive protein. These authors also concluded that patients infected with toxigenic strains of C. difficile but who were toxin-negative should be considered as potential spore excretors. Shimizu et al,¹³ using a combination GDH and toxin as the initial test, found that patients who had toxin in their stool were sicker, and proposed the use of toxigenic culture for discordant samples, although NAAT would be more practical.

Recommendations for testing

In a large meta-analysis of results from numerous publications, the following recommendations were developed and adopted for CDI testing in Europe:¹⁴
All submitted unformed stool samples from patients three years or older should be tested. Samples from children under three should be limited to a physician’s request only.
Formed stool samples should not be tested. If ileus is suspected, a rectal swab can be used for toxigenic culture, NAAT, or GDH.
Diagnosis should be based on clinical signs and symptoms in combination with laboratory tests.
A two-step algorithm should be used, starting with either NAAT or GDH. Samples negative at this stage should be reported as negative for CDI. Samples that are positive should be tested with a toxin A/B EIA. Samples positive for toxin in symptomatic patients should be reported as CDI-positive.
If GDH-negative/toxin-positive results are obtained, these should be considered invalid, and the sample should be retested.
Samples that are NAAT- or GDH-positive but toxin-negative may represent C. difficile carriage.
A test of cure is not recommended.

At the recent American Society for Microbiology (ASM) Microbe 2017 meeting, the interest in molecular testing versus algorithm testing was apparent from overflow attendance at several symposia, including “The C. diff debate: the role of diagnostics in disease determination,” and “Impact of toxin-based testing vs nucleic acid testing. Novel approaches to reduce healthcare-associated C. difficile.” Both featured well-known experts in the field of C. difficile diagnostics who discussed the challenges and importance of lab tests and clinical histories when assessing CDI. The recommendation and implementation of algorithm testing for CDI has recently expanded to China,¹⁵ where the important role of C. difficile in healthcare is being demonstrated. In the United States, the Infectious Disease Society of America guidelines are currently being updated.

In summary, treatment with antibiotics continues to be the primary predisposing factor for CDI. Biomarkers for CDI include toxins A and B, GDH, and the genes for toxins A and B. Algorithm testing is the most accurate approach for CDI, and should include initial screening with GDH or NAAT followed by detection of toxin because toxin testing can help confirm true CDI, as demonstrated by more severe disease, longer duration of diarrhea, and higher levels of inflammatory indicators in toxin-positive patients.

REFERENCES

Lessa FC, Mu Y, Bamberg WM, et al. 2015. Burden of Clostridium difficile infection in the United States. N Engl J Med. 372(9):825-834.
Centers for Disease Control and Infection. Healthcare-associated infections: Clostridium difficile infection the Scientific Core of HAI Prevention. https://www.cdc.gov/hai/organisms/cdiff/cdiff_infect.html.
Davies K, Davis G, Barbut F, Eckert C, Petrosillo N, Wilcox MH. Variability in testing policies and impact on reported Clostridium difficile infection rates: results from the pilot Longitudinal European Clostridium difficile Infection Diagnosis surveillance study (LuCID). Eur J Clin Microbiol Infect Dis. 2016; 35(12):1949-1956.
Dingle KE, Didelot X, Phuong Quan T, et al. Effects of control interventions on Clostridium difficile infection in England: an observational study. Lancet. 2017;17(4):411-421.
Planche TD, Davies KA, Coen PG, et al. Differences in outcome according to Clostridium difficile testing method: a prospective multicenter diagnostic validation study of C. difficile infection. Lancet. 2013;13(11):936-945.
Baker I, Leeming JP, Reynolds R, Ibrahim I, Darley E. Clinical relevance of a positive molecular test in the diagnosis of Clostridium difficile infection. J Hosp Infect. 2013;84(4):311-315.
Polage CR, Gyorke CE, Kennedy MA, et al. Overdiagnosis of Clostridium difficile infection in the molecular test era. JAMA Intern Med. 2015;175(11):1792-801.
Beaulieu C, Dionne L-L Julien A-S, Longtin Y. Clinical characteristics and outcome of patients with Clostridium difficile infection diagnosed by PCR versus a three‐step algorithm. Clin Microbiol Infect. 2014;20(10):1067‐1073.
Koo HL,Van JN, Zhao M, et al. Real‐time polymerase chain reaction detection of asymptomatic Clostridium difficile colonization and rising C. difficile‐associated disease rates. 2014; Infect Control Hosp Epidemiol. 35(6):667‐673.
Patel HJ, Randhawa J, Nanavati S, Marton LR, Baddoura WJ, DeBari VA. Laboratory and clinical features of EIA toxin‐positive and EIA toxin‐ negative community‐acquired Clostridium difficile infection. Annals Clin & Lab Sci. 2015;45(3):333‐339.
Yuhashi K, Yagihara Y, Misawa Y, et al. Diagnosing Clostridium difficile‐associated diarrhea using enzyme immunoassay: the clinical significance of toxin negativity in glutamate dehydrogenase‐positive patients. Infect & Drug Resist. 2016 9:93‐99.
Kumar S, Pollok R, Muscat I, Planche T. Diagnosis and outcome of Clostridium difficile infection by toxin enzyme immunoassay and PCR in an island population. J Gastroenterol & Hepatol. 2017. doi: 10.1111/jgh.13504.
Shimizu H, Mori M, Yoshimoto N. 2015. Clostridium difficile infection is more severe when toxin is detected in the stool than when detected only by a toxigenic culture. Intern Med. 2015;54:2155‐ 2159.
Crobach, M. Planche T, Eckert C, et al. European Society of Clinical Microbiology and Infectious Diseases: update of the diagnostic guidance document for Clostridium difficile infection. Clin Microbiol & Infect. 2016;22.Supplement 4:S63–S81.
Xu Y, Zhang M. Guidance and consensus: expert consensus on diagnosis and treatment of Clostridium difficile infection in Chinese adults. Med J Peking Union Med College Hospital. 2017;8:131-138.