Clostridium difficile has leapfrogged over MRSA as the most common healthcare-acquired infection (HAI). Its ability to cause antibiotic-associated diarrhea (AAD) and colitis (AAC) is well-established. In the United States, there are about 500,000 cases each year, with up to 30,000 deaths attributed to the disease. Healthcare costs for C. difficile infection (CDI) now run more than $6 billion annually. The numbers in Europe probably are similar, although cases in the United Kingdom have been declining. The success in the U.K. is in part due to an extensive educational awareness program about the disease in that nation and its near universal adherence to multistep-algorithmic diagnostic testing, which enables the healthcare system to appropriately find and treat patients with CDI and thereby limit its spread.
Emergence of CDI
Prior to the early 2000s, the incidence of CDI was rising, but it was not considered to be nearly as significant as healthcare-acquired MRSA infections. Recognition of C. difficile as an emerging pathogen quickly coalesced with the appearance of the hypervirulent strain 027, a fluoroquinolone-resistant strain of C. difficile that generally produces more toxin and grows to higher numbers in the intestine than other strains. Because of these increased virulence traits (primarily antibiotic resistance), 027 spread quickly among medical facilities, moving from Europe to Canada and then to the U.S. With the spread of 027, the overall incidence of C. difficile began to increase dramatically. Hospitals, particularly in Canada, noted rising mortality rates. In fact, some Canadian institutions observed a 20 percent mortality rate in patients infected with 027.
C. difficile disease occurs primarily in hospitalized elderly patients, but it has extended beyond this population into the community, infecting younger, healthier patients. As the hypervirulent 027 strain spread, C. difficile began to appear as a community-acquired pathogen. The community-acquired patient population skewed younger and more female than the hospital-acquired population. Some researchers also are categorizing particular C. difficile ribotypes as foodborne pathogens. A more thorough understanding of the emerging pattern of C. difficile is continuing to develop.
CDI sometimes does not cause symptoms, or the infection can be subclinical, or the infection can result in clinically apparent disease. That is one reason why diagnosing CDI is challenging. Consider the following:
- C. difficile can be found in healthy infants at levels that would cause pseudomembranous colitis (the severe stage of CDI) in elderly adult patients; we still do not understand the mechanism of this protection in infants.
- The organism often can be found in hospitalized patients who do not have diarrhea. In fact, a patient’s chance of picking up C. difficile in the hospital and carrying it asymptomatically is greater than developing active CDI while in the hospital.
- C. difficile can be carried in persons who have diarrhea caused by norovirus or Campylobacter or any one of a number of other infections or conditions (for example, other intestinal pathogens, inflammatory bowel disease), but not be involved in causing the diarrhea.
- C. difficile can be a passive bystander in patients with diarrhea for a host of other reasons including laxative use, chemotherapy, liquid diets, stress, and pharmaceutical side effects. These C. difficile carriers should not be treated because the antibiotic regimen puts the patient at greater risk for a true active CDI.
- C. difficile can cause a mild self-limiting diarrhea that probably does not require antibiotic treatment, although the patient most likely will receive such treatment if C. difficile is detected.
- And, of course, C. difficile can cause severe diarrhea and colitis that, if left untreated, can progress to severe pseudomembranous colitis.
Therefore, due to the prevalence of carriers, accurately identifying patients who have true CDI and need treatment remains a significant challenge.
The diagnostic landscape
C. difficile and its ability to cause diarrhea and colitis in patients treated with antibiotics was first recognized more than 40 years ago. Decades later, we still are trying to determine the most accurate approach for diagnosing this disease. Increased awareness has resulted in efforts to improve diagnostic tests for this pathogen. The in vitro diagnostic tests on the market today detect three types of analytes: 1) tcdB and/or tcdA (the genes for toxins A and B); 2) glutamate dehydrogenase (GDH), a metabolic enzyme that is produced in high amounts by actively growing (i.e., vegetative) C. difficile cells; and 3) toxins A and B, which cause the disease.
Nucleic acid amplification testing (NAAT) for tcdB and/or tcdA uses PCR or isothermal amplification and is very sensitive for detecting toxigenic C. difficile in fecal specimens. To produce toxin, strains of C. difficile must have toxin genes and must express them. NAAT tests do not detect toxin, nor do they demonstrate that toxin is being produced and is present in the specimen. Since roughly half of the patients colonized with C. difficile are not colonized with strains producing toxin, this high sensitivity for the toxin gene(s) results in the overdiagnosis of CDI in hospitals, leading to treatment, including antibiotic usage, of patients who do not need treatment.
Additionally, some of these tests detect extremely low numbers of C. difficile in the specimen, numbers that are not clinically relevant. Even if patients only have spores in their intestine, without actively growing cells and therefore without active disease, NAAT tests will give positive results. The ability of C. difficile to live transiently in the patient without causing problems and the fact that it forms very hardy spores that persist in host and hospital environments are reasons why NAAT testing as a standalone test leads to overdiagnosis and low positive predictive values.
GDH is an accurate indicator of actively growing cells and is produced by toxigenic and nontoxigenic cells. Nontoxigenic strains usually are present only at low incidence rates. In many hospitals, they are negligible. Studies have shown that nontoxigenic strains were not present in some institutions. In others, nontoxigenic strains represented less than 10 percent of GDH-positive fecal specimens. Some other organisms that live in the human intestine produce an immunologically related GDH, so optimal performance requires GDH testing performed with immunoassays that have highly specific antibodies for GDH from C. difficile. All ribotypes of C. difficile produce identical GDH molecules except for the GDH of one ribotype that has a single amino acid substitution. This substitution has no effect on detection in GDH assays.1 The positive predictive value using GDH as the biomarker is comparable to that observed with NAAT testing and delivers this performance more cost-effectively.
Toxins A and B are virulence factors that cause disease. The detection of toxin indicates the presence of actively growing toxigenic cells. In symptomatic patients, the detection of toxin correlates closely with true CDI. Over the past five years, studies looking at assay performance have taken a critical look at test results in conjunction with patient history. The analyses have demonstrated strongly that patients who have the presence of toxin have more severe disease, longer bouts of diarrhea, longer hospital stays, higher mortality rates, and more inflammation than patients who were positive by NAAT testing but negative for toxin. In fact, there is no clinical difference between patients who are positive for C. difficile but negative for the presence of toxin and patients who are negative for C. difficile overall.
C. difficile disease often has a primary inflammatory component, noted by increased white cell counts and inflammatory biomarkers such as lactoferrin, and many experts believe it should be considered an inflammatory disease. The inflammation is triggered by the extensive tissue damage caused by the toxins and their chemotactic activity. Although guidelines have not discussed the value of inflammatory biomarkers for C. difficile, studies have shown that these markers help identify severity in CDI.2,3
Most toxin immunoassays detect both toxins. This is important because toxin B-only strains, although uncommon, can cause disease. No toxin A-only strains have been reported. The standard for toxin detection is the tissue culture assay, also referred to as the cell cytotoxicity neutralization assay (CCNA). CCNA offers exquisite sensitivity and detects picogram amounts of toxin due to the high activity of toxin B. However, the assay is tedious and time-consuming. A positive result, noted as cell rounding that is neutralized by specific C. difficile antitoxin, can be noted within eight to 12 hours in samples with medium to high levels of toxin. However, it takes 48 hours to call a specimen negative or to identify samples that contain only very low amounts of toxin. Toxin enzyme immunoassays offer the advantages of ease of use and rapid turnaround time (one hour or less), but they are less sensitive than the tissue culture assay. Trying to develop toxin immunoassays that approach the sensitivity of CCNA is challenging but critical, since toxin testing provides positive predictive values that are considerably higher than those observed with NAAT or GDH testing.
Recommended testing algorithms
For the reasons described above, multiple analyte algorithms rather than single analyte assays have been recommended by two societies well versed in C. difficile testing. The European Society of Clinical Microbiology and Infectious Disease (ESCMID) guidelines that became available two years ago recommended an algorithm approach for CDI.4 Optimal results were obtained using a multistep algorithm. NAAT or GDH tests were recommended as a first step since their performance is comparable, followed by a high-performing toxin assay. This approach offers the most accurate diagnostic strategy for CDI, based on results obtained from large clinical studies in which test results were correlated with patient clinical findings.5,6 Symptomatic patients who have diarrhea from another cause but who are positive by a NAAT or GDH test and negative for toxin are defined as carriers. Symptomatic patients who are positive by a NAAT or GDH test and positive for toxin are identified as patients with true CDI.
The recent guidelines from the Infectious Diseases Society of America (IDSA) and the Society for Healthcare Epidemiology of America (SHEA) similarly support an algorithm approach, although the guidelines include additional information on using a standalone molecular test when the institution had a specimen selection guide in place.7 Although the performance of all testing methods improves from testing the correct samples, algorithm testing does not need to depend upon this prerequisite enhancement for accurate results. As with the ESCMID guidelines, the IDSA/SHEA guidelines state that an algorithm approach that incorporates toxin testing results in the highest predictive positive value of true CDI. In addition, both guidelines note that a negative NAAT or GDH test accurately rules out CDI.
There are a large number of NAAT and immunoassay tests available to the clinical lab. Not all of these tests perform equally well, a point repeatedly noted in the IDSA/SHEA guidelines. The quality of reagents that go into these tests varies. There are differences in purity, specificity, and affinity for C. difficile analytes. Some immunoassay tests use polyclonal antibodies, whereas others use monoclonal antibodies, each with their own advantages and disadvantages for testing fecal specimens.
In this context it is worth noting that fecal specimens are challenging because they are among the most complex clinical samples. Variation is due to bile content, bacterial/host degradative enzymes, diet, and a number of other host factors. Testing fecal specimens is therefore much more complex than testing serum samples, and the development of diagnostic reagents, diluents, and procedures optimized for fecal samples is imperative. Something as simple as mixing a specimen properly can be crucial because in feces, the analyte is not evenly distributed throughout the matrix. Specimens also often contain amplification inhibitors, affecting the performance of NAAT tests. NAAT tests or immunoassays that perform well with “neat” reagents can be totally ineffective with fecal specimens. Therefore, quality and demonstrated performance are critical factors when
choosing a C. difficile test.
The multistep approach
Although there are a large number of tests and formats available, adherence to a multistep algorithm clearly seems to be the best approach. According to recent surveys, it appears that roughly half of the responding labs in the U.S. use NAAT testing as a standalone test or as part of a multistep algorithm. The other half use immunoassays, with more than half of that group using an algorithm approach consisting of GDH and toxin testing. Additional surveys are needed to determine more accurately the current prevalence and adoption of multistep algorithm testing in U.S. labs. An algorithm that incorporates a screening assay (NAAT or GDH) followed by a toxin test with demonstrated performance will provide a higher positive predictive value than standalone testing. The testing can be made more cost-effective by using GDH as the initial screen, since it is comparable in performance to NAAT testing. An algorithm approach is supported by the ESCMID and IDSA/SHEA guidelines and, when paired with the clinical history of the patient, provides the most accurate assessment of true C. difficile infection, resulting in optimal patient care.
- Carman RJ, Wickham KN, Chen L, et al. Glutamate dehydrogenase is highly conserved among Clostridium difficile ribotypes. J Clin Microbiol. 2012;50(4):1425-1426.
- Boone JH, Archbald-Pannone LR, Wickham KN, et al. Ribotype 027 Clostridium difficile infections with measurable stool toxin have increased lactoferrin and are associated with a higher mortality. Eur J Clin Microbiol Infect Dis. 2014;33(6):1045-1051.
- Boone JH, DiPersio JR, Tan MJ, et al. Elevated lactoferrin is associated with moderate to severe Clostridium difficile disease, stool toxin, and 027 infection. Eur J Clin Microbiol Infect Dis. 2013;32(12):1517-1523.
- Crobach MT, Planche C, Eckert F, et al. European Society of Clinical Microbiology and Infectious Diseases: update of the diagnostic guidance document for Clostridium difficile Clin Microbiol Infect. 2016;Aug;22 Suppl 4:S63-81.
- Planche TD, Davies KA, Coen PG, et al. Differences in outcome according to Clostridium difficile testing method: a prospective multicenter diagnostic validation study of difficile infection. The Lancet. 2013;13(11):936-945.
- Polage CR, Gyorke CE, Kennedy MA, et al. Overdiagnosis of Clostridium difficile infection in the molecular test era. JAMA Intern Med. 2015;175(11):1792-80.
- Tschudin-Sutter S, Kuijper EJ, Durovic A, et al. Guidance document for prevention of Clostridium difficile infection in acute healthcare settings. Clin Microbioland Infect. 2018;Mar 2. pii: S1198-743X(18)30195-2.