American Association for Clinical Chemistry
Better health through laboratory medicine
July 2010 Clinical Laboratory News: Clostridium difficile

CLN Banner Logo

July 2010: Volume 36, Number 7


Clostridium difficile
Navigating the Testing Options for Diagnosis

By Glen Hansen, PhD, Stephen Blatt, MD, Stephen M. Brecher, PhD, Erik Dubberke, MD, and Preston Dorsett, PhD


Used with permission of University of Washington Medicine, Department of Laboratory Medicine

Clostridium difficile is a major cause of healthcare–associated infections, the prevalence and severity of which has dramatically increased since the early 2000s when a surge in morbidity and mortality rates occurred. This gram-positive, anaerobic bacillis produces two exotoxins and accounts for 15–25% of all episodes of antibiotic-associated diarrhea. Correct diagnosis of C. difficile infections (CDI) requires both clinical insight and appropriate laboratory testing and is important because numerous pathogens can cause diarrhea and other symptoms of enteric disease. Furthermore, accurate identification of patients with and without CDI is necessary for initiating effective and appropriate patient management, as well as maximizing the positive- and negative-predictive values for C. difficile laboratory tests.

While greater insight into CDI has evolved in the last few years, it remains an area of intense debate and differing clinical viewpoints, even among the authors of this article. Today, a variety of tests are available for detecting C. difficile, but no single testing approach has proven effective for the needs of all patients, laboratories, and hospital settings. Despite the debate over a true, universal testing strategy, here we consider the advantages and limitations of the various diagnostic tests within the profile of various institutions. We also offer discussion and rationale for tailoring testing methods to a given institution.

Clinical Background

In the mid-1930s, scientists first described C. difficile, a gram-positive, anaerobic spore-forming, toxin-producing bacillus, as a component of the intestinal microflora in neonates (1). But it was not until the 1970s that C. difficile was established as a causative agent of pseudomembranous colitis (1). Historically, C. difficile carriage rates in healthy adults have been 1–3%. In addition, up to 50% of individuals with exposure to an inpatient facility may be asymptomatic carriers for C. difficile.

Today, C. difficile is the leading cause of infectious diarrhea in hospitals in the developed world, including up to 20% of reported antibiotic-associated diarrhea and nearly all incidences of pseudomembranous colitis. In the past decade, CDI has become increasingly prevalent and severe globally, especially in patients older than 65 years, among whom the rate of disease is the highest (2). There have also been reports of infections among populations previously thought to be low risk for CDI, including children, pregnant women, and otherwise healthy, non-hospitalized individuals without prior exposure to antibiotic treatment.

Given the increase in CDI, the economic burden of the disease has surged. In the 1990s, it was estimated that hospitals in the U.S. spent in excess of $1 billion annually to treat CDI. Recent projections indicate that the direct cost for managing C. difficile is approaching $3.4 billion per year, not including treatment for serious complications of the disease or lost opportunity costs (3).

Symptoms of CDI, which include fever, abdominal pain, and distension, watery or loose stools more than three times per day, and leukocytosis, typically present 48–72 hours after antibiotic therapy is administered. However, CDI symptomatology may be delayed for as long as 2–3 months after discontinuation of an antimicrobial agent. The severity of CDI may range from mild, uncomplicated diarrhea to moderate with fever, profuse diarrhea, abdominal pain, and leukocytosis. In extreme instances, patients with severe disease have toxic megacolon, dehydration, and sepsis (Table 1).

Table 1
Terminology Associated with Clostridium difficile
Term
Definition
Antibiotic-associated diarrhea

Diarrhea that develops in an individual who is currently taking or has recently taken antibiotics (C. difficile is a cause of this type of diarrhea)

Symptoms include watery diarrhea and abdominal cramping

Asymptomatic colonization/carriage Patient is colonized with C. difficile without signs or symptoms of CDI
CDI

Presence of diarrhea characterized by >3 watery stools per day

Other symptoms can include fever, abdominal pain, cramping, nausea, and loss of appetite

Typically presents in high risk patients (> 65 years, immunocompromised, or severe underlying disease) with exposure to antibiotics

Pseudomembranous colitis

Presence of plaque formations on colon membranes

Considered pathognomonic for CDI in the appropriate clinical setting

Toxic megacolon

Extreme inflammation and distention of the colon often resulting from a severe episode of colitis

Symptoms include abdominal distension and pain, fever, and dehydration

Abbreviation: CDI = Clostridium difficile infection (interchangeable with C. difficile-associated diarrhea).

Diagnosis of CDI

Early diagnosis and appropriate treatment of CDI are important for preventing complications of the infection and decreasing transmission. Clinicians typically rely upon three factors to make a diagnosis of CDI: patient risk profile; symptoms; and analysis of laboratory data. In general, hospitalized patients suffering from symptoms of diarrhea and/or abdominal pain should be tested for CDI only if there is clinically significant diarrhea, defined as more than three watery stools per day; however, patients with recent antibiotic exposures and those in high-risk groups, including patients >65 years, immunocompromised patients, or those with severe underlying disease should be especially targeted for testing. In the community setting, prior antibiotic exposure is less common, particularly if the patient has not been hospitalized in the last 90 days. Testing for CDI should be considered in these patients if the diarrhea is severe or persists longer than 3 days, even in the absence of recent antibiotic exposures.

Although clinicians and hospital staff are well aware of the prevalence of CDI, they may not understand the full implications of laboratory data and what constitutes appropriate testing. Most clinicians understand that there can be a significant level of asymptomatic carriage of C. difficile toxigenic strains, especially in patients residing in long-term care facilities. However, not all heathcare providers realize that asymptomatic carriage of toxigenic C. difficile also has implications in nosocomial transmission of the infection, which can make a definitive diagnosis of CDI difficult. For example, in a landmark paper published in 1989, researchers determined that 21% of hospitalized patients in the study acquired C. difficile via nosocomial transmission, but in 63% of those patients, the recovery of C. difficile was deemed asymptomatic (4).

Use of amplification methods that increase sensitivity such as toxigenic culture or molecular methods should be balanced by measures that maximize positive predictive values of laboratory testing. One such step may include testing non-solid stool specimens, or specimens that take the shape of the container, submitted under appropriate clinical scenarios, as determined with an appropriate clinical history and physical exam. In addition, studies are needed to determine the clinical correlation between toxigenic culture or molecular assay results as well as the potential role of antibody responses to C. difficile toxins. This type of data will further assist healthcare providers in utilizing the outcomes of sensitive testing methods.

Although laboratory practices may vary, clinicians should be wary of the so-called “C. difficile times three testing”, or submission of three specimens with each testing request, as research indicates that positive predictive value decreases significantly with repeat testing (5). Tiered testing, like that proposed in the tailored algorithms discussed below, has been proposed as an alternative solution to optimize sensitivity, specificity, and turn-around times which may be used in place of submitting multiple specimens for the same test (6). Institutions should discourage the practice of testing to detect carrier status in individual patients as well as “test of cures”, as this has not been shown to affect reinfection rates nor provide clinically useful information. Although detecting C. difficile and its toxins is critical to support the diagnosis of CDI, laboratory tests do not make a diagnosis; they help the clinician to confirm the diagnosis.

Laboratory Testing of CDI

A variety of laboratory and non-laboratory tests are available to diagnose CDI. Laboratory tests are geared toward demonstrating the presence of C. difficile either through detecting the toxin via a cytotoxicity neutralization assay (CTN) or toxin enzyme immunoassay (EIA) or by detecting the organism via a toxigenic culture (TC) or glutamate dehydrogenase (GDH) assay. There are also molecular methods to detect the genetic components of toxin production. Non-laboratory tests such as endoscopy and computed axial tomography are used to evaluate the biologic effects of the organism and toxin. Nonspecific laboratory tests for biologic markers of CDI include examination of lactoferrin and stool leukocytes.

Researchers first described the EIA for detection of toxin A in 1983; a monoclonal antibody for toxin A was developed soon after. These initial discoveries set the future standard for CDI testing. Although not as sensitive as CTN or TC, EIA proved to be less labor intensive and decreased turn-around time for results.

While most laboratories use an EIA to detect C. difficile, there has been considerable debate about which testing method constitutes the true laboratory gold standard. Noteworthy is the fact that the Food and Drug Administration (FDA) has historically chosen CTN as the standard for comparison when clearing diagnostic tests for C. difficile. Table 2 outlines the specifics of TC and the CTN assays.

Table 2
Comparison of Gold Standard Assays for Clostridium difficile
 
Cytotoxin neutralization assay
Toxigenic culture
Specifics of the assay

Cell monolayers are exposed to stool filtrates and cells are observed for a cytopathic effect

The specificity of the effect is tested by neutralizing with an antisera

Stool may be pretreated with heat or ethanol

Stool is cultured on pre-reduced selective medium

Colonies of high suspicion are isolated and transferred to broth culture

The supernatant is transferred to cell culture or EIA to detect the presence of toxin

Requirements

Knowledge of cell culturing techniques

Supply/maintain cultured cell monolayers

Knowledge of culturing techniques

Supply/maintain cultured cell monolayers

Drawbacks

Cost

Results can vary depending on the cell line applied, dilution factors, reagents used, storage conditions, and operator expertise

Turnaround time is slow; up to 2 days to demonstrate cytotoxicity

Labor intensive

Cost

Turnaround time is slow; up to 2 days to culture stool and an additional 2 days to determine if isolates are toxigenic (see CTN drawbacks)

Standardization not available

Labor intensive

Confusion about which assay is the gold standard has been driven, in part, by the variation in sensitivity, specificity, and negative- and positive-predictive values that researchers report when comparing GDH, EIA, and molecular assays. Differences in the technical competence and diligence of individual investigators and use of disparate methodologies and growth media also influence the sensitivity of a given assay for detecting C. difficile and toxigenic C. difficile as reported in the literature. Such factors can further contribute to the confusion among clinicians and laboratorians when it comes to understanding the implications of laboratory testing of CDI.

As a result of the differences in CDI test methodology and study design, some investigators have raised concerns about the true accuracy of laboratory testing for CDI. For example, researchers using different case criteria have found that the sensitivity of the CTN assay is substantially lower than traditionally reported (6). Taken collectively, such arguments remind us that the sensitivity of any assay is a direct function of the sensitivity of the gold standard to which it is compared. Table 3 provides an overview of available laboratory CDI diagnostics and highlights the advantages and limitations of the diagnostic methods CTN, TC, EIA, GDH, and the newer molecular assays.

Table 3
Testing Methods for Diagnosing CDI and Associated Characteristics
Parameter
Gold Standard Tests
CTN
TC
Sensitivity (%)
86–97
100
Specificity (%)
97–100
93–96
Negative-predictive value (%)
94–99
100
Positive predictive value* (%)
82–100
68
Cost per assay
$20–30
<$25
Turnaround time (h)
24–48
≥96
Advantages Distinguishes toxin-producing strains Highly sensitive technique
Limitations Labor intensive; requires special equipment; relatively slow technique Labor intensive; requires special materials; slow technique; uncertain specificity for infection
Parameter
Other Tests
EIA
(microwell)
EIA
(rapid card)
GDH
Molecular
24, 32, 36, 40
Sensitivity (%)
67–97
78–95
76–94
93–100
Specificity (%)
91–99
93–100
97–100
94–98
Negative-predictive value (%)
96–99
96–98
95–99
97–100
Positive predictive value* (%)
49–100
55–99
59–100
63–90
Cost per assay
<$10
<$20
<$10
$40–50
Turnaround time (h)
<2
<1
<1
1–<4
Advantages Rapid and simple technique Rapid technique; can be used as screening tool Detects the presence of regions of pathogenic locus
Limitations Sensitivity can vary between manufacturers and laboratories Toxin testing required to verify diagnosis Requires special equipment and specific expertise; uncertain specificity for infection

Abbreviations: CDI = Clostridium difficile infection; CTN = cytotoxicity neutralization assay; TC = toxigenic culture; EIA = enzyme immunoassay; GDH = glutamate dehydrogenase.

*Positive predictive values assume patient has a C. difficile-positive test result.

Selecting a C. difficile Assay or Algorithm

Between the variety of testing methods available and the confusion about the gold standard method, laboratorians may find it challenging to determine the best testing method(s) for their institutions. Consequently, many institutions have developed their own algorithms to optimize patient diagnosis and outcomes, as well as improve laboratory resource utilization. Recently, researchers have reported two-step algorithms that use GDH as an initial screening test followed by either CTN (8) or EIA (9). New data on the sensitivity of GDH, however, has raised concern over use of this method as the initial screening test (10). Another two-step algorithm using EIA as a first-line test followed by a polymerase chain reaction (PCR) assay for samples with positive test results in the low-positive to high-negative range has also been described (11). Finally, researchers have also reported a three-step algorithm that eliminates the CTN assay, which is particularly helpful for laboratories that do not have the facilities or technical expertise to perform this assay (12).

Figure 1
Sample Multistep Algorithm for Diagnosing CDI

Click for figure

  • GDH has a good negative-predictive value and may be useful as an inexpensive screening tool in settings where high volumes of clinically negative patients are expected.
  • Patients considered to be of low clinical suspicion should be monitored closely and retested if diarrhea and/or symptoms persist or worsen.
  • First-line molecular testing is an option for laboratories with appropriate resources.

Abbreviations: CDI = Clostridium difficile infection; GDH = glutamate dehydrogenase; EIA = enzyme immunoassay.

Table 4 (below) lists the factors that laboratorians should consider when selecting a CDI testing method. Based on these types of factors, several laboratories have developed algorithms to achieve the best balance of cost, sensitivity, and specificity of testing modalities against available resources, testing patterns, and the patient population being tested. These algorithms can be modified to meet specific institutional needs. Overall, the selected tests should attempt to balance the parameters of sensitivity, specificity, and positive- and negative-predictive value. Often, a high negative-predictive value (>99%) is an important parameter when selecting a first-line test. Figure 1 (above) shows a sample algorithm and is one of the many potential options for a laboratory to adopt when considering tiered-testing methods.

Table 4
Parameters to Consider in Selecting a Diagnostic Test and Developing a Tailored Algorithm for CDI

  • Impact of C. difficile missed cases on patient care and institution
  • Identification of patient population (low vs. high suspicion)
  • Ordering of test is facilitated by in-house, referral, or specialty department
  • Incidence of CDI at institution
  • Local expertise
  • Resource availability (budget, staffing, and hours of operation)
  • Volume of tests requested
  • Experience and satisfaction with current testing methods
  • Performance range of current testing methods
  • Demand for retests and/or reflex testing
  • Current access to other diagnostic methods (culture and molecular assays) for reflex testing
  • Reporting requirements

Abbreviation: CDI = Clostridium difficile infection.

The Essential Role of the Lab: Communication

Although laboratory tests for CDI are extremely important, they should be used in conjunction with a careful review of the patient’s clinical information. Key clinical symptoms include moderate to severe diarrhea, leukocytosis, low albumin (<2.5 g/dL), and increasing creatinine concentration (double baseline). Severe disease is often associated with a dramatic increase in white blood cell count.

Clinical symptoms are as important as test results. For example, a negative EIA or GDH result in a patient with both significant risk factors such as antibiotic usage and/or being >65 years old and clinical symptoms of CDI should be evaluated carefully. Irrespective of the testing modalities used, good communication between the laboratory and medical staff is important for optimal diagnosis of CDI.

Despite adherence to and use of criteria designed to maximize resulting, exceptional cases can occur. We, as well as others, have seen CDI patients who have negative results on direct-stool EIA testing progress to pancolitis or even death. Consequently, clinicians must recognize that EIA testing may be subject to variables that can reduce sensitivity in some specimens submitted to the laboratory. In such cases, molecular testing may be warranted to identify specific genes involved in C. difficile toxin production. When the clinical findings suggest CDI is likely, antibiotic treatment and infection control protocols should be initiated. Because false-positive and false-negative test results are common, clinical and laboratory data must be reviewed together for optimal diagnosis.

A further challenge to the diagnosis of CDI is the expanded role that testing will have in controlling C. difficile in outpatient populations. Reports show significant increases in CDI among community-based patients. While there have never been doubts that CDI occurs in the community setting, some epidemiologists question whether the incidence of community CDI is increasing or just the result of better detection. Others question whether testing all or some outpatients with loose stools for C. difficile is necessary. Such universal screening would add significant costs but should be considered if clinical symptoms and other laboratory tests make CDI a possibility.

Can we circumvent the laboratory testing issues described here by moving to molecular testing? The answer is maybe. While positive molecular testing results in asymptomatic and colonized patients have been reported, molecular testing in general performs quite well (Table 3). Variable sensitivity between traditional testing modalities such as GDH and EIA, as well as between and within labs, has caused many laboratories to consider molecular testing for C. difficile.

Unfortunately, molecular testing costs significantly more than EIA and sometimes requires dedicated lab personnel. As in much of microbiology and infectious disease testing, recent gains in improved diagnostics have been realized outside the laboratory. This means that the costs of testing, instrumentation, and staffing— particularly for weekend coverage—need to be balanced with detection in otherwise undiagnosed cases, impact on antimicrobial utilization, and role of infection control and transmission through recognition of increased detection. It seems likely that as more molecular tests are developed and costs come down, there will be increased use of these tests in laboratories experiencing challenges with more traditional testing methods.

The Take Home Message

The diagnosis of CDI remains difficult, and there is no one best test that fits everyone’s needs. With new testing methods, epidemic strains, evolving epidemiology and ever growing economic constraints, it is challenging to know which assay to use. EIA alone is often the least expensive option. Two-step algorithms with GDH as the front-end test work well in low-prevalence populations. Molecular tests appear to be more sensitive than non-molecular tests, but the cost is prohibitive for many laboratories. Furthermore, the published literature is, at best, difficult to decipher, particularly with respect to EIA and GDH.

The choice for an individual laboratory should be based on careful evaluation of a selected test in your unique setting. Careful consideration must be given to test performance, cost per test, cost of equipment, ease-of-use, turn-around-time, specimen control, patient population, and prevalence of CDI. The common goal of optimizing patient care and outcome requires the medical community’s careful attention to the details. But for now, these decisions remain “très difficile”.

REFERENCES

  1. Bartlett JG. Historical perspectives on studies of Clostridium difficile and C. difficile infection. Clin Infect Dis 2008;46:S4–S12.
  2. Pepin J, Valiquette L, Alary ME. Clostridium difficile-associated diarrhea in a region of Quebec from 1991 to 2003: a changing pattern of disease severity. Can Med Assoc J 2004;171:466–472.
  3. Dubberke ER, Wertheimer AI. Review of current literature on the economic burden of Clostridium difficile infection. Infect Control Hosp Epidemiol 2009;30:57–66.
  4. McFarland LV, Mulligan ME, Kwok RY, Stamm WE. Nosocomial acquisition of Clostridium difficile infection. N Engl J Med. 1989;320(4):204–10.
  5. Litvin M, Reske KA, Mayfield J, et al. Identification of a pseudo-outbreak of Clostridium difficile infection (CDI) and the effect of repeated testing, sensitivity, and specificity on perceived prevalence of CDI. Infect Control Hosp Epidemiol 2009;30(12):1166–1171.
  6. Peterson LR, Kelly PJ. The role of the clinical microbiology laboratory in management of Clostridium difficile-associated diarrhea. Infect Dis Clin North Am 1993; 7(2) 277–293.
  7. Fang FC, Gerding DN, Peterson LR. Diagnosis of Clostridium difficile Colitis. Annals of Internal Medicine 1996;125(6):515.
  8. Gilligan PH. Is a two-step glutamate dehydrogenase antigen-cytotoxicity neutralization assay algorithm superior to the Premier toxin A and B enzyme immunoassay for laboratory detection of Clostridium difficile? J Clin Microbiol 2008;46:1523–1525.
  9. Fenner L, Widmer AF, Goy G, et al. Rapid and reliable diagnostic algorithm for detection of Clostridium difficile. J Clin Microbiol 2008;46:328–330.
  10. Eastwood K, Else P, Charlett A, et al. Comparison of nine commercially available C. difficile toxin detection assays; a real time PCR assay for C. difficile tcdB and a GSH detection assay, with cytotoxin testing and cytotoxigenic culture. J Clin Microbiol 2009;47:3211–3217.
  11. Brecher SM, O’Brien C, Strymish J et al. Clostridium difficile EIA and PCR testing of EIA high-positives, low-positive, and high-negative stool samples. ICCAC meeting, San Francisco, Calif., Sept. 14, 2009.
  12. Sharp SE, Ivie WM, Buckles MR, et al. A simple 3-step algorithm for improved laboratory detection of Clostridium difficile toxin without the need for tissue culture cytotoxicity neutralization assays. Diagn Microbiol Infect Dis 2009;64:344–346.

Glen Hansen, PhD, is an assistant professor in the Departments of Laboratory Medicine and Pathology and the Department of Medicine Pathology at the University of Minnesota and Director of Clinical Microbiology and Molecular Diagnostics at Hennepin County Medical Center, Minneapolis, Minn. Email: hans2923@umn.edu. (Author to whom correspondence should be addressed.)

Stephen Blatt, MD, is medical director for infectious diseases, TriHealth Infectious Diseases Consultants of Cincinnati Inc., Cincinnati, Ohio.

Stephen M. Brecher, PhD, is the director of the microbiology laboratories of the Boston Veterans Affairs Health Care System, West Roxbury, Mass.

Erik Dubberke, MD, is assistant professor of medicine, Washington University School of Medicine, St Louis, Mo.

Preston Dorsett, PhD, is founder and president emeritus, Meridian Life Sciences, Inc., Memphis, Tenn.

Conflict of Interest Statement
Authors on the paper have all served in advisory roles for Meridian Life Sciences, Inc.