Epidemiology, Diagnosis and Treatment of Clostridium Difficile Infection

Matteo Bassetti; Giovanni Villa; Davide Pecori; Alessandra Arzese; Mark Wilcox


Expert Rev Anti Infect Ther. 2012;10(12):1405-1423. 

In This Article

C. difficile Toxin A & B Detection by EIA

For more than 20 years the detection of toxin A/B from fecal samples by immunoenzymatic methods has been the cornerstone of laboratory CDI diagnosis. Toxin enzyme immunoassays (EIAs) are more rapid and easier to perform than culture or cytotoxin assays, they do not require special equipments or a dedicated technician, and the turnaround time allows the laboratory to refer results within the same day of specimen arrival. However, the performance of immunoenzymatic kits for C. difficile toxin(s) detection available on the market versus gold-standard methods is currently considered suboptimal (both in terms of sensitivity and specificity) if used as standalone assays. Commercial EIA systems produced by several different manufacturers generally show high specificity (~95%), but low to moderate sensitivity, being in the 60–90% range. Importantly, the positive predictive values and negative predictive values of all of the kits vary depending on the prevalence of the condition being detected. In particular, it has been estimated that if the prevalence of C. difficile toxins in stool is relatively low (<10%) the positive-predictive value of these assays is generally too low (<50%) to be acceptable for a reliable diagnosis and clinical management (Table 2).[91]

Therefore, a two-step diagnostic algorithm with a rapid highly sensitive screening method to identify positive samples, to be confirmed by gold standard, should be strongly recommended. Some laboratories routinely test stool specimens for toxin A/B by immunoassay methods and in selected cases perform stool culture including isolate toxin testing (TOXT) or send troublesome specimens to reference laboratory for CCNA confirmatory testing.

In order to be adequate to clinical needs most diagnostic tests were designed to detect biologically active forms of C. difficile toxin A and toxin B, as most CDI-associated strains were thought to produce both of them. Samples from patients with strains showing PaLoc deletion would be missed by diagnostic methods based on detection of toxin A only, as reported in the literature during an outbreak of CDI to a toxin A-/toxin B+ strain of C. difficile,[92] and also from studies of CDI in children.[93] In addition, a clinical C. difficile isolate coding for aberrant toxin B and a biologically active toxin A has been described.[94] These findings show that there are different C. difficile toxin types; strains that produce toxin A but not B have not been confirmed; those that produce toxin B but not A are capable of infecting and causing disease in animals and humans. In conclusion, on the basis of available data for diagnostic purposes the most reliable toxin detection system should target both toxins A and B.