Validation of a Commercial Enzyme Immunoassay for Detection of the Presence of Clostridium difficile Toxins in Feces of Horses with Acute Diarrhea
Clostridium difficile has been implicated as the causative agent of acute diarrhea in adult horses and foals. A fast and accurate diagnosis of C. difficile-associated disease (CDAD) is of the outmost importance for an adequate therapeutic intervention. While the cell cytotoxicity assay (CTA) is considered the "Gold Standard", it is neither readily available nor conductive to routine testing, so enzyme immunoassays are typically used. The C. DIFFICILE TOX A/B IITM ELISA (TECHLAB®, Inc; 2001 Kraft Drive, Blacksburg, VA 24060-6358) has been validated in humans, but has been reported to have low sensitivity and specificity in dogs. While this test is widely used in horses, it performance on equine feces in unclear.
The objective of the present study was to determine the performance of the C. DIFFICILE TOX A/B IITM ELISA, and validate its use as a diagnostic tool for the detection of Clostridium difficile toxins in feces of horses with acute diarrhea.
Fecal samples were collected and processed prospectively from hospitalized horses with acute diarrhea for which the treating clinician requested C. difficile toxin testing. Samples were stored at 4°C within 2 hours of collection, and processed within the following 2 weeks. The ELISA was performed in parallel with the CTA.
Out of a total of 72 fecal samples tested, 19 (26.4%) were positive using the CTA. Sixteen (84.2%) CTA positive samples were positive on the ELISA, and 2 (3.8%) CTA negative results were positive on the ELISA. No significant difference was observed between the ELISA test and the "gold standard" using the McNemar's test [OR = 0.66 (95% confidence interval (CI) = 0.08, 4.28); 2-tailed p = 1.00], indicating that the probability of the ELISA being positive or negative was the same as that of the gold standard. The ELISA had a sensitivity of 84% [(CI = 61, 96%); Median Unbiased Estimates (MUE) = 84%], a specificity of 96% [(CI = 87, 99%); MUE = 96%] and positive and negative predictive values of 89% [(CI = 67, 98%); MUE = 88%] and 94% [(CI = 85, 94%); MUE = 96%], respectively. A good agreement between the ELISA and the CTA was observed [Kappa = 0.82 (CI = 0.67, 0.97) (p-value under Ho of no agreement between tests is <0.0001)].
In conclusion, the overall performance of the test was adequate, being similar to that reported for human samples and superior to that reported for canine samples. The high level of agreement between the ELISA and the CTA indicates that the C. DIFFICILE TOX A/B IITM ELISA test is an adequate diagnostic tool and a reliable, practical test for the clinical diagnosis of CDAD in horses.