Comparative evaluation of commercially available point-of-care heartworm antigen tests using well-characterized canine plasma samples

Abstract

Background: Dirofilaria immitis is a worldwide parasite that is endemic in many parts of the United States. There are many commercial assays available for the detection of D. immitis antigen, one of which was modified and has reentered the market. Our objective was to compare the recently reintroduced Witness® Heartworm (HW) Antigen test Kit (Zoetis, Florham Park, NJ) and the SNAP® Heartworm RT (IDEXX Laboratories, Inc., Westbrook, ME) to the well-based ELISA DiroChek® Heartworm Antigen Test Kit (Zoetis, Florham Park, NJ). Methods: Canine plasma samples were either received at the Auburn Diagnostic Parasitology Laboratory from veterinarians submitting samples for additional heartworm testing (n = 100) from 2008 to 2016 or purchased from purpose-bred beagles (n = 50, presumed negative) in 2016. Samples were categorized as "positive," "borderline" or "negative" using our established spectrophotometric cutoff value with the DiroChek® assay when a sample was initially received and processed. Three commercially available heartworm antigen tests (DiroChek®, Witness® HW, and SNAP® RT) were utilized for simultaneous testing of the 150 samples in random order as per their package insert with the addition of spectrophotometric optical density (OD) readings of the DiroChek® assay. Any samples yielding discordant test results between assays were further evaluated by heat treatment of plasma and retesting. Chi-square tests for the equality of proportions were utilized for statistical analyses. Results: Concordant results occurred in 140/150 (93.3%) samples. Discrepant results occurred in 10/150 samples tested (6.6%): 9/10 occurring in the borderline heartworm (HW) category and 1/10 occurring in the negative HW category. The sensitivity and specificity of each test compared to the DiroChek® read by spectrophotometer was similar to what has been reported previously (Witness®: sensitivity 97.0% [94.1-99.4%], specificity 96.4% [95.5-100.0%]; SNAP® RT: sensitivity 90.9% [78.0-100.0%], specificity 98.8% [96.0-100.0%]). There were significant differences detected when comparing the sensitivities of the SNAP® RT and the Witness® HW to the DiroChek® among the 150 total samples (p = 0.003) and the 50 "borderline" samples (p = 0.001). Conclusions: In this study, the sensitivity of the Witness® HW was higher than the sensitivity of the SNAP® RT when compared with the DiroChek® test results prior to heat treatment of samples.