Comparison of diagnostic sensitivity and specificity of seven Cryptosporidium assays used in the UK

Abstract

To compare the diagnostic sensitivity and specificity of seven Cryptosporidium diagnostic assays used in the UK, results from 259 stool samples from patients with acute gastrointestinal symptoms were compared against a nominated gold standard (real-time PCR and oocyst detection). Of the 152 'true positives', 80 were Cryptosporidium hominis, 68 Cryptosporidium parvum, two Cryptosporidium felis, one Cryptosporidium ubiquitum and one Cryptosporidium meleagridis. The Cryptosporidium spp. diagnostic sensitivities of three Cryptosporidium and Giardia combination enzyme immunoassays (EIA) coupled with confirmation of positive reactions were 91.4-93.4 %, whilst the sensitivity of auramine phenol microscopy was 92.1 % and that of immunofluorescence microscopy (IFM) was 97.4 %, all with overlapping 95 % confidence intervals. However, IFM was significantly more sensitive (P = 0.01, paired test of proportions). The sensitivity of modified Ziehl-Neelsen microscopy was 75.4 %, significantly lower than those for the other tests investigated, including an immunochromatographic lateral flow assay (ICLF) (84.9 %) (P = 0.0016). Specificities were 100 % when the ICLF and EIA test algorithms included confirmation of positive reactions; however, four positive EIA reactions were not confirmed for either parasite. There was no significant difference in the detection of C. parvum and C. hominis by each assay, but the detection of other Cryptosporidium spp. requires further investigation, as the numbers of samples were small. EIAs may be considered for diagnostic testing, subject to local validation, and diagnostic algorithms must include confirmation of positive reactions.