Thirteen studies were included in the review. Reporting deficiencies were noted in terms of patient selection criteria and the spectrum of patients/specimens. Not all studies used the same reference standard. Blinding was uncertain.
Comparison of immunoassays for toxins A and B: Ten studies (16 pairwise comparisons) of seven different immunoassays showed no statistically significant differences between test sensitivities. The authors reported that a lack of precision could not rule out any differences at all. There were no substantial differences for specificities.
Gene detection tests compared with immunoassays for toxins A and B: Four studies (nine direct comparisons) showed that gene detection tests tended to be more sensitive than immunoassays for toxins A and B. The pooled estimate of the difference in sensitivities was statistically significant at 17% (99% CI -3% to 37%; two studies) in one comparison and 25% (99% CI -36% to 86%; two studies) in another. There was no substantial loss of specificity.