A meta-analysis to re-evaluate the sensitivity and specificity of FMDV NSP ELISA tests

Abstract

Foot-and-mouth disease (FMD) serosurveillance is used as one of the disease control instruments. For this goal, it is necessary to differentiate infected animals from the vaccinated ones using non-structural protein (NSP) enzyme-linked immunoassays (ELISA). Various NSP ELISA have been developed. However, their sensitivity and specificity have shown variable results. Therefore, a meta-analysis was conducted to re-evaluate the sensitivity and specificity of these ELISAs. Experimental data were analyzed using R version 4.2.1, employing the rma.mv function in {metafor} package, and the impute covariance_matrix function from the {clubSandwich} package. Commercial kits, the highest sensitivity (0.82, 0.83) and specificity (0.97) were observed in Kit 1 and 3 groups. Differences in performance measures due to animal profiles were not statistically justified. In-house kits, performance measures varied by animal species and NSP protein. Specifically, sensitivity and specificity were lower in pigs (0.62 and 0.81) compared to cattle (0.93 and 0.97) and sheep (0.94 and 0.98) with 3ABC protein. No significant differences were found between sheep and cattle. Additionally, assays using protein 2C showed significantly lower sensitivity and specificity compared to those using protein 3ABC. The highest diagnostic measures were observed in cattle and sheep tested with the 3AB protein, followed by 3B and 3ABC proteins.