Performance Analysis of Anaplasma Antibody Competitive ELISA Using the ROC Curve for Screening of Anaplasmosis in Camel Populations in Egypt

Abstract

Anaplasmosis is a tick-born and potential zoonotic disease caused by Anaplasma (A.) phagocytophilum, A. ovis, A. platys and A. capra. Anaplasma marginale affecting bovines and camels causing significant economic losses. Camels as an integral part of the socio-economic lifestyle of nomads in semi-arid to arid ecosystems are prone to suffer from subclinical Anaplasma infections. This study aimed to determine the performance and adaptation of commercial competitive Anaplasma ELISA (cELISA) as a tool for screening the seroprevalence of anaplasmosis whitin the camel populations in Egypt. This study was based on the serological investigation of 437 camel sera collected between 2015 and 2016 during a Q fever prevalence study in Egypt using commercially available cELISA for the detection of antibodies specific for Anaplasma in bovine serum. The receiver operating characteristic (ROC) curve, an analysis method for optimizing cutoff values in cELISAs, was used to estimate the sensitivity and specificity using 76 true as serological positive (n = 7) and negative (n = 60) for Anaplasma antibodies. ROC curve analysis was done for 7 true positive and 60 true negative bovine samples and 7 true positive and 29 true negative camel samples serum. Real time PCR and/or conventional PCR was applied to confirm Anaplasma spp. specific-DNA in camel serum as an indication of a true positive and true negative for ROC analysis. Chi square analysis was performed to estimate the association between risk factors and anaplasmosis in camels. The cutoff value was determined as 0.42 (p value ≤ 0.001). Data simulation with randomly generated values revealed a cutoff value of 0.417 (p ≤ 0.001) with resulting 58.1% Se and 97.8% Sp. Seven true positive and 29 true negative camel serum samples was confirmed by PCR. Using the estimated cut off, the seroprevalence in the Nile Valley and Delta and the Eastern Desert domain was 47.4% and 46.4%, respectively. The potential risk factors as domains and origin of animals were less significantly associated with the prevalence of anaplasmosis (domains: χ(2) = 41.8, p value ≤ 0.001 and origin: χ(2) = 42.56, p value ≤ 0.001). Raising awareness especially for veterinarians and animal owners will significantly contribute to the best understanding of anaplasmosis in camels in Egypt. Alternative (in silico) validation techniques and preliminary prevalence studies are mandatory towards the control of neglected anaplasmosis in the camel population.