Bayesian Assessment of Accuracy Properties of Rose Bengal Test in REV-1 Vaccinated Small Ruminants

KUJTIM MERSINAJ1*, ARLA JUMA2, LULËZIME HAXHA2, XHELIL KOLECI3

1National Veterinary Epidemiology Unit, Food Safety and Veterinary Institute, Tirana, Albania

2Animal Health Department – Food Safety and Veterinary Institute, Tirana, Albania

3Veterinary Public Health Department, Veterinary Medicine Faculty, Agriculture University of Tirana, Albania

AbstractThe live Brucella melitensis Rev-1 strain is considered the best vaccine available for the prevention of brucellosis in small ruminants.  In Albania, vaccination of small ruminants with REV-1 has been used under various regimes since 2004. The effectiveness of the vaccination campaigns is monitored through a post vaccination monitoring and surveillance system based on testing vaccinated animals with Rose Bengal Plate Test (RBPT). Regrettably the test accuracy properties such as the sensitivity and specificity have not been validated in vaccinated small ruminants. The lack of knowledge on these properties hampers the correct evaluation of the true sero-conversion rate of the post vaccination monitoring at national and at flock scale. This study addressed this issue by using a Bayesian modelling framework to estimate two serological tests the RBPT which is the standard serological test used in the post vaccination monitoring and Complement Fixation Test (CFT). Serum samples from 191 reportedly vaccinated small ruminants were tested in parallel with RBPT and CFT. The estimates of sensitivity and specificity values of RBPT were 91% (95% CrI: 82 -98) and 89% (95% CrI: 70 – 98). For CFT the sensitivity resulted 86% (95% CrI: 70 – 95) and the specificity 95% (95% CrI: 80 – 99). The good sensitivity and acceptable specificity of RBPT support its utilization as screening test for post vaccination monitoring. The interpretation of the post vaccination data with RBPT are well acceptable at national level but not well suited for flock status interpretation. In terms of disease diagnosis, especially for a latter phase of the brucellosis control strategy in Albania, both tests could justify their use in association. Keywords: Brucellosis, Small Ruminants, Bayesian modelling, Screening test, Sensitivity, Specificity

Post Author: MVhDC39RHa