The impact of recent tuberculin skin testing on paratuberculosis antibody responses

There is a recognized interference between the single intradermal cervical comparative tuberculin (SICCT) test used for bovine tuberculosis surveillance and the paratuberculosis milk ELISA. Current UK guidance recommends a minimum interval of 42-60 d between a SICCT test and a paratuberculosis milk ELISA. However, recent research suggested that the post-SICCT antibody response may be greater in cows at high risk of Mycobacterium avium ssp. paratuberculosis (MAP) shedding (i.e, with an increased likelihood of shedding MAP in feces), leading to advisory guidance that high-prevalence herds could shorten the testing interval to improve ELISA sensitivity. If post-SICCT antibody increases also occur in cows not at high risk of MAP shedding, due to nonspecific cross-reactive stimulation rather than anamnestic MAP-specific responses, false-positive results would increase. This observational study aimed to evaluate the association between the SICCT-ELISA interval and the milk ELISA sample-to-positive (S/P) ratio, separately for case cows classified as high risk of MAP shedding and control cows, and to determine whether this association varied across herd-years with high (≥10%) and low (≤2%) annual paratuberculosis antibody prevalence. Individual cow milk ELISA records from 518 UK dairy herds (2012-2023; 1,226,709 tests from 340,839 cows) were matched to SICCT test records. Generalized additive models were fitted separately for case and control cows under 2 case definitions: the industry-standard MAP shedding risk classification (2 positive milk ELISA results in the previous 4 tests), and a more stringent progression-based definition requiring sustained positive or negative testing histories. Additional models were fitted using data from the high- and low-prevalence subsets. Apparent test performance characteristics were estimated across SICCT-ELISA intervals. Contrary to previous findings, both case and control cows exhibited similar antibody responses after a SICCT test, with S/P ratios peaking around 18-22 d in both groups. In the prevalence-stratified analyses, this increase in S/P ratios after the SICCT was observed in control cows from both high- and low-prevalence herd-years. At the 0-14 d SICCT-ELISA interval, apparent positive predictive value was lower compared with the 60-90 d reference across all prevalence levels: 50.8% vs 67.7% in high-prevalence herd-years, 57.4% vs 63.3% in medium-prevalence herd-years, and 39.1% vs 48.9% in low-prevalence herd-years. In high-prevalence herd-years, each additional true positive at this interval was associated with approximately 3.3 additional false positives, whereas in low- and medium-prevalence herd-years, there was no overall increase in true positives. Overall, these results do not support shortening the SICCT-ELISA testing interval below the current guidelines.