Mean effective sensitivity for Mycobacterium avium subsp. paratuberculosis infection in cattle herds

Abstract

BackgroundMycobacterium avium subsp. paratuberculosis (MAP) infections in cattle are generally challenging to detect and cost-effective test strategies are consequently difficult to identify. MAP-specific antibody ELISAs for milk and serum are relatively inexpensive, but their utility is influenced by a number of factors such as herd size, herd composition and diagnostic sensitivity. The sensitivity of the test increases with the age of the tested animal, and therefore the general, or “mean effective sensitivity” (defined as the mean of the sensitivities for all animals within a population, MES), for detecting MAP within a herd is dependent upon the age distribution of the herd. For this study we used a dataset of cattle from 4,259 dairy herds and 4,078 non-dairy herds. The aim was to investigate the MES for groups of cattle considered to be reasonable entities for MAP surveillance and control, in order to assist the decision-makers in planning and optimizing these programs economically. We compared six different groups of cattle (three dairy and three non-dairy) in Denmark by calculating the MES for each herd in each group.ResultsThe distribution of MES showed a large variation within and between groups, and in some groups we found a bimodal distribution of MES. Dairy herds generally showed higher MES than non-dairy herds. Dairy herds in a control programme for paratuberculosis showed a MES similar to all other dairy herds from which animals > 2.0 years were tested (both groups had a median MES = 0.60). For the non-dairy groups, the sensitivity became much higher when animals < 2.0 years and herds with less than 25 cattle were excluded, resulting in a median MES of 0.65.ConclusionThe results showed that MES could indicate the effectivity of testing different cattle groups for MAP, given that the data used are unbiased.

Evaluating the impact of transmission mode, calibration level and farmer compliance in simulation models of paratuberculosis in dairy herds

“…The simulated scenarios presented here reflect a closed herd. However, we believe the results can also reflect an open herd, because the purchase of infected animals would not be expected to increase the within-herd prevalence directly to a high degree, but would instead have more of an impact on eradication strategies, as MAP is continually introduced into the environment 25 . In Fig.…”

“…However, the sensitivity of the test is age dependent, which is incorporated in the iCull model, but is not possible to include in the Rogan-Gladen estimator. We therefore used the mean effective sensitivity (MES) calculated for Danish dairy herds 25 at 60%. The specificity was set to 98.66% 21 .…”

Knowledge gaps that hamper prevention and control ofMycobacterium aviumsubspeciesparatuberculosisinfection

“…Although MAP test comparisons have been the subject of many studies, there is a lack of studies comparing test characteristics in populations that reflect the target population when estimating the prevalence of MAP infection (Nielsen & Toft, ). The sensitivity of MAP tests increases in dairy cattle with increasing days in milk and age but decreases with increased milk yield (Eisenberg, Veldman, Rutten, & Koets, ; Kirkeby, Græsbøll, Halasa, Toft, & Nielsen, ), whereas sensitivity increases with age in sheep and goats (Lybeck, Storset, Djønne, Valheim, & Olsen, ). Therefore, sensitivity estimates need to be adjusted for age in all ruminants, and days in milk and milk yield in dairy cattle when comparing between or within herds and also over time.…”

A data-driven individual-based model of infectious disease in livestock operation: A validation study for paratuberculosis

“…It has a large economic impact for producers due to decreased milk production [13–15], premature culling [16,17], reduced slaughter value [18], low fertility [19,20], and an increased animal replacement rate [21]. However, MAP is difficult to diagnose due to the long incubation period [22,23], lack of early clinical signs [24–26], and imperfect testing [27].…”