Diagnosing subclinical facial eczema in cattle: does combining liver enzyme tests increase the accuracy of diagnosis?

Abstract

ABSTRACT Aims To assess whether adding glutamate dehydrogenase (GDH) activity measurements to measurements of gamma-glutamyl transferase (GGT) activity appreciably increases the accuracy of diagnosis of subclinical facial eczema (FE) in cattle. Methods As part of a larger study on the impact of FE on productivity, GGT and GDH activities were measured in serum samples collected from 426 cattle from one dairy farm in the Taranaki region in April 2018. Bayesian latent class analysis was then used to estimate herd prevalence of subclinical FE as well as the specificity and sensitivity of the activity in serum of GGT or GDH alone, and of GGT and GDH activities combined, as diagnostic tests for subclinical FE. Results The latent class analysis estimated the true prevalence of subclinical FE in the study population as 47.5 (95% probability interval (PI) = 38.3–55.3)%. There was no evidence of any clinically relevant difference between GGT and GDH activities as predictors of subclinical FE; the difference between the areas under the receiver operating characteristic curves for the two measures was 0.005 (95% PI = –0.02 to –0.03). Using the two tests in parallel, with a threshold of 50 IU/L for GGT and 225 IU/L for GDH resulted in specificity and sensitivity of >95%, markedly increasing the accuracy of diagnosis of subclinical FE compared to using GGT or GDH alone at any threshold Conclusions and clinical relevance In this herd, combining the two tests resulted in a clinically relevant improvement in the accuracy of diagnosis of subclinical FE compared to using either test alone, which if used at the individual level will result in fewer cattle being assigned the wrong FE status. This will also apply at the herd level, with combined testing producing fewer false-positive herd test results than using one enzyme alone. This is particularly important for monitoring the efficacy of FE control measures when the expectation should be that the proportion of cattle with FE is very low.