Reasons for performing study: The diagnostic value of several investigative procedures commonly used during the evaluation of suspected equine hepatopathy cases has not been specifically quantified in previous studies.
Hypothesis: No noninvasive procedures would clearly discriminate between horses with and without significant liver disease.
Methods: Histopathology of biopsy samples was used as the 'gold standard' technique for definitive diagnosis of the presence or absence of significant liver disease. Clinical, ultrasonographic and clinicopathological data obtained during the investigation of 82 suspected cases of hepatopathy in mature horses were compared with the results of subsequent biopsy in order to quantify the diagnostic value of each test.
Results: Of the 82 cases, 61 were confirmed to have significant liver disease and 21 were not. Only serum concentrations of gamma-glutamyltransferase (gammaGT), globulins and alkaline phosphatase (AP) were found to be significantly different between the 2 groups of horses. Clinical and ultrasonographic abnormalities were found, when present, to be good indicators of the presence of liver disease. Certain single serum biochemical tests and combinations thereof were found to have high values for sensitivity (SN), specificity (SP), positive predictive value (PPV) and/or negative predictive value (NPV), but none showed high values for all 4 parameters leading to important limitations in their practical diagnostic application. Of all serum biochemical tests studied, the PPVs of only gammaGT and globulins could be clearly improved by considering especially high positive results.
Conclusions: The single positive test results of greatest diagnostic value were presence of hepatic encephalopathy (HE), increased gammaGT, hyperglobulinaemia, hypoalbuminaemia, increased AP, increased total bile acids (TBA) and increased total bilirubin (TBil). Increased aspartate aminotransferase (AST) and increased glutamate dehydrogenase (GLDH) were also of good diagnostic value, but only when used in combination with the above tests. However, no single, combination or sequential test was able to discriminate fully between horses with and without biopsy-confirmed liver disease and reliance on the use of noninvasive tests for the prediction of the presence or absence of significant liver disease may lead to frequent diagnostic errors. Although certain positive results did reliably predict the presence of liver disease, negative test results were invariably poor predictors of the absence of liver disease.
Potential relevance: Application of the findings in this study may not be directly applicable to other case populations. However, the results may help clinicians with their selection and interpretation of appropriate test results in cases of suspected liver disease.