Predictive values and quality control of techniques for the diagnosis of Echinococcus multilocularis in definitive hosts

By the end of the 1980s the only reliable technique for diagnosing the intestinal Echinococcus multilocularis infection of definitive hosts was parasite detection at necropsy. Currently, several techniques for the post mortem and in vivo diagnosis are available, including classical and modern methods. The sedimentation and counting technique (SCT) is used for the exact determination of the worm burden in the intestine after necropsy. The SCT has high sensitivity and specificity values (both close to 100%) and can be regarded as 'gold standard'. The principle of the intestinal scraping technique (IST) is the stereomicroscopic examination of at least 15 intestinal thick smears. This technique has a sensitivity of 78% and a specificity approximating 100%. In recent years, the IST has been successfully used in large post mortem surveys of foxes for E. multilocularis. The newer techniques for detecting coproantigens by ELISA (CA-ELISA) exhibit rather high sensitivities between 84 and 95%, combined with very high specificities of > 96%, the latter regarding non-Echinococcus cestodes and other parasites. However, cross-reactivity may occur with E granulosus. Copro-DNA detection by PCR is also highly sensitive (89-94%) and specific (100%). With the SCT, IST and Copro-PCR one person can only examine about 10-20 animals per day, whereas the CA-ELISA allows the examination of 200 samples. Therefore, the latter test is suited for mass-screening of definitive host populations. Both the CA-ELISA and the Copro-PCR allow the examination of materials from dead and living animals, including faecal samples collected in the field. Quite often diagnostic techniques have been used without adequate quality control and proper definition of their performance characteristics, including diagnostic sensitivity, specificity, predictive values and some other parameters. Examples are presented with the aim to demonstrate the need and the value of calculating the predictive values for assays used to diagnose the E. multilocularis infection in individual animals and in definitive host populations. (C) 2002 Elsevier Science B.V. All rights reserved.