Estimation of the repeatability and reproducibility of three diagnostic tests for infectious salmon anaemia virus.

Reverse transcriptase polymerase chain reaction (RT-PCR), virus isolation (VI) and indirect fluorescent antibody tests (IFAT) are three assays currently used by the salmon industry to identify fish infected with infectious salmon anaemia virus (ISAV). However, no data are available on the repeatability (within-laboratory consistency) and reproducibility (between-laboratory consistency) of these assays and very limited information is available on the effect of freezing samples on test results. In order to evaluate these assays, five laboratories participated in a blinded study of 400 kidney samples representing four populations of farmed Atlantic salmon with different prevalence of ISAV. Each laboratory used its own testing protocols. Repeatability and reproducibility were evaluated using kappa as the measure of agreement. The effect of freezing was evaluated using the McNemar test. Freezing did not affect VI but improved the sensitivity of RT-PCR. The repeatability and reproducibility of VI was almost perfect. There was a substantial difference in repeatability of RT-PCR among the three laboratories with kappa ranging from 0.5 to 0.96. The repeatability for RT-PCR was generally low. The repeatability of IFAT was moderate when the results were analysed using 1+ and above as a positive result. The results of the study show the need to standardize the protocol and interpretation of RT-PCR.

Estimation of specificity and sensitivity of three diagnostic tests for infectious salmon anaemia virus in the absence of a gold standard.

Reverse transcriptase-polymerase chain reaction (RT-PCR), virus isolation (VI) and indirect fluorescent antibody test (IFAT) are three tests currently used by the salmon industry to identify fish infected with the infectious salmon anaemia virus (ISAV). However, very limited information is available on the sensitivity and specificity of these methods. In order to evaluate these tests in fish representing a range of farmed Atlantic salmon populations, five laboratories participated in a blind study of 400 kidney samples from four groups of fish with different prevalences of ISAV. Each laboratory used its own testing protocols. Estimates of the specificity of each test were determined directly from a population assumed to be free of infection. Indirect estimates of the sensitivity and specificity of each test were obtained using maximum likelihood estimation of a latent class model (i.e. no gold standard test result available). There was a substantial difference in sensitivity and specificity of RT-PCR among the three laboratories using this test. If only the best results for the RT-PCR tests are taken into account, the maximum likelihood estimates obtained from this study suggest RT-PCR and VI are of similar high sensitivity (range 92-100%), IFAT is the least sensitive method (range 65-76%) while the three tests have similar high specificities (range 96-100%). The results of the study suggest: (1) RT-PCR tests should be standardized before they are used as a diagnostic test for prevention and control of ISAV, (2) the sensitivity of VI was higher than expected and (3) IFAT has a low sensitivity but might be a good screening test because of its low cost.

Assessment of infectious salmon anaemia virus prevalence for different groups of farmed Atlantic salmon, Salmo salar L., in New Brunswick.

Abstract Infectious salmon anaemia (ISA) virus (ISAV) has been causing disease in New Brunswick since 1996. As a control measure, all fish in an outbreak cage are killed. The objective of this study was to compare ISAV prevalence in cages experiencing an outbreak with healthy cages from the same farm, neighbouring farms and distant farms. Atlantic salmon from five different groups were tested using an RT-PCR test. Groups included moribund fish from a cage experiencing an outbreak (A), healthy fish from an outbreak cage (B), healthy fish from a negative cage from a farm experiencing an outbreak in a different cage (C), healthy fish from a negative farm near an outbreak farm (D), and healthy fish sampled at a negative farm located in an area with only negative farms (E). Apparent prevalences (standard error) for the different groups (A-E) were 0.94 (+/-0.026), 0.41 (+/-0.062), 0.29 (+/-0.040), 0.08 (+/-0.037) and 0.08 (+/-0.037), respectively. All groups were significantly different (P < 0.002) from each other except for groups B and C and groups D and E. Because the prevalence of the virus was significantly higher in the outbreak cage (B) compared with other sites, early harvest of outbreak cages will remove one source of virus. However, ISA negative cages (C) that remain on the positive farm may potentially act as a viral reservoir.