Evaluation of testing strategies for reliable measurement of rates of subclinical mosquito-borne viral infections

Aim Insect-transmitted viral infections remain a significant public health concern in Australia. The positive predictive value (PPV) for detection of anti-viral antibodies in donors was assessed using commercially-available kits. Methods Blood samples collected from Australian blood donors were tested by ELISA for IgM or IgG specific for either Dengue virus (DENV), Ross River virus (RRV) or Barmah Forest virus (BFV). The PPV was determined by comparison with results from reference laboratories. Results For the commercially-available kits: anti-DENV IgG, 225 kit-positive, 213 confirmed, PPV 94.7%; anti-RRV IgG, 315 kit-positive, 239 confirmed, PPV 75.9%; anti-BFV IgG, 127 kit-positive, 47 confirmed, PPV 37.0%; anti-DENV IgM, 118 kit-positive, 7 confirmed, PPV 5.9%; anti-RRV IgM, 138 kit-positive, 45 confirmed, PPV 32.6%; anti-BFV IgM, 142 kit-positive, 71 confirmed, PPV 50.0%. Some potential sources of unexpected kit-positives were investigated, however, these contributed minimally to the unacceptable PPVs. Conclusions These kits are designed for diagnostic testing of symptomatic patients, however, it would be expected that reliability for investigation of sub-clinical infection would be equal. PPVs, with the exception of 94.7% for IgG anti-DENV, were less than 80%. To measure rates of subclinical infection an investigation algorithm including exploration of unexpected positives and reliable confirmatory testing is required.