Short-term adult fish reproductive tests: evaluation, refinement and application

Short-term adult fish reproductive tests are a potentially valuable tool for assessing the potential of chemicals or waste streams to affect the reproductive endocrine system. A short-term reproductive test exposing mummichog (Fundulus heteroclitus) to pulp and paper mill effluent identified high variation in measures of gonad size, endocrine endpoints and egg production that resulted in low power to detect differences among treatments. This outcome led to a meta-analysis of 62 published fish adult reproductive tests using various fish species, in which small samples sizes (mean of 5.7 replicates with a median of five replicates) and high variance in egg production (an average coefficient of variance of 43.8%) were identified as common issues. The resultant ability to only detect reductions in egg production of >50% with confidence is problematic. Based on a case study using mummichog, refinements were proposed to increase the power in short-term adult fish reproductive tests. These include: i) a pre-exposure period followed by tank selection based on a standardized set of criteria, ii) an optimized period of exposure during the peak period of egg production, iii) an extended duration of egg collection, and iv) an increased sample size (eight replicate tanks per treatment). The suggested refinements were tested using pulp and paper mill and municipal effluents and had a six-fold increase in power compared to studies with the original protocol. It is, therefore, concluded that a simple refinement process can address shortcomings in short-term adult fish reproductive protocols. The ability of commonly measured reproductive biomarkers (plasma reproductive hormone levels, vitellogenin levels and gonad size relative to body weight) to predict the reproductive performance of fish in these tests was also evaluated. A qualitative meta-analysis on adult fish short-term reproductive tests found low incidences of false positives (i.e., an effect in biomarkers was not followed by an effect in egg production), but a high occurrence of false negatives (i.e., a reduction in egg production was not accompanied by a biomarker response). The occurrence of false negatives was greatly reduced when using multiple biomarkers, increasing the potential to use biomarkers to predict effects of chemicals and waste streams on reproductive performance in fish.