Variability in organochlorine analysis in fish: an interlaboratory study and its implications for environmental monitoring and regulatory standards.

Samples of muscle, from four types of fish, that had been spiked at 0.25 and 1.0 mg/kg with hexachlorobenzene (HCB), gamma BHC (lindane), technical chlordane, p,p'-DDT, p,p'-DDE and p,p'-DDD were analyzed by eight laboratories. Some laboratories could not reliably detect organochlorine compounds even though they were registered as national testing laboratories. The analytical accuracies of laboratories that were able to reliably detect the spiked compounds were very variable and exceeded acceptable limits. Within-laboratory precision was generally acceptable. There was significant interlaboratory variability among the estimates of percent lipid for each fish type. This variability negated the use of percent lipid for standardization of organochlorine concentrations for comparison between studies. The results highlight the need for standard analytical protocols that are regularly tested by interlaboratory studies. They also indicate that interchangeability of data between studies is limited due to low analytical accuracy. Therefore, comparison of analytical data to absolute environmental and regulatory standards is difficult. The use of "latitudinal confidence ranges" when setting regulatory standards is recommended.