WA2 Plant Toxicology and Soil Interactions
Ballroom F, Level 4
8:00 AM - 12:00 PM, Wednesday, 12 November 2003
Chair: Jackson, Andrew ,
(339) Meta-analysis of algal toxicity test data: Implications for harmonizing test methods and maximizing regulatory use.
Staveley, J.1, Leitman, P.2, Smrchek, J.3, 1 ARCADIS, Inc., Durham, NC, USA2 ARCADIS, Inc., Atlanta, GA, USA3 U.S. EPA, Office of Prevention, Pesticides and Toxic Substances, Washington, DC, USA
ABSTRACT- Standardized toxicity test methods for unicellular microalgae differ in both how the test is conducted and how the test results are expressed. The revision of OECD Guideline 201 (Freshwater Alga, Growth Inhibition Test, 1984) has been complicated by disagreement among the member countries on the appropriate response variables and test endpoints. To provide a better framework for resolving these differences, this study evaluated raw data from 41 algal toxicity tests. These tests were conducted in Europe and the United States, on three identified species as well as one unknown species. Five response variables (final cell counts, biomass integral or area under the growth curve, average growth rate, specific growth rate, and log of biomass integral) were used to calculate some or all of the following test endpoints: EC10, EC20, EC50, and NOEC. The results indicate that EC values based upon growth rate were almost always higher than those based upon biomass integral, but generally within a factor of three. One oft-cited advantage of growth rate is that it is time-independent; however this was demonstrated not to be the case. Factors such as the magnitude of the growth rate, the slope of the dose-response curve, or extrapolation beyond the range of test concentrations did not impact the relationship between biomass integral and growth rate in a predictable manner. It was concluded that both a biomass response variable and a growth rate response variable should be reported, and it is feasible to design an algal toxicity test to obtain endpoints based upon both. Importantly, the interpretation of results must be linked to the response variable. For example, the coefficient of variation in the controls should be <21% for final cell count and biomass integral, but restricted to <12% for average growth rate or specific growth rate. For biomass integral, the NOEC correlates to the EC20, while for growth rate, the NOEC correlates to the EC10.
Key words: growth rate, algal toxicity tests, harmonization, statistical analysis