Methods for Deriving Pesticide Aquatic Life Criteria - Reviews of Environmental Contamination and Toxicology - page 37

ally accepted standards. Relevance is judged by whether a study included appropriate
endpoints, was conducted under relevant conditions, and if the substance tested was
representative of the substance being assessed. The EU criteria derivation guidance,
as described by Bro-Rasmussen (1994), is very general with regard to data quality,
and primarily requires that data include details of tests used.
The UK, the Netherlands, Canada, and Australia and New Zealand evaluate
ecotoxicity data and assign ratings depending on its reliability and/or relevance.
In the UK, primary data are those classified as reliable and relevant, and secondary
data are those for which inadequate details are available. The evaluation is based
purely on expert assessment of experimental procedures, test species, endpoints,
and whether or not a dose-response relationship has been established. Primary data
are used in derivation of EQSs; secondary data are used only as supporting information
(Zabel and Cole 1999).
The Dutch methodology uses a reliability index (RI) to evaluate ecotoxicity data
(RIVM 2001). Reliable data (RI = 1) are from studies conducted and reported in
accordance with internationally accepted test guidelines or those set by Mensink
et al. (1995). Less reliable data (RI = 2) are those from studies in less accord with
accepted guidance or Mensink et al. (1995), and data deemed not reliable (RI = 3)
are from studies not at all in accord with accepted guidance or Mensink et al.
(1995). Data rated 1 or 2 are used in derivation of ERLs; data rated 3 are included
in the final report, but are not used in criteria derivation.
Part of data quality is ensuring that data come from properly conducted, well-
documented studies. In the Netherlands (RIVM 2001), data must come from refer-
enced studies that include specific organism identification, information regarding
purity of the test substance, details of the test, and clearly stated results. For system-
atic evaluation, data are subdivided by type (freshwater, marine, acute and chronic)
and put into data tables. Table headings should include species (including scientific
name), species properties (e.g., age, weight, and lifestage), analysis of test com-
pound (measured or not, Y or N), test type (flow-through, static-renewal, and
static), substance purity, test water, pH, water properties (e.g., hardness, salinity),
exposure time, test criterion (e.g., LC 50 - concentration that is lethal to 50% of
organisms - or NOEC - no-observed-effect concentration - ecotoxicological
endpoint (growth, reproduction, mortality, immobilization, morphological effects,
and histopathological effects), LC 50 values, NOEC values, notes, and reference
information.
For data to be usable in criteria derivation in the Netherlands, specific toxicity
test acceptability requirements must be met (RIVM 2001). These include that the
purity of the test substance must be at least 80%, studies may not use animals col-
lected from polluted sites, concentration of test substance may not exceed 10× the
water solubility, no more than 1 ml/L of carrier solvent can be used, and recovery
of the substance must be at least 80%. For compounds with short half-lives ( t 1/2 ),
the renewal frequency in a static-renewal test is important. In the Dutch methodol-
ogy, if the t 1/2 is shorter than the renewal interval, the data are not used.
The Australia and New Zealand guidelines follow the standard operating procedures
for the AQUIRE system (1994) for rating toxicity studies according to how well they