Critical predicted no effect concentrations (PNECs) should not be based on a single toxicity test

Abstract

Predicted no-effect concentrations (PNECs), which represent the concentration of a substance below which an unacceptable effect most likely will not occur, are widely used for risk assessment and in environmental policy and regulation. They are typically based on single-species laboratory toxicity tests; often, a single test result for the most sensitive endpoints drives the derivation of a PNEC. In the present study, the authors provide a case study emphasizing the importance of determining the reliability of those most sensitive endpoints. Five 21-d Daphnia magna toxicity tests conducted using the same procedures by 2 laboratories gave 20% inhibitory concentration responses to a specific ionic composition of total dissolved solids that varied from 684 mg/L to more than 1510 mg/L. The concentration-response curve was shallow; thus, these differences could have been attributable to chance alone. The authors strongly recommend that the most sensitive endpoints that determine PNECs not be based on a single toxicity test result but rather on the geometric mean of at least 3 test results to adequately assess and bound test variability, especially when the concentration-response curve is shallow.

Development of a total dissolved solids (TDS) chronic effects benchmark for a northern Canadian lake

“…Each batch of synthetic lake water was analyzed for its ionic composition to assess concentrations of the major ions and to calculate TDS concentrations. Results of those analyses confirmed that the calculated TDS concentrations for the synthetic lake water batches ranged from 1380 to 1520 mg/L, within 10% of the target nominal concentration of 1500 mg/L (Chapman and Elphick ), and from 2030 to 2200 mg/L, again within 10% of the target nominal concentration of 2000 mg/L.…”

“…All point estimates are presented as calculated TDS concentrations. Additional details are provided in Golder (); Chapman (, , ); Baker et al (); Marus et al (); and Chapman and Elphick (). Control performance met test acceptability criteria for all test methods, and test results were corrected for mean control responses.…”

“…The geometric mean (CCME ) was greater than 1100 mg/L, still the most sensitive resident taxon tested. Additional details are provided in Chapman and Elphick ().…”

“…The following tests were conducted. A brief summary of methodology is provided below; full details on methodology are provided in Golder (); Chapman (, , ); Baker et al (); Marus et al (); and, Chapman and Elphick (). Phytoplankton: 72‐h growth inhibition of the green alga Pseudokirchneriella subcapitata (Environment Canada ) and 120‐h growth inhibition of the diatom Navicula pelliculosa (ASTM ).…”

Including or excluding toxicity test data for development of a geometric mean

“…The other, Ceriodaphnia dubia, was not a representative of a resident genus or species. Five D. magna 21-d tests gave useful, acceptable data [10]. Six C. dubia 7-d tests that met all guidlines but guideline 7 resulted in no or negligible effects endpoints that ranged from similar to those derived for D. magna and other tested species (3 other invertebrates, 2 algae, 3 fish; all but 2 were sitespecific taxa) to lower than total dissolved solid concentrations that presently are not adversely affecting waterflea populations living in the receiving water body.…”