INTRODUCTION

Environmental risk assessments (ERA) comprise two elements: exposure assessment and effects (or hazard) assessment. In ERA the likelihood of adverse effects of organic chemicals on aquatic organisms is evaluated by comparing exposure estimates with defined effect or no effect endpoints. The effects assessments are generally based on data obtained from a range of standardised toxicity tests of varying duration and employing a range of relevant species. The growing demand for data to support effects assessment underscores the importance of finding efficient approaches to experimental design and data interpretation. It is also important that ecotoxicologists continue to actively pursue the principles of the 3R’s (replacement, reduction and refinement) (Russell and Burch, 1959) of animals used in regulatory studies.

Exposure to aquatic organisms can occur both from the water phase and the diet; however, current guidelines (OECD 203, 202, 201) (OECD, 1992, 2004, 2011) largely derive effects endpoints solely from water-born exposure The concentration in the test medium (water) is generally used to quantify the effect (toxicity) endpoint (e.g. Mackay et al, 1992); however, this exposure medium is only a surrogate for the amount of toxicant that actually reaches the site of toxic action in the organism resulting in the toxic effect at the assessment endpoint. It is generally accepted that the toxic effect is directly attributable to the delivered dose of chemical to a target within the organism and only indirectly to the external exposure (e.g. Escher and Hermens, 2002).

Alternative approaches to the use of these tests have, and are, being explored to establish whether there are more appropriate ways of assessing environmental hazards and whether alternative dose metrics could be more suitable. One approach is the use of critical body burden (CBB) or critical body residue (CBR). McCarty and Mackay (1993) proposed the use of CBRs for use in ecological risk assessment, where exceedance of an effect threshold leads to an observed biological response that is largely proportional to the amount of the chemical at the sites of toxic action. Considerable work has carried on CBRs over the last 20 years (e.g. Meador et al, 2011) and a number of reviews have been made of this concept e.g. Barron et al (1997, 2002), Sijm and Hermens (2000) and Thompson and Stewart (2003). Despite strategies such as lipid normalisation (Di Toro et al, 2000), CBRs tend to be noisy / variable. ECETOC (2005) proposed a multi-tiered approach to using CBB in risk assessment and a number of research projects addressing the value of CBB have been funded by the Cefic Long-range Research Initiative (Cefic LRI). The usefulness of CBB is highlighted by the recognition of a number of toxic modes of action (MoA). Mode of action can be defined as a common set of physiological and behavioural signs that characterise a type of adverse biological response (Escher and Hermens, 2002), where the major (but not all) biochemical steps are understood.

In a series of papers, Verhaar et al (1992, 2000) proposed a framework for the identification of four classes of compounds with different MoA, including two for narcosis with non-polar narcosis defined as baseline toxicity (inert substances) and polar narcosis (less inert chemicals, more toxic than predicted by baseline toxicity estimations), which are commonly identified as possessing a hydrogen bond donor (see Table 1). Another MoA scheme is that described by Russom et al (1997) which classifies substances into one of seven groups. Other studies (Veith et al, 1983) have demonstrated a relationship between the octanol-water partition coefficient (Kow) and non-polar narcosis. The concept has been further developed using approaches that use the Abraham (1994) polyparameter Linear Free Energy Relationships (ppLFERs) to identify non-polar and polar narcotics (Kipka and Di Toro, 2009) instead of Kow. The Kow and ppLFER approaches seek to characterise the same underlying behaviour of chemical partitioning from the aqueous exposure medium to hypothesised target sites in the body, i.e. toxicokinetics.

A second approach considers the link between activity and toxicity, first proposed by Ferguson (1939) for baseline narcotics, has been explored more recently by Mackay et al (1992), Kipka and Di Toro (2009), Mayer and Reichenberg (2006), Reichenberg and Mayer (2006) and Schmidt et al (2013). Precise laboratory exposures can be achieved by passive dosing techniques using solid sorbents as the vehicle for chemical delivery as demonstrated by Schmidt et al (2013). These authors also showed that the toxicity of mixtures can be assessed by addition of activities, as lethality from exposures to individual chemicals and mixtures occurred to springtails at a total activity over a very narrow range from 0.015 to 0.050 with 50% lethality at an activity of approximately 0.03. The chemical ‘activity additivity’ approach is similar in principle to adding toxic units (Escher and Hermens, 2002). Potential additional advantages of expressing toxicity using the activity framework are that it can be applied to air-breathing and water-respiring animals, it avoids the variability in CBR attributable to lipid content differences and it enables measured activities causing baseline toxicity in laboratory studies to be compared with activities that are measured or predicted in the environment (Mackay and Arnot, 2011; Mackay et al, 2011).