Page 147 - The Toxicology of Fishes

Toxicokinetics in Fishes 127

common to start with the simplifying assumption that all elements of an ecosystem are at steady state.
Non-steady-state models may be required, however, when chemical loadings to a system change over
time or when other processes result in a substantial disequilibrium among water, sediment, and biotic
components.
Techniques used to measure, express, model, and predict bioconcentration and bioaccumulation are
described below. The concepts described in these sections are then combined in a simple food web. A
brief description of the fugacity-based approach to chemical modeling is given to provide a basis for
comparison with other modeling approaches. A more detailed description of the fugacity-based modeling
approach is given in Chapter 14.

Bioconcentration
The term used to quantify the magnitude of bioconcentration in aquatic systems is the bioconcentration
factor (BCF), which is defined as a proportionality constant relating the chemical concentration in a fish
(or a portion thereof) to that in water under steady-state conditions. It is a measure of the propensity of
a chemical to accumulate in fish, and BCF values fall between 0 and infinity (Veith et al., 1979). BCFs
have units of water volume/tissue weight (e.g., mL/g) and can be viewed conceptually as the water
volume containing the amount of chemical concentrated in 1 gram of animal tissue (Barron, 1990).
BCFs can be measured directly by exposing fish in water until steady state is attained or estimated using
simple kinetic models in conjunction with short-term uptake and elimination studies. Using these
procedures, BCFs have been determined for a large number of nonionic organic compounds, as well as
some metals, metalloids, and organometallics. Several authors have used these datasets to develop QSARs
that relate the BCF of a compound to one or more physicochemical properties. These QSAR models
are used extensively to estimate BCFs for untested chemicals and by extension to predict their behavior
in aquatic systems.
As indicated previously, hydrophobic compounds tend to partition into tissue lipids. BCFs for such
compounds may be interpreted, therefore, in terms of a chemical distribution between fish lipid and
water. When making these interpretations, it is important to distinguish between a steady-state chemical
distribution and true thermodynamic equilibrium. A steady-state condition is defined as an unchanging
chemical concentration, and a thermodynamic equilibrium is a minimum energy state that reflects the
relative affinity of a chemical for two partitioning phases. A system at steady state is unlikely to be in
thermodynamic equilibrium in the presence of dynamic processes such as biotransformation and
growth.
The lipid content of fish varies widely. Variability among BCFs for the same or similar compounds
in more than one species or life stage can be reduced, therefore, by normalizing for differences in lipid
content, although factors such as lipid composition may also be important (Ewald and Larsson, 1994;
Vigano et al., 1992). The lipid-normalized BCF has units of mL/g lipid.

Steady-State Exposures
Steady-state exposures are conducted by exposing fish to a constant concentration of chemical in water
for an extended period of time (U.S. EPA, 1996). Typically, these studies are conducted by exposing
small fish in a flow-through system for 28 days. The fish are serially sampled, and whole-body concen-
trations of the chemical are quantified. This method may underestimate the true steady-state BCF if
exposure times are too short. The length of time required to reach steady state depends on uptake and
elimination rates; for example, specific PCB congeners with elimination half-lives of 100 days or longer
may require 1 year to reach 90% of steady state (Barron et al., 1994). Steady-state exposures are species
or site specific and allow for the incorporation of environmentally realistic exposure conditions.

Kinetic Modeling
Bioconcentration factors may also be estimated by applying simple kinetic models to uptake and
elimination datasets (Barron et al., 1990; Branson et al., 1975; Landrum et al., 1992; U.S. EPA, 1996).
Uptake from water is generally assumed to be first order with respect to the chemical concentration in