Methods for Deriving Pesticide Aquatic Life Criteria - Reviews of Environmental Contamination and Toxicology - page 67

hydrophobic chemicals by fish have been reported, but the relationship is not con-
sistent (Gobas et al. 1988; Qiao et al. 2000). Gobas et al. (1988) found an inverse
relationship between log K ow and dietary uptake efficiency, with efficiency decreasing
for log K ow values more than 7. In contrast, Qiao et al. (2000) found that gill uptake
accounted for 98% of fish body burden for chemicals with log K ow values of 5 or less;
whereas, for a chemical with a log K ow value of 7.5, 85% of body burden was from
dietary uptake. The Qiao et al. model (2000) determined that food-water concentra-
tion ratios were important predictors of the relative uptake by the two routes. For
ratios more than 10 7 , dietary uptake was predicted to be 100%; at about 10 5 , uptake
was similar for diet and water; and for ratios less than 10 3 , uptake was 100% from
water. The relationship between log K ow and exposure route was modeled using
environmentally relevant food-water concentration ratios (ranging from 191 to
10 5.9 ). Fisk et al. (1998) found a significant curvilinear relationship between log K ow
and dietary uptake efficiency, with efficiency increasing for log K ow values between
5 and 7, and then declining for values above 7. Other studies have found no clear
relationship between log K ow and uptake efficiency. One study (Loonen et al. 1991),
suggested a link between level of chlorination of dioxins, and another reported an
activated transport mechanism for hydrophobic organic chemicals, with uptake effi-
ciency dependent on molecular weight (Burreau et al. 1997).
Although dietary uptake is an important exposure route for many hydrophobic
organic compounds, the theoretical basis for differences in dietary uptake efficiency
is not clearly established. For narcotic chemicals (those exhibiting a non-specific
mode of action), Traas et al. (2004) have developed a food web model for calcula-
tion of environmental quality criteria, based on internal effect concentrations. This
model is based on concentrations of contaminants already in organisms, rather than
from exposure. Thus, all exposure routes are incorporated. However, the model
does not work for chemicals with specific toxic modes of action, a characteristic of
most new pesticides.
Until food web or other models are further developed to incorporate multi-
pathway (multiple routes) exposures into criteria derivation, it is probably best to
continue with water-only assessments. If studies show criteria to be underprotective,
and if a substance has a log K ow between 5 and 7, then dietary uptake studies,
specific to the compound and species affected, should be performed to determine
if exposure has been significantly underestimated. Many modern pesticides tend
to be less hydrophobic (even water soluble) rendering the dietary exposure route
less important.
When deriving water quality criteria, it is appropriate to also consider whether
criteria should be expressed as total chemical or bioavailable chemical, and
whether criteria should be adjusted for other factors (e.g., pH, temperature, inter-
actions with other substances) known to affect toxicity. Criteria that utilize toxic-
ity test data are intrinsically based on bioavailable chemicals, and thus incorporate
bioavailability. However, many laboratory tests are performed in clean water
under controlled water quality conditions. Such tests do not replicate the effects
that natural waters have on toxicity. Bioavailability is addressed in virtually all
existing methodologies, particularly for metals. Few methods, however, quantitatively