dioxins and DLCs. Food choices are compared in Table 5-1. For a 57 kg reference female to make a weekly selection of two 3-ounce servings of seafood and not exceed the RfD level for methylmercury (5.7 µg MeHg per day) requires that no more than one of those selections be white (albacore) tuna. In contrast, levels of DLCs in two 3-ounce servings of white (albacore) or light canned tuna or salmon do not exceed target exposure limits. Making a trade-off of 20 percent fat beef for salmon will not decrease the exposure levels to DLCs, although it will for MeHg. Selecting light canned tuna in place of white (albacore) tuna will decrease exposure levels to both MeHg and DLCs, but will also significantly decrease intake levels of EPA/DHA.
The substitution analyses presented above are based on nutrient and contaminant values. However, there are several sources of uncertainty in the estimates of these nutrient and contaminant values. Mean estimates of nutrients represent best estimates of the value one would expect to find in any specific case. A common indication of how much one can expect individual values to vary from the mean is a confidence band, the calculation of which is based on the individual sample values observed. The committee did not have access to individual sample values or estimates of the variability in those for the estimates reported in Table 5-1.
Table 5-2 characterizes some of what is known about the data from which the estimates in Table 5-1 were derived (i.e., their provenance). As Table 5-2 illustrates, sample sizes and ages vary tremendously. While it is difficult to determine just how significant it is, variability in sample sizes and approaches, together with changes over time in analytical methods (Igarashi et al., 2000; Siddiqui et al., 2003) and in feed (e.g., reductions in the use of fishmeal in poultry feed) are noteworthy sources of uncertainties (Barlow, 2001). The EPA/DHA levels in chicken provide a case in point.
Comparison of the estimates in Table 5-1 with those from other sources also suggests considerable variability and uncertainty. Hamilton et al. (2005) illustrate that levels of omega-3 fatty acids in salmon vary by source. Using samples of farmed, wild, and store-purchased salmon from a large number of locations, they estimated that omega-3 fatty acid levels in farmed Atlantic salmon are more than twice as high as in salmon from other sources. Further, their estimate of omega-3 fatty acid levels in farmed Atlantic salmon is almost twice as high as that shown in Table 5-1.
Importantly, specific population subgroups, e.g., Native Alaskans who previously relied on seafood and marine mammal consumption and followed advice to decrease their intake of these foods to reduce the risks associated with exposure to contaminants, suffered negative consequences in overall nutrition (Wheately and Wheately, 1981; Murphy et al., 1995;