An approach for quantitatively balancing methylmercury risk and omega-3 benefit in fish consumption advisories

Abstract

Background: Nearly all fish consumption advisories for methylmercury (MeHg) are based only on risk. There is a need to also address benefits, especially those from polyunsaturated fatty acids (PUFAs), in neurodevelopmental function and cardiovascular health. However, because MeHg and PUFA generally act on these same end points, disentangling risk and benefit is challenging.

Objectives: We propose an approach for balancing risk and benefit that is based on the use of statistically dissociated measures of risk and benefit.

Discussion: Because of mutual coexposure of MeHg and PUFAs in population-based studies and their opposite effect on many of the same end points, MeHg risk and PUFA benefit are tightly linked statistically, which results in mutual (negative) confounding. Thus, neither MeHg risk nor PUFA benefit can be accurately quantified without taking the other into account. A statistical approach that generates unconfounded risk and benefit coefficients for each end point can permit their subsequent recombination to describe the overall risk-benefit profile of each species of fish or fish diet. However, it appears that some end points may be adversely affected by MeHg without experiencing counterbalancing benefit from PUFAs. Such end points may drive consumption advisories and may preclude balancing of risk and benefit on the basis of other end points.

Conclusions: Our thinking about fish consumption advisories now recognizes the need to balance risk and benefit. However, although statistical analysis of the appropriate data can eliminate mutual confounding, care is required to address the most sensitive end points that may be sensitive to risk and not benefit.

Figure 1

Negative confounding of MeHg risks and PUFA benefits.

Figure 2

Odds ratios for nonfatal myocardial infarction as a function of mercury exposure (A) and DHA intake (B), each nonadjusted and adjusted for the negative confounding of the other. Data from Guallar et al. (2002).

Figure 3

Three-dimensional projection of a hypothetical relationship among IQ, MeHg intake, and PUFA intake with IQ regressed on MeHg and PUFA. The lines show constant values of IQ scores relative to varying levels of MeHg and PUFA. x designates a specific, but indeterminate value.

Figure 4

An example of how the same weekly intake of different species of fish with different characteristic MeHg and PUFA concentrations would influence IQ scores given a hypothetical regression relationship among IQ, MeHg exposure, and PUFA intake (see the description of this figure in the text for a more detailed explanation). x designates a specific, but indeterminate value.