Nutritional advice is usually straightforward. There are things we’re supposed to eat more of: vegetables, fruit, whole grains. There are things we’re supposed to eat less of: trans fats, added sugars, anything coated with salty orange dust.
And then there’s fish.
We’re supposed to eat more of it because it has healthful omega-3 fats. But we’re supposed to eat less of it because it’s full of environmental contaminants. Balancing the risks and benefits is hard, even for the doctors and scientists in the field. Absent sound advice, it’s all but impossible for those of us trying simply to decide what’s for dinner.
Consider salmon. According to the joint advisory issued by the Food and Drug Administration and the Environmental Protection Agency, salmon is low in mercury and safe even for pregnant women. Yet the Environmental Defense Fund, an advocacy group, suggests that all adults — not just pregnant women — limit wild salmon (except for Alaskan) to one serving per month and farmed salmon to no more than two, because of PCB contamination.
This kind of disparity raises two questions: What do we know about fish, and what do we know about the advice we’re getting about fish?
We certainly know that there are benefits from eating fish and risks from eating contaminants. The advantages are mostly attributed to long-chain omega-3s, polyunsaturated fats that are found in almost all fish, and almost exclusively in fish. The best-established benefits are reduction of heart disease risk and, in the case of pregnant women, improved neurodevelopment in fetuses and young children. But there’s also a slate of less well-established benefits, ranging from reduced stroke risk to mood elevation.
The risks come from mercury and PCBs, both byproducts of industrial processes. (The United States banned PCBs in 1979, and environmental levels are persistent but gradually decreasing.) These contaminants accumulate in fish tissue; they become more concentrated as you go up the food chain as the bigger fish eat smaller, contaminated fish. Fish are the only predators we eat regularly, and long-lived predator species such as swordfish and tilefish generally have the highest contaminant levels.
Fish are also contaminated with dioxins, which are a byproduct of incineration. But that concern isn’t fish-specific because dioxins are found up and down the food chain, and we get more dioxins from meat and dairy than from fish.
In recent years, several scientific organizations have tried to weigh the risks of eating fish against the benefits. Notably, two groups — the Institute of Medicine (IOM) in 2007 and the Food and Agriculture Organization of the United Nations working with the World Health Organization (FAO/WHO)in 2010 — convened panels of nutritional, toxicological and epidemiological experts to review the data and do the math.
Both groups issued guidelines for pregnant women and other adults, and their recommendations are remarkably similar.
For pregnant women, their advice is in line with that of the well-known FDA/EPA joint advisory issued in 2004: Eat up to 12 ounces of fish per week, focusing on low-mercury fish and avoiding the four highest-mercury fish (swordfish, shark, tilefish and king mackerel).
For other adults, the advice could be summarized as “eat fish.”
Dariush Mozaffarian of the Harvard School of Public Health, a member of the FAO/WHO panel, sums up the state of fish science: “As long as you’re not a pregnant woman, the evidence suggests that the balance is always toward net benefit.” Although it’s theoretically possible that a fish could be so high in contaminants and low in omega-3s that it could do you more harm than good, no such fish has been found. Even so, if you eat more than two servings of fish per week, both his group and the IOM advise that you don’t always eat the same type, so you’re not eating one contaminant over and over.
So why aren’t we all just eating fish and not worrying about it?
Because of warnings such as those from the Environmental Defense Fund. Its guide to fish consumption lists 21 species of which adults should eat no more than one serving per month, and that’s “assuming no other contaminated fish is consumed.”
So one serving of flounder, perch or blackfin tuna would be your entire monthly allowance, and your choices for the rest of the month would be limited to the least-contaminated fish, such as halibut, sole or yellowfin tuna.
Why does the EDF advice contradict that of the IOM and FAO/WHO? Because, unlike those organizations, the EDF — considers only risks: It compares contaminant levels in fish to the limits set by the FDA and EPA, and advises accordingly.
Timothy Fitzgerald, who developed the EDF guide, acknowledges that the guide does not try to balance risks and benefits. “There’s no widely accepted way to do that, and we tried very hard not to create our own set of equations,” he says.
Mozaffarian disagrees with that approach.
“We have smart, well-meaning scientists who have been educated in this framework, in which you assess risk and come up with tolerable intakes,” he says. But that’s the wrong framework, he says, because you don’t eat pure contaminants, you eat fish, and you can’t get the risks without the benefits.
There is another problem with risk-only advisories: the issues with the FDA/EPA safe thresholds for mercury and PCBs. For mercury, that threshold is 0.1 micrograms per kilogram of body weight per day. For a 150-pound person, that would be fewer than seven micrograms a day. Swordfish, one of the most contaminated fish, can have 100 micrograms — two weeks’ allowance — in a 3.5-ounce serving.
For PCBs, it’s a more complex calculation of cancer risk called a cancer slope factor.
Both those numbers raise problems. The threshold for mercury was developed with reference to fetal neurodevelopment and, because the EPA doesn’t issue separate numbers for separate groups, it set the threshold at the level advised for pregnant women, the most vulnerable subset of the population.
The PCB threshold was developed in 1997, from rat studies. Since then, researchers have questioned the applicability of the data to human risk, and in 2003 a review of all available research concluded that “the weight of evidence does not support a causal association for PCBs and human cancer.” That conclusion was echoed in a follow-up 2009 study.
So not only do the EDF warnings fail to account for the benefits of eating fish, they may also be overstating the risks.
And it’s not just the EDF: The Natural Resources Defense Council, Physicians for Social Responsibility and other groups issue similar warnings, and the EPA thresholds underpin many guides to eating seafood. The Monterey Bay Aquarium uses the EDF’s evaluation to put red asterisks next to fish whose consumption should be limited “due to concerns about mercury or other contaminants.” Many states put out similar warnings, and a 2008 study of those advisories found that not a single one emphasized benefits as much as risks, and 25 percent of advisories didn’t mention benefits at all.
Fitzgerald points out that the data from peer-reviewed research don’t tell the whole story of risk. There is anecdotal evidence for neurological problems from consumption of mercury — mostly subtle neurological problems including impairment of peripheral vision, lack of coordination of movement and muscle weakness. And experts on both sides of this question agree that it is theoretically possible to get mercury poisoning from eating a lot of high-mercury fish. We don’t know how much is too much, but it is more than Americans customarily eat.
But not eating fish is also dangerous. As Mozaffarian and his colleagues at Harvard point out, the risk of dying from heart disease is about 50 percent higher among people who don’t eat fish than among those who get one or two servings of a high-fat fish each week.
Since the average American eats 15.8 pounds of fish and shellfish a year (just under five ounces per week), the risk of too little fish seems to be the bigger threat. And it’s a threat that those risk-only advisories — and even some of the more balanced advisories — may be contributing to by scaring people away from eating fish.
Jay Shimshack, an assistant professor of economics at Tulane University in New Orleans, has studied consumer response to fish advisories, particularly the 2001 FDA warning about mercury in fish, which had the unintended consequence of decreasing overall fish consumption — not just of high-mercury fish, and not just among pregnant women. Even when an advisory is carefully crafted and notes the benefits of fish, people take away a message of danger. “My research suggests that people may be overly focused on the negative aspects and not focused on benefits,” says Shimshack.
Mozaffarian notes that people tend to be more frightened by the threat of a harm than encouraged by the promise of a benefit. “People get confused; they get the wrong message.”
The groups that issue risk-only advisories are using the tools at their disposal to promote public health, and the possibility that the advisories have had the reverse effect is a testament to the inadequacy of those tools to deal with an issue that is complex and nuanced. Public-health professionals are struggling with developing and communicating advisories that encourage optimal fish eating.
While we’re waiting for advisories to improve, the best advice is simply to eat fish; the data show that the benefits outweigh the risks. If you’re pregnant, follow the FDA/EPA guidelines. If you’re not, mix up your species, and be sure to include smaller, bottom-of-the-food-chain fish. If you’re an angler, check with local authorities for contaminant levels where you fish, and limit your consumption if those levels are high.
Haspel writes and blogs (at starvingofftheland.com) about food and health from Cape Cod, where she also fishes.