The argument over whether farmed salmon is safe to eat has been as frenzied as the fish at feeding time. Not only has it alarmed the industry and confused the public, it has also left scientists at odds over how to assess food safety.
Some researchers say the contamination of salmon is within World Health Organization (WHO) safety limits, some say it isn't, and some say stricter limits are required anyway. And to make the confusion worse, there is no clear way of deciding how to balance the health risks of the toxic contaminants against the known benefits of eating fish.
Some facts are not in dispute. Salmon reared in cages contain about 10 times the quantity of PCBs, dioxins and pesticides as wild salmon, because these pollutants are concentrated in the fish oils that make up about 30 per cent of their diet. Paradoxically, one reason the fish oils are fed to salmon is that they are rich in omega-3 fatty acids, which are thought to help reduce the risk of heart attacks in people.
The difficulties come when trying to turn these facts into recommendations for the public. The controversy over last week's study by US researchers (Science, vol 303, p 226) centres on their use of a methodology from the US Environmental Protection Agency (EPA) for estimating the risk of the pollutants causing cancer.
This resulted in the suggestion that people should eat no more than 55 grams a month - less than half a normal 135 gram portion - of the most contaminated salmon from Scotland and the Faeroe Islands if they wanted to keep their additional risk of contracting cancer below 1 in 100,000. This recommendation does not account for health risks such as damage to the immune and reproductive systems.
But the EPA risk estimates have been attacked as "out of date" by the chairman of the UK Food Standards Agency (FSA), John Krebs, because they assume that there is no threshold below which dioxins are safe. Krebs says the correct guidelines come from the WHO, which take into account the mechanism by which dioxins cause cancer. The WHO says that a daily intake of up to 4 picograms per kilogram of body weight is "tolerable". This would allow the eating of up to three normal portions of salmon a week.
But Krebs's line is challenged by other government researchers in the UK. Mel Holmes, Andy Hart and Martin Rose from the Central Science Laboratory in York point out that some of the WHO and FSA's assumptions may underestimate the risk. There are large variations in the potential toxicity of different dioxins, they argue, and if these are taken into account, most people eating a single portion of salmon a week could exceed the WHO limit.
In a paper at the Dioxin 2003 conference in Boston, they say sensitive individuals may be particularly at risk and toxicity data should be re-examined. But the FSA dismisses their analysis as a "worse-case calculation" that "would have the effect of exaggerating the intake of dioxin".
The disagreements over safety limits, however, do nothing to answer the key question of how to balance the damaging impact of the pollutants against the beneficial effects of eating oily fish. This calculation was not done by the US study, and is still being investigated by one of the FSA's expert committees, which is expected to report later this year.
The question is important because it does not just concern salmon. Farmed sea bass are more contaminated with PCBs than wild sea bass, say Portuguese scientists (Chemosphere, vol 54, p 1503). And the same is likely to be true of animals, including sheep and other livestock, that are also fed fish oils to boost their levels of health-giving omega-3 fatty acids.