As time goes by, studies of all sorts of things pile up. At some point no one can keep track of them. Even people with a good memory remember only the studies they like. This calls for a meta-analysis – a calculation of a sort of average of what all reliable studies have shown.
This has now happened in a British study of fish oil. Does fish oil protect the heart? Does it prolong life? Yes, we know that it does!
At least, so we think. According to the new meta-analysis fish oil has "no clear effect on mortality, risk of heart attack or the incidence of cancer" – rest assured, you do not need to eat fish!
As far as cancer goes, one is hardly surprised, but what about the heart? Ever since the two Danish doctors', Bang and Dyerberg's, studies on Greenland 35 years ago, everyone has known that fish oil protects the heart. This is confirmed by numerous studies. How can all these studies amount to a big fat zero when summarized?
Of course they cannot. A meta-analysis can be just as subjective as everything else. This means that others, with the same starting point can arrive at the opposite result. This is the case in a new, far more comprehensive, American report. There is no doubt that fish oil is beneficial for the heart. In fact, it is stated in this report that everyone should be tested to see if they get enough of it. The two important fatty acids in fish oil, DHA and EPA, have "clear beneficial effects". Everyone should ingest at least 1.5 grams of fish oil every day, in case of heart disease, double that. This will lessen the risk of dying from heart disease by 25 %.
That was the Americans. They thoroughly reviewed the extensive biochemical knowledge and conclusions from animal and human studies. From this they made a general conclusion. What did the British do?
The Test Subjects Cheated
They completely ignored all basic knowledge and concentrated on the incidence of heart disease and mortality in humans. But were they neutral?
They reviewed a total of 48 randomized studies. But they did something strange: Only fifteen of the studies were included in the mortality calculations. Why the remaining 33 studies were not included is not known. Maybe some were excluded because of suspicion of being biased. But, in nine of the fifteen studies included, there was "medium or high" risk of bias. For example because the test subjects knew whether they were given fish oil, or, because of the draw, were given something ineffective.
On top of this, an unknown number of the 48 studies were not even with fish oil but with alpha-linoic acid – which is found in flaxseed oil and in rape oil. These oils can be converted to "fish oil" in the human body.
Besides, the most famous study of linoleic acid is missing - the so-called Lyon experiment where rape oil (and a Mediterranean diet with olive oil) lowered the mortality in patients who had had a coronary thrombosis by 73 %.
When everything is boiled down, twelve randomized studies with fish oil remain. But of these, nine are very small and without relevance. This leaves three. The biggest of these three studies, (GISSI) show a massive reduction in mortality in twenty percent of persons who had had a coronary thrombosis. In the table it only says fourteen percent, apparently because of a misunderstanding.
Only one of the larger studies (Burr et al 2003) came out with a negative result. Here the mortality increased by fifteen percent in men with sclerotic coronary arteries, when they were given fish oil.
However, this study is at the least controversial, partly because the participants knew if they were given fish oil or not. Since they had a dangerous heart disease, many of them would have been tempted to take supplements on their own. This could have been checked through blood tests, but only a spot test was taken after six months. After this the study continued for three to nine years. The spot test showed that the difference between the two groups was remarkably small. Moreover, the participants explained that even if they were not in the group given fish oil, they ate so much fat fish (14 grams a day) that this alone would have provided them with the amount recommended in the American report. So, in fact, both groups received "enough", and the value of this study must be questioned.
Still, it was printed in the news paper. However, you should believe the American report.
Fish is good – of course!
- Hooper et al. Risks and benefits of omega 3 fats for mortality, cardiovascular disease, and cancer: Systematic review. BMJ online 24.3.06: BMJ, doi:10.1136/bmj.38755.366331.2F
- Wang C et al. Agency for Health Care Research and Quality. US Department of Health Care Rsearch and Human Services. www.abrq.gov . Evidence report/technology assessment Number 94. Effects of omega -3 fatty acids on cardiovascular disease. March 2004.
- Burr ML, Ashfield-Watt PA, Dustan FD, . , et al.: Lack of benefit of dietary advice to men with angina: results of a controlled trial. Eur J Clin Nutr 2003, 57: 193-200.