There are severe limitations in what can be concluded for the meta-analysis, however. Because the cognitive decline in both treatment and control groups was so small, the trials did not have sufficient power to detect any possible benefit, much less to draw definite conclusions about the ability of fish oils to prevent Alzheimer’s disease. Moreover the fish oil was taken by subjects for at most 40 months; perhaps if it were taken for longer, and the follow-up were also for longer, a difference would manifest itself. Since the trials excluded people with dementia that was already manifest, it was impossible to conclude from the results whether or not fish oil is of benefit in established dementia.
The authors did not conclude that people should not take fish oil supplements, because they said that such fish oils might have benefits other than in the prevention of cognitive decline. And since longer trials might be necessary to establish definitively the uselessness (or otherwise) of fish oils in the prevention of cognitive decline, those who put their faith in them are not yet forced by the evidence to abandon it or risk joining the ranks of the irrational.
For the moment, nutritionists recommend the consumption of fish twice a week, including of oily fish — salmon, herrings, mackerel or sardines — at least once a week. I confess I find such recommendations suspect: how do the nutritionists know that three times a week would not be better, or once a week as good as twice?
The decision as to what to eat cannot be taken on the basis of placebo-controlled double blind trials, first because such trials usually establish very little (and that little is often contradicted by subsequent trials), and second because there are purposes to eating other than the preservation or improvement of health. Meals are not medical procedures, and the dinner table is not, or not yet, an operating table.