by Patrick Holford
One New Year tradition, along with wondering how you could have spent so much on so little, is making healthy resolutions which may involve some sort of supplement. As a nutritional therapist I’ve been helping people for years to handle chronic disorders and stay healthier by recommending specific nutritional supplements based on risk, coupled with substantial changes in people’s diet. There’s plenty of evidence it benefits people but there’s lots we still don’t know and new well-conducted research is always welcome.
However another long standing tradition is the regular publication of studies that don’t help to clarify anything. They claim to have found that certain supplements are ineffective and possibly dangerous, leading to news stories about deadly or at least unnecessary vitamins and expensive urine. I don’t object to negative conclusions – they can be very useful – but these widely publicised studies are invariably done with little understanding of nutrition and their headline-worthy conclusions often bear little relation to what they actually found.
Two classic examples of these toxic studies came out just before Christmas, both published simultaneously in the same journal, presumably for maximum impact. One investigated whether multivitamins could prevent dementia, the other whether they could reduce cardiovascular risk in people with a history of heart disease. No one who knows about nutrition or who is interested in adding to our knowledge of what works and what doesn’t would have conducted either of these poorly conceived trials.
Vitamins can slow brain shrinkage
Why try to prevent dementia by giving a fairly low-dose multi-vitamin when there is already a very good randomised trial showing that it is possible to slow down brain shrinkage and memory loss, the two hallmarks of Alzheimer’s, by giving high doses of B vitamins to those at risk. That’s about half the population over 65.
Giving multivitamins to heart disease patients, who are also on a range of powerful drugs, is equally ignorant. The informed approach would be to use specific supplements known to benefit heart disease on patients who were deficient. For example, giving B vitamins to those with raised levels of the amino acid homocysteine, which is a marker for heart disease, or antioxidants to those with a high level of oxidative stress.
The studies were published in the Annals of Internal Medicine  which ran an editorial that provided the scare headlines but which didn’t remotely reflect what the studies actually found, as will become clear. The key part read: “Antioxidants, folic acid, and B vitamins are harmful or ineffective for chronic disease prevention, and further large prevention trials are no longer justified. Most supplements do not prevent chronic disease or death, their use is not justified, and they should be avoided.”
This is pure prejudice totally disconnected from the results of the trials. These were trials of “multivitamins” and as such tell you nothing about ‘most’ supplements – just the ones on trial. They also came up with no evidence of harm and the cardiovascular trial was not about preventing heart disease (primary prevention), but only reducing further harm once it has developed (secondary prevention).
The study ran into big trouble
So what did these studies actually show? The first was a sub-group of a larger trial. Patients aged 50 or older who had had a heart attack within the previous six weeks were given a multivitamin supplements, at quite high doses or a placebo There were 1708 people at the start of the trial which was intended to see if the vitamins reduced various consequences of heart disease (primary endpoints) such as death, repeated heart attacks , stroke, coronary revascularization, or hospitalization for angina.
But the study ran into big trouble. The patients were meant to take the supplements for at least 3 years but half of them (46%) failed to keep taking the vitamins for that long and 17% dropped out completely. In the end there were just 400 patients in the vitamin group and 426 in the placebo group. The primary endpoint(s) occurred in 27% of patients taking vitamins and 30% of patients taking placebos. Without any treatment you would expect 30% to have one of those cardiac events, so 3% fewer of those on vitamins had them. That’s actually a 10% reduction or a potentially life-saving benefit for one in ten.
However the small size of the trial, made worse by the fact that so many people had stopped taking the vitamins or dropped out, meant that this result couldn’t tell you anything useful at all. Trials have to be a certain size to make sure (technically known as “powered”) that the result didn’t happen by chance. This study was originally powered to detect a 25% reduction in risk so the official ‘statistical’ conclusion was that there was no difference between the risk between those taking the vitamins and a placebo. The positive trend, however, indicates that the trial should be repeated with enough participants to detect a 10% reduction in risk.