Turmeric is a brilliant spice. It is the reason that many curries look so wonderfully colorful, as well as adding that bitter twang that is so redolent of Indian cuisine.
And, if you’ve been watching the news recently, it is apparently the newest cure for poor mood and a wonder-drug that improves memory in older people. A new study has come out apparently proving the benefits of turmeric once and for all!
It sounds amazingly exciting. Maybe we’ve been doing it wrong all along? Maybe the cure to mood and memory disorders isn’t, as we’ve tried before, a complex mixture of drugs and counselling, but actually just a big serving of tasty curry.
It would be pretty amazing.
There’s only one small problem.
Turmeric has no health benefits at all.
The first thing that everyone who is talking about turmeric got wrong is pretty simple: the recent study was not on turmeric at all. The scientists actually studied curcumin, which is a compound that is found in turmeric.
The distinction might seem meaningless, but it turns out that you don’t actually get much curcumin when you eat turmeric. By weight, turmeric is roughly 3% curcumin, and because of the way that turmeric is absorbed by your body, you only actually use about 25% of that amount in the body.
The amount of curcumin that was used in this study is equivalent to about 25 grams (or ~1 ounce for the Americans out there) daily. I had a look at my spice cupboard, and that’s roughly the amount that you’d expect to find in your average supermarket container.
So, anyone who talked about the “health benefits of turmeric” was simply wrong. There aren’t any. There may be benefits of taking large amounts of curcumin, but to do so you have to eat thousands of curries worth of turmeric, making it something of an ineffective treatment.
So what about the study on curcumin?
The recent study that everyone is happily obsessing over was a trial of curcumin supplements that looked at whether older people’s scores on a variety of memory tests improved when they took curcumin vs when they took a placebo. The scientists also used a fancy brain scan to see if there were any changes to their brains during this time.
Basically, the study found that taking curcumin in supplement form twice a day improved scores on some memory tests. They also found that there was a correlation between improved memory scores and lower amounts of Alzheimer’s-associated chemicals in the brain. They hypothesized that the improved test scores were to do with the lower levels of these chemicals, and that the curcumin supplements were somehow preventing them from building up in the people’s brains.
This was also what’s known as a Randomized Controlled Trial, or RCT. They are generally considered to be the highest form of clinical evidence, as they eliminate many of the biases that plague most research.
So why aren’t I celebrating with a massive curry and some memory exercises?
There are several reasons that this study proves nothing whatsoever about curcumin. The first is the number of people in the study. To provide good evidence for health benefits, you’d usually expect a study like this to have at least 70 people in each group, with 140 total*. That wouldn’t be ideal — you would have what statisticians call an “underpowered” study, which isn’t very good at determining cause and effect — but at least it would be something.
This study was done in 40 people total. 20 in each group. Less than a third what you’d need at minimum.
Another thing that makes this study problematic is the number of statistical tests. Stats tests are a bit like flipping a coin: do enough of them and you’ll end up with a positive answer eventually. To their credit, the researchers did take some steps to adjust for this in their analyses, but there were still more than 30 individual tests in the study. More importantly, many of these were negative — you only hear about the positive scores in the media, but the scientists also found, for example, that there was no difference between the placebo group and the curcumin group in terms of visual memory.
Sadly, “Turmeric Doesn’t Aid Visual Memory” doesn’t make a very good headline.
Perhaps even worse than all this is that the leading headline — “Turmeric/Curcumin Improves Mood/Memory” — is simply wrong. This study didn’t find that mood improved at all. It did demonstrate some changes in the parts of the brain that are associated with mood, but that’s a very different thing to actually proving that curcumin is helpful for your mood.
Last but not least, this was an industry-funded trial. That doesn’t mean it’s worthless, but it does mean you have to take the findings with a grain of salt. There is strong evidence that industry-funded trials are more likely to come up with positive conclusions than neutrally-funded ones, and it’s worrying that nowhere in the media do you see a mention of the company that is behind all this hype: Theravalues Corporation, who make curcumin supplements.
So, no, eating turmeric probably won’t improve your health. But really, we knew that already. Curries aren’t some magical ticket to good health, after all. Sadly, that isn’t the worst part of our turmeric adventure.
Virtually every story that was written about this study was wrong.
Some media sources managed to talk about curcumin, rather than turmeric. But even the best story I could find, from Forbes, didn’t mention that actually there’s no reason to believe that curcumin helps with mood at all.
There may be some health benefits from curcumin, although the evidence is very much mixed. They will likely not be particularly big, because if they were we would’ve found them in one of the hundreds of clinical trials that have been conducted on curcumin. One small, industry-funded trial adds almost nothing to the evidence, even if they did make some positive findings. But overall, one thing is clear.
Turmeric is likely worthless for human health.
Curcumin might be useful, but at the moment the jury is very much still out.
Don’t believe the hype.
*This is based on some quick power calculations I did. To detect a difference such as the study found, with a mean difference in scores of ~15 on the tests and with a standard deviation of ~30, you’d need a sample size of 70+ in each group, with an alpha of 0.05 at a power of 80%.