Valentine鈥檚 Day is rolling around and we can expect a flood of articles about chocolate. Some will trot out the tired old trope about phenylethylamine in chocolate stimulating amorous adventures by enhancing the release of dopamine, the 鈥渇eel good chemical鈥� in the brain. Don鈥檛 bank on it! While chocolate does contain phenylethylamine, it is metabolized before it can reach the brain.
We can also expect endless articles about the health benefits of chocolate ranging from lowering blood pressure and neutralizing damage-causing free radicals to clearing plaque from arteries and preventing memory decline. The scientific publications cited in these articles are often funded by the chocolate industry, are peppered with words like 鈥渕ay鈥� and 鈥渃ould,鈥� and end with the phrase 鈥渕ore research is needed.鈥�
Chocolate is not unique in this regard. Scarcely does a day go by without a study that suggests consuming some food or supplement to boost our health or urges us to avoid a food or ingredient that undermines it. This is usually followed by a bevy of media headlines that make more of the results than is warranted.
Such is the case for a study recently published in the British Medical Journal with the rather innocuous title 鈥淐hocolate intake and risk of Type 2 diabetes: prospective cohort studies.鈥� This precipitated numerous headlines along the lines of 鈥淪weet news: Dark chocolate reduces Type 2 diabetes risk.鈥� Some headline writers do try to derail criticism with the usual weasel-qualifying words as in 鈥淗ow Sweet! A daily dose of dark chocolate may cut your risk of diabetes.鈥� Of course, 鈥渕ay not鈥� could apply just as well.
The study in question is typical of many similar ones carried out these days that involve the 鈥渕ining鈥� of data collected from prospective cohort studies. Such studies ask a large number of participants to fill out detailed food frequency questionnaires and then report over a period of years any health issue they encounter. The goal is to detect any association between some component of the diet and disease, but such observational studies can never prove a cause-and-effect relationship. They can, however, serve as a springboard for further studies.
In the case here, the results of three large cohort studies involving a total of 111,654 participants of whom 18,862 were eventually diagnosed with diabetes were pooled and the data 鈥渕ined鈥� for a link between diet and disease. After torturing the data until it yielded some result, the researchers concluded that while milk chocolate had no effect on diabetes, five or more servings of dark chocolate a week, compared with no consumption, reduced the risk of developing diabetes by 21 per cent!
That sounds pretty significant and makes one want to reach for that daily dose of dark chocolate. But before jumping to that conclusion, a little further investigation is warranted. First of all, there is the usual concern about food frequency questionnaires. Memory about what has been consumed is unreliable, and people are notoriously inept at estimating amounts. In this study, participants were asked to report their average frequency of consumption of one ounce of chocolate in the past year, choosing from nine levels ranging from 鈥渘ever, or less than once per month鈥� to 鈥渕ore than six per day.鈥� How reliably can one remember one鈥檚 chocolate consumption over a year? How do you factor in chocolate that was consumed by eating chocolate cake, gulping M&Ms, drinking hot cocoa or plucking a couple of pieces from that box of chocolates?
And the questions don鈥檛 stop there. There isn鈥檛 only one type of dark chocolate. Cocoa content of bars can range from 60 per cent to 90 per cent, meaning significant differences in chemical composition. Lindt even has a 100-per-cent cacao bar that has no sugar at all. I think one would remember how many times that was consumed over a year. For me, it was once. I can even predict my consumption over the next year. None.
Then we come to the different cultivars of the cacao trees that produce the pods containing the cocoa beans from which chocolate is produced. Depending on the cultivar, the chemical composition of the beans can vary. Once the pods are harvested and the beans are extracted, they are allowed to ferment. This means that the naturally occurring yeasts found all around us break down the sugars in the pulp surrounding the beans to produce ethanol. This is first converted by bacteria into lactic acid, which in turn yields acetic acid upon the action of other bacteria.
Aided by the heat produced during fermentation, the acetic acid breaks down the walls of the cacao bean鈥檚 cells and then degrades the proteins released into amino acids and peptides. Through a series of reactions these are converted into several families of polyphenols that are generally believed to be the beneficial components of chocolate. Given all these variables, the chemistry of different dark chocolates can be very different. Consuming a dark chocolate bar with 90-per-cent cocoa that comes from cacao trees grown in Ghana can have quite a different effect on the body than a 60-per-cent dark chocolate bar with cocoa originating from trees grown in Brazil.
Obviously, there is much uncertainty when it comes to amount and type of dark chocolate consumed. To this we can add another degree of uncertainty when we look at the claim of 21-per-cent reduction of diabetes in participants who consumed more than five servings of dark chocolate a week. Reporting results in percentages is always problematic because even a large per-cent change can be quite meaningless. Buying two lottery tickets instead of one increases your chance of winning by 100 per cent, but the chance is still trivial. So, what does a 21-per-cent reduction mean in this case?
I will spare you the tortuous scrutinizing of the data collected by the researchers from which we can glean that if 1,000 people who eat no chocolate were to start eating five servings of dark chocolate a week for a year, one of them would be saved from being diagnosed with diabetes during that year. Even this calculation is prone to significant uncertainty because out of 4,771 people who were diagnosed with Type 2 diabetes, only 132 consumed more than five servings of dark chocolate a week. That鈥檚 a very small number on which to base statistics. Furthermore, as we have seen, we really don鈥檛 know what a 鈥渟erving of dark chocolate鈥� means in terms of the chemical components provided to the body.
There is also the question of lifestyle differences between dark-chocolate consumers and non-consumers. The researchers did make a valiant attempt to correct for this, but that is a challenging task. Estimating levels of physical activity from questionnaires is difficult, as is determining if the results were due to what was not eaten rather than what was. Maybe the five servings of dark chocolate replaced ice cream for dessert and it is the avoidance of the ice cream that resulted in the small benefit.
The bottom line here is that the many hours of research and the supporting funding yielded results that may be of some academic interest but have little practical significance. Media suggestions that adding five servings of dark chocolate a week to the diet with hopes of reducing the risk of Type 2 diabetes are misguided. Maybe if the chocolate is substituted for a sweet dessert. Or maybe not.
