Subscribe to the OSS Weekly Newsletter!

Register for the OSS 25th Anniversary Event

Assessing Reports Linking Espresso and Cholesterol

I was told I needed to contextualize this thing quickly, ideally before dinner. I was able to reassure my friend in short order.

This article was originally posted in the Montreal Gazette.

A friend of mine recently sent me an article titled Espresso Associated with Increased Total Cholesterol. I was told I needed to contextualize this thing quickly, ideally before dinner, after which, my friend noted, good Italians drink their espresso.

I was able to reassure my friend in short order. Whenever you get faced with an article like this you always have to ask the right follow-up question. In this case, “By how much?”

First off, we should remember that cholesterol is a complex thing that is affected partially by diet and partially by genetics, and in many cases, it is genetics that makes the biggest difference. Also, the health effects of coffee vary widely, depending on what you put in it. For example, black coffee has almost no calories in it whereas adding milk and cream and sugar invariably turns it into something more resembling a liquid dessert than a casual beverage. Flavoured coffees in particular are especially high in calories and those looking to lose weight would be well advised to steer clear of them.

Still, given the ubiquity of coffee consumption, any health concern has the potential to have significant public health implications. The study in Open Heart that inspired the article used survey data from Norway to draw links between people’s total cholesterol and their coffee consumption. However, asking people how much coffee they drink will always be prone to some degree of error because of the imperfections of human memory. Also, with a cross-sectional study like this, measuring things at one point in time is problematic because when it comes to diet and health, how things change over time is particularly important.

Two big caveats should limit how much weight we give to this study. First, total cholesterol is not the best metric to assess your blood lipids. While cholesterol is more complex than good (HDL) vs. bad (LDL) cholesterol, it is a useful clinical division and any study claiming to show a difference in cholesterol levels begs the question of whether the bad cholesterol was actually affected at all.

Then, there is the issue of degree. Small changes in cholesterol, blood pressure, blood sugar, and virtually any health parameter are always visible if you track enough data points. We must always remember that some changes might be statistically significant, but not rise to the level of clinical significance. In this study, drinking three to five espressos per day (which is bordering on excessive) increased total cholesterol by 0.09 mmol/L in women and 0.16 mmol/L in men. These changes, even if real, are fairly small to the point of being irrelevant in clinical practice.

What’s more, previous randomized trials don’t support this study’s findings. In general, randomized trials avoid some of the pitfalls that we can see with observational research like the study in Open Heart and tend to return more reliable results. Also, the main focus of any research is to prevent disease. Ideally, one would like to see studies measure things like heart attacks or strokes to demonstrate a clear benefit or harm. By looking only at cholesterol levels, it is hard to link this study back to a meaningful clinical endpoint.

Essentially, there are multiple reasons to be dubious of any significant health risk. Even if we can overlook the methodological limitations of this type of cross-sectional study and the pre-existing data that refutes it, the small differences in total cholesterol seen in those who drank more coffee than others are not important enough to generate any meaningful concern. There is no shortage of claims that some foods will either lower or raise your cholesterol. The important question to ask is whether it lowers or raises it enough to make any real difference to your health. Usually, the answer is no.


Back to top