You may have heard recently that red wine and coffee have been linked to having a healthier or more diverse amount of gut bacteria.
The headlines are the result of a Dutch and Belgian study that each sought to try and narrow down various factors that could influence your intestinal microbiome.
The data is interesting and definitely has potential, but this is a very good example of putting the cart before the horse and highlights a number of pitfalls inherent to trying to understand how your intestinal bacteria influences your health.
Red Wine, Coffee and Microbial Diversity: Study Summaries
- The Dutch study looked at 1,135 adults from the Netherlands and the Belgian study looked at 1,106 adult Belgians
- Fecal samples were taken from participants and lifestyle and dietary questionnaires were administered
- The samples were then sequenced to identify various bacteria levels, which were then compared with many different factors to test for correlations
- The Dutch study found around 126 different factors showed a correlation to gut flora levels, including age, height, weight, sleep, smoking, allergies, molecules in the blood, whether participants were breast fed as babies, method of birth, and diet
- The Belgian study looked at similar factors and has a 92% overlap with the Dutch results
- Red wine and coffee were among the factors linked to an increase in microbial diversity and things like irritable bowel syndrome were associated with lower diversity
- It is worth noting that only 19% of biome changes were correlated to the studied factors in the Dutch study, with the Belgian study showing just an 8% connection
- There is no proposed mechanism of action for some of the observed correlations
Caveats to Keep in Mind
The data is definitely interesting and there may be some useful nuggets to be found, but unfortunately the hype around this research is not warranted. While these are both fairly large studies and the overlap in findings is a good sign, there are still a number of important reasons to remain cautious about the research.
One of the biggest issues with these findings is that the microbial levels were compared to a ton of different factors. It’s important to note that there is no denominator available when looking at the 126 factors the Dutch study found correlations to. This means it’s not currently possible to tell how many total factors were being examined.
As mentioned before in previous research discussions, every time you bring in an additional point of comparison you increase the chance of getting a false positive. The threshold of statistical significance is when calculations show a given finding has a 5% or lower chance of being the result of random luck (i.e.: a false positive). With 126 correlations found in the Dutch study (or about 116 if you go by the 92% overlap), that’s over a hundred sources of erroneous findings.
There are ways to correct for the risks of doing multiple comparisons, but there don’t seem to be signs that this was performed. This doesn’t automatically mean the studies should be tossed, but it does mean their results—even with partial corroboration—need to be taken with a significant grain of salt. Remember, both the Dutch and Belgian studies were basically taking a kitchen sink approach to finding something and science works better the more specific and precisely it operates.
While the studies did their best to maintain the integrity of the fecal samples, they only took one from each participant. This is important since microbial levels can and do fluctuate throughout the day based on a host of different factors, something that people studying over 126 possible influences should know. Taking a single “snapshot” of gut flora means nothing if you can’t verify that these levels are representative of the person’s normal levels.
This last point is less significant than the others but it bears keeping in mind. There is some overlap between the people who wrote the Dutch study and the ones who worked on the Belgian study and subsequent comparison write-up. While not an inherent problem, it’s a possible source of bias to have people with ties to the original research working on its replication.
- Replication is only as good as the rigor involved.
- Large sample sizes can be good, large numbers of comparisons are not.
- There are significant issues affecting the reliability of the studies and concerns about their likelihood of false positives.
- Even if all of the findings are 100% accurate, there’s no evidence any of the observed microbial changes positively or negatively impacted health, meaning any touting of the studies’ “meaning” is premature at best and flat-out wrong at worst.
Yong, E., “Why Are Your Gut Microbes Different From Mine?,” The Atlantic web site, April 28, 2016; http://www.theatlantic.com/science/archive/2016/04/why-are-your-gut-microbes-different-from-mine/480207/, last accessed April 29, 2016.
Zhernakova, A., et. al., “Population-based metagenomics analysis reveals markers for gut microbiome composition and diversity,” Science, 2016; 10.1126/science.aad3369
Falony, G., et. al., “Population-level analysis of gut microbiome variation,” Science, 2016; 10.1126/science.aad3503.