Subscribe to the OSS Weekly Newsletter!

Good Bacteria to Prevent the Flu? Interpreting the Data Is Messy

A recent CBC Marketplace episode may have made you think that eating probiotic yogurt would protect you from the flu. But what the studies actually show is how messy research can be on its way to consensus.

“We have a ton of studies that have shown that eating [Danone’s DanActive yogurt with L. casei probiotics] at a dose of about two containers per day can actually decrease your incidence and severity of things like cold and flu.”

So claimed postdoctoral fellow Mary Scourboutakos on the investigative program CBC Marketplace, which aired on April 6.

“A ton of studies”?

I have to admit: my first impulse was to sneer. This sounded like nonsense. I was ready to debunk it. After all, Danone’s fingers were slapped by the U.S. Federal Trade Commission eight years ago for making this exact claim in the absence of convincing evidence.

But let there be a lesson here. Skepticism is not the same as denialism. I realized my current knowledge might be outdated. Before dismissing a claim as erroneous, I had to investigate it. So I jumped in the confusing mess that is the scientific literature to answer a deceivingly simple question: can probiotics protect us from upper respiratory tract infections (URTIs), like influenza and the common cold? Can I really keep the flu away by eating yogurt?

A probiotic is a bacterium that has beneficial effects in humans and animals. A trip to the grocery store will demonstrate that there’s a lot of interest in these so-called “good bacteria”. They exist in naturally fermented products, like kimchi, kombucha, and kefir, and are added into other foodstuffs, like breakfast cereals and probiotic-infused yogurts.

When wading through the scientific literature on the issue, the first hurdle comes from seeing how many of these studies are supported by and often conducted by industry. The name “Danone” accompanies many of the published trials, as well as companies like Otsu Nutraceuticals (thanked for their “expert assistance”) and Heinz (which provided an “unrestricted grant”). Industry money does not nullify the value of a study; in fact, it is becoming more and more difficult to conduct research on commercial products without the financial assistance of industry these days, a marriage often encouraged by universities themselves. There are also degrees of involvement that must be clearly separated in the interest of fairness. An unrestricted grant reflects an academic liberty that a study funded, planned, executed, and analyzed by industry does not. That being said, an extra degree of skepticism is warranted, because these companies stand to directly benefit if the results are positive.

One way to artificially nurture positive results is to take a lot of measurements. The literature on probiotics for URTIs is rife with this. Researchers are reporting on the number of infections, their average duration, their severity, but also on the total number of common infections  (including the so-called “stomach flu”, which is unrelated to influenza), change in behaviour because of illness, time to first occurence of an infection, number of days absent from school or work because of an infection, and on and on. They might as well measure and compare the dispersal pattern of each sneeze while they’re at it.

Measuring so many outcomes allows for an almost limitless number of analyses. Indeed, these trials—in which half of participants received a placebo and the other half got a probiotic—usually have a “primary outcome measure”: this is the comparison the trial was designed to answer. But, as in the case of the Danone-sponsored French trial conducted by Guillemard and colleagues in 1,000 shift workers, you can decide to add more analyses when your primary outcome measure turns out to be negative. They gave Actimel, a probiotic-containing yogurt, to half of their cohort during the winter months and compared how many common infections they got compared to the placebo group. The result? No significant difference. But they didn’t stop there. They conducted many more analyses, and found that the proportion of participants with at least one infection was lower in the Actimel group. Also, there was a significant finding when they only looked at smokers. The more you look, the more you’re bound to find something, but it may only be due to chance.

Another issue is that not all probiotics are the same. There’s Lactobacillus casei, the species mentioned on Marketplace, but also Lactobacillus pentosus, Bifidobacterium lactis, Lactobacillus paracasei, and others. There are also subspecies and strains. Comparing studies that tested different species may be like comparing apples to oranges. Moreover, some participants had to eat yogurt, while others received probiotic-enriched milk, or tablets, or capsules, or dead probiotics in powder form to be added to a liquid of their choice. While juggling all these variables is daunting, some people have looked at the evidence and tried their best to issue a weighted summary.

The Cochrane Collaboration, a well-respected international body designed to summarize the evidence on health matters, issued a report in 2015 on this very question, analyzing 12 published trials. They found evidence that probiotics did seem to lower the number of people with an acute URTI, as well as the mean duration of an infection and cold-related school absence… but they labelled the evidence for all of this as “low to very low quality”. They highlighted the fact that different bacterial strains were used, many trials were poorly conducted, some samples were very small and, oh yes, industry money was a red flag.

A different meta-analysis published a year earlier and funded by the Global Alliance for Probiotics (ahem) found similar results. But the size of the effect is very small.

What Scourboutakos said on Marketplace is a little exaggerated. To be fair, scientists aren’t typically used to the cameras and, in the heat of the moment, we may end up slipping in our wordage. Yes, there are studies showing a small protective benefit from probiotics with regards to the common cold and the flu, but these studies are far from perfect. Assuming they were reliable studies, though, would we get the same benefit from eating commercial yogurt?

Scourboutakos’ own paper on this, entitled “Mismatch between Probiotic Benefits in Trials versus Food Products”, highlights the gap between what participants are fed in the context of a clinical trial and what consumers can buy in a supermarket. She and her team write that “many of the current probiotic dosages in products were lower than the dosages tested in randomized controlled trials.” Indeed, “consumers would need to eat anywhere from two to twenty-five servings of these products each day.” While eating two dozen containers of yogurt a day may be some people’s idea of culinary heaven, I doubt their waistline would thank them.

I don’t want to dismiss all this evidence, because this small but positive signal seems to persist across many studies. It is indeed possible that certain probiotics we ingest find a way to enhance our immune system to help us fight infections. But given how small the effect appears to be (if it is real at all), I would rather focus on more meaningful interventions. As cardiologist and fellow science communicator Christopher Labos told me, “If you want to protect yourself from the cold or the flu, wash your hands.”


Want to engage with this content? Comment on this article on our Facebook page!

Back to top