top of page

Ask Dr. Benzi: What’s the deal with dairy?

Healthcare Question: Should people with Parkinson’s disease (or people in general) avoid dairy products? I’ve seen several articles recently (shared below) that seem to suggest dairy is bad. – A

Reply: This is a great question and related to several other common questions people ask about diet, illness and health (e.g. Should I avoid gluten? Is sugar bad for cancer….). In addition to giving you my answer to this question (you’re always encouraged to do your own research and come to your own conclusions) I’ll also take you through the steps of how I approached the question.

  1. Reading past the headlines, do the articles come from trustworthy sources?

“A” did a good job of selecting 4 articles from reasonable sources. The articles are: a) Milk Linked to Greater Risk of Parkinson’s, Swedish Study Shows from Parkinson’s News Today; b) Does Consuming Low-fat Dairy Increase the Risk of Parkinson’s Disease? from the American Academy of Neurology; c) Oh, Dear: Dairy? from The Science of Parkinson’s; and d) Dairy Products and the Risk of Parkinson’s Disease from the American Journal of Epidemiology.

Often, when people send me headlines or questions, the “information” comes from a source with a high risk of bias, such as a website linked to a supplement manufacturer or an article from the author of a book promoting a specific diet. A’s articles instead came from: a) A disease-related news-site with sound editorial policies; b) A nationally recognized professional society; c) A blog from a scientist associated with a disease-related charity; and d) A peer-reviewed medical journal.

A second issue arises when the articles come to conclusions based on unreliable evidence, such as quotes from experts, testimonials, or just sciency-sounding babble. Instead, “A” chose articles closely linked to actual research studies.

2. Digging into the studies, what conclusions does the research actually support?

A research study is limited in the questions that it can answer, and our confidence in these answers, based on its design. When a news article’s headline seems too good (or too bad or too simple) to be true it is often because the writer of the article is sensationalizing the story and suggesting a claim that goes beyond what the actual research study design can support. Very commonly the results of a study in animals is headlined with the word cure (e.g. Cheese Cures Cancer, no really) but we know from lots of experience that animal studies are good for identifying potential therapies but rarely predict what will actually work in people (see this Blog for more). The other common sensational headline comes from observational studies, where an association found between two things across a large group of people is suggested to mean that one causes or prevents the other (e.g. Cheese Prevents Dementia, again-really). Again, we know from lots of prior research that observational studies can identify potential preventative or causal factors but cannot prove a causal relationship. In fact, most animal or observational studies lead to disappointing results when they are put to the test in a randomized clinical trial in people.

All four of the studies linked to A’s articles were large observational studies. In these studies, researchers collected data (e.g. diet information, or genes for lactose digestion in one fancy study) and then observed whether a particular health outcome happens (e.g. Parkinson’s). All 4 studies were consistent in finding small associations between certain types of dairy consumption (milk vs. yoghurt, low-fat vs. normal, more dairy vs. less dairy) and risk of Parkinson’s. Article “d” also included a meta-analysis (combining results across multiple studies) suggesting a consistent association in men but not in women.

We need to be very cautious about not over-interpreting observational studies. For one thing, correlations do not prove causation. The reason for this is that there may be other factors that explain the association. Shark attacks are associated with ice cream consumption—but this is better explained by looking at how warm weather influences both swimming and eating of frozen treats, rather than by trying to link sharky aggression to sugar and dairy. Similarly, people’s dietary habits are associated with many things (e.g. income, education, lifestyle & exercise, culture…) that could better explain the apparent link to dairy. Second, there are some inconsistencies within and across studies that deserve further study. For example, why should low-fat dairy (dairy missing milk fats) have a higher risk than full fat? Are milk fats protective?

At this point, we can say that there appears to be a small association between certain types of dairy consumption and Parkinson’s. We cannot (and should not) say based on this evidence that dairy consumption causes Parkinson’s, increases the risk of Parkinson’s, or hastens the progression of Parkinson’s.

3. Putting it into perspective.

So where do we go from here? I would add a couple of comments to put these findings into perspective. First, the strength of the association between dairy and Parkinson’s is small and the absolute risk is low. In study d, the rate of PD for low dairy was approximately 1 in 4,000 compared to 1 in 2,500 in the highest risk group. It is important to view the relative risk in light of the absolute risk because a small change in a small number can appear to be a big percent change. To put things in further perspective, the risks reported in these studies were 1.1-1.4 (10-40%) whereas the relative risk of lung cancer with smoking is 25-30 (2500-3000%). Second, prior association studies in Parkinson’s have found many association, including a mysterious but very consistent correlation between smoking and a lower risk of Parkinson’s. To date, none of these have changed clinical practice (and no one is recommending taking up smoking).

4. Making a recommendation.

Putting everything together, I personally am going to continue to enjoy dairy. For patients who ask, I would place this decision in context with their values. If people are very fearful of Parkinson’s, want to err on the side of doing “everything” (including things that have a low chance of changing their risk), and don’t mind giving up dairy I might recommend that they give up, restrict or limit certain types of dairy. There are no real risks to this decision, and it may help them, particularly if it would give them some peace of mind. For most people, particularly those who like dairy, I would go with the standard advice of “everything in moderation.”

If you’re interested in looking more into the dairy and health story I would recommend this review article from the New England Journal of Medicine or this summary. I particularly love this quote from Marion Nestle, a retired nutrition professor:

“Milk is a food like any other, meaning that its effects depend on everything else people are eating or doing. People who like milk can continue drinking it. Those who don’t like it don’t have to. It’s just a food.”

Take Home Points

  1. Make sure your news is coming from reliable (unbiased) sources.

  2. Be careful to not get drawn in by the headline. Read the research and see what the study design actually supports.

  3. Animal and observation studies are poor predictors of what matter to people.

  4. Always look at both the relative risk and absolute risk. Small changes in low risks can become large percentages.

  5. When making decisions, put the data in perspective and measure the benefits and risks in light of your values.


bottom of page