A large, U.S. study confirmed what you probably suspected all along - taking vitamins is not going to extend your life.

This week – how do you tell what a nephrologist thinks of vitamins? Don’t worry, they’ll let you know.

Yes nephrons like me are fond of telling people that taking vitamins give you expensive pee, but it’s nice to see our flippant attitude bolstered by some real science, as seen in this study appearing in the Annals of Internal Medicine which suggests that taking vitamins has no effect on overall mortality.

Researchers examined data from about 31,000 individuals who were part of the National Health And Nutrition Examination Survey often referred to as NHANES.

Participants in NHANES complete a variety of questionnaires that capture vitamin and supplement use as well as a dietary recall to get a sense of what they are eating.

The researchers linked those responses to overall mortality to explore whether vitamins and supplements prevent, well, death, and if so – which ones do.

Now people who take vitamins may be subject to what is known as the healthy user effect. They are much more likely to do other healthful things too. Just over 50% of the population was taking at least one vitamin or supplement and there were quite a few differences between the takers and non-takers.

Those who partook of encapsulated nutrients tended to be older, more often female, more often white, more highly educated, and were less likely to smoke or drink alcohol.  They also had a higher “healthy eating index” meaning their regular diet was better quality as well.

But the vitamin-users did have more comorbid conditions – 12% had a history of cancer compared to just 6% of the non-users. The authors suggest that this could be due to individuals with chronic conditions taking their health into their own hands. I think it’s also likely that once you are taking a prescription medication, the threshold to add a pill another day is not such a high bar to cross.

After adjusting for all these differences, the authors found… basically nothing.

Out of 43 vitamins and minerals studied, only 1 – lycopene – seemed to have any protective effect. Don’t get excited ketchup lovers – due to the multiple vitamins tested there was around a 90% chance that at least one would pop up as a false positive.

So vitamins don’t do anything.

But – plot twist – food does. When the authors looked at those food diaries they were able to tell who was getting adequate intake of 22 nutrient – like niacin, where nearly 100% of the population has adequate intake, or fiber, where nearly 100% do NOT have adequate intake.

Two nutrients appeared to be protective in terms of all-cause mortality – vitamin K and magnesium, but only if you got them via food.

Here’s a comparison of the food-intake benefit of these nutrients compared to the benefit you get from supplementing these nutrients.

Here’s the thing – this study doesn’t show us that there is something magical about getting these nutrients in food compared to in pill-form that makes them protective. None of these nutrients are likely doing anything. Sure, if you have a severe deficiency of something, the nutrients help, but no one in this study had scurvy, or beriberi, or even rickets. If more Vitamin K is better, we should see a signal from the supplements. 

That we don’t teaches us an important lesson. Dietary research that purports to link a single nutrient to an important outcome is often hopelessly confounded by the strong relationships between the intake of that nutrient and ALL the other things you put in your body and do for your body.

People who get more vitamin K from food live longer not because they eat more vitamin K, but because they eat a host of things – like green leafy vegetables – that correlate with Vitamin K. Vitamin K is just a stand-in for a whole package of behaviors.

What’s a clinician to do? Don’t encourage your patients to eat more vitamin K or magnesium.  Encourage them to eat good, nutritious foods, and engage in all the healthful behaviors – like regular exercise – that really drive these observed benefits.

This commentary first appeared on medscape.com.