It's a Joke

It’s a Joke

Did you hear the one about selenium supplements causing diabetes? It goes like this: The New York Times, ABC News and Reuters walk into a bar. The bartender says, “Want to buy a half-baked study?” And all three say, “Yes!”

A joke isn’t funny if you have to explain it. But I’ll explain this one, and you might end up howling (although, probably not with laughter).

Salvage job

As I’ve noted in previous e-Alerts, selenium has been shown to help produce several important health benefits, including enhancement of the immune system and a reduction of cancer risk: specifically liver, colorectal and prostate cancers. Some research has also suggested that selenium may increase insulin efficiency by improving glucose metabolism. That’s not to say it prevents or cures type 2 diabetes, but rather plays a role – along with other beneficial nutrients – in curbing diabetes risk.

But researchers at the State University of New York at Buffalo (SUNY) didn’t set out to examine a selenium link to diabetes. Their study was designed to assess the effects of selenium supplementation on subjects with nonmelanoma skin cancer. About 1,200 skin cancer patients were recruited. Their average age was 65, and none of the subjects had type 2 diabetes at baseline.

Subjects were randomly divided into two groups. About half took 200 mcg of selenium each day for an average of about eight years. The other half received a placebo. Results showed that subjects who took selenium supplements had a slightly lower risk of dying of cancer, but no specific protection against skin cancer was revealed.

So let’s say you’re on the SUNY team, and when the results are in you have to admit there’s not much to report. Which means you’ve come up just about empty after an expensive eight-year intervention study. So what do you do? Do you throw in the towel, or do you comb through the data looking for something, ANYthing, to salvage the work? If so, you find that the final rates of type 2 diabetes are lopsided. Huh! Maybe selenium causes diabetes! Previous research has found, at best, only scant evidence of such a link, while other studies show a link between high selenium levels and REDUCED diabetes risk. But what the heck? Maybe we can sell this.

And then the New York Times, Reuters, and ABC News walked into a bar

Skin deep

Why don’t sharks eat comedians? Because they taste funny. And that’s why sharks wouldn’t touch this study – it tastes funny and smells bad too.

First: Eligibility for the study was based primarily on cancer history, not on health factors that pertained to diabetes. According to a 2006 report in Diabetes Care, more than 73 million Americans have diabetes or a pre-diabetic condition, but millions of them are not even aware of this. So if you recruit 1,202 subjects with an average age of 65, chances are good that a substantial number don’t realize they’re on the verge of type 2 diabetes.

Second: In the published study, the authors note that out of 1,202 subjects, 97 developed type 2 diabetes. And they add: “This rate is similar to that in other studies of largely white populations.” (All of the subjects in this study were white due to a necessary skin pigmentation restriction because this was a SKIN CANCER STUDY.)

So given that you’ve got mostly older subjects (a demographic that’s at high risk of type 2 diabetes), and given that there’s absolutely no prior proof that selenium raises diabetes risk, and given that the study’s overall rate of diabetes is completely within normal bounds, then the random assignments of subjects to the intervention group and the placebo group could just as easily have produced a greater number of subjects that DIDN’T develop diabetes.

By the way, the difference between the number of diabetes cases in the selenium group (58) compared to diabetes cases in the placebo group (39) was less than 20 out of 1,202 total subjects.

To say that this study reveals a link between selenium intake and increased diabetes risk is really stretching it. In fact, there’s almost nothing to stretch. And without additional research to support these results it’s absurd to start claiming a selenium/diabetes link.

Nevertheless, the study was widely regarded as some kind of dramatic breakthrough. The New York Times, ABC News, and other outlets sounded the alarm, warning their audiences about this newfound selenium danger. Meanwhile, Reuters ran this headline: “Selenium supplements linked to diabetes in U.S. study.” And the Reuters report really warmed to the sales pitch, calling the study “an unusually well-controlled trial.” Really! Are other double-blind, placebo-controlled studies poorly controlled? If so, exactly how is this one “unusually” well-controlled? The article doesn’t explain, of course.

Digging in the dirt

But wait. There’s one more punch line.

The SUNY skin cancer study will be published in an August issue of the Annals of Internal Medicine along with an accompanying editorial written by three doctors from Johns Hopkins University Bloomberg School of Public Health. And they’ve given their editorial an ominous title: “Selenium and Diabetes: More Bad News for Supplements.”

After ratcheting up the fear factor, noting that selenium is lurking in many multivitamins, the authors tell us: “The U.S. public needs to know that most people in this country receive adequate selenium from their diet. By taking selenium supplements on top of an adequate dietary intake, people may increase their risk for diabetes.”

That’s hilarious! But it doesn’t hold water. In the skin cancer study, the SUNY authors note that their subjects were recruited from “dermatology clinics in areas of low selenium consumption of the eastern United States.”

“Low selenium consumption” Does that sound “adequate” to you?

In this case, “eastern United States” would be most of the states east of South Dakota. As I’ve mentioned in previous e-Alerts, fruits and vegetables are good selenium sources, but only when they’re grown in selenium-rich soil. In the U.S., selenium is highly concentrated in the soil of just six states: North and South Dakota, Utah, Colorado, Montana and Wyoming. Bread, fish and meat also deliver some small amounts of selenium.

The Hopkins doctors are probably well aware of just how flimsy their argument is. But when you’re delivering “more bad news for supplements” you don’t want to let annoying reality get in the way.

“Effects of Long-Term Selenium Supplementation on the Incidence of Type 2 Diabetes” Annals of Internal Medicine, Vol. 147, No. 4, 8/21/07,
“Selenium and Diabetes: More Bad News for Supplements” Dr. Joachim Bleys, Dr. Ana Navas-Acien and Dr. Eliseo Guallar, Annals of Internal Medicine, Vol. 147, No. 4, 8/21/07,
“Selenium Supplements Linked to Diabetes in U.S. Study” Reuters, 7/10/07,
“Outcomes: Selenium Supplements May Raise Diabetes Risk” Nicholas Bakalar, The New York Times, 7/10/07,
“Selenium Supplements May Raise Diabetes Risk” Dan Childs, ABC News Medical Unit, 7/10/07,