The Carrot Was Safe. The Carrot Pill Wasn't.

Open your kitchen cupboard. If you're like a lot of people over 50, there's a row of bottles in there. A multivitamin, some vitamin E, maybe an eye formula.

Each bottle rests on one simple idea. Find the good thing in a healthy food, pull it out, and pack a big dose into a pill. You get the benefit without the chewing.

Science put that idea to the test on more than 47,000 smokers. It did not go the way anyone hoped…

The Carrot Pigment Goes On Trial

Carrots have long been tied to the eye. Cut one across and the rings look a bit like an iris. Old folk wisdom read that as a sign.

The orange in a carrot is real enough. It's beta-carotene, a pigment your body turns into vitamin A. Call it the orange pigment from here on.

By the early 1980s, the case for it looked strong. People who ate more carrots, greens and orange fruit kept turning up with less lung cancer. So researchers asked: what if you gave it as a pill?

Finland's National Public Health Institute and the US National Cancer Institute set out to answer that. Starting in 1985, they signed up 29,133 men in southwestern Finland, aged 50 to 69, all smokers. Call it the Finnish trial.

Some men took 20 mg of the orange pigment a day. Others took vitamin E, both, or a dummy pill. They kept at it for five to eight years.

The results ran in April 1994 in the New England Journal of Medicine. Lung cancer did not drop in the pigment group. It rose 18 percent, and deaths from all causes rose 8 percent.

Maybe it was a fluke. America was about to find out…

Then America Saw It Too

A second trial was already running in the US. Its full name was the Beta-Carotene and Retinol Efficacy Trial. It was run out of the Fred Hutchinson Cancer Research Center in Seattle. Retinol is vitamin A itself.

It enrolled 18,314 people: heavy smokers and recent ex-smokers, plus 4,060 men who had breathed asbestos at work. Each day they took 30 mg of the orange pigment plus 25,000 units of vitamin A, or a dummy pill.

On January 18, 1996, the trial was stopped 21 months early. The pill group had 28 percent more lung cancer than the dummy group. Lung cancer deaths ran 46 percent higher. And deaths from any cause ran 17 percent higher.

Gilbert Omenn led the 1996 report.

One fix to a common story. Only the American trial was stopped early. The Finnish one ran to its planned end in 1993.

The Food Never Did This

Here's the strange part. Look inside that same Finnish trial. The men who ate the most of the orange pigment from food at the start had less lung cancer, not more.

That's a link from watching diets, not a test, but it points away from the pill.

And the food side has a clean safety record. The Merck Manual notes that large amounts of carotenoids in food don't cause vitamin A poisoning. Your body converts them very slowly.

So what's the gap between a carrot and a capsule? Let me explain…

A Dose No Carrot Delivers

Start with size. One medium raw carrot holds about 5 mg of the orange pigment, per USDA data. The Finnish dose was four carrots' worth, every day. The American dose was six.

Xiang-Dong Wang and Robert Russell worked at the USDA's aging research center at Tufts University in Boston. They gave ferrets cigarette smoke plus a high dose of the pigment.

In the smoky lungs, the pigment broke down into damaged scraps. Those scraps drained the lungs of a form of vitamin A that helps keep cells in check. Then the lung lining began to change. That ran in 1999 in the Journal of the National Cancer Institute.

A year later the same Tufts group tried two doses. A low dose, about 6 mg a day in human terms, did no harm and may have offered weak protection. That's a carrot and a bit. A high dose, matched to 30 mg, did damage again. Published in 2000 in Carcinogenesis.

Ferrets aren't people. But the pattern fits the trials.

Scientists call this the food matrix. A carrot comes as a package: small amounts of the pigment, bound up in fiber, beside other plant compounds. Pull one piece out, raise the dose, and your body handles it in a new way.

It Wasn't Just One Pill

You might think this was a quirk of one pigment. It wasn't.

The Selenium and Vitamin E Cancer Prevention Trial tested 35,533 men. They took vitamin E, selenium, both, or a dummy pill. Eric Klein of the Cleveland Clinic led the 2011 report in the Journal of the American Medical Association.

Men taking 400 units of vitamin E a day had 17 percent more prostate cancer than the dummy group. That's about 1.6 extra cases per 1,000 men each year. Small, but the reverse of what the trial set out to show.

The pill held one man-made form of vitamin E. Food holds a mix. Qing Jiang and Bruce Ames, at the University of California, Berkeley, made the point in a 2001 review. Most vitamin E in the American diet is a different form, found in seeds and their oils.

Nobody has run a trial of almonds against prostate cancer. So I can't tell you the food form protects you. What I can tell you is that the pill was tested, and it lost.

Where A Pill Did Work

Pills aren't always the villain.

In 1991, Britain's Medical Research Council reported a trial of folic acid. It covered 1,817 women who had already had a pregnancy hit by a neural tube defect. That's a birth defect of the brain or spine. Folic acid cut the repeat rate by 72 percent. Published in the Lancet.

That one worked, and it still stands.

Harvard's Physicians' Health Study gave 22,071 male doctors 50 mg of the orange pigment every other day for about 12 years. It found no benefit, and no harm either. Reported in 1996.

Here's the thing. One nutrient in a big dose acts like a drug, and it needs a drug-sized trial before you trust it. The food version comes with a far longer track record…

Three Things Worth Knowing

The smokers in those trials did nothing wrong. They took what the science of the day said should help.

What they showed is that a nutrient inside a food doesn't act the same way in a pill. The carrot brings the whole package.

So buy a bag of carrots this week. Then read every label in that cupboard before Sunday.

Stay well,
The Healing Plate