Ah how the wheel turns round... Just announced at the American Association for Cancer Research Annual Meeting, dietary supplements increase the risk of heart disease and cancer. "Evidence shows that people who take more dietary supplements than needed tend to have a higher risk of developing cancer." Great.
So basically, too much of a good thing can indeed be bad. If you're eating a healthy diet, at best mega vitamins doses just waste money. At worst, they help your cancer grow. Take away: save money and just eat a well balanced diet. Maybe copy Ray Kurzweil's breakfast (paywall; without paywall) as he aims to live forever.
Linus Pauling.
This vitamin story has a long, very human history to it that I actually got to touch. As a freshman undergraduate, I had the amazing opportunity to work as an intern archivist in the Linus Pauling Papers collection. Pauling is one of only four people to have been awarded two Nobel prizes, one of two who were awarded Nobels in different fields, the only person to have been awarded a Nobel in both Science and Peace and the only person to be awarded two unshared Nobel Prizes. The guy was brilliant: "one of the most influential chemists in history and ranks among the most important scientists of the 20th century." I'll never forget pursuing his documents, notebooks and papers, let alone organizing, cataloging and, in some extremely small way, "discovering" them. [He sent regular letters to his wife and in one, wrote that he wanted to win a Nobel. After he did, he went back, found that letter and in the margin wrote something like "I did it!".]
Pauling is also a classic example of how people get stuck in belief, in this case about vitamin C. He firmly believed that vitamin C would prevent cancer. Note you can also firmly believe that you can jump off a cliff and fly. [Really, don't try it.] Unfortunately for that belief, according to research done in Pauling's own institute, vitamin C helped cancer in everything but near lethal doses [Pubmed ref]. The study was published in 1994, 11 years after a serious falling out with Dr. Arthur Robinson, the study's first author and a colorful scientist who was fired from his tenured position as President at the Pauling Institute at the time of the research. [Incidentally Robinson was awarded a $575,000 lawsuit settlement against Pauling. Online sources say $425,000 was for "slander and libel" and I'm unable to find a primary source on that; if you know one, please post a comment below. More in the June 11, 1979 issue of Baron's: Of Mice and Men - The Linus Pauling Institute is plunged into controversy.] Since then, other studies have shown that supplements "seem to increase overall mortality" (great Atlantic article) and that vitamin C doesn't seem to do anything [good] for cancer. Poor Pauling must be rolling in his grave. Be very wary of anyone, scientist or otherwise, who substitutes "belief" for "reproducible results"!
With the rather current nature of its questions, I read the Robinson paper a bit more in depth. The study itself is straight-forward: they irradiated hairless mice with UV while feeding them 38 diets with different nutrition and vitamins. They then quantified the number and severity of cancerous skin lesions and looked at those as a function of, say, how much vitamin C they ate. A few things stood out to me.
Note the significantly reduced number of lesions for mice on the "Black mixture" diet. Table 1 from Robinson AB, Hunsberger A, Westall FC. Suppression of squamous cell carcinoma in hairless mice by dietary nutrient variation. Mech Ageing Dev. 1994 Oct 20;76(2-3):201-14.
First, because the number of diets they tested was pretty big, the number of mice was what I'd call exploratory, in the 40-50 range. That's actually not so bad (it is in fact 1846 mice all together) except that second, there's no error bar anywhere in the paper, only averages. Error bars are a measure of the scatter of data and directly affect the chances that two groups are statistically "the same." Pretty intuitively, when scatter is high, you need a really big difference to say "they're different." From what's published, we really just can't do that. (I also couldn't find supplemental data listing individual mice to calculate scatter.)
Skin cancer in mice eating a control diet (top row) was significantly worse than those on a diet supplemented with antioxidants (bottom row). This was later found to be due to a sunscreen effect of BHT, which accumulates in skin, not antioxidant activity. Figure 2 from Black HS, Chan JT. Suppression of ultraviolet light-induced tumor formation by dietary antioxidants. J Invest Dermatol. 1975 Oct;65(4):412-4.
Third, speculating on the meaning of their data, the authors the conclude that "perhaps nutrition during cancer therapy should be viewed as the provision of fuel for a race between rapidly growing young tissues and mature older tissues wherein nutrient restriction or malnutrition may favor the older tissues." Cancerous cells are indeed fast growing with unique metabolic needs and this idea that something which is good when you don't have cancer might not be so good if you have it has recently been proposed in the context of antioxidants by Jim Watson. We might paraphrase Jim's thoughts as "take antioxidant as long as you don't have cancer and stop as soon as you've got it." Actually, this makes a lot of sense: the immune system is constantly surveilling (including for cancer) and it uses oxidative bursts to kill things. Lots of antioxidants hanging around? Harder to kill stuff. As it turns out, I've also worked on cancer studies that revealed aspects of the same phenomenon. With the caveat that biology is complex, it seems antioxidants are likely to help cancerous cells more than normal ones.
Fourth and finally, despite the lack of error bars, one thing really did stand out: one diet, the "Black mixture," had around 5-fold less severe lesions and about half the number of lesions in total as any other (see table above). That mixture contained a standard mouse chow supplemented with a bunch of antioxidants: 0.5 g vitamin E, 12 g vitamin C, 5 g butylated hydroxytoluene (BHT, which is in cereal boxes as an antioxidant preservative), and 1 g glutathione per kg of food. The mix was replicated from another study by Black that previously obtained similarly striking results (see mice picture). Robinson et al. broke down that mixture and tested the vitamins independently: vitamins alone didn't show the protective effect. Black later found a mechanism for the (reproducible) observation: BHT accumulates in the skin and acts as a sunscreen, reducing the amount of UV that reaches the lower layers of skin where cancer forms.
In summary: 1) antioxidants are probably not the best thing to take when cancer happens; 2) supplements at best don't do much if you've good nutrition and at worst increase the risk of death; and 3) as so often happens in science, the reproducible anti-cancer activity of a synthetic antioxidant BHT in mice wasn't (as hypothesized) due to its antioxidant properties at all but rater, because it formed a sort of internal sunscreen in skin.
Despite daily multi-gram quantities of vitamin C, Linus Pauling died of prostate cancer in 1994. His wife Ava Helen died 13 years earlier of stomach cancer.
Be careful, biology is complex.