 It all started with this famous study, published in 1980. Johns Hopkins researchers were trying to figure out why states like New Mexico and Arizona had only about half the colon cancer rates of states like New York, New Hampshire, and Vermont. Maybe it's because they got so much sun. And so they proposed that maybe the sunshine vitamin D is a protective factor against colon cancer. Since then, sun exposure has been associated with lower rates of 14 other types of cancer, too. Vitamin D may also affect cancer survival. Higher blood levels of vitamin D were associated with lower mortality of patients with colorectal cancer. How much lower? Like nearly half the mortality. And the higher the D levels, the lower the death rate appeared to fall. This may explain why the survival rate from colon cancer may depend on part on the season of diagnosis. The risk of rapid death is lowest if you're diagnosed in the fall after you've spent the summer building up your vitamin D stores. But look, there are other risk factors that could be seasonal, too. Maybe people are taking advantage of the fall harvest and eating healthier. Maybe that's why the lower risk in the fall season. Or maybe there's more drinking in the winter. And in the summer running around outside, not only are you getting more sun, you're running around outside, getting more exercise, which may itself be protected. So these kind of studies just provide circumstantial evidence, establishing a cause-and-effect relationship between colon cancer and vitamin D deficiency using observational studies is challenging because of confounding factors like the exercise. So-called lurking variables. For example, there may be a tight correlation between ice cream sales and drowning deaths. That doesn't mean ice cream causes drowning. A more likely explanation is that there's a lurking third variable, like hot weather, summertime, that explains why drowning deaths are highest when ice cream consumption is highest. That's kind of a trivial example, but this actually happened with hormone replacement therapy. Women taking drugs like Primerin appear to have 50% less risk of heart disease, and so doctors prescribe it to women by the millions. But if you dig a little deeper into the data, yes, women taking estrogen had 50% lower risk of dying from heart disease, but they also had a 50% lower risk of dying from accidents and homicides. So it probably wasn't the drug. See, the only way to know for sure is to put it to the test. In a randomized clinical trial where you give half the women a drug, see what happens, and a decade later they did. And instead of having a 50% drop in risk, within a year of being given the hormone pills, heart attack and death rates shot up 50%. In retrospect, the lurking variable was likely socioeconomic class. Poor women are less likely to be prescribed hormone replacement therapy and more likely to be murdered and die of heart disease. Because of the lurking variable, a drug we now know to be dangerous had appeared protective. Besides lurking variables, there's also the possibility of reverse causation. Maybe low vitamin D levels didn't worsen the cancer. Maybe the cancer worsened the vitamin D levels. This may be unlikely since tumors don't appear to directly affect vitamin D levels, but cancer treatment might. And simple knee surgery can dramatically drop vitamin D levels within hours thought to be because of just the inflammatory insult of cutting into somebody. So maybe that could help explain the link between lower D and lower survival. And hey, if you've got cancer, maybe you're spending less time running around at the beach. So yes, higher vitamin D levels are associated with improved survival in colorectal cancer and in breast cancer. In fact, about double the risk of breast cancer recurrence in death in women with the lowest vitamin D levels. And vitamin D levels also associated with longer survival with ovarian cancer and other cancers like lymphoma. But bottom line, as we learned with hormone replacement, is that you have to put it to the test. But there weren't a lot of randomized controlled trials on vitamin D supplements in cancer until now. We now have a few randomized controlled trials, and vitamin D supplements do indeed appear to reduce the risk of dying from cancer. But dose, the researchers suggest maybe getting blood levels up to at least around 75 nanomoles per liter, levels not reached by as many as three-quarters of women with breast cancer or a striking 97% of colon cancer patients. Getting up to these kinds of levels, 75 or perhaps even better, 100, might require about 2,000 to 4,000 international units of vitamin D a day, levels of intake for which there appear to be no credible evidence of harm. Regardless what the exact level is, the findings of these kind of studies may have a profound influence on future cancer treatment.