Page 89 - Vitamin D and Cancer

76 E. Giovannucci

of fortified milk (in the USA) contains only 100 IU vitamin D, whereas being
exposed to enough UV-B radiation to cause a slight pinkness to the skin with most
of the skin uncovered (one minimal erythemal dose) produces vitamin D equiva-
lent to an oral dose of 20,000 IU vitamin D [10, 11]. In most populations, with
some exceptions such as in Iceland, much more vitamin D is made from sun
exposure than is ingested. Nonetheless, vitamin D intake is an important contribu-
tor to 25(OH)D levels, especially in winter months in regions at high latitudes,
when it may be the sole contributor. Yet, even with added vitamin D from supple-
mentation and fortification, vitamin D intake at typical levels currently do not
raise 25(OH)D levels substantially, and most variability in populations comes
from sun exposure. One important consideration of studies of vitamin D intake is
that, depending on the specific population, intake of vitamin D may be predomi-
nantly from one or a few sources, such as fatty fish, fortified milk, or supple-
ments. Thus, there will tend to be high correlations with other dietary factors
(e.g., omega-3 fatty acids in fish, calcium in milk, and other vitamins and miner-
als in supplements) increasing the possibility of confounding. One important
issue is that ergocalciferol (D2) is often used in supplements, and ergocalciferol
has been estimated to be only one-fourth as potent as cholecalciferol (D3) in rais-
ing 25(OH)D) [12].

4.2.3 Studies of Predicted 25(OH)D Level

A study can use known predictors of 25(OH)D level based on data on the individual
level to formulate a predicted 25(OH)D score. For example, based on individuals’
reported vitamin D intake, region of residence (surrogate of UV-B exposure), out-
door activity level, skin color, and body mass index, a quantitative estimate of the
expected vitamin D level can be made. The predicted 25(OH)D approach may have
some advantages and disadvantages compared to the use of a single measurement
of circulating 25(OH)D in epidemiologic studies. The measurement of 25(OH)D is
more direct, intuitive, and encompasses some of the sources of variability of
25(OH)D not taken into account by the score. The most important of these is actual
sun exposure behaviors, such as type of clothing and use of sunscreen. However, in
some aspects, the predicted 25(OH)D measure may provide a comparable or supe-
rior estimate of long-term vitamin D status over a single measurement of circulat-
ing 25(OH)D. Most importantly, some factors accounted by the predicted 25(OH)
D score are immutable (e.g., skin color) or relatively stable (region of residence,
body mass index). In contrast, circulating 25(OH)D level has a half-life of
2–3 weeks, and thus a substantial proportion of variability picked up by a single
blood measure would likely be due to relatively recent exposures, which may not
be representative of long-term exposure. The predicted 25(OH)D approach has
been rarely used.