One explanation for discrepant results between epidemiologic studies and randomized trials of
beta-carotene and
cardiovascular disease may be a failure to consider
inflammation as a confounder. To evaluate the potential for such confounding, the authors relate the serum concentrations of five
carotenoids (
alpha-carotene,
beta-carotene,
beta-cryptoxanthin,
lycopene, and
lutein/
zeaxanthin) to levels of three inflammatory markers (
C-reactive protein,
fibrinogen, and white blood cell count) measured during the Third National Health and Nutrition Survey, 1988-1994. The analysis included 4,557 nonsmoking participants aged 25-55 years. Adjusted concentrations of all five
carotenoids were significantly lower in those with
C-reactive protein levels above 0.88 mg/dl (p = 0.001). There was a trend toward lower adjusted
beta-cryptoxanthin concentrations with increasing level of
fibrinogen (p value test for trend = 0.01), but other
carotenoids were not related. Many of the
carotenoid concentrations were lower among participants with high white blood cell counts. After log transformation, only adjusted mean
beta-carotene levels were significantly lower in those with white blood cell counts above 7.85 x 10(9)/liter (p < 0.01). These cross-sectional data do not clarify the biologic relation between
carotenoids and
C-reactive protein but, to the extent that the
carotenoids are associated with
C-reactive protein levels, a
carotenoid-
heart disease association may be, in part, an
inflammation-
heart disease association.