dietary carotenoid intake levels of 8.6 mg/day (RR = 0.68 compared to those consuming 1.1 mg/day of carotenoids).
In the Western Electric cohort study, all-cause mortality was lowest for men who consumed the highest tertile of dietary β-carotene (RR = 0.80 for more than 4.1 mg/day of β-carotene versus less than 2.9 mg/day of β-carotene; p for trend = 0.01) (Pandey et al., 1995).
Because there are literally hundreds of studies of carotenoids and cancer risk, this section emphasizes the results of epidemiological studies of all cancers combined, studies of carotenoids and lung cancer, and a few other selected tumor sites for which an inverse association with carotenoids is commonly seen.
Observational Epidemiological Studies. The Basel Prospective Study evaluated the relationship between plasma carotene concentrations in blood samples obtained in 1971–1973 and subsequent cancer mortality up to 1985 (Stahelin et al., 1991). Results showed that persons who went on to develop any cancer had significantly lower prediagnostic carotene concentrations than persons who remained alive and free of cancer in 1985 (mean plasma total carotenoid concentration 0.34 µmol/L [18 µg/dL] in those with cancer versus 0.43 µmol/L [23 µg/dL] in those free of cancer). The authors state that the reported carotene values represent approximately 80 percent β-carotene and 20 percent α-carotene; thus, plasma β-carotene concentrations of approximately 0.34 µmol/L (0.43 µmol/L × 0.8) (18 µg/dL [23 µg/dL × 0.8]) were typical for the survivors of this cohort. This concentration is within the range associated with lower risk elsewhere as shown in Table 8-3.
Numerous epidemiological studies have shown that individuals who consume a relatively large quantity of carotenoid-rich fruits and vegetables have a lower risk of cancer at several tumor sites (Block et al., 1992). The consistency of the results from observational studies is particularly striking for lung cancer, where carotenoid and fruit and vegetable intake has been associated with lower lung cancer risk in 8 of 8 prospective studies and 18 of 20 retrospective studies reviewed (Ziegler et al., 1996b).
Focusing on prospective blood analyses studies, the study with the largest number of cases (n = 99) was reported by Menkes et al. (1986) as part of the Washington County, Maryland, cohort. The risk of lung cancer increased in a linear fashion with decreasing serum concentrations of β-carotene, with the greatest risk at the