Epidemiology of Vitamin D Insufficiency and Cancer Mortality

UV-B Radiation and Cancer Mortality

In 1941, Frank L. Apperly published data on an inverse association of solar radiation and cancer mortality (8). He concluded that exposure of suitable skin areas to sunlight might reduce cancer deaths, but his work did not significantly stimulate further research in this field at that time (8). About four decades later, in 1980, Garland and Garland were the first in modern literature to hypothesize that low UV-B radiation and subsequently reduced 25(OH)D levels were a risk factor for cancer mortality (9). Their hypothesis was based on the observation that colon cancer mortality in the US was significantly increased in populations with low exposure to solar UV-B radiation. Since then, the knowledge on UV-B radiation and cancer mortality has been extensively increased and was recently reviewed by Grant and Mohr (10). In brief, previous studies have shown that solar UV-B radiation was inversely associated with cancer incidence and mortality of up to 19 cancer sites (11). Importantly, the initial observations of inverse associations of UV-B and cancer mortality, which were largely made in the United States (US), have been further supported in various countries and at different latitudes (10). Furthermore, Grant et al. showed that even after considering the confounding factors of alcohol consumption, Hispanic heritage, urban/rural residence, poverty level and smoking (for lung cancer mortality), there still remained 13 cancer sites whose mortality rates were inversely associated with UV-B radiation: bladder, breast, colon, esophagus, gallbladder, stomach, ovary, pancreas, prostate, rectum, kidney and uterine corpus cancer, as well as Hodgkin's lymphoma (12).

An important aspect concerning the influence of UV-B exposure on vitamin D production and cancer is that due to their higher melanin content in the skin, black persons require more UV-B exposure than whites to produce similar amounts of vitamin D. Hence, blacks have a higher prevalence of vitamin D insufficiency than whites, raising the hypothesis that the higher cancer mortality rate among the US black population compared to US whites might partially be attributable to differences in vitamin D status. Indeed, data from the Health Professionals Follow-Up Study (HPFS), a cohort study among 47,800 men, showed that black men with risk factors for vitamin D insufficiency had a significantly increased risk of cancer mortality compared to white men (relative risk (RR) 2.27; 95% confidence interval (CI) 1.57-3.28) (13). On the other hand, there was no significant difference in cancer mortality when blacks with few risk factors for hypovitaminosis D were compared with whites.

Vitamin D Insufficiency and Cancer Mortality

In 2006, Giovannucci et al. were the first to present results of a prospective study on 25(OH)D levels and cancer mortality by analyzing data from the HPFS (14). 25(OH)D levels were measured in 1,095 out of 47,800 study participants and determinants of 25(OH)D levels, including vitamin D intake, skin pigmentation, geographic residence, sunlight exposure and body mass index, were used to estimate/predict 25(OH)D levels in the remaining study cohort. During the follow-up period from 1986 to January 31, 2000, 4,286 incident carcinomas and 2,025 cancer deaths occurred. After adjustments for possible confounders, an increment of 25 nmol/l in these predicted 25(OH)D levels was associated with a 17% reduction in total cancer incidence (RR 0.83; 95% CI 0.74-0.92) and a 29% reduction in total cancer mortality (RR 0.71; 95% CI 0.60-0.83). Significant inverse associations per increment of 25 nmol/l were also observed for colorectal (RR 0.63; 95% CI 0.48-0.83), pancreatic (RR 0.49; 95% CI 0.28-0.86), esophageal (RR 0.37; 95% CI 0.17-0.80) and oral or pharyngeal cancer mortality (RR 0.30; 95% CI 0.11-0.81).

In 2007, Freedman et al. published results from the Third National Health and Nutrition Examination Survey (NHANES-III), a study among 16,818 participants from the noninstitutionalized US population (15). The investigators of NHANES-III recorded 526 cancer deaths during follow-up. There was no significant association of 25(OH)D levels and total cancer mortality after adjustments for age, sex, race/ethnicity and smoking history. Colorectal cancer mortality was, however, reduced by 72% (95% CI 32-89%) when comparing persons with 25(OH)D levels above 80 nmol/l with those having 25(OH)D levels below 50 nmol/l. Breast cancer mortality was not significantly associated with vitamin D status when using 25(OH)D levels as a continuous variable but was significantly reduced in participants with 25(OH)D levels ≥62.5 nmol/l when compared to those with lower 25(OH)D concentrations (RR 0.28; 95% CI 0.08-0.93).

We also addressed the issue of vitamin D status and cancer mortality in 3,299 patients form the Ludwigshafen Risk and Cardiovascular Health (LURIC) study, a prospective cohort study from Germany among patients referred for coronary angiography (16). In that study, 95 cancer deaths occurred during a follow-up period of 7.75 years. The risk of total cancer mortality was reduced by 55% (RR 0.45; 95% CI 0.22-0.93) in the fourth versus the first 25(OH)D quartile.

Summarizing the results of these three studies we have inconsistent results, with two studies (HPFS and LURIC) which found a significant association of compromised vitamin D status and total cancer mortality and one study (NHANES-III) which observed significant associations only for colorectal and breast cancer mortality but not for 25(OH)D levels and total cancer mortality. Underlying differences of the study populations and study designs may account for these inconsistent results. In particular, the study cohorts of LURIC and NHANES-III were significantly different, with the LURIC cohort being 2 decades older with much lower 25(OH)D levels and a more homogeneous study population compared to the very heterogeneous study cohort of NHANES-III which included various ethnicities from different geographic regions in the US. Of note is that after the LURIC data were published, Freedman et al. performed a reanalysis of NHANES-III for non-Hispanic whites by using a reference group (lowest 5%) with similar 25(OH)D concentrations as the reference group (lowest quartile) in the LURIC study. The risk of cancer mortality was non significantly reduced by 22% in the lowest 5% in NHANES-III (RR 0.78; 95% CI 0.42-1.43), when compared to the highest quartile (17). Hence, this latter analysis might suggest that the association of vitamin D insufficiency and increased cancer mortality is particularly apparent for persons with very low 25(OH)D levels. However, according to the conclusion by Freedman et al., the small numbers with depressed 25(OH)D levels in NHANES-III limit the power of that study to discern effects at very low 25(OH)D levels (17).

Importantly, both the LURIC and NHANES-III studies considered the seasonal variations in 25(OH)D levels in their statistical analyses, because the seasonal changes in UV-B exposure are responsible for circannual differences of the vitamin D status (16, 17). This is well illustrated by observations that 25(OH)D levels are up to twice as high at the end of the summer when compared to winter levels (16, 18-20). To reduce possible confounding by these seasonal variations in 25(OH)D levels, we and others classified the vitamin D status of an individual according to the percentile distribution of 25(OH)D levels within the respective month in which blood was drawn (17, 18). We therefore suggest that in the future the vitamin D status of an individual be classified not only according to absolute 25(OH)D levels but also according to percentiles of population-based 25(OH)D levels for each month of the year. Another issue which is important to consider when interpreting or analyzing data on vitamin D and cancer is the fact that low physical activity is related to both vitamin D insufficiency and increased cancer risk and may therefore be a possible confounder (20). This problem is, however, complex because reduced physical activity with subsequent reduced outdoor activities and lower UV-B exposure is not only a cause of vitamin D insufficiency but might also be a consequence of vitamin D insufficiency. In this context, it has been demonstrated that vitamin D exerts beneficial effects on musculoskeletal health and reduces fractures and falls thereby preventing immobilisation (2, 4). Given that the assessment of physical activity is not trivial, it was not measured and adjusted for in most of the aforementioned studies but the LURIC study did show that the association of cancer mortality and vitamin D insufficiency remained significant even after adjustment for physical activity level (16).

Vitamin D Supplementation and Cancer Mortality

The impact of a daily supplementation of 400 IU vitamin D plus 1,000 mg calcium on cancer mortality was evaluated in the Women's Health Initiative (WHI), a randomized, double-blind, placebo-controlled trial among 36,282 women (26, 27). During an average follow-up period of 7 years, 726 patients died due to cancer. There was a non-significant reduction of cancer mortality of 11% in the vitamin D and calcium versus the placebo group (RR 0.89; 95% CI 0.77-1.03; p=0.12) (26). When interpreting these statistically non significant results of the WHI, it should be underlined that there is a general consensus that daily vitamin D doses of 400 IU (as used in the WHI) are too low to sufficiently raise the vitamin D status to 25(OH)D levels which significantly improve health outcomes and which should be at least as high as 75 nmol/l (30 ng/ml) (28). This notion is in line with other vitamin D-related health outcomes because a recent meta-analysis showed that vitamin D intake had no significant effect on fracture risk in randomized, placebo controlled trials with a daily supplementation of 400 IU vitamin D (29). The incidence of fractures was, however, significantly reduced by vitamin D supplementation in studies using daily vitamin D doses >400 IU (29). Furthermore, in addition to the WHI study, another randomized, double-blind, controlled trial involving a four-monthly supplementation with 100,000 IU vitamin D was performed among 2,686 people aged 65-85 years living in the general community of Great Britain (30). In that study, risk of total cancer mortality was non significantly reduced by 14% in the vitamin D versus the placebo group (RR 0.86; 95% CI 0.61-1.20) (30).

Moreover, Lappe et al. performed a randomized controlled trial among 1,179 healthy community-dwelling postmenopausal women (31). This trial included three groups: placebo, calcium (1,400-1,500 mg) and calcium (1,400-1,500 mg) plus 1,100 IU vitamin D. Fifty women developed cancer during the study duration of 4 years including 13 incident cases diagnosed in the first year of the study. Compared to the placebo group, the risk of incident cancer was reduced by 60% in the vitamin D plus calcium group (RR 0.40; 95% CI 0.20-0.82). This latter analysis was repeated for women free of cancer after 1 year of follow-up in order to reduce possible bias by occult cancer that was already present but not diagnosed at baseline. In this analysis, risk of incident cancer was reduced by 77% in the calcium plus vitamin D versus the placebo group (RR 0.23; 95% CI 0.09-0.60). For both these analyses there was no significant difference for cancer incidence between the placebo and the calcium group. This study by Lappe et al. does not provide data on cancer mortality but the findings of this trial with regard to cancer incidence are of great importance because this was the first study to show to vitamin D supplementation significantly reduced the risk of cancer in a randomized placebo-controlled trial.