The Epidemiology of Vitamin D and Colorectal Cancer: Recent Findings

Edward Giovannucci

Curr Opin Gastroenterol. 2006;22(1):24-29. 

Abstract and Introduction

Purpose of Review: To highlight the human studies published over the past year examining the influence of vitamin D on risk of colorectal cancer.
Recent Findings: Studies over the past year have added more support to the idea that higher levels of vitamin D may decrease risk of colorectal cancer. Further, typical dietary intakes such as 200-400 IU/day may be too low to exert appreciable benefits, and protection may occur with higher levels of vitamin D associated with exposure to sunshine. Recent studies also suggest a potential benefit of vitamin D on other digestive-tract cancers, and that vitamin D status at the time of diagnosis and treatment may influence survival of cancer. However, the evidence for these latter findings is based on limited data and needs to be confirmed. Higher vitamin D levels may also be associated with a higher rate of apoptosis in colorectal mucosa.
Summary: Recent studies add more support to a potential role of vitamin D on risk of colorectal cancer, but suggest that intakes higher than customary are required if solar ultraviolet-B exposure is low. More studies are required to determine the optimal levels and intakes of this vitamin to reduce cancer risk. Potential benefits of vitamin D on other digestive-tract cancers and on survival in patients with colorectal cancer have been suggested by recent studies, but require confirmation.

The suggestion that exposure to sunlight may lower the risk of cancer was first made in the 1930s.[1] An association between latitude and cancer mortality rate was first demonstrated in 1941.[2] This finding was largely ignored until four decades later, when Garland and Garland[3] hypothesized that vitamin D status accounted for the inverse association between solar ultraviolet-B (UV-B) radiation exposure and risk of colon cancer. Subsequently, Garland and colleagues proposed a similar association for breast cancer[4] and ovarian cancer.[5] Remarkably, in 1980, essentially nothing was known about how vitamin D would influence carcinogenesis. Since 1980, a large number of studies have supported the biologic basis of the hypothesis. Many cell types contain vitamin D receptors (VDRs), and when these receptors are activated by 1,25-dihydroxyvitamin D [1,25(OH)2D], the most active metabolite of vitamin D, they induce differentiation and inhibit proliferation, invasiveness, angiogenesis, and metastatic potential. Circulating 25-hydroxyvitamin D [25(OH)D] was also shown to be potentially beneficial because many cells types including cancer cells express 1-α-hydroxylase, and are thus able to convert 25(OH)D into 1,25(OH)2D. This review focuses on recent research of epidemiologic and human studies (published between August 2004 and August 2005) on vitamin D and cancer of the large bowel.

Geographic Studies

The vitamin D hypothesis was initiated by the observation of an inverse correlation between solar UV-B and risk of colon cancer.[3] This relationship has been confirmed subsequently using more sophisticated data and analyses. For example, Grant[6] demonstrated that regional UV-B radiation correlated inversely with mortality rates of numerous cancers, particularly digestive organ cancers; in males, about 80% of the cancers attributable to low solar UV-B were digestive-system cancers. In women, in addition to digestive-tract cancers, cancers of the reproductive system were also correlated inversely with UV-B exposure.

Although such ecologic data are informative, one of the potential weaknesses is the inability of adequately controlling for potentially confounding factors. Although a number of studies have been published in the USA, each adding some refinement, the data are essentially generated from the same source and thus cannot be considered as independent confirmations. A recent study by Mizoue[7*] in Japan is important because it was conducted in an entirely different population. Mizoue calculated Pearson correlation coefficients between averaged annual solar radiation levels for the period 1961-1990 and cancer mortality in the year 2000 in 47 prefectures in Japan. Adjusting for regional per capita income and dietary factors, Mizoue[7*] found an inverse correlation between averaged annual solar radiation levels and mortality from digestive-system cancers (esophagus, stomach, colon, rectum, pancreas, and gallbladder and bile ducts) but not other cancer types in Japan. For colon cancer the correlations were -0.53 in men and -0.46 in women, and for rectal cancer the correlations were -0.53 in men and -0.47 in women.

This study shows independently in an entirely different population a similar correlation between solar radiation levels and lower risk of colorectal cancer and other digestive-tract cancers found in the USA.[6] These cancers tend to have diverse risk factors, and it would be unlikely if an unknown confounding factor would be operative for all these cancers in a similar fashion in both the USA and Japan. This finding suggests that some constant characteristic of the digestive tract makes it particularly sensitive to the effects of vitamin D. The author had speculated that a specific interest in studying solar exposure in Japan is that the high consumption of fatty fish, the only good natural source of dietary vitamin D, may render solar UV-B less important. However, hypovitaminosis D is relatively common in Japan. For example, in the Tokai area, a relatively sunny region, mean serum 25(OH)D level was lowest at the end of winter (15.1 ± 7.1 ng/ml in March), and highest at the end of summer (31.6 ± 5.6 ng/ml in September, and the prevalence of hypovitaminosis D (<20 ng/ml) was 86.7, 33.4, 1.0, and 26.0% in March, June, September, and December, respectively.[8] Thus, even in a country with relatively high intake of vitamin D, solar UV-B is an important source of vitamin D.

Solar Radiation, Vitamin D and Survival Rate of Colon Cancer

Most epidemiologic studies on vitamin D and colorectal cancer risk have focused on incidence or mortality rate, which is a function of incidence rate and survival. A recent study in Norway examined the influence of latitude and season of diagnosis on survival from colon cancer.[9*] This study could be conducted in Norway because the Cancer Registry of Norway has registered all cancer diagnoses since 1953. In the analysis, 12 823 men and 14 922 women with colon cancer were included, and the period of observation was from 1964 to 1992. Both latitude and season variation and survival rate were assessed. There was no significant annual variation in the incidence rates of colon cancer, with 25% of the cancers diagnosed in each season. No significant north-south gradient was found for the death rate for colon cancer. However, death rates at 18, 36, and 45 months were significantly (20-30%) lower in the cancers diagnosed in autumn months compared with those diagnosed in the winter months, with maximal benefit at 18 months after diagnosis. This finding suggests that high vitamin D level at the time of diagnosis, and presumably treatment, may improve survival from colon cancer. Similar findings have also been reported for breast and prostate cancer in the same dataset.[10] Late-stage anticancer effects of vitamin D, such as reduction in metastases, are observed in numerous animal models. Some animal evidence suggests that vitamin D analogues may improve tumor control by radiation treatment, in part by promoting apoptosis.[11]

These data are provocative, but alternative explanations to vitamin D are possible. For example, there may be some other micronutrients related to specific fruits and vegetables that are consumed in the summer and autumn months only. This finding should be considered hypothesis-generating rather than definitive, but nevertheless the potential importance is great as it suggests that a relatively small time window in which vitamin D status is improved, perhaps at the time of treatment, could have an important effect on survival. The lack of association with latitude appears to provide data against the importance of vitamin D, but the authors present some data suggesting that whereas studies do support an approximate doubling of vitamin D levels in the summer/autumn compared with winter in Nordic countries, there may be minimal variation by latitude in these countries. For example, 25(OH)D levels in Tromsø (northern Norway; latitude 70º) are similar as those in Denmark (latitude 56º). In part, this could be explained by higher consumption of fatty fish, a source of vitamin D, in the northern regions.

Vitamin D Levels and Apoptosis in the Rectal Epithelium

Some evidence suggests that a decreased level of apoptosis in the colorectal epithelium may be an indicator of an increased risk of developing an adenoma or cancer.[12,13] In human colorectal adenoma and cancer cell lines 1,25(OH)2D induces apoptosis in a dose-dependent manner.[14] The relationship between vitamin D intake and level, as well as calcium intake and apoptosis, was studied in 498 adenoma patients and 324 adenoma-free individuals undergoing colonoscopies.[15*] Dietary intakes were assessed by food frequency questionnaire, and in one subpopulation (92 patients with adenoma and 188 without), serum 25(OH)D levels were measured. Apoptosis was scored from normal rectal mucosal pinch biopsies, using a morphometric assay and the terminal nucleotidyl transferase-mediated nick-end-labeling method. These measures were correlated and yielded similar results. In a multivariate-adjusted logistic regression model, patients with adenomas who were in the highest versus lowest tertile of dietary calcium intake had 3.4-times higher odds [95% confidence interval (CI), 0.9-12.9] of elevated apoptotic scores, but in adenoma-free patients high calcium intake was not related to elevated apoptosis [odds ratio (OR), 1.2; 95% CI, 0.6-2.7]. In contrast, the highest level of 25(OH)D was associated with a higher apoptosis rate in adenoma-free patients (OR, 2.6; 95% CI, 1.1-6.2) and slightly lower apoptosis levels in patients with adenomas, although this was not statistically significant (OR, 0.6; 95% CI, 0.2-2.2).

These results suggest that calcium and vitamin D may influence apoptosis in colorectal tissue. These results are interesting and suggest that apoptosis is one relevant mechanism for the effect of vitamin D. However, several limitations and inconsistencies merit discussion. Why calcium appeared more relevant in the adenoma cases, whereas vitamin D appeared more important in the adenoma-free individuals, remains obscure. Perhaps chance may have played a role. There were only 92 adenoma patients and 188 non-adenoma patients with available vitamin D levels. The relatively small sample size also precluded an examination of the potential interaction between vitamin D level and calcium intake. Finally, higher 25(OH)D levels are likely found in leaner individuals as well as physically active individuals, who may be more exposed to solar radiation through outdoor activities. Thus, 25(OH)D levels could plausibly be acting as a marker of physical activity and low body mass, two factors that could reduce colon cancer through alternative mechanisms. However, an argument against this interpretation is that the biopsies were from rectal tissue, and rectal cancer in contrast to colon cancer has not been clearly associated with obesity and a sedentary lifestyle.[16]

Of note, individuals with large adenomas (≥1.5 cm) have been found to have both increased proliferation and reduced apoptosis.[12] In a previous study, Holt et al.[17] had shown that serum levels of 25(OH)D3 were associated with a highly significant decrease in whole-crypt labeling index and in the size of the proliferative compartment. Presumably neoplastic growth would be maximal in individuals with simultaneously lower apoptosis rates and high proliferative capacity.

Epidemiologic Studies of Cancer and Adenoma

Other than a long-term, adequately powered randomized trial, the most direct examination of the hypothesis that vitamin D lowers risk of colorectal cancer is a prospective study in which vitamin D is assessed at baseline, and cancer-free individuals are followed for the development of cancer. Until last year only three, relatively small, studies have been conducted, of 34,[18] 57,[19] and 146[20] cancer cases. While generally supportive of the vitamin D-colorectal cancer link, these studies were generally hindered by the small number of cases.[18-20] Recently the relationship between plasma 25(OH)D and colorectal cancer risk was evaluated in the Nurses' Health Study,[21*] based on 193 incident cases of colorectal cancer. Two controls were matched per case on year of birth and month of blood draw. After adjusting for age, body mass index, physical activity, smoking, family history, use of hormone-replacement therapy, aspirin use, and dietary intake the relative risk (RR) decreased monotonically across quintiles of plasma 25(OH)D concentration, with a RR of 0.53 (CI, 0.27-1.04) for quintile 5 versus 1. The median 25(OH)D concentration in quintile 5 was approximately 50 nmol/l higher than that in quintile 1. This inverse association remained strong when limited to women aged 60 years or older at blood collection (P = 0.006), but an association was not apparent among the younger women (P = 0.70). Benefit from higher 25(OH)D concentrations was observed for cancers at the distal colon and rectum (P = 0.02) but was not seen for proximal colon cancer (P = 0.81).

This study adds strong support to a role for 25(OH)D in lowering risk of colorectal cancer. Similar to the second largest study[20] an association was only observed for the distal colorectum. The reason for the site specificity is not known, but Anti et al.[12] had shown that the imbalance between proliferation and apoptosis may be greatest in the left colon, and vitamin D influence may be related to both processes. An important strength of the study was that the findings for 25(OH)D persisted even after controlling for body mass index, physical activity, and multivitamin intake.

In the past year there have been three additional reports on vitamin D intake and colorectal cancer risk, one of which also evaluated adenomas, and another report on adenoma risk. In a study in France[22] the population for adenomas was composed of 516 adenoma cases, including 175 high-risk adenomas, and of 4804 polyp-free subjects confirmed by colonoscopy; the colorectal cancer study population was composed of 172 cases and 67 312 cancer-free subjects at baseline. Diet was assessed using a self-administered questionnaire completed at baseline. This study provided evidence for a benefit of calcium and dairy products, but not for vitamin D intake. Notably, most dairy products in France during the period of dietary assessment were not fortified with vitamin D, and the intake for the 90th percentile (high) intake was only 174 IU/day. This level of intake is unlikely to influence 25(OH)D levels appreciably.

In another prospective dietary cohort study, the US Women's Health Study, 39 876 women aged ≥45 years and free of cardiovascular disease and cancer were enrolled in the study in 1993.[23] During an average follow-up of 10 years, 223 of these women were diagnosed with colorectal cancer. Intakes of calcium and vitamin D from dietary sources and supplements were assessed with a baseline food frequency questionnaire. Neither calcium nor vitamin D intakes were associated with risk of colorectal cancer. In this study, the cut-off points for low and high intake of vitamin D were <161 and >545 IU/day respectively.

One recent study examined vitamin D and calcium intakes in relation to risk of colorectal adenoma in individuals of the Polyp Prevention Trial (PPT), a randomized clinical trial designed to examine the effects of a high-fiber, high-fruit and -vegetable, low-fat diet on the recurrence of colorectal adenomas.[24] In this study, adenoma recurrence was ascertained by complete colonoscopy at baseline and after 1 and 4 years. A new adenoma was diagnosed in 754 of the 1905 trial participants. Total vitamin D intake was inversely associated with adenoma recurrence (OR, 0.84; 95% CI, 0.62-1.13; P trend = 0.03), after adjusting for intervention group, age, gender, nonsteroidal antiinflammatory drug use, total energy intake, and the interaction of gender and intervention group. A slightly stronger association was noted for multiple new adenomas (OR, 0.70; 95% CI, 0.47-1.06; P trend = 0.01), but no association was observed for advanced adenoma. The cut-off points of intake from low to high quintiles were <134 to >468 IU/day. No overall significant associations were observed between adenoma risk and dietary calcium intake (OR for the fifth compared with the lowest quintile, 0.91; 95% CI, 0.67-1.23; P trend = 0.68), or total calcium intake (OR, 0.86; 95% CI, 0.62-1.18; P trend = 0.20), though supplemental calcium use during follow up was inversely associated with adenoma risk (OR for any compared with no use, 0.82; 95% CI, 0.68-0.99).

A large case-control study reported on dietary calcium and vitamin D, as well as some VDR genotypes and risk of colorectal cancer.[25] The main limitation of this study was that as a case-control study diet was assessed retrospectively, which could introduce recall or selection bias. The main strength was the large sample size (n = 2306 cases and 2749 controls), and the simultaneous examination of VDR polymorphisms. Results for colon cancer had been reported previously, with no appreciable associations, except that vitamin D supplements were inversely associated with risk in men (OR, 0.5) and women (OR, 0.6), although confidence limits included the value 1.0.[26] In the new study, high calcium intake was associated with a reduced risk of rectal cancer in women (OR, 0.39; 95% CI, 0.24-0.64) but was not associated with rectal cancer in men (OR, 1.02; 95% CI, 0.66-1.56). Similarly, rectal cancer risk among women was reduced with high intakes of vitamin D (OR, 0.52; 95% CI, 0.32-0.85). An inverse association with vitamin D intake was also observed for individuals ≥60 years (OR, 0.59; 95% CI, 0.40-0.88). The lower and upper cut-off points for vitamin D intake were 124.3 and 332 IU/day, respectively, for women and 168 and 408 IU/day for men. High reported sunshine exposure was suggestively associated with a reduced risk in men (OR, 0.78; 95% CI, 0.57-1.08) but not women (OR, 0.93; 95% CI, 0.63-1.37). Two VDR polymorphisms were evaluated in this study: an intron 8 Bsm1 polymorphism and a 3' untranslated region polyA length polymorphism (designated S for short and L for long). The SS genotype was associated reduced risk of colorectal cancer for men only (OR, 0.71; 95% CI, 0.55-0.92). The authors also examined interactions with the VDR polymorphisms by calcium and vitamin D intakes but results were inconsistent across anatomic subsites.

In a separate report from the same population, interactions between body mass index and energy expenditure were examined in relation to the Bsm and polyA polymorphisms, as well as the Fok1 VDR polymorphism.[27] The Fok1 VDR polymorphism is of special interest because the F allele is associated with a VDR protein that is three amino acids shorter than the f allele. In this study, those with the FF genotype were slightly more likely to develop colon cancer than those with the ff genotype (OR, 1.28; 95% CI, 0.97-1.69). Also of note, higher body mass index and lower long-term vigorous activity levels were much stronger risk factors for colon cancer in individuals with the ff allele; in essence, in highly active, lean individuals, the ff genotype was associated with lowest risk whereas individuals who had the ff genotype were at particularly high risk if they were obese and inactive.

An additional report from a Korean population-based, case-control study[28] examined the VDR Fok1 polymorphism. Compared with Ff heterozygotes, those with the FF genotype were at elevated risk (OR, 1.9; 95% CI, 1.35-2.81), whereas those with the ff genotype were at reduced risk (OR, 0.5; 95% CI, 0.26-0.84). In addition, the FF genotype was particularly strongly associated with cancers that were associated with an elevated serum carcinoembryonic antigen, which has generally been shown to be a poor prognostic factor for recurrence and survival after curative resection of colorectal cancer. Of note, the overall findings of this study for the Fok1 polymorphism were quite similar to those from a previous study in an Asian population.[29]

Discussion

The studies reviewed here have added to the body of evidence that vitamin D status is an important risk factor for colorectal adenoma and cancer. Various approaches were used to assess vitamin D status, including geographical variation in UV-B exposure, season variation in UV-B exposure, circulating levels of 25(OH)D, and dietary and supplementary intakes. In addition, polymorphisms in the VDR have been examined in relation to cancer risk.

Geographical variation is an important source of variations in vitamin D levels, but skin pigmentation, sunlight-exposure behaviors, and adiposity also contribute, in addition to diet and perhaps genetic factors. Diet may be important in some circumstances, but overall contributes only to a modest proportion of total vitamin D in most populations. In general, findings based on circulating 25(OH)D have yielded stronger associations than those based on intake. This pattern is not surprising based on relatively low contribution of intakes on overall vitamin D status. For example, in the Nurses' Health Study, the median 25(OH)D concentration in quintile 5 was approximately 50 nmol/l higher than that in quintile 1. For vitamin D intake-based studies, the difference between high and low intakes was generally not more than about 300-400 IU/day. Such an increment in vitamin D intake would be expected to increase circulating 25(OH)D by less than 10 nmol/l. For example, an increment of approximately 1700 IU of vitamin D was required to raise plasma vitamin D from 50 to 80 nmol/l (30 nmol/l increment) during the winter months in Iowa.[30] Thus the contrast between low and high levels of 25(OH)D in the Nurses' Health Study would correspond approximately to a 3000 IU/day increment from intake alone, almost 10 times the actual range observed in most dietary studies.

Confirming the regional solar UV-B exposure in Japan broadens the ecologic data. This finding is of interest because most previous work had been conducted in the USA. Thus, it was important to study this association in populations with probably different confounding patterns. The finding that season of diagnosis, and presumably treatment, in Norway predicts survival from colon cancer is provocative and could be explained by an influence of vitamin D status at the time of treatment. Although speculative at this time, having a better vitamin D status could potentially enhance treatment effects. For example, higher 25(OH)D was shown to be associated with a higher apoptotic rate in colorectal tissue. Also, vitamin D could potentially influence the capacity of cells to metastasize, possibly during treatment.

Studies of genetic variations in the VDR may offer independent evidence that vitamin D is important for human colorectal cancer. A study conducted in Korea[28] provided strong results in accordance with a previous study in an Asian population.[29] These populations may differ in some ways from US populations; for example, colon cancer rates are lower, populations are leaner, diets differ, and levels and intakes of vitamin D and calcium are likely to be lower than in the USA. In US populations the relationship between VDR and colorectal cancer may be more complex, and may interact with some of these factors.[31] Nonetheless, the consistent findings thus far in Asian populations lend further evidence of a role of vitamin D in colorectal cancer in humans, because if vitamin D were not important we would not expect VDR polymorphisms to be associated with cancer risk.

Conclusion

In vitro, animal and clinical studies strongly indicate that vitamin D may have anticancer benefits, including against progression (such as metastasis) in colorectal cancer and possibly other cancers. Thus improving vitamin D status could be potentially beneficial against either incidence or mortality, or both. High-risk groups for hypovitaminosis D include individuals with low intakes of vitamin D, or who live in regions with low sunlight intensity, avoid sunlight or use sunscreen thoroughly, have darker skin, are old, live in a nursing home, or are overweight or obese. Current recommended intakes of vitamin D (for example, 400 IU/day) may be too low to provide maximal benefits, though the precise optimal dose remains unestablished. For most people solar UV-B exposure remains the major source of vitamin D, particularly since few good sources of vitamin D exist and intakes of vitamin D tend to be low. However, unprotected sun exposure is usually discouraged by health organizations because of the increased risk for skin cancer. Given the potential benefits from this vitamin against colorectal cancer in addition to other malignancies, further research should be a priority.