In 1928, the chemist Adolf Windaus (1876-1959) was awarded the Nobel Prize in recognition of his landmark studies that paved the way to the identification of vitamin D as the antirachitic compound required to prevent and reverse rickets in humans. As a member of fat-soluble secosterols engaged in the absorption of calcium and other minerals in the intestine and kidneys, the vitamin regulates calcium mobilization from the skeleton through enhanced osteoclast formation, a mechanism that allows for fine-tuning of both the expression and enzymatic activity of specific hydrolases in the liver and kidneys, being ultimately responsible for serum levels of the hormonally active 1,25(OH)2D3 metabolite (cholecalciferol). Early radiation experiments in animals provided insights to later understand that the main source of vitamin D in humans is conversion of 7-dehydrocholesterol by ultraviolet B light in the skin; yet, a limited supply of vitamin (vitamin D2 or ergocalciferol) is also provided by dietary components such as fish, liver, and dairy products. Whereas vitamin D exerts its genomic activity through binding to the vitamin D receptor (VDR), an endocrine member of the nuclear receptor superfamily, not unexpectedly, significant individual differences in the metabolism of this vitamin occur as a consequence of polymorphisms of the genes encoding enzymes involved in vitamin D synthesis and activation, vitamin D binding protein, and VDR.[3-5] Though the levels of vitamin D defining insufficiency are debated, a blood level of 25(OH)D3 less than 20 ng/mL (50 nmol/L) constitutes a state of insufficiency, which is common in the industrialized Western countries, mainly as a consequence of limited population exposure to sunlight, a lifestyle that is widely encouraged to prevent skin cancer. Apart from its hormonal activity, vitamin D is credited with being protective against several benign diseases and malignancies. The theory that vitamin D insufficiency may be associated with an increased risk of cancer is biologically plausible, with a wealth of studies supporting the immunologic, proapoptotic, antiproliferative, and antiangiogenic properties of the vitamin, not to mention the expression of VDR in more than 20 human cancer cell lines and the up-regulation of genes of the Toll-like receptor family involved in oxidative stress and inflammation mechanisms associated with vitamin D insufficiency.[7, 8] However, against a consistent and conclusive body of evidence of causality for the association between cancer and vitamin D insufficiency stands more than one analysis of observational studies.[9-12] This is even more so in the field of liver cancer, a tumor that develops in the context of such well-identified risk factors as viral hepatitis, alcohol abuse, and metabolic syndrome, where the risk of neoplastic transformation of the liver is fueled by both constitutional and lifestyle variables. Currently, one case-control study in 226 patients with an incidence of hepatocellular carcinoma (HCC) from a hyperendemic area for liver cancer in China negates any significant association between HCC risk and circulating levels of vitamin D.
Instead, such an association is clearly purported by a nested case-control subanalysis of the multinational prospective study (The European Prospective Investigation into Cancer and Nutrition; EPIC) that was designed to assess the association of diet, lifestyle, and environmental factors with the risk of various types of cancer and other chronic diseases. The scrutiny of a large cohort of 520,000 participants of both sexes, mainly recruited from the general population in 23 centers of 10 European countries, led to the identification of 204 cases of HCC through record linkage with population-based regional cancer registries or by a combination of methods, including surveillance in a minority of cases. Matching 138 cases with serum specimens available for 25(OH)D3 analysis for sex, age, study center, date of collection, fasting status, and menopausal status to 1 cancer-free control selected from the cohort, a significant correlation was demonstrated between vitamin D deficiency, defined as <50 nmol/L, and risk of HCC. After adjustment for smoking, body mass index, coffee, and alcohol drinking, the likelihood of developing HCC in deficient patients was 1.82, compared to reference subjects with physiological levels of 25(OH)D3; however, higher levels of vitamin D were not associated with a reduced risk of developing liver cancer. To minimize the risk of spurious associations caused by the availability of a single sampling for 25(OH)D3, outcomes were carefully controlled for confounding as a result of season and month of collection and the time interval between sampling and diagnosis of HCC. The association between vitamin D deficiency and increased risk of liver cancer was clearly confirmed.
At first glance, it seems impossible to reconcile these findings with the conclusions of analyses of previous observational studies and the only longitudinal study that disqualified the association between vitamin D and risk of liver cancer. It is also hard to reconcile findings of the present EPIC analysis with a recent subset analysis of the EPIC study where the same investigators demonstrated a positive association between consumption of dietary calcium and vitamin D with first incidence of HCC and absence of any association between HCC and vitamin D from nondairy sources. However, mining through the demography of the study population of EPIC presented in Hepatology, some inconsistencies may be identified that challenge the robustness of study conclusions in terms of evidence-based medicine. First, the number of index cases enrolled in this study was reduced by 32% from the original 204 cohort members with HCC as a result of the lack of a serum sample for vitamin D analysis in 66. Among the 138 patients who ultimately were considered for the analysis, 38 (28%) were missing any data on dominant risk factors for HCC, such as viral hepatitis B (HBV) and C virus (HCV), making therefore a proper association analysis between HCC and risk factors possible in only half of the original members of the liver cancer population. One wonders whether a possible approach to compensate for the limited sample size of the study and incomplete identification of all demographic predictors of HCC would be matching the index cases with more than one control member of the EPIC cohort. Second, we were puzzled with the low prevalence (36%) of viral hepatitis markers in HCC patients enrolled in EPIC, compared to the prevalence of the same markers that is officially attributed to HCC patients in the European countries that contributed to EPIC. Conversely, we noticed no substantial differences in the cumulative rate (3%) of HBV and HCV carriers among the control population recruited from the same regions where HCC cases were identified and the general population in most countries of Southern Europe, whereas it was only slightly higher than that reported among the general population in Central Europe. All in all, these discrepancies underlie the risk of a biased process of subject selection in the EPIC study. Finally, whereas we acknowledge that the role of vitamin D in liver cancer may be supported by the finding of each 10-nmol/L decline of serum vitamin D being associated with a 20% increase in HCC risk in the EPIC population, such an association is questioned by the loss of statistical significance of the incidence-rate ratio of HCC in the strata of women, individuals lacking viral hepatitis, and those with normal liver function. In all geographical areas of the world, HCC incidence is increased in men and liver disease patients of any etiology, including viral hepatitis, whereas serum levels of vitamin D are generally lower in individuals with chronic liver diseases and other risk factors that may independently be associated with HCC, as acknowledged by the researchers.[13, 19, 20] All these considerations add to the widely acknowledged argument that epidemiologic associations in general, and even more so between vitamin D and liver cancer risk, do not necessarily imply causality.
All these study inconsistencies notwithstanding, even assuming that EPIC did provide robust evidence for an association between vitamin D insufficiency and an increased risk of liver cancer, the reader is still puzzled with the pragmatic implications of these observations (i.e., whether physiological levels of vitamin D need to be restored in deficient individuals to prevent liver cancer). Intriguingly, the same EPIC study provided an unequivocal answer in having failed to demonstrate any protection against HCC in individuals circulating physiological levels of vitamin D.
In conclusion, there is little doubt that, in cancer, the vitamin D system is altered and this may hold true for HCC as well, with the caveat, however, that the available literature does not clarify whether serum levels of vitamin D predict HCC risk independently on well-recognized host and environmental risk factors for liver cancer. By the same token, causality for the association between vitamin D deficiency and liver cancer remains largely unproven, leaving therefore little chance that restoration of physiological levels of vitamin D in deficient individuals is to be recommended as a measure of primary prophylaxis for liver cancer.
Massimo Colombo, M.D.
Angelo Sangiovanni, M.D.
Department of Liver, Kidney, Lung and Bone Marrow Units and Organ Transplant
Division of Gastroenterology and Hepatology
Centro AM e A Migliavacca for the Study of Liver Disease
Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico
Università degli Studi di Milano