Shedding Light on Colorectal Cancer Prognosis: Vitamin D and Beyond

With more than 153,000 new colorectal cancer cases per year and more than 52,000 patients dying annually from their disease, colorectal cancer is a major cause of cancer death in the United States. Yet, perhaps more importantly, the number of survivors of colorectal cancer in the population increases steadily, attributable in part to the implementation of new screening and treatment regimens. In 2002, the number of cancer survivors in the United States was estimated at 10.1 million, among them more than 1 million colorectal cancer survivors, and this number continues to rise. Yet, surprisingly, very little research has addressed how lifestyle factors, both during and after active treatment, can affect the prognosis of colorectal cancer patients. In this issue of the Journal of Clinical Oncology, Ng et al present data from the Nurses’ Health Study and the Health Professionals Follow-Up Study, showing a strong association between increasing plasma vitamin D levels (25(OH)D) and reduced mortality. These findings are impressive for their impact on overall mortality, suggesting a 48% reduction between the highest and lowest quartile of plasma 25(OH)D, with a linear and statistically significant trend. However, for colorectal cancer mortality, the multivariate hazard ratio was attenuated to a statistically nonsignificant reduction of 39% between the extreme quartiles, with no evident trend. Clearly, the sample size of about 300 patients and only 96 colorectal cancer–specific deaths was a limiting factor. Additional limitations result from the study design, because the study cohorts were not originally established to investigate cancer prognosis. For example, the blood draws used for the 25(OH)D measurement occurred before diagnosis—about 3 years on average— with variable time frames between the exposure measurement and subsequent cancer diagnosis. As the authors note, associations with prediagnostic 25(OH)D level may also reflect the influence of other factors known to be related to vitamin D status. Most important in this context could be physical activity, which is somewhat difficult to measure and quite difficult to adjust for, yet correlates with 25(OH)D levels, presumably due to increased sun exposure associated with outdoor exercise. In this study population, postdiagnosis physical activity was strongly associated with 25(OH)D (ptrend 0.003); physical activity has been associated with improved survival after several types of cancer, including the colorectal cancer in this cohort. A further limitation is the lack of treatment information, a key component of prognostic studies. Nevertheless, despite these drawbacks, the findings by Ng et al are intriguing and important. They highlight the need for well-designed prognostic studies among cohorts of cancer patients, in which exposure assessments are uniformly obtained after diagnosis, treatment regimens are assessed, and information on prognostic tumor characteristics (such as microsatellite instability) is also available. We are just beginning to identify important lifestyle determinants of survival and prognosis, and vitamin D status may emerge as a promising predictor that could be modified. Vitamin D has also received much attention recently as a possible cancer-preventive agent. There is substantial evidence that it suppresses proliferation and promotes differentiation and apoptosis in colorectal cells, through effects on growth factors and cytokines, cellcycle regulation, and apoptosis. Nevertheless, despite intensive study, the role of Vitamin D in the primary prevention of colorectal cancer remains uncertain. Ecologic analyses nearly 30 years ago associated sunlight exposure and resulting vitamin D synthesis with reduced risk of colorectal cancer. Subsequent epidemiologic studies of dietary vitamin D alone have been inconclusive, but more consistent risk reductions have been shown for analyses of both diet and supplements as sources of vitamin D. Plasma 25(OH)D, which reflects body stores of vitamin D from all sources, including diet, supplements, and endogenous synthesis from UV-B exposure, has been associated with reductions in the risk of colorectal cancer of approximately 50% between the highest and lowest quintile, although not all studies have reported decreased risks. Vitamin D synthesis in the skin varies with latitude, skin color, and sun exposure, but almost always has a greater effect on plasma 25(OH)D than dietary intake. Accordingly, observational studies of the relationship between vitamin D and colorectal cancer are susceptible to confounding factors related to sun exposure, which are difficult to assess. As discussed previously, this can include the notoriously difficult to measure amount of physical activity, with exercise known to be a very strong protective factor for colorectal cancer. In addition, race can be a major determinant of vitamin D status, because melanin in skin reduces the rate of synthesis; vitamin D deficiency is particularly widespread among African Americans, including many African American cancer survivors. Randomization can eliminate these sources of confounding, and some randomized trials have been conducted on vitamin D and colorectal and other cancers. The multicenter Women’s Health Initiative included a trial of calcium (1,000 mg calcium/d) plus vitamin D supplementation (400 IU/d) in 36,282 postmenopausal women with a JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 26 NUMBER 18 JUNE 2