Cancer mortality rates vary inversely with geographic latitude and solar ultraviolet-B doses. This relationship may be due to an inhibitory role of vitamin D on cancer development. The relationship between vitamin D and cancer appears to be stronger for studies of cancer mortality than incidence. Because cancer mortality reflects both cancer incidence and survival, the difference may be due to effects of vitamin D on cancer survival. Here we review analytic epidemiologic studies investigating the relation between vitamin D, measured by circulating levels of 25-hydroxyvitamin D (25-OHD), and cancer survival. A relationship between low 25-OHD levels and poor survival is shown by most of the reviewed studies. This relationship is likely to be causal when viewed in light of most criteria for assessing causality (temporality, strength, exposure-response, biological plausibility and consistency). A serum level of 25-OHD around 50 nmol/L appears to be a threshold level. Conversely, there are several mechanisms whereby cancer could lower serum levels of 25-OHD. The severity of disease at the time of diagnosis and time of serum sampling are key factors to clarify the temporal aspect of these relationships. Evidence that vitamin D supplementation could retard the disease process or prolong survival time would be key evidence, but is difficult to generate. However, recent clinical trial results in prostate cancer support a role for vitamin D in this regard.