is quite puzzling as it reports a result at odds with nearly all papers published on the topic. It is contradicted by results in a reference cited in that article, which reported that in sunny countries (Australia, Singapore and Spain), the standardized incidence ratio for prostate cancer following diagnosis of nonmelanoma skin cancer (other than basal cell carcinoma) was 0.43 (95% confidence interval, 0.23-0.73) and that for basal cell carcinoma (BCC) was 0.63 (0.44-0.89). 2 The most important risk factor for squamous cell carcinoma is integrated lifetime UVB irradiance, whereas recreational UV irradiance is the strongest risk factor for BCC. 3 Two recent studies reported that odds ratio for cancer changes according to follow-up interval after serum was drawn, changing signs in one study 4 and decreasing toward no effect in another. 5 A possible reason for the change is that serum 25-hydroxyvitamin D concentrations may change with time.Why did this study find that higher UV exposure was linked to increased risk of prostate cancer? One reason could be that the controls were not properly matched to the cases. Cases were men diagnosed with prostate cancer. Controls were individuals recruited for an earlier study of non-Hodgkin's lymphoma that had with UV exposure as a variable of interest.In the past, there was a similar issue with poorly matched cases and controls in a study of risk factors for pancreatic cancer. 6 Cases had diagnoses of pancreatic cancer, whereas controls were ''273 patients with cancer other than cancers of the pancreas and biliary tract, respiratory tract or bladder and 371 patients with other disorders.'' The study found a modestly increased risk of pancreatic cancer from cigarette smoking and a large increase with respect to drinking coffee [OR ¼ 2.6 (1.2-5.4) for males and 2.3 (1.2-4.6) for females]. A recent metaanalysis found that ''from 37 case-control and 17 cohort studies (10,594 cases), the pooled RR for the highest versus lowest [coffee] intake was 1.13 (95% CI ¼ 0.99-1.29). '' 7 The study from New South Wales could easily have been subject to a similar bias due to unrepresentativeness of the controls of the general population. No assurance is provided that the controls from the earlier non-Hodgkin's disease study were representative of the NSW population.Another bias regarding with the study by Nair-Shalliker et al. is the possibility that patients who had prostate cancer died prematurely due to cardiovascular disease, another disease linked to low serum 25(OH)D concentrations. 8 This is called the Neyman's survival bias, 9 and it has been considered as a bias in prostate cancer studies with respect to cardiovascular disease. 10 There are a number of reasons to not accept the study by Nair-Shalliker et al. as valid. The health benefits of UVB exposure and higher serum 25(OH)D levels in reducing the risk of many types of cancer 2,5 and other diseases greatly outweigh any risks associated with them. If the findings of NairShalliker et al. were to uncritically accepted as correct, de...