We read with interest the recent article by Saiag et al. ( 1 ) on the prognostic value of 25-hydroxyvitamin D3 (25[OH]D3) levels in melanoma patients. The authors reported that both high decrease (-5.25 nmol/L/year or more) and high increase (>4.60 nmol/L/year) in 25(OH)D3 levels during follow-up were associated with a higher risk of relapse after adjusting for other prognostic factors.
In their discussion, the authors stated that these results do not favor the hypothesis made in a previous publication ( 2 ) that suggested that vitamin D (vitD) might mediate the reported lower risk of relapse for melanoma patients who had sunny holidays after diagnosis. Contrarily, we find that the results presented by Saiag et al. ( 1 ) are in fact in line with the hypothesis of Gandini et al. ( 2 ). Indeed, this latter study ( 2 ) did not evaluate the change in sun exposure before and after diagnosis but found that the risk of melanoma recurrence was significantly lower in patients who had sunny holidays after melanoma diagnosis. It is likely that a patient who used to have sunny holidays stopped this behavior after diagnosis, following the common instruction of refraining from sun exposure ( 1 , 3 , 4 ). This behavior may substantially reduce the levels of 25(OH)D3 after diagnosis, therefore increasing the risk of relapse as found in the paper by Saiag et al. ( 1 ). Conversely, patients who had sunny holidays after diagnosis may likely correspond to the reference category of no change in 25(OH)D3 in the study by Saiag et al ( 1 ). Therefore, if a correct comparison between the risk groups presented in the two papers is made ( Figure 1 ), the obtained results are in agreement. The U-shaped hypothesis of an increased risk of mortality for high 25(OH)D3 levels has been investigated in a meta-analysis ( 5 ) of prospective cohorts because it is not easy in a single study to reach the statistical power required to investigate nonlinear trends. This meta-analysis confirmed the U-shaped trend but not a statistically significant risk when a high increase occurred.
Proposed comparison between risk groups in the study by Saiag et al. ( 1 ) and Gandini et al. ( 2 ) assuming changes in 25(OH)D3 as mainly due to changes in patients’ sun exposure behavior. *For a clearer comparison, we changed the reference category from the article by Gandini et al. ( 2 ) and recalculated the hazard ratio (HR) accordingly. Original HRs in the paper ( 2 ) were: 1.00 (ref) for patients with no sunny holidays after melanoma diagnosis and 0.30 (95% confidence interval = 0.10 to 0.87) for patients with sunny holidays after melanoma diagnosis. CI = confidence interval; HR = hazard ratio.
In order to confirm a causal relationship of vitD with cancer recurrence or mortality, randomized clinical trials are warranted. Therefore, we designed a randomized placebo-controlled phase III trial on vitamin D supplementation in melanoma patients (MelaVid). To date, we have randomly assigned 102 stage II melanoma patients, and the median 25(OH)D3 serum concentration at diagnosis was 18ng/mL (interquartile range = 13 to 25ng/mL). Thus, even in a Mediterranean country like Italy, 90% of patients have vitD levels below 30ng/mL, considered the optimal level for cancer prevention ( 6 ).
Because sun exposure is a recognized risk factor for melanoma, the commonly given advice to melanoma patients to reduce their sun exposure after diagnosis could further exacerbate their vitD insufficiency and potentially worsen both cancer prognosis ( 7 ) and overall mortality ( 8 ). According to published guidelines ( 3 , 4 ), we would recommend that melanoma patients have their vitD levels measured at diagnosis. Results from the ongoing clinical trials on vitD supplementation like the MelaVid trial and other large randomized phase III trials will help to determine the cause-effect relationship of vitD with melanoma outcome.
Funding
This work was supported by a Fondazione Umberto Veronesi (FUV-10) grant to SG.
References
4. http://www.genomel.org/wp-content/uploads/2015/08/Advice-for-melanoma-health-care-teams-Vitamin-D-and-melanoma-Leaflet.pdf . Accessed February 10, 2016.
