Background: Low 25-hydroxyvitamin D3 (25(OH)D3) serum concentration at diagnosis of melanoma might be associated with worse survival. We prospectively studied the prognostic value of 25(OH)D3 at diagnosis and during follow-up. Methods: Melancohort is a cohort of patients recruited in Paris hospitals in 2003-08 within 1-3 months after diagnosis of AJCC stage I-II, III or IV invasive melanoma and followed until 06/2011. Blood 25(OH)D3 was measured by mass spectrometry. 25(OH)D3 levels were standardized on month of blood drawn, age, sex and body mass index (BMI). Role of 25(OH)D3 level at inclusion and change through time on disease-free survival (DFS) were analyzed in a time dependent covariates Cox model adjusting for age, sex, BMI and AJCC stage. Results: 1,171 patients were included, with 411 first relapses during follow-up and 303 deaths. On average 3.25 measures per patient were performed. Median 25(OH)D3 serum concentration at inclusion was 47.2 nmol/L (interquartile range 30.65-63.18). 25(OH)D3 levels at inclusion were inversely correlated with prognostic factors such as AJCC stage (p<0.0001 Kruskal-Wallis), Breslow’s thickness (p<0.0001 spearman correlation), and ulceration (p=0.0004 Kruskal-Wallis), but were not associated with DFS in the Cox model, either as a continuous (p=0.747) or categorical variable. Changes in 25(OH)D3 levels during follow-up were significantly associated with worse DFS: with a third quartile Q3 of average change per year (-0.3;4.6nmol/L/Y) used as reference, HR were for Q1 (<-5.25nmol/L/Y) 1.94 95%CI(1.36-2.76), for Q2 (-5.25;-0.30nmol/L/Y) 1.23 95%CI(0.85-1.78), and for Q4 (>4.60nmol/L/Y) 1.61 95%CI(1.14-2.28). In sensitivity analyses, no changes were observed either by AJCC stage or by number of 25(OH)D3 measures performed. Analyses performed on overall survival yielded similar results. Conclusions: We show that 25(OH)D3 variation during follow-up is an independent melanoma prognostic marker, but not its level at diagnosis. The biological significance of this finding should be investigated. Previously reported association between low 25(OH)D3 level at diagnosis and poor prognosis were probably due to insufficient adjustment for prognostic factors. Clinical trial information: NCT00839410.