Vitamin D, skin cancer and all-cause mortality 3rd International UV and Skin Cancer Prevention Conference - Melbourne 2015 Sara Gandini, PhD Senior staff scientist Vice-director Division of Epidemiology and Biostatistics European Institute of Oncology - Milan, Italy
Outline Vitamin D, Cutaneous Melanoma (CM) and Non Melanoma Skin Cancer (NMSC) risk: - Serum level of 25(OH)D - Vitamin D receptors (VDR) polymorphisms - Vitamin D supplementation Sun exposure, Vitamin D and CM Disease Free Survival (DFS) Vitamin D and overall and cancer mortality
Meta-analysis: 25(OH)D and VDR and skin cancer risk First author, PY, cancer Asgari, 2010, BCC Eide, 2011, BCC Liang 1, 2012, BCC Liang 1, 2012, BCC van der Pols, 2013, BCC Summary RR for BCC= 1.82 (1.38, 2.40) I 2 =0% Liang 1, 2012, SCC Liang 1, 2012, SCC Eide, 2011, SCC van der Pols, 2013, SCC Summary RR for SCC= 1.68 (0.44, 6.39) I 2 =81% Tang, 2010, NMSC Afzal, 2013, NMSC Summary RR for NMSC= 1.64 (1.02, 2.65) I 2 =81% Nürnberg, 2009, CM Afzal, 2013, CM van der Pols, 2013, CM Major, 2012, CM Summary RR for CM = 1.46 (0.60, 3.53) 1,420 CM and 2,317 NMSC I 2 =54% 0.3 0.5 0.8 1.3 2.0 3.5 5.0 8.0 12.0 Not surprising: the main source of vitamin D is sun exposure Caini S et al. Eur J Ca 2014. - BsmI and FokI polymorphisms are associated with skin cancer risk Serrano D. et al. Eur J Cancer Prev. 2015. Gnagnarella P. et al. Carcinogenesis. 2014. Raimondi S. et al. Mutation Research, 2014.
Vitamin D supplementation and CM risk: WHI trial WHI trial Tang J Y et al. JCO 2011. Giovannucci. N Engl J Med. 2006. Overall no effect. BUT Vitamin D+cal. vs placebo: HR=0.43 (0.21-0.90) P interaction= 0.04 in women with history of NMSC History of NMSC associated with increased risk of CM. The trial was criticized for the too low dose vitamin D, the low adherence of women to supplementation.
Outline Vitamin D and CM risk: - Serum level of 25(OH)D - Vitamin D receptors polymorphisms - Vitamin D supplementation Sun exposure, Vitamin D and melanoma DFS Vitamin D and overall and cancer mortality
Cumulative Incidence Cumulative Incidence Sunny holidays and DFS in CM pts Cohort of 691 CM pts (median fu-p 4 y) Sunny holidays during f-up associated with reduced DFS. Not sunbed use or exposure during hot hours. 0.3 No sunny holidays Sunny holidays At diagnosis 0.3 No sunny holidays Sunny holidays During follow-up 0.2 Cox model P=0.16 HR=4.55; 95% CI (0.55-37.78) P=0.18 0.2 HR=0.30 (0.10, 0.87) P=0.03 adjusted for age, gender, education, Breslow, ulceration and MD degree Cox model P=0.03 HR=0.27 (95%CI: 0.08-0.90) 0.1 0.1 0.0 0 1 2 3 4 5 6 Years Competing risk framework for N0M0 pts Gandini S. et al. Plos One. 2013 0.0 0 1 2 3 4 5 6 At time of questionnaire Years 2 vs 0 weeks: HR=0.74 (95%CI: 0.16, 3.45) >2 vs 0 weeks: HR=0.28 (95%CI: 0.08, 0.98) P for trend =0.04
Change in 25(OH)D and DFS in CM pts Change in 25OHD in nmol/l HR (95%CI) <5.25 1.94 (1.36, 2.76) <5.30 to -0.3 1.23 (0.85, 1.78) -0.30 to 4.6: 1.00 >4.6: 1.61 (1.14, 2.28) Adj for stage, age, sex, BMI In a cohort of 1171 CM pts change in 25(OH)D in 1 year during follow-up was associated with DFS: very low and very high level may increase risk of relapse Saiag et al. JNCI 2015
MelaViD: No-profit multicenter randomized double blind phase III trial In Italy the incidence of CM has almost doubled in the last 10 years. For Stage II CM no standard treatment is available and the OS at 5 year is 70% Stage II melanoma n=878 R Main endpoint: DFS Sequential trial design with 3 planed interim analyses. Scientific PI: Gandini S. Preliminary analysis at baseline VD 3 2000 IU in average per day Placebo 3 years Seasons n. pts Median* Q1 Q3 August-November 28 23 18 27 December-March 19 14 10 19 April-July 41 17 14 22 Total 88 18 13 25 * 25(OH)D> 30ng/ml Gandini et al Int. J. Cancer. 2011 F-up 2 years
Outline Vitamin D and CM risk: - Serum level of 25(OH)D - Vitamin D receptors polymorphisms - Vitamin D supplementation Sun exposure, Vitamin D and melanoma outcome Vitamin D and overall and cancer mortality
I 2 =0 The reduction was 8% for studies for which the intervention was at least 3 years and for those studies with a placebo control group. Autier and Gandini. Arch. Int. Med. 2007 Meta-Analysis on VD and overall mortality The 18 RCTs included generally healthy people, at high risk for fractures.
Zitterman, Gandini et al. Am J Clin Nutr 2011 top quintile as reference Schöttker et al. BMI 2014 Meta-analysis and Pooled-analysis on 25(OH)D and overall mortality and cancer mortality Meta-analysis of 11 prospective cohort studies: 5,341 died for any cause Pooled-analysis of 7 prospective cohort studies: 2,227 died of cancer 30% (20%-40%) mortality reduction for an increase of 20 ng/ml of 25(OH)D
RCTs: vitamin D supplementation and cancer mortality RR=1.00 (0.94, 1.06) for cancer incidence For cancer mortality: 1,180 deaths I 2 =0% Vitamin D 3 suppl (given singly) decreased cancer mortality and all-cause mortality. Risk of type I errors: low power Risk of attrition bias due to dropout of participants. The Cochrane Collaboration, 2014 Keum and Giovannucci. BJC 2014
Conclusions Some evidence that vitamin D may have a beneficial influence on: - Overall mortality - Cancer mortality - Melanoma recurrence/survival (vitamin D reduction during f-up) Mainly from observational studies and RCTs but with low statistical power, low dose and low adherence Properly designed randomized clinical trials are necessary to demonstrate the effect of vitamin D supplementation on melanoma outcome, cancer mortality and overall mortality