![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
|
|
|
|
|
||
Cancer and Population Studies Unit
Queensland Institute of Medical Research
Brisbane, Australia 4029
School of Population Health
University of Queensland
Brisbane, Australia 4006
Cancer and Population Studies Unit
Queensland Institute of Medical Research
Brisbane, Australia 4029
and School of Population Health
University of Queensland
Brisbane, Australia 4006
Department of Dermatology
University of Nice
Nice, 06202 France
* To whom correspondence should be addressed. E-mail: adele.green{at}qimr.edu.au.
Dear Editor,
Hercberg et al. (1) recently proposed that antioxidant supplementation increases the incidence of skin cancers, including melanomas, in women but not in men. This proposal was based on a secondary analysis of data from the Supplementation in Vitamins and Mineral Antioxidants (SU.VI.MAX) trial designed to assess whether daily supplementation with antioxidant vitamins and minerals could reduce the incidence of chronic diseases and cancer.
We consider this conclusion to be unjustified based on flaws in the methods and in the interpretation of results.
The analysis endpoint was first occurrence of any type of skin cancer and follow-up was thus censored at this first occurrence. For instance, if a person's first skin cancer was a basal cell carcinoma, a subsequent squamous cell carcinoma or melanoma would not have contributed to the assessment of antioxidant effects. Despite this protocol, the published finding rested on a rare post hoc subtype analyzed within a subgroup, namely, melanoma in women (16 events), which was neither an a priori trial hypothesis nor a planned endpoint. (The occurrence of only 2 or 3 extra cases in the placebo group could have tipped the findings to nonsignificance.) Given the possible, even likely, omission of events due to censoring [because of increased risk in those experiencing a first skin cancer (2,3)], as well as ascertainment uncertainties (extent of coverage; whether performed in a blinded manner), considerable doubt exists as to the validity of this finding.
There were some related problems involving the counting of events very early in the intervention period, such as some of the "excess" melanomas in women in the treatment group, that were unlikely to be causally related to the intervention. These melanomas were likely to have been present prior to the intervention in preclinical form given the long latent period of the disease. Their occurrence would have eliminated later events that may have been more relevant to attribution of risk as well as to the overall burden of skin cancers. Also, it is puzzling that skin cancer incidence was higher in women despite the women taking part being 5 y younger on average than the men, raising further doubts about the sex-specificity of the finding.
Manuscript received 20 November 2007.
 |
LITERATURE CITED |
|---|
 TOP LITERATURE CITED |
|---|
1. Hercberg S, Ezzidine K, Guinot C, Preziosi P, Galan P, Bertrais S, Estaquio C, Brian
on S, Favier A, et al. Antioxidant supplementation increases the risk of skin cancers in women but not in men. J Nutr. 2007;137:2098–105.
2. Green AC, O'Rourke MGE. Cutaneous malignant melanoma in association with other skin cancers. J Natl Cancer Inst. 1985;74:977–80.[Medline]
3. Swerdlow AJ, Storm HH, Sasieni PD. Risks of second primary malignancy in patients with cutaneous and ocular melanoma in Denmark, 1943–1989. Int J Cancer. 1995;61:773–9.[Medline]
This article has been cited by other articles:
|
|
 |
|
  M. M. Asgari, S. S. Maruti, L. H. Kushi, and E. White Antioxidant Supplementation and Risk of Incident Melanomas: Results of a Large Prospective Cohort Study Arch Dermatol, August 1, 2009; 145(8): 879 - 882. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
