![[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}
|
|
|
|
|
||
UMR U557 Inserm/U1125 Inra/EA3200 Cnam
Univ Paris 13, Bobigny, France 93017 and
Department of Dermatology, University Hospital Erasme
Université Libre de Bruxelles, Bruxelles, Belgium 1070
EA 3677 and Centre René-Labusquière
(Tropical Medicine and International Health Branch)
Université Victor Segalen Bordeaux 2 and
Department of Internal Medicine and Tropical Diseases
University Hospital Center, Bordeaux, France 33076
EA 3444, Ecole de Santé Publique, Epidémiologie clinique
Faculté de Médecine, CHU Nancy, France 54035
UMR U557 Inserm/U1125 Inra/EA3200 Cnam
Univ Paris 13, Bobigny, France 93017 and
Departement de Santé Publique
Hôpital Avicenne, Bobigny, France 93017
* To whom correspondence should be addressed. E-mail: hercberg{at}cnam.fr.
Dear Editor,
The comments of Green et al. concerning the methods and interpretation of our study (1) raise several issues. These arise from an incorrect understanding of the methodology of the SU.VI.MAX study on 2 important points. First, Green et al. suggest that censoring of followup after an initial cancer event would artificially reduce the number of events observed. Although we acknowledge our method section was not exhaustive on this precise point, no such censoring in fact occurred when estimating melanoma risk. If an individual's first skin cancer was a basal cell carcinoma or a squamous cell carcinoma, a subsequent melanoma was recorded and might have contributed to the assessment of the intervention effect on melanoma occurence (see Fig. 2C of our article). Interestingly, the hazard ratio for melanoma was much higher in the women antioxidant group (4.32) than the hazard ratio for all skin cancers in the same group (1.7). This finding reinforces our principle hypothesis that antioxidant supplemention increases the risk of skin cancer, namely melanoma, in women. Second, Green et al. postulate that the findings rested on a post hoc analysis in a subgroup of patients. Again, this was not the case for melanomas; the post hoc analysis was complementary and restricted to non-melanoma skin cancers. Estimation of the incidence of melanoma in particular was a protocol-specified outcome in the SU.VI.MAX study and this was taken into account in the calculation of the sample size of the study (2). Indeed, the controlled experimental trial design was precisely aimed at reducing inference in interpreting the magnitude of the observed hazard distributions. Thus, these 2 methodological criticisms are unfounded. Green et al. also raise concerns about the scope for incomplete or inaccurate case ascertainment. Given that the SU.VI.MAX study was designed to obtain information on serious medical events, such as cancers, which are unlikely to be underreported due to forgetting events in the intervening time, and that all cases were ascertained by an independent expert adjudication committee of oncologists who were unaware of treatment assignment, the scope for inadequate case assignment seems likely to be very limited.
Another point addressed by Green et al. relates to the counting of very early melanoma events in the intervention period. However, no such melanoma event was recorded in the very early intervention period, particularly among women (see Fig. 2 of our article). We did not claim that the treatment intervention triggered the appearance of melanomas and we explicitly acknowledged that these cancers developed over many years and were likely to be present in a preclinical form prior to the study in some individuals. Rather, we speculated that increased antioxidant exposure might exert a negative influence on antitumoral immunity, angiogenesis, or apoptosis leading to more rapid evolution of preexisting protomelanomas in individuals in the antioxidant group. As we stated, this might explain the more pronounced effect on melanoma, which requires a longer time to develop than other types of skin cancer. Our findings are consistent with the results of many interventional and mechanistic studies suggesting that antioxidant supplementation, especially with β-carotene, may have potentially harmful effects (3,4).
Contrary to the claim of Green et al., the incidence of skin cancers was higher in men (76 cases among 5141 men: 1.5%) than in women (81 among 7876 women: 1.0%) and age and the incidence of skin cancer were significantly correlated in both genders. The link with age was less noticeable for melanoma specifically, which can probably be explained by the fact that melanoma is waited to be more encountered in young adults.
It is also important to keep in mind that different "antioxidant" nutritional supplements vary considerably in their chemistry, metabolism, and biological activities and not all of their properties are necessarily related to their ability to act directly as antioxidants. The link between antioxidant nutrients and carcinogenesis is far more complex than previously assumed. Given the cocktail of vitamin and mineral supplementation used in the SU.VI.MAX study, it is not possible to implicate any particular component in the effects observed. However, whatever the explanation, the findings of the SU.VI.MAX study provide no support for the view that antioxidant supplementation is a useful strategy for cancer prevention.
Manuscript received 15 February 2008.
 |
LITERATURE CITED |
|---|
 TOP LITERATURE CITED |
|---|
1. Hercberg S, Ezzedine K, Guinot C, Preziosi P, Galan P, Bertrais S, Estaquio C, Briancon S, Favier A, Malvy D. Antioxidant supplementation increases the risk of skin cancers in women but not in men. J Nutr. 2007;137:2098–105.
2. Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briancon S. The SU.VI.MAX study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med. 2004;164:2335–42.
3. Bjelakovic G, Nikolova D, Gluud LL, Simonetti RG, Gluud C. Mortality in randomized trials of antioxidant supplements for primary and secondary prevention. JAMA. 2007;297:842–57.
4. Lee BM, Park KK. Beneficial and adverse effects of chemopreventive agents. Mutat Res. 2003;523–524:265–78.[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
