■ site map
melanoma risk: 4-fold
Hercberg et al, Antioxidant Supplementation Increases the Risk of Skin Cancers in Women but Not in Men (Journal of Nutrition, 2007)
PARTICIPANTS - Female: 7,876 (35-60y) Male: 5,141 (45-60y) Total: 13,017
DURATION - October 12th 1994 to November 1st 2002, median follow up 7.5y
TREATMENT GROUP - 6,481 (87.4% completed)
TREATMENT - 120mg
vitamin C (sodium ascorbate)
PLACEBO GROUP - 6,536 (85.3% completed)
RESULTS: - Treatment group: 84
skin cancers, 16 melanomas (13 women, 3 men)
What immediately sticks out is the very small number of melanoma cases: only 9 men and 16 women in both, treatment and placebo group. It is a result of the generally low melanoma incidence rate, and definitely
insufficient for extrapolating statistically reliable associations.
The study authors used these numbers nevertheless, but it does not change the fact that such associations, based on statistically insufficient sample size, do not have realistic grounds.
It is not surprising, then, that the study ended up with an identical supplemental regime more than quadrupling melanoma incidence in women, while halving it in men. Such wild, unpredictable deviations from the large-number-average in a small sample size are rather common. But there are few other points that add to the overall design weakness of this study.
It begins with the supplemental doses. While 120mg of vitamin C a day is likely to prevent scurvy, it is by far insufficient to significantly enhance body's overall antioxidant activity in these times of high oxidative stress. Several times larger dose would be required for such effect. Similarly, 30mg of synthetic vitamin E - amounting to 33 IU - will not appreciably better body's anti-oxidative protection. At least 100 IU would be needed for that, preferably more - 200 to 400 IU - and in the form of natural mixed tocopherols.
The β-carotene dose was close to the DRI value; again, not enough to make significant difference, especially if it is the synthetic form.
Selenium and zinc doses are probably adequate, at nearly twice their DRI values of 55μg and 8/11mg (women/men) a day, respectively. But they
don't guarantee higher body levels
than in individuals not taking any supplements. These doses alone are also unlikely to correct existing serious imbalance of these minerals in the body, especially if it is caused by metabolic inefficiency, higher than average individual need, and/or deficiency or excess of other nutrients and/or pharmacological factors (so called antagonists) that directly influence their body levels.
In order to control this aspect, the study measured body levels of the supplemented nutrients, but the manner in which it was conducted, and results obtained, are rather unconvincing.
The test of choice was serum test, notoriously unreliable for assessing nutrient status, particularly minerals. Blood samples were taken at the beginning (baseline), every 2 years after, and before study's ending, after 7 years, on a randomized sub-sample of 1,134 participants.
These methods are, frankly, not reassuring. Nutrient status can and often does vary significantly within such long periods, and less than 9% sample of all participants is all but reliable representation of the group as a whole.
These four factors - the doses, the test of choice, infrequent measurements and small sample of participants - combined mean that the study
had no means of determining accurate status of these nutrients
averaged over each of the two groups, much less anything coming close to a complete nutritional status, much less an insight in all major factors affecting cellular homeostasis - including body's toxic load - that are significant in developing skin cancer in general, and melanoma in particular.
Without reliable information on the actual nutritional and overall metabolic status of the participants, concluding that the grossly under-sampled number of melanoma cases could be caused by low to moderate antioxidant supplementation is a
double shot in the dark.
In addition, the study relied on subjective accounts and recollections of the degree of recent and lifetime sun exposure as furnished by participants.
No wonder that the numbers came out indicating that taking these limited in dose and number antioxidants, vital for human health, is nearly twice more of a melanoma risk factor for women
than the degree of sun exposure itself!
And, that, on the other side, that same antioxidant supplementation cuts melanoma incidence among men in half.
Since the nutrient with the most significant difference in its body level, as measured in the study, was β-carotene in women vs. men - nearly 50% higher on average - study authors turned to it for speculations of high antioxidant levels reducing the ability of natural killer lymphocytes to destroy cancerous cells, as well as some "possible" obscure gender-based differences in metabolizing the antioxidants.
Again, it is highly unlikely that this level and type of supplementation would result in appreciable suppression of killer lymphocytes activity. And, study's shortcomings mentioned above are quite sufficient to explain its contradictory, illogical results; no need to speculate about antioxidants hurting women in some mysterious way.
Furthermore, considering that the treatment group had "significantly" higher level of supplemented antioxidants only at the 2-year and 7-year measurement, for all supplemented antioxidants except zinc, and
no significant difference at the 4- and 6-year measurements,
the theory of these antioxidants playing a major role in study's numerical outcome becomes even more baseless.
In addition, the authors seem to treat as "significant" any difference in the serum level over 10%, or so, of the nutrients in question (specific criteria for "significant difference" was not specified), an increment well within common individual variations. The single supplemented antioxidant that had potentially significant level differential between women and men at the start of the study, as mentioned, was β-carotene; hence, any suspicion of the possible contribution of antioxidants to the development of melanoma should be limited to this particular nutrient, not all 5 antioxidants.
But, as detailed above, study design does not give sufficient basis for associating supplementation of even that one antioxidant with the increased risk of melanoma.
In all, we can conclude that
this study was poorly designed, controlled, and interpreted,
suffering from symptoms typical of the conventional medicine: lack of acquired and applied knowledge of nutritional and whole-body biochemistry and - directly resulting from it - ignoring the importance of monitoring the entire set of factors potentially important in producing particular outcome.
It is important to understand - and that is one of study's conclusions - that
antioxidants have primarily preventive power
when it comes to diseases like skin cancer. Majority of melanoma cases had it initiated much before they went on antioxidant supplementation, due to UV-radiation-induced cellular damage that could have occurred as far back as early childhood. There's not much that antioxidants can do after the damage is done. From that standpoint, it wasn't to expect that they would significantly reduce skin cancer incidence.
The speculation that antioxidants in this particular study had interfered with the immune system cancer-fighting ability is very unlikely considering low-to-moderate doses used. But this doesn't mean that interference with the immune system, or some other body function, couldn't be among possible unhealthy consequences of a really excessive body level of antioxidants.
Just another reminder that nothing is unconditionally good, and that nutrient excess - especially when imbalanced - can be as damaging as deficiency.