Letters

Sunlight and health

Use of sunscreens does not risk vitamin D deficiency

Article did not help informed debate

Not all sunlight is dangerous, just ultraviolet radiation

Severity of effect depends on where you live

Exposure to sunlight may reduce cancer risk

Authors' reply

Use of sunscreens does not risk vitamin D deficiency

EDITORNess et al misinterpret our work in Australia by stating that we showed that use of sunscreens reduced vitamin D concentrations.¹ In fact, we showed that the use of sunscreens did not prevent the normal summer rise in 25-hydroxy vitamin D concentration (the vitamin D fraction that is used to assess vitamin D deficiency).² Subjects using sunscreens compared with controls using a placebo cream had an equal rise over the summer.

As an extension of the study we also measured concentrations of 1,25-hydroxy vitamin D. This fraction of vitamin D is believed to be regulated by 25-hydroxy vitamin D concentrations via a negative feedback mechanism. We found a rise in 1,25-hydroxy vitamin D concentrations over the summer in people using sunscreen, but the rise was smaller than that among those using placebo. In a small group there was a lower, but not significantly different, concentration of 1,25-hydroxy vitamin D at the end of summer. However, this occurred in the presence of a substantial rise in 25-hydroxy vitamin D concentrations (the vitamin D fraction that controls 1,25-hydroxy vitamin D) in all subjects. We also made it clear in our paper that the 1,25-hydroxy vitamin D concentrations of all the participants in the placebo and sunscreen groups remained within the normal range all the way through the study.

It is mischievous and specious to imply that our work was suggesting that use of sunscreen might lead to vitamin D deficiency. We made the opposite point very strongly in the paper and clearly explained why, in the presence of raised 25-hydroxy vitamin D concentrations, the slightly smaller changes in 1,25-hydroxy vitamin D could not be interpreted as a risk for vitamin D deficiency.

Robin Marks, professor of dermatology. 
St Vincents Hospital, Melbourne, Fitzroy Victoria 3065, Australia ebejerjd{at}svhm.org.au

Article did not help informed debate

EDITORDebating the content of health education messages, the scientific evidence on which they are based, and their likely effects, both bad and good, is important provided that the debate is constructive and based on a sound literature review. Unfortunately the article by Ness et al inadequately addresses the debate about sun protection.¹

The article opens with the statement that certain professionals embrace the notion that "sunlight is bad for health," implying that the message has been to avoid sun exposure. This misrepresents the advice given, which includes "avoiding excessive sun exposure" and "encouraging gradual sun exposure."²

No short article can do justice to the extensive research, both epidemiological and experimental, into the role of exposure to ultraviolet radiation in the aetiology of skin cancer. The limitations and uncertainties about our understanding of the problem have been the subject of international, interdisciplinary debate. The dilemma facing those who wish to prevent extensive morbidity from skin cancer, and rising mortality from melanoma, is that effective primary prevention may take 20 years or more to reduce the incidence of disease. As is the case with the prevention of other diseases, there has been much pressure to promote lifestyle and dietary changes.

Ness et al suggest that early detection and treatment may be more beneficial than primary prevention in reducing mortality from melanoma, but in countries with a low incidence of melanoma we have yet to agree on a cost effective strategy for early detection. ^{3 4} Nor do the authors address the psychological and financial costs of diagnosis and treatment.

Primary prevention undoubtedly brings costs as well as benefits. However, there is insufficient evidence at this stage for Ness et al to suggest that recommendations on sun exposure should be promoted to prevent one disease at the expense of another. It may well be that greater understanding of the problem will enable development of recommendations on sun exposure that can benefit more than one condition.

It is essential that people should have the opportunity to make fully informed choices about their lifestyles and that decisions on prevention are based on the best available evidence. Although Ness et al raise important questions about how best to prevent skin cancer, from an epidemiological view the article is disappointing because it has not adequately reviewed all the health issues and because it has misrepresented at least two studies. ^{2 5}

Jane Melia, epidemiologist. 
Cancer Screening Evaluation Unit, Institute of Cancer Research, Sutton, Surrey SM2 5NG melia{at}icr.ac.uk

Not all sunlight is dangerous, just ultraviolet radiation

EDITORBy referring to just sunlight rather than its individual components, Ness et al apparently fail to understand the message which photobiologists and dermatologists have been seeking to put to the public for many years.¹ Sunlight, as the authors imply, is indeed essential to life, but its warmth and light are all we need. The third component, ultraviolet radiation, is universally harmful to the skin, although usually not noticeably so in the early years of life.

Now that we live much longer, however, exposure to ultraviolet radiation has become of major importance. It causes photoageing of skin in virtually everyone, particularly the fair skinned, inducing a dry, often itchy, wrinkled, blotchy, telangiectatic effect,² and skin cancer in a significant but steadily increasing minority. About 50 000 people in the United Kingdom develop skin cancer each year, nearly 2000 of whom die; however, the number of cases is beginning to fall in countries, such as Australia, where community ultraviolet avoidance programmes are well established.³ The authors assert that these death rates are not very high and that only a small reduction will be achieved by avoiding sunlight. Nevertheless, any skin cancer is detrimental to the sufferer and costly to health services, particularly when preventive measures are readily available. The high outlay on cosmetics and surgery to repair photoageing could also be significantly reduced.

Ness et al further state that mental health, and particularly seasonal affective disorder, is improved by sunlight. But this effect is by exposure to the essentially harmless visible light spectrum through the eye, not to damaging ultraviolet radiation through the skin.⁴ Similarly, vitamin D is readily available in a normal diet, and not sunbathing is unlikely to lead to a deficiency.⁵

What we must understand is that ultraviolet radiation is the sunlight we must avoid and that this is at its most intense between about 11 am and 3 pm in Britain in summer and all year round in tropical climates, even on cloudy or cool days. Adequate protection at such times can easily be achieved by covering up with appropriate clothing, seeking the shade when possible, and using a high protection sunscreen on exposed skin. At other times, however, even in hot and sunny conditions, the ultraviolet intensity is much weaker and less caution is needed; if sunscreens are used the risk is further minimised.

J L M Hawk, professor of dermatological photobiology. 
Department of Photobioology, St Thomas's Hospital, London SE1 7EH

Severity of effect depends on where you live

EDITORNess et al questioned the advice given by public health authorities to reduce exposure to sunlight.¹ Without quantifying the risks and benefits of sun exposure across the population, the authors reasoned, isn't it unethical to advocate a change in behaviour? They then conducted a brief review of known harms and possible benefits of sunlight exposure and concluded that increased exposure to the sun might be beneficial when assessed on a population basis. Predictably, the article has created a storm in the lay press, but what of its scientific content?

The question posed is certainly reasonable. However, the article is flawed by superficial interpretation, a disturbing tendency to equate conjecture with evidence, and a failure to appreciate the adverse effects of sun exposure experienced by people living in other parts of the world. For example, the authors claim that reductions in mortality from melanoma by reducing exposure to the sun will be small and suggest that it would be better to train the public to consult doctors at an earlier stage in the disease process. Even in countries with low rates of melanoma, such as England and Wales, these claims are contentious; when applied to the sun ravaged populations of Australia, New Zealand, and low latitude United States, they are incomprehensible.

Two possible benefits of sunlight exposure were expounded: a reduction in coronary heart disease and improvements in mood and wellbeing, although the evidence proffered for these claims was extremely weak. Assuming that high levels of sunlight exposure are beneficial, then at the crudest level fair skinned Australians might be predicted to have lower cardiovascular mortality and fewer suicides than their northern European cousins. However, death rates for heart disease among Australian men in the MONICA cohorts are higher than those for Iceland, Denmark, or Sweden,² and suicide rates in Australia are among the highest in the world.³ Moreover, people who migrate from England, Wales, and Ireland to Australia commit suicide at higher rates than those who remain behind.⁴

Most people would agree that simple health education messages are blunt tools for addressing complex health problems, but I wonder about the consequences of the high profile strategy adopted by Ness et al. The distilled message of the article (intended or otherwise) that "sunlight is good for you" will echo far beyond the lush pastures of the Avon valley into dusty, sunburnt townships half a world away. It is here that the damage will be done.

David C Whiteman, Nuffield medical research fellow. 
Imperial Cancer Research Fund General Practice Research Group, Institute of Health Sciences, University of Oxford, Oxford OX3 7LF david.whiteman{at}dphpc.ox.ac.uk

Exposure to sunlight may reduce cancer risk

EDITORThe criticism of Ness et al's article questioning the acceptance of sunlight as being bad for health¹ in the lay press underlines how the authors have dared to question one of the axiomatic tenets of modern preventive medicine. Sunlight exposure is viewed as one of the major avoidable causes of cancer, ranking alongside cigarette smoking in the demonology of medicine. It is therefore surprising that the authors did not consider the evidence that, far from causing cancer, sunlight exposure might actually be a potent agent for its prevention.

Several studies have examined the relation between sunlight exposure and internal malignancy. Several of these give sufficient information to allow the effect of changing sunlight exposure on the expected rate of malignancy to be estimated. All the studies show a negative relation of similar magnitude (table), particularly for breast and colon cancer.

In 1995 there were about 30 000 new cases each of breast and large bowel cancer in the United Kingdom and about 30 000 deaths from the two tumours combined. The most conservative of the estimates from the above studies suggest that a 10% decrease in sunlight exposure might lead to a 6% increase in these figures. This would approximate to 1800 extra cases of each of the tumours and 1800 extra cancer deaths. This figure exceeds the total number of deaths due to malignant melanoma, which are unlikely to be totally prevented by such a modest reduction in sunlight exposure.

Thus, reducing exposure to solar radiation, far from preventing cancer, may have the opposite effect. Further research is urgently needed to determine whether this is the case. If the increase is confirmed it will be necessary to determine what aspect of sunlight protects against cancer. Vitamin D or its metabolites may play an important part, offering hope for a strategy of moderating sunlight exposure to minimise the risk of skin cancer but replacing vitamin D to prevent internal malignancy.

Peter L Selby, lecturer in medicine. 
Peter.Selby{at}man.ac.uk

E Barbara Mawer, professor. 
Department of Medicine, Manchester Royal Infirmary, Manchester M13 9WL

Authors' reply

EDITORExposure to sunlight clearly is not unequivocally noxious, as has been observed by others. ^{1 2} Our article was an attempt to consider harm in the context of some potentially protective effects and to suggest that health education messages may be less fragile if a more balanced portrayal of risk and benefit is offered to the public than is sometimes the case. The intemperate tone of some of the responses to our article shows how some issues can become more sectarian than scientific.

Turning to the substantive points raised, Marks and Melia suggest that we have misrepresented the results of a randomised controlled trial of sunscreens.³ In this study 113 adults living in Victoria, Australia, were randomised to either sunscreen or placebo over the course of an Australian summer. Concentrations of 25-hydroxy vitamin D and 1,25-hydroxy vitamin D were measured at the beginning and end of the study. The rise in 25-hydroxy vitamin D concentrations was similar in both groups, but the rise in 1,25-hydroxy vitamin D was lower in those allocated to receive sunscreen: a 1.5% increase from baseline versus a 13.3% increase (P=0.009). To call this "a slightly smaller change," as Marks does, is surely misleading. This study population was exposed to much stronger sunlight than that in temperate climates; furthermore, the sample studied probably experienced a higher than average sun exposure for Australians as they had a history of solar keratoses. Despite these characteristics the study provides evidence that sunscreens may affect vitamin D values. An extreme example of the danger of extrapolating from the Australian experience to less sunny climates is provided by the recent report of a case of rickets in a white child in Toronto who had been covered in sunscreen.⁴

Whiteman points out that we focused on a British population. Clearly, the balance of risks and benefits will differ between climates and populations, and the appropriate public health message will vary. Whiteman also suggests that because exposure to sunlight does not explain several population differences in disease risk it cannot be important. This observation ignores the multifactorial nature of the aetiology of melanoma, and many examplessuch as the high male smoking rate and very low coronary heart disease rate in Japanshow the misplaced reassurance that can follow from such comparisons. Selby and Mawer draw our attention to the evidence that exposure to sunlight may in fact reduce risk of some cancers. Although this evidence is tentative, it supports our contention that more sophisticated risk-benefit analyses are required in the formulation of public health policy.

Andrew R Ness, senior lecturer in epidemiology. 
Stephen J Frankel, professor of epidemiology and public health medicine. 
David J Gunnell, senior lecturer in epidemiology. 
George Davey Smith, professor of clinical epidemiology. 
Department of Social Medicine, University of Bristol, Bristol BS8 2PR