Photoprotection and prevention of melanoma

ARTICLE

Melanoma incidence and mortality rates are rising in most countries where they have been recorded in recent decades. As a result of this, a number of public education programs have been initiated to deal with the problem.

The first approach in most countries has been an early detection program to ensure that those people with melanoma now have it detected and treated at an early curable stage. However, there are data which suggest that melanoma is induced in people constitutionally at risk by excessive exposure to sunlight. On this basis a primary prevention approach to melanoma is possible and many people are now considering or initiating such programs with the aim of preventing melanoma in the long term.

This article briefly reviews some of the evidence relating sunlight exposure to risk of melanoma and then looks at methods of photoprotection which might be of value in reducing the risk in the long term.

The epidemiology of sunlight and melanoma

Although sunlight exposure was first reported to be a potential risk factor for non-melanoma skin cancer (NMSC) in the 1890's, it was not until the 1950's in Australia that it was suggested that it may also be important in the development of melanoma [1, 2]. Since then, there have been an increasing number of epidemiology studies looking at the association between exposure to sunlight and risk of melanoma. Epidemiology studies have formed the basis of our knowledge in this area because, at present, we have no animal model or in vitro system which can replicate with confidence what we believe is happening in humans.

Constitutional factors

There are two components to the development of melanoma. The first is the constitutional or genetic predisposition. This includes having fair skin; having a tendency to burn when exposed unprotected to strong sunlight, rather than develop a tan; having a tendency to develop multiple melanocytic naevi, both common acquired and dysplastic naevi; and having a tendency to develop freckles [3]. There is increasing evidence that sunlight exposure contributes to the development of melanocytic naevi in those who have a constitutional predisposition to do so [4]. There are several clearly genetically predetermined syndromes associated with a very high risk of melanoma, including xeroderma pigmentosum and the familial atypical mole syndrome [5]. As yet, the exact nature of the genetic abnormality predisposing to not only these strongly inherited syndromes, but also the genotype for melanoma risk in the general population has not been clearly established.

Environmental factors

There is a vast volume of epidemiological literature looking at exposure patterns to sunlight and risk of melanoma. Most data suggest that episodic exposures to large doses of sunlight, particularly in childhood, and particularly in doses sufficient to induce sunburn that can be recalled many years later, are associated with increased risk of melanoma in adulthood [6]. Heavy exposures particularly during leisure activities are associated with the risk. Ironically, frequent heavy exposures that might occur as a result of outdoor work seems to be associated with a lower risk of melanoma than infrequent heavy exposures [7]. Thus there are data showing that the melanoma risk is higher in indoor workers than outdoor workers [8]. These are studies on adult exposures thus indicating that it is not only childhood exposure to sunlight which is a risk factor for melanoma.

There are migration studies indicating that the risk of melanoma is lower in people who migrate to hot, sunny climates after their childhood and adolescence than those who were born in that climate and thus were exposed during their early growth and development [9]. Similarly, even within countries with a hot, sunny climate, like Australia, there are latitudinal differences in melanoma incidence consistent with latitudinal differences in sunlight exposure. There are some confounders in those latitudinal studies in that even though the ambient UVR levels are higher at the lower latitudes, the temperature tends to be warmer and the days longer. This allows the population in those areas to be exposed to not only stronger radiation, but also for longer periods and wearing less clothing.

There are some data also that suggest that chronic heavy exposures can contribute to the risk of melanoma. For example, actinic keratoses are a risk factor for melanoma [10]. A study looking at the frequency of melanoma by site relative to total body surface area has shown that when correlated in this way, melanoma frequency is most common in the heavily light-exposed areas. However, overall, melanoma tends to be more frequent in the areas of skin less heavily exposed, e.g. the upper back in men and the lower leg in women.

Most of the data mentioned above relate to the superficial spreading melanoma. The data related to lentigo maligna are slightly different being consistent more with very heavy repeated exposures over many years. These tumours occur most frequently on the sites of maximum exposure as well.

The exact wavelength of radiation in the solar spectrum which induces melanoma is unknown. Nevertheless, what limited data we have suggest that it is the UVR spectrum, particularly UVB and possibly UVA. It is the UVB spectrum which is most biologically active in inducing sunburn. There are some studies published suggesting that the use of UVA sunbeds may increase the risk of melanoma in those countries where the ambient radiation levels are relatively low, but methodological problems make it difficult for these studies to be definite about the role of UVA [11].

Using the latitudinal gradient in melanoma incidence and comparing this with the ambient erythemal UVR at various latitudes, UV dose-incidence models have been developed to indicate the risk of melanoma for increasing doses of radiation [12]. Most of the models suggest a power relationship, i.e. for every 1% increase in UVR received over a lifetime there is between 1-2% increase in the incidence of melanoma.

Despite there having been a large number of epidemiological studies we still cannot be entirely sure of the nature of the exposures, the age at which it is important and how it is received that causes melanoma. In fact, there are almost conflicting data at times, when studies have relied on memory of past exposures, clothing worn, degree of sunburn, etc going back for many decades.

The conflicting or confusing data that have occurred on occasion are probably more a reflection on the epidemiological methods, which despite using increasingly sophisticated analyses are still relying on data which are basically corrupt, i.e. memory of events going back up to 50 years or more [13]. It is probably not until we get more specific biological markers for proving previous exposures that we will have much further advancement in our knowledge of the relationship between UVR exposure and the development of melanoma in those people at risk.

Sunlight and other sources of ultraviolet radiation (UVR)

Natural UVR

The sunlight received at the earth's surface from solar emission is a wide range commencing in the UV spectrum at around 290 nm and running through visible light to infrared radiation at around 2,500 nm. The relative amounts of UVB (290-320 nm) and UVA (320-400 nm) are determined by many variables. These include:

Time of day: the proportion of UVB is higher around the middle of the day when the sun is at its solar zenith and thus there is a shorter distance of atmosphere to absorb UVB.

Time of year: the proportion of UVB is higher during the summer months when the sun is higher in the sky and thus less UVB is absorbed in its passage through a shorter distance of atmosphere.

Latitude: the proportion of UVB is higher at lower latitudes when there is less absorption of UVB due to more direct passage of solar radiation through the atmosphere.

Altitude: there is less absorption of UVB because of shorter passage through the atmosphere. There is approximately a 15% increase in erythemal UV for every 1,000 metres above sea level.

Cloud cover: cloud cover or water vapour differentially absorbs more in the infrared spectrum than in the UV spectrum. Thus with cloud cover there is often a reduced temperature at a time when UV levels have not altered to a large degree. In general, moderately dense cloud cover is required to substantially reduce UV levels.

Reflection: There is different reflection from different surfaces. In general, hard shiny light surfaces tend to reflect more radiation than soft darker surfaces. Water is misleading. With direct radiation vertically above a still water surface, there is relatively little reflection. However, that is not what happens in real life. With turbulent water, there is a very low angle of incidence for the radiation and substantial UVR reflection can occur under these circumstances.

Particulate matter: smog and other particulate matter, such as atmospheric smoke, act as reflectants and absorbants for UVR and thus reduce the terrestial dose received in these conditions.

Stratospheric ozone: stratospheric ozone absorbs the solar emission of radiation predominantly in the UVC and UVB ranges. It does not absorb UVA. There has been increasing concern about stratospheric ozone depletion occurring, particularly at the higher latitudes in the last decade. The estimates of reduction indicate that there may be increases in UVB but not UVC. As yet, nobody has been able to demonstrate a sustained increase in terrestial UVB levels as a result of the ozone depletion. Nevertheless, increasingly stringent controls on the manufacture, distribution, use and release of ozone depleting substances throughout the developed world has led to decreasing atmospheric concentrations of these substances at a rate faster than predicted in some of the original estimates. Therefore, it is likely that the peak damage to the stratosphere and consequent ozone depletion will occur by the end of this century and will improve after that. The relative contribution ot this depletion to the incidence rate of melanoma seen now and in the future is debatable.

All of the above factors combine to contribute to the total dose of UVR and the relative fractions of UVA and UVB available at any time and place. In practice, it is extremely difficult for an individual to estimate what is the ambient UV level and thus estimate their risk of receiving a large dose and subsequent sunburn, i.e. estimate their "sunburn time".

Other sources of UVR

Other sources of UVR include arc welding and quartz halogen sources of lighting. There is an inverse square law of dose being inversely proportional to the square of the distance from the source. This means in practice that the dose of UVR received from quartz halogen lighting is relatively small as the light source tends to be a reasonable distance away. Nevertheless, where there is concern about this form of lighting, the use of a perspex filter in front of the globe is a very effective filter of UVR in both the UVB and UVA range.

Another source of UVR is the industrial process of UV cured ink used in printing. Despite there being potential for moderate doses of UVR to be received from these artificial sources, in practice, under most circumstances their contribution to risk of melanoma is probably relatively small compared with the dose of UVR received from sunlight exposure.

Photoprotection in the prevention of melanoma

It is critical to take into account the circumstances in which past exposures occurred that have led to the melanomas that we are seeing today. In this regard, it was exposure to sunlight, the whole of the terrestial spectrum and not simply UVR, which has been associated with the population-based rises in melanoma incidence seen during this century.

There has been a clear correlation between the rising incidence of melanoma and a change in attitudes and behaviour of the population at risk. That is, it has accompanied the rise in the fashion amongst fair people to change the colour of their skin, i.e. seek a suntan. It has also followed the change in community attitudes to exposure of a large proportion of the body in public open spaces. Those seeking to change their skin colour have been able to do so by taking off their clothes and remaining in the sun for long periods in areas such as the beach or public open spaces during the sunny periods of the year.

Under these circumstances, until we know exactly what spectrum of radiation, what type of exposure and what dose over what age groups is necessary to induce melanoma, the first step in photoprotection should be to reduce exposure to the whole of sunlight, i.e. reverse the previous conditions which have led to the tumours we see now. This is an important concept as a variety of methods of photoprotection, particularly sunscreens, filter out only a proportion of the solar spectrum, particularly in the short wavelength UVR range. Consequently, sunscreens and other methods which filter out only selected wavelengths of sunlight become the second line approach to photoprotection, not the primary one. Barriers, which reflect the whole of the sunlight spectrum, become the primary approach.

The natural approach

Midday sun

The natural approach to photoprotection involves the use of physical barriers to reflect sunlight. This is particularly important at the time of day when sunlight is strongest. The radiation curve shows that around 60% of UVB is received in the two hours either side of the solar zenith (10.00 am-2.00 pm or 11.00 am-3.00 pm during daylight saving time) (Fig. 1). Avoidance of outdoor activities during these periods, if possible, is an important part of the natural approach.

Shade

The use of shade whilst outdoors is also an important component, particularly if it is necessary to be out during the times of peak UVR. Shade can be created naturally in the form of canopies of trees, or it can be created with constructed canopies. The degree of photoprotection afforded by a canopy is a function of not only the material lining the canopy, but also the height of the canopy off the ground. The higher the canopy, the more likely that there will be reflected radiation not only from objects around the canopy, but also from both clear sky and clouds. The atmosphere on a clear sky day reflects and scatters radiation to a certain extent. Cloud certainly scatters radiation, both visible and UVR, so that it is possible to receive a moderate dose of UVR even when sitting below a canopy if the cloud is able to be seen from that position.

Clothing

Wearing clothing is a very effective way of creating a personal shade zone. Clothing is a physical barrier. In contrast to sunscreens, clothing can be seen and it is easy to determine what areas of the body are covered. The degree of protection afforded by clothing is a function mainly of the density of weave of the material, rather than the fibre type [14].

The colour and whether or not the cloth is wet make a small difference to the radiation absorption produced by cloth, particularly if there is a light weave. In the presence of a dense weave, colour makes little, if any, difference. Even with relatively light weaves, the difference in radiation absorption related to different colours (dark colours absorb better than light colours) is only a proportional one, but not a large one in terms of the absolute dose of radiation received [15].

Wetting a light cloth may increase transmission of radiation through it, particularly if the cloth becomes heavy with the moisture and stretches, thus decreasing the density of weave. The penetration of radiation through cloth has been shown to decrease with multiple washings. Under these circumstances, shrinkage of the material has led to an increase in weave density.

Careful design, creating attractive garments that are easy to wear and also allow air movement between the garment and the skin are important in compliance with the wearing of clothing whilst outdoors in hot weather. If there is satisfactory air movement with loose clothing, it is actually cooler to wear clothing when outdoors than to expose the skin to the heat of the sun which is transmitted in the visible and infrared spectrum. UVR does not transmit heat and is not able to be felt. Thus, stratospheric ozone depletion cannot be detected by feeling that the sun is hotter.

Hats

The brim width of a hat determines the amount of radiation to the head and neck (Fig. 2). Unfortunately, the width of the brim also determines the likelihood of the hat being blown off if there is wind. The use of Legionnaire style flaps at the side and back of peaked caps may be a more satisfactory solution in windy conditions.

UV monitors

There are a variety of commercial products designed to measure ambient UVR levels in an attempt to assist people in deciding when it is safe, or otherwise, to be outdoors. Virtually all of these rely on photosensitive patches, some of which give an actual reading while others change colour. These devices are extremely misleading. Firstly, if placed on one part of the body, they do not indicate the dose of radiation received in other parts of the body at different angles and with different exposures. Secondly, calibration with actual levels of radiation may be in doubt. Thirdly, they almost always rely on estimates of burn time to set what is called a "safe limit".

As mentioned above, there are a large number of variables which determine the dose of radiation received at any time or place. Therefore it is almost impossible to estimate an individual's "burn time" with any degree of accuracy. Finally, as mentioned above, we still do not know what is a "safe" radiation dose in regard to risk of melanoma. Therefore, these devices are not recommended as part of the approach to photoprotection in reducing the risk of melanoma.

Sunscreens

Sunscreen active chemicals

Sunscreens are promoted and widely used throughout the world as a form of photoprotection. The sunscreen active chemicals contained in these products are divided into physical reflectants and chemical absorbers. The latter tend to absorb predominantly in the UVB range, but there are some which have activity also in the UVA range (Table I). The physical reflectant containing sunscreens tend to offer protection across a broader spectrum of radiation including not only UVR, but also the visible and infrared spectrum.

The physical reflectants also tend to be particulate matter which is able to be seen on the skin. They tend to be denser, making their use less acceptable to some people. On the other hand, the sunscreen absorbent chemicals tend to be transparent and easily applied to skin in a variety of bases including creams, lotions, waxes for lips, propellant sprays, roll-on sticks and other creative approaches.

Sun protection factor (SPF)

Sunscreens are graded according to their ability to reduce erythemal UVR, which is predominantly UVB. This is the sun protection factor (SPF) grading of sunscreens. On the other hand, because such large doses of UVA are required to induce erythema, a variety of other methods have been attempted to grade the efficacy of these products within the UVA spectrum, including both in vivo and in vitro thin film absorption methods. As yet, there is no internationally acceptable standard for assessing UVA efficacy.

It must be remembered that the SPF grading of a sunscreen is a laboratory measure. It was devised as a measure of relative effectiveness under strictly controlled conditions to assess one product compared with another. Under the laboratory conditions, the dose of UVR from a solar simulator required to produced erythema in unprotected skin is compared with the dose required with use of the sunscreen product. The products are applied in carefully measured amounts per area of skin. The proportional increase in dose of UVR required with the sunscreen is the SPF grading (e.g. a product which requires twice the dose is an SPF2; a product which requires five times the dose is SPF5).

Under these conditions, there is an exponential increase in radiation absorption which leads to a decreased benefit for very large increases in SPF after it reaches around 10
(Fig. 3). Most public health organisations throughout the world have recommended the use of a product with an SPF of at least 15 or more.

There has been considerable controversy in recent years on which particular SPF number should be recommended. In fact, data show that it is not the SPF grading of a sunscreen which is likely to be so important, but its use under practical everyday conditions [16]. Sunscreen is frequently applied in amounts less than those used in the laboratory testing and this will reduce its efficacy [17].

Application

Not only is the amount frequently less than adequate, but the areas covered may be inadequate. Because the sunscreens are often transparent, it is not easy to see whether or not there has been adequate coverage. Frequent rubbing, touching or sweating of the skin may remove the product even when it has been applied correctly. The use of water-resistant bases is an attempt to overcome the problem of removal with sweating or with swimming. Despite this, for all the reasons given above, regular re-application of sunscreens is recommended if people are going to be outdoors for prolonged periods. Application of sunscreen is also recommended up to 30 min before going outdoors, not because the product needs that time to commence working, but to allow adequate dispersal and binding in the skin. People tend to underestimate the time that they are outdoors before they finally decide to apply sunscreen, therefore application before going outdoors is important.

Side-effects

Side-effects from the use of sunscreens are common. An irritant inflammatory reaction is the most common side-effect reported with frequent and prolonged use [18]. Allergic contact dermatitis to sunscreen chemicals is rare. Photoallergic contact dermatitis has been reported with the perfumes and preservatives in the product being more common offenders. Contact urticaria, acneiform eruptions and a variety of other rarer effects have been described.

Sunscreens causing melanoma?

Of more concern are a number of recent epidemiological studies which have reported that sunscreen use itself may be a risk factor for melanoma. The data in these case-controlled studies show only that people with melanoma are more likely to say that they have used sunscreens regularly than the controls without melanoma [19]. It is the interpretation of these data that takes the next step of actually attributing causation of the melanoma to sunscreen use.

There are a variety of confounders possible to explain this response to the question about sunscreen use by people with melanoma. These include recall bias by those with melanoma; more tendency to use sunscreens by people at constitutional risk of melanoma, i.e. people who burn easily; inadequate use of the product by people who said they were using them giving them a false sense of security and inducing them to stay outdoors longer and actually getting a higher dose of radiation than those who do not use sunscreens.

Once again, these studies may reflect the inadequacy of the method used (particularly recall in detail of past behaviours), than a true causative effect of sunscreens in inducing melanoma. Nevertheless, further work is necessary in this area.

What are the chances of reducing melanoma with photoprotection?

As stated above, we do not as yet know exactly what dose of sunlight in susceptible people, nor the way it is received and the age at which it is received, which is necessary to induce a melanoma. There have been some epidemiological studies that show that regular frequent exposure to sunlight is associated with a lower relative risk of melanoma compared with having only occasional exposure. On this basis, it has been postulated that reduction of exposure to sunlight, particularly only an intermediate reduction, may actually increase the risk of melanoma by changing someone from a regular frequent exposure pattern to an episodic exposure pattern. Does this mean that population-based recommendations as part of a public health program could actually increase the risk of melanoma for a community?

Good quality epidemiological studies on migrants who arrive in hot sunny climates after their childhood have clearly demonstrated that protection during childhood affords substantial lifetime protection for melanoma. Therefore, it seems reasonable to recommend substantial protection during childhood and adolescence, if possible, to lower the threshold over a lifetime, even if an indiscretion occurs in subsequent years and a high exposure leads to an episode of sunburn. Without the initiating or priming doses during childhood, it could be that a subsequent exposure is unlikely to have such a critical effect in promotion of the disease.

There are no randomised prospective studies on sunscreen use to indicate that these products can actually reduce risk of melanoma. Nevertheless, they have been shown in such studies to prevent the development of actinic keratoses, and also lead to remission of existing ones [20]. Although these tumours are a different cell line, they do indicate that photoprotection at least has the potential to reduce the risk of carcinogenesis in sunlight-induced skin tumours. Similar studies looking at the value of regular sunscreen use in the prevention of melanocytic naevi in children are currently underway and may help to further our understanding of the value of photoprotection in the prevention of melanoma.

There are animal data in NMSC which show that reducing the dose of radiation to which the skin is exposed reduces not only the number of tumours that occur, but also lengthens the time before they occur [21]. The dose-response curve shows an almost inverse relationship. In other words, halving the dose of radiation doubles the time before the onset of the tumour. Time is an important factor in determining likely development of a tumour, presumably for the requirement of expose to sufficient tumour promoting factors along the carcinogenesis pathway.

In most countries, the average age of onset of melanoma is around 50 years. If the radiation biology related to melanoma is similar to that related to other cutaneous carcinogenesis, then halving the dose of radiation to which individuals are exposed (that is, correctly using the equivalent of an SPF2 sunscreen during a lifetime) would double to the time before onset of the tumour, i.e. to 100 years. Avoiding the sun around the middle of the day could reduce radiation by 60%; wearing good quality clothes with a reasonably tight weave would reduce radiation by at least 95%; using an SPF15 sunscreen would reduce radiation by at least 93%. In other words, a combination of one or more of these factors would substantially reduce radiation and, presumably, prolong the development of a tumour well beyond the normal life span of human beings at the moment.

These comments are theory. What has happened in practice? In Australia, the country with the highest incidence rates of melanoma in the world, substantial primary prevention programs promoting photoprotection as a way of preventing this disease have been undertaken for several decades [22]. It would not be expected that the age-adjusted incidence rate for the whole population would as yet be levelling off or reducing as a result of those programs.

Nevertheless, cohort analysis on recent incidence data reveals that those in the younger cohorts have a levelling off in the incidence of melanoma compared to older cohorts in the population [23]. The younger cohorts are those which behavioural research data show have been influenced by the primary prevention public health programs. It is also expected that the effect of a primary prevention program would be seen initially in the younger cohorts at risk, rather than the older cohorts. The latter may have had substantial exposure both in childhood and adulthood which is less likely to be influenced by the early interventions.

Thus, in the end we do have some population-based human data to indicate that the promotion of photoprotection as a way of reducing risk of melanoma may be of value in the long term.

REFERENCES

1. Unna PG. Die Histopathologie der Hautkrankheiten. Berlin: Verlag Von August Hirschwald, 1894: 719-24.

2. McGovern VJ. Melanoblastoma. Med J Aust 1952; I (suppl.): 139-43.

3. Evans RD, Kopf AW, Lew RA, et al. Risk factors for the development of malignant melanoma. I: Review of case-control studies. J Dermatol Surg Oncol 1988; 14: 393-408.

4. Kelly JW, Rivers JK, Mac Lennan R, et al. Sunlight: a major factor associated with the development of melanocytic nevi in Australian school children. J Am Acad Dermatol 1994; 30: 40-8.

5. Green MH, Clark WH Jr, Tucker MA, et al. High risk of malignant melanoma in melanoma-prone families with dysplastic nevi. Ann Intern Med 1985; 102: 458-65.

6. Marks R, Whiteman D. Sunburn and melanoma: How strong is the evidence. Br Med J 1994; 308: 75-6.

7. Cooke KR, Skegg DC, Fraser J. Socio-economic status, indoor and outdoor work, and malignant melanoma. Int J Cancer 1984; 34: 57-62.

8. Armstrong BK. Epidemiology of malignant melanoma: Intermittent or total accumulated exposure to sun? J Dermatol Surg Oncol 1988; 14: 835-49.

9. Khlat M, Vail A, Parkin M, Green A. Mortality from melanoma in migrants to Australia: variation by age at arrival and duration of stay. Am J Epidemiol 1992; 135: 1130-13.

10. Green AC, O'Rourke MG. Cutaneous malignant melanoma in association with other skin cancers. J Natl Cancer Inst 1985; 74: 977-80.

11. Swerdlow AJ, Weinstock MA. Do tanning lamps cause melanoma? An epidemiologic assessment. J Am Acad Dermatol 1998; 38: 89-98.

12. Elwood JM. The epidemiology of melanoma; Its relationship to ultraviolet radiation and ozone depletion. Trans Menzies Foundation 1989; 15: 95-107.

13. Diffey BL, Gies HP. The confounding influence of sun exposure in melanoma. Lancet 1998; 351: 1101-2.

14. Welch C, Diffey B. The protection against solar actinic radiation afforded by common clothing fabrics. Clin Exp Dermatol 1981; 6: 577-82.

15. Gies HP, Roy CR, Elliott G, Zangli W. Ultraviolet radiation protection factors for clothing. Health Phys 1994; 67: 131-9.

16. Loesch H, Kaplan DL. Pitfalls in sunscreen application. Arch Dermatol 1994; 130: 665-6.

17. Sayre RM, Powell J, Rheins LA. Product application technique alters the sun protection factor. Photodermatol Photoimmunol Photomed 1991; 8: 222-4.

18. Foley P, Nixon R, Marks R, Frowen K, Thompson S. The frequency of reactions to sunscreens: Results of a longitudinal population-based study on the regular use of sunscreens in Australia. Br J Dermatol 1993; 128: 512-8.

19. Westerdahl J, Olsson H, Masback A, Inguar C, Jonsson N. Is the use of sunscreens a risk factor for malignant melanoma? Melanoma Research 1995; 5: 59-65.

20. Thompson SC, Jolley D, Marks R. Reduction of solar keratoses by regular sunscreen use. N Engl J Med 1993; 329: 1147-51.

21. De Grujil FR, Van der Meer JB, Van der Heun. Dose-time dependency of tumour formation by chronic UV exposure. Photochem Photobiol 1983; 37: 53-62.

22. Hill D, White V, Marks R, Borland R. Changes in sun-related attitudes and reduced sunburn prevalence in a population at high risk of melanoma. Europ J Cancer Prev 1993; 2: 447-56.

23. Giles G, Thursfield VJ. Trends in skin cancer in Australia. Cancer Forum 1996; 20: 188-91.