Tanning Devices – Fast Track to Skin Cancer?


* Address reprint requests to Antony R. Young, St John's Institute of Dermatology, Guy's, King's and St Thomas’ School of Medicine, King's College London, Lambeth Palace Road, London SE1 7EH, UK. E-mail: antony.r.young@kcl.ac.uk


The use of UVB and/or UVA emitting devices for cosmetic tanning is widespread in Western populations including young people and is especially prevalent in females. Several epidemiological studies, although not all, have shown a significant relationship between the use of tanning devices and malignant melanoma after, in some cases, adjustment for confounding factors such as solar ultraviolet radiation (UVR) exposure. A relationship between solar exposure, especially intermittent exposure, and malignant melanoma is well established so it is not surprising that a similar connection has been reported for the use of tanning devices. Several epidemiological studies show that childhood exposure to sunlight is a risk factor for malignant melanoma and this may also be the case for the use of tanning devices, especially if sunburns are obtained. Some studies have evaluated the relationship between the use of tanning devices and non-melanoma skin cancer and at least one has suggested an association. The use of tanning devices by a substantial minority of young people is a worrying trend in terms of a likely increased incidence of malignant melanoma, and possibly non-melanoma cancers in the future. Although two recent reviews by epidemiologists conclude that a clear link between tanning devices and malignant melanoma is yet to be proven, there is a strong case for effective legislation to prohibit the use of tanning devices by people under 18 yr of age.

Abbreviations –

basal cell carcinoma


cyclobutane pyrimidine dimers


minimal erythema dose


malignant melanoma


odds ratio


relative risk


squamous cell carcinoma


sun protection factor


solar simulating radiation


320–400 nm radiation


280–320 nm radiation


ultraviolet radiation


Skin cancer is an ongoing major global public health problem in white-skinned populations. There is widespread international consensus that solar ultraviolet radiation (UVR) is the main cause of skin cancer that includes basal cell carcinoma (BCC), squamous cell carcinoma (SCC) and malignant melanoma (MM) (1, 2). Malignant melanoma has a very much lower incidence than BCC and SCC, but is responsible for the vast majority of skin cancer deaths. Epidemiology has shown that SCC arises from cumulative solar radiation exposure whereas BCC and MM have been associated with intermittent patterns of exposure (1, 3, 4). Malignant melanoma has also been associated with sunburning episodes and childhood exposure to sunlight (1). It is widely supposed that the increase in skin cancer over the last few decades is the consequence of changes in behaviour that have resulted in increased exposure to solar UVR.

Mouse and in vitro studies have supported a relationship between solar UVR and non-melanoma skin cancer (1). More recently, molecular epidemiology has confirmed a direct link between solar UVB (c. 295–320 nm) and non-melanoma skin cancer (5). Mouse studies have also shown that UVA (320–400 nm) induces non-melanoma skin cancer but to a much lesser degree than UVB. Extrapolation of these data to humans, with some correction for the differences in the optical properties of the skin of the two species, would predict that UVB is very much more carcinogenic than UVA (6). Until recently, there was no good mammalian model for the study of MM, but recent genetically engineered mouse models show much promise for the study of the molecular photobiology of MM (2). It is also well established, from mouse, in vitro and human studies, that solar UVR causes photoageing (7).

The UVR is thought to result in skin cancer by two main mechanisms; (i) mutation to critical regulatory genes such as p53 and (ii) immunosuppression (8). Mutations arise from damage to epidermal DNA such as the formation of di-pyrimidine lesions caused by direct absorption of UVB and, to a lesser extent, of UVA. Studies have shown that sub-erythemal solar simulated radiation (SSR) induces cyclobutane pyrimidine dimers (CPD) and 6–4 photoproducts in human keratinocytes and melanocytes in vivo (9). Similarly, sub-erythemal SSR suppresses cell-mediated immunity in humans, especially in sun sensitive skin types I and II (10).

During the past decade or so, there has been an increasing trend, especially among young people, to use UVR emitting devices to achieve a tan. This trend has been fostered by the ‘tanning industry’ with annual global revenues estimated by the finance industry at US$2.6 billion mainly, and approximately equally, derived from North America and Europe (11). The popularity of this industry is doubtless due to the emphasis placed on a tan in an increasingly image-conscious society. Tanning devices come in a variety of forms and these have changed over the past three decades with improvements in technology. Typically, current devices are in the form of a sunbed with a battery of fluorescent tubes underneath a UVB-transmitting Perspex support with a top canopy with another battery of tubes. Some of these sunbeds have integral high-pressure facial tanning units whilst others have high-pressure lamps only. Exposure is either at commercial outlets, e.g. tanning salons, where there is the possibility of regulation, or by home-based devices that are not readily subject to regulation.


A major determinant of human skin colour is the presence and balance of different types of melanin in the epidermis. Baseline skin colour depends on the level of constitutive pigmentation that is low in white skinned peoples (skin types I–IV) and high in dark skinned peoples (skin types V and VI), as defined in Table 1. Tanning is the induction of facultative pigmentation by UVR in white-skinned peoples (skin types I, II, III and IV as defined in Table 1), although it also occurs in those with higher levels of constitutive pigmentation. Tanning is one of the main purposes of intentional sun exposure and several studies have shown that people, especially young people, sunbathe with the express purpose of obtaining a tan. Tanning is also the main reason that people use sunbeds. Only one study has examined the kinetics of sunbed tanning in following the Food and Drug Administration (FDA) recommended exposure schedule (12). A good and progressive tanning response, measured spectroscopically and by eye, was seen over an 8-week period with minimal erythema.

Table 1.  Characteristics of different skin phototypes
Skin typeConstitutive skin colourSensitivity to sunburnFacultative tanning abilitySkin cancer risk
IWhiteVery highVirtually nilHigh
IVOliveLowVery goodLow
VBrownVery lowVery goodVery low
VIBlackVery lowVery goodVery low

In the tanning process, melanocytes are stimulated to produce melanin that is transferred to keratinocytes. Constitutive pigmentation is widely regarded as photoprotective because darker skins are more sun-tolerant and have a much lower incidence of skin cancer than white skins. It is widely believed that facultative pigmentation is photoprotective but several studies have shown that the protection afforded by a tan against sunburn and DNA photodamage is equivalent to that of a sunscreen with a sun protection factor (SPF) of about 3 (13). There is evidence to suggest that tans induced by UVB and UVA are qualitatively different and that a UVA-induced tan provides almost no photoprotection (14). One study showed that tanning for at least 6 weeks with UVA sunbeds did not result in any changes in minimal erythema dose (MED), indicating that the tan was not photoprotective (15). Recent studies with SSR have shown that tanning in skin types II and IV is associated with an accumulation of epidermal DNA photodamage (16). Indeed, it has been hypothesized that tanning is the consequence of repair of DNA photodamage (17). Overall, it may be concluded that the benefits from tanning by sunbeds are primarily cosmetic and that this process is associated with genetic damage that is known to give rise to skin cancer. However, there has been recent emphasis on possible benefits, with respect to internal malignancies, derived from vitamin D induction by solar UVR (18).

Sayre and Dowdy (19) have drawn attention to important differences between tanning by solar UVR and from tanning devices. They argue that tanning under the controlled conditions of the tanning industry (12) may be safer than under the uncontrolled practice of outdoor sunbathing, in which the emission spectrum varies considerably during the day and in which sunburn may be more likely. This argument is valid in a case of ‘one or the other’ but less so when both approaches are used as indicated by a correlation between solar exposure and use of tanning devices in Swedish students (20) and in US youth (21) and beachgoers (22).

The importance of spectra

Emission Spectra and Irradiances

McGinley et al. (23) surveyed 100 sunbeds in current (1998) use in Scotland. Average sunbed UVB (280–315 nm) irradiances were similar to that of Glasgow on a sunny day. With the higher power sunbeds, UVB irradiances were similar to those at Mediterranean latitudes. Mean UVA irradiances were three to four times higher than that in Glasgow on a sunny day and up to twice as high as that for the Mediterranean. The authors also compared the amount of UVR received from 20 sessions on a sunbed with sunbathing in Glasgow for 10 days and in the Mediterranean for 1 week (without sunscreen use). They estimated that UVB exposure from the sunbed would be much less in both cases. On the other hand, UVA exposure would be similar to that obtained in Glasgow but less than that received in the Mediterranean.

Gerber et al. (24) made spectral measurements of sunbeds in Switzerland in 2002. The UVB (280–320 nm) irradiances from the sunbeds were in the region of, but mostly lower than, solar UVB measured in July at noon at 47°N (3.08 W/m2). However, UVA (320–400 nm) irradiances from sunbeds ranged from 100 to 538 W/m2 compared with that of 56 W/m2 in sunlight. When weighted with the CIE action spectrum for erythema (25), the effective solar irradiance is 0.21 W/m2 whereas the mean effective irradiance of the most common type of sunbeds is 0.33 W/m2; the latter equivalent to a global UV index of 13 compared with that of 8.5 for high summer noon at intermediate latitudes.

Action Spectra

The action spectrum for CPD induction in human skin (26) is very similar to those for human erythema and tanning, strongly implicating epidermal DNA as a major chromophore for both these endpoints. The action spectra for human erythema and SCC in the mouse are also very similar (6). Overall, these similarities strongly suggest that tanning by any part of the UVR spectrum has an approximately equal risk of non-melanoma skin cancer and there is no such thing as a ‘safe tan’. Some authors have argued that UVA may be relatively more important for MM than for SCC (27) but we lack action spectrum data for MM in a mammalian model although recent mouse studies suggest that this may be possible in the future (2) and there is an urgent need for such data.

Hazard Spectra

A hazard spectrum is the wavelength-by-wavelength product of an emission spectrum with an action spectrum for a given biological endpoint. For example, the c. 5% UVB content of solar UVR accounts for about 80% of its erythemal effect and a similar outcome would be predicted for SSC. One in vitro study showed that the 0.8% of UVB from a UVA sunbed accounted for 75% of its ability to cause CPD and 50% of its ability to cause oxidative damage to DNA (28). This means that the classification of a tanning device by its emission spectrum alone is not useful; action spectra for given endpoints must be taken into account. Hazard spectra as described assume no spectral interaction, i.e. that one waveband does not modulate another. However, mouse studies suggest that this may not be the case for immunosuppression by UVR because prior UVA exposure has been shown to abrogate the immunosuppressive effects of UVB (29).

Use of tanning devices

In 2002, Cokkinides et al. (21) assessed the use of sunbeds in the USA by young people (11–18 yr) and their parents or caregivers. The results, based on a population-based telephone survey of 1192 people, showed that 10% of the young and 8% of their primary carers had used sunbeds in the previous year. Independent predictors of sunbed use were (i) age 17–18 yr, (ii) being female, (iii) having a parent/carer who used sunbeds in the previous year, (iv) non-use of SPF ≥ 15 sunscreen by beach or pool and (v) low sun sensitivity. The conclusions on percentage use are supported in a much larger cross-sectional study, published in the same year, of 4147 boys and 5932 girls aged 12–18 yr from all 50 USA states (30). A total of 9.5% had used a sunbed in the previous year, with much greater use by the girls (OR = 7 95%; CI = 5.7–8.7) and with older girls (15–18) more likely to report use than younger girls (24.6% vs. 4.7%). A smaller study of 210 students, published in 2001 in Texas, reported that 18% had used a tanning bed in the previous 6 months (31). In 1993, Mawn et al. (32) assessed sunbed use in a wider age range group (mean = 33.6 yr ranging from 16–90 yr) in North Carolina and a vacation cruise ship. Of 476 respondents, 34% had used sunbeds, with 45% of these having used sunbeds in the previous 6 months.

Jerkegren et al. (20) reported on sunbed use in a group of 296 Swedish students in 1999. 17% of the females and 14% of the males used sunbeds at least once a month and 11% of the total had used a sunbed in the previous year. 14% of the females and 27% of the males said that they had never used a sunbed, which means that the majority of both sexes had used a sunbed. There was a no relationship between sunbed use and skin type but a highly significant correlation between sunbed use and sunbathing. Overall, in this group the knowledge of risk factors for MM was high. Boldeman et al. (33) compared sunbed use in Swedish adolescents in 1993 and 1999 and reported about a 50% fall in use over the study period and there was also a marked reduction of users who reported erythema. It is possible that these encouraging trends are the consequence of better public education. Overall these studies show that a very significant minority of younger populations use sunbeds.

Behavioural Aspects

Amir et al. (34) evaluated attitudes, beliefs and behaviour with respect to sunbed use in 470 (418F and 52M) healthcare workers in England in 2000. Almost half (46%) of the participants used sunbeds to some extent with 12% reporting frequent use and 78% reporting occasional use. The main reason (93%) for use was ‘to look better’ and ‘to feel healthy’ (78%). Of those who never used a sunbed, 84% agreed that ‘sunbed use might cause skin cancer’ whereas only 57% of frequent users agreed with this statement.

A recent 2002 study sponsored by the Dutch Cancer Society evaluated the behaviour of 349 adult sunbed users in The Netherlands (35). The study sample was not random as participation required the ownership of a sunbed at home. The study, done by telephone survey, showed important differences between peoples’ perceptions of safe sunbed behaviour and their actual behaviour. 94% claimed that their behaviour was safe but in reality only 37% used sunbeds in accordance with safety guidelines. Table 2 shows how the participants responded to positive and negative aspects of sunbed use. It is clear that this, certainly biased, group of users had an unrealistic assessments of risk and benefit.

Table 2.  General beliefs about sunbed use in 349 Dutch adults (35)
Positive aspects%Negative aspects%
Gives you a nice dark tan55Makes your skin dry out a lot27
Is healthy71Makes your skin age a lot16
Is relaxing95Increases your risk of21
Is good for your skin52 getting skin cancer 
Makes you look attractive93  
Makes you look healthy98  
Is pleasantly warm88  

Use of tanning devices and skin cancer

Biomarkers of Skin Cancer

Studies in the late 1980s showed that the use of tanning devices has an adverse effect on human immune function (36, 37). More recently (2000), Whitmore and Morison (38) reported that 10 full-body exposures over a 2-week period suppressed immunity as assessed by the induction and elicitation arms of the contact hypersensitivity response. These authors also studied the effect of 10 full-body tanning exposures in 11 volunteers and, not surprisingly, reported the presence of CPD and p53 protein expression in keratinocytes in vivo (39).

Skin Cancer

Most of the epidemiological studies have been done with melanoma patients using a case–control approach and Table 3 summarizes the results of 12 such studies. The essential results in Table 3 are the odds ratios (OR) which are given as crude (i.e. unadjusted) or adjusted for other possible risk factors, such as solar exposure, which may be confounding. Some of the results are given as relative risk (RR). Table 4 summarizes the results of the only prospective cohort study in which the results are given as RR. All results are given with 95% confidence intervals (CI). It is not possible to give all the data in a single table so additional relevant information is given in the text that follows. Table 3 excludes studies (mostly early) in which no measure of association is given. See Swerdlow and Weinstock (40) for a review of these, and later studies.

Table 3.  Summary of 12 epidemiological case–control studies to assess relationship between sunbed/artificial UVR use and malignant melanoma
AuthorsLocation, year publishedStudy typeCasesControlsOR (crude)OR (adust)Notes
  1. NG, not given. RR is where data are given as relative risk. The reader is advised to see the original publications for full details that are too complex to be included here.

Holman et al. (42)Australia, 1986Population based5115111.1 (0.6–1.8)NG9% of study group reported use
Swerdlow et al. (47)Scotland, 1988Hospital based1801204.1 (0.8–20.3)3.4 (0.6–20.3)OR given for duration of use; never vs. >1 yr. Adjusted for naevi, hair and eye colour, photoype and solar exspoure
Østerlind et al. (43)Denmark, 1988Population based4749260.7 (0.5–1.0)
NGNot clear if RR is crude or adjusted
MacKie et al. (44)Scotland, 1989Hospital based280280M 2.6 (0.9–7.3)
F 1.5 (0.8–2.9)
M 1.3 (0.2–7.9)
F 1.2 (0.5–3.0) RR
Adjusted RR for naevi, freckling, sunburn, tropical residence, phototype
Walter et al. (48)Ontario, Canada, 1990Population based583608M 1.88 (1.20–2.98)
F 1.45 (0.99–2.13)
Authors state about the same as crude ORAdjusted for age, naevi, phototype, socioeconomic status
Garbe et al. (45)Germany, 1993Hospital based8567051.0 (0.7–1.5)
1.5 (0.9–2.4)
Adjusted RR for naevi, hair colour, phototype and different centres
Autier et al. (50)Germany, France, Belgium, 1994Hospital based4204472.71 (1.06–7.78)2.12 (0.84–5.37)Adjusted OR for age, sex, hair colour and sunny holidays
Westerdahl et al. (51)Sweden, 1994Population based400640NG1.3 (0.9–1.8)Adjusted OR for sunburn history, hair colour, naevi and sunbathing history
Holly et al. (46)California, USA, 1995Population based452930NG0.94 (0.74–1.2)Women only, adjusted OR for standard risk factors
Chen et al. (53)Connecticut, USA, 1998Population based6245121.3 (0.97–1.74)1.13 (0.82–1.54)Adjusted OR for sex, age, phototype, sun exposure
Walter et al. (49)Ontario, Canada, 1999Population based583608NGTrunk 1.61 (1.13–2.31)
Non-trunk 1.52 (1.09–2.13)
Adjusted OR for age, sex and phototype
Westerdahl et al. (52)Sweden, 2000Population based5679131.6 (1.1–2.4)1.8 (1.2–2.7)Adjusted OR for naevi, skin type and number of sunburns
Table 4.  Summary of prospective cohort data from Veierød et al. (54) in which relative risk (RR) for multivariate models includes age, location, hair colour and indicators of solar exposure
Age period and tanning device useFrequencies number (%)Number of casesMultivariable RR (95% CI)
10–19 yr  P = 0.44
 Never84 182 (98)1521.00
 Rarely or ≥ once/month1665 (2)  41.52 (0.51–4.12)
20–29 yr  Ptrend = 0.006
 Never71 133 (80)1231.00
 Rarely11 618 (13) 191.11 (0.67–1.85)
 ≥ Once/month6391 (7) 182.58 (1.48–4.50)
30–39 yr  Ptrend = 0.19
 Never44 338 (50) 781.00
 Rarely28 383 (32) 510.93 (0.64–1.34)
 ≥ Once/month15 169 (17) 361.42 (0.93–2.16)
40–49 yr  Ptrend = 0.08
 Never17 345 (42) 271.00
 Rarely14 514 (35) 331.39 (0.82 –2.33)
 ≥ Once/month9550 (23) 221.67 (0.93–2.99)
Combined 10–39 yr  P = 0.04
 Never/rarely65 239 (82)1111.00
 ≥ Once/month14 377 (34) 341.55 (1.04–2.32)

Studies That Do Not Show A Relationship Between Use of Tanning Devices And Malignant Melanoma (1986–1995)

The early publications, in the mid- to late 1980s, of Gallagher et al. (41) (not in Table 3) on MMs and Holman et al. (42) for all melanomas (pre-invasive and invasive), in Western Canada and Australia, respectively, did not show any association with the use of tanning devices. Similarly, the Danish study of Østerlind et al. (43) showed no association. MacKie et al. (44) showed a slight increased risk in a Scottish population, but this risk was very modest compared with other risk factors. In 1993, Garbe et al. (45) showed no association between sunbed use and MM in a German population. These early studies were based on data available about 20 yr ago when it may be speculated that use of tanning devices was much less common than in more recent times. More recently, in a 1995 publication in women in the San Francisco Bay area, Holly et al. (46) did not show any associated between sunbed use and MM, superficial spreading melanoma or nodular melanoma.

Studies That Show A Possible Relationship Between Use of Tanning Devices And Malignant Melanoma (1988–2003)

In a 1988, hospital-based case–control study, Swerdlow et al. (47) reported a significantly increased risk of MM in Scottish patients that was associated with the use, and especially, duration of use, of tanning devices. In 1990, Walter et al. (48) used a population-based, case–control design in southern Ontario, Canada. Overall OR for ever having used a tanning device in males and females were 1.88 (1.20–2.98; P < 0.01) and 1.45 (0.99–21.3; P = 0.6), respectively. Age-adjusted rates of cumulative use showed a significant increased trend in risk associated with a longer duration of use in males (P < 0.01) and females (P = 0.04). In a more recent (1999) analysis of the same populations, Walter et al. (49) also reported a significant association between use of tanning devices and MM.

Autier et al. (50) performed a hospital-based case–control study in Germany, France and Belgium. Participants were asked whether exposure had been for tanning or other purposes, e.g. medical. An increased risk of MM with a crude OR of 2.71 (1.06–7.78) was observed with exposure for 10 hr or more, for tanning purposes only, starting before 1980. The authors also reported a relationship between ‘skin burn’ by tanning devices and MM when this was also associated with at least 10 hr of exposure. Exposure to the UVR sources for any purpose for 10 hr or more showed a crude OR of 4.47 (1.45–13.7). However, this rose to 8.97 (2.10–38.6) when exposure was for tanning purposes only. When adjusted for a variety of factors including holidays spent in sunny resorts, the latter OR was reduced to 7.35 (1.67–32.3) with P ≤ 0.001.

Westerdahl et al. (51, 52) have done population based, case–control studies in southern Sweden. Both studies, published in 1994 and 2000, support the hypothesis that use of tanning devices is a risk factor for MM. In the more recent study, the authors showed an adjusted OR of 1.8 (1.2–2.7; P = 0.05) for MM with regular use of tanning devices. The risk was associated with the frequency of use and the time over which they were used. A dose–response relationship was seen for up to a total of 250 sessions, and thereafter the OR was decreased, although this was not significant. Analyses were done within different age groups, with the highest adjusted OR of 8.1 (1.3–49.5) for regular use of tanning devices in people younger than 36 yr.

Finally, for the case–control studies, Chen et al. (53) reported an association between use of tanning devices and MM in 1998. The risk was particularly high with two or more different types of sunlamp with an adjusted OR of 3.46 (1.32–9.11; P = 0.001 for linear trend). There was also increased risk with home use but not for use on commercial premises.

The most recent study, summarized in Table 4, is prospective cohort study of 106 379 women in Norway and Sweden who were followed for an average of 8.1 yr during which 187 cases of MM were diagnosed (54). The results, after adjustment for measures of sun exposure, provide very strong evidence for a relationship between tanning device use and MM, especially when such exposure occurred between the ages of 20–29 yr in which Ptrend = 0.006 (see Table 4). Additional analyses for use over 10–39 yr suggests that tanning device use is a significant risk factor for MM (P = 0.04). Overall, this study provides the strongest evidence yet for a causal relationship between MM and the use of tanning devices.

Skin cancer mortality is mostly attributable to MM and Diffey (55) has estimated that the current use of tanning devices in the UK results in about 100/1312 (7.6%) deaths per year and he argues that this is modest in public health terms when compared with deaths due to other pleasurable activities such as smoking and drinking.

Studies on Non-Melanoma Skin Cancer

Only a few studies have been done on non-melanoma skin cancer. Two hospital-based case–control studies in Ireland, in the mid- to late 1980s, did not show any relationship between the use of tanning devices and non-melanoma skin cancer (56, 57). A similar conclusion, at about the same time, was reached by Bajdik et al. (58) in British Columbia, Canada, who evaluated 406 controls (population based) against 180 SCC cases and 226 BCC cases. About 10% of each group had ever used a sunlamp. The adjusted OR for BCC and SCC for ever having used a sunlamp were 1.2 (0.7–2.2) and 1.4 (0.7–2.7), respectively, which the authors state are non-significant. One 2002 study used the ‘generalized estimating equation method’ and reported no significant effect of tanning devices for BCC, although the total lifetime exposure to tanning devices was almost twice as high in patients compared with controls (59). In the same year, Karagas et al. (60) assessed the relationship between use of tanning devices and BCC and SCC in a population-based case–control study. In this study, there was greater use of tanning devices ranging from 9.2% (male controls) to 28.4% (female patients). The OR for BCC and SCC were 1.5 (1.1–2.1) and 2.5 (1.7–3.8), respectively and adjustment for a variety of factors made no difference to these results. The results of Karagas et al. (60) suggest that the use of tanning devices is a risk factor for non-melanoma cancer.

Summary of Epidemiological Studies and Future Directions

The more recent studies suggest that the use of tanning devices is a risk factor for MM. This is not surprising, given that the widespread use of tanning devices is a relatively recent phenomenon. Swerdlow and Weinstock (40) and Autier (61) who reviewed the available publications in 1998 and 2002, respectively, drew attention to methodological limitations and outcome contradictions that preclude definitive conclusions from the available studies. Autier (61) stated ‘there is still no conclusive evidence on the influence of sunbed use on melanoma occurrence’. These authors suggested new and more robust studies. This conclusion would also apply to non-melanoma skin cancer. However, some of the problems have been addressed by the most recent (2003) Scandinavian prospective study on MM (54) that provides further evidence that MM is associated with sunbed use.

Studies in Scandinavian countries (54) may also provide the opportunity to address the relationship between MM and emission spectra. The highest risk (RR = 2.58) in the group exposed to tanning devices between the ages 20 and 29 yr (in the 1970s) (see Table 4) may reflect an effect of the mercury–quartz lamps and -panels with a high UVB-emission. Such devices, known as type 4, were common in Sweden and Norway until they were banned from sale in the two countries in the early 1980s with the introduction of legislation on tanning appliances (see section below on regulation). One of the reasons for the ban was to reduce the potential for burns. Unfortunately, the role of sunbed burn was not assessed in the recent study (54) or in other Scandinavian studies (51, 52).

The need for regulation?

Aspects of regulation include the device per se and the use of the device. The European product safety standard for sunbeds EN60335–2-27 (1997) from CENELEC (62) is based on a corresponding International Electrotechnical Commission (IEC)-standard IEC60335–2-27 (1995) (63). This categorizes sunbeds into four different ‘UV-types’ depending on their UV-irradiance level and spectral composition. Types 1–3 are for cosmetic use and type 4, with potentially very high levels of UVB, is for use under medical advice; a situation that has resulted in confusion with respect to modern tanning equipment (64). Types 1 and 2, characterized by extremely high levels of UVA-radiation with no or limited UVB, are ‘intended to be used under supervision of appropriately trained persons’ whereas type 3 has limited irradiance levels both over and under 320 nm with a total CIE-weighted erythemal irradiance equivalent to or less than tropical sun (UV-index = 12). The standard notes that type 3 ‘may be used by unskilled persons’. The US Food and Drug Administration (FDA) has a performance standard for the manufacture of sunlamps (65) but this does not limit the irradiance of such devices. Indoor tanning salon practices are regulated at the state level with approximately about 50% of states having any regulations.

Dellavalle et al. (66) have evaluated the laws as of 2003 that govern youth access to tanning parlors in five English speaking countries (Australia, Canada, New Zealand, UK and US) and in France. The study showed that France has prohibited those under 18 from using tanning facilities since 1997, and that only limited regions of Canada and the US have laws designed to prevent youth access. In the case of the US, Wisconsin, Illinois and Texas prohibit tanning parlor use by those under 16, 14 and 13 yr, respectively. A minority of other states have regulations that allow use with the consent or accompaniment of the guardian. These results show a clear case for national laws to protect minors from using commercial tanning facilities (67) and the International Commission on Non-Ionising Radiation Protection (ICNIRP) has recently recommended that people under 18 yr of age should not use tanning devices (68).

The Swedish government, via the Swedish Radiation Protection Authority (SSI), introduced laws on sunbeds in 1982 (later followed by Norway and Finland) allowing only the sale and commercial use of devices with UVB- and UVA-irradiance limits, later redefined as ‘UV-type 3’. Legislation in Sweden since 1998 (SSI FS 1998:2) (69) also stipulates that information on the risks associated with sunbed use is mandatory and must be posted at sunbeds in tanning salons, etc. Young people under 18 yr are advised not to use sunbeds, although this is not mandatory. The device manufacturer's user information and exposure schedule must also be available. Spain in 2002 introduced legislation with an age-limit (18 yr). It also sets a limit to the irradiance, equivalent to UV-type 3, but with an additional requirement that there must be no radiation with wavelengths lower than 295 nm (Spanish Royal Decree 1002/2002 of 27 September 2002. ‘Sale and use of tanning equipment using ultraviolet radiation’). Diffey (55) has argued that there is no need for outright probation because deaths caused by tanning devices are low compared with other forms of self-inflicted early mortality and that exposure to sunlight, that cannot be regulated, is the major cause of skin cancer.


A significant number of white skinned people use tanning devices, especially younger people and females with the desire for a tan being one of the main reasons for use. These devices contain UVR wavebands that are known to be carcinogenic in mice and humans and users often have a high level of awareness of the relationship between UVR and skin cancer. Some of the studies referred to in Tables 3 and 4, especially the more recent ones show a relationship between the use of tanning devices and MM, even after adjustment for confounding by solar UVR exposure and other risk factors. However, based on current evidence, it is still difficult to categorically state that sunbed use results in MM, although the evidence is increasingly, and perhaps not surprisingly, supporting this view. One study suggests a possible relationship between the use of tanning devices and BCC and SCC. The high current high level of use of tanning devices, especially by young people, raises concerns about future increases of MM and, possibly, non-melanoma skin cancer. Effective laws could limit the use of sunbeds by young people at commercial facilities and an expert panel of the ICNIRP has recommended this.


Acknowledgements– I thank Andy Pearson of the National Radiological Board (NRPB) UK and Ulf Wester of the Swedish Radiation Protection Authority and Sharon Miller of the US Food and Drug Administration (FDA) for their helpful input.