Abstract
Objective
To estimate the separate effects of sunlamp and tanning bed device use on melanoma risk.
Methods
Population-based case–control study of 423 cases of melanoma and 678 controls in the state of New Hampshire. Exposure data, including sunlamp and tanning bed use, were collected by telephone interview. Associations were evaluated using logistic regression analyses.
Results
About 17% of participants ever used a sunlamp, and most use (89%) occurred before 1980. The OR was 1.39 (95% CI 1.00–1.96) for ever using a sunlamp, 1.23 (95% CI 0.81–1.88) for those starting sunlamp use <20 years, and 1.71 (95% CI 1.00–2.92) for those starting ≥20 years. Data suggested increasing risk with number of sunlamp uses and with duration of use (tests of trend p = 0.02). The overall prevalence of tanning bed use was 22% and most use (83%) occurred after 1980. The OR was 1.14 (95% CI 0.80–1.61) for ever using a tanning bed; there was no evidence that risk increased with frequency or duration of use. The OR was 1.96 (95% CI 1.06–3.61) for having used both devices.
Conclusion
Results suggest a modest association between sunlamp use and melanoma risk, and increasing risk with greater frequency and duration of use. No association with tanning bed use was found, but sufficient lag time may not have elapsed to assess a potential effect.
Keywords: Melanoma, Skin cancer, Sunlamp, Tanning bed, Artificial UV
Introduction
Over the last 25 years, cutaneous melanoma has become an increasingly common cause of cancer morbidity and mortality in Caucasian populations worldwide [1–3]. The American Cancer Society estimates that melanoma will account for approximately 59,940 cases and 8,110 deaths in the United States during 2007 [4, 5]. The etiology of cutaneous melanoma is not completely understood, but evidence from numerous studies indicates that host factors, including hair and eye color, nevi, and sensitivity to the sun are related to risk [1, 6–12].
Although much remains to be learned about the type, periodicity, and timing of sun exposure, UV radiation is clearly the predominant environmental, and thus potentially modifiable risk factor for melanoma [13]. Historically, UVB has been thought to account for most of the risk [14–18]. UVB varies considerably in intensity with greatest emission near the equator, and the least near the poles, a pattern corresponding roughly to variation in melanoma incidence [19, 20]. UVA varies much less significantly with latitude. A role for UVA is indicated by studies showing melanomas arising in those who receive UVA for treatment of dermatologic conditions [21–23] as well as by basic scientific evidence of the harmful effects of UVA on DNA, cells, and animals [18].
The influence of artificial sources of UV on risk of melanoma has been examined by numerous studies (for example [24–28]) although findings to date have been heterogeneous, with much of the inconsistency attributed to small sample sizes with low statistical power to detect a plausible association [7, 25–27, 29–39]. Moreover, most studies to date have not been able to distinguish exposures by device type and many did not control for known melanoma risk factors such as pigmentary characteristics and solar exposures [29, 37, 40]. Despite the possibility of increased skin cancer risk, the use of tanning devices is popular among youth, [41, 42] and the demand for tans has produced an exponential growth of tanning facilities in recent years [43, 44]. According to the media, more than a million people daily use tanning beds in the US, [45, 46] and it is estimated that there are 28 million users of tanning facilities and 50,000 tanning facilities nationwide.[43] Modern tanning beds (post-1980) use fluorescent bulbs that emit mostly UVA and smaller than past UVB doses [29, 47, 48]. They deliver up to three times the UVA dose of that delivered by natural sunlight and a UVB dose rate that can approach the rate of bright sunlight [44].
In the modern era, when public health campaigns advocate minimizing or avoiding sun exposure, [5, 49–55] the potential influence of artificial UV sources on melanoma risk is worrisome. The objective of this research was to estimate the association between melanoma risk and use of two types of UV-emitting devices: sunlamps (used mostly pre-1980; higher UVB content) and tanning beds (used mostly post-1980; higher UVA content) while adjusting for a comprehensive vector of melanoma risk factors. Before conducting the present analyses, tanning device use was hypothesized to be associated with an increased risk of melanoma when compared with non– users, adjusting for host characteristics and solar exposures.
Materials and methods
This study was approved by the Committee for the Protection of Human Subjects at Dartmouth College. All participants gave verbal consent for the interview and signed consent for the skin examination and for the release of pathology records.
The methods of this study have been described previously [11]. Case subjects of ages 20 through 69 with a diagnosis of cutaneous malignant melanoma (hereafter referred to as melanoma) occurring between January 1995 and December 1998 were ascertained through the New Hampshire (NH) State Cancer Registry. Those potentially eligible for study enrollment were NH residents with a working telephone number who were able to participate in an interview conducted in English. We sent a letter to the physician of record requesting permission to contact the patient. If an objection was not received within a month, a letter introducing the study was mailed to the case subject, followed within 2 weeks by a telephone call from the interviewer. Using this approach, we enrolled 444 of 579 (77%) potentially eligible cases; 15 (3%) were excluded at their physician’s request, 26 (4%) could not be reached, 30 (5%) had died, and 64 (11%) declined to participate. Twenty-one enrolled cases were deemed ineligible; of these, seven had a previous diagnosis of melanoma, four had an unknown primary site, two had tumors of acral lentiginous histology, and for eight persons the diagnosis of melanoma was not definitive. Thus, 423 cases of first primary melanoma were available for analysis.
Control subjects were ascertained from lists of licensed drivers obtained through the NH Department of Motor Vehicles, and were selected at random to achieve a gender and age (in 5-year age groups) distribution similar to that of case subjects. Controls were also selected to achieve a control:case ratio of 1.6 to allow separate studies of atypical moles within the control group. Potentially eligible control subjects were NH residents with a working telephone, and able to participate in an English-speaking interview. A letter introducing the study was sent to potential control participants, followed within 2 weeks by a telephone call from the interviewer. We enrolled 684 of 1121 (61%) potentially eligible control subjects; 87 (8%) could not be reached, 13 (1%) controls had died, and 337 (30%) declined to participate. Of the 684 control participants, six were deemed ineligible due to a prior diagnosis of melanoma. Thus, 678 controls were available for analysis.
The 40-min telephone interview queried participants for demographic factors, pigmentary characteristics, episodes of sunburn, sun exposure, and use of tanning beds or sunlamps. The details of the questionnaire have been described elsewhere [11]. Briefly, we asked participants to report eye color, natural hair color at age 20, and the reaction of their skin to strong summer sun exposure. Sunburns were assessed separately as episodes of peeling sunburn and blistering sunburn in 10-year age periods starting at age 10. Sun exposure was assessed by asking participants to report outdoor occupational and recreational activities, and sunbathing, defined as relaxing in the sun. Information regarding sunbathing was captured for 10-year age periods starting at age 10; outdoor occupations were assessed starting at age 6 (to accommodate farm work). We queried participants for 11 standard outdoor recreational activities, and an unlimited number of “other activities,” starting at age 10. For all three variables, hours of sun exposure were capped at 10 h per day. Sun exposure variables were derived as cumulative exposures (total hours exposed) for each sun exposure type (recreational, sunbathing, and occupational) and a combined total sun exposure category. Each variable was assessed within life periods categorized as childhood (≤20 years), adulthood (≥20 years), and lifetime (total). Participants were also asked to report their use of sunlamps and tanning beds separately, by answering yes or no to the following questions: “Have you ever used a sunlamp?” or “Have you ever gone to a tanning salon or used a tanning bed?” For each exposure subjects were asked the total number of times the device was used, the age of first use, and the age of last use. No differentiations for medical use or pictures of device types were used to prompt the responders. All exposures were assessed up until the reference date. The reference date was 1 year prior to the date of diagnosis for cases, and randomly assigned to controls based on the frequency of diagnosis dates in the case group.
We first examined descriptive statistics (univariate, proportion, and frequency) on all study variables. All bivariate relations were examined using Pearson or Spearman correlations and chi-square tests. Exposure variables were initially examined in age- and gender-adjusted logistic models of melanoma risk. The final subset of confounders was selected by the model building strategy based on effect changes described by Greenland [56]. Briefly, potential confounders were variables that, when added to age- and sex-adjusted models of melanoma risk, changed the OR of the exposures of interest (sunlamp, tanning bed use) more than 10%. Highly correlated variables were not included simultaneously in the models. When variables were correlated, the final models contained the term for the stronger confounder. The stronger confounder was determined by the magnitude of change in the OR corresponding to the exposure of interest when the variable was added to the logistic model and by the strength of its association with melanoma and the exposures of interest. In the models, continuous variables were initially in quintiles and combined when appropriate to avoid sparsely populated categories. Odds ratios (OR) and 95% confidence intervals (CI) were computed from logistic regression models to examine the influence of sunlamp or tanning bed use on melanoma risk. Tests for trend were based on the categories shown and conducted using the z-statistic defined as the beta coefficient divided by its standard error. Effect measure modification between sunlamps/tanning beds and other variables (e.g., age and pigmentary characteristics) was analyzed using stratification and the Breslow-Day test for homogeneity of ORs. All statistical analyses were conducted in SAS and all p values were based on two-sided tests [57].
Results
Table 1 shows the characteristics of the study population by case–control status. Cases and controls were comparably distributed by age and gender, reflecting the matching scheme. The majority of participants (77.2%) were married; cases had slightly more education than controls, and nearly half of all participants had at least a college education. As expected, blue eye color was more common in the cases (46.6% vs. 34.8%), and brown eye color was more common in the controls (34.6% vs. 23.0%). Red or blonde hair color was more common among the cases, whereas brown, dark-brown, and black were more common in the controls. In response to strong summer sun, a higher proportion of cases reported burning without tanning 48.2% vs. 42.2%), whereas a higher proportion of controls tanned without burning (10.0% vs. 3.3%).
Table 1.
Characteristics of the study population by case and control status
| Case n = 423 n (%) |
Control n = 678 n (%) |
ORcrude (95% CI) |
|
|---|---|---|---|
| Gender | |||
| Female | 200 (47.3) | 330 (48.7) | |
| Male | 223 (52.7) | 348 (51.3) | |
| Age | |||
| 20–29 years old | 22 (5.2) | 26 (3.8) | |
| 30–39 | 72 (17.0) | 132 (19.5) | |
| 40–49 | 97 (22.9) | 134 (19.8) | |
| 50–59 | 117 (27.7) | 197 (29.0) | |
| 60–69 | 115 (27.2) | 189 (27.9) | |
| Mean age (±SD) | 50.1 (±12.2) | 50.3 (±11.6) | |
| Educational Level | |||
| Completed school through grade 8 | 20 (4.7) | 35 (5.2) | 1.0 |
| Completed high school | 144 (34.1) | 261 (38.6) | 0.97 (0.54–1.74) |
| Completed college (2-year degree) | 74 (17.5) | 127 (18.8) | 1.02 (0.55–1.90) |
| Completed college (4-year degree) | 115 (27.3) | 157 (23.2) | 1.28 (0.70–2.34) |
| Completed graduate or professional school | 69 (16.4) | 96 (14.2) | 1.26 (0.67–2.40) |
| Marital Status | |||
| Married | 329 (78.3) | 517 (76.3) | 1.13 (0.84–1.50) |
| Not married | 91 (21.7) | 161 (23.7) | |
| Family history of melanoma | |||
| Yes | 98 (23.2) | 131 (20.2) | 1.25 (0.93–1.70) |
| No | 320 (76.8) | 535 (79.8) | |
| Pigmentary characteristics | |||
| Eye color | |||
| Brown | 97 (23.0) | 234 (34.6) | 1.0 |
| Blue | 196 (46.6) | 235 (34.8) | 2.01 (1.49–2.70) |
| Green/gray/hazel | 128 (30.4) | 207(30.6) | 1.50 (0.93–2.50) |
| Hair color age 20 years | |||
| Brown/dark brown/black | 203 (47.7) | 410 (60.6) | 1.0 |
| Reddish blond/strawberry blond | 19 (4.5) | 13 (1.9) | 5.52 (1.99–15.30) |
| Red | 17 (4.0) | 13 (1.9) | 4.94 (1.76–13.80) |
| Blond | 43 (10.2) | 39 (5.8) | 4.13 (1.80–9.80) |
| Light brown/dark blond | 123 (29.1) | 172 (25.4) | 2.70 (1.25–5.80) |
| Red-brown or auburn | 19 (4.5) | 30 (4.4) | 2.39 (0.94–6.08) |
| Freckles | |||
| Yes | 302 (73.1) | 358 (52.8) | 2.31 (1.77–3.01) |
| No | 114 (26.9) | 312 (47.2) | |
| Sun sensitivity-acute exposure | |||
| Tan without sunburn | 14 (3.3) | 68 (10.0) | 1.0 |
| Sunburn w/peeling, no tan | 204 (48.2) | 282 (42.2) | 3.51 (1.90–6.42) |
| Sunburn with peeling and freckles, no tan | 20 (4.7) | 42 (6.2) | 2.31 (1.10–5.10) |
| Sunburn followed by tan | 185 (43.8) | 286 (41.6) | 3.14 (1.70–5.75) |
| Sunburn history | |||
| Childhood sunburn with peeling | |||
| Never | 32 (7.7) | 95(14.3) | 1.0 |
| Low | 195 (47.1) | 323 (48.6) | 1.79 (1.16–2.78) |
| High | 187 (45.2) | 246 (37.1) | 2.26 (1.45–3.51) |
| Adult sunburn with peeling | |||
| Never | 74(18.1) | 122 (18.4) | 1.0 |
| Low | 156 (38.0) | 284 (42.8) | 0.91 (0.63–1.28) |
| High | 180 (43.9) | 258 (38.8) | 1.15 (0.81–1.62) |
| Lifetime sunburn with peeling | |||
| Never | 6(1.5) | 36 (5.5) | 1.0 |
| Low | 194 (47.8) | 323 (49.2) | 3.60 (1.49–8.70) |
| High | 206 (50.7) | 297 (45.3) | 4.16 (1.68– 10.11) |
| Childhood sunburn with blistering | |||
| Never | 198 (47.5) | 393 (58.2) | 1.0 |
| Low | 121 (29.0) | 160 (23.7) | 1.50 (1.12–2.01) |
| High | 98(23.5) | 122 (18.1) | 1.59 (1.16–2.19) |
| Adult sunburn with blistering | |||
| Never | 260 (62.9) | 443 (66.5) | 1.0 |
| Low | 92(22.3) | 149 (22.4) | 1.05 (0.78–1.42) |
| High | 61(14.8) | 74 (11.1) | 1.41 (1.00–2.03) |
| Lifetime sunburn with blistering | |||
| Never | 144 (35.1) | 286 (43.0) | 1.0 |
| Low | 130 (31.7) | 221 (33.2) | 1.17 (0.89–1.61) |
| High | 136 (33.2) | 158 (23.8) | 1.71 (1.28–2.32) |
| Solar exposure history | |||
| Total childhood solar exposure | |||
| None | 8(1.9) | 11 (1.7) | 1.0 |
| Low | 181 (44.6) | 340 (51.9) | 0.73 (0.29–1.85) |
| High | 217 (53.5) | 304 (46.4) | 0.98 (0.37–2.48) |
| Total adulthood solar exposure | |||
| None | 0(0) | 0 (0) | – |
| Low | 180 (45.6) | 337 (52.5) | 1.0 |
| High | 214 (54.4) | 295 (47.5) | 1.36 (1.11–1.72) |
| Total lifetime solar exposure | |||
| None | 0(0) | 0 (0) | – |
| Lowref | 175 (45.0) | 333 (53.1) | 1.0 |
| High | 214 (55.0) | 294 (46.9) | 1.39 (1.07–1.79) |
| Recreational childhood solar exposure | |||
| None | 46(11.1) | 82 (12.4) | 1.0 |
| Low | 166 (40.2) | 307 (46.5) | 0.96 (0.64–1.44) |
| High | 201 (48.7) | 271 (41.1) | 1.32 (0.88–1.98) |
| Recreational adulthood solar exposure | |||
| None | 8(1.9) | 13 (2.4) | 1.0 |
| Low | 194 (47.0) | 329 (50.1) | 1.18 (0.50–2.81) |
| High | 211 (51.1) | 312 (47.5) | 1.35 (0.57–3.22) |
| Recreational lifetime solar exposure | |||
| None | 0(0) | 0 (0) | – |
| Low | 187 (45.3) | 348 (53.0) | 1.0 |
| High | 226 (54.7) | 309 (47.0) | 1.36 (1.06–1.74) |
| Occupational childhood solar exposure | |||
| None | 196 (46.3) | 314 (46.3) | 1.0 |
| Low | 117 (27.7) | 179 (26.4) | 1.05 (0.78–1.40) |
| High | 110 (26.0) | 185 (27.3) | 0.95 (0.71–1.28) |
| Occupational adulthood solar exposure | |||
| None | 272 (64.3) | 458 (67.8) | 1.0 |
| Low | 74(17.5) | 115 (17.0) | 1.08 (0.78–1.51) |
| High | 77(18.2) | 103 (15.2) | 1.26 (0.90–1.78) |
| Occupational lifetime solar exposure | |||
| None | 160 (37.8) | 266 (29.4) | 1.0 |
| Low | 136 (32.2) | 211 (31.2) | 1.07 (0.80–1.43) |
| High | 127 (30.0) | 199 (39.4) | 1.06 (0.79–1.42) |
| Sunbathing childhood solar exposure | |||
| None | 143 (34.4) | 232 (34.5) | 1.0 |
| Low | 133 (32.0) | 226 (33.6) | 0.96 (0.71–1.29) |
| High | 140 (33.6) | 215 (31.9) | 1.06 (0.78–1.42) |
| Sunbathing adulthood solar exposure | |||
| None | 115 (28.5) | 208 (31.8) | 1.0 |
| Low | 143 (35.4) | 224 (34.4) | 1.16 (0.84–1.57) |
| High | 146 (36.1) | 221 (33.8) | 1.20 (0.88–1.63) |
| Sunbathing lifetime solar exposure | |||
| None | 90(22.6) | 150 (23.1) | 1.0 |
| Low | 152 (38.1) | 252 (38.9) | 1.01 (0.72–1.40) |
| High | 157 (39.3) | 246 (38.0) | 1.06 (0.77–1.48) |
Sunlamp use
Overall, 17.1% of study participants (20.3% of cases, 15.0% of controls) reported ever using a sunlamp. The average age of first use was 20 years for cases and controls. Virtually all sunlamp use (89%) occurred before 1980 (Fig. 1). We found little evidence that the association between sunlamp use and melanoma risk was confounded by measured covariates (Table 2). The covariate-adjusted OR for ever use of a sun-lamp was 1.39 (95% CI 1.00–1.96) overall. Of those who used sunlamps, 61.2% of cases and 66.3% of controls did so before age 20. The covariate-adjusted OR was 1.23 (95% CI 0.81–1.88) for those starting use before age 20, and 1.71 (95% CI 1.00–2.92) for those starting use after age 20. Based on the cutpoints shown, the data suggested an increasing risk with the number of times used (test of trend p = 0.02). Relative to never use, the adjusted OR was 1.29 (95% CI 0.84–1.99) for use less than 6 times, and 1.54 (95% CI 0.93– 2.57) for use 6 or more times. A greater duration of use (years elapsed between last and first use) (≤1,>1) was also associated with an increased melanoma risk (test of trend p = 0.02); the OR for >1 year of use was 1.57 (95% CI 0.91–2.71). Although the estimates were imprecise, the data suggested an elevated risk for those who last used a sunlamp 15 or more years earlier, although the findings were not of statistical significance. The ORs were 1.53 (95% CI 0.84–2.81) and 1.38 (95% CI 0.90–2.12) respectively for 15–24 and ≥25 years since last use, compared to never use.
Fig. 1.
Prevalence of exposure to artificial tanning devices according to device type and decade of first use in a population based, case–control study of incident cutaneous melanoma in New Hampshire between 1995 and 1998
Table 2.
Odds ratios (OR) and 95% confidence intervals (CI) for sunlamp and tanning bed use in relation to melanoma
| Case n (%) |
Control n (%) |
OR* (95% CI) |
OR** (95% CI) |
p value Test of trend |
|
|---|---|---|---|---|---|
| Sunlamp | |||||
| Sunlamp use | |||||
| Never | 337 (79.7) | 576 (85) | 1.00 | 1.00 | |
| Ever | 86 (20.3) | 102 (15.0) | 1.46 (1.06–2.01) | 1.39 (1.00–1.96) | |
| Age at first use | |||||
| Never | 337 (79.9) | 576 (85.1) | 1.00 | 1.00 | p = 0.05 |
| <20 | 52 (12.3) | 67 (9.9) | 1.34 (0.91–1.98) | 1.23 (0.81–1.88) | |
| ≥20 | 33 (7.8) | 34 (5.0) | 1.70 (1.03–2.80) | 1.71 (1.00–2.92) | |
| Frequency of use | |||||
| Never | 337 (79.9) | 576 (85.1) | 1.00 | 1.00 | p = 0.02 |
| <6 times | 52 (11.6) | 67 (9.4) | 1.33 (0.89–1.98) | 1.29 (0.84–1.99) | |
| ≥6 times | 33 (8.5) | 34 (5.5) | 1.69 (1.05–2.73) | 1.54 (0.93–2.57) | |
| Years of use | |||||
| Never | 337 (79.9) | 576 (85.1) | 1.00 | 1.00 | p = 0.02 |
| ≤1 year | 55 (13.0) | 67 (9.9) | 1.42 (0.97–2.08) | 1.30 (0.86–1.98) | |
| >1 year | 30 (7.1) | 34 (5.0) | 1.55 (0.93–2.59) | 1.57 (0.91–2.71) | |
| Years since last use | |||||
| Never | 337 (79.9) | 576 (85.1) | 1.00 | 1.00 | p = 0.07 |
| <15 | 7 (1.7) | 13 (1.9) | 0.92 (0.36–2.32) | 1.10 (0.42–2.96) | |
| 15–24 | 25 (5.9) | 27 (4.0) | 1.58 (0.90–2.78) | 1.53 (0.84–2.81) | |
| ≥25 | 53 (12.6) | 61 (9.0) | 1.53 (1.03–2.27) | 1.38 (0.90–2.12) | |
| Tanning bed | |||||
| Tanning bed use | |||||
| Never | 326 (77.1) | 536 (79.1) | 1.00 | 1.00 | |
| Ever | 97 (22.9) | 142 (20.9) | 1.15 (0.83–1.58) | 1.14 (0.80–1.61) | |
| Age at first use | |||||
| Never | 326 (77.1) | 536 (79.2) | 1.00 | 1.00 | |
| <20 | 18 (4.3) | 17 (2.5) | 1.89 (0.90–3.97) | 1.78 (0.76–4.15) | |
| ≥20 | 79 (18.6) | 124 (18.3) | 1.07 (0.78–1.51) | 1.08 (0.75–1.55) | p = 0.65 |
| Frequency of use | |||||
| Never | 326 (77.1) | 536 (79) | 1.00 | 1.00 | p = 0.42 |
| <10 times | 43 (10.2) | 78 (12) | 0.93 (0.62–1.41) | 1.05 (0.67–1.64) | |
| ≥10 times | 54 (12.7) | 63 (9) | 1.46 (0.96–2.21) | 1.25 (0.79–1.98) | |
| Years of use | |||||
| Never | 326 (77) | 536 (79.2) | 1.00 | 1.00 | p = 0.49 |
| ≤1 year | 50 (12) | 79 (11.5) | 1.06 (0.72–1.68) | 1.17 (0.76–1.79) | |
| >1 year | 47 (11) | 62 (9.3) | 1.28 (0.83–1.98) | 1.09 (0.68–1.76) | |
| Years since last use | |||||
| Never | 326 (77.1) | 536 (79.2) | 1.00 | 1.00 | p = 0.50 |
| <15 | 89 (21.0) | 127 (18.8) | 1.18 (0.85–1.66) | 1.15 (0.79–1.66) | |
| ≥15 | 8 (1.9) | 14 (2.0) | 0.95 (0.40–2.30) | 1.06 (0.42–2.66) | |
| Combined device use | |||||
| Either sunlamp or tanning bed use | |||||
| Never | 156 (37.1) | 218 (32.0) | 1.0 | 1.0 | |
| Ever | 267 (63.2) | 460 (68.1) | 1.27 (0.97–1.71) | 1.22 (0.83–1.80) | |
| Both sunlamp and tanning bed use | |||||
| Never | 267 (63.1) | 460 (67.9) | 1.00 | 1.00 | |
| Ever | 27(6.4) | 26 (3.8) | 1.84 (1.05–3.25) | 1.96 (1.06–3.61) | |
OR adjusted for age and gender
OR adjusted for age, gender, family history of melanoma, hair color, freckles, sun sensitivity, total sun exposure hours
Analyses stratified on age groups (20–29, 30–39, 40–49, 50–59, and 60–69) suggested that melanoma risk was not uniform across age groups (data not shown). In particular, the effect of ever sunlamp use appeared to be most pronounced in the 40–49 and 50–59 year age groups; OR = 1.42 (95% CI 0.75–2.71) and 2.47 (95% CI 1.40–4.38), respectively. However, Breslow-Day tests for homogeneity of the odds ratios indicated that the differences in effects stratified across 10-year age groups for all sunlamp exposures were not statistically significant (p ≥ 0.10).
Additional analyses failed to show an association between sunlamp use and the site of the melanoma. We also found no evidence that the relationship between sun-lamp use was modified by the host characteristics (eye color, hair color, sun sensitivity); all p values were ≥0.30.
Tanning bed use
Overall, 21.9% of study participants (22.9% of cases, 20.9% of controls) reported ever using a tanning bed. The average age of first use was 33 overall, 33 for cases, and 34 for controls. Most use of tanning beds occurred after 1979 (Fig. 1). The data suggested minimal confounding of frequency and duration of use by measured covariates. The covariate-adjusted OR for ever using a tanning bed was 1.14 (95% CI 0.80–1.61) overall (Table 2). Of those who used tanning beds, 18.6% of cases and 12.1% of controls reported first using a tanning bed before age 20. The covariate-adjusted OR was 1.78 (95% CI 0.76–4.15) for those starting use before age 20, and 1.08 (95% CI 0.75–1.55) for those starting at age 20 or older. Based on the cut-points shown, there was no evidence that melanoma risk increased with either the number of episodes of tanning bed use (<10, ≥10), or the duration of use in years (≤1,>1). There was also no indication that time since last tanning bed use (<15, ≥15 years) was associated with an increased melanoma risk, and numbers were sparse in the latter grouping.
We conducted extensive sub-analyses according to the era and decade of tanning bed use (exposure before or after 1980), and by age of use. Aside from increased imprecision, the results were similar by strata corresponding to era or decade of use. In analyses stratified by age (in 10-year age-groups), the OR for ever use was highest in the 60–69-year age-group [1.56 (95% CI 0.66–3.66)]. However, we found no statistical evidence of heterogeneity of odds ratios across the 10-year age strata for any of the tanning bed exposures, including ever use (Breslow-Day test of homogeneity p ≥ 0.20).
We found no association between tanning bed use and the site of the melanoma tumor. There was also no evidence of effect modification by pigmentary characteristics (eye color, hair color, sun sensitivity, p ≥ 0.25).
Any device use and combined device use
Approximately one-third of both cases and controls reported using either a sunlamp or a tanning bed; the covariate-adjusted OR for use of either device was 1.22 (95% CI 0.83–1.80). Only 6.4% of cases and 3.8% of controls had used both sunlamps and tanning beds. When use of both UV devices was combined into one exposure variable, the covariate-adjusted OR was 1.96 (95% CI 1.06–3.61). The data provided no indication of increasing risk with either the frequency of use, or duration of use of both types of UV device combined. (data not shown)
Discussion
Although UV radiation from the sun is an accepted risk factor for all types of skin cancers, the evidence is far less clear for UV produced from artificial sources. Our findings support an association between sunlamp use and melanoma risk. Because sunlamps emit predominantly UVB, [47] which is known to increase skin cancer risk, this association is not surprising. Our findings also suggest that melanoma risk increases with more frequent and a longer duration of sunlamp use, although most use was over a short time period (on average less than 3 years). Consistent with a latency period of tumor development, risk appeared to be elevated for those who last used a sunlamp at least 15 years earlier, but not for those who discontinued use within the preceding 14 years.
We did not find evidence of an association between tanning bed use, which occurred mainly after 1980 in this study population, and melanoma risk. Modern tanning beds emit predominantly UVA, which may have less influence than UVB on melanoma risk [25, 29, 40, 47]. However, the global distribution of UVA radiation is associated with melanoma mortality rates, [14] medical use of UVA has shown an increase in skin cancers including melanoma, [23] and laboratory studies show that UVA can induce DNA damage in cells and melanoma precursors in animals [18]. In our study, tanning bed users started use later (33 vs. 20 years), and most use had occurred less than 20 years before the melanoma diagnosis date. Thus, tumor-induction time may have been insufficient, resulting in an underestimate of the association between tanning bed use and melanoma risk. This possibility is supported by Stern et al. who found that melanomas associated with UVA treatment of dermatologic conditions appeared to have a latency of at least 10–15 years [18, 23, 58]. Also, the most recent meta-analyses concluded that exposure at younger ages to tanning devices may have the biggest impact on melanoma risk, whereas in our study less than 15% of tanning bed users started use before the age of 20 [25, 29].
Similar to Chen et al. the strongest association was seen in subjects reporting use of both sunlamps and tanning beds. We explored the possibility this might be explained by the overall increased ‘sun-seeking’ behaviors among individuals that used both devices, but found no correlation between sun exposures and use of both devices, or with duration of use. Thus, it is possible the increased odds, although mainly arising from sunlamp exposures, may reflect the increased overall artificial UV exposure. Of course it is also possible the larger OR may be the result of small numbers or simply an artifact, which we were unable to assess.
A number of epidemiologic investigations have attempted to determine the nature of the putative association between artificial UV and melanoma with more recent studies generally suggesting use of artificial tanning devices is a risk factor [25, 28, 29]. However, three recent European studies reported conflicting results; the U.K. and multi-center European case–control studies found no association and the only cohort study to date in Scandinavia found the highest association among the youngest users with the longest lag time [26, 32, 59]. The difference in findings may be due to insufficient differentiation by device type, era of use, age at use of device type (i.e., not accounting for greater lag-time and use of devices with higher UVB-emissions) in the case–control studies. As with our findings the authors and reviewers of the Scandinavian study suggested that the increased risk was likely due to use of devices with higher UVB-emission and sufficient lag time to melanoma development. Swerdlow et al.’s 1998 review of the epidemiologic literature concluded there is insufficient evidence to determine whether or not tanning devices cause melanoma [37]. However, more recent meta-analyses by Gallagher et al. and the IARC identified a modestly increased risk of melanoma related to ever using a tanning device, despite between-study significant heterogeneity. Control for sun exposure, differentiation of device type, and era of use was not possible in most of the studies they assessed [25, 29]. Our findings, which allowed adjustment for solar exposures and differentiation of era of use, support the association with sunlamp use but are inconclusive for tanning bed use.
If an association between tanning bed use and melanoma risk is established, our data suggest the consequences at the population level may be greater than those for sunlamp use. Compared to sunlamp use, a slightly larger proportion of individuals reported using tanning beds, first used them at a younger age, used them more frequently and for a longer duration. Also, tanning bed use continues to gain popularity especially among teenagers and young adults, whereas sunlamps are largely of the past [41, 43, 44, 46, 48].
Strengths of the study include the amount of information available to assess tanning device use and adjust for a robust set of covariates. Many previous studies used only ever/ never exposure categories and potential confounders, such as sun exposure, sun sensitivity, and socioeconomic status, were often not included in the previous analyses [28, 37, 46]. We were able to assess the frequency and duration of tanning device use, along with age at first use and time since last use. We were also able to adjust for a comprehensive vector of potential confounders and to conduct exploratory analyses involving sun exposures and sunburn histories. The latter analyses showed no correlation between combined device use and sun-exposure variables, reducing the likelihood that sunlamp or tanning bed use is a marker of sun-seeking behaviors [60]. The sunburn histories were also similar for users versus nonusers of tanning devices. However, differences in sun exposure histories did exist with tanning device non-users typically having the greatest number of hours of sun exposure.
There are several limitations of this research. Participants were retrospectively interviewed to obtain self-reported artificial tanning device use and other covariates, including solar exposures. We cannot exclude the possibility of recall bias, and in particular, cases might be more likely than controls to report exposures. It seems unlikely, however, that recall bias would affect self-reported sun-lamp use without similarly affecting self-reported tanning bed use. It is conceivable that there is a differential level of accuracy in the information provided because cases may be more likely to recall the actual details (frequency, duration, and age at use) of their exposure more accurately than controls. However, it seems unlikely that cases or controls would differentially or incorrectly report ever/never use of tanning devices. Because no preexisting records exist for tanning device use, there are no alternatives to the use of self-reported exposure data. It is conceivable that the small decrease in the OR when adjusted for covariates may be due to imprecise measurement. However, previous studies showed overall comparable small changes in effects from adjustments including solar exposures [26, 27, 33, 39, 61]. A potential limitation of our study is that the interview rate among the controls was lower than that in cases. If cases or controls that did not participate differed from interviewed subjects with regard to exposure history, then selection bias would be introduced. Unfortunately, no data on characteristics of non-participants was available.
As previously stated, we do not know the intensity or spectral outputs of the devices to which participants were exposed. Two recent studies have reported risk by detailed device type [32, 33]. Similar to Chen et al. but unlike Bataille et al. we found the strongest effect for sunlamp and combined device use rather than for tanning bed use [32, 33]. The nature of lamps and the type of use in the US (UVA-dominated versus UVB-dominated, commercial versus home) have changed over time [25, 29, 47]. Diffey and Farr showed that, since the mid-1980s, modern tanning devices available primarily in commercial settings use fluorescent bulbs that emit mostly UVA with smaller UVB doses [47]. In our study, most sunlamp exposures occurred before the changeover and most tanning bed exposures occurred afterward. Consequently, the association we observed with sunlamps probably reflects UVB exposure. In an effort to compensate for the lack of detailed information about UV emissions in sunlamps and tanning beds, we examined both the decade of first use and the type of device used in relation to risk. Separating the device categories (sunlamp vs. tanning bed) in this study provided a proxy, although not perfect, for higher (sunlamp) and lower (tanning bed) ratios of UVB to UVA emission [27, 62, 63]. It would be difficult, if not impossible, in a retrospective study to collect detailed information on output spectra and intensity of devices because the technical features of tanning devices are variable and have changed over time. Another limitation of our study is that our duration measure, which was based on time of first and last use, is only a proxy for actual duration of use.
In summary, our results suggest an association between sunlamp use (higher UVB content) and melanoma risk, and an increasing risk with greater frequency and duration of use; sufficient lag time may not have elapsed to assess a potential effect of tanning bed use (higher UVA content). The dependent association between artificial UV and melanoma is complicated by the wide variety of emission spectra that characterize devices and lag time between device exposure and melanoma diagnosis. Prospective research designed to incorporate sufficient lag times as well as to examine the UV emissions of artificial tanning device exposures would provide critical information for understanding the potential relation of UVA to melanoma risk.
Acknowledgments
We are grateful to Rebecca A. Silliman, Professor of Medicine and Public Health and Timothy L. Lash Associate Professor of Public Health and Medicine both at Boston University, for their informed advice and thoughtful critic of this study. We would like to thank the physicians in New Hampshire for their cooperation with this study and the men and women of New Hampshire, who generously gave their time to ensure the success of this project.
This work was supported by grant RO1CA 66032 from the National Cancer Institute and conducted at Dartmouth Medical School. This manuscript contains original material that has not been previously presented.
Footnotes
None of the authors have a conflict of interest.
Contributor Information
Kerri M. Clough-Gorr, Geriatrics Section, Department of Medicine, Boston University, Medical Center, 88 East Newton Street, Robinson 2, Boston, MA 02118, USA, kmclough@bu.edu
Linda Titus-Ernstoff, Departments of Community & Family Medicine and of Pediatrics, Dartmouth Medical School and the Norris Cotton Cancer Center, Lebanon, NH, USA.
Ann E. Perry, Department of Pathology, Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA
Steven K. Spencer, Department of Dermatology, Dartmouth-Hitchcock Medical, Center, Lebanon, NH, USA
Marc S. Ernstoff, Department of Medicine, Dartmouth-Hitchcock Medical Center, Lebanon, NH, USA
References
- 1.Berwick M, Wiggins C. The current epidemiology of cutaneous malignant melanoma. Front Biosci. 2006;11:1244–1254. doi: 10.2741/1877. [DOI] [PubMed] [Google Scholar]
- 2.Brochez L, Naeyaert JM. Understanding the trends in melanoma incidence and mortality: where do we stand? Eur J Dermatol. 2000;10:71–75. quiz 76. [PubMed] [Google Scholar]
- 3.de Vries E, Bray FI, Coebergh JW, Parkin DM. Changing epidemiology of malignant cutaneous melanoma in Europe 1953–1997: rising trends in incidence and mortality but recent stabilizations in western Europe and decreases in Scandinavia. Int J Cancer. 2003;107:119–126. doi: 10.1002/ijc.11360. [DOI] [PubMed] [Google Scholar]
- 4.American Cancer Society. Cancer facts & figures 2007. Atlanta, GA: 2007. http://www.cancer.org. [Google Scholar]
- 5.The American Cancer Society website. 2005 http://www.cancer.org.
- 6.Desmond RA, Soong SJ. Epidemiology of malignant melanoma. Surg Clin North Am. 2003;83:1–29. doi: 10.1016/s0039-6109(02)00092-0. [DOI] [PubMed] [Google Scholar]
- 7.Gallagher RP, Elwood JM, Hill GB. Risk factors for cutaneous malignant melanoma: the western Canada melanoma study. Recent Results Cancer Res. 1986;102:38–55. doi: 10.1007/978-3-642-82641-2_4. [DOI] [PubMed] [Google Scholar]
- 8.Gandini S, Sera F, Cattaruzza MS, et al. Meta-analysis of risk factors for cutaneous melanoma: I. Common and atypical naevi. Eur J Cancer. 2005;41:28–44. doi: 10.1016/j.ejca.2004.10.015. [DOI] [PubMed] [Google Scholar]
- 9.Gandini S, Sera F, Cattaruzza MS, et al. Meta-analysis of risk factors for cutaneous melanoma: III. Family history, actinic damage and phenotypic factors. Eur J Cancer. 2005;41:2040–2059. doi: 10.1016/j.ejca.2005.03.034. [DOI] [PubMed] [Google Scholar]
- 10.Titus-Ernstoff L. An overview of the epidemiology of cutaneous melanoma. Clin Plast Surg. 2000;27:305–316. vii. [PubMed] [Google Scholar]
- 11.Titus-Ernstoff L, Perry AE, Spencer SK, et al. Pigmentary characteristics and moles in relation to melanoma risk. Int J Cancer. 2005;116:144–149. doi: 10.1002/ijc.21001. [DOI] [PubMed] [Google Scholar]
- 12.Weinstock MA. Issues in the epidemiology of melanoma. Hematol Oncol Clin North Am. 1998;12:681–698. doi: 10.1016/s0889-8588(05)70018-6. [DOI] [PubMed] [Google Scholar]
- 13.Gandini S, Sera F, Cattaruzza MS, et al. Meta-analysis of risk factors for cutaneous melanoma: Ii. Sun exposure. Eur J Cancer. 2005;41:45–60. doi: 10.1016/j.ejca.2004.10.016. [DOI] [PubMed] [Google Scholar]
- 14.Garland CF, Garland FC, Gorham ED. Epidemiologic evidence for different roles of ultraviolet a and b radiation in melanoma mortality rates. Ann Epidemiol. 2003;13:395–404. doi: 10.1016/s1047-2797(02)00461-1. [DOI] [PubMed] [Google Scholar]
- 15.Jhappan C, Noonan FP, Merlino G. Ultraviolet radiation and cutaneous malignant melanoma. Oncogene. 2003;22:3099–3112. doi: 10.1038/sj.onc.1206450. [DOI] [PubMed] [Google Scholar]
- 16.Kligman LH, Akin FJ, Kligman AM. Sunscreens prevent ultraviolet photocarcinogenesis. J Am Acad Dermatol. 1980;3:30–35. doi: 10.1016/s0190-9622(80)80221-0. [DOI] [PubMed] [Google Scholar]
- 17.Knox JM, Griffin AC, Hakim RE. Protection from ultraviolet carcinogenesis. J Invest Dermatol. 1960;34:51–58. [PubMed] [Google Scholar]
- 18.Wang SQ, Setlow R, Berwick M, et al. Ultraviolet a and melanoma: a review. J Am Acad Dermatol. 2001;44:837–846. doi: 10.1067/mjd.2001.114594. [DOI] [PubMed] [Google Scholar]
- 19.Elwood JM, Lee JA, Walter SD, Mo T, Green AE. Relationship of melanoma and other skin cancer mortality to latitude and ultraviolet radiation in the united states and canada. Int J Epidemiol. 1974;3:325–332. doi: 10.1093/ije/3.4.325. [DOI] [PubMed] [Google Scholar]
- 20.Lancaster HO. Some geographical aspects of the mortality from melanoma in europeans. Med J Aust. 1956;43:1082–1087. [PubMed] [Google Scholar]
- 21.Hannuksela-Svahn A, Sigurgeirsson B, Pukkala E, et al. Trioxsalen bath puva did not increase the risk of squamous cell skin carcinoma and cutaneous malignant melanoma in a joint analysis of 944 Swedish and Finnish patients with psoriasis. Br J Dermatol. 1999;141:497–501. doi: 10.1046/j.1365-2133.1999.03044.x. [DOI] [PubMed] [Google Scholar]
- 22.Lindelof B, Sigurgeirsson B, Tegner E, et al. Puva and cancer risk: the Swedish follow-up study. Br J Dermatol. 1999;141:108–112. doi: 10.1046/j.1365-2133.1999.02928.x. [DOI] [PubMed] [Google Scholar]
- 23.Stern RS. The risk of melanoma in association with long-term exposure to puva. J Am Acad Dermatol. 2001;44:755–761. doi: 10.1067/mjd.2001.114576. [DOI] [PubMed] [Google Scholar]
- 24.Autier P. Perspectives in melanoma prevention: the case of sunbeds. Eur J Cancer. 2004;40:2367–2376. doi: 10.1016/j.ejca.2004.07.018. [DOI] [PubMed] [Google Scholar]
- 25.Gallagher RP, Spinelli JJ, Lee TK. Tanning beds, sun-lamps, and risk of cutaneous malignant melanoma. Cancer Epidemiol Biomarkers Prev. 2005;14:562–566. doi: 10.1158/1055-9965.EPI-04-0564. [DOI] [PubMed] [Google Scholar]
- 26.Veierod MB, Weiderpass E, Thorn M, et al. A prospective study of pigmentation, sun exposure, and risk of cutaneous malignant melanoma in women. J Natl Cancer Inst. 2003;95:1530–1538. doi: 10.1093/jnci/djg075. [DOI] [PubMed] [Google Scholar]
- 27.Westerdahl J, Ingvar C, Masback A, Jonsson N, Olsson H. Risk of cutaneous malignant melanoma in relation to use of sunbeds: further evidence for uv-a carcinogenicity. Br J Cancer. 2000;82:1593–1599. doi: 10.1054/bjoc.1999.1181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Young AR. Tanning devices–fast track to skin cancer? Pigment Cell Res. 2004;17:2–9. doi: 10.1046/j.1600-0749.2003.00117.x. [DOI] [PubMed] [Google Scholar]
- 29.IARC The association of use of sunbeds with cutaneous malignant melanoma and other skin cancers: a systematic review. Int J Cancer. 2007;120:1116–1122. doi: 10.1002/ijc.22453. [DOI] [PubMed] [Google Scholar]
- 30.Autier P, Dore JF, Lejeune F, et al. Cutaneous malignant melanoma and exposure to sunlamps or sunbeds: an eortc multicenter case–control study in Belgium, France and Germany. Eortc melanoma cooperative group. Int J Cancer. 1994;58:809–813. doi: 10.1002/ijc.2910580610. [DOI] [PubMed] [Google Scholar]
- 31.Autier P, Joarlette M, Lejeune F, et al. Cutaneous malignant melanoma and exposure to sunlamps and sunbeds: a descriptive study in Belgium. Melanoma Res. 1991;1:69–74. doi: 10.1097/00008390-199104000-00009. [DOI] [PubMed] [Google Scholar]
- 32.Bataille V, Boniol M, De Vries E, et al. A multicentre epidemiological study on sunbed use and cutaneous melanoma in europe. Eur J Cancer. 2005;41:2141–2149. doi: 10.1016/j.ejca.2005.04.038. [DOI] [PubMed] [Google Scholar]
- 33.Chen YT, Dubrow R, Zheng T, et al. Sunlamp use and the risk of cutaneous malignant melanoma: a population-based case-control study in connecticut, USA. Int J Epidemiol. 1998;27:758–765. doi: 10.1093/ije/27.5.758. [DOI] [PubMed] [Google Scholar]
- 34.Holly EA, Aston DA, Cress RD, Ahn DK, Kristiansen JJ. Cutaneous melanoma in women. I. Exposure to sunlight, ability to tan, and other risk factors related to ultraviolet light. Am J Epidemiol. 1995;141:923–933. doi: 10.1093/oxfordjournals.aje.a117359. [DOI] [PubMed] [Google Scholar]
- 35.Osterlind A, Tucker MA, Stone BJ, Jensen OM. The danish case–control study of cutaneous malignant melanoma. II. Importance of uv-light exposure. Int J Cancer. 1988;42:319–324. doi: 10.1002/ijc.2910420303. [DOI] [PubMed] [Google Scholar]
- 36.Swerdlow AJ, English JS, MacKie RM, et al. Fluorescent lights, ultraviolet lamps, and risk of cutaneous melanoma. Bmj. 1988;297:647–650. doi: 10.1136/bmj.297.6649.647. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Swerdlow AJ, Weinstock MA. Do tanning lamps cause melanoma? An epidemiologic assessment. J Am Acad Dermatol. 1998;38:89–98. doi: 10.1016/s0190-9622(98)70544-4. [DOI] [PubMed] [Google Scholar]
- 38.Walter SD, Marrett LD, From L, et al. The association of cutaneous malignant melanoma with the use of sunbeds and sunlamps. Am J Epidemiol. 1990;131:232–243. doi: 10.1093/oxfordjournals.aje.a115493. [DOI] [PubMed] [Google Scholar]
- 39.Westerdahl J, Olsson H, Masback A, et al. Use of sunbeds or sunlamps and malignant melanoma in southern sweden. Am J Epidemiol. 1994;140:691–699. doi: 10.1093/oxfordjournals.aje.a117317. [DOI] [PubMed] [Google Scholar]
- 40.Gallagher R. Sunbeds–do they increase risk of melanoma or not? Eur J Cancer. 2005;41:2038–2039. doi: 10.1016/j.ejca.2005.06.011. [DOI] [PubMed] [Google Scholar]
- 41.Geller AC, Colditz G, Oliveria S, et al. Use of sunscreen, sunburning rates, and tanning bed use among more than 10,000 US children and adolescents. Pediatrics. 2002;109:1009–1014. doi: 10.1542/peds.109.6.1009. [DOI] [PubMed] [Google Scholar]
- 42.O’Riordan DL, Field AE, Geller AC, et al. Frequent tanning bed use, weight concerns, and other health risk behaviors in adolescent females (united states) Cancer Causes Control. 2006;17:679–686. doi: 10.1007/s10552-005-0453-9. [DOI] [PubMed] [Google Scholar]
- 43.Cyr WH. Cdrh evaluation of UV-emitting sunlamp products; San Diego, CA. In: Proceedings of the CDC and prevention conference.1999. [Google Scholar]
- 44.Kwon HT, Mayer JA, Walker KK, et al. Promotion of frequent tanning sessions by indoor tanning facilities: two studies. J Am Acad Dermatol. 2002;46:700–705. doi: 10.1067/mjd.2002.119560. [DOI] [PubMed] [Google Scholar]
- 45.Dangerous rays: tanners flock to beaches and booths despite cancer risks. 2003 http://www.abcnews.com.
- 46.Levine JA, Sorace M, Spencer J, Siegel DM. The indoor uv tanning industry: a review of skin cancer risk, health benefit claims, and regulation. J Am Acad Dermatol. 2005;53:1038–1044. doi: 10.1016/j.jaad.2005.07.066. [DOI] [PubMed] [Google Scholar]
- 47.Diffey BL, Farr PM. Tanning with uvb or uva: an appraisal of risks. J Photochem Photobiol B. 1991;8:219–223. doi: 10.1016/1011-1344(91)80062-m. [DOI] [PubMed] [Google Scholar]
- 48.Shuttleworth D. Sunbeds and the pursuit of the year round tan. Bmj. 1993;307:1508–1509. doi: 10.1136/bmj.307.6918.1508. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.National institutes of health summary of the consensus development conference on sunlight, ultraviolet radiation, and the skin. Bethesda, maryland, may 8–10, 1989. Consensus development panel. J Am Acad Dermatol. 1991;24:608–612. doi: 10.1016/s0190-9622(08)80159-4. [DOI] [PubMed] [Google Scholar]
- 50.Cokkinides VE, Weinstock MA, Cardinez CJ, O’Connel MA. Sun-safe practices in US. Youth and their parents: role of caregiver on youth sunscreen use. Am J Prev Med. 2004;26:147–151. doi: 10.1016/j.amepre.2003.10.002. [DOI] [PubMed] [Google Scholar]
- 51.de Vries E, Boniol M, Severi G, et al. Public awareness about risk factors could pose problems for case–control studies: the example of sunbed use and cutaneous melanoma. Eur J Cancer. 2005;41:2150–2154. doi: 10.1016/j.ejca.2005.04.042. [DOI] [PubMed] [Google Scholar]
- 52.Demierre MF. Epidemiology and prevention of cutaneous melanoma. Curr Treat Options Oncol. 2006;7:181–186. doi: 10.1007/s11864-006-0011-z. [DOI] [PubMed] [Google Scholar]
- 53.American academy of dermatology website. 2005 http://www.aad.org/public.
- 54.The national institute of health website. 2005 http://www.cancer.gov/cancertopics.
- 55.Weinstock MA. Public health messages regarding skin cancer. J Invest Dermatol. 2004;123:xvii–xix. doi: 10.1111/j.0022-202X.2004.23524.x. [DOI] [PubMed] [Google Scholar]
- 56.Greenland S. Introduction to regression modeling, chapter 21. In: Rothman KGS, editor. Modern epidemiology. Philadelphia, PA: Lippincott-Raven; 1998. pp. 401–434. [Google Scholar]
- 57.The SAS Institute: Statistical Analysis System (SAS) Cary, NC: 1999. [Google Scholar]
- 58.Stern RS, Nichols KT, Vakeva LH. Malignant melanoma in patients treated for psoriasis with methoxsalen (psoralen) and ultraviolet a radiation (puva). The puva follow-up study. N Engl J Med. 1997;336:1041–1045. doi: 10.1056/NEJM199704103361501. [DOI] [PubMed] [Google Scholar]
- 59.Bataille V, Winnett A, Sasieni P, Newton Bishop JA, Cuzick J. Exposure to the sun and sunbeds and the risk of cutaneous melanoma in the uk: a case–control study. Eur J Cancer. 2004;40:429–435. doi: 10.1016/j.ejca.2003.09.030. [DOI] [PubMed] [Google Scholar]
- 60.Buckel TB, Goldstein AM, Fraser MC, Rogers B, Tucker MA. Recent tanning bed use: a risk factor for melanoma. Arch Dermatol. 2006;142:485–488. doi: 10.1001/archderm.142.4.485. [DOI] [PubMed] [Google Scholar]
- 61.Walter SD, King WD, Marrett LD. Association of cutaneous malignant melanoma with intermittent exposure to ultraviolet radiation: results of a case–control study in ontario, canada. Int J Epidemiol. 1999;28:418–427. doi: 10.1093/ije/28.3.418. [DOI] [PubMed] [Google Scholar]
- 62.Miller SA, Hamilton SL, Wester UG, Cyr WH. An analysis of UVA emissions from sunlamps and the potential importance for melanoma. Photochem Photobiol. 1998;68:63–70. [PubMed] [Google Scholar]
- 63.Spencer JM, Amonette RA. Indoor tanning: risks, benefits, and future trends. J Am Acad Dermatol. 1995;33:288–298. doi: 10.1016/0190-9622(95)90263-5. [DOI] [PubMed] [Google Scholar]