|Home | About | Journals | Submit | Contact Us | Français|
To examine the relationship between tanning and nevus development in very-light-skinned children.
Prospective cohort nested within a randomized controlled trial. Skin examinations in 3 consecutive years (2004, 2005, and 2006) included full-body counts of nevi, skin color and tanning measurement using colorimetry, and hair and eye color evaluation by comparison with charts. Telephone interviews of parents provided sun exposure, sun protection, and sunburn history.
Large managed-care organization and private pediatric offices in the Denver, Colorado, metropolitan area.
A total of 131 very-light-skinned white children without red hair and 444 darker-skinned white children without red hair born in Colorado in 1998.
Full-body nevus counts at ages 6 to 8 years.
Among very-light-skinned white children, geometric mean numbers of nevi for minimally tanned children were 14.8 at age 6 years; 18.8 at age 7 years; and 22.3 at age 8 years. Mean numbers of nevi for tanned children were 21.2 at age 6 years; 27.9 at age 7 years; and 31.9 at age 8 years. Differences in nevus counts between untanned and tanned children were statistically significant at all ages (P < .05 for all comparisons). The relationship between tanning and number of nevi was independent of the child’s hair and eye color, parent-reported sun exposure, and skin phototype. Among darker-skinned white children, there was no relationship between tanning and nevi.
Very-light-skinned children who tan (based on objective measurement) develop more nevi than children who do not tan. These results suggest that light-skinned children who develop tans may be increasing their risk for developing melanoma later in life.
Of all cancers in the United States, cutaneous melanoma ranks sixth in incidence among men and women.1 For those born in 2008, it is estimated that 1 in 55 will be diagnosed with melanoma in their lifetimes,2 an increase from 1 in 250 in 1980 and 1 in 1500 in 1935.3 Several studies have shown that much of the preventable risk for melanoma is established in childhood, with sun exposure and sun protection habits being particularly important.4,5 In Colorado, where the environmental factors of high altitude and 300 or more annual days of sunshine contribute to high UV exposure levels, the melanoma incidence rate is 30% higher than the national average.6
The strongest risk factors for the development of melanoma are the presence of numerous melanocytic nevi and the presence of atypical or dysplastic nevi.7 The number and size of nevi are often used for determining the risk of developing melanoma.8–11 The risk factors for melanoma and factors associated with higher nevus counts are the same: lighter hair color, eye color, and skin color; greater UV exposure; higher frequency and severity of sunburns; male sex; and freckling.9,12–23 Sun exposure resulting in sunburns increases the risk of developing melanoma by as much as 2-fold, depending on the age at which the sunburns occurred.4
However, research results are mixed about whether tanning is risk producing or protective for melanoma. Some research has shown that chronically exposed individuals (eg, outdoor workers) are less likely to develop melanoma than those who have a pattern of intermittent exposure resulting in sunburn.4 Other research has shown that both chronic exposure and intermittent exposure convey risk for melanoma.24,25 Studies on this issue have primarily used case-control designs and thus are hindered by the recall bias inherent in the time span between the childhood exposure period and the onset of the disease.
Children develop nevi rapidly, depending on the amount and intensity of their sun exposure.14,20,26–28 Studies demonstrate that light-skinned children acquire more nevi than dark-skinned children, which puts them at a greater risk of developing melanoma.21,26 While the exact mechanism for the relationship between sun exposure and nevus development is unknown, it has been suggested that UV light may have a direct action on melanocytes and an indirect effect through the immune system12 or through the generation of reactive oxygen species.29 Nicholls30 has proposed that UV radiation causes cellular damage resulting in mutation of melanocytes, which in turn may form nevi. In line with this, the recent discovery that BRAF mutations appear in up to 80% of common acquired nevi suggests that the RAS/RAF/MAPK pathway is a common target for UV-induced mutations that initiate nevus formation.31,32 Melanocytic nevi have been described as clonal amplifications of nevomelanocytes that have undergone oncogene-induced, p16-mediated senescence.33,34 The skin-UV interaction is thus essential to nevus formation, with UV both initiating and potentially promoting nevus development through p53-mediated growth and tanning factor expression.35,36 Since both tanning and mutational activation of nevomelanocytes are consequences of UV exposure, a clear understanding of the relationships between UV exposure, tanning, and nevus development is essential to better characterize the potential causative pathways.
Studies of the relationship between skin pigmentation and nevi generally have been conducted within white populations and have contrasted the number of acquired nevi between light-skinned children and dark-skinned children. One study found that in a white population, children who tan more readily tend to develop fewer nevi than those who do not tan, supporting the argument that tanning is protective.21 Tanning ability and darker base skin color tend to go together, and these studies have not investigated the relationship between tanning and nevi among those with the lightest skin.
The present study explored the relationship between tanning and nevi in a cohort of young children. Because UV exposure causes tanning, and UV exposure is related to the development of nevi, we hypothesized that among the lightest-skinned children, those who tan would develop more nevi than those who do not tan.
This study was reviewed and approved by the Colorado multiple institutional review board. In 2003 and 2004, as part of a randomized controlled trial, a cohort of 1145 children born from January 1998 through September 1998 was recruited from a large managed care organization, private pediatric offices, and community locations in the Denver metropolitan area.23 Skin examinations were conducted in 3 consecutive years: 2004, 2005, and 2006.
Skin color was measured using a Chroma Meter CR-400 (Konica Minolta Sensing Americas Inc, Ramsey, New Jersey), and examiners were blinded to the colorimetry results. The Chroma Meter CR-400 measures across the visible light spectrum using the Hunter Laboratory parameters. The L scale represents the color spectrum from total black to total white. Decreasing values on the L scale (more black) are indicative of darkening skin color.37,38 Following the procedures used by Buller et al,39 a ruler was placed with the 0 mark in the center of the child’s axilla to obtain an inner arm measurement representing untanned skin. Five readings were taken at a point 3 in (7.5 cm) below the axilla. To obtain an outer arm measurement representing tanned skin, a ruler was placed with the 0 mark in the center of the child’s antecubital fossa. The child then rolled his or her arm over so that the palm of the hand was facing down and the forearm was facing up. Five readings were taken on the dorsal aspect of the forearm at an adjacent point 3 inches (7.5 cm) below the antecubital fossa. The average of 5 readings in each location was used. Tanning was defined as the difference between the inner arm and outer arm mean values.
For this analysis, we sought to explore the relationship between tanning and nevi among the lightest-skinned children in this cohort. To establish the subset of children with the lightest skin, we used the untanned skin color measurements from the Chroma Meter CR-400 and examined our data values for the L scale in conjunction with flesh-toned color samples. Higher L values are associated with lighter skin color. Consensus was reached among our group of investigators and dermatologists that those with L scale values above the cut point of 60 would universally be considered light skinned within a white population. Children were then placed into 2 groups, light skinned or dark skinned. To qualify as light skinned, we required them to have L readings of 60 or higher for all 3 years of data collection.
From the cohort of 1145 children, we included only those children who completed all 3 skin examinations (n = 696). Of these, 160 children met the criteria for lightest skin at all 3 skin examinations (L reading ≥60) (Figure). We excluded redheads (n = 20) because numerous previous studies have reported significantly fewer nevi among individuals with red hair compared with all other hair colors in white populations, 12,15,19,20,22,40–44 and because red hair is determined (at least partially) by the MC1R gene, which has been found to suppress nevus development.41 For example, in this cohort of light-skinned children, children with red hair (n = 20) had a geometric mean of 15.9 nevi, compared with 23.8 for blonde hair (n = 18), 29.8 for light brown hair (n = 56), and 33.0 for medium brown hair (n = 57). Because the development of nevi is fundamentally different in redheads, they should be examined separately, which is not possible with only 20 in our cohort. A further concern with the inclusion of redheads in the present analysis is the significantly higher presence of freckling,40 which can obscure the Chroma Meter–based measure of tanning. We also excluded 9 children with incomplete data for nevus counts in 2005 and 2006. The total number of light-skinned white participants available for analysis was 131.
For comparison purposes, we also investigated all white children (Hispanic or non-Hispanic) who did not fall into our lightest-skinned cohort. We applied the same exclusion criteria to this group as for our lightest-skinned cohort (Figure).
Examinations were conducted by a team of dermatologists, pediatricians, and pediatric nurses or nurse practitioners during the summer months (June through early September) of each study year to allow observation of differences in skin pigmentation due to summer tanning. Skin examinations provided full-body melanocytic nevus counts (excluding the buttocks, genitals, and scalp). Nevi were differentiated from freckles and café au lait macules by whether they were raised (only nevi are raised). If the evaluated lesion was flat, it was further identified by other characteristics: early junctional nevi are dark brown, have regular edges, and do not occur in patches, as freckles do.21 This protocol did not distinguish between normal and atypical (or dysplastic) nevi. However, a previous study suggests that such nevi are extremely rare in children.45 Warts were distinguished from nevi by their verrucous nature.21 Congenital nevi were excluded from total nevus counts. As in other studies, plastic stencils were used to measure nevus size. Nevi were recorded as being either raised or flat and coded by size (<2 mm, ≥2 mm to <5 mm, or ≥5 mm). Placement of nevi was recorded on a body map.21,46
Prior to the beginning of each summer skin examination season, nevus counting protocols were reviewed as part of skin examination retrainings conducted by the lead study dermatologist. Interrater reliability was assessed over the duration of each summer period to allow any discrepancies in procedures to be corrected. Interrater reliabilities for examinations conducted in 2004, 2005, and 2006 were 0.89, 0.80, and 0.85, respectively.
In the summer months of 2006, sunburn history, skin phototype (ie, propensity to tan and burn), weekly time spent in the sun, and sun protection behaviors were assessed in a telephone interview of the parent or legal guardian who was the primary caregiver. Sunburn history included whether the child had been sunburned in the past year (regardless of severity). Skin phototype was assessed using a 4-category scale of the subject’s expected reaction to 1 hour of strong sunshine exposure at the beginning of the summer (degree of sunburn the following day/degree of tan 1 week later: painful/none, painful/light, slight/little, none/good).23 The number of hours spent per week in the sun was assessed using 2 variables: the number of days per week the child spent more than 15 minutes outside between 11 am and 3 pm and the usual number of hours spent outside during these occasions. These variables were multiplied to create a single variable representing the total number of hours outside per week between 11 am and 3 pm. A sun protection index variable was created by taking the mean of 4 variables assessing the frequency of using sunscreen, shade, clothes covering most of the body, and hats (5, all of the time; 4, most of the time; 3, about half the time; 2, not very often; 1, never) while outside between 11 am and 3 pm for more than 15 minutes. A score of 3 or above on the scale was considered a high score, and below 3 was considered a low score.
Vacation sun exposure, which has been reported to be related to nevus development in several previous studies, 14–16,47–49 was assessed in annual interviews (2004–2006) by inquiring about any vacations to sunny locations, including the location, season, and length of trip (in days). Using methods established previously in this cohort,47 vacation locations were classified as waterside if they had a body of water and were known to be associated with recreational activities that would lead the participant to get significant sun exposure (eg, surfing, water skiing, boating). Due to seasonal temperatures, within the United States, only Hawaii and locations below latitude 27.5°N qualified as waterside year round. Other water destinations in the United States were deemed waterside only during the summer months of June, July, and August. Under the expectation that a time lag of at least 1 year is necessary for intense sun exposure to stimulate nevus development, vacations over each child’s life were used to create 2 variables: total number of waterside vacations more than 1 year prior to summer 2006 and total number of waterside vacations 1 year or less prior to summer 2006.47,50
We examined the geometric mean number of nevi across all 3 years of data by 1-point increments in tanning based on the L scale (Table 1). We looked specifically for evidence of 3 possible types of relationships between tanning level and the number of nevi: (1) linear or approximately linear; (2) curvilinear, in which those with the lowest and highest levels of tanning had fewer nevi than those with a moderate level of tanning; and (3) a clear jump in the number of nevi at a particular level of tanning. The data suggested a jump in number of nevi at a tanning level cut point of 5 L scale units. We applied this cut point to the 2006 tanning measurement to define 2 groups of children, those who tanned and those who did not tan. To compensate for a skewed distribution, the nevus counts were log transformed.23 We used analysis of variance to verify that the number of acquired nevi (log transformed) was significantly related to tanning using the cut point. We repeated this analysis for darker-skinned white children.
Because previous research has shown that skin phototype is related to both nevus development and melanoma18,21,48,51 and because skin phototype is likely related to and might be confounded by both base skin color and tanning, we examined the relationship between skin phototype and the colorimeter-based measures of skin color and tanning using Pearson correlation coefficient and χ2 analysis. To verify that the relationship between tanning and number of nevi was not due to or confounded by other variables, multiple linear regression was used to simultaneously control for base skin color, hair color, eye color, hours per week in the sun, skin phototype, level of sun protection used when outside, past-year sunburns, and waterside vacation history. A significance level of P < .05 with 2-sided tests was used throughout; SPSS software, version 16 for Windows (SPSS Inc, Chicago, Illinois) was used to perform the analyses.
As summarized in Table 2, for each of the 3 years, the geometric mean number of nevi was significantly higher for children with high tanning vs low tanning: age 6 years, 2004, 21.2 vs 14.8 nevi (P = .02); age 7 years, 2005, 27.9 vs 18.8 nevi (P = .02); and age 8 years, 2006, 31.9 vs 22.3 nevi (P = .03).
When both light- and dark-skinned white children were included, skin phototype was moderately correlated with base skin color as measured by the L scale, with lighter skin color (ie, higher L readings) associated with lower skin phototype (r = −0.32) (P < .001). Skin phototype was also significantly correlated, but less strongly, with tanning (r = 0.17) (P < .001). Among the children we designated as light-skinned children, their parents rated their skin phototypes as type 1, 18%; type 2, 30%; type 3, 44%; and type 4, 10%, which was significantly different from children designated as darker skinned: type 1, 5%; type 2, 22%; type 3, 49%; and type 4, 24% (P < .001). Of the children we designated as both light skinned and low tanners (n = 20), 35% were rated by their parents as skin type 1 (n = 7). Among the light-skinned children that we designated as high tanners (n = 111), 14% were rated by parents as skin type 1 (P < .001).
Among the lightest-skinned children, univariate analysis of risk factors that have been found in other studies to be related to number of acquired nevi (including base skin color, hair color, eye color, hours per week in the sun, skin phototype, sun protection behavior, past sunburns, and vacation history) revealed no significant relationships between any of these variables and the number of nevi (unadjusted P>.05 for all) (Table 3). Because observed tanning is likely a combination of the child’s natural propensity to tan and the amount of sun exposure received, in multiple linear regression analysis we examined whether the relationship between tanning and nevi (log transformed) could be explained by hours per week in the sun, skin phototype, sun protection behavior, past sunburns, and vacation history. After adjustment for all covariates outlined in Table 3, the significant association between tanning level and nevi persisted (P = .005). In addition, blonde hair and higher skin phototype were significantly associated with fewer nevi in the multivariate analysis, while all other covariates remained unassociated with number of acquired nevi. To verify that these results were not specific to the children who had skin examinations in all 3 years, we repeated the analysis on all light-skinned children examined in 2006 (n = 225), and the results were not changed.
For comparison, we applied the same tanning cut point to darker-skinned white children (n = 444) and examined the relationship between tanning and nevi. As summarized in Table 2, there was no relationship between tanning level and number of nevi for the darker-skinned white children. For example, in 2006 at age 8 years, those children who did not tan had a mean of 27.8 nevi compared with 25.2 nevi among those who tanned (P = .31). Also noteworthy is that darker-skinned children who tanned had consistently fewer nevi than the lightest-skinned children who tanned. For example, in 2006 the darker-skinned children who tanned had a mean of 25.2 nevi, whereas the lighter-skinned children who tanned had a mean of 31.9 nevi (Table 2).
Our findings support our hypothesis that very-light-skinned children who tan develop more nevi than those who do not tan, independent of base skin color, hair color, eye color, hours per week in the sun, skin phototype, sun protection behavior, past sunburns, and vacation sun exposure. To our knowledge, this is the first study to examine this relationship in very-light-skinned children, those at highest risk for melanoma. Compared with darker-skinned white children, those with light skin have 2 to 3 times greater risk for melanoma.51,52 The presence of high numbers of nevi has been established as an important marker for melanoma risk.7 Many or perhaps most melanomas may not arise in existing nevi,53,54 suggesting that nevus development is not required for melanoma development. Thus, numerous nevi may be a marker for UV-induced skin damage and/or genetic susceptibility to melanoma.7 Whether nevus development is directly in the pathway for melanoma or a surrogate for melanoma susceptibility, prevention of nevus development may reduce the risk for melanoma.18,21,48 Thus, tanning avoidance may reduce the risk for melanoma in the lightest-skinned, non-redhead children. In darker-skinned white children, we found that tanning had neither a risk nor a protective effect in the development of nevi.
It was somewhat unexpected to find that higher parent-reported skin phototype was protective against nevus development, while observed tanning conveyed risk for nevus development, and these effects were independent. Our measures of skin color and tanning are similar but not equivalent to parents’ ratings of skin phototype. Skin phototype includes both burning and tanning and is based on perception rather than objective measurement, and in some cases parents may have believed that their child would burn and not tan but never actually observed it because of careful sun protection practices such as use of sunscreen, clothing, and sun avoidance. Our analysis focuses on tanning while controlling for sunburn experience as reported by parents. Thus, while correlated with skin phototype, our measures of skin color and tanning provide somewhat different information, and our results regarding these measures are independent of skin phototype, as demonstrated in the multiple regression analysis. This indicates that regardless of a child’s parent-reported propensity to tan, the child’s actual tanning is a risk factor for nevus development.
We did not find evidence for an effect of number of hours spent outside in midday during the summer on nevus development, which was somewhat unexpected. Several previous studies have found an effect for daily sun exposure,16,18,55 but 1 study has not.15 The lack of relationship may be because our measure, which asked parents to estimate the amount of time the child is usually outdoors between 11 am and 3 pm in the summer, lacked the necessary level of sensitivity to discriminate levels of routine sun exposure. It is also possible that intermittent intense UV exposure (rather than routine daily exposure) is required for nevus development, as has been suggested by Autier and colleagues.48 Numerous studies have noted a relationship between vacation sun exposure and nevus development, including an analysis of our full cohort at age 6 years,14–16,47–49 but such a relationship was not detected in this subgroup of light-skinned 8-year-olds. Likewise, sunburns were found to be related to nevus development in previous studies, including our full cohort,12,13,15,16,18,23,48 but this relationship was not found in the present analysis. None of these previous studies focused specifically on the lightest-skinned children.
Prior studies that investigated nevus development in white children used designations such as fair skin, light skin reflectance, or white to describe skin pigmentation. These studies concluded that children with fair skin, light skin reflectance, or white skin acquired more nevi than those with darker skin.21,26 Gallagher et al21 looked at white school children and found that participants who acquired deeper tans tended to have fewer nevi than those who did not tan, suggesting a protective effect of tanning. However, that study asked parents of participants to specify via questionnaire the degree of summer tan (none, light, moderate, or deep) in September. The discrepancy in results between our study and that of Gallagher et al21 may be due to several factors. First, the earlier study included all white children, while we specifically looked at the lightest-skinned white children. Second, in the earlier study, the measure of tanning was based on parents’ perceptions of depth and extent of tanning. Our measure of tanning was objective, using a Chroma Meter CR-400, during the summer months. We have found, anecdotally, that perception of tanning is not highly related to the objective measure of tanning, primarily because those with a lighter base skin color can have a greater difference between their tanned and untanned skin but appear to have a lighter tan than those who started out with a darker base color. Thus, the tanning measure used by Gallagher et al21 reflects the perceived darkness of the tan, while ours reflects a difference between tanned and untanned skin. This phenomenon is present in our own data, as parental reports of greater tanning ability were associated with fewer nevi, while objective measurements of tanning using the colorimeter showed the opposite relationship (Table 3).
Among very-light-skinned children, lack of observable tan may be due to either lack of sun exposure or lack of ability to tan. These 2 factors may further be confounded by a tendency of those who are unable to tan to spend significantly less time in the sun to avoid a sunburn. It is further possible that a subset of those who are unable to tan are also predisposed to develop fewer nevi, and thus the lower number of nevi in these individuals is due to physiologic factors rather than lack of sun exposure. For example, there may be genetic variation in the responsiveness of melanocytes to UV exposure, and in some individuals, UV exposure may stimulate both tanning and nevus development to a lesser degree than in other individuals.
The only way to definitively disentangle ability to tan from amount of UV exposure would be to randomize these lightest-skinned children to receive or not receive sun exposure and determine their skin’s response, including whether or not individual children develop a tan and how this is related to nevus development. There are obvious ethical issues with allowing exposure to UV light to perform such a study. Future studies should use larger samples to verify our findings. This issue should also be explored in populations in other geographic areas where sun exposure patterns may be different. Studies with precise measures of sun exposure, including use of sun protection such as clothing, hats, and sunscreen, could help to further clarify our findings regarding tanning and nevus development, as could use of UV dosimeters.
In conclusion, UV tanning promotes nevus development in non-redhead children with the lightest skin pigmentation. Whether nevus development is directly in the pathway for melanoma development or a surrogate marker for UV-induced skin damage and/or genetic susceptibility to melanoma, our results suggest that tanning avoidance should be considered as a measure for the reduction of melanoma risk in this population.
Funding/Support: This study was supported in part by grant RO1-CA74592 from the National Cancer Institute (Dr Crane).
Role of the Sponsors: The sponsors had no role in the design or conduct of the study; in the collection, analysis, or interpretation of data; or in the preparation, review, or approval of the manuscript.
Author Contributions: Ms Aalborg and Drs Asdigian and Crane had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Aalborg, Morelli, Byers, and Crane. Acquisition of data: Aalborg, Morelli, Mokrohisky, Asdigian, and Crane. Analysis and interpretation of data: Aalborg, Morelli, Mokrohisky, Asdigian, Byers, Dellavalle, Box, and Crane. Drafting of the manuscript: Aalborg, Morelli, Byers, and Crane. Critical revision of the manuscript for important intellectual content: Aalborg, Morelli, Mokrohisky, Asdigian, Byers, Dellavalle, Box, and Crane. Statistical analysis: Aalborg, Byers, and Crane. Obtained funding: Crane. Administrative, technical, and material support: Aalborg, Mokrohisky, Asdigian, Box, and Crane. Study supervision: Morelli, Mokrohisky, and Crane.
Financial Disclosure: None reported.
Additional Contributions: We are indebted to H. Alan Arbuckle, MD, Joanna Burch, MD, Brenda Mokrohisky, RN, Cathi Sommer, MS, CPNP, and Laura Wilson, RN, CPNP, for conducting skin examinations.