To investigate the body-site distribution of melanocytic nevi (MN) with respect to habitually and intermittently sun-exposed surfaces.
Cross-sectional survey of MN prevalence.
Townsville (19.16°S), Queensland, Australia.
Random sample of 506 1- to 6-year-old white children who were born and raised in Townsville (response, 87.6%).
Main Outcome Measures
Site-specific counts and densities (number per square meter) of MN.
Densities of MN of all sizes were highest on the outer forearms, followed by the outer upper arms, neck, and face. The feet had the lowest density of MN. Densities of MN of 2 mm or greater were highest on the upper arms and trunk. Boys had higher densities of MN of all sizes on the neck than girls (P = .002). Girls had higher densities of MN of 2 mm or greater on the lower legs (P=.006) and thighs (P = .005) than boys. Habitually sun-exposed body sites had higher densities, particularly of small MN, than relatively sun-protected sites, and larger MN were most prevalent on the intermittently exposed skin of the trunk.
These children have higher total body and site-specific MN counts and densities than children from elsewhere, and their MN are distributed over the body in a way that implicates exposure to sunlight. As sun exposure in childhood and MN are risk factors for melanoma, intervention studies are required to determine if MN can be prevented.
THE NUMBER of melanocytic nevi (MN) is an important risk marker for cutaneous melanoma (CM).1 Although their causative role is uncertain, MN may be risk markers sharing similar causative factors with CM, or direct precursors for a substantial proportion of CMs.2-6Epidemiological studies1-5 suggest that sun exposure is a common causative factor for both MN and CM. While total lifetime or intermittent sun exposure is difficult to estimate for adults, sun exposure during childhood or adolescence has been confirmed as a causative factor for CM,7-10 with MN counts in children increasing with the amount and intensity of sun exposure.5,11-13
Attempts have been made to determine whether MN and CM have similar body-site distributions,6,14-16 but have been inconclusive. Studies4,14,16 in temperate climates have shown that MN counts increase in the first 2 decades, peak in the third, and decline after the fourth decade. In contrast, incidence rates for CM are highest from age 50 years and older.17Thus, the body-site distribution of MN in patients with CM may not reflect the site distribution relevant for the development of CM, as initiation of the disease may date back to childhood or adolescence.
This study provides baseline information on the body-site distribution of MN in young Australian children raised in an intense UV radiation environment. Analysis focuses on sun-exposed and sun-protected body sites in relation to models of the natural history of CM and MN18-20 and the body-site distribution of the latter in children.
In 1991, 516 one- to six-year-old children who were born and raised in Townsville (19.16°S), Australia, were included in a cross-sectional study of MN prevalence (response, 87.6%). All children had white parents, but 10 children who had 2 or more non-European grandparents were excluded from analysis. Further details of this sample are published elsewhere.5 The 506 children (259 boys, 247 girls) were examined for MN according to a standard international protocol.21 Thirty different body sites that differentiated between sun-exposed and less-exposed areas were distinguished. The scalp, buttocks, and genitals were not examined. Informed consent was given by the children's mothers, who also completed a standardized questionnaire regarding their child's ancestry and sun exposure.
Melanocytic nevi were defined as brown- to black-pigmented macules or papules of any size, darker in color than the surrounding skin, excluding lesions with the clinical characteristics of freckles, solar lentigines, or café au lait spots.21 As it is not possible to distinguish between MN and lentigo simplex clinically, and biopsy is not feasible in children, lentigo simplex may have been misclassified as MN. Heavy freckling may lead to undercounting of MN that may be difficult to distinguish in an area of freckling. However, these caveats apply to all observational studies of MN.
Skin-colored palpable lesions with the morphological features of compound or intradermal MN, halo nevi, nevi spili, congenital nevuslike nevi, and blue nevi were counted separately and are included in total MN counts. Café au lait spots were recorded separately. The distribution of freckling was assessed on the face, arms, and shoulders using a standardized semiquantitative scale.
Melanocytic nevi were measured with a clear plastic film imprinted with circles of 2-, 3-, 4-, and 5-mm diam-eters attached to an illuminated (×3) magnifying glass. Lesions were measured with the skin unstretched and judged to be of a specific size if the lesion touched both sides of the circle. All children were examined by one of us (S.L.H.), who was previously trained in the recognition of MN by dermatologists.13 Replicate counts on 9% of the subjects showed 93% intraobserver reliability.
Height and weight of the children were measured to estimate the total surface area of the body.22 Proportions of total surface area attributed to the head, trunk, and upper and lower extremities were based on estimations by Boyd23 and modified according to results of Berkow24 and Lund and Browder25 to allow further differentiation of body sites (Table 1). Surface proportions of the inner and outer upper arms and forearms, and the anterior and posterior thighs and lower legs were calculated under the assumption that each subsite covers 50% of the surface proportion of the upper arms, forearms, thighs, and lower legs. To compare the effects of exposure at different sites, it is necessary to compare numbers of MN per unit area (density). Melanocytic nevus densities were calculated as follows:
For total body MN densities, adjustments were made to account for the scalp, buttocks, and genitals, reducing the body surface area on average by 14.1% (Table 1).
Since total and site-specific MN counts and densities were positively skewed, median values and interquartile ranges are given. Sex differences were assessed with nonparametric Wilcoxon-Mann-Whitney tests.26 Multiple linear regression analysis was performed with log-transformed MN counts and densities to test for influence of age (in months) and sex. Differences in site-specific counts and densities of MN were assessed by nonparametric-paired Wilcoxon and Friedman tests.26 Statistical analysis was performed using SPSS for Windows, Version 6.1.3 (SPSS Inc, Chicago, Ill). A significance level of .05 was assumed.
Of the 506 children, 98.8% had MN. The 2 boys and 4 girls without MN were aged between 1 and 2 years. Forty-eight of 103 children between 1 and 2 years of age had MN of 2 mm or greater. Total MN counts, and counts of MN of 2 mm or greater, increased with age (P<.001, respectively) (Table 2) and there was no significant sex difference in either count.
Counts of MN of all sizes and MN of 2 mm or greater counted separately on the face, neck, trunk, upper arms, forearms, hands, feet, lower legs, and thighs increased steadily with age (P<.001, respectively)(Table 2). Across all ages, counts of MN of all sizes were highest on the trunk, followed by the upper arms, thighs, and forearms (overall test comparing body sites: P<.001, respectively, for MN of all sizes and ≥2 mm). Hands and feet had fewer MN than other sites, but only comprise a small proportion of the body's total surface area. Boys had more MN of all sizes on the neck than girls (P<.05). In contrast, girls had more MN of 2 mm or greater on the lower legs (P<.05) and thighs (P<.01). These sex differences remained statistically significant after adjusting for age (P<.01, respectively). More girls had MN of 2 mm or greater on the lower legs (43.3% vs 33.2%) and thighs (51.0% vs 36.7%) than boys. In total, 73 children had 82 MN of 5 mm or greater. Most large MN were on the trunk (35 [42.7%] of 82), arms (15 [18.3%] of 82; 8 on forearms and 7 on upper arms), and thighs (12 [14.6%] of 82).
Overall, densities of MN of all sizes were highest on the forearms, upper arms, neck, and face. The feet had the lowest MN densities (overall test, P<.001); (Table 3; Figure 1). The highest densities of MN of 2 mm or greater were found on the upper arms and trunk (overall test, P<.001). The results of MN density considered by age and sex reflect those for MN counts. That is, densities of MN of all sizes, and 2 mm or greater, increased significantly with age for all body sites (P<.001, respectively); densities of MN of all sizes on the neck were higher for boys than girls (P<.01, adjusted for age), particularly the back of the neck (P<.005); and densities of MN of 2 mm or greater on the lower legs and thighs were higher for girls than boys (P<.01, respectively, adjusted for age). Densities of MN of all sizes, and 2 mm or greater, were highest (among whole body sites) for the neck of 5- and 6-year-olds. Only 1 of the 6-year-olds had no MN on the neck, 22.2% had no MN of 2 mm or greater on the neck, 92.6% had a neck-specific density of 100 or greater for MN of all sizes, and 38.9% presented with a neck-specific density of 100 or greater for MN of 2 mm or greater.
The median number of MN of all sizes on the inner upper arms as well as the inner forearms (median, 1, respectively) was significantly lower than on the outer upper arms (median, 4) and outer forearms (median, 3) (P<.001, respectively). There was no significant difference between the anterior and posterior lower legs (median, 1; P = .09, respectively). However, the difference between the anterior (median, 3) and posterior (median, 1) thighs was highly significant (P<.001). Densities of MN of all sizes, and 2 mm or greater, were significantly lower on the inner upper arms and inner forearms than on the outer upper arms and outer forearms (P<.001, respectively) and were significantly lower on the posterior lower legs and thighs than on the anterior lower legs and thighs (lower legs: all sizes, P<.05; ≥2 mm,P<.01; thighs: all sizes, P<.001; ≥2 mm,P<.001) (Table 4). The density of MN of all sizes was also higher on the dorsal surface of the feet (median, 0; mean, 32.3) and hands (median, 47.2; mean, 65.0) compared with the soles of the feet (median, 0; mean, 1.3), and the palms of the hands (median, 0; mean, 5.3) (P<.001, respectively). Densities were highest on the outer forearms and outer upper arms, consistently for all ages (Table 4). One 6-year-old girl had 31 nevi on the outer upper arm and 24 nevi on the outer forearm, a surface equivalent to 6.7% of the total surface area at this age.
These children had extremely high total body MN counts, with the highest MN densities occurring on the outer forearms, outer upper arms, neck, and face, all of which are highly exposed to the sun. Children from Townsville have total body and site-specific MN counts and densities that are higher than those reported for children the same age or slightly older from other Australian states13,27 and abroad,11,12,28-30 even though the prevalence of MN is similar at birth.31,32 This may be partly explained by geographic differences in ambient solar UV radiation and annual temperature variation, which influences the length of the season during which clothing offering little protection from the sun is worn. Using the DNA Action Spectrum of Setlow33 and monthly ratios for cloud cover,34 the calculated average annual total DNA-weighted UV radiation received in Townsville (19°S, 1.91 × 106 J/m2) is higher than for other Australian cities (Perth, 32°S; 1.48 × 106 J/m2; Sydney, 34°S; 1.1 × 106 J/m2) where MN surveys have been conducted13,27 (Aurel Moise, MSc, written communication, 1997).
Local climatic conditions influence the type of clothing worn. Townsville experiences minimal fluctuations in daily and annual temperatures, and has an average temperature range of 15° to 26°C in winter and 24°C to 31°C in summer (Australian Bureau of Meteorology, written communication, April 1997). Consequently, people mostly wear lightweight summer clothing that exposes the forearms, neck, and most or all of the upper arms and legs. As in children from Perth,27 the sites with the highest MN densities are those exposed to the sun in summer clothing and not protected by the structure of the human body, or its movement. The arms are often held so that the medial surface of the upper limb is protected by the trunk, while the lateral surface is exposed. When sitting, the posterior thighs and the soles of the feet are shielded from the sun, whereas the anterior thighs may be exposed (depending on clothing). In keeping with this pattern, we found more MN on the anterior than the posterior surface of the lower limbs that was caused by an excess of MN on the anterior thighs compared with the posterior thighs, rather than the lower legs where there was little difference in MN density between the anterior and posterior surface. Likewise, MN were less concentrated on the soles than on the dorsa of the feet. Also, boys had significantly more MN on the posterior neck than girls that may be because of girls having longer hair.
Apparent anomalies in nevus densities on the anterior and posterior lower legs may be attributable to these sites having equal opportunities for exposure when standing and perhaps even when sitting. The only markedly different opportunity for sun exposure occurs when lying down. However, sunbaking is a relatively rare pastime in Townsville, particularly for children. Melanocytic nevi were more concentrated on the dorsum than the palm, but as reported by others,27 there were fewer MN than expected on this heavily exposed surface. This deviation from the sun-exposure hypothesis is more difficult to explain. Site-specific CM rates in Queensland35 and other populations36,37 also show few cases at this site. Green and MacLennan35 suggest that this may be due to a protective factor or a differential susceptibility according to site. The fact that in Queensland the proportion of CMs with adjacent nevi at different body sites is not readily explained by site-specific MN densities19 suggests that MN are more likely to be risk markers for CM than obligatory precursors.
The model proposed by Armstrong18 suggests 3 pathways linking a normal melanocyte to CM: via a melanocytic nevus cell; via an MN; or by some other undefined mode. For the first 2 pathways, nevus cells are first-step mutations in a sequence thought to be triggered by exposure to UV radiation in a susceptible host. Our results provide support for Armstrong's pathway from melanocyte to MN, which is said to result from mutations caused by exposure to sunlight, followed by a period of clonal expansion as a result of the same. However, the model states that not all altered melanocytes (nevocytes) will undergo sufficient clonal expansion to produce a visible nevus before undergoing malignant transformation, and nevocytes can be destroyed by biological mechanisms or lose their ability to progress. Thus, the nevus cell pathway explains why precursor lesions cannot be identified for all melanomas.18
As we found that large (≥5 mm) and moderately large (≥3 mm) MN were concentrated on the intermittently exposed skin of the trunk and that small MN were concentrated on habitually exposed body sites, we agree with Richard and coworkers20 who suggest that intense episodic sun exposure, characteristic of intermittently exposed sites, causes small MN to grow. We propose that acute episodic exposure of unprotected skin may cause MN to grow in diameter, and that the degree of exposure for MN growth may be higher than the dose causing proliferation of melanocytes to form a nevus. Intense sun exposure is more likely in environments with high ambient UV radiation and may explain why large MN are more concentrated on the trunk of children from Townsville than from elsewhere.38 Perhaps larger MN are less prevalent on habitually exposed sites because the melanocytes are protected by the thickening of the stratum corneum and/or deposition of melanin in children who can tan. It also is conceivable that a sufficient dose of UV radiation could be received by the melanocyte, through tanned skin, in individuals who spend a lot of time outdoors in environments with intense solar UV radiation.
The development of atypical features in a susceptible benign nevus may also be associated with an accumulation of mutations in response to sunlight exposure in a susceptible host18 or may be associated with a differential site-dependent susceptibility of the melanocytes to dysplasia and/or malignant transformation as proposed by Green.19 The latter could be incorporated into the host's response in the model of Armstrong18 to explain why atypical nevi are found primarily on the trunk of older fair-skinned children from Townsville38 and why a high proportion of nevus-associated melanomas are located on the back.19
Despite a lack of congruence in all aspects of the anatomical distribution of CM in adults and MN in children, the similarities deserve consideration. In Queensland, MN and CM are common on several body sites that are always exposed to the sun.35 In 1987, the highest density of invasive CM occurred on the ears in men and the face in women, followed by the shoulders for both sexes.35 High densities of MN were found on the face (including ears) of children born in Townsville between 1985 and 1990, although the densities were not as high as they were for the forearms, upper arms, and neck. Nevertheless, the predilection of MN for the forearms of children is similar to the increase in incidence of CM at this site between 1979/1980 and 1987.35 Melanocytic nevi of all sizes were more prevalent on the neck of boys than girls. Similarly, over the same period, the incidence of CM on the neck doubled in males and declined significantly in females.35Girls had more MN of 2 mm or greater on the lower limbs than boys. Significant increases were also seen for incidence of CM on the thighs of women.35 While we are forced to compare CM incidence in adults and MN prevalence in children, the similarities that exist may reflect similar sun-related behaviors.
Our study is unique because it focuses on young white children living in an intense solar UV radiation environment. Our results suggest that MN develop preferentially on maximally sun-exposed sites in young children and bear some resemblance to the site-specific incidence of CM in Queensland.35 As sun exposure in childhood39and MN are risk factors for CM,1 the development of MN should be prevented as much as possible. Intervention studies are required to determine whether a significant proportion of MN can be prevented by reducing sun exposure in early childhood.
Accepted for publication June 18, 1998.
This work was supported by the Queensland Cancer Fund, Queensland, Australia.
We thank Rick Speare, PhD, Madeleine Nowak, and Lynne Raw for critical revision of the manuscript; Aurel Moise, MSc, for ambient UV radiation data; Christine Bailiff for assistance with data collection; the staff of the Townsville Mater Hospital and the Kirwan Hospital, Townsville, Australia, for Women for their assistance with subject recruitment; and in particular, we thank the participating children and their families, without whom, none of this research would have been possible.
Reprints: Simone L. Harrison, MPH, School of Public Health and Tropical Medicine, James Cook University, Townsville, Queensland, Australia 4811 (e-mail: email@example.com).
DR Epidemiologic studies. Balch
S-Jeds. Cutaneous Melanoma: Clinical Management and Treatment Results Worldwide
2nd ed. Philadelphia, Pa JB Lippincott1992;12- 26Google Scholar
P Melanocytic nevi in schoolchildren in Queensland. J Am Acad Dermatol.
1989;201054- 1060Google ScholarCrossref
et al. Suntan, sunburn and pigmentation factors and the frequency of acquired melanocytic nevi in children: similarities to melanoma: the Vancouver Mole Study. Arch Dermatol.
1990;126770- 776Google ScholarCrossref
et al. Associated factors in the prevalence of more than 50 common melanocytic nevi, atypical melanocytic nevi, and actinic lentigines: multicenter case-control study of the Central Malignant Melanoma Registry of the German Dermatological Society. J Invest Dermatol.
1994;102700- 705Google ScholarCrossref
I Sun exposure and melanocytic nevi in young Australian children. Lancet.
1994;3441529- 1532Google ScholarCrossref
CE Epidemiologic evidence for the role of melanocytic nevi as risk markers and direct precursors of cutaneous malignant melanoma: results of a case-control study in melanoma patients and nonmelanoma control subjects. J Am Acad Dermatol.
1992;26920- 926Google ScholarCrossref
N Sun exposure habits in patients with cutaneous melanoma: a case control study. J Dermatol Surg Oncol.
PJ Relationship of cutaneous malignant melanoma to individual sunlight-exposure habits. J Natl Cancer Inst.
et al. Nonfamilial cutaneous melanoma incidence in women associated with sun exposure before 20 years of age. Pediatrics.
et al. Increase of melanocytic nevus counts in children during 5 years of follow-up and analysis of associated factors. Arch Dermatol.
1996;1321473- 1478Google ScholarCrossref
V Nevi in schoolchildren in Scotland and Australia. Br J Dermatol.
1994;130599- 603Google ScholarCrossref
BJ Sunlight: a major factor associated with the development of melanocytic nevi in Australian schoolchildren. J Am Acad Dermatol.
1994;3040- 48Google ScholarCrossref
G Acquired melanocytic nevus: epidemiologic clinical study of a healthy population. Giorn Ital Dermatol Venereol.
1990;125231- 236Google Scholar
et al. Overall and site-specific risk of malignant melanoma associated with nevus counts at different body sites: a multicenter case-control study of the German Central Malignant Melanoma Registry. Int J Cancer.
1995;62393- 397Google ScholarCrossref
P The number and distribution of benign pigmented moles (melanocytic nevi) in a healthy British population. Br J Dermatol.
1985;113167- 174Google ScholarCrossref
J Cancer Incidence in Five Continents. Vol. 6 New York, NY Oxford University Press Inc1992;
BK The epidemiology of melanoma: where do we go from here? Gallagher
JMeds. Epidemiological Aspects of Cutaneous Malignant Melanoma
Norwell, Mass Kluwer Academic Publishers1994;307- 309Google Scholar
A A theory of site distribution of melanomas: Queensland, Australia. Cancer Causes Control.
1992;3513- 516Google ScholarCrossref
et al. Role de l'exposition solaire sur les nevus melanocytaires bénins: première étude dans des populations controlées pour l'âge, le sexe et le phenotype. Ann Dermatol Venereol.
1994;121639- 644Google Scholar
et al. Epidemiological Studies of Melanocytic Nevi: Protocol for Identifying and Recording Nevi. Lyon, France International Agency for Research on Cancer1990;Internal report No. 90/002
EF A formula to estimate the approximate surface area if height and weight be known. Arch Intern Med.
1916;17863- 871Google ScholarCrossref
E The Growth of the Surface Area of the Human Body. Minneapolis University of Minnesota Press1935;
SG A method of estimating the extensiveness of lesions (burns and scalds) based on surface area proportions. Arch Surg.
1924;8138- 148Google ScholarCrossref
NC The estimation of area of burns. Surg Gynecol Obstet.
1944;79352- 361Google Scholar
DG Practical Statistics for Medical Research. New York, NY Chapman & Hall1991;
BK Melanocytic nevi in children, I: anatomic sites and demographic host factors. Am J Epidemiol.
1994;139390- 401Google Scholar
F Frequency of acquired melanonevocytic nevi and their relationship to skin complexion in 939 schoolchildren. Dermatologica.
1989;179123- 128Google ScholarCrossref
D Benign pigmented nevi in children from Kidderminster, England: prevalence and associated factors. J Am Acad Dermatol.
1990;22747- 750Google ScholarCrossref
et al. Anatomic distribution of acquired melanocytic nevi in white children: a comparison with melanoma: The Vancouver Mole Study. Arch Dermatol.
1990;125466- 471Google ScholarCrossref
LB The incidence and significance of birthmarks in a cohort of 4,641 newborns. Pediatr Dermatol.
1983;158- 68Google ScholarCrossref
C A prevalence survey of dermatoses in the Australian neonate. J Am Acad Dermatol.
1990;2377- 81Google ScholarCrossref
RB The wavelength in sunlight effective in producing skin cancer: a theoretical analysis. Proc Natl Acad Sci.
1974;713363- 3366Google ScholarCrossref
IJ Erythemal ultraviolet radiation over Australia: calculations, detailed results and input data including frequency analysis of observed Australian cloud cover. CSIRO Australian Division of Atmospheric Physics.
Melbourne, Australia CSIRO1978;1- 48Techinical paper No. 33Google Scholar
R Etiological clues from the anatomical distribution of cutaneous melanoma. Gallagher
JMeds. Epidemiological Aspects of Cutaneous Malignant Melanoma
Norwell, Mass Kluwer Academic Publishers1994;67- 78Google Scholar
W Time trends in malignant melanoma of the upper limb in Connecticut. Cancer.
1991;681854- 1858Google ScholarCrossref
et al. The Eastern Australian childhood nevus study: prevalence of atypical nevi, congenital nevus-like nevi, and other pigmented lesions. J Am Acad Dermatol.
1995;32957- 963Google ScholarCrossref
BK Cutaneous malignant melanoma and indicators of total accumulated exposure to the sun: an analysis separating histogenic types. J Natl Cancer Inst.
1984;7375- 82Google Scholar