SEARCH

SEARCH BY CITATION

Keywords:

  • melanocytic nevi;
  • malignant melanoma;
  • sunburns;
  • pigmentary traits;
  • epidemiology;
  • South European population

Abstract

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In several studies from northern Europe, north America and Australia, melanocytic nevi are correlated with pigmentary traits and with intense sun exposure in a way similar to malignant melanoma. However, it is unclear if these data can be extrapolated to populations in other geographic locations and with different prevalent phenotypes. Our study was conducted among schoolchildren aged 13–14 years in 16 Italian cities. The parents of 3,127 children of a total of 3,160 (99%) consented to our study. A structured questionnaire was used to collect information about sun exposure and lifetime history of sunburns. Children were also examined by trained dermatologists to assess pigmentary traits and to make a count of melanocytic nevi. The median nevus density was higher among boys than girls. Areas that are usually chronically exposed to the sun exhibited a higher density of nevi compared to intermittently and rarely exposed areas. A higher density of nevi was found in children with lighter skin, blond hair and blue eyes. Red-haired children had a remarkably lower nevus density compared to the other color categories. The density of nevi increased with an increased number of reported episodes of sunburns. The results concerning nevi ≥6 mm in diameter paralleled those obtained for the total nevus density. However, at variance with total nevus density, a significant relation was also observed between larger nevi and freckling. Our study confirms that, in Italian schoolchildren, there is a relation between pigmentary traits, history of sunburns and the density of melanocytic nevi. Melanocytic nevi and malignant melanoma share a similar risk factor profile. © 2002 Wiley-Liss, Inc.

A large number of common melanocytic nevi and the presence of large atypical nevi are strongly associated with cutaneous malignant melanoma.1, 2, 3, 4, 5, 6 The biologic correlates of such an association are still poorly understood, but it is plausible that factors that affect the density of melanocytic nevi may also influence the development of malignant melanoma later in life. Risk factors for malignant melanoma have been widely explored in epidemiologic studies and, besides the number of nevi, include pigmentary traits, e.g., light skin and hair color, propensity to sunburns and indicators of intense and intermittent exposure to solar ultraviolet (UV) radiation especially during early life.7, 8 A number of epidemiologic studies confirm that the number of acquired melanocytic nevi is correlated with pigmentary traits, similar to those associated with malignant melanoma, and with intense sun exposure and history of sunburns.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19

Data concerning factors associated with melanocytic nevi have been mostly derived from populations in northern Europe, north America and Australia, and it is unclear if these data can be extrapolated to populations in other geographic locations and with different prevalent phenotypes. The aim of our study was to investigate the density of melanocytic nevi among Italian schoolchildren and to assess constitutional and environmental factors associated with density.

MATERIAL AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Our study was conducted during the spring of 1998 among schoolchildren aged 13–14 years, attending the third class of a number of secondary schools in different Italian cities. The cities were a convenience sample selected according to the presence of a dermatologic centre participating in the clinical network of the Italian Group for Epidemiological Research in Dermatology (GISED). The cities were located in the northern part of Italy, latitude 45.6–44.5 (Milan, Bergamo, Cremona, Ferrara, Verona, Reggio Emilia), middle part, latitude 44.2–43.6 (Florence, Ravenna, Cesena, Ancona, San Marino Republic Perugla) and southern part of Italy, latitude 41.5–37.4 (Neaples, Benevento, Bari, Faggia, Catania). The study was coordinated locally by dermatologist members of GISED. The district directors of education were briefed on the purpose of our study and gave the local GISED coordinator a list of the available schools. Two or 3 schools per district, according to a preliminary agreement by the local study coordinator, were randomly selected from the list, giving a total of 38 secondary schools taking part in our study. Within each school, all the children attending the third class were eligible. The school principals gave permission to conduct the study by contacting the parents and their children. A direct and active involvement of teachers and families was pursued. The project purpose, i.e., to estimate the number of nevi and associated factors, was explained. A standard questionnaire was distributed to parents and written permission to examine children was asked for. Educational materials were distributed to families at the end of our study. The parents of 3,185 children out of a total of 3,218 (99%) consented to our study. There were 3,127 (98%) children of European Caucasoid race and a total of 58 (2%) of other races. Only European Caucasoids were considered for the analyses.

The questionnaire sought information about parents' education, family residence, children's pattern of reaction to first sun exposure and lifetime history of sunburns. Sunburns were defined as episodes of intense erythema with or without blisters causing pain and discomfort for more than 2 days. Children were examined individually by trained dermatologists in the school infirmary. Besides skin examination with an assessment of pigmentary traits, count of melanocytic nevi and evaluation of freckles, height and weight were obtained. An atlas was developed for the recognition of pigmentary lesions and an instrument called a “nevometer,” i.e., a clear piece of plastic with open circles of 2 mm and 6 mm in diameter, was used to determine the sizes of the lesions. Nevi equal to or greater than 2 mm and 6 mm in diameter were recorded separately. Nevi were counted according to predefined anatomic sites. The scalp, genitalia, and, limited to girls, breast areas were not considered. Congenital nevus-like nevi and blue nevi were counted separately and not included in the total nevus count. Freckles, defined as pinpoint brown macules occurring in clusters on exposed body sites, were assessed on the face, upper limbs and shoulders, by comparison with drawings showing schematic patterns and classified according to a 5-point ordinal scale from none to dense. Skin color was evaluated according to a 3-grade scale (dark, medium, fair) based on the examiner's judgement and comparison to representative sample photographs. Eye color was evaluated according to a 5-category scale (black-brown, hazel, brown-green, green-gray, blue) and hair color according to a 5-category scale (black, dark-medium brown, light brown, blond, red). The inter- and intrarater reliability for the judgement about pigmentary traits and nevus counts was assessed in preliminary duplicate exercises and was judged to be excellent with intra- and interobserver intraclass correlation values ranging from 0.70–0.97.

Whole-body nevus counts were calculated by adding the site-specific counts and were expressed both as totals and as counts per unit of body surface area (referred to as “density”). Height and weight were used to estimate body surface area.20 Surface areas of specific body sites were calculated by reference to the charts of Lund and Browder.21 Arithmetic mean, geometric mean and median values were calculated to summarize nevus count data. The distribution of nevus counts was skewed by a small number of high counts, and the adoption of a logarithmic transformation allowed us to better approximate gaussian distribution. Because of a few zero values, the logarithm of 1 plus the nevus count was used. The antilog of the mean of this variable was taken and 1 was subtracted to produce the geometric mean. To analyze correlates and determinants of nevus density adjusting simultaneously for the effect of a number of potential confounding factors, multiple linear regression analysis was performed on the logarithmically transformed nevus counts. Relative geometric means were calculated, together with their 95% confidence intervals (CI), as the ratio of the predicted geometric mean in a specified category to the predicted geometric mean in the baseline category. Estimates of the coefficient of determination (R2) and of p-values of the F test were also given. To evaluate risk factors associated with the presence of nevi 6 mm in diameter or larger, children were separated into 2 categories according to the presence or absence of at least 1 large nevus. Logistic regression-derived odds ratios, together with their 95% CI, were obtained. Different logistic models were fitted. We presented the model including terms for sex and total nevus density only, since adding other potential confounding variables did not materially change the estimates (except for wider confidence intervals).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

There were 1,600 (51.2%) males and 1,527 (48.8%) females. The mean weight was 55.6 Kg (standard deviation [SD] 11.3) in boys and 51.9 Kg (SD 9) in girls and the mean height was 164.2 cm (SD 8.7) and 160.7 cm (SD 6.5), respectively. The mean body surface area was 1.59 (SD 0.21) in males and 1.52 (SD 0.15) in females.

The median count of melanocytic nevi ≥2 mm in diameter was 11 (25th–75th percentiles, 5–22), and the corresponding median density was 7.1 (25th–75th percentiles, 3.2–14.3). The median nevus density was higher among boys than girls (7.72 vs. 6.77) with a relative geometric mean adjusted for phenotypic characteristics of 1.11 (95% CI 1.03–1.19).

The distribution of nevus density according to the anatomic site is shown in Table I. In both sexes, the density was highest on the face followed by the back. The lowest density was observed on the lower limbs. Areas that are usually chronically exposed to the sun, e.g., face, exhibited a higher density of nevi compared to intermittently exposed areas, e.g., trunk, and rarely exposed areas, e.g., palms, soles, inner arms and buttocks.

Table I. Median Values Together With 25th–75th Percentiles of the Density of Melanocytic Nevi ≥ 2 mm at Different Anatomic Sites in Body Areas Grouped According to Their Propensity to Sun Exposure and by Geographic Areas
 Median density (25th–75th percentiles)
BoysGirlsBoth
  • 1

    Chronically exposed areas include the face, neck and dorsal aspects of the hands, intermittently exposed areas include the trunk, the lateral surfaces of upper limbs and the anterior surface of lower limbs, and infrequently exposed areas include buttocks, the medial surface of upper limbs, the posterior surface of lower limbs, palms and soles. Genitalia and the scalp are excluded.

Anatomic site   
 Face-neck14.614.314.4 (0–32.8)
 Upper limbs6.87.57.3 (0–14.6)
 Chest-abdomen6.95.55.9 (1.9–11.0)
 Back10.78.110.0 (0–25.5)
 Lower limbs2.72.82.7 (0–3.8)
Body areas according to their propensity to sun exposure1   
 Chronically exposed areas14.710.113.8 (0–28.4)
 Intermittently exposed areas6.76.16.4 (2.5–13.3)
 Infrequently exposed areas2.62.82.7 (0–8.2)
Geographic areas (latitude)   
 Northern Italy (45.6–44.5)7.86.97.6 (5–25.7)
 Central Italy (44.2–43.6)7.26.66.9 (2–22.4)
 Southern Italy (41.5–37.4)8.37.17.7 (5–26.2)

No correlation was found between the residence and the density of nevi either taking into account the latitude or the geographic location by the seaside. Such a result did not change after adjustment for skin phenotype, i.e., skin color, in a multiple linear regression analysis. For comparison to other studies, Table I shows variations with the median density of nevi according to latitude. There was no clear relation between indicators of socioeconomic status and nevus density (data not shown).

Table II presents the median density of nevi and relative geometric means according to pigmentary traits, modality of sun reaction and history of sunburns. A higher density of nevi was found in children with lighter skin color. Concerning hair color, children with blond hair had the highest density of nevi followed by children with light brown, medium-dark brown and black hair, in that order. Red-haired children had a remarkably lower nevus density compared to the other color categories. Concerning eye color, children with brown/green and blue eyes had the highest density of nevi, followed by green/gray, hazel and black/brown, in that order.

Table II. Median Density and Relative Geometric Mean of Nevi in Relation to Pigmentary Characteristics, Modality of Reaction to Sun Exposure and History of Sunburns Among 3,127 White Children
 No. (%)Median density (25th–75th percentiles)Relative geometric mean1 (95% CI)
  • 1

    Ratio of predicted geometric mean in specified category to predicted geometric mean in baseline category adjusted for center and, in turn, according to all the other variables listed in the table.

Skin complexion   
 Dark429 (13.7)4.67 (2.0–10.0)1
 Medium1,796 (57.4)7.13 (3.3–13.7)1.36 (1.20–1.55)
 Fair902 (28.8)9.13 (4.1–17.6)1.60 (1.36–1.87)
   R2 = 0.0074 p < 0.0001
Eye color   
 Black/brown1,340 (42.9)5.90 (2.5–11.8)1
 Hazel653 (20.9)8.13 (3.7–15.8)1.15 (1.03–1.28)
 Brown/green353 (11.3)8.80 (4.4–16.2)1.20 (1.05–1.37)
 Green/gray415 (13.3)8.26 (3.9–15.0)1.03 (0.90–1.17)
 Blue366 (11.7)9.62 (4.4–18.4)1.19 (1.03–1.38)
   R2 = 0.0028 p = 0.013
Hair color   
 Black295 (9.4)5.84 (2.3–11.4)1
 Dark/medium brown1,480 (47.3)6.36 (2.9–12.9)1.12 (0.97–1.29)
 Light brown988 (31.6)8.22 (3.8–16.2)1.14 (0.97–1.33)
 Blond336 (10.7)10.07 (5.8–19.4)1.30 (1.07–1.57)
 Red28 (0.9)5.25 (1.2–5.3)0.58 (0.37–0.89)
   R2 = 0.0040 p = 0.0012
Burn   
 Never589 (18.8)5.53 (2.5–11.0)1
 Seldom1,208 (38.6)6.97 (3.0–13.2)1.05 (0.95–1.18)
 Sometimes869 (27.8)8.44 (3.9–16.3)1.09 (0.96–1.23)
 Always461 (14.7)8.64 (4.1–18.8)1.13 (0.98–1.32)
   R2 = 0.0007 p = 0.04
Tan   
 Dark1,794 (57.4)6.42 (2.8–12.1)1
 Medium1,266 (40.5)8.57 (4.2–17.5)1.21 (1.11–1.32)
 No67 (2.1)7.59 (3.5–18.8)0.97 (0.74–1.28)
   R2 = 0.0046 p < 0.02
Sunburns   
 None1,377 (42.8)6.40 (2.9–12.3)1
 1815 (26.1)7.26 (3.3–14.7)1.11 (1.01–1.23)
 2673 (21.5)8.31 (3.7–16.3)1.16 (1.04–1.29)
 ≥3302 (9.7)9.46 (4.8–16.9)1.29 (1.11–1.49)
   R2 = 0.0033 p < 0.0001
Freckles   
 None2,449 (78.3)6.63 (2.9–13.5)1
 Few416 (13.3)8.71 (3.5–15.7)1.31 (0.89–1.60)
 Many262 (8.3)7.60 (2.3–16.1)1.01 (0.71–1.30)
   R2 = 0.0002 p = 0.12

Subjects who burned easily on a first exposure to the sun and did not tan had a higher nevus density compared to those who burned infrequently or rarely and tanned easily. Moreover, the density of nevi increased with the number of reported episodes of sunburns. No relation was documented with the year of the first episode of sunburn (data not shown). A weak, nonsignificant relation was observed with the presence of freckles.

A total of 653 (20.9%) children had at least 1 melanocytic nevus ≥6 mm in diameter. No difference in the frequency of these larger nevi was documented between boys and girls (20.6% vs. 21.3%). Table III presents the distribution of children according to the presence of at least 1 melanocytic nevus ≥6 mm in diameter and selected pigmentary traits, modality of reaction to sun exposure and history of sunburns. The results paralleled those obtained for the total nevus density with an increased risk of observing at least 1 melanocytic nevus ≥6 mm in children with lighter pigmentary traits, a tendency to burn easily and to tan with difficulty on sun exposure and according to a history of sunburns. At variance with total nevus density, a significant relation was observed between larger nevi and freckling.

Table III. Distribution of Children According to the Presence of at Least 1 Melanocytic Nevus 6 mm or Larger in Diameter and Selected Pigmentary Traits, Modality of Reaction to Sun Exposure and History of Sunburns
 Nevi ≥ 6 mmOdds ratio1(95% CI)
NoneYes (%)
  • 1

    Multiple logistic regression estimates adjusted for sex and the total density of nevi.

  • 2

    χ2 test for trend p-value.

Skin complexion    
 Dark37455 (12.8)1 
 Medium1,428368 (20.5)1.75(1.0–3.19)
 Fair672230 (25.5)2.33(1.29–2.37)
    p < 0.0012
Eye color    
 Black/brown1,113227 (16.9)1 
 Hazel519134 (20.5)1.26(1.0–1.61)
 Brown/green27875 (21.3)1.32(0.99–1.77)
 Green/gray297118 (28.4)1.95(1.51–2.51)
 Blue26799 (27.1)1.82(1.39–2.38)
    p < 0.0012
Hair color    
 Black25243 (14.6)1 
 Dark/medium brown1,194286 (19.3)1.40(0.99–1.99)
 Light brown762226 (22.9)1.74(1.22–2.47)
 Blond24195 (28.3)2.31(1.56–3.43)
 Red253 (10.7)0.70(0.21–2.42)
    p < 0.0012
Burn    
 Never481108 (18.3)1 
 Seldom976232 (19.2)1.06(0.82–1.36)
 Sometimes668201 (23.1)1.34(1.03–1.74)
 Always349112 (24.3)1.43(1.06–1.92)
    p = 0.0162
Tan    
 Dark1,463331 (18.5)1 
 Medium960306 (24.2)1.41(1.18–1.68)
 No5116 (23.9)1.39(0.78–2.46)
    p = 0.0012
Sunburns    
 None1,044238 (18.5)1 
 1605176 (22.5)1.17(0.94–1.44)
 2553135 (19.6)1.23(0.98–1.54)
 ≥3272104 (27.6)1.40(1.01–1.81)
    p = 0.0072
Freckles    
 None1,980469 (19.3)1 
 Few293123 (29.5)1.33(1.12–1.43)
 Many20161 (23.2)1.00(0.70–1.41)
    p = 0.042

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Our study confirms that the density of nevi and the presence of large nevi (≥6 mm) in children from a Southern European population correlates with pigmentary traits and modality of sun reactions and increases with an increased number of reported sunburns. In addition, our study provides further quantitative evidence for gender and body site variations in the density of nevi with a higher density in boys compared to girls and a higher density of nevi in those body areas that are chronically exposed to the sun. These data largely agree with those obtained in populations from Northern Europe, North America and Australia.9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19

Our study was based on convenience rather than a genuine random sample of Italian cities. However, within the individual cities, schools were chosen randomly and almost all the children were recruited, with a minimum of nonparticipation. In addition, our analyses were based on comparisons of subgroups internal to the study sample. To reduce variability with nevus counts and with information on history of sunburns, we decided to restrict the study to children of 13–14 years. A limitation of our study is its cross-sectional nature, which allows only factors associated with the prevalence rather than the incidence of nevi to be considered, and the retrospective collection of the history of sun exposure. It should be noted, however, that we concentrated on episodes of sunburns, which can be taken as an objective indication that a biologically effective dose of ultraviolet radiation has been received and as a good “after-the-fact” indicator of inadequate sun protection behaviour. Our definition of sunburn involved the existence of pain or discomfort lasting for more than 2 days. Such a specific definition was intended to restrict attention to more severe manifestations, with minor episodes characterised by transient symptoms being missed. It should be recognised that there is a lack of standardised approaches to measuring self-reported or parent-proxy-reported sunburns.22 Our definition of sunburns was similar to those adopted in other studies of nevi and melanoma.16, 23, 24, 25 Our intent was to document variations in nevus density according to a history of sunburns, reducing the recall bias that is expected to mainly influence the report of minor sunburn episodes. We experienced a higher rate of participation in our study compared to similar ones. Our study was integrated into the educational curriculum of the participating schools and an effort was made to actively involve teachers and families with its conduct. Such an effort may have been responsible for the very high response rate we observed.

Recent studies point to genetic-environmental interaction as a plausible explanation for the origin of melanocytic nevi. Clustering of nevi in families26 and higher concordance rate of nevus counts between homozygotic compared to dyzygotic twins have been documented.27 Most children are born without nevi and develop them later in life.13, 28 Data from cross-sectional studies suggest that, while the number of nevi increases progressively with age up to adulthood, the ratio of the number of nevi to the body surface area (usually referred to as “nevus density”) reaches a plateau between ages 9 and 14, pointing to nevus density as a possible genetic trait.10, 12, 14 However, cross-sectional studies, especially if conducted over broad age ranges, may reflect cohort differences in nevus prevalence and degree of exposure to purported risk factors, e.g., ultraviolet light radiation, rather than and in addition to genuine age effects. The relation between the density of nevi in childhood and the density in adult life has not been fully elucidated. In adult life, the rate of regression may affect the total number of nevi.

Gender variations in nevus density, with higher nevus counts in boys compared to girls, have been repeatedly documented.11, 12, 14 Even if discrepant data exist,10, 13 these variations suggest a potential role for hormonal factors. In adults, nevus counts have been greater in women or equal between the sexes,29 and in several countries, including Italy, melanoma incidence (but not mortality) is higher in women than in men. Higher incidence or persistence rates of nevi in adult life among women compared to men may explain these gender differences. To clarify these issues, however, cohort studies are clearly needed.

Among constitutional characteristics, nevus density has been consistently correlated with pigmentary traits and propensity to sunburns.10, 13, 14, 15, 16 In many studies, including the present one, the density of nevi is inversely related to the degree of constitutional skin pigmentation. In addition, children with blue eyes and blond hair tend to have a higher nevus density compared to darker phenotypes. Interestingly, as documented elsewhere,12, 13, 14 although not uniformly,10 our study shows that red-haired children have fewer nevi than other children. This observation points to genetic factors affecting the balance between eumelanins and pheomelanins (the pigment of red hair) as influencing the density of nevi.

Although red hair is rich in pheomelanin, the skin of redheads does not have more pheomelanins than that of people with other hair colors.30 With the exception of red hair, which is associated with an increased risk of melanoma but lower nevus density, the relation between pigmentary traits, propensity to sunburns and the density of nevi parallels risk factors documented for malignant melanoma. At variance with other studies,10, 11, 12, 13, 14 we failed to show an association between freckles and total nevus density, even if we did document such an association when the analysis was restricted to large nevi. Apart from the role of chance, we do not have an explanation for such a discrepancy. A complex relation has been shown between freckling and the density of nevi, with the highest risk documented in those with moderate freckling, and the lowest seen in those with no freckles at one end and heavy freckling at the other.14

The main environmental factor associated with the development of melanocytic nevi is sunlight exposure.10, 12, 13, 14, 15, 16, 17, 18, 19 There are indications that both continuous and intense intermittent exposure are important. Topographic variations in the density of nevi with higher values in body areas chronically exposed to the sun, as documented by us and several other studies,14, 19, 31 point to a role for continuous sun exposure. On the other hand, the relation between nevus density and history and number of sunburns in infancy10, 12, 13 indicates that intense intermittent exposure, a modality that appears to also play a major role in the development of malignant melanoma, may be important as well. In our study, a history of sunburns was associated with the density of nevi, and more than 50% of children reported at least 1 sunburn episode. Previous studies in other countries have documented a relation between nevus density and latitude.15, 16, 17 We failed to document such an association. The lack of correlation may be the combined effect of genetic factors, poorly accounted for by phenotype adjustment, and of variations with attitudes to sun exposure between the northern and southern part of Italy. Interestingly, at variance with other countries of the northern hemisphere where north-south gradients in melanoma mortality rates are observed, in Italy, higher rates have been documented in the north than in the south.32

It has been suggested that the relation between sun exposure and melanocytic nevi may be complex with a parabolic dose-response curve.33 Ultraviolet light radiation may have both a direct action on melanocytes and an indirect one through the immune system. Ultraviolet light radiation has been reported to produce immunosuppression in laboratory animals and in humans34 and it is well known that the density of nevi is elevated in different conditions associated with immunosuppression, e.g., childhood leukemia and organ transplantation.35, 36 Acute and reversible changes in the clinical, dermatoscopic and histologic aspects of melanocytic nevi have been repeatedly documented after intense ultraviolet light radiation. Irradiated nevi became more irregular, darker in color, with an increased number of suprabasally located melanocytes, indicating activation of these cells.37, 38, 39

To summarize, our study confirms that in Italian schoolchildren there is a relation between pigmentary traits, history of sunburns, and the density of melanocytic nevi. The above-mentioned factors are similar to those associated with malignant melanoma in the same population,6 indicating that melanocytic nevi and malignant melanoma share a similar risk factor profile. Since the number of melanocytic nevi is the strongest indicator of the melanoma risk, preventive modalities aimed at reducing the density of nevi should, in principle, result in a reduced incidence of malignant melanoma. Educational interventions to promote safe sun exposure in young children should be evaluated as a way to reduce nevus density and, in the long run, to prevent malignant melanoma in adult life.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIAL AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES