close

Вход

Забыли?

вход по аккаунту

?

632

код для вставкиСкачать
Int. J. Cancer: 67,485-491 (1996)
0 1996 Wiley-Liss, Inc.
Publicationof the InternationalUnion Against Cancer
Publicationde I Union InternationaleContre le Cancer
NAEVI AND PIGMENTARY CHARACTERISTICS AS RISK FACTORS
FOR MELANOMA IN A HIGH-RISK POPULATION:
A CASE-CONTROL STUDY IN NEW SOUTH WALES, AUSTRALIA
Andrew E. GRULKH’.~,Veronique BATAILLE?,
Anthony J. sWERDLOW3,Julia A. NEWTON-BISHOP4, Jack CUZICK~,
Peter HERSEY~
and William H. MCCARTHY’
‘Depai-tment of Public Health, University of Sydney, Sydney, Australia; ?St. Georges Hospital; 3London School of Hygiene
and Tropical Medicine; 4Royal London Hospital Medical College; 51mperial Cancer Research Fund, London, UK;
6Royal Newcastle Hospital, Newcastle; ’Royal Prince Alfred Hospital, Sydney, Australia.
The relationship between risk of cutaneous malignant melanoma and total body and site-specific naevus counts and other
host factors was investigated in a Caucasian population aged
15-84 years in New South Wales, Australia. The study sample
comprised 244 cases with melanoma diagnosed in 1989-1 993,
and 276 controls. The strongest relationshipwas with total body
naevus count. Risk of melanoma was raised I 2 times in those
with more than 100 naevi compared with those with less than
10. There were also strong risks, with odds ratios of 5 or more,
associated with having multiple atypical naevi, multiple large
naevi, high naevus counts in sun-exposed or sun-protected
areas and being unable to tan on repeated sun exposure. The
effect of inability to tan was stronger at younger than older ages.
Lesser risks, with odds ratios of 2-3, were associated with being
prone to burn on sun exposure, having many freckles as a child
and having red hair. The site distribution of naevi in males
compared with females resembled the distribution of melanoma by sex. Risk of melanoma of the back was significantly
more closely related to back naevus count than naevus count for
the remainder of the body. For other anatomical sites, naevus
count was non-significantly more closely related to naevus
counts at that site than counts over the remainder of the body.
Naevus count declined with age in both cases and controls. In
those aged under 40, having I00 or more naevi was associated
with an aetiological fraction (AF) of 41%. In those aged 60 and
over, however, the AF associated with this number of naevi was
only 5%.
o 1996 Wiley-Liss,Inc.
Melanoma is the third most common cancer in both men
and women in NSW (Coates et al., 1995). Over the past 10
years, case-control studies in Europe (Swerdlow et al., 1986;
0sterlind et al., 1988; Garbe et al., 1989; Augusstson et al.,
1991), and the United States (Holly et al., 1987) have revealed
that a large number of benign melanocytic naevi on the body
surface is a strong risk factor for melanoma. Most of these
studies were performed in areas of low environmental ultraviolet (UV) radiation exposure. U V radiation exposure is a
known risk factor for melanoma and is also probably related to
the development of melanocytic naevi (Harrison et al., 1994).
The relationship between naevi and melanoma in Australia,
areas of which have the highest rates of melanoma in the world
(MacLennan et al., 1992), has been shown with limb naevus
counts (Holman and Armstrong, 1984; Green et al., 1985), but
the only study that counted naevi on most of the body surface
found a weak relationship with total naevus count and did not
examine the relationship of site-specific counts to risk of
melanoma (Nordlund et al., 1985). We performed a casecontrol study of melanoma in the region of Sydney, Australia.
POPULATION AND METHODS
Cases were consecutively presenting white melanoma patients during 1992-1993 at the Sydney and Newcastle melanoma clinics. They were aged 15-84 years and were first
diagnosed with melanoma in 1989 or later. Sydney and
Newcastle are coastal cities, situated close to each other
between latitudes 33 and 34 degrees South, with sub-tropical
climates. In all patients the pathology slides were reviewed by a
specialist pathologist and the diagnosis of melanoma confirmed. The study comprised a total of 244 cases. Refusals to
participate were rare. Two patients otherwise eligible were
excluded because they were non-white.
Two hundred and seventy-six white controls aged 15-84
years were selected from outpatient clinics in the same
hospitals as the cases, and from general practitioners’ surgeries
in the catchment areas of the hospitals; they had non-chronic
conditions judged to b e unrelated to sun exposure. Thirty-five
percent were attending hospital clinics, and the largest sources
were ear, nose and throat (9%), rheumatology (8%) and
orthopaedic (8%) clinics. Forty-six percent were attending
their general practitioner, and 20% were friends o r relatives of
patients (with non-sun exposure-related conditions) attending
a hospital or general practitioner. Controls were selected so as
to have a similar age, sex and ethnicity distribution to cases.
Fewer than 20% of potential controls refused to participate.
Data were collected by one of two medical practitioners, a
dermatologist (V.B.) and a medical epidemiologist (A.G.) who
had been trained by the dermatologist. As one of the primary
aims was to enable comparison with a similar study being
conducted in England (Newton et al., 1993), the naevus
counters had either participated in the English study (V.B.) or
been trained at it (A.G.).
A questionnaire was administered eliciting a range of
information about risk factors for melanoma. Ethnicity was
determined by the countries of birth of the patients and their
parents and grandparents, as well as their predecessors if the
grandparents had been born in Australia. Patients were asked
whether they would tan if they exposed themselves to the early
summer sun for several days. Patients could usually estimate
their tanning ability with confidence, because a high level of
sun exposure is a part of daily life in New South Wales. They
were asked if they would burn if they went out into the sun, at
midday, with no sun protection, for 1 hr, in early summer
before they had acquired a tan. They estimated their natural
hair colour at the age of 15 years, as well as freckling as a 15
year old from a series of diagrams of the face. The rating of
freckling was on a scale of 0-100 but was reduced to 4
categories for analysis. Other data were sought on sun and
artificial U V radiation exposure, and these will be examined
elsewhere.
Patients were then examined for naevi on the whole body
except the genitalia, breasts (in women), and posterior scalp.
Raised or flat naevi with a largest diameter of 2 mm or more
were counted. Naevi from 5 to 9 mm and 2 1 cm in diameter
were tabulated separately. Atypical naevi were defined as
being 2 5 mm in diameter, with an irregular or blurred edge
BTo whom correspondence and reprint requests should be addressed. at the National Centre in HIV Eoidemioloev and Clinical
Research, 376 Victoria Street, Darlinghurs’t, NSW a10,Australia.
Fax: (61)2 332 1837.
Received: January 11, 1996 and in revised form April 1,1996.
486
GRULICH ETAL.
and irregular pigmentation. A naevus was termed congenital if
the patient recalled it being present in early childhood and if it
was 2 1 cm in diameter and had a well-defined even border
and colouration. Each type of naevus was counted in 19
separate body sites. Eye colour was noted, and pigmented
lesions of the iris were counted but were not confirmed by slit
lamp examination.
Formal testing of repeatability of naevus counting was
undertaken on a sample of 15 patients, and the inter-observer
correlation of the total naevus count was 0.88 (95% CI = 0.311.0, Fisher’s z transformation).
Data entry and initial tabulations were performed with
Epi-Info. All odds ratios (ORs) presented were calculated by
unmatched logistic regression. They were adjusted for age, sex
and ethnic origin (Northern European vs. other) using SAS.
Although on stratification there were no systematic differences
in the relationship between naevus count and melanoma risk
by examiner, all analyses were adjusted for examiner.
To analyse the relationship between site-specific naevus
count and site-specific melanoma risk, odds ratios were calculated for the association between site-specific naevus counts
and melanomas at that site and were then compared with odds
ratios for the association between the site-specific counts and
all other melanomas. To determine the significance of these
differences, odds ratios were calculated for the site-specific
naevus count using melanomas at that site as cases, and
melanomas at all other sites as controls; the resulting odds
ratios were then tested for linear trend.
To analyse the potential public health implications of the
results, multiple exposure level aetiological fractions (the
proportion of melanoma cases in the population statistically
attributable to each level of naevus count) were calculated
(Schlesselman, 1982). These were stratified by age, as this was
noted during analysis to be related to naevus count.
RESULTS
Cases and controls were similar with respect to mean age (51
years in cases, 50 in controls), sex (57% and 55% male,
respectively), and ethnicity (95% and 92%, respectively, of
Northern European origin). Compared with all melanomas
occurring in New South Wales from 1989 to 1991, the case
melanomas were a little thicker (Table I). Most cases were
superficial spreading melanoma (SSM, 62%) or nodular melanoma (NM, 20%). Lentigo maligna melanoma (LMM) comprised only 2% of case melanomas, compared with 9% of all
melanomas in New South Wales.
TABLE I - COMPARISON OF CASES WITH ALL INCIDENT ME\ANOMAS,
NSW, 1989-1991 (EXCLUDINGMELANOMA IN
Characteristic
Sex
Male
Female
Age
Mean (SD)
Thickness
Mean (SD)
Median
Histology
SSM
Nodular
Desrnoulastic
Lentigd maligna
Other
Cases ( n = 236)*
sm)
All melanomas,
NSW, 1989-1991
(n = 6,107)
57 (134)
43 (102)
56
44
51.0 (15.7)
55.4 (17.4)
1.57 (1.53)
1.10
62 (146)
20 47 ’
6 (131
2 (5)’
11 (25)3
1.42 (1.97)
0.75
65
17
1
9
7
‘Data are percents, with numbers in parentheses except for
means and medians.-zExcluding 6 cases of in situ melanoma which
is not a registrable cancer in NSW.-3Including 19 unspecified, 1
acral lentiginous, 1 amelanotic and 4 other.
The most common site of melanoma in males was the back
(45% of all melanomas, compared with 21% in females). In
females, the most common sites were the legs (35%, compared
with 12% in males) and arms (29%, cornparedwith 11%). The
female excess of limb melanomas was mainly below the knee
(24%, vs. 7% in males) and on the upper arm (21% vs. 8%).
Head and neck melanomas constituted a greater proportion of
cases in males (22%) than in females (lo%), and the anterior
trunk was the site of melanoma in less than 10% of instances in
each sex.
All ORs were initially examined separately by histology
(SSM, NM, and other melanoma), and by tertile of thickness of
melanoma (0-0.60 mm, 0.61-1.59 mm and 21.6 mm), but
there were no systematic differences (data not shown).
Of the non-naevus phenotypic factors (Table 11), inability to
tan was the strongest risk factor. Those who were unable to tan
had a risk 5 times that of those who tanned easily and darkly.
Those who burnt easily had a 3-fold risk compared with those
who were resistant to burning. Red hair and blue eyes were
also risk factors, but the magnitude of association was smaller.
Extensive freckling on the face as a child was associated with a
2.6-fold increased risk compared with those who did not
freckle. When these factors were adjusted for each other, only
inability to tan remained significant. The effect of inability to
tan appeared to be stronger at younger than older ages, with
those who were unable to tan having an 11-fold increased risk
of melanoma in the bottom tertile of age ( 1 6 4 1 years)
compared with a 3-fold increase in risk in those in the top
tertile of age (age 2 60 years).
The magnitude of the association with naevi was greater
than that with other phenotypic variables (Table 111). Persons
with 100 or more naevi had an OR of 12 compared with those
with less than 10 naevi. Individuals with 5 o r more naevi of at
least 5 mm in diameter had a 5-fold increased risk compared
with those with none of these naevi. Having at least 1 naevus of
diameter 2 1 cm was associated with a non-significantly
increased risk. Persons with 5 or more atypical naevi had a
9-fold increased risk. Neither blue naevi nor “congenital”
naevi were related to risk. Those with at least one iris naevus
had a 2-fold increase in risk. When total naevus count was
adjusted for atypical naevi, and all other naevus variables were
adjusted for total naevus count, only total count and the
presence of iris naevi remained significantly related to melanoma risk. The odds ratios for high total naevus counts were
non-significantly higher in younger than in older people (data
not shown).
Raised ORs were seen with increased naevus counts on
sun-exposed areas (the head and neck, trunk, arms and legs),
but also on rarely sun-exposed areas such as the buttocks and
the scalp (Table IV). Site-specific naevus counts were more
predictive of melanoma in that site than for melanoma
elsewhere for all sites except the arms (Table V). This
tendency was statistically significant for the back ( p = 0.02,
trend test) and nonsignificantly positive for all other sites
except the arm.
In cases and controls, site-specific naevus counts by sex
showed the same site pattern as melanoma (Table VI). Counts
were higher on the legs and arms in women than men, and
higher on the back in men than in women, in both cases and
controls, although these differences did not reach statistical
significance. Counts on the head and neck were similar in
males and females.
The naevus count declined with age in both cases and
controls (Fig. 1). It was also lower in redheads than in those
with other colours of hair, both in cases (mean of 35 in
redheads vs. 64 in those with other hair co1our.p < 0.05) and
controls (15 vs. 30,p < 0.05).
The aetiological fraction (AF) in those with 50 or more naevi
was 30%, and in those with 100 or more it was 20% (Table
VII). In persons aged under 40 years, the A F was 49% for
NAEVI AND MELANOMA IN NEW SOUTH WALES
487
TABLE I1 - ODDS RATIO OF MELANOMA IN RELATION TO PHENOTYPIC VARIABLES
Factor
Number
of cases
Category
(%)
Easily and deeply
Easily
Moderately
Very slightly
Never
p value, trend test
Propensity to burn Never/rarely
Sometimes
Usually
Always
p value, trend test
Hair colour
Blackidark brown
Medium brown
Light brown
Blonde
Red-brown Ired
p value, heterogeneity
Eye colour
Brown
Hazel
Green
Blueigrey
p value, heterogeneity
Freckles as a child None
Few
Moderate
Many
I, value, trend test
Number
of controls
(8)
Ability to tan
23 (10
44 j181
80 33
94 (39)
53 (20
71 j261
79 29
66 (25)
$$!giy
Odds ratio
(95% CI)’
1 (referent)
1 (referent)
1.3 (0.7-2.7)
p < 0.0001
1 (referent)
1.3 07-25
2.3 $1:3-4:11
3.1 ( 1 . ~ 7
D < 0.0001
i (referent)
1.0 (0.6-1.7)
0.9 (0.5-1.5)
p = 0.041
1 (referent)
36 (15
p = 0.020
1 (referent)
1.1 (0.6-2.3
1.3 j0.7-2.51
1.6 1.0-2.6
p =‘ 0.024 ’
1 (referent)
1.8 (1.1-3.0)
124 51)
42 17)
32 113
44 18{
p =
0.0002
) 1.3 (0.6-3.0)
D = 0.406
l‘(referent1
1.0 (0.5-1.7)
0.8 0.5-1.4
1.1 [0.6-2.0]
1.1 (0.5-2.4)
p = 0.820
1 (referent)
1.1 (0.5-2.2)
1.2 0.6-2.3
1.2 iO.7-2.11
p = 0.983
1 (referent)
1.4 (0.8-2.5
D = 0.142
‘Adjusted for age, sex, examiner and ethnicity.-*Adjusted for the above variables and all other
variables in the Table.
TABLE Ill - ODDS RATIO OF MELANOMA IN RELATION TO TOTAL BODY NAEVUS VARIABLES
AND IRIS NAEVI
Factor
Cdlegory
Total body naevus count 0-9
10-24
25-49
50-99
loo+
p value, trend test
Naevi of size 5-9 mm
0
Number
of(%\
cases
Number
of controls
(9%)
54 22
53 [22]
34 1 6 1
41 17
56 23)
107 40 1 (referent)
76 [28 1.7 1.1-2.9
39 jl5( 2.9 [1.6-5.4]
31 12 4.3 2.2-8.1
16 6) 12.1 (5.7-25.6)
p < 0.0001
181
45 /68]
17 11.4
(rjferent)
0.8-2.4)
g
4giF
Odds ratio
(95% CI):
1 (referent)
1.7 (1.0-2.8
2.8 r.5-5.21
3.7 1.9-7.3
7.4 2.9-18.6)
p < 0.0001
1.1
1
(riferent)
0.6-1.9
1
Naevi of size
2
1 cm
Atypical naevi
“Congenital” naevi
Blue naevi
Iris naevi
2-4
5+
p value, trend test
0
1+
0
1-2
3-4
5+
p value, trend test
0
1+
0
1+
0
If
26 10
37 {lh) 14 5)
I
I I
202 (85) 238 (89
35 15 28 11
149 63 208 78
I
~
I
~
201 (85 221 (83
36 15 45 17
232 98 260 98
44 (18)
31 (12)
2.9 1.6-5.2)
1.4 0.7-2.81
5.5 2.7-11.2) 1.7 0.7-4.1)
p < 0.0001
p = 0.176
1 (referent) 1 (referent)
1.6 (0.9-2.7)
1 (referent)
1.0 (0.6-1.8)
1 (referent)
p < 0.0001
1 (referent)
1.1 (0.7-1.9)
1 (referent)
1.0 (0.3-3.4)
1 (referent)
2.0 (1.2-3.3)
p = 0.097
1 (referent)
0.9 (0.5-1.6)
1 (referent)
0.8 (0.2-3.0)
1 (referent)
1.8 (1.0-3.0)
‘Adjusted for age, sex, examiner and ethnicity.-2Adjusted for the above variables. Odds ratios for
total body naevus count are also adjusted for number of atypical naevi, and all other variables are
adjusted for total body naevus count.
those with SO or more naevi and 41% for those with 100 or
more naevi. It declined with age to 15% and 5%, respectively,
in those 60 or older.
DISCUSSION
Our study found a strong relationship between total body
naevus count and risk of melanoma. We found increased risks
of melanoma also associated with atypical, large and iris naevi,
site-specific naevus counts and a sun-sensitive phenotype.
In interpreting these results, it must be acknowledged that
the study was not population-based. To allow for future
comparisons with a similar study being conducted in England,
the same medical observers were used in both studies, precluding the additional time required to obtain population-based
GRULICH ETAL
488
TABLE Tv- ODDS RATIO O F MELANOMA IN RELATION T O SITE-SPECIFIC NAEVUS COUNTS
Andlomical site
Arms
Legs (excluding feet)
Naevus
count
0
1-4
5-9
10-19
20 +
p value, trend
0
1-4
5-9
10-19
20
p value, trend
0
1-4
5-9
10-19
20 +
p value, trend
0
1-4
5-9
10-19
20 +
p value, trend
0
1
2+
p value, trend
0
1
2+
p value, trend
0
1
2+
p value, trend
0
1-4
5-9
10+
p value, trend
+
Anterior trunk
Posterior trunk (excluding
buttocks)
Scalp
Feet
Buttocks
Face and neck
Number
of cases
(%))
Number
Odds ratio
(95% CI)'
of controls
(%)
1
1.3 (0.8-2.31
3.3 1.7-6.2'
2.2 1.1421
8.4 4.2-16.7)
p < 0.0001
1
I
41 (17
47 20
33 [ld
38 16
82 (341
77 (29
96 (36
109 1411
39 15
19 7.1
6 [2.2]
71 (26)
98 36)
45 17
28 [lo]
27 (10
203 84
27 [ l l j
12 (5.0)
245 91)
21 17.8
3 (1.11
167 69
37 1151
38 16)
235 87)
22 18.21
12 4.4
148 62
32 [13]
59 (25)
223 83)
19 t7.1
26 (9.71
2.9 1.5-5.4
7.6 r4.0-14.3)
p < 0.0001
1
I
1.6 11.0-2.6)
3.2 (1.8-5.9j
5.6 2.7-11.6
15.1 [5.5-41.1]
p < 0.0001
1
p < 0.0001
1
1.8 (1.0-3.3)
5.0 (1.3-18.5)
p = 0.0033
1
2.7 (1.5-4.9)
5.0 (2.4-10.2)
p < 0.0001
1
3.0 (1.6-5.7)
3.8 (2.2-6.6)
p < 0.0001
1
31 13
20 (8.3)
7 (2.6)
1.2 (0.8-1.8
2.0 (1.1-3.91
4.2 11.5-11.6)
p =' 0.002 '
'Adjusted for age, sex, examiner and ethnicity
cases and controls. However, the close similarity of the risks we
have found associated with phenotypic variables to those
found in Australian population-based studies (Holman and
Armstrong, 1984; Green el al., 1985) suggests that our study
sample was representative of the population. It is unlikely that
survival bias would have had a major effect, since a condition of
inclusion was diagnosis in the last 3 years. No survival data
were available for New South Wales, but in South Australia
during 1986-1992 the 3 year survival from invasive melanoma
was 94% (South Australian Cancer Registry, 1994). Observer
bias was also possible since the interviewers were aware of the
case-control status of the subjects. A large degree of interobserver bias was not present, as indicated by the close
correlation between the 2 examiners' naevus counts.
Non-naevus variables
The magnitude of the relationships of the non-naevus
variables with melanoma risk was similar to that in other
published studies (Bliss et al., 1995). Hair and eye colour were
only weakly related to melanoma risk, and a history of many
freckles as a child was moderately associated with melanoma
risk. Persons who were likely to burn were at approximately
3-fold increased risk. Inability to tan showed the strongest
relationship to melanoma risk of the non-naevus factors. It was
also the strongest non-naevus risk factor in population-based
studies in Queensland (Green et al., 1985) and Western
Australia (Holman and Armstrong, 1984). Somewhat lower
risks associated with tanning ability have been found in
Denmark (0sterlind et al., 1988), Germany (Garbe et al., 1989)
and Canada (Elwood et al., 1985), although the measures used
have not been identical. Ability to tan may be more accurately
recalled in an environment where exposure to UV radiation
sufficient to cause tanning is usual, compared with countries
where such levels of UV radiation are uncommon. A decrease
of the effect of inability to tan at older ages has not been
reported previously.
Naevus variables
The magnitude of the relationship we have found between
melanoma risk and naevus count is difficult to compare with
the findings of others, because different studies have used
different techniques of counting naevi, and differing sampling
schemes for cases and controls have been used. We have
compared our results only with studies which counted naevi of
2 2 mm. Cautious comparisons were possible by re-grouping
our total naevus counts to the categories used by others (data
not shown). In 2 Northern European studies (Swerdlow et al.,
1986; Garbe et al., 1989), much higher ORs for the melanomanaevus relationship were found than in the present study. In a
Scandinavian study (Augustsson et al., 1991), ORs were only
slightly higher than in the present study, but there were
exceedingly large differences in both case and control naevus
NAEVI AND MELANOMA IN NEW SOUTH WALES
'TABLE V - 0III)S
R,\TIO O t Mtl.A?OhlA IN K t L 4 ' I ' I O N
489
TO SITE-SPECIFIC NAEVIJS COUSIS FOR AILLANOMA
ATTILYI S l r E A Y D hIEL,\NO.M,\ A I ' A L L OTIIERSlT'tS
-
Odds ratio
Anatomical site (number of cases)
Arms (n
N
~
~
~ (95%
~ 'I)' ~for
melanoma
at this site
s
Odds ratio (95% CI),'
Oddsratio
(95ro
for
cases = melanoma
~
~
t
at that site,
m e , $ ~ all
~ ~ ~controls
t ~ ~= melanoma
at all other sites
1
1
0
1-4
0.86 (0.30-2.5) 1.5 (0.83-2.6)
3.9 1.9-76
5-9
1.7 (0.50-5.61
2.2 [1.0-4:5]
10-19
9.0 (4.3-19)
20 +
p value, trend
Le s (excluding feet)
0
0
1
1.4 0.49-3.9)
1.2 0.66-2.2
fn = 56)
1-4
1.9 t0.94-3.81
3.2 [Ll-9.4)
5-9
3.4 1.1-10
2.6 1.3-5.3)
10-19
8.5 t2.9-251
6.9 [3.5-14)
20 +
p value, trend
1
Anterior trunk (n = 16)
0
1
14
0.34 (0.04-3.3) 1.7 (1.0-2.7
3.3 (1.8-6.01
2.0 0.29-13)
5-9
10-19
4.8 (2.3-10)
12.3 12.2-69)
31 (3.3-289)
20 +
14 (5.2-39)
p value, trend
1
1
Posterior trunk (excluding
0
2.6 1.1-6.2)
buttocks) (n = 85)
14
5-9
1.9 0.65-5.6)
10-19
5.9 2.0-17)
11 (4.1-31)
20 +
p value, trend
0
1
Face and neck (n = 40)
1
1-4
1.0 (0.44-2.2
1.9 0.43-8.31
5-9
6.2 11.3-29)
10+
I, value, trend
=
45)
p
p
p
i
p
D
1
0.56
0.68
1.06
0.84
= 0.98
1
1.60
1.94
1.47
1.36
= 0.77
1
0.18
0.55
2.47
1.09
= 0.24
1
1.64
1.07
1.39
3.88
= 0.02
1
0.74
0.65
1.54
= 0.24
'Adjusted for age, sex, examiner and ethnicity.
TABLE VI - SITE-SPECIFIC NAEVUS COUNTS BY CASE-CONTROL STATUS AND SEX FOR THOSE SITES OF
MELANOMA THAT SHOWED MARKED DIFFERENCES IN DISTRIBUTION BY SEX (MEAN AND 95% CI)
Cases
Site
Males
Legs
Arms
Back
Head and neck
Controls
Females
Males
Female5
17 13-20)
13 f9.7-16)
14 (11-17)
3.0 (2.14.0)
counts (higher in Scandinavia) from the present study, suggesting that the criteria for counting naevi may have been different
from ours. A California study found ORs only slightly higher
than ours (Holly et a/., 1987). There is a suggestion in these
data that the naevus-melanoma relationship is somewhat less
strong in countries with higher U V exposure.
Naevus counts in our controls were higher than in 2 of 3
other European studies, consistent with the hypothesised role
of U V radiation exposure in naevus formation. In Germany
(Garbe et a / . , 1989), controls of similar mean age to ours had
mean naevus counts of 18, while ours had 29. In Scotland
(Swerdlow et a/., 1986), 5% of controls had more than 50 naevi,
compared with 17% in the present study. However, higher
mean counts in controls (of 67) were found in 30 to 50 year olds
in a Swedish study (Augustsson et al., 1991). The mean naevus
count in controls in the same age group in the present study
was 44. The reason for this difference remains unclear, but
difference in naevus counting technique remains a possibility.
High mean naevus counts (of 36 in controls) were also found in
a study in California (Holly et al., 1987), where U V radiation
will have been much higher than in Europe.
Comparisons with the results of others were also performed
for atypical naevi. On re-grouping our counts to those of
others, and restricting our age range to similar ranges, the ORs
we have found were lower than in Sweden, Germany and
California (Augustsson et al., 1991; Garbe et al., 1989; Holly et
al., 1987). However, atypical naevus counts in controls in our
study were similar to others' findings. Differences in techniques and criteria for naevus counting may have occurred
between the studies.
Only 2 studies of identical design, using the same observers,
in areas of high and low U V exposure can resolve whether the
patterns described of the naevus-melanoma relationship in
different countries are due to U V radiation or study design. It
will therefore be of interest to compare the results of this study
with those of the comparable study conducted in England,
when data are available to do so. In addition, future studies of
the relationship of naevi and melanoma should present their
data stratified by variables known to influence naevus count,
such as age, and possibly pigmentary traits, to facilitate
comparisons with other investigators' results.
Few case-control studies have reported site-specific naevus
counts as risk factors for melanoma. Our results were consistent with those of Swerdlow et al. (1986), who found that
naevus counts in nearly all body sites, including chronically,
intermittently, and rarely sun-exposed areas, were associated
with an increased risk of melanoma. In addition, we found
higher naevus counts on the arms and legs of females than
GRULICH ETAL.
490
Naevus count
/
80
~
-
1 ~.'
\ I
\
601
40
20
...-..
__I
1.
1
1-
\
\
\
\
0
TABLEVII - AETIOLOGICAL FRACTION OF MELANOMA FOR SELECTED
NAEVUS COUNTS BY AGE
Age (years)
Prevalence
in controls
[% (numberitotal)]
Aetiological
fraction
[% (95% CI)]
< 40
t 50
L 100
40-59
t 50
2
60 +
100
2 50
L 100
Total study population
2 50
2 100
14 14/99)
7h99)
3 3/91
1 [I1911
18 471269
6 f16/269]
15 6.6-23)
5 1-0.2-9.4)
30 22-39
20 114-261
males, and on the backs of males than females, consistent with
the distribution of melanoma. A similar sex distribution of
naevi has been reported in a Swedish case-control study (legs
and back only, Stierner et af., 1992). We also found anatomical
site-specific naevus counts to b e significantly more strongly
related to melanoma at the same site than melanoma at other
sites for the back, and a similar non-significant relationship for
other sites except the arm. This finding would suggest that
local factors might affect risk of melanoma rather than simply a
systemic effect. Whether these patterns may be related to
site-specific patterns of sun exposure will be examined in a
future report.
Use of naevus count as a marker of a high-riskpopulation
The great majority of studies that have looked at naevus
count have identified naevus count as the strongest risk factor
for melanoma in light-skinned people. The possibility of using
naevus count as a marker of high-risk individuals, so they can
be closely followed for development of melanoma, has been
raised (English and Armstrong, 1988). In this study, 20% of all
melanomas occurred in the 6% of the population with > 100
naevi. In those aged under 40, over 40% of all melanomas
occurred in the 10% of the population with > 100 naevi. In
those aged 60 or more, however, a high naevus count was less
predictive. The presence of > 100 naevi occurred in only 1%of
controls and was associated with only 5% of melanomas in this
age group. Although using a lower cut-off in the elderly would
raise the AF, a lower naevus count would be a less useful
clinical marker of risk because of the confusion with nonmelanocytic pigmented lesions, which are common in this age
group.
ACKNOWLEDGEMENTS
The authors thank the staff of the melanoma units and
other medical clinics at which the study took place,
Mrs. M. Coates of the NSW Central Cancer Registry, who
supplied data on melanoma incidence in NSW, and Dr. A.
Green of the Queensland Institute of Medical Research, who
provided useful comments on early versions of the questionnaire.
REFERENCES
AUGUSTSSON,
A,, STIERNER,
U., ROSDAHL,
I. and SUURKULA,M.,
Common and dysplastic naevi as risk factors for cutaneous malignant
melanoma in a Swedish Population.Acta Derm. Venereal., 71,518-524
(1991).
BLISS,J.M., and 13 OTHERS,
for the International Melanoma Analysis
Group (IMAGE), Risk of cutaneous melanoma associated with
pigmentation characteristics and freckling: systematic overview of 10
case-control studies. Int. J. Cancer, 62,367-376 (1995).
COATES,
M., DAY,P., MCCREDIE,
M. and TAYLOR,
R., Cancer in New
South Wales, incidence and mortality 1992. NSW Central Cancer
Registry, Kings Cross, Sydney (1995).
ELWOOD,J.M., GALLAGHER,
R.P., DAVISON,
J. and HILL, G.B.,
Sunburn, suntan and the risk of cutaneous malignant melanoma-the
Western Canada Melanoma Study. Brit. J. Cancer, 51,543-549 (1985).
ENGLISH,
D.R., and ARMSTRONG,
B.K., Identifying people at high risk
of cutaneous malignant melanoma: results from a case-control study in
Western Australia. Brit. med. J., 296,1285-1288 (1988).
GARBE,
C., KRUGER,S., STADLER,
R., GUGGENMOOS-HOLZMANN,
I.
and ORFANOS,
C.E., Markers and relative risk in a German population
for developing malignant melanoma. Int. J. Dermatol., 28, 517-523
(1989).
GREEN,A., MACLENNAN,
R. and SISKIND,
V., Common acquired nevi
and the risk of malignant melanoma. Int. J. Cancer, 35,297-300 (1985).
NAEVI AND MELANOMA IN NEW SOUTH WALES
HARRISON,
S.L., MACLENNAN,
R., SPEARE,R. and WRONSKI,I., Sun
exposure and melanocytic naevi in young Australian children. Lancer,
344,1529-1532 (1994).
HOLLY,E.A., KEELY,J.W., SHPALL,S.N. and CHIU,S.H., Number of
melanocytic nevi as a major risk factor for malignant melanoma. J.
Amer. Acad. Dennatol., 17,459-468 (1987).
HOLMAN,C.D.J. and ARMSTRONG,
B.K., Pigmentary traits, ethnic
origin, benign nevi, and family history as risk factors for cutaneous
malignant melanoma. J. nut. Cancer Inst., 72,257-266 (1984).
MACLENNAN,
R., GREEN,A.C., MCLEOD,G.R.C. and MARTIN,
N.G.,
Increasing incidence of cutaneous melanoma in Queensland, Australia. J. nat. Cancer Inst., 84,1427-1432 (1992).
NEWTON,
J.A., BATAILLE,
V., GRIFFITHS,
K., SQUIRE,
J.M., SASIENI,
P.,
CUZICK,J.. BISHOP,T. and SWERDLOW,
A., How common is the
atypical mole syndrome phenotype in apparently sporadic melanoma?
J. Amer. Acad. Dprmatol., 29, 989-996 (1993).
NORDLUND,
J.J., KIRKWOOD,J., FORGET,B.M., SCHEIBNER,
A.,
ALBERT,D.M., LERNER,
E. and MILTON,G.W., Demographic study of
491
clinically atypical (dysplastic) nevi in patients with melanoma and
comparison subjects. Cuncer Rex, 45,1855-1861 (1985).
P)STERLIND, A., TUCKER,M.A., HOU-JENSEN,
K., STONE,B.J., ENGHOLM,G. and JENSEN,O.M., The Danish case-control study of
cutaneous malignant melanoma. I. Importance of host factors. Int. J.
Cancer, 42,200-206 (1988).
SCHLESSELMAN,
J.J., Case-control studies. Oxford University Press,
New York (1982).
SOUTHAUSTRALIAN
CANCERREGISTRY.Epidemiology of cancer in
South Australia. South Australian Cancer Registry, Adelaide (1994).
STIERNER,U., AUGUSTSSON,
A., ROSDAHL,I. and SUURKULA,
M.,
Regional distribution of common and dysplastic naevi in relation to
melanoma site and sun exposure; a case-control study. Melanoma Res.,
1,367-375 (1992).
SWERDLOW,
A.J., ENGLISH,J., MACKIE,R.M., O'DOHERTY,C.J.,
HUNTER,
J.A.A., CLARK,
J. and HOLE,D.J., Benign melanocytic naevi
as a risk factor for malignant melanoma. Brit. med. J., 292, 1555-1559
(1986).
Документ
Категория
Без категории
Просмотров
2
Размер файла
654 Кб
Теги
632
1/--страниц
Пожаловаться на содержимое документа