Between the mid-1950s and the mid-1980s, mortality from cutaneous malignant melanoma (CMM) has been rising in both young and middle-aged adults in most European countries (Franceschi et al.,1991, 1992), as well as in North America, Australia and New Zealand (La Vecchia et al.,1993a,b). In southern Europe, the upward trends have been more marked at younger age, suggesting a future expansion.
Estimates of future trends of the disease were, however, inconsistent. Further rises in CMM mortality up to the first decades of the next century have been projected for Sweden (Thörn et al.,1992), Canada (MacNeill and Elwood, 1995) and the European Union (Balzi et al.,1997). However, mortality from melanoma appears to have stopped rising in Scotland (MacKie et al.,1997; MacKie, 1998), the United States (Lee, 1997) and Australia (Giles et al.,1996) in the early 1990s and to have fallen in New Zealand for cohorts born after 1950 (Bulliard and Cox, 1996).
We have, therefore, systematically reviewed death-certification rates from skin cancer between 1955 and 1984 and between 1985 and 1995 in countries providing data to the World Health Organization (WHO) database.
MATERIAL AND METHODS
Numbers of death certifications from skin cancer over the period 1955–1995 were available for 18 countries from Europe, 2 from North America, Australia and New Zealand from the WHO database. All classifications used were re-coded according to the 9th Revision of the International Classification of Diseases (ICD-IX; World Health Organization, 1977).
On account of the inadequacy of mortality data to distinguish reliably between melanomatous (ICD-IX: 172) and non-melanomatous skin cancer (ICD-IX: 173), for the present analysis only young (i.e., age 20–44 years) and middle-aged (i.e., 45–64 years) adults were considered. In these age groups, over 90% and 80%, respectively, of deaths attributed to skin cancer are due to CMM (Levi et al.,1996). Such a definition does not include ocular melanoma. Estimates of the resident populations, based on official censuses, were obtained from the same WHO database.
From the matrices of certified deaths and resident populations, age-specific rates for each 5-year age group and for the individual years and calendar periods considered were computed. Age-standardized rates were based on the world standard population. When a single year was missing within a period, numerators and denominators were interpolated linearly from the previous and subsequent calendar years. The percent change in rates was derived from a log-linear model (based on single calendar year rates).
Table I gives skin cancer death rates in males and females aged 20 to 44 years in 1955–1959, 1980–1984 and 1990–1995 together with the corresponding percent change in rates per year for 2 separate calendar periods, 1955–1984 and 1985–1995. Figure 1 gives in graphical form the trends in mortality across each subsequent calendar period, from 1955 to 1995. Four different scales were adopted (1.0, 1.5, 2.0 and 4.0/100,000). Over the period 1955–1984, mortality from CMM increased in all countries considered for males, the increase ranging between 0.4% per year in Australia (which had the highest rates in 1955–1959 already) and 4.8% per year in Spain. In females, all countries except Australia and Portugal showed upward trends, ranging between 0.3% per year in the United States and 5.1% per year in Poland. Over the last decade, however, mortality from CMM declined in males in 12 of 22 and in females in 11 of 22 countries considered. The pattern of trends was similar for females and males in most European countries, and the trends were more favourable for females in the United States, Australia and New Zealand (Fig. 1).
Table I. TRENDS IN MORTALITY1 FROM SKIN CANCER AT AGE 20 TO 44 IN SELECTED COUNTRIES, 1955–1984 AND 1985–1995
Table II and Figure 2 give the corresponding figures for the age group 45 to 64. In Figure 2, three different scales were adopted (5, 10 and 15/100,000). Between 1955 and 1984, CMM mortality rates increased in both sexes and in all countries considered, with rates of increase between 2% and 4% per year. The rates of increase declined in both sexes and in several countries considered over the most recent calendar period, and in 7 countries in males and 10 in females a decline of CMM mortality was registered over the period 1985–1995. As for the younger age groups, trends were more favourable for females than for males in the United States, Australia and New Zealand.
Table II. TRENDS IN MORTALITY1 FROM SKIN CANCER AT AGE 45 TO 64 IN SELECTED COUNTRIES, 1955–1984 AND 1985–1995
Our present data confirm that mortality from CMM is still rising in middle-aged (45–64) males in several countries, though the change in rates has been smaller over the last decade in several countries of northern Europe, North America, Australia and New Zealand, i.e., where mortality from CMM had reached its peak levels and rates appear therefore to level off. The trends were more favourable in middle-aged women since in most countries (particularly in central and northern Europe and in North America) rates among women have been declining over the last decade.
Even more interesting are the changes in CMM mortality trends at age 20 to 44, when substantial rises had been observed in most countries between 1955 and 1985 (Franceschi et al.,1992). However, over the last decade, mortality from CMM declined in several countries, particularly in northern Europe.
Most single rates are unstable, due to a substantial random variation. However, the overall picture that emerged from a systematic analysis of worldwide CMM mortality at age 20 to 44 is broadly consistent, and indicates that mortality from CMM is now levelling off and even declining in these younger age groups. Since cohort effects have a major influence on skin cancer mortality (La Vecchia et al.,1998), the favourable mortality trends in the young are important indicators of a likely decline in mortality from CMM in developed countries over the next few years.
These trends may reflect modification in risk-factor exposure (Hill et al.,1993; Melia, 1997; Autier et al.,1996; Elwood and Jopson, 1997; MMWR, 1998). They also likely reflect better access to health care and better earlier prevention of melanoma (Mackie and Hole, 1992; Bulliard et al.,1992; Cristofolini et al.,1993; Thörn et al.,1994; Levi et al.,1998).