Throughout the 39-year investigation period, the incidence of uveal melanoma in Sweden was remarkably high from a global perspective. In the first period (1960–1964) the age-standardized incidence was 11.7 and 10.3 cases/million in males and females, respectively. The incidence declined to 8.1 and 8.6 cases/million during the period 1995 to 1998, a level similar to the incidence recently reported for Jews of Polish and Romanian ancestry.
13
The incidence of uveal melanoma in Scandinavia was once recognized as among the highest worldwide. In Denmark, Jensen
3 reviewed the period from 1943 to 1952 and found a crude, but not age-standardized incidence of uveal melanoma of 7.4 cases/million. Østerlind
7 later analyzed the Danish population of 1943 to 1982 with age standardization and found a stable incidence of ocular melanoma (conjunctival melanomas included) of 7.5 and 6 cases/million in males and females, respectively, during that period. In the United States with its heterogeneous population, the age-standardized incidence of ocular melanoma was estimated at 6 cases/million, based on the Third National Cancer Survey during the period from 1969 to 1971.
20 Because different standard populations, and, hence, stratum weights, were used in these studies, the figures are not fully comparable. Other investigations, mainly in white populations have resulted in similar data.
21 22 23 However, in many earlier studies, the incidence of ocular malignancies in the population older than 15 years has been used as an approximation for the incidence of uveal melanoma, suggesting inaccuracies stemming from inclusion of too broad a group. The temporal trends in incidence have often been regarded as stable, although with small fluctuations in many populations.
5 13 24 25 A recent investigation in the United States (Singh A, personal communication, 2001) supports the impression of a stable incidence of uveal melanoma during the period 1973 to 1997, with an overall rate of 4.3 cases/million inhabitants.
However, in our data, we detected a statistically significant reduction in the incidence of uveal melanoma of 1% yearly for the male population in Sweden during the period from 1960 to 1998. In the female population, the incidence rates declined by 0.7% annually, but the reduction did not quite reach a statistically significant level (P = 0.108). A slight predominance of males has been found in several epidemiologic studies on uveal melanoma. To our knowledge no statistically significant age-dependent gender difference has previously been demonstrated for this tumor. However, we discovered a significantly higher incidence in males than in females in the group more than 45 years of age, although the reasons are unknown.
Until brachytherapy became a preferred treatment for smaller melanomas, almost every eye with uveal melanoma was enucleated and a specimen submitted for pathologic analysis. Ocular pathology in the earlier decades of the study period was not centralized, and therefore various pathologists at different laboratories examined the specimens. In an evaluation to determine whether the diagnostic criteria for uveal melanoma changed over the investigative period, reexamination of specimens identified misdiagnoses in only 0.33%. The diagnostic accuracy appears to be in the same low range as reported in the Collaborative Ocular Melanoma Study (COMS) in which the rate of misdiagnoses was 0.3%.
26 The risk of not reporting uveal melanoma cases could have increased in later years, because more eyes are treated without obtaining a pathology specimen, and therefore the cases were not reported dually by both clinician and pathologist. To reduce this risk, hospital files from the only two centers treating patients with eye-sparing techniques (brachytherapy and proton beam irradiation) were checked and missing patients included in the study.
Although the Swedish population was once considered ethnically homogeneous, a small but growing immigration from both European and non-European countries has occurred in the past few decades. However, the net immigration rate remains comparatively low: the Swedish rate is estimated at 0.86/1000 population compared with 3.5/1000 population for the United States during 2000.
27 Data from the Swedish National Bureau of Statistics
16 indicate that the proportion of the population of non-European origin in 1960 was 0.2% and in 1998 had increased to 4.0%.
Furthermore, the age distribution of the immigrant population differed, because it was proportionally younger than the total Swedish population for every period investigated. The influence on the crude incidence of uveal melanoma for each gender over time when taking into account a presumably lower incidence in the immigrant population was estimated and found to be insignificant. Therefore, immigration alone probably does not account for the decreasing incidence of uveal melanoma documented in this study.
The stable incidence rates of uveal melanoma published since the 1950s contrasts sharply with the well acknowledged increase in incidence of skin melanoma in populations with fair complexions.
28 Our results concur with those found in a recently published population-based Swedish survey of incidence rates for vulvar melanoma.
29 At this extracutaneous UV-light protected site, the incidence declined 3% annually during the investigation period of 1960 to1984.
The average density of cutaneous melanoma is lower than that of uveal melanoma when the relationship between the skin area and the uveal area is considered. Specifically, uveal melanomas are 50 times more common in men and 41 times more common in women than cutaneous melanomas. Also, vulvar melanomas were found to be overrepresented by a factor of 2.5, compared with cutaneous melanomas.
30
The potential role of UV light in the development of uveal melanoma is a matter of controversy. Some studies indicate that sunlight exposure may be a risk factor,
10 31 32 but others found no clear latitudinal gradient.
20 Nearly all UV light in the adult eye is absorbed before reaching the uvea, although greater transmission is possible in childhood,
33 and a blue or light-colored iris is also suggested to transmit more UV-light.
Our results in a national survey in Sweden during a 39-year period point out a stable, and, in males, even decreasing, incidence of uveal melanoma during the same period that the incidence of cutaneous melanoma increased four to five times. The results therefore suggest that the initiation of uveal melanoma is less dependent on UV light exposure than that of cutaneous melanoma.