Trends in the incidence of malignant melanoma in Denmark 1978-2007

The incidence of cutaneous malignant melanoma (CMM) in Denmark has risen more than 17% during the 1988-1997 period [1]. CMM is responsible for approximately 90% of all skin cancer-related deaths. On a global scale, CMM is one of the most rapidly growing cancer types [2]. In order to reduce the incidence rate and to improve the survival rate, it is necessary to understand in depth the aetiology of melanoma.

At the Department of Plastic Surgery, Breast Surgery and Burns Treatment at the Copenhagen University Hospital (Rigshospitalet), we have observed a pronounced increase in the number of patients referred with CMM from the island of Bornholm over recent years. The number of referred patients has been markedly higher than expected taking into account the island’s population.

Bornholm is an island in the Baltic Sea, situated to the South East of Denmark. It consists of bedrock and enjoys climatic conditions slightly different from those of the rest of Denmark. Its population is comparable to that of the rest of the country, but probably more prone to outdoor activities and exposure to ultraviolet (UV) light. Differences in melanoma incidence within the Danish population could help us understand the contribution of UV exposure to the documented increase in incidence.

The aim of this study was to quantify and to analyse the difference in melanoma incidence between Bornholm and the rest of Denmark.

MATERIAL AND METHOD

Data were extracted from the Danish National Cancer Registry [3] which is managed by the Danish Board of Health (DBH). Data were available for the 1978-2007 period and were validated by the DBH.

Age and sex-adjusted incidence data for ‘‘Bornholm’s municipality’’ and for ‘‘Denmark in total’’ were extracted (Figure 1). Data concerning the population size was extracted from the Danish National Survey Institute [4]. All statistical analyses were performed in the SAS 9.1 software.

ANALYSES

Incidence rates, per 100,000 per year, show a parabolic shape with an accelerating annual increase (Figure 1). Incidence data from both Bornholm and ‘‘Denmark in total’’ show an exponential development. The data fulfil the criteria of the Poisson distribution and a statistical test for trends and differences was made.

RESULTS

During the 1978-2007 period, an annual increase in incidence rates was documented for ‘‘Denmark in total’’ (Table 1). For ‘‘Denmark in total’’, we calculated an annual increase of 3.4% and for Bornholm the annual increase was 4%. Both estimates are significant, p < 0.001.

By comparing incidence rates on the basis of the Poisson distribution parameters, we found that melan-oma incidence rates on Bornholm are significantly higher than the corresponding Danish rates. The difference was estimated to constitute 20.5% (Table 1).

Based on the trend in melanoma incidence in ‘‘Denmark in total’’ from 1978 to 2007, an extrapolation for the coming years was made (Figure 2). Given a steady increase in incidence rate of 3.4% per year, the expected incidence rate will be more than twice as high (114% increase) in 2030 as in 2007 (Table 2).

DISCUSSION

Data from the Danish National Cancer Database was only available in an age- and sex-standardized form if extracted as incidence rates. The incidence rates were transformed into incidences by multiplying the incidence rate by the annual population of Bornholm and Denmark, respectively. The data are therefore encumbered with more unreliability than estimates based on raw data. On the other hand, estimates are more reliable when adjusted for the influence of age and sex. Both age and sex range among the strongest risk factors for melanoma [5].

A corresponding increase in incidence rate was found in a large-scale register study performed on a US population (290 mill. person years, 70,596 cases). In the US study, the annual increase was estimated to 3.1%.

The Bornholm incidence data are included in the incidence data for ‘‘Denmark in total’’. This will cause any difference between the two groups to be underesti- mated. In the 1978-2007 period, the population of Born-holm on average represented 0.9% of the entire Danish population [4].

The extrapolation into the coming years is based on incidence rates for ‘‘Denmark in total’’. The trend remained constant during the 1978-2007 period, and there was limited variation in the annual incidence rates, which improves the reliability of the extrapolation. The extrapolation should, however, be interpreted with some caution because it was not possible to account for all conditions. The extrapolation becomes less reliable the further into the future it predicts. Nevertheless, the data may serve as the basis for estimation of the effect of future campaigns and other prophylactic initiatives.

Various conditions might be involved in the overall incidence pattern for Bornholm and ‘‘Denmark in total’’. Three main factors contribute to melanoma aetiology: sun-related behaviour, constitutional factors and en- vironmental conditions.

Behavioural risk factors include pattern and amount of sun exposure, artificial sunlight and sunburns. Both exposure to artificial sunlight [6] from sunbeds and sunburns [7-9] are categorized as peak intermittent sun exposure. Intermittent sun exposure is believed to be associated with an increased risk of malignant melanoma compared with more regular sun exposure [7, 8, 10].

Differences in behaviour result in different sun exposure patterns and amounts. A person with sun-seeking behaviour can increase the exposure many folds compared with a person with sun-avoiding behaviour. Intermittent sun exposure patterns are typical for indoor workers. Danish studies have documented that indoor workers experience more sunburns and more intermittent exposure than outdoor workers [11, 12]. No data exist on behavioural differences between regions of Denmark, and it is not possible to compare behavioural differences between the population of Bornholm and ‘‘Denmark in total’’ based on data provided by the Danish National Survey Institute. It is expected that the population of Bornholm is more prone to outdoor activ-ities than people in other parts of Denmark. Whether this results in regular or intermittent sun exposure patterns remains unknown.

Constitutional factors [9, 13, 14] associated with melanoma are widely known (light phenotype, e.g. red or blond hair, fair skin and freckle and nevus density). A family history of melanoma is strongly associated with melanoma. No indications support that the population of Bornholm should differ from the population of ‘‘Denmark in total’’. During the observed period of time, the population of Bornholm has declined from approximately 47,000 in 1978 to 43,000 in 2006. Because of the introduction of a new municipality structure in 2003, it is not possible to achieve a complete overview of local migration [15]. The period 2003 to 2005 saw an annual net emigration of 140 persons (a little over 1,000 immigrants and nearly 1,200 emigrants) [15]. Compared with Denmark, Bornholm receives less immigrants per capita per year which is not consistent with a hypothesis that the melanoma cases may be imported from other countries. Higher immigration rates in ‘‘Denmark in total’’ compared with Bornholm could lead to an increasing share of people with darker skin and thereby a lower risk of melanoma in ‘‘Denmark in total’’ relative to Bornholm. The national distribution of ethnicity and phenotype characteristics is unknown. However, continuing emigration from Bornholm to other regions of Denmark will tend to equalize the differences in melanoma incidence between Bornholm and Denmark.

The environmental conditions of relevance include climatic differences and geological differences.

Denmark has a limited geographical extension and the ultraviolet index (UVI), sun ray angle and thickness of the ozone layer are therefore assumed to be the identical across the country. Moreover, both UVI and ozone layer thickness have been assessed too briefly to document any constant change with ensuing biological effects.

‘‘Hours of sunshine’’ is the one climatic factor where there are local differences within Denmark as it depends on cloud cover and aerosols. For Denmark there has been a 6.3% increase in sunshine hours from 1968 to 2007 [16]. The average difference in sunshine hours between Bornholm and Denmark was 5.4% in the 1961-1990 period [17].

Bornholm consists of bedrock which contains ra-dium. With radium decay, radon is formed. Radon gas up-concentrates in basements and buildings without sufficient protection [18]. Radon is the primary contributor to the mean radiation doses received by the Danish population, radiotherapy excluded. Ionizing radiation is to some extent associated with the development of melanoma, but the association is neither strong nor well-documented [19, 20].

CONCLUSIONS

Annual incidence rates for both Bornholm (4.0%) and ‘‘Denmark in total’’ (3.4%) have risen significantly from 1978 to 2007. The Bornholm incidence rate is 20% higher than the incidence in ‘‘Denmark in total’’. This difference is significant. A similar annual increase in melanoma incidence was found in a recent American study.

The cause of the difference in melanoma incidence rates between Bornholm and ‘‘Denmark in total’’ is probably associated with differences in sun exposure. In this context, the personal sun exposure pattern is essential, as it may dramatically affect UV exposure. Climatic and geological conditions may also play a role. The inter-regional differences and the changes over time in sunshine hours are too small to explain the continuing increase in melanoma incidences. More evidence is needed on the relation between melanoma and UV radiation and socio-demographic patterns within Denmark.

It is estimated that the melanoma incidence rate in Denmark will rise by more than 50% by 2020 and by more than 100% by 2030 compared with 2007. In order to reduce the increase in melanoma incidence rate, further prevention campaigns are needed. Campaigns should primarily focus on high-risk individuals according to phenotypic characteristics and on groups displaying risk behaviour.

Correspondence: Anders Klit, Platanvej 24, 1., -4., 1810 Frederiksberg C, Denmark. E-mail: andersklit@gmail.com

Accepted: 9 November 2010

Conflicts of interest: None