Incidence of melanoma in four English counties, 1989-92BMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6978.502 (Published 25 February 1995) Cite this as: BMJ 1995;310:502
- a Unit of Health-Care Epidemiology, Department of Public Health and Primary Care, University of Oxford OX3 7LF
- b Oxford Cancer Intelligence Unit, Anglia and Oxford Regional Health Authority, Oxford OX3 7LF
- Correspondence to: Julia Redburn.
- Accepted 30 December 1994
The age standardised incidence of malignant melanoma in England seems to have been increasing in both sexes by more than 5% a year from 1979 to 1988.1 More recent data from the Scottish cancer registry suggest that incidence in women has not increased there since 1988.2 Routine mortality data for England and Wales show no increase in the death rate from melanoma in women from 1989 to 1992, whereas for men it continued to rise. In view of the lack of national incidence data after 1988, we analysed registrations for malignant melanoma in a single regional registry from 1981 to 1992.
Methods and results
The Oxford Cancer Intelligence Unit collects and analyses information on a population of approximately 2.5 million people in Oxfordshire, Berkshire, Buckinghamshire, and Northamptonshire. The histological verification rate for melanomas is high at 96%. The completeness of melanoma registration was validated by using a variety of methods. Annual registration rates for malignant melanoma of the skin (International Classification of Diseases, ninth revision (ICD-9) code 172) for men and women were directly age standardised with the World Health Organisation's European standard population; 95% confidence intevals were calculated by a standard method.3 Similar rates were calculated and analysed separately for carcinoma in situ of the skin (ICD-9 232) that were histologically confirmed as malignant melanoma.
In the period 1981-92, 2068 melanomas and 296 melanomas in situ were registered. The figure shows that for men annual registration rates for melanoma (ICD-9 172 only) increased over this period, with an additional peak in 1988. In women, however, the increase from 1981 to 1988 was followed by a fall in the registration rate from 11.41 (95% confidence interval 9.5 to 13.3) per 100000 in 1988 to 7.53 (6.0 to 9.0) in 1992. The proportion of melanomas that were registered as in situ increased from 1985 to 1992 (P=0.09, χ2 test for linear trend, both sexes combined). This increase was more obvious in women than men. When melanomas in situ were analysed together with other melanomas, the annual registration rates for women still showed a decline after 1989 that was not seen in men.
The year on year increase in the incidence of melanoma up to 1988 has not continued beyond that year in women, at least in the area covered by the Oxford Cancer Intelligence Unit. This reversal of trend is unlikely to be artefactual as a change in ascertainment levels over time would not produce such a trend only in women.
Exposure to the ultraviolet radiation in sunlight is the most important known risk factor for melanoma.4 Healthy attitudes to sun exposure are becoming more common, particularly among women and those in higher social classes (G B Hastings and D R Eadie, unpublished data). Melanoma registration rates may be falling in women because they have been more receptive than men to health promotion messages. The increasing proportion of in situ melanomas suggests that both sexes are presenting earlier with these cancers.
There is evidence that public health campaigns give rise to a short term increase in the apparent incidence of melanoma.5 The rise in incidence before 1988 in both sexes was probably partly the result of campaigns that took place mainly in 1987. A preliminary examination of registrations for 1993 suggests that the decline in melanoma in women from 1988 to 1992 is about to be partially reversed, presumably as a result of greater public and professional awareness of skin cancers because of the current Health of the Nation initiative. Fortunately, the additional cancers registered in 1993 seem to include many in situ and thin malignant melanomas; both of these tumour types have a relatively good prognosis.
We thank Stephen Walters for data preparation and statistical help. The Unit of Health-Care Epidemiology is funded by the Department of Health and the Anglia and Oxford Regional Health Authority; the Oxford Cancer Intelligence Unit is funded by the latter.