Seasonal variation in the diagnosis of cutaneous melanoma and non-cutaneous malignancies: an Italian population-based study

Abstract

The aim of the present study was to investigate whether or not the diagnosis of cutaneous melanoma and non-cutaneous tumours shows seasonal variations. Data from the Tuscany Cancer Registry, central Italy, for malignant cutaneous melanomas (no. 2062) and non-skin cancers (no. 114,755), diagnosed during 1985-2000, were used to evaluate the monthly variation in the number of diagnoses. The chi-squared test was used to evaluate the heterogeneity in the monthly number of cancer diagnoses throughout the year and to compare monthly proportions between groups. Poisson multivariate regression was used to estimate the monthly risk of cancer diagnosis. Clear monthly variations were evident for both melanoma and non-cutaneous cancers, but the monthly diagnosis distributions differed. Statistically significant peaks in June [relative risk (RR), 1.54] and July (RR, 1.40) and a lower risk for January (RR, 0.76), in comparison with the mean monthly value, were documented for melanoma. Monthly risks were similar for thin (< or =1 mm) and thick (>1 mm) melanomas and for males and females. A peak in May-June and October and significantly lower numbers of cases in August and December were detected for all non-skin cancers. The holiday periods may influence the seasonality in the diagnosis of melanomas and non-cutaneous cancers. However, the monthly fluctuation for melanoma differed from that for non-cutaneous cancers; the peak in July suggests an additional effect of summer clothing, with fewer clothes worn and more skin visible, on melanoma diagnosis.