Basal Cell Carcinoma: An Emerging Epidemic in Women in Iceland

J.A. Adalsteinsson; D. Ratner; E. Olafsdóttir; J. Grant-Kels; J. Ungar; J.I. Silverberg; A.K. Kristjansson; J.G. Jonasson; L. Tryggvadottir


The British Journal of Dermatology. 2020;183(5):847-856. 

In This Article

Patients and Methods

The ICR is a high-quality registry that contains complete records of all cases of pathologically confirmed skin cancers nationwide from 1981, with histological verification.[7] Our study cohort contained all patients diagnosed with first and subsequent BCCs in the years 1981–2017, with an associated International Classification of Diseases code and pathological diagnosis. We divided Iceland into two regions: (i) Reykjavik and the adjacent Reykjanes peninsula, and (ii) the rest of the country, which is composed mainly of small towns and rural areas. All BCCs diagnosed in individuals younger than 30 years of age were re-reviewed by a dermatopathologist accredited by the Accreditation Council for Graduate Medical Education before being included in the study.

For BCC incidence and join-point analysis, we used data from the population-based ICR. All new BCCs that occurred between January 1981 and December 2017 were included. Incidence rates of confirmed BCCs were presented for the 37-year period using world-standardized rates (WSRs)[11] expressed per 100 000 person-years, as well as cumulative risk. Cumulative risk of BCC occurrence before ages 40, 65 and 80 years (the latter defined as the lifetime risk) was calculated using age-specific rates, multiplied by the proportion of survivors and expressed as a percentage: cumulative risk = 1 − exp(− cumulative rate).[12]

For single and multiple BCCs, we present age-specific incidence rates. Due to the small population of Iceland (350 000 individuals), random variation was prominent. For example, the average yearly number of cases was only between one and four for sites other than the head/neck, trunk and legs (women). Therefore we used 5-year moving averages when showing changes with time, instead of looking at individual years. When calculating multiplicity, because the same tumour might be histologically diagnosed twice, all records of apparently multiple BCCs diagnosed within a 4-month period in the same person, and at the same anatomical location, were excluded. We also calculated the median interval between first and second BCCs, as well as the overall median interval between all BCCs that occurred.

Trends and join points were calculated for the three most common sites (head/neck, trunk and legs; in women) using Joinpoint version 4·6·0·0.[13] We stratified according to age (< 50 years and ≥ 50 years) in order to be able to compare the data with those from a previous Icelandic study on melanoma.[14] Trends were assessed by calculating the annual percentage change and the corresponding 95% confidence interval (CI).[13] Due to the large sample size in this whole population with independent observations, P-values were calculated using the c2-test, which met all assumptions. P-values < 0·05 were considered statistically significant. Due to the chance of BCCs in the early years of the study being second or third lesions we conducted a sensitivity analysis.