Known and potential new risk factors for skin cancer in European populations: a multicentre case-control study

Abstract

BACKGROUND: \ud \ud During recent years numerous studies have suggested that personal and environmental factors might influence cancer development.\ud \ud OBJECTIVES: \ud \ud To investigate environmental and personal characteristics associated with skin cancer risk.\ud \ud METHODS: \ud \ud A multicentre hospital-based case-control study was performed in Finland, Germany, Greece, Italy, Malta, Poland, Scotland and Spain, including 409 patients with squamous cell carcinoma (SCC), 602 with basal cell carcinoma (BCC) and 360 with cutaneous malignant melanoma (CMM) and 1550 control persons. Exposures were assessed by questionnaires that were partly self-administered, partly completed by dermatologists. Unconditional logistic regression modelling was used to assess associations including the influence of certain drugs and food items on skin cancer risk.\ud \ud RESULTS: \ud \ud The usual associations were observed for sun exposure and pigmentation characteristics, with chronic sun exposure being most strongly associated with SCC risk, and naevi and atypical naevi with CMM risk. Use of ciprofloxacin was associated with a decreased risk of BCC [odds ratio (OR) 0·33] and use of thiazide diuretics was associated with an increased risk of SCC (OR 1·66). Ciprofloxacin was also associated with SCC (OR 0·34) and thiazines with BCC (OR 2·04), but these associations lost significance after correction for multiple testing. Consumption of pomegranate, rich in antioxidants, was associated with decreased BCC and SCC risk, also after correcting for multiple testing. Recent experience of stressful events was associated with increased risk, particularly of CMM.\ud \ud CONCLUSIONS: \ud \ud In this large case-control study from across Europe the expected associations were observed for known risk factors. Some new potential protective factors and potential risk factors were identified for consumption of certain food items, medication use and stress, which deserve further investigation in future studies

Potential impact of interventions resulting in reduced exposure to ultraviolet (UV) radiation (UVA and UVB) on skin cancer incidence in four European countries, 2010-2050

“…As the Scottish population was projected to grow slightly after 2033, but then decrease to levels slightly below 2033, we believe our solution of retaining the forecasted population sizes as in 2033 will have limited effects on numbers and proportions of potentially avoidable cases for 2050. Prevalence data and RRs were all derived from EPIDERM data, 16 ensuring comparisons with exactly the same grouping and measurements of data. However, if some selection has played a role in the original datasets, some estimates of both prevalence and RR might be biased.…”

The patient journey: a report of skin cancer care across Europe

“…The EPIDERM project investigating skin cancers across Europe 13 has identified and confirmed variations in skin cancer incidence 14,15 and treatment 16 that exist between the participating centres. These variations may be attributed to inherent differences in the populations under study (for example skin colour, dietary variations, sun exposure practices etc.)…”

Assessing physicians’ preferences on skin cancer treatment in Europe

“…Information on the demographic distribution and tumoral types of the overall sample of the EPIDERM study is published elsewhere. 9 Surgery was the first treatment option in 76AE5% (n = 1307) of the patients with skin cancer, the remaining group (23AE4%, n = 401) being treated with nonsurgical methods (P = 0AE001).…”

“…More details are described elsewhere. 9 Information on the study variables was obtained from questionnaires designed by the EPIDERM investigators and clinical data gathered by the participant dermatologists at face-to-face visits. Variables demonstrating a statistically significant association (univariate OR with P < 0AE05) in the univariate model were analysed in a multivariate model by unconditional binary logistic regression to identify independent predictors of treatment preference (surgery vs. nonsurgery), resulting in ORs with 95% CIs.…”