Estimating the attributable fraction for cancer: A meta-analysis of nevi and melanoma

Abstract

Epidemiologic research has shown convincingly that certain phenotypic attributes are associated with increased relative risks of melanoma. Although such findings have intrinsic utility, there have been few attempts to translate such knowledge into estimates of disease burden suitable for framing public health policy. We aimed to estimate the population attributable fraction (PAF) for melanoma associated with melanocytic nevi using relative risk estimates derived from a systematic review and meta-analysis. We identified eligible studies using citation databases, followed by manual review of retrieved references. Of 49 studies identified, 25 and 23, respectively, were included in meta-analyses of atypical and common nevi. For people with > or =1 atypical nevi, the summary relative risk was 3.63 (95% confidence interval, 2.85-4.62), with a PAF of 0.25. The relative risk increased by 1.017 (95% confidence interval, 1.014-1.020) for each common nevus; however, significant heterogeneity in risk estimates was observed. We estimated that 42% of melanomas were attributable to having > or =25 common nevi (PAF 25-49 nevi = 0.15; PAF > or =50 nevi = 0.27), whereas PAFs for low nevus counts were modest (PAF 0-10 nevi = 0.04; PAF 11-24 nevi = 0.07). We modeled PAF under scenarios of varying nevus prevalence; the highest melanoma burden was always among those with high nevus counts (PAF range of 0.31-0.62 for > or =25 common nevi). Patients with > or =25 common nevi and/or > or =1 atypical nevi are a high-risk group, which might be targeted for identification, screening, and education. This work is the necessary first step in designing targeted preventive strategies for melanoma, which must now be overlaid with information about cost and utility.