Lung cancer mortality and fine particulate air pollution in Europe

Abstract

Inspired by the recent report from the EPIC cohort1 and 5 other prospective studies,2, 3, 4, 5 on the association between lung cancer and fine particulate air pollution, we plotted the mortality rates of lung cancer of 15 European countries against their average air concentration of fine particulates (PM2.5). Age standardized rates for lung cancer were derived from the EUCAN database (International Agency on Research on Cancer) for the year 1998, which included data for Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, Netherlands, Portugal, Spain, Sweden and the United Kingdom.7 We standardized for smoking by the indirect method using the reported smoking prevalence across European countries.8 Mean annual PM2.5 concentrations were based on grid-averaged PM2.5 (annual mean, μg/m3), which was estimated for the year 2002 by the International Institute for Applied Systems Analysis.9 It is reasonable to assume that the mean annual PM2.5 across countries measured in 2002 is representative for exposures in the past decennia. The age and smoking standardized lung cancer rates across the countries ranged from 15 to 115 per 100,000 men and from 7 to 57 per 100,000 women, in line with previous estimates.10 Mean annual PM2.5 concentrations ranged from 1 to 22 μg/m3. Spearman correlation coefficients between the countries' age and smoking standardized mortality rates for lung cancer and annual mean PM2.5 were 0.57 for men (p = 0.028, Fig. 1a), while the association was not significant for women (r = 0.37, p = 0.18; Fig. 1b). For men, mortality standardized for age and smoking prevalence at the country level showed that a 5 μg/m3 higher mean annual PM2.5 was associated with an increased risk for lung cancer of 17.7% (95% CI: 3.7–31.7). Mean annual PM2.5 explained 36% of the variation in the age and smoking standardized lung cancer rates in men across the 15 European countries. Age and smoking standardized lung cancer rates for men (a) and women (b) plotted by mean annual fine particulate air pollution (PM2.5) for Austria (AT), Belgium (BE), Denmark (DK), France (FR), Finland (FI), Germany (DE), Greece (EL), Ireland (IE), Italy (IT), Luxembourg (LU), The Netherlands (NL), Portugal (PT), Spain (ES), Sweden (SE) and United Kingdom (UK). Our analysis is based on aggregated data, which are prone to ecological bias, associated with attempting to deduce individual-level effects from group level data. Note that the large countries (FR, DE, UK and IT) where the within country heterogeneity in air pollution is likely to be greatest, are situated in the middle of the distribution, and thus, do not greatly influence the association (Fig. 1). In these larger countries, it would be interesting to do the analysis at a regional level, but we do not have such data. Obviously, our analysis cannot prove causality, but it may nevertheless provide guidance on how to design further detailed epidemiological studies, based on individual level data. Most of the prospective cohort studies have been conducted in countries with relatively low pollution levels, e.g. United States1, 2, 3 and Norway.5 Our analysis shows the relevance to include Belgium in future large scale European studies, given that this region has the highest mean annual PM2.5 concentration of Europe9, 11 and shows high standardized mortality rates for lung cancer. In conclusion, at an ecological level the heterogeneity in lung cancer mortality across European countries could not be entirely explained by differences in the age distribution and/or smoking habits across the countries. Our ecological analysis suggests that some of the differences in lung cancer mortality in Europe could be explained by fine particulate air concentrations. Yours sincerely, Tim S. Nawrot, Kristiaan Nackaerts, Peter H.M. Hoet, Benoit Nemery.