This general population study described the proportions of subjects working in sectors of activity exposed to crystalline silica over a period of 45 years and highlighted a decrease in exposure to crystalline silica since the 1980s. This phenomenon had been highlighted in France by Delabre et al. (6.2% in 1982 against 4.1% in 2007 for the whole French population) [5]. The proportion of subjects exposed in 2010 remains significant with 4.4% of the total sample exposed and 9% in men only.

The analysis among men shows a high proportion of exposed workers in the construction activities (particularly in the specialised construction activities including electrical and plumbing installation activities), which is also highlighted by Delabre et al. who found 628,000 subjects in the construction sector in 2017 in France, i.e. 64% of all exposed subjects [5]. Abroad, this predominance of the construction sector is also found, notably in Sweden, where Gustavsson et al. estimated that in 2013, approximately 50% of the subjects exposed to crystalline silica worked in the construction sector [9]. There was also a very high proportion of workers in the manufacture of basic metals, which peaked between 1975 and 1985 (5.48% and 5.5% respectively) and then gradually decreased to 3.01% in 2010. This result can be partly explained by the fact that part of our sample lives in the metropolitan area of Dunkirk, which is a zone particularly rich in metallurgical industries. It should nevertheless be noted that the industrial sector, and in particular the steel industry, was the 3rd most frequent sector of activity in 2017 for exposure to silica according to Delabre et al. [5].

In our sample, exposure was mainly found among men (only 9 out of 1548 women had been exposed to silica during their career in our study). Here again, this phenomenon was also reported by Delabre et al. [5] (93% of the subjects exposed to silica in 2017 were men) but also outside France, notably in Australia [10] (10.2% exposed among men versus 1.2% exposed among women in a sample of 4993 Australians) and Sweden [9] (about 90% of exposed subjects were men in the 2013 census data). As the number of women exposed here is very low, the interest of comparing their sectors of activity to those of men is limited, as each sector only concerns 1 to 2 subjects at most.

One of the major strengths of our study is the presentation of data on the sectors of activity that expose to crystalline silica over the entire career of subjects randomly selected from the general population. The application of the Matgéné Silice JEM on career history allowed a description of these sectors over 45 years, starting in 1965, with a presentation of the proportions per 5-year period, making it possible to follow the evolution of each exposing sector over time and consequently to provide original information, not limited by the population census periods used in other studies [5, 9]. The drawback of our methodology in relation to the use of census data is the limitation of the number of subjects and the geographical scope. Indeed, we included subjects from the general population residing only in Lille and Dunkirk (a city with a relatively high level of industrial activity), which provided a good representation of the population of northern France, but, as we have shown, we find similar results to those obtained in studies with larger sample sizes, based on the proportion of exposed subjects alone, which shows the value of the additional information provided by our study (proportions of exposed subjects by sector of activity at different times during their career, from 1965 onwards) [5]. Another limitation is the transversal design, which does not allow us to eliminate the cohort and age effects, which may lead to a survival bias. Our cross-sectional study is limited to subjects aged between 40 and 65 at the time of the study, so there are few subjects who worked in 1965 and 1970 and the results must be interpreted with caution for these years, given the small numbers involved. The 40–65 age group was chosen in the ELISABET study, from which the sample was drawn, partly in order to assess the respiratory effects of repeated occupational and environmental exposure, with the lower limit at 40, and partly to limit the effects of greater geographical mobility after retirement, which may depend on health status and socio-economic level, with the upper limit at 65. Nevertheless, the death rates remain low at these ages. Finally, the use of the career history collected in a cross-sectional manner may have introduced a recall bias. This bias should be qualified by the fact that the subject was asked about his or her career as a whole and not about the presence of silica exposure, which was assessed afterwards.

This study shows that the proportion of workers exposed to crystalline silica has been decreasing since the 1980s, but still reaches significant levels, particularly in the construction sector, making occupational exposure to crystalline silica and its health consequences an ongoing issue. Our results also confirm the importance of collecting and tracing workers’ occupational careers as part of their occupational health monitoring, since the frequency of a given occupational exposure may vary over time on a population scale, as we have shown here for crystalline silica, all the more so if these exposures are at risk of causing pathologies with a long latency period.