Objectives In the 1990s there were great expectations that the use of markers of genetic susceptibility would allow the identification of new occupational risks, a more complete characterisation of dose-response relationships and improved risk assessment. Several interaction studies were conducted examining candidate genes, for example on isocyanate exposure, genes in immune pathways (e.g. HLAII group) and occupational asthma, or studies on cancer, aromatic amines and the NAT2 gene.

Method Very few replicated in more than one population. Part of the problem was the small sample size and the selective evaluation of candidate genes/SNPs. The availability of GWAS should, theoretically, solve the second problem. We have very few GEWIS (genome-wide interaction study) on occupational exposures and they suffer even more than candidate-gene studies from sample size and also from lack of available studies for replication.

Results Apart from exposures, GEWIS may help identify new diseases genes since many may only have an effect in combination with exposure. In addition GEWIS should provide a much more complete evaluation of specific pathways, e.g. oxidative stress. The use of a specific pathway-based approach is still valuable if based on biological knowledge. Contrary to major advances brought by molecular epidemiology, studies on gene-environment interactions have proven to be complex and with few well established findings.

Conclusions I will discuss reasons for this, discuss recent changes from the use of genome-wide analyses and compare with earlier approaches based on evaluations of interactions of occupational exposures with few candidate genes. I will provide examples from a recent GEWIS on occupational asthma.