Asthma and bronchial hyper-responsiveness (BHR), an asthma-related phenotype, result from many genetic (G) and environmental (E) factors. Passive exposure to tobacco smoke (ETS) in early life is one of these risk factors. Following a genome scan for asthma and associated phenotypes conducted in 295 French Epidemiological study on the Genetics and Environment of Asthma, our present aim was to investigate interactions between genetic susceptibility to asthma and to BHR with passive ETS using two different methods: the predivided sample test (PST) and the mean interaction test (MIT). PST and MIT consider the identical by descent (identity by descent) distribution at all markers in affected sib-pairs with 2, 1 or 0 sib(s) exposed to ETS. While the PST allows detection of both linkage and G x E interaction, the MIT tests for linkage by taking into account a possible interaction. Among the six regions detected at P</=0.005, three of them (1q43-q44, 4q34, 17p11) were revealed by both PST and MIT for BHR. The three other regions were detected by only one method: 5p15 for BHR using PST; 14q32 and 17q21 for asthma using MIT, underlying the importance of using concomitantly different approaches. None of these regions was revealed for asthma and bronchial responsiveness by previous linkage analyses of these data, supporting the idea that taking into account gene-environment interactions can substantially increase power of linkage detection. Our results also showed evidence for G x ETS interactions underlying BHR in all four regions detected by PST.