Crohn's disease (CD) is a complex genetic disorder for which aetiology is unknown. Recently, genetic factors for susceptibility have been described. Several genetic loci have been mapped and partially explain the familial aggregations of the disease. However, environmental factors may also contribute to these aggregations. We considered that if the role of non-genetic factors was negligible, CD patients would be randomly distributed in sibships with multiple affected siblings. On the other hand if there was a significant environmental contribution, the siblings would be affected non-randomly over exposure status. In order to test this hypothesis, we studied 102 sibships with two or more affected siblings. A statistical test, named Cluster of Affected Sibling Test or CAST, was developed, based on the exact calculation of the probability of observing a given number of clusters of affected siblings in multiplex families. The null hypothesis of a random distribution of affected siblings was rejected (P=0.005). The observed excess of affected sibling clusters indicates that birth order influences the disease status. Considering that an adjacent order of birth is a global estimate of environmental sharing, this observation strongly suggests that environmental factors contribute to the observed familial aggregations of the disease. This observation provides evidence that familial CD is a relevant tool for further studies of environmental factors and gene-environment interaction. More generally, the CAST statistics may be widely applicable to estimate the involvement of environmental factors in the aetiology of other binary traits which may be observed in multiple members of the same sibship.