Fragile X in families multiplex for autism and related phenotypes: prevalence and criteria for cytogenetic diagnosis

There is uncertainty regarding the necessary and sufficient criteria for cytogenetic diagnosis of the fragile X syndrome. Some have made the diagnosis when 1% or more of cells are anomalous, whereas others have used 3-4% as a threshold. The choice of threshold level has implications for the extent to which fragile X accounts for the genetic findings in autism. In this study, we tested the first degree relatives of families that were multiplex for autism and related phenotypes, and investigated the reliability and validity of different cytogenetic thresholds for fragile X diagnosis. Clinical diagnoses were made using standardized assessment procedures and operationalized criteria. Cytogenetic investigations, blind to clinical evaluations, were performed in two laboratories, utilizing a variety of induction procedures. Latent class analysis supported a three class solution to the cytogenetic data, corresponding to no, low (1-3%) and high (> 3%) levels of fragile X expression. High level expression was reliably identified and associated with other phenotypic features of the fragile X syndrome. It was found in 2 of the 24 multiplex families tested and consequently, could not account for the majority of familial aggregation of autism and related phenotypes. Low level expression was less reliably identified and, although occasionally found in apparently normal individuals, was associated with disorders of cognitive function. The aetiology of low level expression was unclear, but the overall pattern of findings suggested that it was different to high level expression. The findings highlight the need to interpret low positive counts cautiously.