Several reports since the mid 1990s have demonstrated causal mutations in COL5A1 and COL5A2 in Ehlers-Danlos syndrome (EDS) Classic Type. However, the presence of rare type I collagen variants, and the absence of a large consecutive series with fully annotated clinical findings and mutation analysis, has not permitted the same definitive molecular causation for classic EDS as for the Marfan syndrome. Symoens et al. (Hum Mutat 33: 1485–1493, 2012) have now reported what is likely to be a benchmark clinical/molecular series on classic EDS. One hundred twenty-six patients with suspected classic EDS underwent sequence analysis of COL5A1 and COL5A2 and haplo-insufficiency studies of COL5A1, of which 102 patients displayed all three cardinal features of classic EDS: laxity of small and large joints, velvety hyperextensible skin, and widened atrophic scars. Ninety-three patients were found to have evidence of type V collagen mutations, with more than half of these resulting in haploinsufficiency of COL5A1. While 93 of 102 patients with all three major features had evidence of type V collagen deficiency, none of the 24 patients having only two of the major clinical features but lacking fragile skin and atrophic scars did. The latter features might then point to a more specific tissue requirement for type V collagen, whereas the other two features are likely to be more causally heterogeneous, since they are also found in other forms of EDS.
If one were to assume that the presence of all three cardinal features defines the phenotype, then the exon-centric approach used by Symoens et al. has 92% (93/102) sensitivity. The authors argue, with some plausibility, that since their methods do not detect mutations in the promoter region or deep within introns, the undetected mutations lurk there. It is just as likely, however, that undetected type COL5A1 mutations in those regions account only for the 7/102 patients with molecular evidence of COL5A1 haplo-insufficiency but without a detected mutation, and that causal mutations occur elsewhere. The final answer will require a full genome approach, which is now within reach.