In the news

Twin studies have often been used to gauge the magnitude of the genetic contribution to susceptibility to common diseases. The reasoning is that monozygotic ("identical") and dizygotic twins should share similar environments, so the degree to which monozygotic twins share disease status (i.e. are "concordant") more often than dizygotic twins should give an indication of the size of the genetic component. Svendsen et al have applied this method in a study on rhematoid arthritis but rather than relying on twin volunteers, as most previous studies have done, they took a population approach, using the Danish twin registry to find twins and then measuring the "probandwise concordance rate" [Svendsen, A.J. et al (2002) BMJ 324, 264-268]. This counts only those twin pairs ascertained through a proband who independently reported rhematoid arthritis and fulfilled the study's clinical criteria. They claim that this approach minimises the ascertainment bias that is likely in other studies, in which concordant monozygotic twins tend to be over-represented, and also avoids reliance on subjects' own perception of whether they have the disease. Among 37,338 twins they found no concordant monozygotic twins by these criteria, and two concordant dizygotic twin pairs. They conclude that genes have little or no influence on susceptibility to rheumatoid arthritis.

Comment: As pointed out in an accompanying commentary by Silman, the stringent criteria for inclusion used by Svendsen et al mean that the total number of concordant twin pairs was very low and the confidence intervals very large, so firm conclusions are difficult. Other studies have implicated variants of immune system genes in susceptibility to rheumatoid arthritis but their influence appears to be more modest than in other autoimmune diseases such as multiple sclerosis.