Can we generalise from findings in twins?

A number of recent studies have used data from twins to shed light on the causal pathways underlying the observed association between birthweight and cardiovascular risk factors or coronary heart disease. The issue of whether findings from twin studies are generally informative, or whether factors associated with twinning preclude generalisation, is considered here. It is concluded that the association between birthweight and later health may differ quantitatively between twins and singletons, but evidence regarding blood pressure suggests it may not differ qualitatively. However, more information is needed on a number of gestational and maternal factors, and on measures of health other than blood pressure. Placentation and issues relating to infertility and its treatment need to be recorded and, together with gestation length, may need to be taken into account in analyses.