We were very interested to read the recent, thought-provoking editorial by Andrews et al (Reference Andrews, Szabo and Burns2002) on the prevention of depression in young people. However, we are concerned that they have understated the important role of genetics in early-onset depression. Contrary to their assertion that the children of parents with depression are likely to be at heightened risk for psychological rather than genetic reasons, available evidence suggests that childhood-onset depression represents a strongly genetic subtype of affective disorder (Reference Neuman, Geller and RiceNeuman et al, 1997; Reference Sullivan, Neale and KendlerSullivan et al, 2000). Up to 50% of pre-pubertal children with depression eventually develop bipolar disorder (Reference Geller, Zimmerman and WilliamsGeller et al, 2001) and recurrent, early-onset depression (defined as two or more episodes before age 25) is recognised as a malignant form of affective disorder characterised by high genetic loading, frequent recurrence and poor long-term outcome (Reference Zubenko, Zubenko and SpikerZubenko et al, 2001). Furthermore, one recent study suggests that the inheritance of depression in these families is compatible with a single major locus (Reference Maher, Marazita and ZubenkoMaher et al, 2002).

Preliminary findings from our own study of early-onset depression in a university population support the view that early age at onset defines a subgroup at very high genetic risk. Using the Family Interview for Genetics Studies (FIGS; National Institute of Mental Health, 1999), 76% of the subjects seen so far (36 out of 47) report at least one first-degree relative with affective disorder, with 87% (41 out of 47) reporting either a first- or second-degree relative affected. The mean age at onset in this group is 15.6 years (s.d.=2.6).

Population-based interventions are unlikely to reduce the prevalence of depression in young people as long as we have an incomplete understanding of how genetic and non-genetic risk factors interact to bring about the depressive phenotype. Interventions such as the cognitive therapy programme described by Clarke and colleagues (Reference Clarke, Hornbrook and LynchClarke et al, 2001) might be cost-effective strategies if they can be targeted to high-risk groups. Unfortunately, we are not yet in a position to reliably identify those individuals at high risk.