Psychological Medicine

Original Article

Generalized anxiety disorder and depression: childhood risk factors in a birth cohort followed to age 32

a1 Institute of Psychiatry, King's College London, London, UK, and Duke University, Durham, NC, USA
a2 University of Wisconsin, Madison, WI, USA
a3 Institute of Psychiatry, King's College London, London, UK
a4 Dunedin Multidisciplinary Health and Development Research Unit, University of Otago, New Zealand

Article author query
moffitt te   [PubMed][Google Scholar] 
caspi a   [PubMed][Google Scholar] 
harrington h   [PubMed][Google Scholar] 
milne bj   [PubMed][Google Scholar] 
melchior m   [PubMed][Google Scholar] 
goldberg d   [PubMed][Google Scholar] 
poulton r   [PubMed][Google Scholar] 


Background. The close association between generalized anxiety disorder (GAD) and major depressive disorder (MDD) prompts questions about how to characterize them in future diagnostic systems. We tested whether risk factors for MDD and GAD are similar or different.

Method. The representative 1972–73 Dunedin birth cohort of 1037 males and females was followed to age 32 with 96% retention. Adult GAD and MDD were diagnosed at ages 18, 21, 26, and 32 years, and juvenile anxiety/depression were also taken into account. Thirteen prospective risk measures indexed domains of family history, adverse family environment, childhood behavior, and adolescent self-esteem and personality traits.

Results. Co-morbid MDD+GAD was antedated by highly elevated risk factors broadly across all domains. MDD+GAD was further characterized by the earliest onset, most recurrence, and greatest use of mental health services and medication. Pure GAD had levels of risk factors similar to the elevated levels for co-morbid MDD+GAD; generally, pure MDD did not. Pure GAD had risks during childhood not shared by pure MDD, in domains of adverse family environment (low SES, somewhat more maltreatment) and childhood behavior (internalizing problems, conduct problems, somewhat more inhibited temperament). Pure MDD had risks not shared by pure GAD, in domains of family history (of depression) and personality (low positive emotionality).

Conclusions. Specific antecedent risk factors for pure adult MDD versus GAD may suggest partly different etiological pathways. That GAD and co-morbid MDD+GAD share many risk markers suggests that the presence of GAD may signal a pathway toward relatively more severe internalizing disorder.

(Published Online January 4 2007)

c1 PO8O Institute of Psychiatry, De Crespigny Park, London SE5 8AF, UK. (Email: