The elegant study of 1920 participants from the Baltimore Epidemiologic Catchment Area programme concluded that ‘the implied one-way causal relationship between spontaneous panic attacks and agoraphobia in DSM–IV appears incorrect’ (Reference Bienvenu, Onyike and SteinBienvenu et al, 2006). Bienvenu et al echo the arguments of many researchers, beginning with Marks (Reference Marks1987), that agoraphobia without panic attacks (primary agoraphobia) should be reinstated in DSM–V as a stand-alone diagnosis as in ICD–10.

It has been argued that evolutionary biological reasoning predicts the existence of a ‘hard-wired’ primary stand-alone agoraphobia, which should be classified with other specific phobias (Reference BrachaBracha, 2006). Specific phobias have been considered as conserved traits that enhanced survival during the human era of evolutionary adaptedness (Reference Nesse and StearnsNesse, 1999; Reference BrachaBracha, 2006). Primary agoraphobia may similarly be traced back to the fact that humans relied on arboreality as a major escape response long after they diverged from chimpanzees. Homo sapiens expanded beyond its densely forested East-African indigenous niche into sparsely wooded habitats (savannahs and water-front dunes) only about 70 000 years ago. In sparsely wooded habitats, anxiety in wide-open spaces was arguably a survival-enhancing trait since opportunities for arboreal escape from large predators were limited (Reference BrachaBracha, 2006). These arguments may be relevant to psychiatric classification and contribute to the ‘neuroscience research agenda to guide development of a pathophysiologically based classification system’ emphasised in the research agenda for DSM–V (Reference Kupfer, First and RegierKupfer et al, 2002).

If, as one of us (Reference BrachaBracha, 2006) has argued, the two types of agoraphobia have different modes of acquisition, there might be some clinical implications. Primary agoraphobia might, like other specific phobias, be especially amenable to virtual reality exposure treatment. In contrast, agoraphobia secondary to panic attacks can be classified in DSM–V and treated along with post-traumatic stress disorder (and other fear–memory–overconsolidation disorders, which are misclassified as specific phobias in DSM–IV–TR, e.g. hospital phobia, dentist phobia, dog phobia, bird phobia, and bat phobia).

Finally, contrary to myth, predictions based on brain evolution are eminently testable/falsifiable (Reference Nesse and StearnsNesse, 1999). Some 30 such predictions are elaborated elsewhere (Reference BrachaBracha, 2006).