The paper by Thomas et al 1 (this issue, pp 83–87) furthers the concept of “vascular depression”. The concept is a venerable one, dating back at least to the scholarly work of Felix Post.2 It found new vigour in the work of Robinson and others,3 whose investigations provided important empirical support for older clinical findings. New in Robinson's studies was their demonstration of regional specificity, with the greatest burden of depression in patients whose cortical infarcts were small and/or located in the left frontal region (thus refuting “meaningful” explanations of depression in reaction to the extent of functional incapacitation). More recent work has shown associations between depression and cardiovascular risk factors or neuroimaging evidence suggestive of microvascular pathology (reviewed by Thomas et al 1) These later studies have tended to emphasise the importance of subcortical lesions, and have led to widespread speculation that microvascular change can provoke demyelination and other lesions that in turn disrupt long association fibres and, possibly, the circuitry that regulates mood and various drives.3

Thomas et al have tested and extended these notions by showing a relatively specific association between late life clinical depressions and atherosclerotic cerebrovascular disease confirmed at postmortem. Surprisingly, their investigations did not show an association of depression with cerebrovascular risk factors, nor with direct pathological evidence of microvascular pathology. We do not know whether these patients had MRI evidence of such pathology, as has been shown often before. Thus, we do not know whether some unusual attribute of the present sample explains its absence of microvascular disease, or whether there is instead a disjunction in this regard between suggestive MRI evidence and direct pathological examination. We need also to note that this was a small study in terms of statistical inference, and that its negative findings especially must therefore be regarded as tentative. Surely more work of this sort, preferably including MRI correlation, would be of enormous interest.

Whatever its underlying cause or mechanism, the emerging entity of “vascular depression” should serve as a reminder that most important psychiatric diagnoses are syndromes that can have multiple aetiologies and pathogenetic pathways. This principle helps us to understand why “major depression” in late life (for which cerebrovascular disease is probably a prominent provoking element) may differ importantly from early life depressive disorders that have similar defining criteria. Contrasting aetiologies and mechanisms may explain, for example, why the later life disorders may have a different clinical appearance and may tend toward chronicity and refractoriness to treatments that commonly succeed with younger patients,4 including in particular the selective serotonin reuptake inhibitors.