Gainotti et al (Reference Gainotti, Azzoni and Marra1999) concluded that post-stroke major depression may fit with a psychological model rather than with a neurological model based on their findings that post-stroke patients with major depression displayed more reactive symptoms (emotionalism, catastrophic reaction, anxiety) and fewer endogenous symptoms than patients with endogenous depression.

There are some concerns regarding the conception and the methodology of the study. The major drawback in the methodology is the bias in selecting the control group. It is not surprising that patients with endogenous depression will have more endogenous symptoms than patients with post-stroke major depression, as major depression can be diagnosed using DSM-III-R (American Psychiatric Association, 1987) operational criteria in the absence of endogenous or melancholic symptoms. Further, it is also expected that affective symptoms related to brain damage, such as emotionalism and catastrophic reaction, will be more prevalent in post-stroke major depression than in patients with endogenous depression.

Gainotti et al seek to create an impression that there is no association between endogenous depression and psychological stressors, and that post-stroke major depression with symptoms such as anxiety and hyperemotionalism are the representation of a psychological reaction to stressful situations. The available literature fails to support the validity of dichotomous endogenous/reactive and endogenous/non-endogenous classifications (Reference Farmer and McGuffinFarmer & McGuffin, 1989). Further, emotionalism observed in post-stroke patients is often precipitated by non-specific stimuli, and crying or tearfulness observed in emotionalism may not be associated with alterations in mood (Reference PoeckPoeck, 1969). No attempt was made to examine whether the symptoms assumed to be stress-related, in stroke patients with major depression, were correlated with the severity of functional impairment or with subjective perception of stressful situations.

It would be of interest to investigate the differences in prevalence of endogenous or melancholic symptoms between post-stroke major depression and major depressive disorder. Moreover, using the same classificatory procedure in sub-typing post-stroke depression and depressive disorders may cause limitations and constraints when phenomenological comparisons are made between these two disorders. To overcome this problem, classification of post-stroke depressive disorders could be approached from the bottom up, with the identification of depressive symptoms in stroke patients. By applying multivariate analysis, these symptoms can be grouped into clusters or syndromes, which can then be validated.