We agree with Dr Fielding that neuroimaging in dementia is controversial (Psychiatric Bulletin, January 2005, 29, 21-23). Guidelines vary between the relatively restrictive Royal College of Psychiatrists (1995) statement, referred to by Fielding, to the all-inclusive consensus statement from the American Academy of Neurology (2001) in which computed tomography/magnetic resonance imaging (CT/MRI) is recommended.

We conducted a small audit which has some similar findings to those of Fielding. Out of 32 patients scanned in the past year whose notes were readily accessible, 25 (79%) were referred for CT scan according to College guidelines. Only 1 (3%) potentially reversible cause of dementia was found: an incidental meningioma which was not treated. This rate compares closely with Fielding's report. We also found a very high prevalence of cerebrovascular disease: ischaemic changes or infarcts being found in 27 patients (85%). This prevalence is much higher than in Fielding's report, perhaps reflecting variation in radiological reporting and/or geographical variation in the prevalence of this disorder.

The very low incidence of potentially reversible causes may reflect the patient group presenting to old age psychiatry. This may be higher in neurology clinics and other settings. The high prevalence of cerebrovascular disease is perhaps of much greater clinical significance to the old age psychiatrist. There may be treatment implications arising from an emerging view that vascular and Alzheimer pathology co-exist (Reference Langa, Foster and LarsonLanga et al, 2004), and this might be justification for CT as a routine test, as advocated by the draft Scottish Intercollegiate Guideline Network (SIGN).