Liddell et al (Reference Liddell, Lovestone and Owen2001) reviewed what knowledge we have on the genetic epidemiology of Alzheimer's disease for the purpose of informing relatives of patients about their own risks. We read their review with interest; however, we disagree on several points.

First, in many cases of late-onset dementia, differentiating between the common causes of Alzheimer's disease and vascular dementia is difficult. In everyday clinical practice even differentiating Alzheimer's disease from Lewy-body disease and frontal—temporal dementia is not always feasible. To what extent these distinctions are relevant to genetic counselling with respect to late-onset dementia is not clear.

Second, the very high prevalence of dementia found in centenarians (Reference Asada, Yamagata and KilnoshitaAsada et al, 1996; Reference Blansjaar, Thomassen and van SchaickBlansjaar et al, 2000) is not the only argument against a slowing down of the rate of increase in dementia over 85, 90 or 95 years of age. Meta-analyses, not included in the review, did not find evidence for such a slowing down (Reference Gao, Hendrie and HallGao et al, 1998; Reference Jorm and JolleyJorm & Jolley, 1998). Therefore, the prevalence of dementia almost certainly increases substantially, exceeding 15% from the age of 85.

Most investigations attributed some three-quarters of late-onset dementia to Alzheimer's disease. We agree that the literature indicates a three— to fourfold risk in first-degree relatives of patients with late-onset dementia (seven— to eightfold with two affected first-degree relatives). We can only conclude that this leads to a risk of one in three, if not higher, for those first-degree relatives who reach the age of 85 years. Obfuscating this information by showing graphs to anxious relatives is, in our opinion, not an appropriate reassurance. We feel that better consolation can be effected by proffering the view that most people do not reach the age of 85, and by explaining the slowly progressive course of most cases of late-onset dementia.