Answers are grounded in this paper and cite their sources.
To the Editor:—Some conclusions of the recent paper by Huff et al1 prompted us to share our preliminary findings on age at onset of Alzheimer's disease (AD). Our results, drawn from data coming from the population-based study IMAGE (Investigations de la Maladie d'Alzheimer: génétique et épidémiologie) which is being conducted in the Saguenay –Lac-Saint-Jean (SLSJ) area in the province of Québec (Canada) are quite different from those presented by Huff et al.1 The question as to whether our current definition of AD refers to a heterogeneous condition which can be divided in two distinct diseases on the basis of age at onset is still controversial. If this were the case, one might expect a bimodal distribution of AD cases by age at onset or, as Ruenberg,2 a discontinuous age-specific incidence of the two conditions. It has been stressed already that none of the population-based studies that estimated age-specific incidence rates for AD supported the age-related heterogeneity hypothesis.3 On the other hand, some of the studies that were based on a series of patients attending a particular hospital suggest that age at onset is an explanatory variable for the apparent heterogeneity of AD,4 but there is no consensus on this point.5 A registry of over 400 presumptive cases of senile dementia of the Alzheimer's type from the SLSJ area is now being screened for AD. The findings reported in this letter are based on definite, probable or possible cases (see reference 6 for definitions), and those for whom the age at onset is known. For most cases, age at onset was estimated by a next of kin at the time of medical examination. Onset was associated with the first signs of progressive memory loss in 55 cases (68%). A majority of cases were recruited from the participating hospital centers and through the office of the regional neurologists, whereas the remainder came from nursing homes of the SLSJ area. Figure 1 shows the distribution of cases by age at onset of the 81 patients of IMAGE and of those reported by Huff et al.1 Both sets of data were grouped by 5-year age intervals. Grouping seems a more appropriate way of summarizing this type of data, because it can be expected that estimates of age at onset are only accurate within a 2- to 3-year interval. Also, because so few cases (about 100 to 160 in this instance) were distributed in a range of more than 40 values (ie, years), on average there are less than four cases per year. . Distribution of Alzheimer's disease cases by age at onset to 74 years (P = 0.1), and none for the age-group 55 to 59 years (P = 0.24). Therefore, there is no statistical confirmation of the apparent bimodality reported in their study. The distribution of the IMAGE cases presents a single peak in the 65 to 69 year age-group. The Haldane test for bimodality7 shows a peak at 65 to 69 years (P = .07). The same test was applied to the data reported by Huff et al.1 The P-values for the two apparent peaks of the nongrouped data are: P = 0.88 at 60 years of age; and P = 0.28 at 68 years of age. For the grouped data, the P-values suggest a single peak for the age-group 70 In conclusion, it can be said that our data, as well as that reported by Huff et al, have a unimodal distribution. However, since a relatively small number of cases distributed over a large range of ages was analyzed, we cannot claim that these results are decisive. Nevertheless, we consider to be of interest to draw your attention to the distribution of the age at onset of our data, as well as to the use of the Haldane method when testing for bimodality.7
No takes yet. Share an insight, caveat, or question.
Froda et al. (1988) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: