Does Preventing Stroke Prevent All Kinds of Dementia?
To the Editor:
The recent publication in Stroke regarding dementia after stroke1 makes interesting observations. However, there are some points in the article to which I would like to draw attention.
First, the subgroup analysis in the text mentions that the ε4 allele for apolipoprotein E (ApoE) genotype was present in 17.3% of the cases and in 22% of the controls. Table 1 in the article seems to have a printing error in that the above percentages have been reversed in the table (ie, 22% of cases and 17% of controls have ε4 allele for ApoE genotype).
Second, the changing denominators in Table 1 show that a lot of baseline data were actually missing. A mention of this would have been appropriate in the limitations of the study.
Last, it seems logical to conclude that the prevention of stroke would reduce the burden of vascular dementia. The authors conclude that primary and secondary prevention of stroke should significantly decrease the risk of all dementia. I wonder how valid that observation could be. It is true that a proportion of stroke cases had Alzheimer disease (AD), but the proportion of AD still remained higher among controls. There is no statistical analysis in this particular study to suggest that stroke increases the incidence of AD or precipitates AD in a vulnerable individual; ie, it is possible that the stroke patients with AD would have developed AD even if the stroke were prevented. Perhaps a study to compare AD patients with and without a history of stroke could look into the question of how the pathology in AD is influenced by the occurrence of a stroke.
Ivan CS, Seshadri S, Beiser A, Au R, Kase CS, Kelly-Hayes M, Wolf PA. Dementia after stroke: the Framingham Study. Stroke. 2004; 35: 1264–1268.
We thank Dr Chakraborty for her interest in our study and for raising an important issue. We agree that our study does not (and as currently designed can not) address “the question of how the pathology of Alzheimer disease (AD) is influenced by the occurrence of a stroke.” However, previous studies have suggested that ischemia may increase β-amyloid deposition in mouse models of AD1 and that the severity of cognitive impairment detected prior to death was greater in subjects who had brain infarcts at autopsy as compared with infarct-free subjects with equivalent degrees of Alzheimer pathology.2 In our study, we observed that the overall risk of developing a dementia (regardless of type; ie, “all dementia”) doubled in subjects who had sustained a stroke as compared with those who had not sustained a stroke; we thus concluded that “prevention of stroke should significantly reduce the risk of all dementia.” We chose not to report the relative risks of developing individual subtypes of dementia because we did not have a sufficient number of cases to permit such analyses. When we compared the relative risk (RR) of developing either AD or mixed dementia in stroke cases and controls, there was no difference between the groups, but the wide confidence intervals do not permit a definite conclusion. The crude RR was 1.0 (CI, 0.6 to 1.9), the RR adjusted for age, sex, and education was 1.0 (CI, 0.6 to 1.7). These compare to the RRs of 2.2 and 2.0 for all dementias. Our data can thus neither confirm nor refute Dr Chakraborty’s suggestion that those who developed AD after a stroke would have done so even in the absence of a stroke. As more of these subjects donate their brains for post-mortem studies we hope to address this interesting question in future studies.
We also thank Dr Chakraborty for detecting a typographical error in Table 1 of our article. Both the table and the text correctly state the actual number of subjects (cases and controls) with at least one apolipoprotein E (ApoE) ε4 allele as 22/127 (17%) subjects with a stroke and 153/694 (22%) of subjects without a stroke. The equivalent percentage was incorrectly stated in Table 1 (although it was correctly noted in the text).
We acknowledge the absence of baseline covariate data in some subjects, and have listed in Table 2 the actual denominator used for each subgroup analysis. ApoE genotype data were missing in 37% of subjects; many of these subjects were enrolled (and died) prior to initiation of ApoE genotyping in the Framingham study population.3 However, we consider it unlikely that the missing data affect the validity of our results because the unadjusted analysis, the analysis adjusting for age, sex, and education alone (data available in 97% of the subjects), and the analysis adjusting for stroke type and presence or absence of second stroke (data available in 97%) were all significant and these relative risks were also ≥2.0.
Sadowski M, Pankiewicz J, Scholtzova H, Ji Y, Quartermain D, Jensen CH, Duff K, Nixon RA, Gruen RJ, Wisniewski T. Amyloid-beta deposition is associated with decreased hippocampal glucose metabolism and spatial memory impairment in APP/PS1 mice. J Neuropathol Exp Neurol. 2004; 63: 418–428.
Myers RH, Schaefer EJ, Wilson PW, D’Agostino R, Ordovas JM, Espino A, Au R, White RF, Knoefel JE, Cobb JL, McNulty KA, Beiser A, Wolf PA. Apolipoprotein E ε4 association with dementia in a population-based study: the Framingham Study. Neurology. 1996; 46: 673–677.