Haroon et al. (1) conducted a cohort study targeting seniors aged 69–78 years with and without diabetes to elucidate the predictors of dementia. They reported that the risk of dementia in patients with diabetes was 16% higher than that in patients without diabetes. I have some queries on their study.
First, hypertension was a protective factor for the incidence of dementia. Although the significance disappeared after adjusting some information on medication, other baseline disorders, such as cerebrovascular, coronary artery, peripheral vascular, and chronic kidney diseases, were risk factors for dementia. The authors speculated that the lack of association or protective effect of hypertension on dementia was partly mediated by medication such as statins and calcium channel blockers, although there was no special way to elucidate the net effect of hypertension on dementia. Depression is a risk factor for dementia and hypertension (2), and psychological factors in the elderly should be considered for the analysis.
Second, Morris et al. (3) conducted a 2-year follow-up study to monitor the effect of impaired glycemia on disease progression in patients with mild cognitive impairment and concluded that baseline glycemia is related to cognitive decline and progression to Alzheimer disease. The mean age of subjects was about 75 years, and this report presented the adverse effect of impaired glycemia on dementia.
Finally, Cheng et al. (4) reported that the relative risks (95% CI) of Alzheimer disease and vascular dementia for diabetes were 1.46 (1.20–1.77) and 2.49 (2.09–2.97), respectively. In addition, the relative risk (95% CI) of mild cognitive impairment for diabetes was 1.22 (1.0–1.45). These data are calculated by random-effect models.
I suppose that Haroon et al. (1) can present the risk of diabetes for dementia by separating the type of dementia, such as Alzheimer disease and vascular dementia, just as Cheng et al. (4) explained the lower risk of diabetes for mild cognitive impairment than for dementia by insufficient number of samples. In any case, sensitivity analysis by separating the type of dementia as dependent variable should be considered for the analysis of the study by Haroon et al. (1).
Article Information
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
