Body mass index and risk of dementia: Analysis of individual-level data from 1.3 million individuals

Abstract

Introduction: Higher midlife body mass index (BMI) is suggested to increase the risk of dementia, but weight loss during the preclinical dementia phase may mask such effects.

Methods: We examined this hypothesis in 1,349,857 dementia-free participants from 39 cohort studies. BMI was assessed at baseline. Dementia was ascertained at follow-up using linkage to electronic health records (N = 6894). We assumed BMI is little affected by preclinical dementia when assessed decades before dementia onset and much affected when assessed nearer diagnosis.

Results: Hazard ratios per 5-kg/m² increase in BMI for dementia were 0.71 (95% confidence interval = 0.66-0.77), 0.94 (0.89-0.99), and 1.16 (1.05-1.27) when BMI was assessed 10 years, 10-20 years, and >20 years before dementia diagnosis.

Conclusions: The association between BMI and dementia is likely to be attributable to two different processes: a harmful effect of higher BMI, which is observable in long follow-up, and a reverse-causation effect that makes a higher BMI to appear protective when the follow-up is short.

Keywords: Bias; Body mass index; Cohort study; Dementia; Obesity.

Fig. 1

Conceptual model: Effect of reverse causation (preclinical disease reduces weight) on BMI at different etiological periods before dementia diagnosis. Abbreviation: BMI, body mass index.

Fig. 2

Age-, sex-, and ethnicity-adjusted HR for incident dementia per 5-kg/m² increase in BMI after progressive exclusion of the follow-up period in all studies (A)^∗ and in studies with dementia ascertainment using dementia morbidity data (B)^†. ^∗39 studies, total N = 1,349,857. ^†5 studies, total N = 95,851. The figure shows that risk of bias due to preclinical dementia decreases with increasing exclusion of the follow-up period. Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio.

Fig. 3

Shape of the association between BMI and dementia before (A) and after (B) exclusion of the first 20 years of follow-up. ^∗39 studies, total N = 1,349,857. ^†10 studies, total N = 391,596. The figure shows that risk of bias due to preclinical dementia is smaller after exclusion of follow-up.

Fig. 4

Age-, sex-, and ethnicity-adjusted hazard ratio for dementia per 5-kg/m² increase in BMI in analysis stratified by follow-up period in all studies (A) and in studies with dementia ascertainment using dementia morbidity data (B). ^∗39 studies, total N = 1,349,857. ^†5 studies, total N = 95,851. The figure shows that risk of bias due to preclinical dementia is smaller at later follow-up periods. Abbreviations: BMI, body mass index; CI, confidence interval; HR, hazard ratio.