Skip to main content
. 2015 Mar 24;181(9):672–679. doi: 10.1093/aje/kwu331

Table 2.

Associations of Midlife Overweight and Obesity With Measures of Brain Structure and Vascular Damage Assessed via Magnetic Resonance Imaging 26 Years Later in Late Life (n = 3,864), the AGES-Reykjavik Study, 1967–2006

Model and Weight Categorya % TBV
% WMLV
BI
CMB
Dementia
Meanb (SE) βc 95% CI Meanb (SE) βc 95% CI RRd 95% CI RRd 95% CI RRd 95% CI
Model 1e
 Normal weight 72.19 (0.07) 0.92 (0.02) 1.00 Referent 1.00 Referent 1.00 Referent
 Overweight 72.10 (0.06) 0.90 (0.02) 1.21 1.04, 1.41 1.04 0.76, 1.41 1.24 0.93, 1.65
 Obese 72.02 (0.13) 0.89 (0.03) 1.06 0.79, 1.41 0.90 0.50, 1.62 0.86 0.46, 1.59
 BMI, continuous 0.29 0.03, 0.56 −0.02 −0.09, 0.04 1.21 0.95, 1.55 1.20 0.78, 1.84 1.27 0.9, 1.79
Model 2f
 Normal weight 72.14 (−0.07) 0.93 (0.02) 1.00 Referent 1.00 Referent 1.00 Referent
 Overweight 72.14 (0.06) 0.89 (0.02) 1.16 0.99, 1.36 0.98 0.72, 1.33 1.23 0.92, 1.64
 Obese 72.13 (0.13) 0.86 (0.03) 0.97 0.72, 1.29 0.82 0.44, 1.50 0.85 0.46, 1.57
 BMI, continuous 1.13 −0.38, 2.64 −0.04 −0.11, 0.03 1.20 0.95, 1.53 1.17 0.77, 1.79 1.26 0.89, 1.78
Model 3g
 Normal weight 72.11 (0.07) 0.94 (0.02) 1.00 Referent 1.00 Referent 1.00 Referent
 Overweight 72.17 (0.06) 0.89 (0.02) 1.08 0.92, 1.27 0.91 0.67, 1.25 1.22 0.92, 1.63
 Obese 72.22 (0.13) 0.84 (0.03) 0.82 0.61, 1.10 0.69 0.37, 1.32 0.91 0.49, 1.72
 BMI, continuous 0.90 −0.58, 2.38 −0.04 −0.11, 0.03 1.17 0.92, 1.50 1.16 0.75, 1.79 1.27 0.89, 1.81

Abbreviations: BI, infarct-like brain lesion; BMI, body mass index; CI, confidence interval; CMB, cerebral microbleed; RR, relative risks; SE, standard error; TBV, total brain volume; WMLV, white matter lesions volume.

a Normal weight was defined as a body mass index (measured as weight in kilograms divided by height in meters squared) of 18.5–24.9; overweight was defined as a body mass index of 25–29.9; and obese was defined as a body mass index ≥30.

b Mean of brain volumes were adjusted for intracranial volume and for the covariates included in models.

c β coefficients from linear regression models of percentage of TBV and percentage of WMLV per unit difference in body mass index.

d Relative risks representing the risk of ≥2 BIs, ≥2 CMBs, and dementia from Poisson models with robust 95% confidence intervals.

e Model 1 was adjusted for sex, age in midlife, no. of years of follow-up, and intracranial volume.

f Model 2 was adjusted for the variables in model 1 and Apolipoprotein ε4 allele, educational level, physical activity level, systolic and diastolic blood pressures, and serum cholesterol level in midlife.

g Model 3 was adjusted for the variables in model 2 and coronary artery calcium, prevalent coronary heart disease, hypertension, diabetes, and depression, alcohol drinking status, and smoking habits in late life. The models for percentage of TBV were further adjusted for BI, and the models for dementia were also adjusted for percentage of WMLV.