Lower body mass index is associated with greater Alzheimer pathology in asymptomatic individuals


      Large-scale cross-sectional and longitudinal epidemiological studies indicate that chronic metabolic dysfunction such as diabetes, insulin resistance, and obesity increases the risk of developing Alzheimer disease (AD) (Akomolafe et al., 2007; Crane et al, 2013). This relationship has led some researchers to characterize AD as a metabolic syndrome. Conversely, a subset of work has demonstrated an almost protective relationship between AD pathology and weight or insulin resistance (Burns et al., 2007; 2012).


      Participants were 100 cognitively normal individuals (mean age of 72.1, stdev 5.1, 50 females) with a Clinical Dementia Rating (CDR) of 0. All participants had measures of body mass index (BMI), APOE genotyping, and a lumbar puncture to measure cerebrospinal fluid (CSF) levels of Aβ42, tau, and ptau181. 64 individuals also had positron emission tomography (PET) amyloid imaging using 11[C]Pittsburgh Compound B (PiB). Regression modeling examined the main effects of age, gender, APOE genotype(ε4+/-) and BMI.


      In all models advancing age and the presence of the ε4 allele were associated with greater AD pathology, while increasing BMI was associated with less pathology (Figure 1). In the model examining amyloid PET there was a significant effect of age (B=0.06, F1,59=11.0, p<0.01), APOE genotype (B=0.7, F1,59=16.6, p<0.01), and BMI (B=-0.05, F1,59=6.4 p<0.01). For ptau181 there was a significant effect of age (B=1.3, F1,95=4.4, p<0.05) and trends for both APOE (B=13.9, F1,95=3.6, p=0.06), and BMI (B=-1.2, F1,94=3.7, p=0.06). For tau there were significant effects for age (B=10.9, F1,95=13.1 p<0.001), APOE genotype (B=74.3, F1,95=4.1, p<0.05), and BMI (B=-9.4, F1,95=6.8, p<0.05). For CSF Aβ42 there was a significant effect of age (B=-17.9, F1,95=10.7, p<0.01) and APOE genotype (B=-211.7, F1,95=13.3, p<0.001).


      Lower values of BMI were associated with elevated levels of AD pathology in cognitively normal older adults. The mean BMI for the sample was 26.8 and only three individuals were underweight (BMI<18.5), suggesting individuals with very low weight did not drive this effect. Instead the effects may represent a subtle loss in appetite or behavioral alterations in meal preparation as AD pathology accumulates.
      Figure thumbnail fx1
      Figure thumbnail fx2