Background Recent evidence suggests that cumulative low-level lead exposure has adverse effects on cognitive function in the elderly. To date, the few studies that have measured bone lead exposure relied on K-X-ray fluorescence (KXRF), methods that are mostly unavailable in large community-based studies. Here, we employ a methylation-based estimation method for bone and blood lead in the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort.

Methods Tibia and patella bone as well as blood lead levels were estimated using blood DNA methylation (DNAm) biomarkers in 625 participants from the ADNI cohort. Cognitive function was assessed using the Montreal Cognitive Assessment (MoCA) test. Longitudinal analyses of bone lead and cognitive decline were conducted using linear mixed-effect regression models, as participants had different years of baseline (range from 2010-2014) and follow-up visits (years of last visits range from 2014-2015).

Results DNAm derived tibia and patella lead levels were negatively associated with MoCA scores throughout follow-up, while DNAm derived blood lead level was not associated with MoCA scores. On average, we observed lower MoCA scores with increasing DNAm tibia lead (per interquartile range (IQR): β=-0.23; 95%CI: -0.44, -0.03), and DNAm patella lead, albeit the later was weaker (per IQR: β=-0.19; 95%CI: -0.41, 0.04). When stratifying by gender, women showed a stronger decrease in cognitive function with increasing DNAm tibia lead (per IQR β=-0.34; 95%CI: -0.65, -0.04) than men (per IQR β=-0.15; 95%CI: -0.42, 0.13). The estimated decrease in MoCA scores per DNAm tibia lead IQR increase was stronger among participants with one or two APOE4 alleles (per IQR β=-0.37; 95%CI: -0.74, -0.01) than those with zero alleles (per IQR β=-0.14; 95%CI: -0.38, 0.10).

Conclusion These findings strengthen the evidence that cumulative long-term lead exposures levels are associated with decreased cognitive function in the elderly, especially women and carriers of one or two APOE4 alleles. These findings based on whole blood methylation data corroborate previous epidemiologic studies that used KXRF for measuring bone lead.