Your brain doesn’t look a day past 70! Cross-sectional associations with brain-predicted age in the cognitively-intact oldest-old

The cognitively-intact oldest-old (85 +) may be the most-resilient members of their birth cohort; due to survivorship effects (e.g., depletion of susceptibles), risk factors associated with brain aging biomarkers in younger samples may not generalize to the cognitively-intact oldest-old. We evaluated associations between established aging-related risk factors and brain-predicted age difference (brainPAD) in a cross-sectional cognitively-intact oldest-old sample. Additionally, we evaluated brainPAD-cognition associations to characterize brain maintenance vs. cognitive reserve in our sample. Oldest-old adults (N = 206; 85-99 years; Montreal Cognitive Assessment > 22 or neurologist evaluation) underwent T1-weighted MRI; brainPAD was generated with brainageR, such that more-positive brainPAD reflected more-advanced brain aging. Sex, education, alcohol and smoking history, exercise history, BMI, cardiovascular and metabolic disease history, and anticholinergic medication burden were self-reported. Global cognitive z-score and coefficient of variation were derived from the UDS 3.0 cognitive battery; crystallized-fluid discrepancy was derived from the NIH Toolbox Cognitive Battery. Mean brainPAD was -7.99 (SD: 5.37; range: -24.50, 6.03). Women showed more-delayed brain aging than men (B = -2.9, 95% CI = -4.6, -1.1, p = 0.002). No other exposures were significantly associated with brainPAD. BrainPAD was not associated with any cognitive variable. These findings suggest that cognitively-intact oldest-old adults may be atypically-resistant to risk factors associated with aging in younger samples, consistent with survivorship effects in aging. Furthermore, brainPAD may have limited explanatory value for cognitive performance in cognitively-intact oldest-old adults, potentially due to high cognitive reserve. Overall, our findings highlight the impact of survivorship effects on brain aging research.