A longitudinal study spanning up to 27 years identified accelerated brain atrophy as a primary predictor of cognitive decline, with type 2 diabetes and abnormal amyloid-β ratios presenting as significant synergistic risk factors for progression from normal cognition to mild cognitive impairment.

Published in JAMA Network Open, the study tracked 185 cognitively normal, middle-aged and older adults, providing insight into brain structure changes that foreshadow cognitive deterioration. Participants with increased rates of white matter atrophy and ventricular enlargement exhibited earlier mild cognitive impairment (MCI) progression. The study found that type 2 diabetes and low CSF abnormal amyloid-β Aβ₄₂/β₄₀ ratios independently raised the risk of developing MCI (HR 1.41 and 1.48, respectively). However, the combination of these factors produced an even higher hazard ratio of 1.55, indicating a synergistic effect.

Researchers analyzed 951 MRI scans across a median follow-up of 20 years, with participants undergoing an average of 5 scans. Using advanced deep learning–based brain parcellation (OpenMAP-T1) and longitudinal data harmonization, they measured annual volume changes:

• Cortical gray matter: -0.0053%
• White matter: -0.0229%
• Ventricles: +0.0089%
• Sulci: +0.0144%

Change-point analyses identified white matter atrophy acceleration beginning at age 62.4, and ventricular enlargement beginning at age 68.9.

Among the 185 participants (mean age 55.4 years, 63% women), 60 progressed to MCI, with 8 of these later developing dementia. The cohort was predominantly White and well-educated. No significant associations were identified between MCI progression and other vascular risk factors, such as hypertension, dyslipidemia, smoking, or with CSF tau biomarkers (p-tau181 and t-tau) and APOE ε4 status.

The study’s limitations included a relatively small sample size, an imbalanced sex distribution, and potential imaging variability due to MRI scanner upgrades over time. The homogeneity of the study population limits generalizability to more diverse groups.

In an invited commentary, Shohei Fujita, MD, PhD, noted the need for more diverse study populations and further investigation into gender and racial differences in neurodegeneration patterns. He highlighted the potential of long-term brain imaging data to inform preventive strategies and the importance of improving imaging methods for consistent long-term use.

Full disclosures are available in the study and commentary.