: We conducted a prospective cohort study to examine the risk of incident mild cognitive
impairment (MCI) as predicted by baseline neuropsychiatric symptoms (NPS) and brain
regional glucose metabolic dysfunction.
: 1363 cognitively unimpaired (CU) individuals (52.8% males) aged ≥ 50 years were
followed for a median of 4.8 years to the outcome of incident MCI. NPS were assessed
using Beck Depression and Anxiety Inventories (BDI-II, BAI) and Neuropsychiatric Inventory
Questionnaire. Glucose hypometabolism was measured by FDG-PET and defined as SUVR
≤ 1.47 in regions typically affected in Alzheimer’s disease. Cox proportional hazards
models were adjusted for age, sex, education and
APOE ε4 status.
: Participants with regional glucose hypometabolism and depression (BDI-II ≥ 13) had
a more than threefold increased risk of incident MCI (HR [95% CI], 3.66 [1.75, 7.65],
p < 0.001, χ
2 = 11.83, df = 1) as compared to the reference group (normal regional glucose metabolism
and no depression), and the risk was also significantly elevated (7.21 [3.54, 14.7],
p < 0.001, χ
2 = 29.68, df = 1) for participants with glucose hypometabolism and anxiety (BAI ≥
10). Having glucose hypometabolism and ≥ 1 NPS (3.74 [2.40, 5.82], p < 0.001, χ
2 = 34.13, df = 1) or ≥ 2 NPS (3.89 [2.20, 6.86], p < 0.001, χ
2 = 21.92, df = 1) increased the risk of incident MCI by more than three times, and
having ≥ 3 NPS increased the risk by more than four times (4.12 [2.03, 8.37], p <
2 = 15.39, df = 1).
: Combined presence of NPS with regional glucose hypometabolism is associated with
an increased risk of incident MCI, with FDG-PET appearing to be a stronger driving
force of cognitive decline than NPS.