The etiology of long‐term stable mild cognitive impairment

Abstract

Background Of those diagnosed with mild cognitive impairment (MCI), 5‐10% convert annually to dementia (Mitchell & Shiri‐Feshki, 2009). However, 5‐30% can remain cognitively stable (sMCI) for up to 10 years (Alves et al., 2018), suggesting a non‐neurodegenerative etiology in these patients. This study aims to examine the etiological underpinnings of those who remain cognitively stable versus those who decline over time. Method Participants from the National Alzheimer’s Coordinating Center were selected if they were nondemented at baseline, had available clinical data ≥5 years following baseline, and had available autopsy data <1.5 years following last visit. They were defined as stable normal controls (sNC; n = 106) or sMCI (n = 28) if cognitive status was unchanged between first and last visit (≥5 years later); declined normal controls (dNC; n = 128) if cognitive status at first visit was normal and MCI or demented by last visit; or declined MCI (dMCI; n = 139) if cognitive status was MCI at first visit and demented by last visit. Neuropathological features were dichotomized to reflect clinically significant levels (≥moderate neuritic plaque density; Braak neurofibrillary degeneration stage ≥4; Thal amyloid plaque phase ≥3; any Lewy bodies; ≥1 large arterial infarct; ≥1 lacune; ≥1 microbleed; ≥1 hemorrhage). Baseline psychiatric symptoms were quantified using the Geriatric Depression Scale (GDS; ≥6 considered clinically significant) and the Neuropsychiatric Inventory Questionnaire (NPI‐Q; grouped into ‘mood’, ‘agitation’ and ‘psychosis’ factors based on previous factor analysis: Aalten et al., 2003, Siafarikas et al., 2018). The groups were compared on frequencies of neuropathological features at autopsy and baseline psychiatric symptoms using chi‐square (χ 2). Result Demographics are presented in Table 1, with main analysis results in Table 2. Relative to other groups, sMCI had significantly fewer amyloid plaques (χ 2 = 44.2, p < 0.001) and more microinfarcts (χ 2 = 9.9, p = .019). GDS and NPI‐Q showed no significant relationship with sMCI. Conclusion Previous research indicates microinfarcts are independent of neurodegenerative disease (Launer et al., 2011). In line with this, our results suggest that sMCI may be caused by microinfarcts, rather than frank neurodegenerative disease. This study was limited as only neuropsychiatric symptoms were reported, and future studies should examine sMCI in relation to diagnosed psychiatric disorders.