For the estimated 15–20% of older adults living with mild cognitive impairment, the question of who progresses to dementia is not purely neurological — it may be deeply psychiatric. Identifying modifiable psychological risk factors at this transitional stage could open a critical window for intervention before irreversible neurodegeneration takes hold. Depression, long associated with cognitive decline, is now under sharper quantitative scrutiny as a potential accelerant.
This systematic review and meta-analysis synthesized 17 longitudinal cohort studies, spanning follow-up periods from six months to twelve years, to quantify depression's role in MCI-to-dementia conversion. Unadjusted analyses found that baseline depression was associated with a 66% elevated risk of overall progression to dementia (HR 1.66, 95% CI 1.22–2.26). When stratified by dementia subtype, the association with Alzheimer's disease specifically reached statistical significance at a 57% increased hazard (HR 1.57, 95% CI 1.15–2.15), while the estimate for all-cause dementia did not achieve significance, likely due to extreme variability across studies. Critically, heterogeneity was exceptionally high across analyses — I² values exceeding 99% — substantially undermining confidence in the pooled estimates.
The heterogeneity problem here is not a minor caveat; it is the central limitation. When studies diverge this dramatically in their individual effect sizes, a pooled number risks being statistically misleading. Differences in how depression was defined and measured — clinical diagnosis versus validated scales such as the GDS or CES-D — likely contributed substantially to this variance, as did varying MCI diagnostic criteria and population demographics across studies. From a mechanistic standpoint, the depression-dementia link is biologically plausible: chronic depressive states elevate cortisol, promote neuroinflammation, and are associated with hippocampal volume reduction — all pathways relevant to Alzheimer's pathophysiology. Whether depression is a true causal accelerant, an early symptom of subclinical neurodegeneration, or both remains unresolved. This meta-analysis is confirmatory in direction but inconclusive in magnitude, underscoring the need for standardized depression measurement protocols in future MCI cohort studies.