Early cognitive instability predicts persistent decline: a 10 year longitudinal study in a rural, low resource community.

INTRODUCTION

Cognitive stability, defined as the maintenance of cognitive performance within expected variability, may represent an early marker of resilience during aging. However, its determinants remain poorly understood, particularly in low resource settings. This 10 year population based study examined predictors of cognitive stability and long term cognitive trajectories in a homogeneous rural cohort.

METHODS

Residents of rural Ecuador (median age: 60 years) completed the Montreal Cognitive Assessments (MoCA) at baseline, short‑term, and long‑term follow‑ups. Cognitive stability was defined as ≤2 points decline between the first and second MoCA (short‑term) and ≤3 points between the first and third MoCA (long‑term), thresholds aligned with test-retest variability. Baseline determinants included demographics, cardiovascular health metrics, sleep quality, oily fish intake, depressive symptoms, and social risk. Mixed effects models evaluated continuous cognitive trajectories and categorical stability.

RESULTS

Of 830 participants with baseline MoCA, 637 completed the second assessment and 523 the third. Short‑term cognitive stability was common (74%), but long‑term stability persisted in only 40%. No participant classified as unstable at the short‑term regained stability at long‑term follow‑up, indicating that early instability marked the onset of irreversible decline. In multivariable models, younger age, higher educational attainment, and low social risk were independently associated with higher cognitive performance or greater odds of stability.

CONCLUSION

Cognitive stability is frequent over short intervals but declines sharply across a decade. Early instability appears prognostic of persistent decline. Social context emerged as a robust determinant of resilience, underscoring the need to integrate social and biological factors when addressing cognitive aging in underserved rural populations.