Baseline and longitudinal trajectories of peak expiratory flow in relation to incident cognitive impairment among older adults: Evidence from CHARLS

Background Previous studies, including CHARLS analyses, have linked lower peak expiratory flow (PEF) with cognitive decline. Evidence regarding incident score-based cognitive impairment, longitudinal PEF patterns, and incremental prediction remains limited. Objective To examine associations of baseline PEF and exploratory PEF trajectories with incident score-based cognitive impairment and evaluate the incremental predictive value of PEF. Methods The primary analysis included 2604 CHARLS participants aged ≥60 years without score-based cognitive impairment at baseline. Cognitive impairment was defined as a global cognitive score ≥1 SD below the wave- and age-specific mean and was not a clinical diagnosis of mild cognitive impairment or dementia. Cox models, sensitivity analyses, prediction analyses, and exploratory trajectory modeling were performed. Results Over a median follow-up of 9.0 years, 636 participants developed score-based cognitive impairment. Each 1-SD increase in PEF was associated with lower risk in fully adjusted models (HR = 0.801, 95% CI 0.728–0.880) and after adjustment for baseline cognition (HR = 0.846, 95% CI 0.768–0.930). Nine-year cumulative incidence decreased from 35.6% to 16.6% across PEF quartiles. Adding PEF modestly improved 9-year AUC (ΔAUC = 0.011, p = 0.042). In the exploratory trajectory analysis, the moderate PEF group was not significantly associated with risk versus the high group (HR = 1.294, 95% CI 0.860–1.948), whereas the low PEF group had higher risk (HR = 1.817, 95% CI 1.172–2.817), with no clear graded pattern. Conclusions Lower baseline PEF was associated with wave-specific score-based cognitive impairment, but not under stricter definitions requiring subsequent confirmation with wave-specific or fixed thresholds. PEF trajectories were exploratory, and incremental predictive value was limited.