Nonlinear association between age and verbal fluency and moderation by adiposity: cross-sectional evidence from the Look Up 8+ study

Background – Verbal fluency (VF) declines with age, but the shape of this association and its modifiers remain debated. In community-dwelling adults aged 30–80 years from the Look Up 8+ study, we investigated the nonlinear association between age and phonemic VF and evaluated the impact of educational attainment, lifestyle factors, central adiposity, and muscle function on VF performance. Methods – Verbal fluency was assessed using a 30-s phonemic fluency task. Generalized additive models (GAMs) with negative binomial distribution characterized nonlinear age-related trajectories. Sequential models adjusted for sex, educational attainment, physical activity (PA), healthy diet, smoking status, blood pressure, total cholesterol, blood glucose, body mass index, and sleep quality. Central adiposity was evaluated using waist-height ratio (WHtR), with waist-hip ratio (WHR) included in sensitivity analyses. Moderation analyses tested whether central adiposity modified the association between age and VF. Associations between VF and muscle- parameters were also examined. Results – Among 3, 615 adults (54% female; median age 58 years; median VF score 10 words), the age-VF association was significantly nonlinear (edf = 3.36, p < 0.001), peaking at 45–50 years and showing progressively lower values thereafter. Higher educational attainment was strongly associated with greater VF: high school/BSc (IRR = 1.218, 95% CI 1.168–1.271, p < 0.001) and master's/postgraduate (IRR = 1.326, 95% CI 1.269–1.385, p < 0.001) versus lower education. Healthy diet (IRR = 1.044, p < 0.001), PA (IRR = 1.038, p = 0.005), and normal blood glucose (IRR = 1.050, p = 0.022) were independently associated with higher VF; the dietary association was confirmed for Mediterranean diet adherence. Higher central adiposity was associated with lower VF for both WHtR (IRR = 0.973, p < 0.001) and WHR (IRR = 0.979, p = 0.003) and significantly moderated the age-VF association (Age × WHtR: IRR = 0.985, p = 0.014; Age × WHR: IRR = 0.986, p = 0.021). Among muscle-related parameters, longer chair stand test time was associated with lower VF (IRR = 0.964, p < 0.001). Conclusion – Age and VF are nonlinearly associated across adulthood, with educational attainment being a major determinant. Higher central adiposity is associated with lower VF and moderates the associated with age. Lower-limb functional performance was independently associated with VF, potentially reflecting a functional pathway linking adiposity, mobility, and cognitive aging.