Additive Burden of Low Peak Expiratory Flow and Sleep Disturbance on Incident Depressive Symptoms in Later Life: A Four-Cohort Longitudinal Study of 49,201 Participants

Background: Lung function and sleep are each prospectively associated with late-life depressive symptoms, but whether their co-occurrence carries risk beyond the sum of the two components has not been tested across national populations. Methods: We quantified the joint association of low peak expiratory flow (PEF) and sleep disturbance with incident depressive symptoms in four ageing cohorts and formally tested additive and multiplicative interaction. We analysed 49,201 at-risk participants aged 50 to 100 years from the Health and Retirement Study (US), the English Longitudinal Study of Ageing, the China Health and Retirement Longitudinal Study and the Survey of Health, Ageing and Retirement in Europe (16 countries). Low PEF was defined as the lowest quartile within sex and age bands; sleep disturbance used each cohort’s established measure. Incident elevated depressive symptoms were modelled with Poisson generalized estimating equations with robust variance and random-effects meta-analysis. Results: The joint-exposure group had a higher adjusted rate of incident elevated symptoms in every cohort (pooled aRR 1.40, 95% CI 1.28 to 1.52; I2 = 41%), corresponding to 7 to 25 additional events per 1000 person-years; component associations were smaller (low PEF only aRR 1.14; sleep disturbance only 1.27). No evidence of interaction was found on either scale, with complete cross-cohort consistency (multiplicative interaction RR 0.95, 0.89 to 1.02; additive RERI −0.03, −0.11 to 0.06; both I2 = 0%). Findings persisted in zero-symptom-baseline, sex-stratified, attrition-weighted and competing-risk analyses and were replicated in ELSA when low lung function was defined by spirometric FEV1 (aRR 1.33) or FVC (aRR 1.23) instead of PEF. Conclusions: Co-occurring low PEF and sleep disturbance mark a consistent, additive burden of incident depressive symptoms in later life; combined assessment should be read as simple risk addition, not as a synergistic high-risk phenotype.