PURPOSE: We examined the association between childhood respiratory infections and adult lung function and how this association varies depending on the age at infection. METHODS: The Barry Caerphilly Growth study collected information on childhood upper and lower respiratory tract infections (URTI, LRTI) from birth to 5 years on 14 occasions. Subjects were traced prospectively and had lung function measured at age 25 years. RESULTS: A total of 581 subjects had acceptable data for both forced expiratory volume in one second (FEV1) and forced vital capacity (FVC). Childhood LRTIs (0-5 years) but not URTIs were negatively associated with all lung function measures except FVC and showed a dose-response effect. In the first year of life, a two-fold increase in the number of LRTIs experienced was associated with a reduction in FEV1 (78 mL; 95% confidence interval [95% CI], 3-153), FEV1/FVC (1.23%; 95% CI 0.25-2.22), forced expiratory flow 25%-75% (0.25 l/sec; 95% CI 0.08-0.41), and peak expiratory flow (0.30 l/sec; 95% CI 0.11-0.49) after adjustment for confounders. Few associations were found after the first year of life. There was evidence that age at infection effect modifies the association between LRTIs and FEV1, forced expiratory flow 25%-75%, and peak expiratory flow. CONCLUSIONS: LRTIs are associated with an obstructive lung function deficit. Furthermore, the first year of life may be a sensitive period for experiencing LRTIs.