Background Ambient air pollution has been consistently associated with exacerbation of respiratory diseases in schoolchildren, but the role of early exposure to traffic-related air pollution in the first occurrence of respiratory symptoms and asthma is not yet clear. Methods We assessed the association between indexes of exposure to traffic-related air pollution during different periods of life and respiratory outcomes in a birth cohort of 672 newborns (Rome, Italy). Direct interviews of the mother were conducted at birth and at 6, 15 months, 4 and 7 years. Exposure to traffic-related air pollution was assessed for each residential address during the follow-up period using a Land-Use Regression model (LUR) for nitrogen dioxide (NO2) and a Geographic Information System (GIS) variable of proximity to high-traffic roads (HTR) (>10 000vehicles/day). We used age-specific NO2 levels to develop indices of exposure at birth, current, and lifetime time-weighted average. The association of NO2 and traffic proximity with respiratory disorders were evaluated using logistic regression in a longitudinal approach (Generalised Estimating Equation). The exposure indexes were used as continuous and categorical variables (cut-off points based on the 75th percentile for NO2 and the 25th percentile for distance from HTRs). Results The average NO2 exposure level at birth was 37.2 μg/m3 (SD 7.2, 10–90th range 29.2–46.1). There were no statistical significant associations between the exposure indices and the respiratory outcomes in the longitudinal model. The odds ratios for a 10-µg/m3 increase in time-weighted average NO2 exposure were: asthma incidence OR=1.09; 95 CI% 0.78 to 1.52, wheezing OR=1.07; 95 CI% 0.90 to 1.28, shortness of breath with wheezing OR=1.16; 95 CI% 0.94 to 1.43, cough or phlegm apart from cold OR=1.11; 95 CI% 0.92 to 1.33, and otitis OR=1.08; 95 CI% 0.89 to 1.32. Stronger but not significant associations were found considering the 75th percentile of the NO2 distribution as a cut-off, especially for incidence of asthma and prevalence of wheeze (OR=1.41; 95 CI% 0.88 to 2.28 and OR=1.27; 95 CI% 0.95 to 1.70, respectively); the highest OR was found for wheezing (OR=2.29; 95 CI% 1.15 to 4.56) at the 7-year follow-up. No association was found with distance from HTRs. Conclusions Exposure to traffic-related air pollution is only weakly associated with respiratory symptoms in young children in the first 7 years of life.