Reduced pulmonary function is an important predictor of cardiovascular morbidity and mortality. The mechanisms underlying this association are unknown but may involve systemic inflammation. We assessed the cross-sectional and longitudinal relationships between C-reactive protein (CRP) levels and forced expiratory volume in 1s (FEV1) and its decline in the general population, over a period of 8.5 years. The analyzes were based on 531 subjects (mean age at baseline: 37+/-7 years, 50% women and 42% non-smokers), recruited at two French centers participating in the European Community Respiratory Health Survey. Lung function was expressed as a percentage of predicted FEV1. CRP was measured centrally, by means of a highly sensitive assay. In cross-sectional analysis, FEV1 as a % of predicted values was negatively associated with serum CRP concentration (P=0.002). Multivariate adjustment did not alter these results (P=0.002). In longitudinal analysis, annual FEV1 decline tended to increase from the lower to the upper tertile for baseline CRP concentration but the association was borderline significant (P=0.14). Mean values of annual FEV1 decline were 26+/-32, 31+/-32, and 34+/-32 ml/year for the lower, middle and upper tertiles of baseline CRP concentration, respectively, after adjusting for potential confounders (P=0.09). Changes in CRP levels during follow-up were associated with annual FEV1 decline. The mean annual FEV1 declines in subjects with increasing CRP, in those with stable CRP and in those with decreasing CRP were 36+/-31, 30+/-31 and 24+/-31 ml/year, respectively (P<0.001). These findings were not affected by adjustment for potential confounders (P=0.002). In conclusion, increases in CRP levels over time were associated with a steeper FEV1 decline.