Background: As little is known about the long term relationship between respiratory symptoms and mortality from non-malignant respiratory diseases, a study was undertaken to investigate the predictive value of respiratory symptoms and symptom load for mortality from obstructive lung disease (OLD) and pneumonia in the long term in a Norwegian population. Methods: In 1972, 19 998 persons aged 15-70 years living in Oslo were randomly selected for a respiratory survey. The response rate was 89%. All were followed for 30 years. The association between cough, asthma-like symptoms, two levels of dyspnoea on exercise, a symptom score, and mortality from OLD and pneumonia were investigated separately for men and women by multivariable analyses, with adjustment for age, occupational exposure to air pollution, and smoking habits. Results: OLD accounted for 43% and pneumonia for 50% of all deaths from respiratory causes. In men the hazard ratio for mortality from OLD varied from 4.0 (95% confidence interval (CI) 2.4 to 6.5) for cough to 9.6 (95% CI 5.1 to 18.3) for severe dyspnoea, and in women from 5.1 (95% CI 2.3 to 11.3) for moderate dyspnoea to 13.0 (95% CI 6.0 to 28.3) for severe dyspnoea. The symptom score was strongly predictive of death from OLD in a dose-response manner. Conclusions: There is a significant, positive, strong association between respiratory symptoms and 30 year mortality from OLD. The association between respiratory symptoms and mortality from pneumonia is weaker and not significant.