Respiratory symptoms as predictors of 27 year mortality in a representative sample of British adults.

Abstract

To examine associations between reported respiratory symptoms (as elicited by questionnaire) and subsequent mortality. Prospective cohort study. 92 General practices in Great Britain. A nationally representative sample of 1532 British men and women aged between 40 and 64. Mortality from all causes, cardiovascular disease, lung cancer, and chronic bronchitis. Subjects were interviewed in 1958 regarding various respiratory symptoms (including cough, phlegm, breathlessness, and wheeze) by using a questionnaire which formed the basis of the Medical Research Council's questionnaire on respiratory symptoms. By the end of 1985, 889 deaths had been reported, including 51 in men due to chronic bronchitis. After adjustment for differences in age and smoking habits death rates from chronic bronchitis in men who reported symptoms were greater than those in men who did not for each of the symptoms examined. The adjusted mortality ratios were 3.4 (95% confidence interval 1.8 to 6.5) for morning cough, 3.7 (2.0 to 6.9) for morning phlegm, 6.4 (3.0 to 13.8) for breathlessness when walking on the level, and 10.5 (4.4 to 24.6) for wheeze most days or nights. Mortality ratios were also significantly raised for four episodic symptoms not usually included in more recent respiratory symptom questionnaires--namely, occasional wheeze (mortality ratio 6.0; 95% confidence interval, 2.4 to 15.1), weather affects chest (5.7; 3.1 to 10.3), breathing different in summer (4.9; 2.8 to 8.6), and cold usually goes to chest (3.7; 2.0 to 6.8). The excess mortality associated with these symptoms remained significant after further adjustment for breathlessness or phlegm. Ratios for all cause mortality in men and women were also significantly raised for most respiratory symptoms, death rates being some 20-50% higher in people reporting symptoms after adjustment for age, sex, and smoking. Breathlessness was the only symptom significantly associated with excess mortality from cardiovascular disease (mortality ratio 1.4 (95% confidence interval 1.0 to 1.9) for breathlessness when walking on the level). Ratios were generally around unity and not significant for mortality due to lung cancer. The results suggest that episodic symptoms, which often do not appear in standard respiratory questionnaires, predict subsequent mortality from chronic obstructive airways disease. This supports the hypothesis that reversible airflow obstruction may be a precursor of progressive and irreversible decline in ventilatory function.