Prediction curves for growth and decline of lung function: effects of smoking and respiratory symptoms

Aim. To evaluate differences in growth and decline of lung function over time in four categories of subjects: asymptomatic never smokers (AN) asymptomatic current smokers (AS), symptomatic never smokers (SN), and symptomatic current smokers (SS). Methods. We analyzed 2666 measurements of FEVI provided by 1830 male subjects older than 14 yr observed in either 1980/82 (n = 620) or 1988/91 (n = 352), or both (n = 847) as part of the Po river delta study in Northern ltaly. Those subjects who reported any respiratory symptom or disease were defined as symptomatic. Linear regressions were estimated in the four categories separately. To obtain a continuous curve for the growth and decline of FEVI, age entered by means of one break point natural cubic splines. Body mass index (as second order polynomial) and height were also included as covariates. Random effects models were applied to account for potential dependency of repeated. observations. Results. AN showed the greatest (4127 ml) and longest lasting (through 29 yr) growth, and the slowest decline (10 ml/yr average loss from 30 to 60 yr) followed by SN (4096 ml; through 29 yr; 11 ml/yr), AS (4041 ml; through 27 yr13 ml/yr), and SS (4022 ml; through 25 yr; 15 ml/yr). Based on regression models, for average height and weight subjects, a smoker, even though asymptomatic, has at 50 the same FEV1 predicted value that a non smoker has at 56. We obtained 95% confidence bands for the prediction curves of AN and AS that were separated for ages over 40, indicating that differences between the predicted mean values were significant. Conclusions. Regardless of smoking respiratory symptoms are associated to slightly reduced predicted FEVI values. However, smoking is far more detrimental to lung function than reporting respiratory symptoms.