Period and cohort effects: consequences on lung function in Norway during the 20th century

Abstract

<b>Background:</b> Several factors can influence lung function over time, such as changes in spirometry methodology or in socio-economic conditions affecting physical development. <b>Aims and objectives:</b> We investigated the presence of period effects (changes in lung function with period of examination) and cohort effects (changes in lung function with year of birth) in a large pooled general population sample in Norway during the 20th century. <b>Methods:</b> 36,466 subjects (born 1894-1969) from four cross-sectional surveys conducted between 1965 and 1999 were included, with harmonized data on smoking habits, respiratory symptoms, lung diseases, education and lung function (z-scores calculated according to GLI equations). Changes in FEV1 and FVC z-scores in healthy subjects across surveys were explored to investigate period effects. Linear mixed-effects models of FEV1 and FVC z-scores on year of birth, with survey as random effect, were used to investigate cohort effects. <b>Results:</b> Relatively higher FEV1 and FVC z-scores in healthy subjects were present in the first survey (1965-1970) compared to the more recent ones (1988-1999) suggesting period effects. FEV1 and FVC z-scores increased with year of birth from 1894 to 1969 after adjustment for confounders. Z-score change/year of birth in the total population: FEV1 0.006 (95% CI 0.004-0.007, p&lt;0.001) and 0.012 (95% CI 0.010-0.014, p&lt;0.001), FVC 0.009 (95% CI 0.007-0.011, p&lt;0.001) and 0.011 (95% CI 0.009-0.013, p&lt;0.001), in men and women, respectively. <b>Conclusions:</b> Increased lung function with year of birth was observed in Norwegian subjects born between 1894 and 1969. The impact of period effects on lung function seemed to have decreased from 1965 to 1999.