Level of FEV₁ as a predictor of all-cause and cardiovascular mortality: an effect beyond smoking and physical fitness?

In 1970, Higgins and Keller 1 showed that low spirometric values predict mortality independently of smoking in the general population of Tecumseh, MI, USA. Furthermore, in the general population of Renfrew and Paisley, Scotland, a forced expiratory volume in one second (FEV₁) level in the lowest quintile explains a substantial fraction of all-cause mortality occurring in the subsequent 15 yrs (31% in males and 26% in females) 2. This observation was mainly mediated through a higher risk of dying from cardiovascular conditions, as it has also been observed in a number of other cohorts. In line with this, a significant association has also been reported between baseline levels of lung function and the incidence of coronary heart disease 3 and stroke 4; and, furthermore, annual decline of FEV₁, independent of baseline FEV₁, has also been related to cardiovascular mortality 5.

Recent evidence suggests that individuals in the community with moderate airflow limitation may have co-existing systemic inflammation 6, and this may therefore be the mechanism underlying the observed association between lung function and mortality. However, after adjusting for several peripheral inflammatory markers, the follow-up of 14,480 patients in the Atherosclerosis Risk in Communities study (ARIC) revealed a steady increase in the incidence of cardiac events (related to coronary artery disease) with decreasing quartiles of FEV₁ 3.

Residual confounding by smoking, physical fitness, body mass index (BMI) or social class offer possible alternative explanations for the observed relationship between lung function and mortality. Although most of the reported findings from population studies have been adjusted for smoking habits and BMI, only very few have incorporated changes in these variables beyond baseline. Furthermore, an association has also recently been reported between physical inactivity and lower FEV₁ 7, but this aspect has only been considered in very few studies. Moreover, even fewer studies have considered the role of socio-economic factors for the level of lung function 8, and these factors therefore remain as an alternative explanation for the observed association.

Stavem et al. 9 in this issue of the European Respiratory Journal report findings from a 26-yr follow-up study of 1,223 healthy males, aged 40–59 yrs at enrolment, from five companies in Oslo, Norway. In the total sample, the relative risk of all-cause mortality increased by 10% for each 10% decrease in FEV₁ % predicted, after adjusting for smoking habits, physical fitness, blood pressure, BMI and total cholesterol at baseline, as well as changes in smoking habits during the first 8 yrs of the observation period. The corresponding relative risk increased by 7% for cardiovascular deaths. However, surprisingly, after stratifying by smoking category, these associations were no longer significant among the 398 never-smokers (relative risk = 1.01 for all and 0.93 for cardiovascular causes).

This study contributed new knowledge primarily because it included a reliable objective measure of physical fitness at baseline. Few of the previously published studies on mortality have incorporated physical fitness, which is known to be a predictor of mortality 10, and to our knowledge none of these studies have included an objective test. Furthermore, the observed association between FEV₁ and mortality was not confounded by physical fitness. This study, therefore, is likely to rule out residual confounding by physical inactivity as the driving mechanism behind the association between level of lung function and mortality. Another question of interest not addressed by Stavem et al. 9 is whether exercise may modify the role of impaired FEV₁ on mortality.

The lack of association among never-smokers may suggest residual confounding by smoking. However, the likelihood of residual confounding by smoking in the association between level of FEV₁ and mortality seems unlikely given the large number of studies finding a positive association with cardiovascular mortality in never-smokers 2, 11–13. The observed lack of association between FEV₁ and mortality in never-smokers reported from the study by Stavem et al. 9 is, therefore, most likely due to the relatively limited sample size. Nevertheless, larger studies or pooling data from a number of studies may be necessary to obtain a definitive answer to this question.

The lack of repeated measures of BMI and information on socio-economic determinants are the major limitations of this study, the former not least because changes in BMI may be considerable after changes in smoking habits 14. However, general health and quality of life are related with socio-economic conditions, in part because of poorer access to education and health services 8, higher exposures to air pollutants or occupational exposures 15, and the consequences of early life events, such as maternal diet or pre- and post-natal exposures, including diet 16 and environmental tobacco smoke 17, for the development of lung function.

In conclusion, the study by Stavem et al. 9 provides new insight into the association between level of lung function and all-cause and cardiovascular mortality by documenting that this relationship cannot be explained solely by tobacco exposure and poor physical fitness.