Commentary: Lung function and risk of fatal and non-fatal stroke—The Copenhagen City Heart Study

The findings of Truelsen and colleagues1 confirm earlier reports of the inverse association of lung function with risk of both fatal and non-fatal stroke.2,3 Is a link between markers of abnormal vascular biology and airway function surprising? The relationship is attenuated when adjustment is made for other potential confounders, particularly age, sex and smoking. This shows that much of the relationship is explained by these factors. While there remains a possibility that residual confounding, due to imprecise measurement of lifetime tobacco exposure, is a sufficient explanation for the findings, among those who have never smoked, a relationship between low forced expiratory volume in one second (FEV1) and cardiovascular disease is still seen.4,5 Sick people tend to have low FEV1, in part because of the muscle weakness associated with the cachexia of illness and old age, and in part because of inability to make a maximal expiratory effort. Analysis of the FEV1-stroke relationship comparing FEV1-stroke associations for strokes occurring within 5 years of measurement of lung function and for those occurring subsequently might detect a potential ill-health bias. Traditionally FEV1 is adjusted for height (as taller people have bigger lungs), but this adjustment hides the independent association between adult height and stroke.6 It would be preferable to examine FEV1-stroke relationships making progressive adjustment for age, smoking and height rather than using an adjusted FEV1 as the exposure variable, as has been done recently.7 As height is composed of trunk and leg lengths, it is likely that height may have an effect on risk of stroke by two mechanisms—as an easier and more accurate measure of lung volume (i.e. a proxy for FEV1) and as an indirect measure of leg length. Leg length is a marker of childhood nutritional status and growth and has been shown to correlate closely with risk of cardiovascular disease in adult life.8 The maternal-fetal environment determines susceptibility to reduced FEV1, asthma, chronic obstructive pulmonary disease, diabetes mellitus, obesity, hypertension, and atherosclerosis in later life.9–14 Hence, the link between FEV1 and vascular pathology may reflect common early life determinants. It is now important to establish the underlying mechanisms that are responsible and discover how they operate throughout the life course. Truelsen and colleagues propose that increased susceptibility to infections found in individuals with reduced lung function results in raised blood viscosity which increases the risk of vascular occlusion and consequently, stroke. Pelkonen and colleagues suggest that the mechanism is hypoxaemia resulting from airway obstruction due to repeated inflammation.4 Both these hypotheses seems inadequate to explain the graded relationship seen in this and other studies. Haematocrit, blood viscosity and hypoxaemia only increase at very low levels of FEV1 and it is likely that infections would be no more or less likely among people in the mid-range and bottom of the FEV1 distribution. Furthermore, as both reduced FEV1 and vascular disease develop over many decades, either may precede the other, making attribution of causality difficult. Therefore hypotheses, also suggested by other workers,15 that reduced lung function is a stimulant of pathogenetic mechanisms in different organ systems seem improbable. It is known that FEV1 is associated with carotid artery intimal-medial thickness16 and it is possible that FEV1 is acting as a marker for smooth muscle hyperplasia and an increased likelihood of developing carotid plaques. Abnormalities of shared underlying mechanisms controlling the balanced interactions between cells and the extracellular matrix in the blood vessels and airways should therefore be considered. Previously, within the same population sample, Lange and colleagues have shown an inverse association between plasma glucose and FEV1.17 Furthermore, in the Normative Aging Study, a negative correlation was found between FEV1 and fasting insulin and the fasting insulin resistance index.18 Raised glucose, raised insulin and insulin resistance are risk factors for increased cardiovascular morbidity and mortality and are regulated by the growth hormone/insulin-like growth factor-1 (GH/IGF-1) axis.19,20 Hence, in the development of vascular and airways disease, altered activity of the GH/IGF-1 axis at different stages of the life course may be relevant. Advanced glycation end-products (AGE) could also be a plausible part of this conundrum. Advanced glycation end-products accumulate on long-lived extracellular matrix proteins and have been implicated in diabetic and non-diabetic atherosclerotic complications21 and are considered contributory to reduced FEV1 found in diabetics and in relation to ageing.22 In arterial sections taken at the time of coronary artery by-pass in euglycaemic normolipidaemic individuals, using immunoassays, levels of serum AGE-modified lipoproteins were found to correlate with arterial tissue AGE and atherosclerosis.23 Advanced glycation end-products can mediate upregulation of different growth factors such as the IGF and transforming growth factor-beta (TGF-beta) through a receptor-mediated mechanism.24 Insulin-like growth factors and TGF-beta have been identified in asthmatic airways and can stimulate pro-collagen production, fibroblast proliferation, and possibly subepithelial fibrosis that will eventually result in lung function impairment.25,26 Insulin, glucose, IGF and their binding proteins are growth promoters for arterial cells and act together in stimulating smooth muscle migration and proliferation implicated in the development of experimental atherosclerotic lesions.27,28 Future work should explore the role of glucose, insulin, growth factors such as GH/IGF and TGF-beta and AGE as components in causal pathways responsible for health-disease links between blood vessels and airways. With this population sample and others, examination of the rate of decline in FEV1 and stroke would be particularly informative. If relationships between stroke are stronger with rate of decline than with baseline levels, this would suggest that the mechanisms responsible operate in adult life. It would also be useful to examine familial aggregation of markers of atheroma, stroke, myocardial infarction and reduced FEV1 since this may indicate an additional genetic influence.