This study examined the association of lung function with body composition and systemic and airway inflammation, in overweight and obese males and females with asthma. The findings of this study demonstrate a negative mechanical effect of adipose tissue on lung function within both the android and thoracic regions, with this relationship being particularly evident in females. Interestingly, while lean mass within the android and thoracic regions was associated with improved lung function in males, it had a negative or neutral effect in females. Furthermore, the mechanical effect of obesity appears to have a more significant impact on static lung function, whereas both systemic and airway inflammation have greater impact on dynamic lung function. We also found neutrophilic airway inflammation was positively associated with gynoid and leg lean mass but not fat mass in females.
In females, we found that for every 100 g decrease in android fat mass there was a corresponding increase in ERV of 20 mL. This indicates that central fat mass reduction is an important mechanism for improving static lung function in females. However the inflammatory influences we observed indicate that the relationship between respiratory function and obesity in females is more complex. The inflammatory effect of obesity on lung function is occurring via innate immune pathways, as we observed a positive association between sputum neutrophils, and lean mass, as well as lung function. This is in agreement with other authors, who show that it is not airway eosinophilia driving the association between obesity and asthma [24–27]. This also supports our previous study, where we observed a positive association between sputum neutrophils and BMI in asthmatic females . While other authors have reported no significant association between obesity and neutrophilic airway inflammation, they do illustrate a clear trend of increasing %neutrophils as BMI category increases [25–27]. The lack of statistical significance in these studies may possibly be due to sex effects not being examined [25, 27], the small sample size enrolled [25, 26] and the high variability in the measurement of sputum neutrophils. Interestingly, the positive relationship we observed in this study was between neutrophilic airway inflammation and lean mass, but not fat mass. Sood et al  previously observed a positive association between asthma incidence and lean mass in 1422 females who underwent body composition analysis using DXA. The authors found that asthma incidence was more strongly associated with lean mass than fat mass. The authors hypothesised that this may be due to intramyocellular lipid; that is, fat droplets within skeletal muscle that are highly metabolically active . Intramyocellular lipid levels are higher in females and are highly correlated with circulating leptin and adiponectin levels [29, 30]. Therefore, although lean tissue itself is not proinflammatory, the presence of elevated intramyocellular lipid within lean tissue may exert proinflammatory effects. Interestingly, the association we observed was between airway neutrophils and lower body lean tissue. This observation suggests that the relationship may be mediated by leptin. Leptin is secreted 2-3 fold higher from subcutaneous compared with visceral adipose tissue  and increases activation of neutrophils via TNF-α . This may explain the negative association we observed between respiratory function and lean tissue, the positive association between sputum neutrophils and lower body lean tissue and the positive trend between leptin and lower body lean tissue in females. DXA scanning cannot distinguish intramyocellular lipid from lean tissue, hence additional analysis would be required to test this hypothesis. However, examination of intramyocellular lipids may provide further insight into the obese-asthma association.
In contrast, we observed a positive association between central (android and thoracic) lean mass and static lung function in males. A number of DXA studies have cited a similar relationship in healthy subjects [6, 33], with loss of lean mass an important predictor of lung function decline in the elderly . We found that for every 100 g increase in android lean mass, there was a corresponding increase in ERV of 29 mL. This finding suggests an important role for central lean mass in preserving static lung function in males. Surprisingly we did not observe significant associations between fat mass and lung function in males, nor was fat mass a significant predictor of lung function in either of the regression models. This is in contrast to females, for whom fat mass was a considerable predictor of reduced lung function. This supports previous research by Sutherland et al  who also observed a greater effect of adipose tissue on the lung function of females compared with males. In contrast to the female subjects, there was no association between neutrophilic airway inflammation and body composition or lung function in males. We did, however, observe a weak positive association between sputum eosinophils and fat mass in males, while sputum eosinophils were a negative predictor of FEV1/FVC. This result has not been previously described and further examination of the role of eosinophilic airway inflammation in overweight and obese males may be warranted.
Simple anthropometric measurements such as body weight or waist circumference alone do not accurately describe body composition. We did not find body weight or waist circumference to be significantly related to dynamic or static lung function in males. We propose that this may be due to the opposing effects that lean mass and fat mass have on lung function and the fact that body weight and waist measurement cannot discriminate between these two tissue types. In females, however, both body weight and waist circumference were negatively associated with static lung function. These findings suggest that if lung function prediction equations were to be adjusted for obesity, waist circumference would be a suitable measure in females but not males. This poses a problem in males, as more accurate measures of body composition (i.e. those that quantify lean mass) are not practical within the clinical setting.
A limitation of this trial is the limited sample size and its cross-sectional nature which does not allow for the establishment of causality. However, this study has generated some novel findings that extend existing knowledge in the field and suggest an interesting direction for future research. Another limitation is that DXA is not able to differentiate between visceral and subcutaneous fat, nor can it distinguish intramyocellular lipid from lean tissue. More hazardous and expensive investigational techniques such as computerised tomography (CT) and magnetic resonance imaging (MRI) would be required to evaluate the importance of these measures of fat distribution, however would be advantageous for future trials. Finally, all subjects were within a limited BMI range (28-40 kg/m2) and as such these results cannot necessarily be generalised to all individuals.