This study shows that obese patients with the metabolic syndrome have a higher proportion of blood monocytes and eosinophils and a lower FEV1/FVC ratio, indicating airway obstruction, than obese patients without the metabolic syndrome. Blood eosinophils (%) in morbidly obese subjects were related to FEV1/FVC, whereas monocytes were not. After adjustment for multiple variables, the relationship between FEV1/FVC ratio and eosinophils remained intact. Although the differences are small, it strengthens our hypothesis that the presence of the metabolic syndrome could play a role in lung function impairment, through the induction of systemic inflammation, in particular mediated by blood eosinophils. Whether this also leads to asthma on the long term still remains to be elucidated.
To our knowledge, this is the first study concerning lung function and the metabolic syndrome in a cohort of only morbidly obese subjects. Secondly, comorbid conditions associated with obesity such as OSAS and GERD were taken into account in the current study. Thirdly, adiposity was not only assessed with BMI, but also with abdominal circumference. Although one would expect that all our subjects have the metabolic syndrome, we found a 65% prevalence of the metabolic syndrome in our group, which corresponds with 60% of the morbidly obese in the NHANES III cohort [16
Various studies have shown that obesity causes a modest reduction in total lung capacity (TLC) and a larger reduction in functional residual capacity (FRC) [17
]. However, we were unable to perform a TLC measurement in all subjects to rule out a restrictive pattern. The FEV1
/FVC ratio is usually well preserved in obese subjects, or even increased. This could explain the high mean FEV1
/FVC ratio in our subjects, and therefore we have used the FEV1
/FVC ratio as a continuous variable and did not use a cutoff value of 70% predicted. Furthermore, in contrast to the FEV1
/FVC, the FEV1
is influenced by BMI; hence, our focus is on the FEV1
The prevalence of self-reported asthma was similar in the two groups. We did not perform methacholine-provocation tests; so, a definitive diagnosis of asthma was often not possible. Since misdiagnosis of asthma is a relevant issue, also in the obese [18
], we felt more comfortable using objective parameters instead of a presumed diagnosis.
Since the metabolic syndrome was associated with the FEV1
/FVC ratio in univariate analysis, we investigated which of the components of the metabolic syndrome contributed to this relationship. Hypertension was the only component that was significantly associated with a reduced FEV1
/FVC ratio after multiple backward regression analysis. Hypertension may have the strongest association with systemic inflammation, as proposed by Irace et al. [19
]. We could also confirm the results of Leone et al. [10
] showing that there is indeed a relationship between abdominal circumference (in cm) and the FEV1
/FVC ratio. Interestingly, we found a correlation between blood eosinophil percentage and abdominal circumference, where we found no correlation between blood eosinophil percentage and BMI, indicating that it is mainly the place of the fat that matters. This suggests that the increased abdominal circumference of subjects with the metabolic syndrome causes the relative eosinophilia.
Obesity is a state of chronic low-grade systemic inflammation. Leukocyte count is considered a marker of systemic inflammation. Several epidemiological studies have already noted a relationship between some components of metabolic syndrome and leukocytes [20
]. Several studies showed an increased eosinophil percentage in obesity [22
] or in the metabolic syndrome [20
]. C-reactive protein (CRP) is a traditional marker of inflammation and is well correlated with BMI [23
]. Even though our study found no difference in CRP between the two groups—we did not use high-sensitivity CRP measurements—monocyte and eosinophils percentage in the blood were higher in those with the metabolic syndrome. Our study suggests that increased numbers or circulating eosinophils could be a specific manifestation of the systemic inflammation associated with the metabolic syndrome.
The question remains why there is relative eosinophilia in the obese. Traditionally, eosinophils are related to allergic diseases, asthma, and parasitic infections. The fat tissue is a source of adipokines, which are considered to play a role in the low-grade systemic inflammation in obesity [24
]. Leptin—mainly produced by adipose tissue—is a proinflammatory agent. Serum leptin is markedly increased in obese humans, correlating to BMI [25
]; it activates eosinophils [26
] and increases their survival [27
]. Serum leptin levels are elevated in adults of normal weight with impaired lung function [28
]. Eotaxin is an eosinophil-specific chemokine that is increased in obesity [29
]. Adipocytes produce more eotaxin when stimulated by leptin [30
]. Plasma adiponectin—an anti-inflammatory hormone—is reversely correlated to BMI [31
]. The leptin/adiponectin ratio is a marker of insulin resistance [32
] and the metabolic syndrome [33
]. Thus, both leptin and possibly eotaxin could contribute to relative eosinophilia in the obese. Unfortunately, we do not have data on leptin, adiponectin, and eotaxin to further support their role in the observed eosinophilia.
Another question is whether this relative eosinophilia in the peripheral blood also has local effects on the bronchial tissue and causes or enhances the bronchial inflammation as seen in asthma. Asthma-like symptoms are common in patients with the metabolic syndrome, and their pulmonary function is impaired [34
]. Although we did find differences in pulmonary function between subjects with and without the metabolic syndrome, there was no difference in prevalence of self-reported asthma between the two groups. This is, however, unreliable in the morbidly obese [18
]. The characteristics of airway inflammation in asthma (exhaled nitric oxide (eNO) and inflammatory cells in induced sputum) have been investigated in obesity and are still subject to debate. de Winter-de Groot et al. showed that a higher BMI was associated with a higher eNO in healthy patients [36
]. Several studies have found no—or even an inverse—relationship between the number of sputum eosinophils and BMI [37
]. We found no difference in eNO between subjects with and without the metabolic syndrome. Induced sputum or bronchial biopsies could have helped to solve this question but were unfeasible in this study. Since there was no difference in the use of inhaled corticosteroids, this could not have influenced our results of eNO.
Our study includes several limitations. We did not measure TLC values; so, we were not able to firmly assess a restrictive lung function pattern. Also, we did not perform methacholine-provocation tests; so, a definite diagnosis of asthma was often not possible. The FEV1
/FVC ratio, however, is easy to measure and widely used in clinical practice. Although the differences in FEV1
/FVC and eosinophils between subjects with and without the metabolic syndrome were small, the difference was significant despite the fact that our study groups consisted of unselected subjects. There probably is a selection bias since we only included patients who were willing to undergo bariatric surgery. It is widely known that most of the subjects who undergo bariatric surgery are females. This explains the female predominance among our subjects. Furthermore, we did not measure adipocytokines or high-sensitivityCRP. We did not include a non-obese control group as comparison for low-grade inflammation, and we cannot fully exclude that smoking might have contributed to a state of low-grade inflammation. However, we corrected for smoking in the multiple regression analysis. Finally, we realize that a cross-sectional association between metabolic syndrome and lung function cannot establish causality. However, Naveed et al. [12
] have recently shown that the metabolic syndrome predicts a steeper FEV1
decline over time, suggesting that the systemic inflammation produced by metabolic syndrome may impact the progression to abnormal lung function in a longitudinally followed cohort.
In summary, our study shows that there is a small, but statistically significant, difference in eosinophils and FEV1/FVC between subjects with and without the metabolic syndrome. After correction for other variables, an association between blood eosinophils and FEV1/FVC remained. Although the differences we have found were relatively small, it might support our hypothesis that the presence of the metabolic syndrome may influence lung function impairment, through the induction of systemic inflammation, in particular, mediated by blood eosinophils. Further research with a firm diagnosis of asthma and assessment of peripheral airway inflammation is necessary. Moreover, in order to establish causality between the metabolic syndrome and lung function impairment, longitudinal studies in morbidly obese patients before and after bariatric surgery are needed.