Published ahead of print on June 15, 2006, doi:10.1164/rccm.200602-243OC
© 2006 American Thoracic Society doi: 10.1164/rccm.200602-243OC
Inverse Association between Pulmonary Function and C-Reactive Protein in Apparently Healthy SubjectsDepartments of Cardiology, Pulmonology, and Internal Medicine; and Center for Preventive Medicine, Rambam Medical Center and Rappaport Faculty of Medicine, Haifa, Israel Correspondence and requests for reprints should be addressed to Doron Aronson, M.D., Department of Cardiology, Rambam Medical Center, POB 9602, Haifa 31096, Israel. E-mail: daronson{at}techunix.technion.ac.il
Rationale: Increased levels of systemic markers of inflammation have been reported in patients with impaired lung function due to obstructive or restrictive lung disease. Objective: We tested the hypothesis that a decline in lung function within the normal range may be associated with a systemic subclinical inflammation. Methods: Pulmonary function tests, cardiorespiratory fitness, components of the metabolic syndrome, and high-sensitivity C-reactive protein (CRP) were determined in 1,131 subjects without known pulmonary disease. Measurements and Main Results: Ninety-six of the study participants (8.5%) had FEV1 of less than 80% of predicted values. There was a strong inverse association between CRP levels and quartiles of FEV1. The median CRP levels in nonsmoking participants were 2.5, 1.8, 1.7, and 1.3 mg/L in the first, second, third, and forth FEV1 quartiles, respectively (p < 0.0001). A similar inverse association was present in smoking subjects (median CRP levels were 3.8, 2.3, 2.0, and 1.9 mg/L in the first, second, third, and fourth FEV1 quartiles, respectively; p < 0.0001). These associations remained highly significant after adjustment for age, sex, components of the metabolic syndrome, and fitness level (p = 0.0005). Conclusions: An inverse linear relationship exists between CRP concentrations and measures of pulmonary function in subjects without pulmonary disease and in never-smokers. These results indicate that systemic inflammation may be linked to early perturbations of pulmonary function.
Key Words: C-reactive protein • forced vital capacity • obesity • physical fitness • systemic inflammation Several population-based studies have shown that impaired lung function as measured by FVC or FEV1 is a powerful predictor of nonfatal ischemic heart disease and of mortality due to cardiovascular disease (1–4). The relationship between reduced FEV1 and cardiovascular mortality also exists in lifetime nonsmokers (2), and even a modest decline in FEV1 is associated with a substantial increase in death from coronary artery disease (2). Furthermore, annual decline of FEV1, independent of baseline FEV1, is related to cardiovascular mortality (5). The majority of patients with reduced FEV1 have asthma, chronic obstructive pulmonary disease (COPD), or fibrotic lung disease (6). In these conditions, cytokines are overexpressed in lung tissue, potentially resulting in systemic low-grade inflammation (6–9). This led to the suggestion that inflammation is an important pathway between lung disease and vascular disease (2, 6). However, increased levels of systemic markers of inflammation have only been demonstrated mainly when patients with overt airflow limitation were compared with subjects without pulmonary impairment (6, 7, 10). The lung is dependent on tightly regulated immunologic and inflammatory processes because it is exposed to a large and varied burden of infectious agents and to a diverse group of noxious gases and particulates during the process of gas exchange. Consequently, the lung has a large number of resident macrophages that generate various inflammatory mediators and cytokines (11–13). Thus, the lung may be a source for low-grade systemic inflammation in the absence of overt pulmonary disease. The relationship between ventilatory function and markers of systemic inflammation, such as C-reactive protein (CRP), is affected by several factors that are associated with subclinical systemic inflammation, including smoking, obesity, and reduced cardiorespiratory fitness (14–16). In the present cross-sectional study, we sought to determine whether differences in pulmonary function within the normal range are associated with CRP concentrations and whether this association is independent of factors that affect ventilatory function and plasma CRP.
A more detailed description of the methods appears in the online supplement.
Subjects
Definitions Fitness was quantified using a maximal exercise test with the Bruce protocol (19) as previously described (18). Cardiorespiratory fitness was categorized based on sex-specific tertiles. High-sensitivity CRP was measured as previously described (18).
Pulmonary Function Tests
Statistical Methods The distribution of CRP levels was highly skewed. Therefore, CRP values were transformed to their natural logarithm (ln CRP) for all other analyses. The adjusted mean values of ln CRP in relation to quartiles of pulmonary function tests were calculated by analysis of covariance (ANCOVA), with results expressed as geometric means. Geometric means of CRP were adjusted for age, sex, smoking status, use of statins and aspirin, history of coronary artery disease, components of the metabolic syndrome (presence of obesity, glucose intolerance, hypertension, low high-density lipoprotein cholesterol, and elevated triglycerides), and fitness level. Geometric means of CRP were calculated using two-way ANCOVA, with ln CRP as the dependent variable, quartiles of FEV1 as one factor, and categories of BMI as the other. Similar models were fitted with quartiles of FEV1 and fitness level. Multivariate logistic regression was used to examine the association between the metabolic syndrome and high-risk CRP, defined as CRP > 3.0 mg/L (22), in relation to quartiles of pulmonary function tests. All multivariate models were adjusted for smoking status. To exclude the possibility of confounding by smoking-induced respiratory disease, all multivariate models were repeated in the subgroup of never-smokers (n = 734). All statistical analyses were performed using the SPSS statistical software version 12.0 (SPSS, Inc., Chicago, IL).
The study population included 1,131 subjects (mean age, 50 ± 9 yr [range, 25–82 yr]; 24% female). Ninety-six of the study participants (8.5%) had FEV1%pred of less than 80%, and four participants had evidence of airflow obstruction (0.4%). The clinical characteristics of the study participants according to FEV1%pred are presented in Table 1. Study participants with lower FEV1%pred values were more likely to be men and to have a history of smoking. The prevalence of some of the characteristics of the metabolic syndrome, including obesity, elevated triglyceride levels, and glucose intolerance, was higher in participants with lower FEV1%pred. Positive criteria for the diagnosis of the metabolic syndrome were higher in subjects with lower FEV1%pred values (Table 1).
Figure 1A shows that the percentage of study participants who were in the upper fitness tertile was lowest among subjects with the lowest FEV1%pred and highest among subjects with the highest FEV1%pred. In an ANCOVA model adjusting for age, sex, smoking status, and BMI, fitness level increased with increasing quartiles of predicted FEV1 (p value for trend, < 0.0001; Figure 1B).
Relationship between CRP and Pulmonary Function Tests CRP levels decreased with increasing quartiles of FEV1 (Figure 2), and the strong inverse relationship of CRP and FEV1 was present in never-smokers (p < 0.0001) and smokers (p < 0.0001). Median CRP levels in the lower FEV1 quartile compared with the upper FEV1 quartile were 90% (2.5 mg/L [interquartile range, 1.3–5.5] vs. 1.3 mg/L [interquartile range, 0.6–2.5]) and 100% (3.8 mg/L [interquartile range, 1.6–7.5] vs. 1.9 mg/L [interquartile range, 1.0–3.4]) higher in never-smokers and smokers, respectively. A similar inverse association between pulmonary function and CRP was present when the analyses were repeated using quartiles of FVC (see Figure E1 of the online supplement) and quartiles of PEF (p < 0.0001) but not with quartiles of the FEV1/FVC ratio (p = 0.59).
Adjusted geometric mean CRP was computed in analyses in which study participants were stratified into 12 groups according to FEV1 quartile and three BMI categories. Two-way ANCOVA main effects indicated a significant inverse association between CRP and FEV1 quartile (p < 0.0001) and a direct association between CRP and BMI (p < 0.0001). For each category of BMI, the adjusted geometric mean CRP was lowest among subjects with the highest FEV1 and highest among subjects in the lowest FEV1 quartile (Figure 3). Similar results were obtained when never-smokers (n = 734; p = 0.004 for the main effect of FEV1 quartile) and subjects without diabetes (n = 1015; p = 0.003 for the main effect of FEV1 quartile) were analyzed separately.
The relationship between FEV1 and BMI was also tested in three separate models for each BMI category. The inverse relationship between CRP levels and FEV1 quartiles was present in all three BMI groups (subjects with normal BMI, p = 0.01; overweight subjects, p < 0.0001; and obese subjects, p = 0.001). Figure 4 shows adjusted geometric mean CRP levels obtained from the two-way ANCOVA model with the main effects of FEV1 quartile (p < 0.0001) and tertiles of cardiorespiratory fitness (p < 0.0001). For each level of cardiorespiratory fitness, the adjusted geometric mean CRP was lowest among subjects with the highest FEV1. The strong inverse association between FEV1 and CRP was also present in never-smokers (p = 0.005 for the main effect of FEV1 quartile).
The results of the multivariate linear regression analysis of the fully adjusted model (age, sex, drug use, history of coronary disease, smoking, metabolic abnormalities, and fitness level) are shown in Table 2. A highly significant inverse association was observed between CRP and quartiles of FEV1. Similar results were obtained when the analyses were repeated using continuous rather than categorical variables for FEV1 (p < 0.0001), cardiorespiratory fitness (p = 0.03), and components of the metabolic syndrome, including BMI (p < 0.0001) and high-density lipoprotein cholesterol (p < 0.0001).
The frequencies of high-risk CRP (> 3 mg/L) were 50, 38, 29, and 26% in the first, second, third, and fourth FEV1 quartile, respectively (p < 0.0001). Multivariate logistic regression models were developed to determine the relationship between quartiles of FEV1 and the presence of high-risk CRP (Figure 5). Compared with subjects in the highest FEV1 quartile, the adjusted odds for high-risk CRP level were 2.5 in subjects in the first FEV1 quartile (95% confidence interval, 1.50–4.0; p < 0.0001). In never-smokers, the adjusted odds for high-risk CRP level were 2.3 in subjects in the first FEV1 quartile (95% confidence interval, 1.4–3.8; p = 0.001).
The results of the present study show that measures of pulmonary function are inversely related to plasma concentrations of CRP. The inverse association between pulmonary function and CRP was independent of factors known to influence CRP levels, such as smoking, components of the metabolic syndrome, and cardiorespiratory fitness. Inverse linear associations between quartiles of FEV1 and CRP were observed within subsets of individuals with different degrees of metabolic perturbations and subjects with different level of cardiorespiratory fitness. The inverse association between pulmonary function and CRP was also present in the subgroup of never-smokers. The lung has a large number of resident macrophages involved in innate protection of the pulmonary tree against inhaled microorganisms and other noxious particulate matter. These include tobacco smoke, infectious agents, occupational dusts and chemicals, and particulate air pollution. Subjects with lower FEV1 may have higher exposure to tobacco smoke or to environmental insults that lead to a subtle decline in lung function and, in parallel, induce a low-grade inflammatory response. It has long been recognized that smoking is associated with airway inflammation (12, 23). In never-smokers, other mechanisms may induce subclinical inflammation. For example, there is a systemic response to inhalation of fine atmospheric particles, initiated by cytokines generated by lung macrophages that phagocytose particles deposited on the lung surface. These cytokines modulate local inflammatory responses and enter the circulation, where they stimulate the liver to release acute-phase proteins (24, 25). Such a mechanism may explain the finding of elevated levels of CRP in people during times of increased particulate air pollution (24, 26, 27). Airspace inflammation leading to the release of proinflammatory cytokines (e.g., interleukin [IL]-6) from alveolar macrophages may also occur due to subclinical respiratory infection and lead to systemic inflammation (28). However, because objective evidence for pulmonary inflammation was not available, our study cannot determine whether the inverse association between CRP and measures of lung function is due to intrapulmonary inflammation. In addition to environmental exposures, individual variations in ventilatory function are likely to be influenced by genetic factors. People with a genetic predisposition for an exaggerated inflammatory response to an environmental or infectious insult (29) may be at higher risk for deterioration of their pulmonary function. Thus, the inverse association between pulmonary function and CRP may also be partly mediated by genetic characteristics conferring increased susceptibility for inflammation-mediated lung injury.
Although the association between impaired lung function and cardiovascular disease is well recognized (1–4), the mechanism is unclear. Recently, evidence for systemic inflammation has been demonstrated in patients with overt airflow limitation. Several studies have shown an inverse association between pulmonary function tests and levels of inflammation sensitive plasma proteins (fibrinogen, The results of the present study extend these findings by demonstrating an inverse linear relationship between CRP and measures of pulmonary function in subjects without pulmonary disease and in never-smokers. These findings are important because they may increase our understanding of the link between respiratory impairment and cardiovascular risk. Inflammation is a common pathway for obstructive and several forms of restrictive lung disease. For example, COPD is characterized by chronic inflammation throughout the airways, parenchyma, and pulmonary vasculature, with inflammatory cells producing a variety of proinflammatory cytokines (13). Although subjects in the present study are unlikely to develop COPD, our findings suggest that low-grade inflammation accompanies the reduction in ventilatory function in the preclinical stage of COPD. Thus, because clinically overt lung disease is preceded by a long period of progressive decline in ventilatory function, elevation of CRP and other proinflammatory mediators may contribute to the subsequent risk of coronary heart disease even before overt lung disease develops (31). Previous studies have shown that obesity is negatively associated with lung function in healthy adults (14, 15). In addition, there is an association between physical activity and slower rate of age-related FEV1 decline (16). Obesity (32) and low cardiorespiratory fitness (33) are strongly related to increased CRP levels. Although the relationship between ventilatory function and CRP was independent of these factors, the coexistence of reduced pulmonary function and obesity or reduced cardiorespiratory fitness greatly increases CRP concentrations. Inflammation is important in the initiation, progression, and clinical outcome of atherosclerosis (31). Prospective studies have shown an association between CRP levels and the long-term risk of cardiovascular disease (34). CRP is believed to be a distal indicator of inflammation that reflects the consequence of distinct inflammatory processes of nonvascular or vascular origin (35). Such inflammatory processes lead to elevated levels of proinflammatory cytokines or other products that have a causal role in atherogenesis or instability of atherosclerotic plaques (35, 36). However, the stimuli that trigger CRP production and the causes of low-grade inflammation in apparently healthy subjects are incompletely understood (35, 36). Raised CRP concentrations are strongly associated with BMI (32), reflecting the fact that adipose tissue is an important source of cytokines (37) and produces approximately 25% of basal circulating IL-6 (37), the principal cytokine that induces the acute-phase response. Other potential low-grade inflammatory processes that lead to increased CRP include fatty infiltration of the liver (38) and inflammation within the arterial wall. The results of the present study suggest that intrapulmonary inflammation is another possible nonvascular trigger for low-grade inflammation.
Study Limitations There was no measurement of local lung inflammation, such as bronchoalveolar lavage or sputum analyses, in our study. Thus, our data do not directly show that lung inflammation underlies the association between reduced pulmonary function tests and CRP. Our study used a cross-sectional design and lacked cardiovascular morbidity and mortality endpoints. Further understanding of the complex interaction between pulmonary function, low-grade inflammation, and their association with cardiovascular disease should be obtained from longitudinal studies that measure pulmonary function, inflammatory markers, and cardiovascular morbidity and mortality endpoints. We measured CRP level at a single time point and therefore cannot exclude the possibility of intraindividual changes in CRP level with repeated measurements (22). However, such misclassification of should lead to an underestimation of the relationship between pulmonary function and CRP.
An inverse linear relationship exists between CRP concentrations and measures of pulmonary function in subjects without pulmonary disease and in never-smokers. These results indicate that systemic inflammation may be linked to early perturbations of pulmonary function.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200602-243OC on June 15, 2006 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form February 19, 2006; accepted in final form June 7, 2006
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
2 statistic for noncontinuous variables. 
1-antitrypsin, haptoglobin, and ceruloplasmin) (
3 mg/ was higher in patients with moderate or severe COPD and in patients with restrictive lung disease (
