Systemic inflammation in chronic obstructive pulmonary disease: a population-based study

The present study is part of the EPI-SCAN study, a multicentre, cross-sectional, population-based, observational study conducted at 11 sites throughout Spain [15, 16]. The final population recruited was formed by 4,274 non-institutionalized participants from 40-80 years old. The study was approved by the corresponding ethics committees and all participants gave written informed consent.

In accordance with current GOLD guidelines, COPD was defined by a postbronchodilator FEV₁/FVC ratio < 0.70 [9]. COPD severity was determined by the GOLD criteria and the BODE index [9, 17]. Subjects with a postbronchodilator FEV₁/FVC ratio ≥ 0.70 were considered not to have COPD.

All participants classified as COPD were selected for the systemic biomarker analysis. To avoid excessive testing of the non-COPD study population, an equal number of non-COPD subjects were consecutively selected in each centre. Exclusion criteria for this analysis included a previous diagnosis of acute myocardial infarction, angina, congestive heart failure, cancer, hepatic cirrhosis, chronic renal failure, rheumatoid arthritis or any other systemic inflammatory disease. In addition, specific exclusion criteria from the non-COPD cohort were any respiratory symptoms as per the European Coal and Steel Community (ECSC) questionnaire, any associated concomitant disease, and regularly prescribed medications. The reference group obtained after applying these selection criteria was considered to be of healthy subjects.

Fieldwork and all methods have been described previously [15, 16]. Self-reported exposure was identified initially through a questiondeveloped for the European Community Respiratory Health Survey: "Have you ever worked in a job which exposed you to vapors, gas, dust, or fumes?" The question was followedby a list of 23 individual exposures considered a priori risk factors for COPD, subsequently grouped into three categories: biological dusts, mineral dusts and gases or fumes. Baseline dyspnea was assessed by the Modified Medical Research Council (MMRC) scale, and subjects completed the ECSC questionnaire of respiratory symptoms, the London Chest Activity of Daily Living (LCADL) scale, the EQ-5D questionnaire and the St. George's Respiratory Questionnaire.

Blood samples were collected using standardized procedures and stored at -80°C. Samples were shipped to a single laboratory (Hospital Clinic, Barcelona) for centralized analysis approximately every 2 months. TNF-α, IL-6 and IL-8 were determined in duplicate with a high sensitivity enzyme-linked immunosorbent assay (Biosource, Nivelles, Belgium) with lower detection limits of 3 pg/ml for total TNF-α, 2 pg/ml for IL-6 and 0.7 pg/ml for IL-8. The intra-assay coefficients of variation were 3.7% for TNF-α, 2.2% for IL-6 and 2.3% for IL-8. C-reactive protein (CRP) was assessed by latex-enhanced immunonephelometry (Siemens, Dublin, Ireland) with a lower detection limit of 0.4 mg/l and an intra-assay coefficient of variation of 1.2%. Alpha-1 antitrypsin (A1AT) was measured by a particle-enhanced immunonephelometry (Siemens, Malburg, Germany), with detection limits ranged from 0.0095 to 0.3040 g/l and an intra- and interassay variability or 3.9% and 2.0%, respectively.

Albumin levels were estimated by the bromocresol green method (Siemens, Dublin, Ireland), with a detection limits from 10 to 60 g/l and an intra-assay coefficient of variation of 1.5%. Fibrinogen was assessed using a coagulation analyzer (Roche, Mannheim, Germany) according to the Clauss method and calculated from ethylenediamine tetra-acetic acid to citrate plasma values. The detection range was 0.5 to 12.0 g/L and the intra-assay variability 2.8%. Nitrites and nitrates (NOx) were determined by a chemiluminescence detector in an NO analyser (Sievers Instruments, Inc., Boulder, CO, USA). The lower detection limit was 1 pmol and the intra-assay coefficient of variation was 10%.

Baseline and post-bronchodilator spirometries were performed at each site using the same equipment according to current recommendations [18]. The predicted values used were those of the Spanish reference population [19]. A 6-min walk test was performed twice, with an interval between testing of 30 minutes, according to the ATS guidelines [20].

Variables are presented as a percentage, mean ± SD or median (interquartile range) as required depending on their distribution. Statistical analysis was performed with SPSS 14.0 for Windows (SPSS, Inc., Chicago, IL) and with SAS statistical package (version 9.1, Cary, NC). A two-sided p value < 0.05 was considered statistically significant.

Pearson's chi-square test, Mann-Whitney U test or Student's t test were used for two-group comparisons, depending on data distribution. The effect of the possible confounding factors was assessed using generalised linear model analysis [21]. In this analysis, a logarithmic transformation was used in those variables to reduce their skewness. We constructed a multivariate model, including group and gender as fixed factors and age, BMI and smoking history as a dichotomous variable (≥ 10 pack-years, yes/no) as covariates. The link function used was the identity. For each systemic biomarker, we chose the normal distribution because it was more fitting than inverse Gaussian or gamma distribution, according to the plausibility criteria, Pearson's chi-square and analysis of deviance. Comparisons by differing severity within the COPD group were performed using ANOVA analysis, with post-hoc analysis by the Bonferroni test. In the COPD group, the correlations between the serum levels of systemic biomarkers and the clinical and functional parameters were estimated using Pearson's linear bivariate correlation coefficient.

Data are presented according to current recommendations for observational studies in epidemiology (STROBE).

To adequately evaluate the effect of COPD on systemic biomarkers, several risk factors associated with COPD should be considered. COPD is an age-related disorder and the normal process of aging appears to be associated with a similar low-grade systemic inflammatory process [16, 22]. The importance of gender is given by the fact that females have a more vigorous inflammatory reaction and generate more oxidative stress in the airways than males [23]. Although an abnormal systemic inflammatory reaction is detected in most smokers, it has been demonstrated that some systemic biomarkers remain persistently high after smoking cessation [24], suggesting the contribution of other factors. For this reason, some authors propose to evaluate the impact of tobacco on systemic biomarkers depending on whether a dose threshold (10 pack-years) has been reached [25]. Obesity is associated with low-grade systemic inflammation and it has been suggested that the distribution of body compartments might originate a different behaviour of some inflammatory markers [26, 27]. In concordance with previous reports [28], a direct correlation was found between BMI and CRP (r = 0.242, p = 0.0001) in the COPD participants of our study.

After adjusting for possible confounding factors, we report that COPD patients showed higher levels of TNF-α, IL-6, IL-8, CRP and nitrites/nitrates than control subjects. The origin of systemic inflammation in COPD is not completely clear. The hypothesis that systemic inflammation is originated by spill over from the pulmonary compartment has not yet been proven [3]. It has been suggested that some common genetic or constitutional factors may predispose individuals with COPD towards both systemic and pulmonary inflammation [29]. Lung hyperinflation, tissue hypoxia and skeletal muscle and bone marrow alterations have also been implicated in the induction of systemic inflammation [3].

Although an increased production of NO in COPD patients could constitute a host defense mechanism, a high level of NO can also cause injury and thus contribute to the respiratory and systemic features of the disease. In an inflammatory environment, exaggerated production of NO in the presence of oxidative stress may produce the formation of strong oxidizing reactive nitrogen species, such as peroxynitrite, leading to nitration, which provokes inhibition of mitochondrial respiration, protein dysfunction and cell damage [30]. The activation of various heme peroxidases by hydrogen peroxide can promote oxidation of nitrites to intermediates that are capable of nitrating aromatic substratesand proteins [30].

Although the COPD severity classification according to the BODE index shows a great capacity for discriminating among the systemic biomarker levels, as expected from its multicomponent character, the GOLD classification also shows differences in biomarker levels. However, the selection of a small number of severe patients in our population sample may reduce the strength of a possible association between biomarkers and GOLD stage. In some previous studies, the relation between plasma CRP levels and the severity of the disease has already been suggested [5, 31]. De Torres and colleagues reported the usefulness of CRP in predicting clinical and functional outcomes in stable COPD, with similar correlation coefficients to those of our study [27].

Nevertheless, one of the major implications of systemic inflammation in COPD is its contribution to a proatherosclerotic state. The relationship between COPD, systemic inflammation, and cardiovascular diseases may be especially relevant as over half of patients with COPD die from cardiovascular causes [32]. A Copenhagen City Heart Study cohort study showed that the incidence of COPD hospitalization and COPD death was higher in individuals with baseline CRP above 3 mg/L, with an absolute 10-yr risk for death of 57% [33]. In fact, it has been suggested that CRP can be considered as the sentinel biomarker [32, 33]. Interesting, in our COPD patients, serum CRP levels were related to concentrations of IL-6 (r = 0.333, p < 0.001), IL-8 (r = 0.125, p = 0.039), fibrinogen (r = 0.356, p < 0.001) and A1AT (r = 0.194, p < 0.001).

In our COPD patients, CRP and IL-6 were inversely related to postbronchodilator FEV₁ (% predicted). However, the contribution of systemic inflammation to lung function decline is less clear. While crossectional studies show that systemic inflammatory markers are inversely related to lung function [6, 13, 25], a prospective evaluation of lung function decline in a randomly selected population did not identify this negative effect over a 9-year period [34].

Finally, we found that exercise tolerance, as assessed by the distance walked in the 6-minute test was inversely related to serum CRP, IL-6 and IL-8 levels. IL-6 is produced by contracting muscles and released into the blood, acting as an energy sensor. When contracting muscles are low in glycogen, IL-6 gene transcription is increased and IL-6 is released to increase glucose uptake and induce lipolysis [35]. When muscles are exposed to oxidative stress, both IL-6 mRNA and IL-6 protein expression are enhanced [35]. It is known that COPD patients with high plasma levels of CRP had more impaired energy metabolism, increased disability and more distress due to respiratory symptoms than patients with normal CRP levels [35]. Moreover, the relation between serum CRP levels and exercise tolerance seems to be independent of other factors such as age, sex, and smoking history [36]. Whether skeletal muscle dysfunction is a direct consequence of the systemic effects of the COPD or an independent process that contributes to the systemic inflammatory load of the disease, our results indicate that systemic biomarker levels could indirectly reflect the exercise capacity of these patients.

The strengths of this study include its population-based design, the use of post-bronchodilator spirometry as diagnostic criteria and the detailed characterization of the participants which allowed us to investigate factors associated with circulating biomarker levels. In addition to the post-bronchodilator spirometric criteria confirming the existence of an irreversible obstruction, in our never-smoker COPD patients there was evidence of the existence of a particular exposure history that could support the diagnosis.

Although it has been reported that the fixed FEV₁/FVC ratio method results in a greater proportion of COPD diagnoses than other alternative methods, especially in the elderly, it still continues to be the criterion established in the GOLD guidelines and is certainly the most widely used in clinical practice. The current debate about LLN as an alternative to the fixed ratio is based more on the opinions of experts [37, 38] than on the existence of clear evidence. In this situation, the information provided by the Cardiovascular Health Study is especially relevant, demonstrating that a cohort of elderly subjects classified as ''normal'' using the LLN but abnormal using the fixed ratio were more likely to die and to have a COPD-related hospitalization during an 11-year follow-up [39]. Thus, a fixed FEV₁/FVC ratio < 0.70 may identify at-risk patients, even among older adults.

There are several potential limitations of our study worth discussing. Firstly, our COPD patient sample does not turn out to be necessarily representative of the COPD regularly seen in clinical practice. Due to the design of the present study, and to avoid the confusion by comorbidities, patients with several associated illnesses were excluded from the analysis. This strategy probably reduces differences between the two study groups, but it assures a stricter evaluation of these. Secondly, there were significant differences between COPD patients and reference subjects for anthropometric characteristics and smoking status. These are consequences of the population-based extraction of our study subjects and are partly a reflection of the fact that some differences are likely manifestations of the disease. Although the statistical model aims to adjust for these possible confounding factors, the existence of some uncontrolled effects cannot be excluded. And finally, this is a cross-sectional study, and inference of causality is not possible. Our results could also be affected by other drawbacks. Of the COPD patients in our study, 23% were using inhaled corticosteroids. This could contribute to the underestimation of the difference in systemic biomarkers between COPD patients and control subjects. Nevertheless, the effect of inhaled corticosteroids on inflammatory biomarkers is still controversial. In patients with moderate to severe COPD, it has been reported that one month of fluticasone did not reduce serum CRP or IL-6 levels [28]. Other chronic conditions, such as chronic heart failure or diabetes, also appear to be associated with a similar low-grade systemic inflammatory process [22]. Several studies have described that these disorders are more frequent in COPD patients and, therefore, might also contribute to their proinflammatory state. Nevertheless, and in order to analyze only the effect of COPD as an independent factor, we have carefully excluded these conditions from the present study.