Association of chronic inflammation with cardiovascular risk in chronic obstructive pulmonary disease—A cross‐sectional study

Abstract Background and Aims COPD is progressive lung disease with known higher cardiovascular (CV) risk, mainly attributed to smoking of cigarettes as the main etiological factor of disease. The aim of this study was to compare CV risk in patients with COPD to control groups of smokers and non‐COPD and to investigate the relation of lung function variables, COPD severity, and smoking with Systemic Coronary Risk Estimation (SCORE) risk calculation, arterial stiffness (AS) values, and biological systemic inflammatory markers. Methods A total of 208 subjects were included in this study: 61 subjects diagnosed with COPD, 83 smokers without COPD, and 64 nonsmokers without COPD. Medical history and clinical data were recorded, including assessment of pulmonary function and AS, calculation of ankle‐brachial index, blood analysis, and CV risk assessment by SCORE risk calculation. Results Subjects with COPD had significantly higher values of SCORE calculation of risk, central aortic pressure, AS, and markers of systemic inflammation compared to control groups of smokers and nonsmokers without COPD (p < 0.001). Furthermore, statistically significant increase in hs‐CRP concentration was found between the COPD group and the control group of non‐COPD smokers (p < 0.001), and a statistically significantly higher SCORE calculation was found in the COPD group compared to control groups of smokers and nonsmokers without COPD (p < 0.001). Conclusion The results of the research support further identification and research of biological markers and simple specific tests such as arteriography that will enable progress in personalized treatment of patients with COPD and better primary and secondary prevention of comorbidities with the aim of improved treatment outcome.


| INTRODUCTION
Although COPD is considered to be a lung disease, systemic manifestations of COPD associated with increased morbidity and mortality are increasingly being recognized. 1 Comorbidities were previously thought to occur only in the later stages of COPD, however, research in recent years indicates a high proportion of patients with comorbidities, even in patients with mild bronchoobstruction. 2 Accurately determining the type of comorbidities that occur with COPD regulates treatment choices, occurrence, frequency, and types of patients' future health complications and affect survival rate and time. This is important because today it is known that optimal and early treatment as well as primary and secondary prevention of comorbidities have a clear positive effect on the clinical outcome of COPD. Several prospective studies have described an association between impaired lung function and cardiovascular (CV) morbidity and mortality. [3][4][5][6] Furthermore, epidemiological data suggest that COPD patients are at higher risk of CV disease compared to control groups by age and sex without COPD. 7 Also, the systemic inflammatory response is thought to play a possible role in explaining this association. 8 Besides the already mentioned, patients with mild to moderate COPD activity have been found to die more often from lung cancer and CV disease, such as coronary heart disease, than from the respiratory effects of COPD itself. [9][10][11] Although it has long been established that inflammation of the small airway mucosa is an initial event in the pathogenesis of COPD and that the severity of inflammation is related to the degree of obstruction, recent studies have indicated that systemic inflammation in COPD may accelerate atherosclerosis. 12,13 Since atherosclerosis is the result of multiple risk factors, all currently valid guidelines for CV disease prevention in clinical practice recommend an assessment of the overall CV risk. Thus, the Systemic Coronary Risk Estimation (SCORE) chart of fatal CV disease is recommended in the European guidelines for the prevention of CV disease in clinical practice from 2016. 14,15 Since classical risk factors only indirectly suggest atherosclerotic processes that induce CV changes, the interest of the profession for a simple and noninvasive way of detecting increased CV risk in a subclinical stage at the individual level has increased. SCORE risk assessment is based on classical risk factors (age, smoking, cholesterol, systolic blood pressure [SBP]) and is effective at the population level, but is less accurate for determining specific, individual risk exposure of an individual. 16 Of the noninvasive methods of CV risk assessment, the method of determining the elasticity of the arterial wall, that is, the assessment of arterial stiffness (AS), has become the most widespread in the last decade, with methods such as pulse wave analysis (PWA), pulse wave velocity (PWV), and aortic augmentation index (AIx) due to their reproducibility and ease of performance. 17,18 These tests have been shown to be associated with coronary microvascular endothelial function 19 and aortic PWV as an independent predictor of the CV disease. 20 Previous research suggests that aortic stiffness has been shown to be an independent predictor of overall and CV morbidity and mortality in hypertensives and healthy subjects in the elderly. [21][22][23][24] Research by Zureik et al. demonstrates that PWV is significantly and negatively associated with the spirometric parameter of forced expiratory volume in the first second (FEV1). 25 Furthermore, Sabit et al. compared PWV in COPD patients with healthy smokers and ex-smokers who did not suffer from the CV disease. PWV was higher in patients with COPD and inversely related to FEV1 values. 26 A review of the available literature did not find a study that comprehensively investigated the relationships of pulmonary function, CV risk assessment, and AS measured by the oscillometric method as an independent predictor of CV risk in patients with COPD and smokers without COPD and nonsmokers without COPD as control groups.
The main objective of this study was to determine and compare A total of 208 subjects were included in this study divided into three groups: 61 subjects diagnosed with COPD, 83 smokers without COPD, and 64 nonsmokers without COPD. All subjects in the study, including smokers and nonsmokers without COPD were volunteers and all the diagnostic measures were made in the study time from July 2017 to April 2018.
Criteria for inclusion of subjects with COPD were as follows: men or women between 40 and 70 years of age, confirmed diagnosis of COPD without exacerbation in the last month; nonsmokerssomeone who has not smoked more than 100 cigarettes in their lifetime and does not currently smoke; smokers-current whose packyears index (PYI) ≥ 10 or former-someone who has smoked more than 100 cigarettes in their lifetime, but has not smoked in the last 28 days; patients receiving adequate therapy for COPD and patients without initiating therapy. Criteria for inclusion of non-COPD smokers were: men or women between 40 and 70 years of age, smokers-current or former with an intensity index with PYI ≥ 10.
Criteria for inclusion of non-COPD nonsmokers were: men or women between the ages of 40 and 70 years.
The excluded study criteria applied both to cases and controls were: persons suffering from lung diseases (active tuberculosis, bronchiectasis, pneumonia, lung cancer, lung fibrosis), patients on continuous oxygen therapy, inability to perform lung function tests, proven coronary heart disease or atherosclerotic disease of peripheral arteries, tachyarrhythmia or bradyarrhythmia of the heart and clinically manifest heart failure, unregulated diabetes, chronic renal failure, active rheumatic disease, autoimmune disease, and unregulated or poorly regulated arterial hypertension with mean values above 140/90 mmHg, respectively.
Anamnestic data were recorded in all subjects: physical examination was performed, and physical and anthropometric measurements were performed on calibrated devices: body weight, height, arterial pressure measurement with an oscillometric pressure gauge, pulse rate, and hemoglobin oxygen saturation. In the groups of subjects with COPD, the disease was assessed by BODEx index (BMI, FEV1, mMRC scale of dyspnea, frequency of exacerbations of COPD), and the method of pharmacological treatment of COPD was recorded in each subject. 27 3 | METHODS

| Ethical approval details and informed consent
In conducting the research, the laws of the Republic of Croatia and international conventions were fully respected. The study was approved by the Ethics Committee of the Medical Faculty of the University of Zagreb and the Ethics Committee of the Clinical Hospital Dubrava (No. 380-59-10106-16-20/269). All subjects were familiar with the conduct of the study and signed an informed consent to participate in the study.

| Assessment of pulmonary function
The assessment of pulmonary function was determined and GOLD criterion was applied and applied criterion was assessed postbronchodilator by spirometry according to ISO standards IS9001 and ISO13485 using a Minispir ® Light spirometer in accordance with the recommendations of the European Respiratory Society. 28 The following parameters were measured by spirometry: forced expiratory volume in the first second (FEV1), forced vital capacity (FVC), Tiffeneau-Pinelli index (FEV1/FVC), and airflow through the small airways. Using disposable, factory-calibrated nozzle airflow sensors, FlowMir ® , it was not necessary to calibrate the device. The obtained data were analyzed using the computer program Winspiro PRO ® PC.

| Arterial blood gas analysis
Blood for gas analysis of arterial blood was obtained by taking blood from the arteria radialis. 29 Blood was analyzed in a Gem premier 3000 analyser, Vetusi. Arterial blood gas analysis was used to analyze partial oxygen pressure (pO 2 /kPa), carbon dioxide partial pressure (pCO 2 /kPa), pH value, hydrogen carbonate ion concentration (mmol/L), and oxygen saturation (%).

| Analysis of hematological and biochemical parameters
Blood for analysis of hematological and biochemical parameters was sampled by venepuncture of the cubital vein using Vacutainer ® tubes with K 3 EDTA with anticoagulant (hematology) and without anticoagulant (biochemistry). Complete blood count (CBC) was determined in a Siemens Advia 2120i hematology analyser (Siemens Healthcare Diagnostics). For the assessment of biochemical parameters, the serum was isolated by centrifuging the blood at 1300 rpm for 10 min using a 32 Rotofix A centrifuge (Andreas Hettich GmbH & Co). From the biochemical parameters, the levels of glucose, urea, creatinine, triglycerides, total cholesterol, HDL-cholesterol, LDLcholesterol, and hs-CRP were determined. These parameters were analyzed in Beckman Coulter AU2700 plus and AU680 biochemical analysers. Serum hs-CRP concentration was determined by a highly specific immunoturbidimetric method on latex particles, 30

| Assessment of AS
AS was determined by measuring the aortic pulse wave velocity (PWVAo) and the AIX using a noninvasive TensioMed Arteriograph device and associated software (TensioMed Software v.1.10.0.2, TensioMed). The distance from the aortic arch to the iliac bifurcation was approximated by measuring the distance between the sternal jugulum and the pubic symphysis. 32 PWVAо and AIX values were presented as the mean values of the two measurements. The standard deviation (SD) was calculated for each beat when measured for 8 s.

| Ankle-brachial index (ABI)
The ABI is the ratio of the SBP measured at the ankle to that measured at the brachial artery of the arm. 33 The blood pressure cuff LJUBIČIĆ ET AL. is inflated proximal to the brachial artery of the arm and foot's posterior tibial or dorsalis pedis artery and systolic pressure on the foot determined by the Doppler Ultrasound with 8 megahertz peripheral probe. The highest ABPI ratio An ABPI between and including 0.90 and 1.29 considered normal, while a lesser than 0.9 indicates arterial disease, and an ABPI value of 1.3 or greater suggests severe calcification of the walls.

| CV risk assessment by SCORE risk calculation
In addition to the above methods, the assessment of increased CV risk was determined using the SCORE table of calculations for highrisk countries according to the guidelines for the prevention of CV disease in clinical practice from 2016. 15  Quantitative values were analyzed by Kolmogorov-Smirnov test and in the further analysis appropriate parametric statistical tests and data display methods were applied. Quantitative values are presented as mean and SD and 95% confidence intervals (95% CIs), while categorical values are presented in absolute numbers and corresponding proportions. One-way ANOVA was performed to establish significant differences between all three study groups (COPD, smokers without COPD, and nonsmokers without COPD).

| Statistical analyses
After the analysis of variance, a post hoc analysis according to Bonferroni with age and sex adjustment was additionally performed, to show the significance of individual interrelationships between each of the examined groups. The association of pulmonary parameters with CV risks was determined by Pearson's correlation coefficients (r), where the absolute value of the correlation coefficient >0.600 was rated as a strong correlation, from 0.300 to 0.599 as a moderate correlation, and <0.300 as a weak correlation.
All p values less than 0.05 were considered significant. Spearman's correlation coefficients rho (r s ) were used in the calculation of correlations of COPD phenotype and COPD diagnosis with CV risk parameters, given the nonparametric distribution of COPD phenotype and diagnosis according to GOLD.

| Socio-demographic and clinical characteristics of the examinees
According to the analysis of the socio-demographic and clinical characteristics of the respondents (Table 1), over two-thirds of respondents (69.2%) in the nonsmoking group were female, while in the smoking group there were 51 (61.4%) women and in the COPD group only 20 (32.8%) female (p < 0.001). In the COPD group, the majority of the respondents were between 55 and 64 years of age (54.1%), while in the groups of smokers and non-smokers without COPD symptoms, the age group was younger than 55 years. In the COPD group, 43 subjects (70.5%) actively smoked, and this prevalence was statistically significantly different from the group of smokers who did not have COPD symptoms (67.5%, p = 0.019).
Within the COPD group, the level of obstruction according to GOLD type 2: 23 prevails (37.7%). Most subjects with COPD were assigned to Groups A, B, and D, while the fewest were in Group C ( Table 2).
The most common phenotype in the study population was a nonexacerbator (72.1%), and the least was an exacerbator with chronic bronchitis (6.6%) and subjects with asthma-overlap syndrome and COPD (3.3%). 21.3% of subjects were not included in pharmacological treatment, which also coincides with the percentage of subjects in the GOLD A group (21.3%). Shortness of breath was rated by grade 50.0% of respondents according to the mMRC scale, while over 50% of them did not have an exacerbation of the disease in the previous year. Compared to active smokers without COPD symptoms, smokers with COPD symptoms had a statistically significantly longer smoking history (p < 0.001), smoked on average almost seven cigarettes more (p = 0.001), and had twice the PYI (p < 0.001). The analysis of anthropometric variables (height, body weight, BMI, jugulum-symphysis distance) did not reveal statistically significant differences between the examined groups.

| Differences between examined groups
By analyzing the results of spirometric parameters (Table 3), it was found that subjects in the COPD group had statistically significantly lower values of all spirometric parameters compared to control groups of smokers and nonsmokers without COPD (p < 0.001), while there were no statistically significant differences observed between smokers and nonsmokers.
Furthermore, analyses of the results of blood pressure, AS and ejection fraction parameters in the COPD group indicate statistically significantly increased values of SBP, mean arterial pressure, PWVAо and SBPAо compared to both control groups (p < 0.001), while LVETand P-wave return times indicate statistically significantly less (p < 0.001) ( Table 4). The average SD in all PWVAo measurements using the Arteriograph was below 1.1 m/s, which indicates excellent measurement quality. Also, the COPD group indicated statistically significantly increased values of pulse pressure and brachial pressure compared to the control group of smokers without COPD and diastolic pressure and heart rate per minute compared to the control group of nonsmokers without COPD (p < 0.05). No statistically significant differences were observed between the control groups of smokers and nonsmokers without COPD.
By analyzing the biochemical parameters (Table 5), the statistically most significant changes were visible in the decrease in hemoglobin oxygen saturation and the increased concentration of fibrinogen in the COPD group compared to both control groups (p < 0.001). Also, a statistically significant increase in hs-CRP concentration was found between the COPD group and the control group of non-COPD smokers (p < 0.001), while statistically significant differences in serum triglyceride concentrations were found only between control groups of smokers and nonsmokers without COPD (p < 0.05).
The analysis of hematological parameters (Table 6)     nonsmokers without COPD (p < 0.001), while statistically significant differences with respect to ABI index values were found only between control group of smokers and nonsmokers without COPD (p < 0.05) ( Table 7).
The level of obstruction according to GOLD was significantly associated with almost all the parameters of AS, and mostly with the values of PWVAo (r = 0.496, p < 0.001), P2 wave return time and SCORE by risk calculation 0.542, p < 0.001) (Tables 8 and 9).
These results suggest that a higher level of obstruction is associated with higher AS values, decreased O 2 concentration, increased CO 2 concentration, and higher SCORE calculation of CV risk (Figures 1-3).
A significant correlation between the parameters of pulmonary function and the level of obstruction according to GOLD was also found with the measured markers of systemic inflammation (Table 10) (Table 12).
Considering the nonparametric distribution of COPD phenotype and GOLD diagnosis (ABCD tool) ( CV disease has fallen by more than 30%, and a further increase in the incidence of COPD is projected in the next 20 years. 34  By analyzing the clinical characteristics of the group diagnosed with COPD, the most common phenotype of COPD in this study was a nonexacerbator with emphysema (54.1%), while the total percentage of subjects who were nonexacerbators (including nonexacerbators with chronic bronchitis) was 72% in our sample. The reason for such a high share of nonexacerbators can be seen from two aspects.
One is that these patients receive adequate therapy in regular pulmonary examinations, given that our sample consists entirely of subjects treated by pulmonologists and we can conclude that there is a good agreement in the distribution of clinical phenotypes with the FENEPOC study in the group of pulmonologically controlled patients. 36 Another reason is that one of the aims of this study was to detect an increased CV risk in patients with COPD without proven comorbidities that would interfere with the measurements, therefore we did not include patients with comorbidities. Namely, such patients are more prone to exacerbations, which is confirmed by a recent study by Dutch authors which showed that patients with COPD, who have one or more comorbidities, more often have ≥2 exacerbations per year, 37 therefore based on a sample of patients with COPD in this study cannot draw conclusions about the distribution of the COPD phenotype outside the sample. Also, it is particularly worth noting the fact that this study analyzed data from subjects with COPD without proven comorbidities, which according to population-based studies in which they make up about 10% of the sample, are the exception rather than the rule. 37 The analysis of anthropometric parameters (height, body weight, BMI, length of the jugulum symphysis) did not show statistically significant differences, which differs from previous studies in which a negative correlation between BMI and obstruction levels according to GOLD was proven. 38 This result can be explained by applying the criteria included in this study. Namely, patients with advanced COPD more often have lower BMI and more comorbidities, as well as obese patients, and then BMI contributes to the predictive power of the BODEx index, hence in the population of this study, which consisted of patients with "isolated" COPD without comorbidities, BMI had no predictive value.
Although this study included subjects without a confirmed diagnosis of arterial hypertension, which was the sole criterion, significant differences in systolic and diastolic pressure, heart rate per minute, and mean arterial pressure were noted between the study groups. The mean value of systolic pressure in all groups of the examined population does not exceed the value above 140 mmHg.
One-way analysis of variance showed that the value of systolic pressure was significantly higher in the group of subjects with COPD   By comparing the values of mean arterial pressure and central systolic aortic pressure between the examined groups, in the COPD group, statistically significantly higher values of these parameters were proved compared to the control groups of smokers and nonsmokers without COPD. Also, additional analysis by Pearson correlation coefficients confirmed a weak correlation between mean arterial pressure and GOLD obstruction level (r = 0.250, p < 0.001).
These results confirm the previously presented research data in which significant differences in blood pressure parameters between the examined groups were investigated. 41 However, in addition to using a different device in the study, it also included subjects with COPD who had CV comorbidities and did not show whether there was a difference between smokers and nonsmokers without COPD.
The data of the second group of researchers did not indicate significant differences in the examined and control group in the values of mean arterial pressure. 26 There was no control group of The results of several studies indicate that the level of aortic stiffness in patients with COPD is increased compared to control groups (after correction for age and sex) with positive anamnestic data on smoking. 26,43,44 Aortic stiffness, measured using PWVAo, is an independent factor in the prediction of CV disease, but it is still not a diagnostic method implemented in everyday practice. 22,24,45 To date, an association between AS and a reduction in the incidence of CV events has been reported in only one study in a limited sample of patients with advanced renal disease. 46 Also, in the Framingham study cohort, it has been shown that aortic stiffness further contributes to traditional CV risk factors in predicting the degree of risk. 47 In this study, the value of PWVAо, as an immediate marker of AS, was also statistically significantly higher in the COPD group   Numerous studies indicate a two-way relationship between COPD and the CV disease. 51,52 In one direction, patients with the coronary artery disease suffering from COPD have twice the risk of the CV disease compared to patients without COPD. In the second direction, patients with COPD have a higher risk of morbidity and mortality than CV disease. 53,54 In this study, observing the level of TNF-α, and fibrinogen in addition to CRP. 62 Also, studies indicate that CRP concentration is associated with increased mortality from COPD and that it negatively correlates with FEV1 values. 63,64 In this study, subjects in the COPD group also had a statistically significantly higher value of hs-CRP concentration. The COPD group had average hs-CRP values of 8.81 g/L, the nonsmoking group 3.56 g/L, while the smoking group had an average value of 1.61 g/L. Despite expectations, the nonsmoking group had higher hs-CRP values than the smoking group, but this difference was not statistically significant.
This can be explained by the fact that hs-CRP is a highly sensitive marker of inflammation and it is possible that in some subjects in the nonsmoking group hs-CRP was elevated for some other reasons that were not covered by the exclusive criteria or were obvious when taking medical history or physical examination. These results are consistent with the GENOA (The Genetic Epidemiology Network of Arteriopathy study) study, which, among other things, investigated the association between smoking and inflammation. The authors did not find a significant association between smoking intensity and any of the inflammatory markers studied. 65 Also, in addition to hs-CRP, in this study, a statistically significantly higher value and concentration  Since the CR3 complement receptor, detected on neutrophils, NK cells, and macrophages, has the ability to bind fibrin, the authors hypothesize this mechanism as a possible option to explain the association between fibrinogen concentration and pneumonia. 69 The results of the post hoc analysis of correlation coefficients indicate a significant correlation between the concentration of hs-CRP and fibrinogen with the level of obstruction according to GOLD and all measured spirometric parameters, which confirmed the results of previous studies (Table 10). 63,64 Also, by post hoc analysis of the results of all smokers in the sample, a significant association between hs-CRP and fibrinogen with the pack-year index was found. From the above, we can conclude that twice as much exposure to tobacco smoke is an important factor in the group of subjects with COPD compared to the group of smokers without COPD, therefore the higher the exposure the higher the concentration of hs-CRP is.

| LIMITATIONS OF THE STUDY
There is a lack of generalizability of findings for age and sex differences since a relatively small clinical study sample (n = 208) was analyzed and the sample size calculation was not sufficient to allow the adequate gender stratification of the analyses, but this is a wellknown fact that most COPD subjects are males and over 55 years, as this is accordingly presentable in such a manner in this study.
Comparison analyses were not adjusted for pack-years in COPD cases when compared with two different control groups of smokers without COPD and nonsmokers without COPD, but this presents no clinical obstacle in obtaining conclusions regarding the fact that in calculations, including CV risk assessment, it is taken in a binary mode.
In this study, patients with comorbidities such as coronary heart disease or atherosclerotic disease of peripheral arteries, arrhythmias or manifest heart failure, unregulated diabetes or hypertension, chronic renal failure, active rheumatic disease, and autoimmune T A B L E 12 Correlation coefficients of arterial stiffness, SCORE risk calculation, ABI index, and PWVAо with inflammatory markers and arterial blood gas analysis parameters Abbreviations: ABI, ankle-brachial index; hs-CRP, high sensitivity protein; PWVAo, aortic pulse wave velocity; r, correlation coefficient.
T A B L E 13 Spearman's correlation coefficients of COPD phenotype and COPD assessment according to GOLD in relation to SCORE risk calculation, PWVAо, hs-CRP, and fibrinogen Abbreviations: hs-CRP, high sensitivity protein; PWVAo, aortic pulse wave velocity; r, correlation coefficient. disease since it is recognized that these comorbidities alone affect biological markers of systemic inflammation. However, such a choice introduces a selection bias that renders the study subjects not representative of the universe of the COPD patients, most of whom have comorbidities.