Smoking, Systemic Inflammation, and Airflow Limitation: A Mendelian Randomization Analysis of 98 085 Individuals From the General Population

Smoking, Systemic Inflammation, and Airflow Limitation: A Mendelian Randomization Analysis of... Abstract Introduction Smoking is associated with systemic and local inflammation in the lungs. Furthermore, in chronic obstructive pulmonary disease, which is often caused by smoking, there is often systemic inflammation that is linked to lung function impairment. However, the causal pathways linking smoking, systemic inflammation, and airflow limitation are still unknown. We tested whether higher tobacco consumption is associated with higher systemic inflammation, observationally and genetically and whether genetically higher systemic inflammation is associated with airflow limitation. Methods We included 98 085 individuals aged 20–100 years from the Copenhagen General Population Study; 36589 were former smokers and 16172 were current smokers. CHRNA3 rs1051730 genotype was used as a proxy for higher tobacco consumption and the IL6R rs2228145 genotype was used for higher systemic inflammation. Airflow limitation was defined as forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) <70%. Results Difference in plasma level of C-reactive protein was 4.8% (95% CI = 4.4% to 5.2%) per 10 pack-year increase and 1.6% (95% CI = 0.4% to 2.8%) per T allele. Corresponding differences were 1.2% (95% CI = 1.1% to 1.3%) and 0.5% (95% CI = 0.3% to 0.8%) for fibrinogen, 1.2% (95% CI = 1.2% to 1.3%) and 0.7% (95% CI = 0.5% to 1.0%) for α1-antitrypsin, 2.0% (95% CI = 1.8% to 2.1%) and 0.7% (95% CI = 0.4% to 1.1%) for leukocytes, 1.9% (95% CI = 1.8% to 2.1%) and 0.8% (95% CI = 0.4% to 1.2%) for neutrophils, and 0.8% (95% CI = 0.7% to 1.0%) and 0.4% (95% CI = 0.1% to 0.7%) for thrombocytes. The differences in these levels were lower for former smokers compared with current smokers. The IL6R rs2228145 genotype was associated with higher plasma acute-phase reactants but not with airflow limitation. Compared with the C/C genotype, the odds ratio for airflow limitation was 0.95 (95% CI = 0.89 to 1.02) for A/C genotype and 0.94 (95% CI = 0.87 to 1.01) for A/A genotype. Conclusions Higher tobacco consumption is associated with higher systemic inflammation both genetically and observationally, whereas systemic inflammation was not associated with airflow limitation genetically. Implications The association between higher tobacco consumption and higher systemic inflammation may be causal, and the association is stronger among current smokers compared to former smokers, indicating that smoking cessation may reduce the effects of smoking on systemic inflammation. Systemic inflammation does not seem to be a causal driver in development of airflow limitation. These findings can help to understand the pathogenic effects of smoking and the interplay between smoking, systemic inflammation, and airflow limitation and hence development and progression of chronic obstructive pulmonary disease. Introduction Smoking is considered to be the most important risk factor for development of chronic obstructive pulmonary disease (COPD), characterized by persistent airflow limitation that is usually progressive, by inducing local inflammation in the lungs.1 However, smoking is not only associated with an increased degree of inflammation in the lungs but also in the systemic circulation.2 Furthermore, individuals with COPD often have systemic inflammation with high levels of C-reactive protein and fibrinogen3 and high levels of these biomarkers have been independently associated with lung function impairment.4–8 Despite the many studies on the importance of systemic inflammation in COPD,3,9 it is still unknown whether systemic inflammation plays a role in development and progression of the disease. Thus, the causal pathways linking smoking, systemic inflammation, and airflow limitation are still unknown. Observational studies are usually prone to confounding and reverse causation. Although randomized clinical trials are the gold standard in establishing causal relationships, some exposures such as smoking would be unethical or impractical to study using such a study design. An alternative approach is to use genetics in a Mendelian randomization analysis.10 By using the genetic variant CHRNA3 rs1051730, where the T allele is associated with higher tobacco consumption, it is possible to assess the association of higher tobacco consumption with systemic inflammation and lung function largely free of confounding and reverse causation.11–15 Due to the random distribution of alleles at conception, genetic variants should not be associated with potential confounders and, as genes are present at birth, genetic variants are not susceptible to reverse causation.10 However, as the CHRNA3 rs1051730 genotype may be associated with several tobacco consumption behaviors, instrumental variable estimates will not be calculated as the relevant exposure is difficult to define unambiguously. Thus, our approach will be akin to carrying out an intention-to-treat analysis in a randomized intervention trial and not estimating the effect of the intervention but rather of the randomization, and is the preferable approach when the assumptions for calculating instrumental variable estimates may be in doubt.16 Similarly, by using the genetic variant IL6R rs2228145, where the A allele is associated with an increased activity of the interleukin-6 receptor in hepatocytes and leucocytes, which causes increased plasma levels of C-reactive protein and other acute-phase reactants,17–24 it is possible to assess consequences of systemic inflammation by circumventing confounding and reverse causation. Since C-reactive protein and other acute-phase reactants are downstream products of inflammation, we used a more proximal regulator of inflammation, that is, interleukin-6 receptor, to investigate the association more generally rather than using individual markers alone. The purpose of the present study was to investigate the causal pathways linking smoking, systemic inflammation, and airflow limitation. Firstly, we tested whether higher tobacco consumption is associated with higher systemic inflammation, observationally and genetically (Supplementary Figure S1, Panel A). Secondly, we tested whether genetically higher systemic inflammation is associated with airflow limitation (Supplementary Figure S1, Panel B). For this purpose, we used the Copenhagen General Population Study, including 98 085 individuals, of whom 36589 were former smokers and 16172 current smokers. Methods Study Design and Subjects We recruited 98 085 individuals aged 20–100 years from the Copenhagen General Population Study, a population-based prospective cohort study.25–28 All individuals in Denmark are assigned a unique identification number at birth or immigration and registered in the national Danish Civil Registration System. Thus, participants were randomly selected from the national Danish Civil Registration System to reflect the adult white Danish population of Danish descent. All participants completed a comprehensive questionnaire, underwent a physical examination, and gave blood for biochemical and genetic analyses in the period of 2003–2013. Questionnaires were reviewed in detail at the day of attendance by a health care professional together with the participant. The study was approved by the Herlev and Gentofte Hospital and by a Danish ethical committee and conducted according to the Declaration of Helsinki (approval number: H-KF-01-144/01). Written informed consent was obtained from all participants. Smoking, Systemic Inflammation, and Airflow Limitation Smoking status was self-reported and defined as never, former, or current smokers. Tobacco consumption was defined as cumulative tobacco consumed through smoking and measured in pack-years based on information on duration of tobacco consumption and current amount of consumed tobacco: one pack-year was 20 cigarettes or equivalent smoked daily for 1 year. Nevertheless, tobacco consumption was also investigated in the form of daily amount of consumed tobacco in current smokers, thus also investigating immediate exposure. Systemic inflammation was assessed by acute-phase reactants (C-reactive protein [high-sensitivity assay], fibrinogen, and α1-antitrypsin) and cells of the immune system (leukocytes, neutrophils, and thrombocytes), and measured in the systemic circulation with standard hospital assays. Spirometry was performed with a Vitalograph (Maids Moreton, Buckinghamshire, UK) in the first 14625 participants and with an EasyOne Spirometer (ndd Medical Technologies, Zurich, Switzerland) in the remaining participants with measurements of pre-bronchodilatatory forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC).29 Predicted values were calculated using internally derived reference values based on a subsample of healthy asymptomatic never-smokers with age and height as covariates separately for men and women and separately for the two spirometers (Supplementary Figure S2).26 Presence of airflow limitation was defined according to (1) a fixed ratio defined as FEV1/FVC<70% and (2) lower limit of normal (LLN) defined as FEV1/FVC<LLN, calculated as the mean predicted value for age and sex multiplied by −1.645 standard error. Genotyping Genotyping was conducted blind to information on smoking status, tobacco consumption, degree of systemic inflammation, and presence of airflow limitation. The ABI PRISM 7900HT Sequence Detection System (Applied Biosystems Inc.) was used to genotype CHRNA3 rs1051730 with TaqMan assays. Call-rates were above 99.8% as we performed re-runs twice. IL6R rs2228145 was genotyped similarly to CHRNA3 rs1051730. Information on the CHRNA3 rs1051730 genotype was available in 98 085 individuals whereas the IL6R rs2228145 genotype was available in 81397 individuals. Potential Confounders Information on potential confounders was acquired from the questionnaires and physical examinations. Body mass index (BMI) was measured weight divided by measured height squared (kg/m2). Alcohol consumption was reported in units per week and converted to grams (1 unit=12 grams). Occupational exposure to dust and/or fumes, daily exposure to passive smoking, and asthma was self-reported. Physical activity in leisure time was reported according to hours per week and degree of activity. Education was based on years attending school. Income was reported as annual household income. Statistical Analyses STATA/SE 13.1 was used in the statistical analyses. Deviation from Hardy–Weinberg equilibrium was investigated using the chi-square tests to assess the frequency of allele distribution across the individual genotypes; deviation may suggest genotyping or population sampling errors.30 Cuzick’s non-parametric trend test was used to assess trend across ordered groups to investigate association with potential confounders. In addition, association with potential confounders were also investigated using linear regression for continuous covariates and logistic regression for dichotomous covariates. In observational analyses, we adjusted for potential confounders that could be associated with tobacco consumption, inflammatory biomarkers, and/or airflow limitation, that is, for age, sex, BMI, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. In the genetic analyses, we adjusted for age and sex. Furthermore, studies indicate that the CHRNA3 rs1051730 genotype may also be associated with BMI,11,28 and the association may be modified by smoking, where the genotype is associated with lower BMI in current smokers and higher BMI in never-smokers (Supplementary Figure S3).31 In addition, higher BMI is associated with higher degree of systemic inflammation.32,33 Thus, the effects on BMI could be considered to be a pleiotropic effect of the genotype, which could bias the results and, therefore, we performed the genetic analyses with and without additional adjustment for BMI. Also, the analyses stratified for smoking status can be used to assess potential pleiotropy, as different associations in ever-smokers versus never-smokers may hint to other biological effects of genetic variants in the CHRNA3. Association of tobacco consumption and the genotypes with each other and with the inflammatory biomarkers was investigated using multiple linear regressions. Levels of the inflammatory biomarkers were logarithmic transformed in order to approach a normal distribution. Association of tobacco consumption and the individual genotypes with airflow limitation was investigated using logistic regression. Association of airflow limitation with the inflammatory biomarkers was investigated using multiple linear regressions. Results Among 98 085 individuals from the Copenhagen General Population Study, 36589 were former smokers and 16172 were current smokers (Supplementary Figure S4). Tobacco consumption was associated with almost all of the potential confounders in former and current smokers (Table 1 and Supplementary Table S1), as well as in the entire study population (Supplementary Table S2). In contrast, the CHRNA3 rs1051730 genotype was not associated with most potential confounders in former and current smokers (Table 1), or in the entire study population (Supplementary Table S3). However, the CHRNA3 rs1051730 genotype was associated with age, being more often a current smoker, having a lower BMI as well as FEV1 and FVC as percentage of predicted values and FEV1/FVC in former and current smokers, as shown previously11,28; the associations with BMI and lung function measurements should be interpreted as downstream effects of tobacco consumption rather than potential confounders. Similarly, the IL6R rs2228145 genotype was not associated with any of the potential confounders in former and current smokers (Supplementary Table S4), or in the entire study population (Supplementary Table S5). There was no clear evidence of deviation from Hardy–Weinberg equilibrium; p = .99 for CHRNA3 rs1051730 genotype and p = .80 for IL6R rs2228145 genotype. Table 1. Baseline Characteristics of Former and Current Smokers in the Copenhagen General Population Study According to the CHRNA3 rs1051730 Genotype   CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)    CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)  BMI = body mass index; CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity. Data are summarized as median with the 25th and 75th percentiles, or %, or effect sizes with 95% confidence intervals. aWhen p for trend is adjusted for 14 number of individual trend analyses according to the Bonferroni method, p = .05 is equivalent to p = .05/14 = .004. bEffect sizes report beta coefficients on linear regression for continuous covariates and on logistic regression for dichotomous covariates. cThis p value indicates that the number of current smokers was higher among those with higher tobacco consumption (please see Supplementary Table S1). View Large Table 1. Baseline Characteristics of Former and Current Smokers in the Copenhagen General Population Study According to the CHRNA3 rs1051730 Genotype   CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)    CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)  BMI = body mass index; CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity. Data are summarized as median with the 25th and 75th percentiles, or %, or effect sizes with 95% confidence intervals. aWhen p for trend is adjusted for 14 number of individual trend analyses according to the Bonferroni method, p = .05 is equivalent to p = .05/14 = .004. bEffect sizes report beta coefficients on linear regression for continuous covariates and on logistic regression for dichotomous covariates. cThis p value indicates that the number of current smokers was higher among those with higher tobacco consumption (please see Supplementary Table S1). View Large Association of the CHRNA3 rs1051730 Genotype With Tobacco Consumption Collectively, the results show that the CHRNA3 rs1051730 genotype was associated with higher risk of being current smokers among ever smokers and in the general population (Table 1 and Supplementary Table S1), higher cumulative tobacco consumption in former and current smokers, and higher daily tobacco consumption among current smokers (Supplementary Figure S5). Cumulative tobacco consumption in former and current smokers was 20.3 pack-years (95% confidence interval [CI] = 20.1 to 20.6 pack-years) for the C/C genotype, 21.8 pack-years (95% CI = 21.6 to 22.1 pack-years) for the C/T genotype, and 23.2 pack-years (95% CI = 22.7 to 23.7 pack-years) for the T/T genotype. In addition, daily amount of smoked tobacco in current smokers was 14.8 g/day (95% CI = 14.6 to 15.1 g/day) for the C/C genotype, 15.7 g/day (95% CI = 15.5 to 15.9 g/day) for the C/T genotype, and 16.8 g/day (95% CI = 16.3 to 17.2 g/day) for the T/T genotype. Association of Tobacco Consumption and CHRNA3 rs1051730 With Systemic Inflammation Higher tobacco consumption in former and current smokers was associated with higher levels of acute-phase reactants, including C-reactive protein, fibrinogen, and α1-antitrypsin, and cellular markers of inflammation, including leukocytes, neutrophils, and thrombocytes, in the systemic circulation with a dose-dependent relationship (Figure 1 and Supplementary Figures S6 and S7). In former and current smokers, the multivariable adjusted percent difference in the level of inflammatory biomarkers per 10 pack-year increase was 4.8% (95% CI = 4.4% to 5.2%) for C-reactive protein, 1.2% (95% CI = 1.1% to 1.3%) for fibrinogen, 1.2% (95% CI = 1.2% to 1.3%) for α1-antitrypsin, 2.0% (95% CI = 1.8% to 2.1%) for leukocytes, 1.9% (95% CI = 1.8% to 2.1%) for neutrophils, and 0.8% (95% CI = 0.7% to 1.0%) for thrombocytes (Figure 1). When former and current smokers were analyzed separately, the percent differences in the level of inflammatory biomarkers for former smokers were lower compared with current smokers. Interaction analyses between former and current smokers suggested that these associations were strongest among current smokers. Figure 1. View largeDownload slide Association of tobacco consumption with systemic inflammation. Levels of inflammatory biomarkers are summarized as percentage difference with 95% confidence interval. Models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. p values were obtained from the likelihood ratio tests. CI = confidence interval. Figure 1. View largeDownload slide Association of tobacco consumption with systemic inflammation. Levels of inflammatory biomarkers are summarized as percentage difference with 95% confidence interval. Models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. p values were obtained from the likelihood ratio tests. CI = confidence interval. Similarly, the CHRNA3 rs1051730 genotype in former and current smokers was associated with higher levels of acute-phase reactants and cellular markers of inflammation with a dose-dependent relationship (Figure 2 and Supplementary Figure S6). After adjustment for BMI, the association seemed slightly stronger, especially for current smokers (Figure 2). In former and current smokers, the age and sex adjusted percent difference in the level of inflammatory biomarkers per T allele was 1.6% (95% CI = 0.4% to 2.8%) for C-reactive protein, 0.5% (95% CI = 0.3% to 0.8%) for fibrinogen, 0.7% (95% CI = 0.5% to 1.0%) for α1-antitrypsin, 0.7% (95% CI = 0.4% to 1.1%) for leukocytes, 0.8% (95% CI = 0.4% to 1.2%) for neutrophils, and 0.4% (95% CI = 0.1% to 0.7%) for thrombocytes (Figure 2). With regard to never-smokers, the CHRNA3 rs1051730 genotype was positively associated with C-reactive protein and negatively associated with leukocytes and neutrophils; although, these associations were weak and no associations with other inflammatory biomarkers could be observed. Interaction analyses between never and former and current smokers suggested that these associations were strongest among former and current smokers and that with regard to C-reactive protein and fibrinogen, there seems to be no difference between the different smoking categories. When former and current smokers were analyzed separately, the percent differences in the levels of inflammatory biomarkers were lower for former smokers compared to current smokers. Figure 2. View largeDownload slide Association of the CHRNA3 rs1051730 genotype with systemic inflammation. Levels of inflammatory biomarkers are summarized as % difference with 95% confidence interval. Models were adjusted for age and sex and when indicated additionally for body mass index. P values were obtained from the likelihood ratio tests. CI = confidence interval. Figure 2. View largeDownload slide Association of the CHRNA3 rs1051730 genotype with systemic inflammation. Levels of inflammatory biomarkers are summarized as % difference with 95% confidence interval. Models were adjusted for age and sex and when indicated additionally for body mass index. P values were obtained from the likelihood ratio tests. CI = confidence interval. Association of IL6R rs2228145 With Systemic Inflammation and Tobacco Consumption Collectively, the results show that the IL6R rs2228145 genotype was associated with higher levels of acute-phase reactants but unclear with regard to levels of cellular markers of inflammation in the systemic circulation (Supplementary Figures S8 and S9). In former and current smokers, the age and sex adjusted percent difference in the level of C-reactive protein was 8.3% (95% CI = 5.6% to 11%) for the A/C genotype and 15% (95% CI = 12% to 18%) for the A/A genotype. Corresponding levels were 1.2% (95% CI = 0.7% to 1.8%) and 1.8% (95% CI = 1.2% to 2.4%) for fibrinogen, 0.5% (95% CI = −0.05% to 1.0%) and 0.6% (95% CI = 0.03% to 1.2%) for α1-antitrypsin, −0.2% (95% CI = −0.9% to 0.4%) and 0.3% (95% CI = −0.4% to 0.9%) for leukocytes, −0.02% (95% CI = −0.8% to 0.8%) and 0.6% (95% CI = −0.2% to 1.5%) for neutrophils, and −0.4% (95% CI = −1.1% to 0.2%) and −1.0% (95% CI = −1.6% to −0.4%) for thrombocytes (Supplementary Figure S8). As seen, the IL6R rs2228145 genotype was associated with lower levels of thrombocytes. These results suggest that the IL6R rs2228145 genotype can be used as a genetic instrument for the acute-phase reactants but not for the cellular markers of inflammation. The IL6R rs2228145 genotype was not associated with any tobacco consumption phenotype (Supplementary Tables S4 and S5). Association of Tobacco Consumption, CHRNA3 rs1051730, and IL6R rs2228145 With Airflow Limitation Higher tobacco consumption and the CHRNA3 rs1051730 genotype were associated with a higher risk of airflow limitation in former and current smokers regardless of the chosen criterion for airflow limitation, with a dose-dependent relationship (Figure 3). Compared to <15 pack-years, the odds ratio (OR) for airflow limitation according to FEV1/FVC < 70% was 1.63 (95% CI = 1.53 to 1.74) for 15–29.9 pack-years and 2.61 (95% CI = 2.46 to 2.78) for ≥30 pack-years. Compared with the C/C genotype, the OR for airflow limitation according to FEV1/FVC < 70% was 1.10 (95% CI = 1.05 to 1.16) for C/T genotype and 1.27 (95% CI = 1.18 to 1.37) for T/T genotype. The IL6R rs2228145 genotype was not associated with airflow limitation regardless of the chosen criterion for airflow limitation (Figure 3 and Supplementary Figure S10). Compared to the C/C genotype, the OR for airflow limitation according to FEV1/FVC < 70% was 0.95 (95% CI = 0.89 to 1.02) for A/C genotype and 0.94 (95% CI = 0.87 to 1.01) for A/A genotype (Figure 3). Figure 3. View largeDownload slide Association of tobacco consumption, CHRNA3 rs1051730, and IL6R rs2228145 with airflow limitation in former and current smokers. In the observational analyses, models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. In the genetic analyses, models were adjusted for age and sex. p values were obtained from the Wald tests from the logistic regression models. CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity; OR = odds ratio; LLN = lower limit of normal. Figure 3. View largeDownload slide Association of tobacco consumption, CHRNA3 rs1051730, and IL6R rs2228145 with airflow limitation in former and current smokers. In the observational analyses, models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. In the genetic analyses, models were adjusted for age and sex. p values were obtained from the Wald tests from the logistic regression models. CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity; OR = odds ratio; LLN = lower limit of normal. Sensitivity Analyses The CHRNA3 rs1051730 genotype was, as expected, not associated with airflow limitation in never-smokers (Supplementary Figure S11). Among all individuals, there was an interaction of the CHRNA3 rs1051730 genotype with smoking status on risk of airflow limitation with highest risk estimates for current smokers (p for interaction = 1 × 10−4). Airflow limitation was associated with higher levels of acute-phase reactants and cellular markers of inflammation regardless of the chosen criterion for airflow limitation (Supplementary Figure S12). Discussion Using a large sample from the general population, we found that higher tobacco consumption was associated with higher systemic inflammation in observational and genetic analyses, whereas systemic inflammation was not associated with airflow limitation in genetic analyses. Our approach supports the notion that the association between higher tobacco consumption and higher systemic inflammation may be causal, and the association is stronger among current smokers compared to former smokers, indicating that smoking cessation may reduce the effects of smoking on systemic inflammation. Furthermore, systemic inflammation does not seem to be a causal driver in development of airflow limitation. These are novel findings that can help to understand the pathogenic effects of smoking and the interplay between smoking, systemic inflammation, and airflow limitation and hence development and progression of COPD. There may be several potential mechanisms involved in the association between smoking, systemic inflammation, and airflow limitation. Although smoking is associated with local inflammation in the lungs34 and local inflammation in the lungs is associated with increased lung function decline and disease severity,35,36 the mechanisms behind systemic inflammation in smokers and in COPD patients are unclear. Our results in combination with previous studies show that smoking induces systemic inflammation, whereas variation in the IL6R and CRP genes, which genetically determined higher systemic inflammation or markers of such, is not associated with an increased risk of COPD in previous Mendelian randomization studies of our and other populations.37,38 This supports the notion that systemic inflammation in individuals with COPD is a marker of smoking exposure and possibly presence of more severe disease rather than a driving force of disease progression. However, our results also point to that a direct pathway may exist from smoking to diseases where systemic inflammation may be a contributing cause, for example, coronary heart disease.20,39 In a Mendelian randomization analysis, where the causal association between tobacco consumption with cardiovascular risk factors was investigated, the association with C-reactive protein was not statistically significant; however, the observational and genetic estimates were congruent and in the same direction as observed in the present study but C-reactive protein measurements were only available in a subsample.40 Nevertheless, further genetic studies investigating the causality of smoking through systemic inflammation for other diseases are still needed. Interestingly, the IL6R rs2228145 genotype was associated with higher levels of acute-phase reactants but not clearly associated with cellular markers of inflammation. A potential explanation may be that the IL6R rs2228145 genotype primarily affects the acute-phase reactants while the effects on cellular markers of inflammation are weak. Thus, though the study was well powered to investigate the association of the acute-phase reactants with airflow limitation, the cellular markers deserves further investigation. Potential limitations in genetic studies include population stratification and genetic pleiotropy. However, as we had an ethnically homogenous population, the complicating effects of population stratification are likely to have been avoided. Since genotype distributions did not appear to differ from Hardy–Weinberg equilibrium, we have also likely avoided genotyping and population sampling errors. It is also important to note that stratification on smoking status may introduce the possibility of collider stratification bias.28,41,42 Yet, stratification by smoking status also provides an important test of the no pleiotropy assumption of Mendelian randomization studies; we observed no clear association between genotype and outcomes in never-smokers, supporting the notion that pleiotropy does not seem to be biasing our results; however, the CHRNA3 rs1051730 genotype was marginally associated with higher levels of C-reactive protein and lower levels of leukocytes and neutrophils in never-smokers, something that clearly should be investigated further in future studies. Former and current smokers were also extensively investigated according to the CHRNA3 rs1051730 genotype with regard to potential confounders without observed differences. While we were able to show that smoking induces systemic inflammation and that systemic inflammation may not induce airflow limitation, an important limitation is the lack of a specific genetic variant for development of airflow limitation. It could very well be that airflow limitation in itself could induce systemic inflammation, but to investigate this hypothesis we would require genetic variants associated with airflow limitation independently of smoking behaviors and systemic inflammation. Unfortunately, such an investigation is not feasible in the present study. Lastly, both observational and Mendelian randomization studies could potentially be influenced by survival bias, for example, heavier smokers are likely to die earlier and therefore not be included in the analysis. Strengths of the present study are use of a large sample with 98 085 individuals from the general population, including a large sample of smokers. Furthermore, we had extensive information on genotypes, inflammatory biomarkers in the systemic circulation, lung function, and several tobacco consumption variables. In conclusion, higher tobacco consumption is associated with higher systemic inflammation both genetically and observationally, while systemic inflammation was not associated with airflow limitation genetically. These results are consistent with the notion that the association between higher tobacco consumption and higher systemic inflammation may be causal, whereas systemic inflammation does not seem to be a causal driver in development of airflow limitation. Furthermore, the association of smoking to systemic inflammation was stronger among current smokers compared to former smokers, indicating that smoking cessation may reduce the effects of smoking on systemic inflammation. Supplementary Material Supplementary data are available at Nicotine & Tobacco Research online. Funding This work was supported by the Lundbeck Foundation, Danish Lung Association, the Danish Cancer Society, Herlev and Gentofte Hospital, Copenhagen University Hospital, and University of Copenhagen. The funders had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication. Declaration of Interests YÇ reports personal fees from Boehringer Ingelheim and AstraZeneca outside of the submitted work. PL reports grants from AstraZeneca and GlaxoSmithKline and personal fees from Boehringer Ingelheim, AstraZeneca, Novartis, and GlaxoSmithKline outside of the submitted work. SA and BGN have nothing to disclose. References 1. Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Lung Disease. Global Initiative for Chronic Obstructive Lung Disease (GOLD). http://www.goldcopd.org/gold-2017-global-strategy-diagnosis-management-prevention-copd/. Accessed February 23, 2017. 2. Yanbaeva DG, Dentener MA, Creutzberg EC, Wesseling G, Wouters EF. Systemic effects of smoking. Chest . 2007; 131( 5): 1557– 1566. Google Scholar CrossRef Search ADS PubMed  3. Gan WQ, Man SF, Senthilselvan A, Sin DD. Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and a meta-analysis. Thorax . 2004; 59( 7): 574– 580. Google Scholar CrossRef Search ADS PubMed  4. Aronson D, Roterman I, Yigla M, et al.   Inverse association between pulmonary function and C-reactive protein in apparently healthy subjects. Am J Respir Crit Care Med . 2006; 174( 6): 626– 632. Google Scholar CrossRef Search ADS PubMed  5. Dahl M, Tybjaerg-Hansen A, Vestbo J, Lange P, Nordestgaard BG. Elevated plasma fibrinogen associated with reduced pulmonary function and increased risk of chronic obstructive pulmonary disease. Am J Respir Crit Care Med . 2001; 164( 6): 1008– 1011. Google Scholar CrossRef Search ADS PubMed  6. Donaldson GC, Seemungal TA, Patel IS, et al.   Airway and systemic inflammation and decline in lung function in patients with COPD. Chest . 2005; 128( 4): 1995– 2004. Google Scholar CrossRef Search ADS PubMed  7. Shaaban R, Kony S, Driss F, et al.   Change in C-reactive protein levels and FEV1 decline: A longitudinal population-based study. Respir Med . 2006; 100( 12): 2112– 2120. Google Scholar CrossRef Search ADS PubMed  8. Yende S, Waterer GW, Tolley EA, et al.   Inflammatory markers are associated with ventilatory limitation and muscle dysfunction in obstructive lung disease in well functioning elderly subjects. Thorax . 2006; 61( 1): 10– 16. Google Scholar CrossRef Search ADS PubMed  9. Yanbaeva DG, Dentener MA, Creutzberg EC, Wouters EF. Systemic inflammation in COPD: Is genetic susceptibility a key factor? COPD . 2006; 3( 1): 51– 61. Google Scholar CrossRef Search ADS PubMed  10. Smith GD, Ebrahim S. ‘Mendelian randomization’: Can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol . 2003; 32( 1): 1– 22. Google Scholar CrossRef Search ADS PubMed  11. Freathy RM, Kazeem GR, Morris RW, et al.   Genetic variation at CHRNA5-CHRNA3-CHRNB4 interacts with smoking status to influence body mass index. Int J Epidemiol . 2011; 40( 6): 1617– 1628. Google Scholar CrossRef Search ADS PubMed  12. Munafò MR, Timofeeva MN, Morris RW, et al.  ; EPIC Study Group. Association between genetic variants on chromosome 15q25 locus and objective measures of tobacco exposure. J Natl Cancer Inst . 2012; 104( 10): 740– 748. Google Scholar CrossRef Search ADS PubMed  13. Kaur-Knudsen D, Nordestgaard BG, Bojesen SE. CHRNA3 genotype, nicotine dependence, lung function and disease in the general population. Eur Respir J . 2012; 40( 6): 1538– 1544. Google Scholar CrossRef Search ADS PubMed  14. Thorgeirsson TE, Geller F, Sulem P, et al.   A variant associated with nicotine dependence, lung cancer and peripheral arterial disease. Nature . 2008; 452( 7187): 638– 642. Google Scholar CrossRef Search ADS PubMed  15. Thorgeirsson TE, Gudbjartsson DF, Surakka I, et al.  ; ENGAGE Consortium. Sequence variants at CHRNB3-CHRNA6 and CYP2A6 affect smoking behavior. Nat Genet . 2010; 42( 5): 448– 453. Google Scholar CrossRef Search ADS PubMed  16. Nitsch D, Molokhia M, Smeeth L, DeStavola BL, Whittaker JC, Leon DA. Limits to causal inference based on Mendelian randomization: A comparison with randomized controlled trials. Am J Epidemiol . 2006; 163( 5): 397– 403. Google Scholar CrossRef Search ADS PubMed  17. Danik JS, Paré G, Chasman DI, et al.   Novel loci, including those related to Crohn disease, psoriasis, and inflammation, identified in a genome-wide association study of fibrinogen in 17 686 women: The Women’s Genome Health Study. Circ Cardiovasc Genet . 2009; 2( 2): 134– 141. Google Scholar CrossRef Search ADS PubMed  18. Müllberg J, Oberthür W, Lottspeich F, et al.   The soluble human IL-6 receptor. Mutational characterization of the proteolytic cleavage site. J Immunol . 1994; 152( 10): 4958– 4968. Google Scholar PubMed  19. Swerdlow DI, Holmes MV, Kuchenbaecker KB, et al.   The interleukin-6 receptor as a target for prevention of coronary heart disease: A mendelian randomisation analysis. Lancet . 2012; 379(9822): 1214– 1224. 20. Sarwar N, Butterworth AS, Freitag DF, et al.   Interleukin-6 receptor pathways in coronary heart disease: A collaborative meta-analysis of 82 studies. Lancet . 2012; 379(9822): 1205– 1213. 21. Boekholdt SM, Stroes ES. The interleukin-6 pathway and atherosclerosis. Lancet . 2012; 379( 9822): 1176– 1178. Google Scholar CrossRef Search ADS PubMed  22. Elliott P, Chambers JC, Zhang W, et al.   Genetic loci associated with C-reactive protein levels and risk of coronary heart disease. JAMA . 2009; 302( 1): 37– 48. Google Scholar CrossRef Search ADS PubMed  23. Galicia JC, Tai H, Komatsu Y, Shimada Y, Akazawa K, Yoshie H. Polymorphisms in the IL-6 receptor (IL-6R) gene: Strong evidence that serum levels of soluble IL-6R are genetically influenced. Genes Immun . 2004; 5( 6): 513– 516. Google Scholar CrossRef Search ADS PubMed  24. Reich D, Patterson N, Ramesh V, et al.  ; Health, Aging and Body Composition (Health ABC) Study. Admixture mapping of an allele affecting interleukin 6 soluble receptor and interleukin 6 levels. Am J Hum Genet . 2007; 80( 4): 716– 726. Google Scholar CrossRef Search ADS PubMed  25. Afzal S, Brøndum-Jacobsen P, Bojesen SE, Nordestgaard BG. Genetically low vitamin D concentrations and increased mortality: Mendelian randomisation analysis in three large cohorts. BMJ . 2014; 349: g6330. Google Scholar CrossRef Search ADS PubMed  26. Çolak Y, Afzal S, Nordestgaard BG, Lange P. Characteristics and prognosis of never-smokers and smokers with asthma in the Copenhagen general population study. A prospective cohort study. Am J Respir Crit Care Med . 2015; 192( 2): 172– 181. Google Scholar CrossRef Search ADS PubMed  27. Zacho J, Tybjaerg-Hansen A, Jensen JS, Grande P, Sillesen H, Nordestgaard BG. Genetically elevated C-reactive protein and ischemic vascular disease. N Engl J Med . 2008; 359( 18): 1897– 1908. Google Scholar CrossRef Search ADS PubMed  28. Winsløw UC, Rode L, Nordestgaard BG. High tobacco consumption lowers body weight: A Mendelian randomization study of the Copenhagen general population study. Int J Epidemiol . 2015; 44( 2): 540– 550. Google Scholar CrossRef Search ADS PubMed  29. Løkke A, Marott JL, Mortensen J, Nordestgaard BG, Dahl M, Lange P. New Danish reference values for spirometry. Clin Respir J . 2013; 7( 2): 153– 167. Google Scholar CrossRef Search ADS PubMed  30. Lawlor DA, Harbord RM, Sterne JA, Timpson N, Davey Smith G. Mendelian randomization: Using genes as instruments for making causal inferences in epidemiology. Stat Med . 2008; 27( 8): 1133– 1163. Google Scholar CrossRef Search ADS PubMed  31. Taylor AE, Morris RW, Fluharty ME, et al.   Stratification by smoking status reveals an association of CHRNA5-A3-B4 genotype with body mass index in never smokers. PLoS Genet . 2014; 10( 12): e1004799. Google Scholar CrossRef Search ADS PubMed  32. Rode L, Nordestgaard BG, Weischer M, Bojesen SE. Increased body mass index, elevated C-reactive protein, and short telomere length. J Clin Endocrinol Metab . 2014; 99( 9): E1671– E1675. Google Scholar CrossRef Search ADS PubMed  33. Varbo A, Benn M, Smith GD, Timpson NJ, Tybjaerg-Hansen A, Nordestgaard BG. Remnant cholesterol, low-density lipoprotein cholesterol, and blood pressure as mediators from obesity to ischemic heart disease. Circ Res . 2015; 116( 4): 665– 673. Google Scholar CrossRef Search ADS PubMed  34. Quint JK, Wedzicha JA. The neutrophil in chronic obstructive pulmonary disease. J Allergy Clin Immunol . 2007; 119( 5): 1065– 1071. Google Scholar CrossRef Search ADS PubMed  35. O’Donnell RA, Peebles C, Ward JA, et al.   Relationship between peripheral airway dysfunction, airway obstruction, and neutrophilic inflammation in COPD. Thorax . 2004; 59( 10): 837– 842. Google Scholar CrossRef Search ADS PubMed  36. Vestbo J, Edwards LD, Scanlon PD, et al.  ; ECLIPSE Investigators. Changes in forced expiratory volume in 1 second over time in COPD. N Engl J Med . 2011; 365( 13): 1184– 1192. Google Scholar CrossRef Search ADS PubMed  37. Dahl M, Vestbo J, Zacho J, Lange P, Tybjærg-Hansen A, Nordestgaard BG. C reactive protein and chronic obstructive pulmonary disease: A Mendelian randomisation approach. Thorax . 2011; 66( 3): 197– 204. Google Scholar CrossRef Search ADS PubMed  38. van Durme YM, Lahousse L, Verhamme KM, et al.   Mendelian randomization study of interleukin-6 in chronic obstructive pulmonary disease. Respiration . 2011; 82( 6): 530– 538. Google Scholar CrossRef Search ADS PubMed  39. Freitag DF, Butterworth AS, Willeit P; Interleukin 1 Genetics Consortium. Cardiometabolic effects of genetic upregulation of the interleukin 1 receptor antagonist: A Mendelian randomisation analysis. Lancet Diabetes Endocrinol. 2015;3:243–253. https://www.thelancet.com/pdfs/journals/landia/PIIS2213-8587(15)00034-0.pdf; https://www.ncbi.nlm.nih.gov/pubmed/?term=Cardiometabolic+effects+of+genetic+upregulation+of+the+interleukin+1+receptor+antagonist%3A+a+Mendelian+randomisation+analysis 40. Åsvold BO, Bjørngaard JH, Carslake D, et al.   Causal associations of tobacco smoking with cardiovascular risk factors: A Mendelian randomization analysis of the HUNT study in Norway. Int J Epidemiol . 2014; 43( 5): 1458– 1470. Google Scholar CrossRef Search ADS PubMed  41. Banack HR, Kaufman JS. The obesity paradox: Understanding the effect of obesity on mortality among individuals with cardiovascular disease. Prev Med . 2014; 62: 96– 102. Google Scholar CrossRef Search ADS PubMed  42. VanderWeele TJ, Tchetgen Tchetgen EJ, Cornelis M, Kraft P. Methodological challenges in Mendelian randomization. Epidemiology . 2014; 25( 3): 427– 435. Google Scholar CrossRef Search ADS PubMed  © The Author 2018. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nicotine and Tobacco Research Oxford University Press

Smoking, Systemic Inflammation, and Airflow Limitation: A Mendelian Randomization Analysis of 98 085 Individuals From the General Population

Loading next page...
 
/lp/ou_press/smoking-systemic-inflammation-and-airflow-limitation-a-mendelian-TX7xN2NfQp
Publisher
Oxford University Press
Copyright
© The Author 2018. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.
ISSN
1462-2203
eISSN
1469-994X
D.O.I.
10.1093/ntr/nty077
Publisher site
See Article on Publisher Site

Abstract

Abstract Introduction Smoking is associated with systemic and local inflammation in the lungs. Furthermore, in chronic obstructive pulmonary disease, which is often caused by smoking, there is often systemic inflammation that is linked to lung function impairment. However, the causal pathways linking smoking, systemic inflammation, and airflow limitation are still unknown. We tested whether higher tobacco consumption is associated with higher systemic inflammation, observationally and genetically and whether genetically higher systemic inflammation is associated with airflow limitation. Methods We included 98 085 individuals aged 20–100 years from the Copenhagen General Population Study; 36589 were former smokers and 16172 were current smokers. CHRNA3 rs1051730 genotype was used as a proxy for higher tobacco consumption and the IL6R rs2228145 genotype was used for higher systemic inflammation. Airflow limitation was defined as forced expiratory volume in 1 s (FEV1)/forced vital capacity (FVC) <70%. Results Difference in plasma level of C-reactive protein was 4.8% (95% CI = 4.4% to 5.2%) per 10 pack-year increase and 1.6% (95% CI = 0.4% to 2.8%) per T allele. Corresponding differences were 1.2% (95% CI = 1.1% to 1.3%) and 0.5% (95% CI = 0.3% to 0.8%) for fibrinogen, 1.2% (95% CI = 1.2% to 1.3%) and 0.7% (95% CI = 0.5% to 1.0%) for α1-antitrypsin, 2.0% (95% CI = 1.8% to 2.1%) and 0.7% (95% CI = 0.4% to 1.1%) for leukocytes, 1.9% (95% CI = 1.8% to 2.1%) and 0.8% (95% CI = 0.4% to 1.2%) for neutrophils, and 0.8% (95% CI = 0.7% to 1.0%) and 0.4% (95% CI = 0.1% to 0.7%) for thrombocytes. The differences in these levels were lower for former smokers compared with current smokers. The IL6R rs2228145 genotype was associated with higher plasma acute-phase reactants but not with airflow limitation. Compared with the C/C genotype, the odds ratio for airflow limitation was 0.95 (95% CI = 0.89 to 1.02) for A/C genotype and 0.94 (95% CI = 0.87 to 1.01) for A/A genotype. Conclusions Higher tobacco consumption is associated with higher systemic inflammation both genetically and observationally, whereas systemic inflammation was not associated with airflow limitation genetically. Implications The association between higher tobacco consumption and higher systemic inflammation may be causal, and the association is stronger among current smokers compared to former smokers, indicating that smoking cessation may reduce the effects of smoking on systemic inflammation. Systemic inflammation does not seem to be a causal driver in development of airflow limitation. These findings can help to understand the pathogenic effects of smoking and the interplay between smoking, systemic inflammation, and airflow limitation and hence development and progression of chronic obstructive pulmonary disease. Introduction Smoking is considered to be the most important risk factor for development of chronic obstructive pulmonary disease (COPD), characterized by persistent airflow limitation that is usually progressive, by inducing local inflammation in the lungs.1 However, smoking is not only associated with an increased degree of inflammation in the lungs but also in the systemic circulation.2 Furthermore, individuals with COPD often have systemic inflammation with high levels of C-reactive protein and fibrinogen3 and high levels of these biomarkers have been independently associated with lung function impairment.4–8 Despite the many studies on the importance of systemic inflammation in COPD,3,9 it is still unknown whether systemic inflammation plays a role in development and progression of the disease. Thus, the causal pathways linking smoking, systemic inflammation, and airflow limitation are still unknown. Observational studies are usually prone to confounding and reverse causation. Although randomized clinical trials are the gold standard in establishing causal relationships, some exposures such as smoking would be unethical or impractical to study using such a study design. An alternative approach is to use genetics in a Mendelian randomization analysis.10 By using the genetic variant CHRNA3 rs1051730, where the T allele is associated with higher tobacco consumption, it is possible to assess the association of higher tobacco consumption with systemic inflammation and lung function largely free of confounding and reverse causation.11–15 Due to the random distribution of alleles at conception, genetic variants should not be associated with potential confounders and, as genes are present at birth, genetic variants are not susceptible to reverse causation.10 However, as the CHRNA3 rs1051730 genotype may be associated with several tobacco consumption behaviors, instrumental variable estimates will not be calculated as the relevant exposure is difficult to define unambiguously. Thus, our approach will be akin to carrying out an intention-to-treat analysis in a randomized intervention trial and not estimating the effect of the intervention but rather of the randomization, and is the preferable approach when the assumptions for calculating instrumental variable estimates may be in doubt.16 Similarly, by using the genetic variant IL6R rs2228145, where the A allele is associated with an increased activity of the interleukin-6 receptor in hepatocytes and leucocytes, which causes increased plasma levels of C-reactive protein and other acute-phase reactants,17–24 it is possible to assess consequences of systemic inflammation by circumventing confounding and reverse causation. Since C-reactive protein and other acute-phase reactants are downstream products of inflammation, we used a more proximal regulator of inflammation, that is, interleukin-6 receptor, to investigate the association more generally rather than using individual markers alone. The purpose of the present study was to investigate the causal pathways linking smoking, systemic inflammation, and airflow limitation. Firstly, we tested whether higher tobacco consumption is associated with higher systemic inflammation, observationally and genetically (Supplementary Figure S1, Panel A). Secondly, we tested whether genetically higher systemic inflammation is associated with airflow limitation (Supplementary Figure S1, Panel B). For this purpose, we used the Copenhagen General Population Study, including 98 085 individuals, of whom 36589 were former smokers and 16172 current smokers. Methods Study Design and Subjects We recruited 98 085 individuals aged 20–100 years from the Copenhagen General Population Study, a population-based prospective cohort study.25–28 All individuals in Denmark are assigned a unique identification number at birth or immigration and registered in the national Danish Civil Registration System. Thus, participants were randomly selected from the national Danish Civil Registration System to reflect the adult white Danish population of Danish descent. All participants completed a comprehensive questionnaire, underwent a physical examination, and gave blood for biochemical and genetic analyses in the period of 2003–2013. Questionnaires were reviewed in detail at the day of attendance by a health care professional together with the participant. The study was approved by the Herlev and Gentofte Hospital and by a Danish ethical committee and conducted according to the Declaration of Helsinki (approval number: H-KF-01-144/01). Written informed consent was obtained from all participants. Smoking, Systemic Inflammation, and Airflow Limitation Smoking status was self-reported and defined as never, former, or current smokers. Tobacco consumption was defined as cumulative tobacco consumed through smoking and measured in pack-years based on information on duration of tobacco consumption and current amount of consumed tobacco: one pack-year was 20 cigarettes or equivalent smoked daily for 1 year. Nevertheless, tobacco consumption was also investigated in the form of daily amount of consumed tobacco in current smokers, thus also investigating immediate exposure. Systemic inflammation was assessed by acute-phase reactants (C-reactive protein [high-sensitivity assay], fibrinogen, and α1-antitrypsin) and cells of the immune system (leukocytes, neutrophils, and thrombocytes), and measured in the systemic circulation with standard hospital assays. Spirometry was performed with a Vitalograph (Maids Moreton, Buckinghamshire, UK) in the first 14625 participants and with an EasyOne Spirometer (ndd Medical Technologies, Zurich, Switzerland) in the remaining participants with measurements of pre-bronchodilatatory forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC).29 Predicted values were calculated using internally derived reference values based on a subsample of healthy asymptomatic never-smokers with age and height as covariates separately for men and women and separately for the two spirometers (Supplementary Figure S2).26 Presence of airflow limitation was defined according to (1) a fixed ratio defined as FEV1/FVC<70% and (2) lower limit of normal (LLN) defined as FEV1/FVC<LLN, calculated as the mean predicted value for age and sex multiplied by −1.645 standard error. Genotyping Genotyping was conducted blind to information on smoking status, tobacco consumption, degree of systemic inflammation, and presence of airflow limitation. The ABI PRISM 7900HT Sequence Detection System (Applied Biosystems Inc.) was used to genotype CHRNA3 rs1051730 with TaqMan assays. Call-rates were above 99.8% as we performed re-runs twice. IL6R rs2228145 was genotyped similarly to CHRNA3 rs1051730. Information on the CHRNA3 rs1051730 genotype was available in 98 085 individuals whereas the IL6R rs2228145 genotype was available in 81397 individuals. Potential Confounders Information on potential confounders was acquired from the questionnaires and physical examinations. Body mass index (BMI) was measured weight divided by measured height squared (kg/m2). Alcohol consumption was reported in units per week and converted to grams (1 unit=12 grams). Occupational exposure to dust and/or fumes, daily exposure to passive smoking, and asthma was self-reported. Physical activity in leisure time was reported according to hours per week and degree of activity. Education was based on years attending school. Income was reported as annual household income. Statistical Analyses STATA/SE 13.1 was used in the statistical analyses. Deviation from Hardy–Weinberg equilibrium was investigated using the chi-square tests to assess the frequency of allele distribution across the individual genotypes; deviation may suggest genotyping or population sampling errors.30 Cuzick’s non-parametric trend test was used to assess trend across ordered groups to investigate association with potential confounders. In addition, association with potential confounders were also investigated using linear regression for continuous covariates and logistic regression for dichotomous covariates. In observational analyses, we adjusted for potential confounders that could be associated with tobacco consumption, inflammatory biomarkers, and/or airflow limitation, that is, for age, sex, BMI, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. In the genetic analyses, we adjusted for age and sex. Furthermore, studies indicate that the CHRNA3 rs1051730 genotype may also be associated with BMI,11,28 and the association may be modified by smoking, where the genotype is associated with lower BMI in current smokers and higher BMI in never-smokers (Supplementary Figure S3).31 In addition, higher BMI is associated with higher degree of systemic inflammation.32,33 Thus, the effects on BMI could be considered to be a pleiotropic effect of the genotype, which could bias the results and, therefore, we performed the genetic analyses with and without additional adjustment for BMI. Also, the analyses stratified for smoking status can be used to assess potential pleiotropy, as different associations in ever-smokers versus never-smokers may hint to other biological effects of genetic variants in the CHRNA3. Association of tobacco consumption and the genotypes with each other and with the inflammatory biomarkers was investigated using multiple linear regressions. Levels of the inflammatory biomarkers were logarithmic transformed in order to approach a normal distribution. Association of tobacco consumption and the individual genotypes with airflow limitation was investigated using logistic regression. Association of airflow limitation with the inflammatory biomarkers was investigated using multiple linear regressions. Results Among 98 085 individuals from the Copenhagen General Population Study, 36589 were former smokers and 16172 were current smokers (Supplementary Figure S4). Tobacco consumption was associated with almost all of the potential confounders in former and current smokers (Table 1 and Supplementary Table S1), as well as in the entire study population (Supplementary Table S2). In contrast, the CHRNA3 rs1051730 genotype was not associated with most potential confounders in former and current smokers (Table 1), or in the entire study population (Supplementary Table S3). However, the CHRNA3 rs1051730 genotype was associated with age, being more often a current smoker, having a lower BMI as well as FEV1 and FVC as percentage of predicted values and FEV1/FVC in former and current smokers, as shown previously11,28; the associations with BMI and lung function measurements should be interpreted as downstream effects of tobacco consumption rather than potential confounders. Similarly, the IL6R rs2228145 genotype was not associated with any of the potential confounders in former and current smokers (Supplementary Table S4), or in the entire study population (Supplementary Table S5). There was no clear evidence of deviation from Hardy–Weinberg equilibrium; p = .99 for CHRNA3 rs1051730 genotype and p = .80 for IL6R rs2228145 genotype. Table 1. Baseline Characteristics of Former and Current Smokers in the Copenhagen General Population Study According to the CHRNA3 rs1051730 Genotype   CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)    CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)  BMI = body mass index; CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity. Data are summarized as median with the 25th and 75th percentiles, or %, or effect sizes with 95% confidence intervals. aWhen p for trend is adjusted for 14 number of individual trend analyses according to the Bonferroni method, p = .05 is equivalent to p = .05/14 = .004. bEffect sizes report beta coefficients on linear regression for continuous covariates and on logistic regression for dichotomous covariates. cThis p value indicates that the number of current smokers was higher among those with higher tobacco consumption (please see Supplementary Table S1). View Large Table 1. Baseline Characteristics of Former and Current Smokers in the Copenhagen General Population Study According to the CHRNA3 rs1051730 Genotype   CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)    CHRNA3 rs1051730 genotype  Tobacco consumption    C/C n = 24073  C/T n = 23117  T/T n = 5571  p for trenda  Per allele effect sizeb (95% CI)  p for trenda  Per pack-year effect sizeb (95% CI)  Men (%)  48  48  49  0.72  0.005 (−0.021; 0.031)  <1 × 10−300  0.026 (0.025; 0.027)  Age (years)  59 (50–68)  59 (50–68)  59 (50–68)  0.003  −0.216 (−0.378; −0.053)  <1 x 10–300  0.147 (0.142; 0.152)  BMI (kg/m2)  26 (23–29)  26 (23–29)  26 (23–28)  4 × 10–5  −0.113 (−0.168; −0.059)  <1 × 10–300  0.027 (0.025; 0.029)  Alcohol consumption (grams/week)  120 (48–204)  120 (48–204)  120 (48–204)  0.51  0.771 (−0.995; 2.537)  <1 × 10–300  1.266 (1.211; 1.322)  Current smokers (%)  30  31  32  2 × 10–7  0.073 (0.046; 0.101)  <1 × 10−300c  0.028 (0.027; 0.029)  Occupational exposure to dust and/or fumes (%)  13  13  13  0.30  0.020 (−0.018; 0.059)  <1 × 10−300  0.021 (0.020; 0.022)  Daily exposure to passive smoking (%)  19  19  19  0.18  0.023 (−0.010; 0.056)  9 × 10−147  0.011 (0.011; 0.012)  Self-reported asthma (%)  6  6  7  0.004  0.078 (0.025; 0.132)  0.04  0.002 (−0.00005; 0.003)  Low leisure-time physical activity (%)  7  7  8  0.009  0.066 (0.017; 0.116)  3 × 10−107  0.014 (0.013; 0.015)  Low education (%)  29  29  29  0.33  −0.014 (−0.043; 0.014)  <1 × 10−300  0.024 (0.023; 0.025)  Low annual household income (%)  14  14  14  0.27  −0.021 (−0.058; 0.016)  8 × 10−177  0.016 (0.015; 0.017)  FEV1 % of predicted  94 (84-104)  94 (83-104)  93 (82-103)  3 × 10−14  −0.905 (−1.126; −0.685)  <1 × 10−300  −0.256 (−0.263; −0.249)  FVC % of predicted  97 (87-106)  96 (86-106)  96 (86-106)  3 × 10−6  −0.494 (−0.692; −0.296)  <1 × 10−300  −0.165 (−0.171; −0.158)  FEV1/FVC (%)  77 (72-81)  76 (71-81)  76 (70-81)  1 × 10−6  −0.341 (−0.449; −0.232)  <1 × 10−300  −0.104 (−0.108; −0.101)  BMI = body mass index; CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity. Data are summarized as median with the 25th and 75th percentiles, or %, or effect sizes with 95% confidence intervals. aWhen p for trend is adjusted for 14 number of individual trend analyses according to the Bonferroni method, p = .05 is equivalent to p = .05/14 = .004. bEffect sizes report beta coefficients on linear regression for continuous covariates and on logistic regression for dichotomous covariates. cThis p value indicates that the number of current smokers was higher among those with higher tobacco consumption (please see Supplementary Table S1). View Large Association of the CHRNA3 rs1051730 Genotype With Tobacco Consumption Collectively, the results show that the CHRNA3 rs1051730 genotype was associated with higher risk of being current smokers among ever smokers and in the general population (Table 1 and Supplementary Table S1), higher cumulative tobacco consumption in former and current smokers, and higher daily tobacco consumption among current smokers (Supplementary Figure S5). Cumulative tobacco consumption in former and current smokers was 20.3 pack-years (95% confidence interval [CI] = 20.1 to 20.6 pack-years) for the C/C genotype, 21.8 pack-years (95% CI = 21.6 to 22.1 pack-years) for the C/T genotype, and 23.2 pack-years (95% CI = 22.7 to 23.7 pack-years) for the T/T genotype. In addition, daily amount of smoked tobacco in current smokers was 14.8 g/day (95% CI = 14.6 to 15.1 g/day) for the C/C genotype, 15.7 g/day (95% CI = 15.5 to 15.9 g/day) for the C/T genotype, and 16.8 g/day (95% CI = 16.3 to 17.2 g/day) for the T/T genotype. Association of Tobacco Consumption and CHRNA3 rs1051730 With Systemic Inflammation Higher tobacco consumption in former and current smokers was associated with higher levels of acute-phase reactants, including C-reactive protein, fibrinogen, and α1-antitrypsin, and cellular markers of inflammation, including leukocytes, neutrophils, and thrombocytes, in the systemic circulation with a dose-dependent relationship (Figure 1 and Supplementary Figures S6 and S7). In former and current smokers, the multivariable adjusted percent difference in the level of inflammatory biomarkers per 10 pack-year increase was 4.8% (95% CI = 4.4% to 5.2%) for C-reactive protein, 1.2% (95% CI = 1.1% to 1.3%) for fibrinogen, 1.2% (95% CI = 1.2% to 1.3%) for α1-antitrypsin, 2.0% (95% CI = 1.8% to 2.1%) for leukocytes, 1.9% (95% CI = 1.8% to 2.1%) for neutrophils, and 0.8% (95% CI = 0.7% to 1.0%) for thrombocytes (Figure 1). When former and current smokers were analyzed separately, the percent differences in the level of inflammatory biomarkers for former smokers were lower compared with current smokers. Interaction analyses between former and current smokers suggested that these associations were strongest among current smokers. Figure 1. View largeDownload slide Association of tobacco consumption with systemic inflammation. Levels of inflammatory biomarkers are summarized as percentage difference with 95% confidence interval. Models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. p values were obtained from the likelihood ratio tests. CI = confidence interval. Figure 1. View largeDownload slide Association of tobacco consumption with systemic inflammation. Levels of inflammatory biomarkers are summarized as percentage difference with 95% confidence interval. Models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. p values were obtained from the likelihood ratio tests. CI = confidence interval. Similarly, the CHRNA3 rs1051730 genotype in former and current smokers was associated with higher levels of acute-phase reactants and cellular markers of inflammation with a dose-dependent relationship (Figure 2 and Supplementary Figure S6). After adjustment for BMI, the association seemed slightly stronger, especially for current smokers (Figure 2). In former and current smokers, the age and sex adjusted percent difference in the level of inflammatory biomarkers per T allele was 1.6% (95% CI = 0.4% to 2.8%) for C-reactive protein, 0.5% (95% CI = 0.3% to 0.8%) for fibrinogen, 0.7% (95% CI = 0.5% to 1.0%) for α1-antitrypsin, 0.7% (95% CI = 0.4% to 1.1%) for leukocytes, 0.8% (95% CI = 0.4% to 1.2%) for neutrophils, and 0.4% (95% CI = 0.1% to 0.7%) for thrombocytes (Figure 2). With regard to never-smokers, the CHRNA3 rs1051730 genotype was positively associated with C-reactive protein and negatively associated with leukocytes and neutrophils; although, these associations were weak and no associations with other inflammatory biomarkers could be observed. Interaction analyses between never and former and current smokers suggested that these associations were strongest among former and current smokers and that with regard to C-reactive protein and fibrinogen, there seems to be no difference between the different smoking categories. When former and current smokers were analyzed separately, the percent differences in the levels of inflammatory biomarkers were lower for former smokers compared to current smokers. Figure 2. View largeDownload slide Association of the CHRNA3 rs1051730 genotype with systemic inflammation. Levels of inflammatory biomarkers are summarized as % difference with 95% confidence interval. Models were adjusted for age and sex and when indicated additionally for body mass index. P values were obtained from the likelihood ratio tests. CI = confidence interval. Figure 2. View largeDownload slide Association of the CHRNA3 rs1051730 genotype with systemic inflammation. Levels of inflammatory biomarkers are summarized as % difference with 95% confidence interval. Models were adjusted for age and sex and when indicated additionally for body mass index. P values were obtained from the likelihood ratio tests. CI = confidence interval. Association of IL6R rs2228145 With Systemic Inflammation and Tobacco Consumption Collectively, the results show that the IL6R rs2228145 genotype was associated with higher levels of acute-phase reactants but unclear with regard to levels of cellular markers of inflammation in the systemic circulation (Supplementary Figures S8 and S9). In former and current smokers, the age and sex adjusted percent difference in the level of C-reactive protein was 8.3% (95% CI = 5.6% to 11%) for the A/C genotype and 15% (95% CI = 12% to 18%) for the A/A genotype. Corresponding levels were 1.2% (95% CI = 0.7% to 1.8%) and 1.8% (95% CI = 1.2% to 2.4%) for fibrinogen, 0.5% (95% CI = −0.05% to 1.0%) and 0.6% (95% CI = 0.03% to 1.2%) for α1-antitrypsin, −0.2% (95% CI = −0.9% to 0.4%) and 0.3% (95% CI = −0.4% to 0.9%) for leukocytes, −0.02% (95% CI = −0.8% to 0.8%) and 0.6% (95% CI = −0.2% to 1.5%) for neutrophils, and −0.4% (95% CI = −1.1% to 0.2%) and −1.0% (95% CI = −1.6% to −0.4%) for thrombocytes (Supplementary Figure S8). As seen, the IL6R rs2228145 genotype was associated with lower levels of thrombocytes. These results suggest that the IL6R rs2228145 genotype can be used as a genetic instrument for the acute-phase reactants but not for the cellular markers of inflammation. The IL6R rs2228145 genotype was not associated with any tobacco consumption phenotype (Supplementary Tables S4 and S5). Association of Tobacco Consumption, CHRNA3 rs1051730, and IL6R rs2228145 With Airflow Limitation Higher tobacco consumption and the CHRNA3 rs1051730 genotype were associated with a higher risk of airflow limitation in former and current smokers regardless of the chosen criterion for airflow limitation, with a dose-dependent relationship (Figure 3). Compared to <15 pack-years, the odds ratio (OR) for airflow limitation according to FEV1/FVC < 70% was 1.63 (95% CI = 1.53 to 1.74) for 15–29.9 pack-years and 2.61 (95% CI = 2.46 to 2.78) for ≥30 pack-years. Compared with the C/C genotype, the OR for airflow limitation according to FEV1/FVC < 70% was 1.10 (95% CI = 1.05 to 1.16) for C/T genotype and 1.27 (95% CI = 1.18 to 1.37) for T/T genotype. The IL6R rs2228145 genotype was not associated with airflow limitation regardless of the chosen criterion for airflow limitation (Figure 3 and Supplementary Figure S10). Compared to the C/C genotype, the OR for airflow limitation according to FEV1/FVC < 70% was 0.95 (95% CI = 0.89 to 1.02) for A/C genotype and 0.94 (95% CI = 0.87 to 1.01) for A/A genotype (Figure 3). Figure 3. View largeDownload slide Association of tobacco consumption, CHRNA3 rs1051730, and IL6R rs2228145 with airflow limitation in former and current smokers. In the observational analyses, models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. In the genetic analyses, models were adjusted for age and sex. p values were obtained from the Wald tests from the logistic regression models. CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity; OR = odds ratio; LLN = lower limit of normal. Figure 3. View largeDownload slide Association of tobacco consumption, CHRNA3 rs1051730, and IL6R rs2228145 with airflow limitation in former and current smokers. In the observational analyses, models were adjusted for age, sex, body mass index, alcohol consumption, smoking status, occupational exposure to dust and/or fumes, daily exposure to passive smoking, self-reported asthma, physical activity in leisure time, education, and annual household income. In the genetic analyses, models were adjusted for age and sex. p values were obtained from the Wald tests from the logistic regression models. CI = confidence interval; FEV1 = forced expiratory volume in 1 s; FVC = forced vital capacity; OR = odds ratio; LLN = lower limit of normal. Sensitivity Analyses The CHRNA3 rs1051730 genotype was, as expected, not associated with airflow limitation in never-smokers (Supplementary Figure S11). Among all individuals, there was an interaction of the CHRNA3 rs1051730 genotype with smoking status on risk of airflow limitation with highest risk estimates for current smokers (p for interaction = 1 × 10−4). Airflow limitation was associated with higher levels of acute-phase reactants and cellular markers of inflammation regardless of the chosen criterion for airflow limitation (Supplementary Figure S12). Discussion Using a large sample from the general population, we found that higher tobacco consumption was associated with higher systemic inflammation in observational and genetic analyses, whereas systemic inflammation was not associated with airflow limitation in genetic analyses. Our approach supports the notion that the association between higher tobacco consumption and higher systemic inflammation may be causal, and the association is stronger among current smokers compared to former smokers, indicating that smoking cessation may reduce the effects of smoking on systemic inflammation. Furthermore, systemic inflammation does not seem to be a causal driver in development of airflow limitation. These are novel findings that can help to understand the pathogenic effects of smoking and the interplay between smoking, systemic inflammation, and airflow limitation and hence development and progression of COPD. There may be several potential mechanisms involved in the association between smoking, systemic inflammation, and airflow limitation. Although smoking is associated with local inflammation in the lungs34 and local inflammation in the lungs is associated with increased lung function decline and disease severity,35,36 the mechanisms behind systemic inflammation in smokers and in COPD patients are unclear. Our results in combination with previous studies show that smoking induces systemic inflammation, whereas variation in the IL6R and CRP genes, which genetically determined higher systemic inflammation or markers of such, is not associated with an increased risk of COPD in previous Mendelian randomization studies of our and other populations.37,38 This supports the notion that systemic inflammation in individuals with COPD is a marker of smoking exposure and possibly presence of more severe disease rather than a driving force of disease progression. However, our results also point to that a direct pathway may exist from smoking to diseases where systemic inflammation may be a contributing cause, for example, coronary heart disease.20,39 In a Mendelian randomization analysis, where the causal association between tobacco consumption with cardiovascular risk factors was investigated, the association with C-reactive protein was not statistically significant; however, the observational and genetic estimates were congruent and in the same direction as observed in the present study but C-reactive protein measurements were only available in a subsample.40 Nevertheless, further genetic studies investigating the causality of smoking through systemic inflammation for other diseases are still needed. Interestingly, the IL6R rs2228145 genotype was associated with higher levels of acute-phase reactants but not clearly associated with cellular markers of inflammation. A potential explanation may be that the IL6R rs2228145 genotype primarily affects the acute-phase reactants while the effects on cellular markers of inflammation are weak. Thus, though the study was well powered to investigate the association of the acute-phase reactants with airflow limitation, the cellular markers deserves further investigation. Potential limitations in genetic studies include population stratification and genetic pleiotropy. However, as we had an ethnically homogenous population, the complicating effects of population stratification are likely to have been avoided. Since genotype distributions did not appear to differ from Hardy–Weinberg equilibrium, we have also likely avoided genotyping and population sampling errors. It is also important to note that stratification on smoking status may introduce the possibility of collider stratification bias.28,41,42 Yet, stratification by smoking status also provides an important test of the no pleiotropy assumption of Mendelian randomization studies; we observed no clear association between genotype and outcomes in never-smokers, supporting the notion that pleiotropy does not seem to be biasing our results; however, the CHRNA3 rs1051730 genotype was marginally associated with higher levels of C-reactive protein and lower levels of leukocytes and neutrophils in never-smokers, something that clearly should be investigated further in future studies. Former and current smokers were also extensively investigated according to the CHRNA3 rs1051730 genotype with regard to potential confounders without observed differences. While we were able to show that smoking induces systemic inflammation and that systemic inflammation may not induce airflow limitation, an important limitation is the lack of a specific genetic variant for development of airflow limitation. It could very well be that airflow limitation in itself could induce systemic inflammation, but to investigate this hypothesis we would require genetic variants associated with airflow limitation independently of smoking behaviors and systemic inflammation. Unfortunately, such an investigation is not feasible in the present study. Lastly, both observational and Mendelian randomization studies could potentially be influenced by survival bias, for example, heavier smokers are likely to die earlier and therefore not be included in the analysis. Strengths of the present study are use of a large sample with 98 085 individuals from the general population, including a large sample of smokers. Furthermore, we had extensive information on genotypes, inflammatory biomarkers in the systemic circulation, lung function, and several tobacco consumption variables. In conclusion, higher tobacco consumption is associated with higher systemic inflammation both genetically and observationally, while systemic inflammation was not associated with airflow limitation genetically. These results are consistent with the notion that the association between higher tobacco consumption and higher systemic inflammation may be causal, whereas systemic inflammation does not seem to be a causal driver in development of airflow limitation. Furthermore, the association of smoking to systemic inflammation was stronger among current smokers compared to former smokers, indicating that smoking cessation may reduce the effects of smoking on systemic inflammation. Supplementary Material Supplementary data are available at Nicotine & Tobacco Research online. Funding This work was supported by the Lundbeck Foundation, Danish Lung Association, the Danish Cancer Society, Herlev and Gentofte Hospital, Copenhagen University Hospital, and University of Copenhagen. The funders had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication. Declaration of Interests YÇ reports personal fees from Boehringer Ingelheim and AstraZeneca outside of the submitted work. PL reports grants from AstraZeneca and GlaxoSmithKline and personal fees from Boehringer Ingelheim, AstraZeneca, Novartis, and GlaxoSmithKline outside of the submitted work. SA and BGN have nothing to disclose. References 1. Global Strategy for the Diagnosis, Management, and Prevention of Chronic Obstructive Lung Disease. Global Initiative for Chronic Obstructive Lung Disease (GOLD). http://www.goldcopd.org/gold-2017-global-strategy-diagnosis-management-prevention-copd/. Accessed February 23, 2017. 2. Yanbaeva DG, Dentener MA, Creutzberg EC, Wesseling G, Wouters EF. Systemic effects of smoking. Chest . 2007; 131( 5): 1557– 1566. Google Scholar CrossRef Search ADS PubMed  3. Gan WQ, Man SF, Senthilselvan A, Sin DD. Association between chronic obstructive pulmonary disease and systemic inflammation: a systematic review and a meta-analysis. Thorax . 2004; 59( 7): 574– 580. Google Scholar CrossRef Search ADS PubMed  4. Aronson D, Roterman I, Yigla M, et al.   Inverse association between pulmonary function and C-reactive protein in apparently healthy subjects. Am J Respir Crit Care Med . 2006; 174( 6): 626– 632. Google Scholar CrossRef Search ADS PubMed  5. Dahl M, Tybjaerg-Hansen A, Vestbo J, Lange P, Nordestgaard BG. Elevated plasma fibrinogen associated with reduced pulmonary function and increased risk of chronic obstructive pulmonary disease. Am J Respir Crit Care Med . 2001; 164( 6): 1008– 1011. Google Scholar CrossRef Search ADS PubMed  6. Donaldson GC, Seemungal TA, Patel IS, et al.   Airway and systemic inflammation and decline in lung function in patients with COPD. Chest . 2005; 128( 4): 1995– 2004. Google Scholar CrossRef Search ADS PubMed  7. Shaaban R, Kony S, Driss F, et al.   Change in C-reactive protein levels and FEV1 decline: A longitudinal population-based study. Respir Med . 2006; 100( 12): 2112– 2120. Google Scholar CrossRef Search ADS PubMed  8. Yende S, Waterer GW, Tolley EA, et al.   Inflammatory markers are associated with ventilatory limitation and muscle dysfunction in obstructive lung disease in well functioning elderly subjects. Thorax . 2006; 61( 1): 10– 16. Google Scholar CrossRef Search ADS PubMed  9. Yanbaeva DG, Dentener MA, Creutzberg EC, Wouters EF. Systemic inflammation in COPD: Is genetic susceptibility a key factor? COPD . 2006; 3( 1): 51– 61. Google Scholar CrossRef Search ADS PubMed  10. Smith GD, Ebrahim S. ‘Mendelian randomization’: Can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol . 2003; 32( 1): 1– 22. Google Scholar CrossRef Search ADS PubMed  11. Freathy RM, Kazeem GR, Morris RW, et al.   Genetic variation at CHRNA5-CHRNA3-CHRNB4 interacts with smoking status to influence body mass index. Int J Epidemiol . 2011; 40( 6): 1617– 1628. Google Scholar CrossRef Search ADS PubMed  12. Munafò MR, Timofeeva MN, Morris RW, et al.  ; EPIC Study Group. Association between genetic variants on chromosome 15q25 locus and objective measures of tobacco exposure. J Natl Cancer Inst . 2012; 104( 10): 740– 748. Google Scholar CrossRef Search ADS PubMed  13. Kaur-Knudsen D, Nordestgaard BG, Bojesen SE. CHRNA3 genotype, nicotine dependence, lung function and disease in the general population. Eur Respir J . 2012; 40( 6): 1538– 1544. Google Scholar CrossRef Search ADS PubMed  14. Thorgeirsson TE, Geller F, Sulem P, et al.   A variant associated with nicotine dependence, lung cancer and peripheral arterial disease. Nature . 2008; 452( 7187): 638– 642. Google Scholar CrossRef Search ADS PubMed  15. Thorgeirsson TE, Gudbjartsson DF, Surakka I, et al.  ; ENGAGE Consortium. Sequence variants at CHRNB3-CHRNA6 and CYP2A6 affect smoking behavior. Nat Genet . 2010; 42( 5): 448– 453. Google Scholar CrossRef Search ADS PubMed  16. Nitsch D, Molokhia M, Smeeth L, DeStavola BL, Whittaker JC, Leon DA. Limits to causal inference based on Mendelian randomization: A comparison with randomized controlled trials. Am J Epidemiol . 2006; 163( 5): 397– 403. Google Scholar CrossRef Search ADS PubMed  17. Danik JS, Paré G, Chasman DI, et al.   Novel loci, including those related to Crohn disease, psoriasis, and inflammation, identified in a genome-wide association study of fibrinogen in 17 686 women: The Women’s Genome Health Study. Circ Cardiovasc Genet . 2009; 2( 2): 134– 141. Google Scholar CrossRef Search ADS PubMed  18. Müllberg J, Oberthür W, Lottspeich F, et al.   The soluble human IL-6 receptor. Mutational characterization of the proteolytic cleavage site. J Immunol . 1994; 152( 10): 4958– 4968. Google Scholar PubMed  19. Swerdlow DI, Holmes MV, Kuchenbaecker KB, et al.   The interleukin-6 receptor as a target for prevention of coronary heart disease: A mendelian randomisation analysis. Lancet . 2012; 379(9822): 1214– 1224. 20. Sarwar N, Butterworth AS, Freitag DF, et al.   Interleukin-6 receptor pathways in coronary heart disease: A collaborative meta-analysis of 82 studies. Lancet . 2012; 379(9822): 1205– 1213. 21. Boekholdt SM, Stroes ES. The interleukin-6 pathway and atherosclerosis. Lancet . 2012; 379( 9822): 1176– 1178. Google Scholar CrossRef Search ADS PubMed  22. Elliott P, Chambers JC, Zhang W, et al.   Genetic loci associated with C-reactive protein levels and risk of coronary heart disease. JAMA . 2009; 302( 1): 37– 48. Google Scholar CrossRef Search ADS PubMed  23. Galicia JC, Tai H, Komatsu Y, Shimada Y, Akazawa K, Yoshie H. Polymorphisms in the IL-6 receptor (IL-6R) gene: Strong evidence that serum levels of soluble IL-6R are genetically influenced. Genes Immun . 2004; 5( 6): 513– 516. Google Scholar CrossRef Search ADS PubMed  24. Reich D, Patterson N, Ramesh V, et al.  ; Health, Aging and Body Composition (Health ABC) Study. Admixture mapping of an allele affecting interleukin 6 soluble receptor and interleukin 6 levels. Am J Hum Genet . 2007; 80( 4): 716– 726. Google Scholar CrossRef Search ADS PubMed  25. Afzal S, Brøndum-Jacobsen P, Bojesen SE, Nordestgaard BG. Genetically low vitamin D concentrations and increased mortality: Mendelian randomisation analysis in three large cohorts. BMJ . 2014; 349: g6330. Google Scholar CrossRef Search ADS PubMed  26. Çolak Y, Afzal S, Nordestgaard BG, Lange P. Characteristics and prognosis of never-smokers and smokers with asthma in the Copenhagen general population study. A prospective cohort study. Am J Respir Crit Care Med . 2015; 192( 2): 172– 181. Google Scholar CrossRef Search ADS PubMed  27. Zacho J, Tybjaerg-Hansen A, Jensen JS, Grande P, Sillesen H, Nordestgaard BG. Genetically elevated C-reactive protein and ischemic vascular disease. N Engl J Med . 2008; 359( 18): 1897– 1908. Google Scholar CrossRef Search ADS PubMed  28. Winsløw UC, Rode L, Nordestgaard BG. High tobacco consumption lowers body weight: A Mendelian randomization study of the Copenhagen general population study. Int J Epidemiol . 2015; 44( 2): 540– 550. Google Scholar CrossRef Search ADS PubMed  29. Løkke A, Marott JL, Mortensen J, Nordestgaard BG, Dahl M, Lange P. New Danish reference values for spirometry. Clin Respir J . 2013; 7( 2): 153– 167. Google Scholar CrossRef Search ADS PubMed  30. Lawlor DA, Harbord RM, Sterne JA, Timpson N, Davey Smith G. Mendelian randomization: Using genes as instruments for making causal inferences in epidemiology. Stat Med . 2008; 27( 8): 1133– 1163. Google Scholar CrossRef Search ADS PubMed  31. Taylor AE, Morris RW, Fluharty ME, et al.   Stratification by smoking status reveals an association of CHRNA5-A3-B4 genotype with body mass index in never smokers. PLoS Genet . 2014; 10( 12): e1004799. Google Scholar CrossRef Search ADS PubMed  32. Rode L, Nordestgaard BG, Weischer M, Bojesen SE. Increased body mass index, elevated C-reactive protein, and short telomere length. J Clin Endocrinol Metab . 2014; 99( 9): E1671– E1675. Google Scholar CrossRef Search ADS PubMed  33. Varbo A, Benn M, Smith GD, Timpson NJ, Tybjaerg-Hansen A, Nordestgaard BG. Remnant cholesterol, low-density lipoprotein cholesterol, and blood pressure as mediators from obesity to ischemic heart disease. Circ Res . 2015; 116( 4): 665– 673. Google Scholar CrossRef Search ADS PubMed  34. Quint JK, Wedzicha JA. The neutrophil in chronic obstructive pulmonary disease. J Allergy Clin Immunol . 2007; 119( 5): 1065– 1071. Google Scholar CrossRef Search ADS PubMed  35. O’Donnell RA, Peebles C, Ward JA, et al.   Relationship between peripheral airway dysfunction, airway obstruction, and neutrophilic inflammation in COPD. Thorax . 2004; 59( 10): 837– 842. Google Scholar CrossRef Search ADS PubMed  36. Vestbo J, Edwards LD, Scanlon PD, et al.  ; ECLIPSE Investigators. Changes in forced expiratory volume in 1 second over time in COPD. N Engl J Med . 2011; 365( 13): 1184– 1192. Google Scholar CrossRef Search ADS PubMed  37. Dahl M, Vestbo J, Zacho J, Lange P, Tybjærg-Hansen A, Nordestgaard BG. C reactive protein and chronic obstructive pulmonary disease: A Mendelian randomisation approach. Thorax . 2011; 66( 3): 197– 204. Google Scholar CrossRef Search ADS PubMed  38. van Durme YM, Lahousse L, Verhamme KM, et al.   Mendelian randomization study of interleukin-6 in chronic obstructive pulmonary disease. Respiration . 2011; 82( 6): 530– 538. Google Scholar CrossRef Search ADS PubMed  39. Freitag DF, Butterworth AS, Willeit P; Interleukin 1 Genetics Consortium. Cardiometabolic effects of genetic upregulation of the interleukin 1 receptor antagonist: A Mendelian randomisation analysis. Lancet Diabetes Endocrinol. 2015;3:243–253. https://www.thelancet.com/pdfs/journals/landia/PIIS2213-8587(15)00034-0.pdf; https://www.ncbi.nlm.nih.gov/pubmed/?term=Cardiometabolic+effects+of+genetic+upregulation+of+the+interleukin+1+receptor+antagonist%3A+a+Mendelian+randomisation+analysis 40. Åsvold BO, Bjørngaard JH, Carslake D, et al.   Causal associations of tobacco smoking with cardiovascular risk factors: A Mendelian randomization analysis of the HUNT study in Norway. Int J Epidemiol . 2014; 43( 5): 1458– 1470. Google Scholar CrossRef Search ADS PubMed  41. Banack HR, Kaufman JS. The obesity paradox: Understanding the effect of obesity on mortality among individuals with cardiovascular disease. Prev Med . 2014; 62: 96– 102. Google Scholar CrossRef Search ADS PubMed  42. VanderWeele TJ, Tchetgen Tchetgen EJ, Cornelis M, Kraft P. Methodological challenges in Mendelian randomization. Epidemiology . 2014; 25( 3): 427– 435. Google Scholar CrossRef Search ADS PubMed  © The Author 2018. Published by Oxford University Press on behalf of the Society for Research on Nicotine and Tobacco. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Nicotine and Tobacco ResearchOxford University Press

Published: Apr 24, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off