Effects of chronic obstructive pulmonary disease (COPD) treatments on cardiovascular (CV) risk
| Author/study | Design and participants | Duration | Population | Treatment(s) | Primary outcome | Key results | Other findings |
| β2-agonists | |||||||
| Suissa (2003) [71] | Population-based, nested case–control (n=12 090) | Follow-up to 1999 | Newly diagnosed patients with COPD aged >55 years (identified from the Saskatchewan Health Services databases, 1980–1997) | SABAs | Cases of acute MI occurring during cohort follow-up | 1127 evaluable patients identified with fatal or non-fatal acute MI SABA use not associated with increased risk of acute MI (rate ratio for any use: 1.06, 95% CI 0.92–1.23) | No significant increase in risk when analysis is restricted to patients with CV risk factors |
| Salpeter (2004) [72] | Meta-analysis of randomised and placebo-controlled trials 13 RCTs of single-dose treatment (n=232) 20 trials of longer treatment duration (n=6623) | Mean of 4.7 months (range: 3 days to 1 year) | Patients with obstructive airway disease (COPD or asthma) Mean age 56.6 years (single-dose) and 52.2 years (longer treatment studies) | β2-agonists versus placebo | Short-term effect on heart rate and potassium concentrations Long-term effect on adverse CV events | Single-dose β2-agonist versus placebo increased heart rate by 9.12 bpm and reduced potassium concentration by 0.36 mmol·L−1 Longer β2-agonist treatment significantly increased risk for CV events (relative risk 2.54, 95% CI 1.59–4.05) | The relative risk for sinus tachycardia was 3.06 (95% CI 1.70–5.50) For all other events, relative risk was 1.66 (95% CI 0.76–3.6) |
| Cazzola (2007) [27] | Randomised, double blind, double dummy (n=20) | COPD and sPAP >20 mmHg (rest) | Salmeterol 50 µg Formoterol 12 µg | Acute haemodynamic response | Mean sPAP significantly decreased versus baseline at 15, 30 and 60 min post-inhalation (p<0.05) | No correlation between maximum increase in FEV1 and maximum decrease in sPAP after inhalation of salmeterol or formoterol | |
| Santus (2015) [73] | Randomised, double blind, placebo controlled, crossover (n=40) | COPD (FEV1 ≤70% pred; RV ≥135% pred) Absence of CV comorbidities Age 50–85 years | Indacaterol 150 µg versus placebo (1:1) | Effect of reduction of right-ventricular and FRC on right heart systolic/diastolic functional indices | Significant improvements in right-ventricular compliance/cardiac performance with indacaterol versus placebo (p≤0.05 after 180 min treatment) as follows: 1) TAPSE: 0.41 mm versus 0.02 mm 2) DT-TR: 11.9 ms versus 3.8 ms 3) ≤Heart rate: −2 bpm versus 0.6 bpm | Indacaterol associated with significant (p<0.05) increases versus placebo in FEV1, VC and IC | |
| Muscarinic antagonists | |||||||
| Kesten 2006 [74] | Pooled safety analysis of 19 randomised, double blind, placebo controlled trials (including two asthma trials) (n=7819) | COPD diagnosis with airflow limitation Age ≥40 years Smoking history ≥10 pack-years | Tiotropium 18 µg once daily (n=4435) Placebo (n=3384) | AEs occurring during the study | Tiotropium not associated with increased risk of serious cardiac events versus placebo. Data as follows: 1) CV mortality (relative risk 0.57, 95% CI 0.26–1.26) 2) Cardiac arrest (relative risk 0.90, 95% CI 0.26–3.15) 3) MI (relative risk 0.74, 95% CI 0.26–2.07) | Relative risk of tachycardia with tiotropium versus placebo: 1.68 (95% CI 0.69–4.11) | |
| Travers (2007) [75] | Randomised, double blind, placebo controlled, crossover (n=18) | 7–10 day treatment period + 35 day washout period | COPD (FEV1 ≤65% pred; FRC ≥120% pred; modified baseline dyspnoea index score ≤6) | Tiotropium 18 µg once daily versus placebo | Effect of tiotropium on CV response to exercise | Tiotropium improved cardiac function during exercise versus placebo: 1) Significantly reduced heart rate (105 bpm versus 112 bpm, p<0.05) 2) Higher O2 pulse (10.9 mL·beat−1 versus 10.1 mL·beat−1, p<0.05) 3) Lower SBP (148 mmHg versus 156 mmHg, p<0.05) | Tiotropium significantly improved measures of dynamic hyperinflation during exercise: IRV was significantly greater with tiotropium versus placebo (0.60 L versus 0.44 L, p<0.05) |
| Singh (2008) [76] | Systematic review and meta-analysis (17 RCTs; 12 on tiotropium, 5 on ipratropium) (n=14 783) | COPD of any severity | Tiotropium or ipratropium (n=7472) versus control (placebo/active control) (n=7311) | Composite of non-fatal MI, non-fatal stroke and CV death | Tiotropium or ipratropium significantly increased risk of CV death, MI or stroke versus control (1.8% versus 1.2%; relative risk 1.58 (95% CI 1.21–2.06), p<0.001) | Tiotropium or ipratropium did not significantly increase risk of all-cause mortality versus control (2.0% versus 1.6%; relative risk 1.26 (95% CI 0.99–1.61), p=0.06) | |
| Rodrigo (2009) [77] | Systematic review and meta-analysis (19 RCTs) (n=18111) | 7 trials (28–48 months) 12 trials (8 weeks–6 months) | Patients with COPD (average baseline FEV1: 41% pred normal) Mean age 64.8 years | Tiotropium versus placebo (n=15#), SFC (n=2#), salmeterol (n=1#), salmeterol/placebo (n=1#) | Composite of MACE, CV mortality and non-fatal MI or stroke | No difference in incidence of MACE versus control groups (relative risk 0.96, 95% CI 0.82–1.12) | Compared with control groups, tiotropium did not significantly increase risk of: 1) CV death (relative risk 0.93, 95% CI 0.73–1.20) 2) Non-fatal MI (relative risk 0.84, 95% CI 0.64–1.09) 3) Non-fatal stroke (relative risk 1.04, 95% CI 0.78–1.39) |
| Celli (2009, UPLIFT) [78] | Multicentre, randomised, double blind, placebo-controlled, parallel group (n=5993) Mortality analysis | 4 years | Patients aged >40 years with COPD (post-bronchodilator FEV1 ≤70% pred normal; FEV1/FVC ≤70%) | Tiotropium (n=2987) Placebo (n=3006) | Effect of tiotropium on survival | Reduced risk of cardiac mortality with tiotropium versus placebo (HR 0.86, 95% CI 0.75–0.99) | Reduced risk of all-cause mortality for tiotropium versus placebo (HR 0.84 (95% CI 0.73–0.97), p=0.016) |
| Nojiri 2012 [79] | Single-centre, prospective, pilot (n=21) | 12 weeks | Patients with COPD (FEV1/FVC <0.70) and prior (≥1 year) pulmonary resection for lung cancer ECOG status 1 >20 pack-year smoking history | Tiotropium 18 µg once daily | Pulmonary function and left-ventricular diastolic dysfunction (E/e′ ratio¶) | Significant improvements after tiotropium (versus before tiotropium) in: 1) FEV1 (1.84 L versus 1.60 L, p<0.001) 2) E/e′ ratio (7.59 versus 8.97, p<0.001) | No significant differences before/after tiotropium in BP, heart rate, FVC, left-ventricular mass and LVEF Significant improvement in PASP versus before tiotropium (33.0 mmHg versus 38.5 mmHg, p<0.01) |
| Pepin (2014) [80] | Multicentre, randomised, double dummy, parallel group, blinded (n=257) | 12 weeks | COPD patients aged ≥40 years Smoking history ≥10 pack-years Post-bronchodilator FEV1 ≤70% pred normal FEV1/FVC ≤0.70 aPWV ≥11 m·s−1 | Tiotropium 18 µg once daily (n=130) versus fluticasone furoate/vilanterol combination 100 µg/25 µg once daily (n=127) | Change from baseline in arterial stiffness (aPWV)+ at 12 weeks | At 12 weeks there was a comparable reduction from baseline in aPWV with tiotropium (−1.118 m·s−1) and fluticasone furoate/vilanterol combination (−0.859 m·s−1, p=ns) | No significant differences between tiotropium and fluticasone furoate/vilanterol combination for change from baseline in trough FEV1 (0.080 L versus 0.117 L) or IC (0.019 L versus 0.089 L) |
| Suissa (2017) [65] | UK observational, population-based prior ICS-matched cohort analysis (primary care, CPRD) (n=115 397, base cohort (new users of a long-acting bronchodilator); n=70 550, sub-cohort (linked to HES database)) | 1 year | New users of long-acting bronchodilators (LABA or tiotropium) for COPD Aged >55 years ≥2 years medical history First LABA/tiotropium prescription on/after September 25, 2003 | Full cohort (matched by propensity score): tiotropium (n=26 442), LABA (n=26 442) HES sub-cohort: tiotropium (n=15 427), LABA (n=15 427) | Incidence of acute MI, stroke, HF (full cohort) and incidence of arrhythmia and pneumonia (HES sub-cohort) following 1 year of treatment with tiotropium versus LABA | No difference between tiotropium versus LABA in CV events, as follows: 1) Acute MI: HR 1.10 (95% CI 0.88–1.38) 2) Stroke: HR 1.02 (95% CI 0.78–1.34) 3) Heart failure: HR 0.90 (95% CI 0.79–1.02) 4) Arrhythmia: HR 0.81 (95% CI 0.60–1.09) | Risk of pneumonia significantly reduced with tiotropium versus LABA (HR 0.81, 95% CI 0.72–0.92) |
| β2-agonists and muscarinic antagonists | |||||||
| Berton (2010) [81] | Double-blind, placebo-controlled, crossover (n=12) | Moderate-to-severe COPD (FEV1/FVC <0.7; post-bronchodilator FEV1 <60% pred) Resting PaO2 >60 mmHg | Salbutamol 120 µg + ipratropium 20 µg/actuation versus placebo | Key determinants of O2 delivery and uptake during high-intensity, constant work rate cycling exercise | Compared with placebo, bronchodilators accelerated central haemodynamic response at exercise onset, as follows: 1) t1/2QT: 75.9±10.3 s versus 58.9±18.9 s (p=0.02) 2) t1/2heart rate: 78.2±13.0 s versus 62.5±15.5 s (p=0.03) 3) t1/2SV: 51.0±8.1 s versus 40.6±10.3 s (p=0.02) | Bronchodilators led to lung deflation and increased exercise tolerance versus placebo (454±131 s versus 321±140 s; p<0.05) | |
| Wilchesky (2012–part 1) [82] | Retrospective, cohort (healthcare databases, Province of Saskatchewan, Canada) (n= 6018) | COPD patients aged ≥55 years Newly treated COPD (three or more prescriptions for a bronchodilator, on two different dates, within any 1-year period) | LABA SABA Methylxanthines Ipratropium bromide | Rate of cardiac arrhythmias | Rate of arrhythmia increased with new use of: 1) ipratropium (relative risk 2.4, 95% CI 1.4–4.0) 2) LABA (relative risk 4.5, 95% CI 1.4–14.4) Rate of arrhythmia was not increased with new use of: 1) SABA (relative risk 0.9, 95% CI 0.5–1.6) 2) Methylxanthines (relative risk 1.6, 95% CI 0.7–3.7) | ||
| Wilchesky (2012–part 2) [83] | Retrospective, cohort (healthcare databases, Quebec, Canada) (n=76 661) | COPD patients aged ≥67 years Newly treated COPD (three or more prescriptions for a bronchodilator, on two different dates between January 01, 1990 and December 31, 2002) | LABA SABA Methylxanthines Ipratropium bromide | Rate of cardiac arrhythmias | Rate of arrhythmia increased with new use of: 1) SABA (relative risk 1.27, 95% CI 1.03–1.57) 2) LABA (relative risk 1.47, 95% CI 1.01–2.15) Rate of arrhythmia slightly (not significantly) increased with new use of: 1) ipratropium bromide (relative risk 1.23, 95% CI 0.95–1.57) 2) Methylxanthines (relative risk 1.28, 95% CI 0.93–1.77) | ||
| Gershon (2013) [84] | Population-based, nested case–control analysis of a retrospective study (n=191 005) | COPD patients aged ≥66 years Receiving treatment (September 2003–March 2009) 53 532 had a CV event 26 628 matched to control§ | LABA LAMA | Hospitalisation or ED visit for a CV event | New users of LABAs and LAMAs more likely versus non-users to have a CV-related hospitalisation/ED visit LABAs: OR 1.31 (95% CI 1.12–1.52), p<0.001 LAMAs: OR 1.14 (95% CI 1.01–1.28), p=0.03 | No significant difference in CV events between LABAs and LAMAs (OR 1.15 (95% CI 0.95–1.38), p=0.16) | |
| Wang (2018) [123] | Nested case-control study (healthcare database, Taiwan) (n=284 229) | COPD patients aged ≥40 years LABA–LAMA naïve 37 719 had a CV event Matched with 146 139 randomly selected controls | LABA LAMA | Inpatient or ED visit for CAD, HF, ischaemic stroke or arrhythmia | New LABA and LAMA use were associated with an increased risk of a CV event within 30 days (OR 1.50 (95% CI 1.35–1.67), p<0.001 and OR 1.52 (95% CI 1.28–1.80), p<0.001, respectively). The risk was absent, or even reduced, with prevalent use | No difference in risk was observed between individual LABA agents, LAMA dosage forms, or concomitant COPD regimens | |
| LABA/LAMA combinations | |||||||
| Suissa (2017) [66] | UK observational, population-based cohort analysis (primary care, CPRD) (n=463 899, base cohort) 31 174 patients on combined bronchodilator therapy matched to 31 174 patients on bronchodilator monotherapy | 1 year | New users of long-acting bronchodilators (LABA or tiotropium) for COPD Aged >55 years ≥2 years medical history First LABA/tiotropium prescription on/after September 25, 2002 | LABA/tiotropium initiation + second long-acting bronchodilator (n=31 174) Bronchodilator monotherapy (n=31 174) | Incidence of acute MI, stroke, HF and arrhythmia following 1 year of treatment | Combination of two long-acting bronchodilators was not associated with increased risk of: 1) Acute MI (HR 1.12, 95% CI 0.92–1.37) 2) Stroke (HR 0.87, 95% CI 0.69–1.10) 3) Arrhythmia (HR 1.05, 95% CI 0.81–1.36) | Two long-acting bronchodilators in combination were associated with increased risk of HF (HR 1.16, 95% CI 1.03–1.30) |
| Samp (2017) [85] | Retrospective, observational cohort (using health insurance claims data) (n=19 078 matched patients) | Patients with COPD initiating LABA/LAMA or ICS/LABA | LABA/LAMA (n=3844) ICS/LABA (n=15 234) | CCV outcomes: hospitalisations for ACS, HF, cardiac dysrhythmia, stroke, or TIA | LABA/LAMA treatment was associated with fewer CV events versus ICS/LABA treatment (HR 0.794, 95% CI 0.623–0.997) | No difference between treatments in cerebrovascular events (HR 1.166, 95% CI 0.653–1.959) | |
| Hohlfeld (2017) [86] | Randomised, double-blind, single centre, placebo-controlled, two period, crossover (n=62) | Patients with COPD and increased RV (>135% predicted) without CVD | Indacaterol/glycopyrronium 110 µg/50 µg once daily Placebo | Change in LVEDV measured with MRI on day 14 | Compared with placebo, indacaterol/glycopyrronium resulted in a significant increase in LVEDV (10.27 mL, p<0.0001) and QT (0.337 L·min−1, p=0.0032) | Compared with placebo, indacaterol/glycopyrronium was associated with: 1) Improved lung function (peak FEV1 increased by 0.42 L, p<0.0001) 2) Reduced lung hyperinflation (RV −0.75 L, p<0.0001) | |
| ICS and ICS/LABA combinations | |||||||
| Huiart (2005) [87] | Nested case–control analysis (health services databases, Saskatchewan, Canada) (n=5648) | Follow-up until first MI | Patients ≥55 years with new-onset COPD who had not received any bronchodilator, anti-asthma drug or nasal/ICS in the previous 5 years | ICS | First fatal or non-fatal acute MI | Results based on 371 cases with first acute MI matched to 1864 controls Low-dose ICS (50–200 µg·day−1) significantly reduced risk of acute MI by 32% (rate ratio 0.68, 95% CI 0.47–0.99) | Overall, current use of ICS was not associated with a significant decrease in risk of acute MI (rate ratio 0.82, 95% CI 0.57–1.16) |
| Loke (2010) [88] | Systematic review of 23 RCTs (n=23 396) and 12 observational studies | RCTs ≥24 weeks | Patients with COPD of any severity | ICS versus placebo or ICS/LABA versus LABA | Risk of fatal and non-fatal MI and CV death | Findings from RCTs indicated that ICS were not associated with significantly reduced risk of MI (relative risk 0.95, 95% CI 0.73–1.23), CV death (relative risk 1.02, 95% CI 0.81–1.27) or mortality (relative risk 0.96, 95% CI 0.86–1.07) | Findings from observational studies indicated that ICS use was associated with a significant reduction in CV death (two studies: relative risk 0.79 (95% CI 0.72–0.86), p<0.0001) and mortality (11 studies: relative risk 0.78 (95% CI 0.75–0.80), p<0.0001) |
| Calverley (2010, TORCH) [89] | Multicentre, randomised, double-blind, placebo-controlled, parallel group (post hoc analysis) (n=6184) | 3 years | Patients (current/former smokers) with COPD (pre-bronchodilator FEV1 <60% pred and FEV1/FVC ≤0.70) Aged 40–80 years | Salmeterol/fluticasone propionate combination 50 µg/500 µg (n=1546) Salmeterol 50 µg (n=1542) Fluticasone propionate 500 µg (n=1552) Placebo (n=1544) (all twice daily) | CV AEs and SAEs | The probability of patients having a CV AE by 3 years was lowest for salmeterol/fluticasone propionate combination (20.8%) versus placebo (24.2%), salmeterol (22.7%) and fluticasone propionate (24.3%) | Treatment with salmeterol/fluticasone propionate combination was associated with a significant reduction versus placebo in probability of a CV AE by 3 years in patients receiving CV therapy at baseline (27.9% versus 33.5%, respectively; p<0.05) |
| Dransfield (2011) [90] | Multicentre, randomised, double-blind, placebo-controlled (n=249) | 12 weeks | Patients with COPD (post-bronchodilator FEV1 <80% pred and FEV1/FVC ratio ≤0.70) Aged ≥50 years Smoking history of ≥10 pack-years | Salmeterol/fluticasone propionate combination 50 µg/250 µg twice daily (n=123) Placebo (n=126) (Both arms received open label tiotropium 18 µg once daily for 4 weeks after a 12 week treatment period) | aPWV change from baseline at 12 weeks | For patients that remained on treatment for 12 weeks (n=96 in each group), salmeterol/fluticasone propionate combination was associated with a significant reduction in aPWV versus placebo (−0.49 m·s−1, p=0.045) | No significant changes in aPWV for salmeterol/fluticasone propionate combination + tiotropium versus tiotropium from 12–16 weeks (mean change 0.18 m·s−1) |
| Stone (2016) [91] | Single-centre, randomised, double-blind, placebo-controlled, crossover (n=45) | Two 7-day treatment periods separated by a 7±2-day washout period | COPD (FEV1 <70% pred) Smoking history ≥15 pack-years Aged >40 years MRC score >1 Lung hyperinflation (RV >120% pred) which improved ≥7.5% after salbutamol | Fluticasone furoate/vilanterol combination 100 µg/25 µg once daily Placebo | Change in RVEDVI from baseline versus placebo after 7 days treatment (maximum 14 days) | Mean increase in change from baseline in RVEDVI of 5.8 mL·m−2 (95% CI 2.74–8.91), p<0.001 versus placebo | Improved lung hyperinflation and airflow limitation from baseline with fluticasone furoate/vilanterol combination relative to placebo, as follows: 1) RV: 429 mL reduction (p<0.001) 2) Increased IC, IC/TLC, FEV1 and FVC (261 mL, 4.6%, 220 mL and 350 mL, respectively; all p<0.001) |
| Vestbo (2016, SUMMIT) [92] | Multicentre, randomised, double-blind, placebo-controlled, parallel group, event-driven (n=16 485) | Maximum follow-up was 4 years | COPD (post-bronchodilator FEV1 50–70% pred and FEV1/FVC ≤0.70) Smoking history ≥10 pack-years mMRC dyspnoea scale ≥2 History/increased risk of CVD | Fluticasone furoate/vilanterol combination 100 µg/25 µg once daily (n=4121) Fluticasone furoate 100 µg once daily (n=4135) Vilanterol 25 µg once daily (n=4118) Placebo (n=4111) | All-cause mortality | All-cause mortality did not differ significantly between fluticasone furoate/vilanterol combination and placebo (HR 0.88 (95% CI 0.74–1.04); 12% relative reduction, p=0.137) or components | Fluticasone furoate/vilanterol combination had no effect on composite CV events (CV death, MI, stroke, unstable angina, TIA) compared with placebo (HR 0.93, 95% CI 0.75–1.14) |
| Bhatt (2017) [93] | Multicentre, randomised, double-blind, parallel group, placebo-controlled (stratified by COPD exacerbation history) (n=430) | 24 weeks | Patients aged ≥40 years with a history of COPD ≥10 pack-year smoking history FEV1/FVC ≤0.70 Post-bronchodilator FEV1 ≤70% pred aPWV ≥11 m·s−1 | Fluticasone furoate/vilanterol combination 100 µg/25 µg once daily (n=135) Vilanterol 25 µg once daily (n=154) Placebo (n=141) | Change from baseline in aPWV after 24 weeks with fluticasone furoate/vilanterol combination versus placebo | No significant difference in mean change from baseline in aPWV at 24 weeks with fluticasone furoate/vilanterol combination (−1.75 m·s−1) versus placebo (−1.97 m·s−1) | Post-hoc analysis indicated a greater proportion of respondersƒ in the fluticasone furoate/vilanterol combination versus placebo groups when withdrawn patients were classified as non-responders (50% versus 36%, respectively) |
| Other COPD pharmacological treatments | |||||||
| Suissa (1996) [94] | Population-based cohort from Saskatchewan, Canada (n=12 301) | Patients with asthma aged 5–54 years | Theophylline and β-agonists versus control | Identified 30 CV deaths in which acute asthma did not appear to be a contributing factor | Rate of CV death was greater with theophylline (rate ratio 2.7, 95% CI 1.2–6.1) and β-agonists administered orally or by nebuliser (rate ratio 2.4, 95% CI 1.0–5.4) | Rate of CV death was not greater with β-agonists administered by MDI (rate ratio 1.2, 95% Cl 0.5–2.7) | |
| Huerta (2005) [95] | Nested case-control (population-based cohort, UK General Practice Research Database after January 01, 1994) (n=5710) | Patients aged 10–79 years with asthma or COPD (710 cases and 5000 controls) | Theophylline (and other therapies: β-agonists, oral steroids, ICS) | Rhythm disorders | Short-term theophylline use was weakly associated with: 1) Arrhythmia (relative risk 1.8, 95% CI 1.0–3.3) 2) AF (relative risk 1.8, 95% CI 0.9–3.7) | Short-term theophylline use was associated with supraventricular tachycardia (relative risk 4.0, 95% CI 0.9–18.1) | |
| White (2013) [96] | Pooled analysis of 14 intermediate and long-term trials (n=12 054) | Range: 12–52 weeks | Patients with moderate-to-very-severe COPD | Roflumilast (n=6563) Placebo (n=5491) | MACE (CV death, non-fatal MI and stroke) | MACE composite rate was significantly lower with roflumilast versus placebo (HR 0.65 (95% CI 0.45–0.93), p=0.019) | MACE experienced by 14.3/1000 patient-years (roflumilast) and by 22.3/1000 patient-years (placebo) |
ACS: acute coronary syndrome; AE: adverse event; AF: atrial fibrillation; aPWV: aortic pulse wave velocity; BP: blood pressure; CAD: coronary artery disease; CCV: cardiovascular and cerebrovascular; CPRD: Clinical Practice Research Datalink; CVD: cardiovascular disease; DT-TR: tricuspid E-wave deceleration time; ECOG: Eastern Cooperative Oncology Group; ED: emergency department; E/e′: ratio of peak early diastolic mitral flow velocity to peak early diastolic mitral annual movement velocity; FEV1: forced expiratory volume in 1 s; FRC: functional residual capacity; FVC: forced vital capacity; HES: hospital episode statistics; HF: heart failure; HR: hazard ratio; IC: inspiratory capacity; ICS: inhaled corticosteroid; IRV: inspiratory reserve volume; LABA: long-acting β2-agonist; LAMA: long-acting muscarinic antagonist; LVEDV: left-ventricular end-diastolic volume; LVEF: left-ventricular ejection fraction; MACE: major adverse cardiovascular event; MDI: metered dose inhaler; MI: myocardial infarction; mMRC: modified Medical Research Council; MRI: magnetic resonance imaging; ns: not significant; OR: odds ratio; PaO2: arterial oxygen tension; PASP: pulmonary arterial systolic pressure; QT: cardiac output; RCT: randomised controlled trial; RVEDVI: right-ventricular end-diastolic volume index; RV: residual volume; SABA: short-acting β2-agonist; SAE: serious adverse event; SFC: salmeterol/fluticasone combination; SBP: systolic blood pressure; sPAP: systolic pulmonary arterial pressure; SV: stroke volume; TAPSE: tricuspid annular plane systolic excursion; TIA: transient ischaemic attack; TLC: total lung capacity; VC: vital capacity. #: number of studies; ¶: increases with elevation of left-ventricular filling pressure that occurs due to impaired left-ventricular diastolic function; +: marker of CV risk; §: for each case, one control was randomly selected, matched for age (±1 year), sex, duration of COPD, HF and history of hospitalisation for ACS, HF, ischaemic stroke, cardiac arrhythmia and acute respiratory disease (acute exacerbation of COPD, pneumonia, influenza, or acute bronchitis); ƒ: patients with aPWV reduction from baseline of ≥1 m·s−1 on day 168.