Abstract
Chronic obstructive pulmonary disease (COPD) and cardiovascular disease (CVD) frequently occur together and their coexistence is associated with worse outcomes than either condition alone. Pathophysiological links between COPD and CVD include lung hyperinflation, systemic inflammation and COPD exacerbations. COPD treatments may produce beneficial cardiovascular (CV) effects, such as long-acting bronchodilators, which are associated with improvements in arterial stiffness, pulmonary vasoconstriction, and cardiac function. However, data are limited regarding whether these translate into benefits in CV outcomes. Some studies have suggested that treatment with long-acting β2-agonists and long-acting muscarinic antagonists leads to an increase in the risk of CV events, particularly at treatment initiation, although the safety profile of these agents with prolonged use appears reassuring. Some CV medications may have a beneficial impact on COPD outcomes, but there have been concerns about β-blocker use leading to bronchospasm in COPD, which may result in patients not receiving guideline-recommended treatment. However, there are few data suggesting harm with these agents and patients should not be denied β-blockers if required. Clearer recommendations are necessary regarding the identification and management of comorbid CVD in patients with COPD in order to facilitate early intervention and appropriate treatment.
Abstract
CVD and COPD are often comorbid. Their pathophysiology and treatment may affect each other and health outcomes. http://ow.ly/v18p30lxmms
Introduction
Chronic obstructive pulmonary disease (COPD) is a complex respiratory disorder characterised by chronic airflow limitation and an increased inflammatory response of the lung [1]. COPD is associated with many comorbidities [2, 3] (figure 1) and can be one of multiple chronic or acute diseases and medical conditions present within one person [4]. In particular, cardiovascular disease (CVD) and COPD share similar risk factors such as ageing, history of cigarette smoking (or other exposures) and a sedentary lifestyle, and frequently coexist [2, 3, 5].
Historically, many terms and definitions have been inconsistently used to describe the co-existence of diseases [6]. Recently, it was proposed that the term “index disease” should be used to describe the main condition of interest and “comorbidity” for any other medical conditions present at diagnosis of the index disease or later [6]. By contrast, “multimorbidity” is defined simply as the co-existence of two or more chronic diseases [6, 7].
Patients with comorbid COPD and CVD experience high rates of morbidity, including worse quality of life, dyspnoea and exercise tolerance [8], and a higher risk of hospitalisation for COPD and for CVD [9]. In addition, the presence of CVD or cardiovascular (CV) conditions (such as heart failure (HF), ischaemic heart disease (IHD), diabetes or atrial fibrillation (AF)) increases the risk of frequent exacerbations [10] and mortality [2, 8, 11] (figure 2). Further, COPD exacerbations and lung function decline are associated with increased CV risk and mortality [12, 13]. To minimise the risk of poor outcomes, it is therefore important to ensure that patients with comorbid COPD and CVD are managed effectively. However, drug therapies for COPD could have both beneficial and potential adverse effects on CVD and vice versa.
In the current review we provide a brief overview of key pathophysiological mechanisms, which may help explain comorbid COPD and CVD and inform the rationale for treatment. In the light of the plethora of recent literature considering the CV risks of pharmacotherapy in patients with COPD [14–18] we then provide a synthesis of the research from clinical trials and observational studies to establish the evidence for the main classes of drugs used to treat COPD and CVD. Finally, reflecting on our findings, we make some recommendations for the management of COPD patients with CVD.
Pathophysiological links between COPD and CVD
The mechanisms which underlie the association between COPD and CVD are not well understood but several processes are thought to be important and may interact with each other [19, 20]. These include lung hyperinflation, hypoxaemia, pulmonary hypertension (PH), systemic inflammation and oxidative stress, exacerbations, shared risk factors and shared genetics (figure 3), as well as COPD phenotype.
Hyperinflation, characterised by abnormally elevated residual gas in the lungs following spontaneous exhalation [21], is a major driver of COPD burden and mortality [1]. It is the cardinal pathophysiological mechanism affecting the mechanics of breathing and can be either static (resulting from destruction of the lung parenchyma and subsequent loss of lung elastic recoil) or dynamic (occurring when a patient inhales before exhaling fully, trapping air with each additional breath) [21]. Hyperinflation significantly reduces the efficiency of the respiratory muscles [22] and is increasingly recognised as a major cause of dyspnoea (shortness of breath) [23].
Abnormal lung function, including hyperinflation, is also thought to compromise cardiac function through various means [24]. Notably, airflow limitation caused by lung hyperinflation may cause increased pressures in the cardiopulmonary system, right-ventricular dysfunction, impaired left-ventricular filling and reduced cardiac output (QT) [25–29]. Emphysema is associated with static hyperinflation [30] and hyperinflation may therefore be a key risk factor for CVD in patients with emphysema-predominant COPD. In addition, progressive airflow limitation and emphysema in COPD lead to a ventilation/perfusion mismatch that is a key contributor to the development of hypoxaemia, which can be further exacerbated by exercise and sleep disordered breathing [31]. Hypoxaemia in patients with COPD can lead to pulmonary vasoconstriction and vascular remodelling, resulting in right-ventricular diastolic dysfunction [32]. Indeed, PH, common in patients with severe COPD, can lead to right HF, which is in turn associated with left HF [33, 34]. In addition, altered cardiac repolarisation in patients with COPD may be related to hypoxaemia and could increase the risk of ventricular arrhythmias and sudden cardiac death [35].
Chronic or intermittent hypoxia may also increase systemic inflammation, which is known to play a role in the pathogenesis of CVD [36, 37] and has been linked to the development of arterial stiffness which has a strong predictive value for CV events [38]. Markers of pulmonary inflammation (such as surfactant protein D) and of systemic inflammation (such as C-reactive protein (CRP)) are elevated in patients with stable COPD [8, 39–41], while patients with COPD and CVD have higher blood concentrations of inflammatory markers, such as fibrinogen, interleukin (IL)-6 and IL-8, than those without CVD [8]. Moreover, coronary artery calcification scores (a marker of coronary atherosclerosis) correlate with markers of systemic and pulmonary inflammation, such as IL-6, IL-8, surfactant protein D and peripheral blood neutrophil count, in patients with COPD [42]. These findings suggest that COPD is either a systemic inflammatory state or that inflammatory processes spill over from the lungs into the systemic circulation [43], contributing to the development of CVD.
The occurrence of COPD exacerbations has been shown to increase the risk of subsequent CV events in patients with CVD or CV risk factors [44]. This may be related to lung inflammation which is heightened during COPD exacerbations [24]. Furthermore, lower respiratory tract infections, a common cause of COPD exacerbation, are associated with increased inflammatory markers such as fibrinogen and IL-6 which are linked with thrombosis and CV events [12, 45]. High levels of inflammation and oxidative stress that occur during and after an exacerbation may reduce circulating CD34+ cells (which are involved in vascular repair) and increase platelet activation and arterial stiffness [45–49]. Of note, the sudden increase in airway resistance increases that occurs during an exacerbation further limits expiratory flow and lung emptying. Consequently, patients experiencing an exacerbation tend to adopt a rapid, shallow breathing pattern, resulting in a vicious cycle of diminishing lung emptying time and increasing dynamic hyperinflation [22].
The source of CVD risk may be associated with the dominance of either chronic bronchitis or emphysema within the COPD disease profile. The presence of emphysema may reflect accelerated ageing of the lung, which may also result in HF due to downregulation of anti-ageing molecules such as sirtuins [50], while the presence of chronic bronchitis could reflect a distinct inflammatory subtype of COPD requiring specific anti-inflammatory interventions. The genetic background of chronic bronchitis and emphysema also differ, which may influence the processes underlying the association between COPD and CVD [51].
The physiological effects of CVD, for example HF-associated dyspnoea (due to pulmonary oedema) and reduced exercise capacity (due to reduced QT and impaired perfusive and diffusive oxygen transport) [52, 53] may add to the effects of COPD caused by “pulmonary” mechanisms (dyspnoea due to hyperinflation and resulting exercise avoidance and deconditioning [1]). In HF, cardiomegaly may be involved in causing a restrictive lung pattern and reduced alveolar volume [54], with alveolar gas diffusion progressively worsening due to reduction in the lung tissue participating in gas exchange [55]. Ventilatory response to exercise is greater than normal for a given metabolic rate, due to an increase in physiological dead space to tidal volume ratio (driven by high ventilation/perfusion mismatching, increased carbon dioxide production relative to oxygen uptake from lactate buffering and a decrease in partial pressure of carbon dioxide) [56]. The clinical consequences are illustrated in a study in which patients with COPD and IHD had significantly worse health status, dyspnoea and exercise capacity than those without IHD; additionally, although exacerbations were not more frequent, recovery time was longer in those with IHD [57]. CVD is also a leading cause of hospitalisation and mortality in patients with COPD [9].
Although the causal or consequential role of associated systemic inflammation is still not clear in COPD [58, 59], as described earlier, there are a number of mechanisms by which COPD-associated systemic inflammation could contribute to CVD and low-grade systemic inflammation in patients with airflow obstruction has been associated with increased risk of cardiac injury [40]. Systemic inflammation is also a feature of CVD, likely playing a role in its development and progression [60]. While the effect of CVD-associated systemic inflammation on COPD is unknown, a study in COPD patients with and without hepatitis C found a steeper decline in lung function with hepatitis C versus controls [61], suggesting an effect on the lungs due to systemic inflammation. Furthermore, risk factors of systemic and vascular inflammation, such as visceral obesity, diabetes and inactivity are also associated with reduced pulmonary function, airway hyper-reactivity and eventually COPD [58], supporting the idea that systemic inflammation associated with CVD may affect COPD.
In summary, many of the pathophysiological mechanisms underlying COPD may increase the risk of CVD and vice versa; however, more studies are needed to provide evidence that the mechanisms uncovered are relevant to the clinical manifestations of these diseases when they occur together.
Pharmacological management of COPD and CVD
With the close association between COPD and CVD, it is possible that treatments for one condition may influence the other and appropriate treatment of all conditions is therefore essential. Improvements in survival following hospital discharge for acute COPD exacerbation have been attributed to better management of COPD and associated comorbidities, including the use of CV therapies such as angiotensin-converting enzyme inhibitors (ACEIs), angiotensin receptor blockers (ARBs), β-blockers and statins [62]. However, the treatment of CV conditions in patients with COPD is associated with therapeutic challenges, most notably the co-administration of (selective) β1-blockers and β2-agonists. β-blockers are widely prescribed in the treatment of CVD and β-agonists represent a cornerstone of COPD treatment; however, the two have opposing pharmacological actions and physicians may be concerned that treatment of one condition may worsen the other [63].
A number of randomised controlled trials (RCTs) and observational studies provide insight into the effect of COPD treatments on CVD and vice versa, although both study types are subject to limitations. For instance, results from observational studies may be affected by lack of blinding of study treatments and confounding bias [64]. Also, observational studies are often reliant upon patient databases, which may lack accuracy (for instance, the definition of COPD and the presence of comorbidities such as asthma) [64–66]. In addition, a number of observational analyses discussed below rely upon prescription data, which may not accurately reflect actual medication intake [65, 66]. Conversely, RCTs are associated with limitations of their own, such as their limited size and duration of exposure, as well as a lack of generalisability to the general COPD population due to stringent inclusion criteria and trial setting [64]. Moreover, RCTs may introduce bias due to the deleterious effect of treatment withdrawal at randomisation, the inclusion of a run-in period, the use of untreated placebo groups and by truncating follow-up at treatment discontinuation [67–69].
Effects of COPD treatments on CV risk
Among COPD patients, CV causes are the second most common cause of death (following pulmonary causes) [70]. In this section, we discuss the effects of different classes of COPD treatment on CV risk and mortality, followed by the effects of different classes of CV treatment on COPD. Details from key studies [27, 65, 66, 71–124] are summarised in tables 1 and 2.
Bronchodilators
There is conflicting evidence regarding the CV safety of bronchodilators. An increased CV risk has been observed with bronchodilators in some studies [72, 76, 84, 125], whereas others found no evidence of increased risk or even some evidence of CV risk reduction [89, 126].
β2-agonists
While a systematic review of long-acting β2-agonists (LABAs), in combination with inhaled corticosteroids (ICS), concluded that ICS/LABA products may have a good CV safety profile in asthma patients [127], evidence for the CV safety of LABAs in COPD is less definitive. LABAs have been associated with an increased risk of CV events, possibly due to stimulation of sympathetic drive [72, 84], which may also increase the risk of arrhythmias and myocardial ischaemia in patients with CVD (particularly HF). In clinical trials, β2-agonists have been shown to increase heart rate, reduce potassium concentrations and increase the risk of CV events versus placebo in patients with obstructive airways disease [72]. Furthermore, studies of Canadian healthcare databases indicate that new use of LABAs is associated with an elevated rate of cardiac arrhythmia in patients with COPD [82, 83]. In addition, a recent analysis of patient cohort data from the Taiwan National Health Insurance Research Database found that new LABA use was associated with an increased risk of a severe CV event within 30 days of therapy (OR 1.50, p<0.001 compared with non-use of LABA or long-acting muscarinic antagonists (LAMAs)) [123]. However, there was no increase in risk with prevalent (long-term) use. Furthermore, in the SUMMIT (Study to Understand Mortality and Morbidity in COPD) RCT, which was conducted among patients with heightened CV risk, there was no excess of cardiac disorders or arrhythmias among patients treated with the LABA vilanterol compared with a placebo [92]. Observational data regarding new use of short-acting β2-agonists (SABAs) are also mixed and, in one Canadian database analysis, SABAs did not increase fatal or non-fatal acute myocardial infarction (MI) [71].
There are also data to suggest that β2-agonists can produce effects that may have a positive impact on CV risk. For example, LABAs provide significant improvements in inspiratory capacity (IC), a marker of lung hyperinflation, versus placebo [128–130]. Lung deflation and improvements in QT kinetics and microvascular oxygen delivery have also been observed [81], while improvements in dyspnoea with indacaterol were associated with improvements in indices of right-ventricular compliance [73]. Inhaled LABAs may also have direct beneficial effects on pulmonary haemodynamics (e.g. pulmonary arterial pressure (PAP)) [27] and reduce the rate of COPD exacerbations [70, 131, 132], which are associated with an increased CV risk and mortality [12].
Muscarinic antagonists
Similar to LABAs, LAMAs have been associated with both positive and negative CV effects. Anti-muscarinic agents may suppress parasympathetic control of heart rate, which could increase the risk of tachyarrythmias [133]. However, there are various potential mechanisms whereby muscarinic antagonists could lower CV risk, such as by reducing lung hyperinflation [134, 135].
The LAMA tiotropium was shown to reduce lung hyperinflation and improve CV responses to exercise (such as reducing heart rate and blood pressure (BP)) [75]. The reduction in heart rate correlated with an increase in inspiratory reserve volume (IRV) [75], consistent with an observed improvement in CV effects due to mechanical unloading of ventilator muscles. Tiotropium also reduces arterial stiffness [80]. Other potential CV benefits of LAMAs include reduction in exacerbations and an improvement in left-ventricular diastolic function, although not left-ventricular ejection fraction (LVEF) [79].
There are inconsistent results regarding the effects of muscarinic antagonists on CV risk and mortality in COPD. Meta-analyses and cohort analyses reported that muscarinic antagonists, and tiotropium specifically, are not associated with an increase in CV events, risks, or deaths [65, 77]. In contrast, an analysis of Canadian healthcare databases indicated that new ipratropium use was associated with an elevated rate of cardiac arrhythmia in patients with COPD [82, 83] and another meta-analysis reported that ipratropium and tiotropium increased the risk of the composite endpoint of CV mortality, MI and stroke compared with control therapy [76]. However, it has been suggested that the latter meta-analysis had methodological flaws [136], such as the inclusion of many patients from one study and that many CV deaths occurred in patients who were non-compliant with ipratropium. Tiotropium may also have pro-ischaemic and pro-arrhythmic effects [133]. Further data from Taiwan found that new LAMA use (as well as new LABA use) was associated with a 1.52-fold increase in the risk of a severe CV event within 30 days of therapy initiation compared with non-use of a LABA or a LAMA (p<0.001). It was hypothesised that this effect, observed for both LAMAs and LABAs, may be related to sympathetic over-activation or an increase in inflammatory cytokine levels [123]. However, in the same study, there was no risk associated with long-term use [123] and tiotropium has been associated with a trend towards reduced risk for all-cause, CV and respiratory mortality in RCTs [74, 78, 137].
LABA/LAMA combinations
The CV risk associated with combining long-acting bronchodilators is largely uncertain. A combination of the short-acting bronchodilators salbutamol and ipratropium reduced lung hyperinflation and provided faster QT kinetics as well as larger improvements in microvascular oxygen delivery versus placebo [81]. Furthermore, in the CLAIM study [86], the LABA/LAMA combination indacaterol/glycopyrronium significantly improved left-ventricular and right-ventricular end diastolic volumes, QT and peak forced expiratory volume in 1 s (FEV1), and induced lung deflation compared with placebo.
Large RCTs have not identified additional CV-associated safety concerns with combining long-acting bronchodilators versus monotherapy alone, although studies were not designed to investigate this outcome [64, 138]. Real-world, primary care data indicated that the addition of a second long-acting bronchodilator to existing LABA or tiotropium treatment was not associated with an increased risk of acute MI, stroke or arrhythmia after 1 year compared with monotherapy, but the risk of developing HF was significantly elevated [66]. Furthermore, HF risk increased by 21% when patients with prior HF were removed from the analysis [66]. The reasons for this elevation are unknown, and further investigation and monitoring are required. However, when comparing LABA/LAMA and ICS/LABA therapy in a retrospective, observational cohort study, CV events were less frequent with LABA/LAMA treatment, with no significant difference observed in the risk of cerebrovascular events [85].
ICS, ICS/LABA and ICS/LABA/LAMA combinations
The effects of ICS on systemic inflammation [41, 139, 140] and cardioprotection [87, 88] are unclear. Systemic corticosteroids appear to promote progression of atherogenesis, but may also improve recovery from occlusive vascular events and intravascular injury [141, 142]. Lung deflation with fluticasone furoate/vilanterol improved cardiac function (right-ventricular, left-ventricular and left atrial volumes) versus placebo [91]. Treatment with ICS/LABA also reduced arterial stiffness to a similar extent as tiotropium in a study of 257 patients with COPD, suggesting that the long-acting bronchodilator component of the combination drives this effect [80]. However, no reduction in aortic pulse wave velocity (aPWV) was observed compared with placebo [90, 93].
A hypothesis-generating post hoc analysis of TORCH (Towards a Revolution in COPD Health) indicated that the ICS/LABA combination salmeterol/fluticasone propionate might have beneficial effects on reducing CV events in certain subgroups of patients, such as those receiving CV medications at baseline or with moderate COPD [89]. The SUMMIT study investigated this hypothesis by comparing fluticasone furoate/vilanterol, respective monotherapies, and placebo in patients with moderate COPD and CVD or multiple CV risk factors [92]. Fluticasone furoate/vilanterol did not significantly reduce the risk of mortality versus placebo and there were no significant differences when data were analysed by age, sex, baseline therapy or presence of CVD [92]. In addition, there was no effect on the composite CV endpoint [92]. Why SUMMIT did not reproduce the results of the TORCH post hoc analysis is unclear. This could reflect differences in the efficacy of the treatments used, or other methodological considerations related to RCTs [68], as discussed above.
Data on the CV risks or benefits of ICS or ICS/LABA combinations in observational studies are also mixed, with reports that low-dose ICS (50–200 µg·day−1) may be associated with a reduction in the risk of acute MI in patients with COPD [87]. Although observational studies have indicated that ICS have benefits on CV mortality, RCTs have not shown any significant effect on CV death [88]. To date, no studies have been designed to investigate the effect of ICS/LABA/LAMA combination therapies on CV events; however, RCTs comparing ICS/LABA/LAMA with LAMA/LAMA or ICS/LABA combinations have provided no evidence of an excess CV risk with triple therapy [143, 144].
Other COPD treatments
Phosphodiesterase-4 inhibitors, such as roflumilast, are recommended as an additional therapy on top of triple therapy (a combination of ICS, LABA and LAMA) in inadequately controlled patients in Global Initiative for Chronic Obstructive Lung Disease (GOLD) Group D [1]. Phosphodiesterase inhibitors have anti-inflammatory effects, can reduce the risk of exacerbations [145] and may therefore have beneficial effects on CV events in patients with COPD and increased CV risk. Available data are limited, however, a lower rate of major CV adverse events (AEs) with roflumilast versus placebo has been reported [96]. A more recent meta-analysis did not identify any particular CV safety signal with roflumilast, but it did report an increased incidence of some (non-CV) AEs, which led the authors to conclude that further, long-term safety studies were needed [146].
Theophylline, a further possible treatment for COPD, [1] has anti-inflammatory effects [147] and intravenous theophylline reduces pulmonary vascular resistance (PVR) and improves ventricular function [148]. However, high doses of theophylline have been linked to changes in cardiac electrophysiology [149]. Weak associations of theophylline and arrhythmias, AF and supraventricular tachycardias have been suggested by a UK case-control study [95], and a cohort study of 12 301 subjects from Canada indicated that the rate of CV death was greater in users of theophylline (with most deaths occurring in individuals with pre-existing conditions) [94]. While this latter analysis was not specific to patients with COPD, the balance of available data suggests that theophylline should be used with caution in patients with COPD and CVD.
Aside from pharmacological strategies, pulmonary rehabilitation (PR) is recommended for patients with COPD as part of integrated patient management [1]. For those with CVD, it is important to consider both diseases when creating PR programmes [150]. Retrospective analyses and prospective studies provide conflicting insight into the effect of CVD on PR efficacy, with the former suggesting a reduction in the ability to achieve a clinically-important difference in 6-minute walking distance and health status in patients with metabolic diseases versus those without [151]. However, when studied prospectively, this effect was not observed [152]. Furthermore, the effect of PR on CVD risk is largely unknown. A systematic review of studies investigating arterial stiffness in response to PR or an exercise-training programme in COPD patients only found three eligible studies, with conflicting results [153]; however, there may be subpopulations of patients who benefit from PR with regards to CV risk [153].
Finally, as an intervention, smoking cessation has the greatest potential to influence the natural history of COPD [1]. A network meta-analysis of pharmacological and behavioural smoking cessation interventions in CVD patients indicated that varenicline and bupropion were associated with greater abstinence than placebo, with increased efficacy also noted with telephone therapy and individual counselling versus usual care [154]. However, a further network meta-analysis found no association between cessation medications and major CV events, although nicotine replacement therapy was associated with an increase in CV events overall (driven predominantly by less serious events) compared with smoking cessation advice alone [155, 156].
In summary, the clinical data suggests that inhaled therapies used in the treatment of COPD are not associated with significant CV risk while smoking cessation remains a core strategy for both COPD and CVD, reducing the overall risk of premature mortality from smoking-related diseases. However, more data are required to establish the long-term safety of COPD pharmacotherapies, particularly among patients at high risk of CV events who are often excluded from COPD clinical trials. Observational studies suggest that new users of LABAs and LAMAs with COPD are at higher risk of CV events compared with non-use of LABAs and LAMAs, which indicates the need to monitor patients with CV comorbidities, particularly when bronchodilator therapies are initiated.
Effects of CV treatments on COPD
β-blockers
The European Society of Cardiology (ESC) advocates the use of β-blockers in HF patients irrespective of the presence of COPD [52], but advises caution in patients with stable coronary artery disease (CAD) and concurrent COPD [157]. The use of β-blockers in patients with COPD and concurrent CVD has historically been avoided because of concerns about potential adverse pulmonary effects. Bronchial smooth muscle contains adrenergic receptors, primarily of the β2-subtype, and activation of these receptors by agonists causes bronchodilation [158]. Consequently β-blockers are contraindicated in patients with COPD due to concerns about the potential for acute bronchospasm [159]. Indeed, the non-selective blockade of β-receptors by agents such as propranolol has been shown to inhibit the bronchodilator response to β2-agonists in patients with COPD [160]. However, concerns over the safety of β-blocker therapy in patients with COPD have resulted in sub-optimal therapy in patients with CVD and comorbid COPD [161, 162].
Data which address these safety concerns are increasing. Selective β1-blockers (e.g. atenolol, bisoprolol and metoprolol) have a 20-fold greater affinity for β1-receptors versus β2-receptors; thus, this subclass of agents are significantly less likely to induce bronchoconstriction. Clinical trials and large meta-analyses have shown that single-dose and long-term use of selective β1-blockers does not have a significant effect on FEV1, β-agonist response, respiratory symptoms, or overall patient condition compared with patients with COPD not receiving β1-blockers [163–165]. Indeed, selective β1-blockers and β2-agonists may have complementary effects, as the use of β1-blockers in patients with COPD can sensitise β2-receptors to β2-agonists [166]. Clinical hesitance to administer selective β1-blockers when respiratory conditions are present is reflected in review articles and practice guidelines; however, a Cochrane review found that β1-blockers in mild-to-moderate reversible airway disease or COPD did not produce adverse respiratory effects [163]. Furthermore, selective β1-blockers do not increase the risk for moderate or severe exacerbations in patients with asthma [167]. This finding is particularly noteworthy given that airflow obstruction reversibility (following β2-agonist inhalation) is characteristically greater in asthma than in COPD. Thus, one might expect that patients with asthma would be more vulnerable to the adverse effects of β-blockade than patients with COPD; however, with selective β1-blockade this appears not to be the case.
Evidence from observational studies suggests that β-blockers are associated with various benefits in patients with COPD with or without CVD, such as reductions in mortality, hospital admissions, emergency room (ER) visits and COPD exacerbations [97–99, 101–103], although these studies may have been affected by biases from immortal and immeasurable time [168–170].
Furthermore, the continued use of β-blockers in COPD patients hospitalised for exacerbation did not result in an increase in in-hospital mortality, 30-day readmission or late mechanical ventilation [100]. However, the use of β-blockers may have little effect on lung function and dynamic hyperinflation [104]. Such data must be interpreted cautiously in the absence of RCT data, particularly given the recent discrepancy between observational and RCT data in CV therapy for COPD discussed below in relation to statins.
Renin–angiotensin–aldosterone system inhibitors ACEIs and ARBs
The renin–angiotensin–aldosterone system (RAAS) has been implicated in various processes in the lungs that may be important in the pathogenesis of COPD, including the induction of pro-inflammatory modulators, the generation of reactive oxygen species and the development of pulmonary fibrosis [171]. Data on the effects of RAAS inhibitors in patients with COPD are limited, although a couple of small studies have indicated a potential benefit on pulmonary function and haemodynamics [105, 106]. More recently, an analysis of the Multi Ethnic Study of Atherosclerosis (MESA), including individuals in the general population aged 45–84 years who had no clinical evidence of CV disease, found that baseline use of an ACEI or ARB protected against the progression of emphysema, especially when prescribed at high doses [124]. The authors attributed the effects of such RAAS inhibitors to inhibition of transforming growth factor-β signalling in the lung, thereby reducing the progression of airspace enlargement.
Unfortunately, the most common side effect of therapy with ACEIs is cough, which develops in 5–20% of patients and may be problematic for patients with COPD. Although uncommon, worsening airflow obstruction has been associated with ACEI treatment, leading to suggestions that these agents be used with caution or as a second-line in patients with COPD [158]. No such safety concerns have been reported with ARBs [158]; indeed, data from an observational cohort study among patients with COPD who used ACEIs or ARBs found that ARBs were associated with fewer COPD complications, including severe exacerbations, pneumonia and mortality, than ACEIs [122]. Although these findings require further confirmation, this may suggest that ARBs are a better choice for patients with COPD requiring treatment with a RAAS inhibitor compared with ACEIs [122].
Statins
Given the high prevalence of CVD and CV risk factors in COPD [3], many patients with COPD receive statin therapy for the primary or secondary prevention of CVD. The lipid-lowering effects of statins are well documented; however, anti-inflammatory effects have also been observed in the airways and CV tissue [172–176] and therefore these agents could have beneficial effects in COPD. Indeed, statin therapy has been shown to improve Borg dyspnoea scores versus placebo in patients with COPD and PH [109] and data from retrospective analyses and prospective observational studies support a role for statins in patients with COPD in terms of reductions in exacerbations, hospitalisations and mortality after an exacerbation [107, 110–112, 114, 117, 177, 178]. However, many of these studies may have been affected by biases from immortal and immeasurable time [179].
In the Rotterdam study, long-term statin use had a beneficial effect on mortality in patients with COPD compared with never use [115]. Observational studies suggest that statins may also be associated with a reduced prevalence of airflow limitation [113] and a reduction in FEV1 decline [108]. However, an analysis of patients with chronic HF and history of COPD found no favourable effect between statins and all-cause mortality, CV death, non-CV death or all-cause hospitalisation [118], Indeed, in the STATCOPE study (a randomised placebo-controlled trial of simvastatin in the prevention of COPD exacerbations), statin therapy did not decrease exacerbation rates [116]. However, patients with diabetes and coronary heart disease were not included in this study and therefore a benefit of statin therapy on patients at CV risk cannot be excluded.
Anticoagulants
Anticoagulants (e.g. warfarin and newer agents such as apixaban, dabigatran, edoxaban and rivaroxaban) are often used to prevent future thrombotic events in patients with CVD. It is well known that smoking increases the risk of venous thromboembolism (VTE) [180], but patients with severe COPD are also at increased risk of secondary VTE and mortality is higher in patients who have COPD and VTE versus COPD alone [181]. Furthermore, a recent meta-analysis revealed that pulmonary embolism was present in approximately one-sixth of patients who had an acute COPD exacerbation of unknown cause and that these emboli were often in regions indicated for anticoagulant therapy [182]. Thus, use of anticoagulants in at-risk COPD patients may help to reduce future thrombotic-related morbidity and mortality.
Antiplatelet therapy
Thrombocytosis has been reported in an observational cohort of 1343 patients hospitalised for an acute exacerbation of COPD and antiplatelet therapy correlated with lower 1-year mortality in this study [120]. A national prospective multicentre study also suggests a positive effect of antiplatelet therapy on mortality in patients with COPD [119].
Other CV medications
Data are limited for the effects of other CV medications in patients with COPD. There are no safety concerns or contraindications regarding the use of calcium channel blockers (CCBs) or aldosterone receptor blockers, such as spironolactone, in patients with COPD [158]. In patients with COPD and hypertension, the use of a thiazide diuretic in combination therapy did not affect the risk of COPD exacerbations [121]. However, respiratory acidosis, a common condition in patients with COPD, may be further complicated by the effects of diuretics on electrolyte levels and acid–base balance, leading to development of mixed acid–base disorders [183]. For example, treating patients with COPD and HF with high-dose diuretics can result in metabolic acidosis and metabolic alkalosis, in addition to pre-existing respiratory acidosis [183]. Patients with HF with reduced ejection fraction may also be treated with sacubitril/valsartan, a first-in-class angiotensin receptor–neprilysin inhibitor (ARNI) [184]. However, the potential impact of ARNIs on COPD outcomes has not been studied.
Overall, the evidence to support the use of CV therapies in patients with COPD is reassuring. Indeed, there are some data to suggest that RAAS inhibitors and statins have a protective effect on the progression of COPD although outcomes studies specifically designed to assess the effect of therapy on patients with COPD and CV would be welcome. Furthermore, historic concerns about the safety of β-blockers are not borne out by the data, which indicates that this class is beneficial in patients with COPD.
Recognising and managing comorbid COPD and CVD
Comorbidity in COPD is common and can be fatal. Therefore, in order to treat patients appropriately, physicians should proactively search for prevalent and clinically important comorbidities such as CVD. The three most commonly occurring cardiac comorbidities of COPD are AF, HF and IHD [185]. However, non-specific symptoms such as dyspnoea and fatigue are common to all four diseases and acute exacerbations of each of these conditions can result in exacerbation of respiratory symptoms. Indeed, acute respiratory symptoms invariably have mixed pulmonary and cardiac origin [185]. Diagnosing and managing comorbid COPD and CVD and their exacerbations thus remain challenging [186].
Recognition
Given the impact of CVD in COPD, it is important to recognise and treat CVD and CV risk factors (such as smoking, cholesterol and BP) as early as possible, but practical guidance is limited. Numerous guidelines have been published to assist in the differential diagnosis of COPD and CVD in isolation, but few exist in the setting of comorbidity or multimorbidity. However, Roversi et al. [185] have made important progress in this regard, publishing a set of diagnostic and screening procedures to help differentiate COPD from AF, HF and IHD.
When used in addition to lung function tests, CV risk scores significantly improve the prediction of CV events and mortality in patients with COPD [187]. Over and above guidelines-recommended diagnostic procedures for CVD, further subclinical markers and biomarkers may also assist the assessment and diagnosis of patients in clinical practice. For example, arterial stiffness (as measured by aPWV) has been shown to be predictive of CV events independent of classic CV factors and has been proposed as a surrogate marker of severity [38]. In addition, a prospective cohort study indicated that angiopoietin-like protein 4 was independently associated with CV function in patients with COPD [188].
There is also potential for pulmonologists to learn from other specialities that recognise the importance of identifying and managing key comorbidities. For example, diabetes guidelines from the American Diabetes Association recognise the need to look beyond blood glucose, highlighting the importance of effectively managing high BP in patients with diabetes [189].
Management
Key international, regional and local guidelines provide limited recommendations on how to manage patients with COPD and CVD [1, 190–192]. The GOLD strategy document states that the presence of comorbidities should not, in general, alter COPD treatment and that comorbidities should be treated as per usual standards, irrespective of the presence of COPD [1].
In approaching the treatment of COPD in HF patients, the ESC advocate the use of β-blockers without contraindication for COPD [52]. Unfortunately, despite no consistent evidence contraindicating concurrent β-blocker and LABA administration, some patients with COPD and CVD are not receiving guideline-based therapies due to historical concerns [63, 193] and patients with HF are less likely to receive β-blockers if they also have COPD [63, 193].
In 2016, the National Institute for Health and Care Excellence published a set of clinical guidelines for the assessment, prioritisation and management of patients with commonly occurring multimorbidities [7]. However, more extensive, integrated recommendations specifically concerning CV risk assessment and management in patients with COPD are necessary to optimise management of comorbid COPD and CVD. Indeed, in order to ascertain the most appropriate treatment in patients with further exacerbations despite treatment with a LABA/LAMA, recent proposals for alternative treatment algorithms advocate the assessment of comorbidities [186]. Adoption of strategies aimed at improving outcomes in these “cardiopulmonary” patients (e.g. via joint management by respiratory and cardiac health professionals in a cardiopulmonary outpatient clinic) is to be encouraged [185].
Our recommendations
CV diseases are often underdiagnosed and under-treated in patients with COPD [18]. Furthermore, treatment with therapies targeted to treat comorbidities such as HF, IHD, AF and hypertension have the potential to modify the natural history of patients with COPD [115, 124]. We therefore support the active detection and management of comorbidities in COPD and suggest using relevant treatment guidelines for patients without COPD in the absence of more specific information. Our recommendations for clinicians regarding management of patients with both COPD and CVD in clinical practice are as follows: 1) multimorbidity is often overlooked during the diagnosis of an initial (i.e. index) chronic disease; however, all patients who have COPD, CVD or another index chronic disease should be evaluated thoroughly to rule out the presence of additional chronic diseases. This is especially important for index diseases such as COPD, in which treatment is purely symptomatic and does not address underlying pathophysiological causes; 2) consider and measure CV risk in every patient with COPD using a validated score (e.g. QRISK3 [194]).Where this gives intermediate results, consider an individualised assessment of risk, such as coronary calcium score or pulse wave velocity. Assess diffusion capacity for the differential diagnosis of dyspnoea in COPD; 3) in patients with COPD, manage comorbid CVD and CV risks according to guidelines. Prescribe β-blockers only in accordance with approved indications, use cardioselective β1-blockers (e.g. atenolol, bisoprolol and metoprolol), initiate treatment at the lowest dose and up-titrate slowly. Engage/interact with other medical disciplines to address treating patients with multimorbidity; 4) in patients with COPD, monitor lung hyperinflation routinely via IC assessments; 5) reassess risk and control in response to changes in intervention
Conclusion
COPD and CVD are complex disorders that frequently co-exist and are associated with worse outcomes than either condition alone. Potential mechanisms have been discussed whereby COPD and CVD may interact and treatments for COPD may help to reduce the risk of CVD. These include mechanical offloading of the CV system through reductions in lung hyperinflation, anti-inflammatory effects and the prevention of exacerbations. Current guidelines offer only limited recommendations for the management of CVD in patients with COPD and there is an urgent need for more extensive and specific recommendations to guide physicians in clinical practice.
Acknowledgements
The authors were assisted in the preparation of the manuscript by Graham Allcock and Elizabeth Andrew (professional medical writers, CircleScience (an Ashfield company, part of UDG Healthcare plc), Tytherington, UK) and Laura Elson (a professional medical writer contracted to CircleScience). Medical writing support was funded by Novartis Pharma AG (Basel, Switzerland).
Footnotes
This article has been amended according to the erratum published in the December 2018 issue of the European Respiratory Review.
Provenance: Submitted article, peer reviewed.
Conflict of interest: K.F. Rabe reports personal fees (consultant and speakers fees) from AstraZeneca, Boehringer Ingelheim, Novartis, Sanofi, Teva, Intermune, Chiesi Pharmaceuticals and Berlin Chemie, as well as grants from the Ministry of Education and Science, Germany, outside the submitted work.
Conflict of interest: J. Hurst reports grants, personal fees and non-financial support from pharmaceutical companies that make medicines to treat COPD, outside the submitted work.
Conflict of interest: S. Suissa reports grants and personal fees (board membership and research grant) from Novartis and Boehringer Ingelheim, and personal fees (lectures) from AstraZeneca, outside the submitted work.
Support statement: This study was funded by Novartis. Funding information for this article has been deposited with the Crossref Funder Registry.
- Received June 19, 2018.
- Accepted August 20, 2018.
- Copyright ©ERS 2018.
ERR articles are open access and distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0.