- Email: [email protected]
Chronic obstructive pulmonary disease and ventricular arrhythmia—Association or causation?
Accepted Manuscript Chronic obstructive pulmonary disease and ventricular arrhythmia – association or causation? Nathaniel M. Hawkins, MBChB MD MPH PII:
S1547-5271(17)31211-0
DOI:
10.1016/j.hrthm.2017.10.018
Reference:
HRTHM 7351
To appear in:
Heart Rhythm
Received Date: 10 October 2017
Please cite this article as: Hawkins NM, Chronic obstructive pulmonary disease and ventricular arrhythmia – association or causation?, Heart Rhythm (2017), doi: 10.1016/j.hrthm.2017.10.018. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Chronic obstructive pulmonary disease and ventricular arrhythmia – association or causation?
Nathaniel M Hawkins MBChB MD MPH 1
Division of Cardiology, University of British Columbia, Vancouver, Canada.
RI PT
1
Correspondence to:
SC
Nathaniel M Hawkins University of British Columbia
Tel: 604 875 4111 Fax: 604 875 5504
TE D
e-mail: [email protected]
M AN U
St. Paul’s Hospital, 1081 Burrard Street, Vancouver, British Columbia V6Z 1Y6
AC C
Conflicts of Interest
EP
Word Count: 1470 including references
Conflicts of interest: none.
ACCEPTED MANUSCRIPT
In this edition of the journal, Konecny et al. retrospectively examine the association between airflow obstruction and ventricular tachycardia (VT) in 6351 patients with documented spirometry, Holter monitoring and echocardiography at the Mayo clinic.1 Non-sustained VT (NSVT) was more common in patients with compared to without obstruction (23% versus 13%), correlated with
RI PT
percent predicted forced expiratory volume in 1 second (FEV1), and remained independently associated with obstruction after adjustment for confounders including left ventricular ejection fraction (LVEF) (adjusted odds ratio 1.33 (95% CI 1.15 – 1.54).
SC
Chronic obstructive pulmonary disease (COPD) is a leading cause of hospitalization in developed countries, and the only major cause of mortality for which death rates continue to rise.
M AN U
One third of deaths in COPD relate to cardiovascular disease, the majority of which are classified as sudden.2 Previous studies have suggested an increased burden of ventricular arrhythmia in COPD, but have been limited in terms of sample size, stratification of COPD severity, and ability to adjust for confounding factors. The present study in part addresses these limitations and the classic Bradford Hill criteria for assessing causation,3 demonstrating for the first time a clear dose response
TE D
relationship between exposure and effect, independent of the foremost confounders (LVEF and coronary disease). This gradient was consistent for premature ventricular contractions,
EP
idioventricular rhythm and NSVT. Moreover, no significant relationship between severity of airflow obstruction and ambient heart rate was apparent, suggesting a lesser etiological role for
AC C
sympathetic tone and ventricular automaticity. Though pointing towards a causative relationship between COPD and VT, our interpretation must be tempered by considerations of study design, bias and confounding. First, the study employs airflow obstruction and non-sustained arrhythmia as surrogates for COPD and VT respectively. The former arguably impairs generalizability, given that fixed spirometric criteria over-diagnose airflow obstruction relative to lower limit of normal criteria.4 Notably, only 22% of patients with mild obstruction were prescribed beta-agonists, suggesting a significant proportion may not have a clinical diagnosis of COPD. Nevertheless, spirometry provides the objective measure central to
ACCEPTED MANUSCRIPT
defining the biological gradient, a key strength of the study. The adoption of NSVT as surrogate for ventricular arrhythmia is more challenging, as the authors acknowledge. While PVC burden and NSVT are associated with all-cause mortality in patients with structural heart disease, the relationship with sudden death is more complex, and suppression with anti-arrhythmic drug therapy
RI PT
has been unrewarding. Secondly, retrospective linkage of three separate and temporally dissociated investigations (spirometry, Holter and echocardiography) creates selection or indication bias whose significance is
SC
difficult to quantify, interpret or control. A reasonable assumption is that patients receiving all three tests are more likely to have pathological findings. Moreover, the impact of medical events between
M AN U
investigations (e.g. myocardial infarction occurring after a normal echocardiogram but before an abnormal Holter) is almost impossible to quantify in these circumstances. Considering again the Bradford Hill criteria, there is certainly biological plausibility and coherence between epidemiological and laboratory observations. The proarrhythmic milieu in patients with COPD includes hypoxia, acid-base and electrolyte disturbance, beta-agonists,
TE D
autonomic neuropathy, QTC prolongation, and underlying cardiac disease.5 Nevertheless, the inevitable question for all observational studies is whether unmeasured confounders may account
EP
for the observed relationship. Konecny et al. adjust for multiple factors, included LVEF as a continuous variable and presence of coronary disease. As alluded to earlier, the observed correlation
AC C
between severity of obstruction and arrhythmia burden infers a degree of causality. However, confounders themselves may also exhibit proportionality, such as ventricular remodelling in relation to cumulative smoking history,6 severity and control of risk factors (not simply presence/absence), extent of scar (LVEF being a relatively crude surrogate), and severity of underlying coronary disease.7 All these factors may be worse in more severe lung disease, contribute to increased ventricular arrhythmia, and be incompletely captured in standard models. These thoughts return to the first Bradford Hill criteria, namely strength of association.3 The magnitude of the unadjusted odds ratio, and the degree of attenuation with multivariable regression
ACCEPTED MANUSCRIPT
(odds ratio 1.33 [1.15-1.54]), suggests the relationship is relatively weak and in large part attributable to underlying cardiovascular disease (as opposed to the COPD disease state per se). Further characterization of unmeasured confounders, or the severity of cardiovascular disease, could well render the independent association non-significant. This raises a somewhat philosophical
RI PT
debate regarding adjusted versus actual risk. While research extols adjusted risk as a marker of causal inference, patients care about their absolute risk in unadjusted terms. Treatment decisions in clinical practice are also based on absolute thresholds defining high risk groups e.g. impaired LVEF
SC
for beta-blockers or implantable cardioverter defibrillators (ICDs). What is clear in this study, and many others, is that patients with COPD suffer a high burden of arrhythmia.
M AN U
The authors focus the discussion on ICD eligibility, posing the question whether COPD conveys sufficient risk to justify ICD therapy beyond established guideline criteria. Figure 3 of the analysis is particularly informative, depicting the incidence of NSVT stratified by both COPD severity and LVEF.1 In patients with no, mild, moderate or severe COPD, compared to those with normal LVEF, NSVT was consistently twice as common in patients with mid-range LVEF, and
TE D
three times more common in those with LVEF <40%. It may be helpful in future work to stratify the arrhythmic burden using an LVEF threshold of 35%, to determine the proportion of patients
EP
already eligible for ICDs. An equally pertinent question is how many patients fulfilling guideline criteria actually have an ICD implanted. COPD is a recognized barrier to care, with implantation
AC C
rates consistently lower in those with pulmonary disease. A common misconception is that higher baseline all-cause mortality risk precludes substantial benefit from ICD therapy. However, three key factors determine ICD benefit: annual total mortality; the ratio of arrhythmic to non-arrhythmic death; and time horizon of follow-up. Providing a sufficient proportion is arrhythmic in origin, the highest risk populations in fact derive significant gain.8 Referring once more to Figure 3, the challenge for COPD an a ‘standalone’ marker is that arrhythmia burden is only markedly elevated in very severe obstruction (FEV1 percent predicted <0.3). Not only does this group represent a small proportion of the overall population with COPD (6%, n=166), but the competing risk of non-cardiac
ACCEPTED MANUSCRIPT
death reduces the ratio of arrhythmic to non-arrhythmic mortality, and hence the lifespan-gain and cost-effectiveness of ICD therapy. Acknowledging this, possibly the simplest opportunity for improvement is evident in Table 1 of the analysis. Despite the high prevalence of coronary artery disease, arrhythmia burden and heart
RI PT
failure, beta-blocker utilization was only 29% in patients with mild through to severe airflow obstruction. Evidence has steadily amassed that beta-blockers, particularly cardio-selective, are safe, well tolerated, and associated with improved outcomes in patients with COPD and
SC
cardiovascular disease.5 There is also a consistent interaction between beta-blockers and betaagonists, the former attenuating the risk of sudden death and adverse cardiovascular events 9
To conclude, the thoughtful work of Konecny et al. is another
M AN U
associated with the latter.5,
important advance in our understanding of arrhythmic risk in pulmonary disease, but should also prompt us to consider how best to reduce this risk with the tools at our disposal.
TE D
References
Konecny T, Somers KR, Park JY, John A, Orban M, Doshi R, Scanlon PD, Asirvatham SJ, Rihal CS, Brady PA. Chronic obstructive pulmonary disease as a risk factor for ventricular arrhythmias independent of left ventricular function. Heart Rhythm 2017.
2.
McGarvey LP, John M, Anderson JA, Zvarich M, Wise RA. Ascertainment of cause-specific mortality in COPD: operations of the TORCH Clinical Endpoint Committee. Thorax 2007;62:411-415.
3.
Hill AB. The environment and disease: association or causation? Proc R Soc Med May 1965;58:295-300.
4.
Roberts SD, Farber MO, Knox KS, Phillips GS, Bhatt NY, Mastronarde JG, Wood KL. FEV1/FVC ratio of 70% misclassifies patients with obstruction at the extremes of age. Chest Jul 2006;130:200-206.
5.
Hawkins NM, Petrie MC, Macdonald MR, Jhund PS, Fabbri LM, Wikstrand J, McMurray JJ. Heart failure and chronic obstructive pulmonary disease the quandary of Beta-blockers and Beta-agonists. J Am Coll Cardiol 2011;57:2127-2138.
6.
Nadruz W, Jr., Goncalves A, Claggett B, Querejeta Roca G, Shah AM, Cheng S, Heiss G, Ballantyne CM, Solomon SD. Influence of cigarette smoking on cardiac biomarkers: the Atherosclerosis Risk in Communities (ARIC) Study. Eur J Heart Fail Jun 2016;18:629-637.
AC C
EP
1.
ACCEPTED MANUSCRIPT
Dursunoglu N, Dursunoglu D, Yildiz AI, Uludag B, Alacam ZN, Saricopur A. Severity of coronary atherosclerosis in patients with COPD. The clinical respiratory journal Nov 25 2015.
8.
Raphael CE, Finegold JA, Barron AJ, Whinnett ZI, Mayet J, Linde C, Cleland JGF, Levy WC, Francis DP. The effect of duration of follow-up and presence of competing risk on lifespan-gain from implantable cardioverter defibrillator therapy: Who benefits themost? European Heart Journal 2015;36:1676-1688.
9.
Narayanan K, Reinier K, Uy-Evanado A, Teodorescu C, Zhang L, Chugh H, Nichols GA, Gunson K, Jui J, Chugh SS. Chronic Obstructive Pulmonary Disease and Risk of Sudden Cardiac Death. JACC: Clinical Electrophysiology 2015;1:381.
RI PT
7.
SC
Funding
AC C
EP
TE D
M AN U
The author received no financial support in preparation of the manuscript.
S1547-5271(17)31211-0
DOI:
10.1016/j.hrthm.2017.10.018
Reference:
HRTHM 7351
To appear in:
Heart Rhythm
Received Date: 10 October 2017
Please cite this article as: Hawkins NM, Chronic obstructive pulmonary disease and ventricular arrhythmia – association or causation?, Heart Rhythm (2017), doi: 10.1016/j.hrthm.2017.10.018. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Chronic obstructive pulmonary disease and ventricular arrhythmia – association or causation?
Nathaniel M Hawkins MBChB MD MPH 1
Division of Cardiology, University of British Columbia, Vancouver, Canada.
RI PT
1
Correspondence to:
SC
Nathaniel M Hawkins University of British Columbia
Tel: 604 875 4111 Fax: 604 875 5504
TE D
e-mail: [email protected]
M AN U
St. Paul’s Hospital, 1081 Burrard Street, Vancouver, British Columbia V6Z 1Y6
AC C
Conflicts of Interest
EP
Word Count: 1470 including references
Conflicts of interest: none.
ACCEPTED MANUSCRIPT
In this edition of the journal, Konecny et al. retrospectively examine the association between airflow obstruction and ventricular tachycardia (VT) in 6351 patients with documented spirometry, Holter monitoring and echocardiography at the Mayo clinic.1 Non-sustained VT (NSVT) was more common in patients with compared to without obstruction (23% versus 13%), correlated with
RI PT
percent predicted forced expiratory volume in 1 second (FEV1), and remained independently associated with obstruction after adjustment for confounders including left ventricular ejection fraction (LVEF) (adjusted odds ratio 1.33 (95% CI 1.15 – 1.54).
SC
Chronic obstructive pulmonary disease (COPD) is a leading cause of hospitalization in developed countries, and the only major cause of mortality for which death rates continue to rise.
M AN U
One third of deaths in COPD relate to cardiovascular disease, the majority of which are classified as sudden.2 Previous studies have suggested an increased burden of ventricular arrhythmia in COPD, but have been limited in terms of sample size, stratification of COPD severity, and ability to adjust for confounding factors. The present study in part addresses these limitations and the classic Bradford Hill criteria for assessing causation,3 demonstrating for the first time a clear dose response
TE D
relationship between exposure and effect, independent of the foremost confounders (LVEF and coronary disease). This gradient was consistent for premature ventricular contractions,
EP
idioventricular rhythm and NSVT. Moreover, no significant relationship between severity of airflow obstruction and ambient heart rate was apparent, suggesting a lesser etiological role for
AC C
sympathetic tone and ventricular automaticity. Though pointing towards a causative relationship between COPD and VT, our interpretation must be tempered by considerations of study design, bias and confounding. First, the study employs airflow obstruction and non-sustained arrhythmia as surrogates for COPD and VT respectively. The former arguably impairs generalizability, given that fixed spirometric criteria over-diagnose airflow obstruction relative to lower limit of normal criteria.4 Notably, only 22% of patients with mild obstruction were prescribed beta-agonists, suggesting a significant proportion may not have a clinical diagnosis of COPD. Nevertheless, spirometry provides the objective measure central to
ACCEPTED MANUSCRIPT
defining the biological gradient, a key strength of the study. The adoption of NSVT as surrogate for ventricular arrhythmia is more challenging, as the authors acknowledge. While PVC burden and NSVT are associated with all-cause mortality in patients with structural heart disease, the relationship with sudden death is more complex, and suppression with anti-arrhythmic drug therapy
RI PT
has been unrewarding. Secondly, retrospective linkage of three separate and temporally dissociated investigations (spirometry, Holter and echocardiography) creates selection or indication bias whose significance is
SC
difficult to quantify, interpret or control. A reasonable assumption is that patients receiving all three tests are more likely to have pathological findings. Moreover, the impact of medical events between
M AN U
investigations (e.g. myocardial infarction occurring after a normal echocardiogram but before an abnormal Holter) is almost impossible to quantify in these circumstances. Considering again the Bradford Hill criteria, there is certainly biological plausibility and coherence between epidemiological and laboratory observations. The proarrhythmic milieu in patients with COPD includes hypoxia, acid-base and electrolyte disturbance, beta-agonists,
TE D
autonomic neuropathy, QTC prolongation, and underlying cardiac disease.5 Nevertheless, the inevitable question for all observational studies is whether unmeasured confounders may account
EP
for the observed relationship. Konecny et al. adjust for multiple factors, included LVEF as a continuous variable and presence of coronary disease. As alluded to earlier, the observed correlation
AC C
between severity of obstruction and arrhythmia burden infers a degree of causality. However, confounders themselves may also exhibit proportionality, such as ventricular remodelling in relation to cumulative smoking history,6 severity and control of risk factors (not simply presence/absence), extent of scar (LVEF being a relatively crude surrogate), and severity of underlying coronary disease.7 All these factors may be worse in more severe lung disease, contribute to increased ventricular arrhythmia, and be incompletely captured in standard models. These thoughts return to the first Bradford Hill criteria, namely strength of association.3 The magnitude of the unadjusted odds ratio, and the degree of attenuation with multivariable regression
ACCEPTED MANUSCRIPT
(odds ratio 1.33 [1.15-1.54]), suggests the relationship is relatively weak and in large part attributable to underlying cardiovascular disease (as opposed to the COPD disease state per se). Further characterization of unmeasured confounders, or the severity of cardiovascular disease, could well render the independent association non-significant. This raises a somewhat philosophical
RI PT
debate regarding adjusted versus actual risk. While research extols adjusted risk as a marker of causal inference, patients care about their absolute risk in unadjusted terms. Treatment decisions in clinical practice are also based on absolute thresholds defining high risk groups e.g. impaired LVEF
SC
for beta-blockers or implantable cardioverter defibrillators (ICDs). What is clear in this study, and many others, is that patients with COPD suffer a high burden of arrhythmia.
M AN U
The authors focus the discussion on ICD eligibility, posing the question whether COPD conveys sufficient risk to justify ICD therapy beyond established guideline criteria. Figure 3 of the analysis is particularly informative, depicting the incidence of NSVT stratified by both COPD severity and LVEF.1 In patients with no, mild, moderate or severe COPD, compared to those with normal LVEF, NSVT was consistently twice as common in patients with mid-range LVEF, and
TE D
three times more common in those with LVEF <40%. It may be helpful in future work to stratify the arrhythmic burden using an LVEF threshold of 35%, to determine the proportion of patients
EP
already eligible for ICDs. An equally pertinent question is how many patients fulfilling guideline criteria actually have an ICD implanted. COPD is a recognized barrier to care, with implantation
AC C
rates consistently lower in those with pulmonary disease. A common misconception is that higher baseline all-cause mortality risk precludes substantial benefit from ICD therapy. However, three key factors determine ICD benefit: annual total mortality; the ratio of arrhythmic to non-arrhythmic death; and time horizon of follow-up. Providing a sufficient proportion is arrhythmic in origin, the highest risk populations in fact derive significant gain.8 Referring once more to Figure 3, the challenge for COPD an a ‘standalone’ marker is that arrhythmia burden is only markedly elevated in very severe obstruction (FEV1 percent predicted <0.3). Not only does this group represent a small proportion of the overall population with COPD (6%, n=166), but the competing risk of non-cardiac
ACCEPTED MANUSCRIPT
death reduces the ratio of arrhythmic to non-arrhythmic mortality, and hence the lifespan-gain and cost-effectiveness of ICD therapy. Acknowledging this, possibly the simplest opportunity for improvement is evident in Table 1 of the analysis. Despite the high prevalence of coronary artery disease, arrhythmia burden and heart
RI PT
failure, beta-blocker utilization was only 29% in patients with mild through to severe airflow obstruction. Evidence has steadily amassed that beta-blockers, particularly cardio-selective, are safe, well tolerated, and associated with improved outcomes in patients with COPD and
SC
cardiovascular disease.5 There is also a consistent interaction between beta-blockers and betaagonists, the former attenuating the risk of sudden death and adverse cardiovascular events 9
To conclude, the thoughtful work of Konecny et al. is another
M AN U
associated with the latter.5,
important advance in our understanding of arrhythmic risk in pulmonary disease, but should also prompt us to consider how best to reduce this risk with the tools at our disposal.
TE D
References
Konecny T, Somers KR, Park JY, John A, Orban M, Doshi R, Scanlon PD, Asirvatham SJ, Rihal CS, Brady PA. Chronic obstructive pulmonary disease as a risk factor for ventricular arrhythmias independent of left ventricular function. Heart Rhythm 2017.
2.
McGarvey LP, John M, Anderson JA, Zvarich M, Wise RA. Ascertainment of cause-specific mortality in COPD: operations of the TORCH Clinical Endpoint Committee. Thorax 2007;62:411-415.
3.
Hill AB. The environment and disease: association or causation? Proc R Soc Med May 1965;58:295-300.
4.
Roberts SD, Farber MO, Knox KS, Phillips GS, Bhatt NY, Mastronarde JG, Wood KL. FEV1/FVC ratio of 70% misclassifies patients with obstruction at the extremes of age. Chest Jul 2006;130:200-206.
5.
Hawkins NM, Petrie MC, Macdonald MR, Jhund PS, Fabbri LM, Wikstrand J, McMurray JJ. Heart failure and chronic obstructive pulmonary disease the quandary of Beta-blockers and Beta-agonists. J Am Coll Cardiol 2011;57:2127-2138.
6.
Nadruz W, Jr., Goncalves A, Claggett B, Querejeta Roca G, Shah AM, Cheng S, Heiss G, Ballantyne CM, Solomon SD. Influence of cigarette smoking on cardiac biomarkers: the Atherosclerosis Risk in Communities (ARIC) Study. Eur J Heart Fail Jun 2016;18:629-637.
AC C
EP
1.
ACCEPTED MANUSCRIPT
Dursunoglu N, Dursunoglu D, Yildiz AI, Uludag B, Alacam ZN, Saricopur A. Severity of coronary atherosclerosis in patients with COPD. The clinical respiratory journal Nov 25 2015.
8.
Raphael CE, Finegold JA, Barron AJ, Whinnett ZI, Mayet J, Linde C, Cleland JGF, Levy WC, Francis DP. The effect of duration of follow-up and presence of competing risk on lifespan-gain from implantable cardioverter defibrillator therapy: Who benefits themost? European Heart Journal 2015;36:1676-1688.
9.
Narayanan K, Reinier K, Uy-Evanado A, Teodorescu C, Zhang L, Chugh H, Nichols GA, Gunson K, Jui J, Chugh SS. Chronic Obstructive Pulmonary Disease and Risk of Sudden Cardiac Death. JACC: Clinical Electrophysiology 2015;1:381.
RI PT
7.
SC
Funding
AC C
EP
TE D
M AN U
The author received no financial support in preparation of the manuscript.