Author + information
- Received May 17, 2014
- Revision received July 5, 2014
- Accepted July 8, 2014
- Published online November 11, 2014.
- Steven M. Kawut, MD, MS∗,
- Hooman D. Poor, MD†,
- Megha A. Parikh, MS†,
- Katja Hueper, MD‡,
- Benjamin M. Smith, MD, MS†,
- David A. Bluemke, MD, PhD§,
- João A.C. Lima, MD‖,
- Martin R. Prince, MD, PhD¶,
- Eric A. Hoffman, PhD#,
- John H.M. Austin, MD¶,
- Jens Vogel-Claussen, MD, PhD‡ and
- R. Graham Barr, MD, DrPH†∗∗∗ ()
- ∗Departments of Medicine and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania
- †Department of Medicine, College of Physicians and Surgeons, Columbia University, New York, New York
- ‡Department of Radiology, Hannover Medical School, Hannover, Germany
- §Radiology and Imaging Sciences, NIH Clinical Center, Bethesda, Maryland
- ‖Departments of Medicine and Radiology, Johns Hopkins University, Baltimore, Maryland
- ¶Department of Radiology, College of Physicians and Surgeons, Columbia University, New York, New York
- #Department of Radiology, University of Iowa, Iowa City, Iowa
- ∗∗Department of Epidemiology, Mailman School of Public Health, Columbia University, New York, New York
- ↵∗Reprint requests and correspondence:
Dr. R. Graham Barr, Presbyterian Hospital, 9 East 105, Columbia University Medical Center, 630 West 168th Street, New York, New York 10032.
Background The classic cardiovascular complication of chronic obstructive pulmonary disease (COPD) is cor pulmonale or right ventricular (RV) enlargement. Most studies of cor pulmonale were conducted decades ago.
Objectives This study sought to examine RV changes in contemporary COPD and emphysema using cardiac magnetic resonance (CMR) imaging.
Methods We performed a case-control study nested predominantly in 2 general population studies of 310 participants with COPD and control subjects 50 to 79 years of age with ≥10 pack-years of smoking who were free of clinical cardiovascular disease. RV volumes and mass were assessed using magnetic resonance imaging. COPD and COPD severity were defined according to standard spirometric criteria. The percentage of emphysema was defined as the percentage of lung regions <−950 Hounsfield units on full-lung computed tomography; emphysema subtypes were scored by radiologists. Results were adjusted for age, race/ethnicity, sex, height, weight, smoking status, pack-years, systemic hypertension, and sleep apnea.
Results Right ventricular end-diastolic volume (RVEDV) was reduced in COPD compared with control subjects (−7.8 ml; 95% confidence interval: −15.0 to −0.5 ml; p = 0.04). Increasing severity of COPD was associated with lower RVEDV (p = 0.004) and lower RV stroke volume (p < 0.001). RV mass and ejection fraction were similar between the groups. A greater percentage of emphysema also was associated with lower RVEDV (p = 0.005) and stroke volume (p < 0.001), as was the presence of centrilobular and paraseptal emphysema.
Conclusions RV volumes are lower without significant alterations in RV mass and ejection fraction in contemporary COPD, and this reduction is related to the greater percentage of emphysema on computed tomography.
Chronic obstructive pulmonary disease (COPD), the third leading cause of death in the United States (1), is defined by accelerated, age-related loss of lung function resulting in incompletely reversible airway obstruction (2). Overlapping partially with COPD, pulmonary emphysema is described as airspace enlargement distal to the terminal bronchioles with destruction of their walls. Hypoxic pulmonary vasoconstriction, hypercapnia, acidosis, and pulmonary vascular remodeling in COPD can all contribute to increased pulmonary vascular resistance and greater right ventricular (RV) afterload, resulting in cor pulmonale or RV hypertrophy and dilation (3).
Cor pulmonale was once considered common in COPD (4), with resting pulmonary hypertension seen frequently and exertional pulmonary hypertension occurring in as many as 58% of COPD patients without resting pulmonary hypertension (5), both of which contribute to increased RV mass and volume. However, the contemporary literature on cor pulmonale in COPD is limited by small sample sizes of highly selected patients, perhaps due to the difficulties in assessing the right ventricle using transthoracic echocardiography in COPD and performing cardiac catheterization in large cohorts. Some of these small studies demonstrated increased RV mass and RV dysfunction in COPD (6), whereas others showed reduced RV size and intrathoracic blood volumes (7–13). Hence, changes in RV structure and function in contemporary COPD remain poorly defined.
Therefore, in the MESA (Multi-Ethnic Study of Atherosclerosis) COPD study, we assessed RV morphology in COPD and emphysema in COPD patients and control subjects drawn predominantly from the general population undergoing cardiac magnetic resonance (CMR) imaging, the standard of reference for noninvasively assessing the RV. Given our previous findings showing a small right ventricle in patients with emphysema and historical autopsy studies (14–16), we hypothesized that COPD and a greater percentage of emphysema-like lung on computed tomography (CT) would be associated with lower RV volumes.
Between 2009 and 2011, the MESA COPD study enrolled participants with COPD and normal control subjects predominantly from a prospective population-based cohort study (MESA) (17) and a lung cancer screening and emphysema progression study (EMCAP [Emphysema and Cancer Action Project]) (18). Participants were 50 to 79 years of age with ≥10 pack-years of smoking and without a clinical diagnosis of coronary heart disease, heart failure, cerebrovascular disease, asthma before 45 years of age, other lung disease or cancer, previous lung resection, stages IIIb to V kidney disease, allergy to gadolinium, claustrophobia, metal in the body, pregnancy, or weight >300 lb. Recent COPD exacerbation was a temporary exclusion criterion. We selected all eligible participants at 4 sites in the MESA Lung Study (19) and oversampled participants with COPD or emphysema from the remainder of MESA and EMCAP, in addition to a small number from neither study.
Protocols were approved by the institutional review boards of the participating institutions and the National Heart, Lung, and Blood Institute. Written informed consent was obtained from all participants.
Pulmonary function testing
Spirometry was conducted in accordance with American Thoracic Society/European Respiratory Society guidelines (20) before and after inhalation of albuterol. COPD was defined as a post-bronchodilator ratio of the forced expiratory volume in 1 s (FEV1) to the forced vital capacity (FVC) <0.70 (2). COPD severity was classified as mild, FEV1 ≥80% predicted; moderate, FEV1 ≥50% to 79% predicted; and severe, FEV1 <50% predicted (21).
All participants underwent full-lung CT scans on 64-slice helical scanners following the MESA-Lung/SPIROMICS (Subpopulations and Intermediate Outcome Measures in COPD Study) full-inspiration protocol (22). The percentage of emphysema was defined using Apollo software (Vida Diagnostics, Coralville, Iowa) as the percentage of total voxels within the lung field that fell below −950 Hounsfield units (23). In addition, the presence or absence of emphysema and predominant emphysema subtype was assessed visually on all CT scans by an experienced thoracic radiologist using a standardized protocol (24) blinded to other clinical information.
Magnetic resonance imaging
The CMR protocol was that of the fifth examination of MESA modified to include assessment of the pulmonary vasculature (25); the protocol and methods for interpretation were previously reported (26,27). RV image analysis was performed by 2 independent analysts, who were unaware of other clinical information, using QMASS software version 4.2 (Medis, Leiden, the Netherlands). Right ventricular end-diastolic volume (RVEDV) and RV end-systolic volume were calculated using Simpson's rule by summation of areas on each slice multiplied by the sum of slice thickness and image gap. RV mass was determined at end-diastole as the difference between end-diastolic epicardial and endocardial volumes multiplied by the specific gravity of the heart (1.05 g/cm3). RV stroke volume was calculated by subtracting RV end-systolic volume from RVEDV; right ventricular ejection fraction (RVEF) was calculated by dividing RV stroke volume by RVEDV. Trabeculations were excluded from RV mass and included in RV volumes, as in previous large-scale studies using CMR to assess the ventricles (27). The RV outflow tract was included in the RV volumes, and the septum was considered part of the left ventricle for this study (Online Figure 1). Our method of interpretation has been shown to have high intra- and inter-reader reproducibility (28).
Pulmonary perfusion was assessed in a subset of participants (n = 141) using dynamic first-pass, contrast-enhanced CMR of the thorax. Pulmonary microvascular perfusion was assessed on a coronal slice at the level of the trachea in the peripheral 2 cm of the lung, as previously described (29,30).
See the Online Appendix for details of other pulmonary function measures, the CT and CMR protocols, and microvascular perfusion measures.
Age, sex, race/ethnicity, educational attainment, smoking status, pack-years, and medical history were self-reported. Medication use was assessed by medication inventory (31). Height, weight, and blood pressure were measured following standardized MESA protocols (25). Participants with a blood cotinine level >25 ng/ml, a urine level >500 ng/ml, or a self- report of current smoking were classified as current smokers. Oxygen saturation was measured with a pulse oximeter off oxygen, if used.
Data were summarized as mean ± SD or median (interquartile range), as appropriate. Linear regression models included categories of COPD status and severity as independent variables and continuous RV parameters as dependent variables. A test of trend across categories of COPD severity was performed; if positive, Holm’s step-down procedure was performed.
Linear regression models also were constructed for continuous independent variables, weighted according to cohort-specific probabilities of selection and enrollment into the MESA-COPD study to account for sampling (32). Robust SEs were used. Results are shown by quintiles of the percentage of emphysema for descriptive purposes; statistical tests were based on the continuous variable in the regression model. The nonlinearity of these associations was tested in generalized additive models with the same weighting and robust SEs.
The minimally adjusted models included age, sex, height, weight, race/ethnicity, and the cohort from which the participant was recruited. The fully adjusted models also included smoking status, pack-years, hypertension, and sleep apnea (and milliamperes for the percentage of emphysema). The full models were also then adjusted for the respective LV parameters to better understand whether the associations were right ventricle specific. There were no missing covariate data except for 10 participants (3%) who did not know whether they carried a diagnosis of sleep apnea; they were assumed not to have sleep apnea.
The p values were 2-tailed with statistical significance defined as p < 0.05. Analyses were performed with SAS version 9.2 (Cary, North Carolina) and R version 2.14.1 (Vienna, Austria).
The study enrollment and reasons for exclusion are shown in Online Figure 2. Of the 310 participants, 47% had COPD: 60 with mild, 67 with moderate, and 20 with severe disease. The mean age of the participants was 67.9 ± 7.0 years, 39.7% were women, and 27.7% were current cigarette smokers. The race/ethnic distribution was 53.2% non-Hispanic white, 26.1% African American, 14.2% Hispanic, and 6.5% Chinese American.
Patients with COPD were more likely to be male and white and somewhat more likely to be current smokers with more pack-years (Table 1). Those with COPD demonstrated characteristic spirometric, volumetric, and diffusing capacity for carbon monoxide patterns compared with smoking controls and had more emphysema. Only 7 of the 147 patients (4.7%) with COPD used long-term supplemental oxygen (all of whom had severe COPD).
COPD and the right ventricle
RVEDV was lower in patients with COPD compared with control subjects (p = 0.02) in minimally adjusted models that accounted for differences in age, sex, race/ethnicity, height, weight, and cohort. Results were similar in fully adjusted models (−7.8 ml; 95% confidence interval: −15.0 to −0.5 ml; p = 0.04).
Increasing severity of COPD was significantly and monotonically associated with lower RVEDV in minimally and fully adjusted models (Table 2). The decrement in RVEDV was directly related to the FEV1/FVC ratio (Figure 1A) without evidence of nonlinearity, such as a J shape (p for nonlinearity = 0.65). The association of COPD severity with lower RVEDV persisted after additional adjustment for LV end-diastolic volume (Online Table 1) (p = 0.02).
A similar association was present for RV end-systolic volume, although its significance was marginal in the fully adjusted model (Table 2, Online Table 1). RV stroke volume was significantly reduced with increasing severity of COPD.
RV mass was similar between patients with COPD and controls in minimally and fully adjusted models (−0.7 g; 95% confidence interval: −2.7 to 1.3; p = 0.51) and across categories of COPD severity (Table 2, Online Table 1). There was no evidence of nonlinearity in the association of the FEV1/FVC ratio and RV mass (Figure 1B) (p for nonlinearity = 0.13). The ratio of RV mass to RVEDV was slightly but significantly increased in severe COPD due to reduced volume; however, RVEF was not associated with the presence or severity of COPD (Table 2).
The association of RVEDV with COPD was largely unchanged with 1) further adjustment for long-term oxygen use, 2) restriction to participants sampled from MESA and EMCAP or without systemic hypertension, diabetes, or sleep apnea, and 3) stratification by sex (Online Figure 3). There were no differences in former smokers compared with current smokers (p for interaction = 0.48), but there may have been a more pronounced decrement in African-American participants compared with other race/ethnic groups (p for interaction = 0.09).
Pulmonary emphysema and the right ventricle
RVEDV was monotonically lower across increasing categories of percentage of emphysema on CT. This association was significant in minimally and fully adjusted models (Table 3, Online Table 2) and was linear (Figure 2A). Increasing percentage of emphysema was also associated with a significantly lower RV stroke volume. The percentage of emphysema was not associated with RV mass (Figure 2B), RV mass/RVEDV ratio, or RVEF (Table 3, Online Table 2).
The presence of emphysema assessed by radiologist interpretation was also associated with lower RVEDV and RV stroke volume (p = 0.003 and 0.004, respectively) in fully adjusted models. Whereas centrilobular emphysema and paraseptal emphysema were significantly related to lower RVEDV (p < 0.001 and p = 0.003, respectively), there was no evidence of such a relationship for panlobular emphysema (p = 0.39) (Online Table 3).
The findings for percentage of emphysema and RVEDV were not sensitive to additional adjustment or subset analyses (Online Figure 4). We found no interaction between percentage of emphysema and smoking status or race in terms of the relationship with RVEDV (p for interaction = 0.78 and 0.18, respectively).
Among participants with gadolinium-enhanced CMR (n = 141) (Online Table 4), lower RVEDVs were associated with reduced pulmonary microvascular blood flow and reduced pulmonary microvascular blood volume (Online Table 5). Higher diffusing capacity for carbon monoxide was associated with larger RVEDV, RV end-systolic volume, and RV stroke volume. We saw no association between lung volumes and RV parameters (Online Table 6). Subset analysis with only MESA and EMCAP participants showed similar results (data not shown).
For all of the models, we performed routine checks for model fit to make sure the conclusions were not based on a few “influential” subjects. These procedures provided reassurance that the models had adequate fit. Adjustment for arterial oxygen saturation by pulse oximetry did not change the results (data not shown).
Patients with COPD displayed, on average, lower RV volumes compared with control subjects in this multicenter study that used CMR to assess RV structure and function. More severe COPD was associated with lower RV volumes, but there was no consistent relationship of COPD severity with RV mass. Furthermore, the presence and severity of emphysema on CT scan were associated with lower RV volumes and no change in RV mass.
The finding of reduced RV volumes contradicts the classic paradigm of cor pulmonale in COPD; that is, increased RV mass and volume in the setting of lung disease with or without RV dysfunction (3,33). Our results for RV volumes are consistent with those of a recent small study that showed lower RV volumes but greater RV mass in 25 patients with COPD and mild hypoxemia (12); other investigations in COPD (all with <20 patients) have demonstrated similar findings (9–11). These studies, however, included only a few highly select patients and did not adjust for potential confounders, issues that were addressed in this study.
A previous study in the parent MESA cohort showed that a greater percentage of emphysema in the lower two-thirds of the lungs was associated with lower RV volumes and mass, specifically in smokers, with more complicated relationships after adjustment for the respective LV measures (16). A lower FEV1/FVC ratio was associated with lower RV mass after adjustment for LV mass in that study, but in the current study, increasing COPD severity was linked to increased RV mass/RVEDV ratio, suggesting increased RV afterload. However, this association was attributable to lower RVEDV in more severe COPD patients (rather than increasing RV mass), which is difficult to attribute to increased RV wall stress.
Our findings may differ from the traditional manifestation of increased RV volumes and increased RV mass in COPD for several reasons. First, we sampled participants predominantly from the general population, resulting in a sample of patients whose COPD may be less severe than those in previous studies but making the results more generalizable. Our sample size of patients with severe COPD was similar to (or larger than) those in most previous studies. While there was a suggestion of a J-shaped relationship of RV mass with COPD severity, there was no such suggestion for RVEDV or in relation to the percentage of emphysema. Hence, the disease spectrum alone does not account for these differences.
Another explanation may be the predominant COPD subphenotypes studied. Burrows et al. (14) found that COPD patients with emphysema on chest radiography were less likely to demonstrate RV hypertrophy on electrocardiography than other COPD patients, given the same pulmonary vascular resistance. Furthermore, early autopsy studies showed that patients dying of emphysema did not have evidence of RV hypertrophy, which was more common in COPD patients with airway inflammation and mucus gland hypertrophy suggestive of chronic bronchitis (15). More recently, increasing airway wall size was more strongly associated with increased pulmonary artery diameter in COPD than was the percentage of emphysema (34). Other studies of the “comorbidome” of COPD have suggested that cardiac complications are linked to more prominent airway disease and less parenchymal destruction, supporting a stronger link between the “chronic bronchitis” subphenotype and traditional cardiac sequelae than the “emphysema” subphenotype (35,36). We showed that increasing centrilobular and paraseptal emphysema were associated with lower RV volumes, whereas panlobular emphysema was not. Therefore, rather than being inconsistent with the classic paradigm of cor pulmonale, our findings may reflect the current phenotype of COPD in the general population in the United States and may not apply to selected patients with severe chronic bronchitis or marked gas trapping.
Consistent with this, Hilde et al. (6) recently published a study of Norwegian COPD patients with residual volume of ∼200% of predicted normal undergoing echocardiography and right heart catheterization. These patients demonstrated increased RV mass (without correction for body size) and more RV dysfunction compared with smoking controls. In contrast, patients in our study generally had normal residual volumes, which were not significantly associated with RV morphology. The different mix of subphenotypes of COPD may explain these population differences.
Third, treatments for COPD may secondarily affect afterload and preload and the RV sequelae. Long-acting bronchodilators reduce hyperinflation with gas trapping and supplemental oxygen could affect pulmonary vascular disease and RV morphology. We were not able to test these hypotheses; however, the very small number of patients in our study receiving long-acting bronchodilators and oxygen (<5%) and the persistence of the results despite adjustment for supplemental oxygen make use of these therapies unlikely to explain the results.
The mechanism of reduced RV filling in emphysematous COPD may relate to several factors (Central Illustration). Watz et al. (7) suggested that pulmonary hyperinflation reduces right atrial and RV filling in moderate to severe COPD and lung volume reduction surgery for very severe COPD (which decreases hyperinflation) (37,38) is associated with increased oxygen pulse (39,40), RVEDV, and RV stroke volume, and reduced atrial pressure (41). However, measures of hyperinflation were not associated with RV volumes in the current study. Other mechanisms that reduce blood return to the thorax in severe COPD, affecting filling of both ventricles and the pulmonary vasculature, include a reduced gradient in pressure between the abdomen and chest and abnormal diaphragm orientation compressing the vena cava (42). Alternatively, intrinsic RV diastolic dysfunction, possibly related to increased inflammation or endothelial dysfunction, could decrease RV filling (43,44). These novel mechanisms may suggest innovative approaches that target diastolic function (e.g., soluble guanylate cyclase activators) to treat the cardiac component of exercise limitation in these patients.
Although we included a modest number of patients with severe COPD, the findings appeared to be similar to those with mild and moderate COPD. We excluded patients with clinically apparent heart failure, potentially leading to selection bias. However, the number of participants excluded for heart failure was very small, lessening the impact on the results. Assessment of RV morphology can be challenging due to the presence of the tricuspid valve; however, previous studies using the same methodology in our Reading Center have shown outstanding intra- and inter-reader reliability for RV measurements. Measurement error (unless differential by COPD status) would bias to the null, so that our findings may underestimate the actual differences in RV morphology between COPD and control subjects. The percentage of emphysema measured quantitatively reflects pulmonary emphysema and other factors; nevertheless, radiologist-determined emphysema showed similar findings. We did not acquire invasive hemodynamic measures, which would be difficult to justify in a large population-based sample, or transthoracic echocardiography, which is technically challenging in patients with COPD, our main population of study. Our perfusion measures may have certain technical limitations due to the determination of the arterial input function and the bolus dose. Although we adjusted for most potential confounders in multivariate analyses, residual confounding in multivariable analyses is possible.
In the MESA COPD study, more severe COPD and emphysema on CT were associated with lower RV volumes without changes in RVEF or RV mass. Smaller rather than larger RV size appears to be the more common RV phenotype in COPD in the U.S. general population without overt cardiovascular disease, perhaps better termed “cor pulmonale parvus.” Future studies of COPD treatment should consider the role of this novel cardiac comorbidity.
COMPETENCY IN MEDICAL KNOWLEDGE: More severe emphysema in patients with COPD is associated with smaller RV dimensions.
TRANSLATIONAL OUTLOOK: Further studies are needed to elucidate the mechanisms by which changes in right ventricular dimensions and function are related to the morbidity and mortality associated with COPD.
The authors thank the investigators, staff, and participants of the MESA COPD study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org. This manuscript was reviewed by the MESA investigators for scientific content and consistency of data interpretation with previous MESA publications, and significant comments were incorporated before submission for publication.
For supplemental material including tables and figures, please see the online version of this article.
The MESA COPD study is supported by the National Institutes of Health R01-HL093081, R01-HL077612, and R01-HL075476. MESA is supported by N01-HC95159-HC95169 and UL1-RR024156. Dr. Kawut is supported by K24-HL103844 and R01-HL086719. Dr. Hoffman is the founder of and a shareholder in Vida Diagnostics Inc. Dr. Prince has patent agreements with GE, Siemens, Philips, Toshiba, Hitachi, Bayer HealthCare, Bracco, Mallinkrodt, Medrad, and Lantheus. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- cardiac magnetic resonance
- chronic obstructive pulmonary disease
- computed tomography
- forced expiratory volume in 1 s
- forced vital capacity
- left ventricular
- right ventricular
- right ventricular end-diastolic volume
- right ventricular ejection fraction
- Received May 17, 2014.
- Revision received July 5, 2014.
- Accepted July 8, 2014.
- American College of Cardiology Foundation
- Heath D.,
- Brewer D.B.,
- Hicken P.
- Hilde J.M.,
- Skjorten I.,
- Grotta O.J.,
- et al.
- Turnbull L.W.,
- Ridgway J.P.,
- Biernacki W.,
- et al.
- Bild D.E.,
- Bluemke D.A.,
- Burke G.L.,
- et al.
- Hankinson J.L.,
- Kawut S.M.,
- Shahar E.,
- Smith L.J.,
- Stukovsky K.H.,
- Barr R.G.
- Celli B.R.,
- MacNee W.
- Couper D.,
- LaVange L.M.,
- Han M.,
- et al.,
- for the SPIROMICS Research Group
- Smith B.M.,
- Austin J.H.M.,
- Newell J.D.,
- et al.
- ↵Exam 5 field center procedures manual of operations. MESA–NHLBI, 2010. Available at: http://www.Mesa-nhlbi.Org/publicdocs/2011/mesae5_mopjanuary2011.pdf. Accessed August 27, 2014.
- Kawut S.M.,
- Lima J.A.,
- Barr R.G.,
- et al.
- Hueper K.,
- Parikh M.A.,
- Prince M.R.,
- et al.
- Psaty B.M.,
- Lee M.,
- Savage P.J.,
- Rutan G.H.,
- German P.S.,
- Lyles M.
- Smith B.M.,
- Kawut S.M.,
- Bluemke D.A.,
- et al.
- World Health Organization
- Kawut S.M.
- Lammi M.R.,
- Ciccolella D.,
- Marchetti N.,
- Kohler M.,
- Criner G.J.
- Nakhjavan F.K.,
- Palmer W.H.,
- McGregor M.