Influence of Left Ventricular Hypertrophy on In-Hospital Outcomes in Acute Exacerbation of Chronic Obstructive Pulmonary Disease

ABSTRACT

Left ventricular hypertrophy (LVH) is associated with worse outcomes in chronic obstructive pulmonary disease (COPD); however, its role in an acute exacerbation of COPD (AECOPD) has not been reported. This was a retrospective cohort study during 2008–2012 at an academic medical center. AECOPD patients >18 years with available echocardiographic data were included. LVH was defined as LV mass index (LVMI) >95 g/m² (women) and >115g/m² (men). Relative wall thickness was used to classify LVH as concentric (>0.42) or eccentric (<0.42). Outcomes included need for and duration of non-invasive ventilation (NIV) and mechanical ventilation (MV), NIV failure, intensive care unit (ICU) and total length of stay (LOS), and in-hospital mortality. Two-tailed p < 0.05 was considered statistically significant. Of 802 patients with AECOPD, 615 patients with 264 (42.9%) having LVH were included. The LVH cohort had higher LVMI (141.1 ± 39.4 g/m² vs. 79.7 ± 19.1 g/m²; p < 0.001) and lower LV ejection fraction (44.5±21.9% vs. 50.0±21.6%; p ≤ 0.001). The LVH cohort had statistically non-significant longer ICU LOS, and higher NIV and MV use and duration. Of the 264 LVH patients, concentric LVH (198; 75.0%) was predictive of greater NIV use [82 (41.4%) vs. 16 (24.2%), p = 0.01] and duration (1.0 ± 1.9 vs. 0.6 ± 1.4 days, p = 0.01) compared to eccentric LVH. Concentric LVH remained independently associated with NIV use and duration. In-hospital outcomes in patients with AECOPD were comparable in patients with and without LVH. Patients with concentric LVH had higher NIV need and duration in comparison to eccentric LVH.

KEYWORDS:

• Acute exacerbation of chronic obstructive pulmonary disease
• left ventricular hypertrophy
• left ventricular mass index
• concentric hypertrophy
• non-invasive ventilation

Introduction

Chronic obstructive pulmonary disease (COPD) is primarily a respiratory disease characterized by heterogeneous inflammatory changes and obstructive airway physiology (1,2). Patients experiencing an acute exacerbation of COPD (AECOPD) are often managed in the hospital, commonly in the intensive care unit (ICU), leading to significantly higher health care costs and morbidity (1–3). However, the adverse outcomes and mortality in these patients are often due to cardiovascular dysfunction rather than pulmonary causes (2–4). Prior studies, dating as early as 1970s, have demonstrated concomitant left ventricular hypertrophy (LVH) in COPD. However, it has been assumed to be a consequence of coronary artery disease and other chronic cardiovascular co-morbidities (2,3,5). Recent evidence has demonstrated that COPD results in chronic low-grade systemic inflammation with activation of the renin-angiotensin-aldosterone system (RAAS) resulting in systemic end-organ damage, including LVH (1). Increase in LV mass has been demonstrated to result in worse long-term outcomes in stable COPD, independent of right ventricular (RV) structure and function (2,6,7). In 2013, Anderson and colleagues (3) demonstrated an independent correlation between LVH and COPD.

Patients with COPD have significantly worse cardiac function during AECOPD due to: (a) changes in intra-thoracic pressures and dynamic hyperinflation (7); (b) acute or chronic hypoxemia-induced acute pulmonary circulation dysfunction and cor pulmonale with a rise in B-type natriuretic peptide (BNP) levels (8); and (c) acute left heart failure in 20–30% patients, which is either a trigger or a consequence of AECOPD (9). Due to these changes, patients with AECOPD are significantly different from a pathophysiological standpoint in comparison to those with stable COPD. All prior studies focusing on the role of LVH evaluated patients with stable COPD (2,3,10); however, the role of LVH in AECOPD patients has not been elucidated. Short and colleagues demonstrated that LVH was not associated with increased risk of hospital admissions; however, this study did not differentiate AECOPD vs. other causes for admission (10). To address this knowledge gap, we sought to evaluate the influence of LVH on in-hospital outcomes in patients admitted with AECOPD.

Materials and methods

This was a 5-year retrospective study from January 1, 2008 to December 31, 2012 at a tertiary care academic medical center. The Creighton University Biomedical Institutional Review Board approved the study and waived the need for informed consent due to its retrospective nature. All research activities were conducted in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki). All adult patients (>18 years) admitted to ICU, stepdown unit, and general medical floors with a primary diagnosis of AECOPD were included in our study. The diagnosis of AECOPD was based on clinical evaluation, vital signs (specifically peripheral capillary oxygen saturation), arterial blood gas analysis, chest radiography, and evaluation for suspected pulmonary infection by Anthonisen criteria. All patients were required to have echocardiographic data from the index hospital stay. Pregnant female patients and patients without calculable LV mass index (LVMI) data were excluded from our study. Since LVH is a known consequence of aortic and mitral valve pathology, we excluded patients with moderate or greater aortic stenosis/regurgitation and/or mitral regurgitation. This study extracted data for our cohort of interest from a larger data set that was previously collected and analyzed for other purposes (4).

Patient information on demographics, co-morbidities, home medications, clinical examination, pulmonary function testing (PFTs), echocardiographic variables, and laboratory parameters were abstracted from the medical records. “A priori” selected outcomes included need for non-invasive ventilation (NIV), invasive mechanical ventilation (MV), failure of NIV, reintubation at 48 hours, duration of NIV and/or MV, length of stay (LOS) in the ICU, total LOS, and in-hospital mortality. Failure of NIV was defined as continued hypoxemia, severe respiratory acidosis, and/or signs of clinical distress. In an attempt to control for factors that can contribute to LVH either independently or in combination (2), we performed an exploratory analysis of AECOPD patients without prior chronic kidney disease (CKD), hypertension, and diabetes mellitus type II (DM-2).

Echocardiographic analysis

The echocardiogram was obtained during the index hospital admission. Detailed information on LV dimensions, LV ejection fraction (LVEF), and semi-quantitative RV size and function was collected. The LVEF was calculated by Simpson's biplane method or visual estimation method. Diastolic dysfunction was classified by a semi-quantitative method into grades I (mild), II (moderate), and III (severe). RV systolic pressure (RVSP) was extrapolated from the tricuspid regurgitant jet velocity and right atrial pressure.

The LV dimensions of interventricular septal thickness, posterior wall thickness, and end-diastolic diameter were measured at end of diastole with M-mode. The LV mass was calculated using the cube formula using M-mode in the linear method. LVH was defined as LVMI >95 g/m² and >115 g/m² in women and men, respectively, which represent the upper limit of normal using the American Society of Echocardiography criteria (11,12). Relative wall thickness (RWT) was calculated as (2 × posterior wall thickness)/(LV internal diameter at end diastole) and was used to classify LVH as concentric (RWT > 0.42) vs. eccentric (RWT ≤ 0.42) (12).

Statistical analysis

Continuous data are presented as mean ± standard deviation (SD) and categorical data are presented as totals (percentages). All variables were assessed for normality of distribution prior to statistical analysis. When appropriate, the concomitant non-parametric test was used for analysis. The outcomes of total LOS, ICU days, MV days, and NIV days were recorded as whole days, and thus treated as counts and were compared using unpaired t-test or Mann–Whitney U test. The outcomes of need for NIV, need for MV, NIV failure, reintubation at 48 hours, and in-hospital mortality were treated as categorical variables and were compared using Chi-square test or Fisher's exact test.

Multivariate logistic regression analysis was performed to evaluate the influence of concentric hypertrophy on the categorical clinical outcomes of NIV, MV, NIV failure, reintubation at 48 hours, and in-hospital mortality. Linear regression was performed for the continuous variables of NIV days, MV days, ICU LOS, and total LOS. Independent variables of age, sex, hyperlipidemia, coronary artery disease, BNP, and LVEF were selected after univariate and correlational analysis. Odds ratio (OR) with their corresponding 95% confidence interval (CI) was used to report categorical variables in univariate and multivariate analysis. Estimate (Est.) and standard error (SE) was used to report multivariate linear regression analysis. Two-tailed p < 0.05 was considered statistically significant. All statistical analyses were performed with JMP version 9.0.1 (SAS Institute, Cary, NC).

A part of the study findings were presented as a slide presentation and published as an abstract as a part of the proceedings of CHEST 2015, American College of Chest Physicians, Montreal QC (October 2015) (13).

Results

From January 2008 to December 2012, a total of 802 patients with AECOPD were admitted to our institution. Of these patients, 104 (13.0%) had insufficient data to calculate LVMI (absent or incomplete echocardiogram) and 83 (10.4%) had moderate or greater aortic stenosis/regurgitation and/or mitral regurgitation and were subsequently excluded. The final cohort comprising 615 (76.6%) patients formed our study subjects with LVH noted in 264 (42.9%) patients. Detailed baseline characteristics of patients with and without LVH are presented in Table 1. The cohort of patients with LVH had a higher prevalence of hypertension, DM-2, hyperlipidemia, and coronary artery disease. Renal dysfunction as manifest by CKD, end-stage renal disease, and ongoing need for hemodialysis was higher in the LVH cohort. Patients with LVH had higher use of cardiovascular medications and lower use of pulmonary medications. PFTs were available for 291 (47.3%) patients. Between the cohorts, forced expiratory volume at 1 second (FEV₁)/diffusion capacity of the lung for carbon monoxide (DLCO) data were available for 126 (47.7%)/102 (38.6%) and 165 (47.0%)/126 (35.9%) patients, respectively. Mean BNP and creatinine were significantly higher in the LVH cohort. On echocardiography, the LVH cohort had a higher LVMI (141.1 ± 39.4 g/m² vs. 79.7 ± 19.1 g/m²; p < 0.001) and lower LVEF (44.5 ± 21.9% vs. 50.0 ± 21.6%; p < 0.001). Diastolic dysfunction was calculable for 207/264 (78.4%) and 264/351 (75.2%) patients in the LVH and no LVH cohorts. Diastolic dysfunction grade II and above was higher in the LVH cohort [63/207 (30.4%) vs. 29/264 (11.0%); p < 0.001]. Detailed echocardiographic parameters for the total cohort are presented in Table 1.

Table 1. Baseline patient characteristics.

Parameter	LVH (n = 264)	No LVH (n = 351)	p-value
Male sex	125 (47.3)	181 (51.6)	0.33
Age (in years)	65.2 ± 12.1	64.2 ± 11.5	0.29
Co-morbidities
HTN	212 (80.3)	253 (72.1)	0.02
DM-2	111 (42.0)	103 (29.3)	0.001
Hyperlipidemia	133 (50.4)	119 (33.9)	<0.001
CAD	118 (44.7)	113 (32.2)	0.002
Home oxygen	98 (37.1)	142 (40.5)	0.41
CKD	68 (25.8)	55 (15.7)	0.002
ESRD	13 (4.9)	2 (0.6)	0.001
Hemodialysis	13 (4.9)	2 (0.6)	0.001
OSA	62 (23.5)	60 (17.1)	0.05
Current smoking	151 (57.2)	186 (53.0)	0.28
Past smoking	152 (57.6)	224 (63.8)	0.13
Pack-years	51.6 ± 34.7	49.5 ± 31.2	0.96
Pulmonary function tests
FEV1 (%)	40.7 (27.2–57.2)	38.0 (25.0–55.4)	0.24
DLCO (%)	42.2 (21.0–54.2)	36.3 (22.0–50.1)	0.40
Home medications
LABA	139 (52.7)	215 (61.3)	0.01
LAAC	107 (40.5)	187 (53.3)	0.002
ICS	147 (55.7)	214 (61.0)	0.22
ACE-i/ARB	126 (47.7)	125 (35.6)	0.003
Beta-blockers	128 (48.5)	120 (34.2)	<0.001
Loop diuretics	127 (48.1)	113 (32.2)	<0.001
Statins	131 (49.6)	121 (34.5)	<0.001
Physical examination
HR (beats/minute)	94.5 ± 19.3	97.9 ± 22.0	0.02
SBP (mm Hg)	136.1 ± 26.0	130.0 ± 28.1	0.004
BMI (kg/m^2)	30.5 ± 10.0	28.1 ± 10.3	0.001
SpO2 (%)	94.1 ± 9.6	94.0 ± 8.7	0.05
Laboratory parameters
Hemoglobin (g/dL)	12.4 ± 2.4	12.8 ± 2.2	0.09
pH	7.30 ± 0.48	7.36 ± 0.17	0.14
pCO2 (mm Hg)	58.6 ± 21.9	58.2 ± 28.6	0.24
BNP (pg/mL)	607.6 ± 912.8	299.1 ± 499.3	<0.001
Creatinine (mg/dL)	1.3 ± 1.1	1.0 ± 1.0	<0.001
GFR (ml/min/1.73 m^2)	46.5 ± 88.9	68.1 ± 152.4	0.19
Echocardiographic parameters
LVMI (g/m^2)	141.1 ± 39.4	79.7 ± 19.1	<0.001
LVEF (%)	44.5 ± 21.9	50.0 ± 21.6	<0.001
DD (Grades I–III)	187/207 (90.3)	242/264 (91.7)	0.63
LAD (cm)	4.2 ± 0.8	3.9 ± 0.8	<0.001
Abnormal RV size	75 (28.4)	113 (32.2)	0.33
Abnormal RV function	52 (19.9)	63 (18.4)	0.68
RVSP (mm Hg)	24.6 ± 26.5	22.2 ± 27.2	0.23
RAP (mm Hg)	8.6 ± 6.4	10.4 ± 7.6	0.04

Represented as: Mean ± standard deviation, median (interquartile range) or number (percentage).

ACE-i, angiotensin-converting enzyme inhibitors; ARB, angiotensinogen receptor blockers; BMI, body mass index; BNP, B-type natriuretic peptide; CAD, coronary artery disease; CKD, chronic kidney disease; DD, diastolic dysfunction; DLCO, diffusion capacity of the lung for carbon monoxide; DM-2, diabetes mellitus, type II; ESRD, end-stage renal disease; FEV₁, forced expiratory volume at 1 second; GFR, glomerular filtration rate; HR, heart rate; HTN, hypertension; ICS, inhaled corticosteroids; LAAC, long-acting anti-cholinergics; LABA, long-acting beta-agonists; LAD, left atrial diameter; LVEF, left ventricular ejection fraction; LVH, left ventricular hypertrophy; LVMI, left ventricular mass index; OSA, obstructive sleep apnea; pCO₂, partial pressure of carbon dioxide; PFTs, pulmonary function tests; RAP, right atrial pressure; RVSP, right ventricular systolic pressure; SBP, systolic blood pressure; SpO₂, peripheral capillary oxygen saturation.

Clinical outcomes

In the total cohort, NIV and MV were used in 215 (35.0%) and 64 (10.4%) patients with mean duration of 0.9 ± 1.8 and 0.7 ± 2.4 days, respectively. NIV failure was noted in 39 (6.3%) patients with 6 (1.0%) patients requiring reintubation within 48 hours. Mean ICU LOS and total LOS were 1.4 ± 3.1 and 5.5 ± 5.2 days. In-hospital mortality was noted in 22 (3.6%) patients. Detailed outcomes between cohorts are presented in Table 2. The LVH cohort had higher use of NIV and MV with longer duration of both modalities, higher rates of reintubation, and longer ICU LOS in comparison to patients without LVH; however, these did not attain statistical significance. For the categorical outcomes, LVMI was not a significant influence with receiver operating characteristic (ROC) curves demonstrating area under the curve (AUC) of 0.46–0.58. AUC for reintubation at 48 hours ROC curve was 0.70; however, there were less number of absolute events (6/615; 0.1%).

Table 2. Clinical outcomes classified by LVH.

Clinical outcomes	LVH (n = 264)	No LVH (n = 351)	OR (95% CI)	p-value
NIV	98 (37.1)	117 (33.3)	1.2 (0.9–1.7)	0.35
MV	30 (11.4)	34 (9.7)	1.2 (0.7–2.0)	0.51
NIV days	0.9 ± 1.8	0.8 ± 1.8	–	0.38
MV days	0.8 ± 2.8	0.6 ± 2.2	–	0.60
NIV failure	12 (4.6)	27 (7.7)	0.6 (0.3–1.2)	0.13
Reintubation 48	5 (1.9)	1 (2.9)	6.8 (0.8–58.2)	0.09
ICU LOS (days)	1.5 ± 3.5	1.3 ± 2.8	–	0.92
Total LOS (days)	5.4 ± 4.9	5.5 ± 5.4	–	0.69
In-hospital mortality	7 (2.7)	15 (4.3)	0.6 (0.3–1.5)	0.38

Represented as: mean ± standard deviation or number (percentage).

CI, confidence interval; ICU, intensive care unit; LOS, length of stay; LVH, left ventricular hypertrophy; MV, mechanical ventilation; NIV, non-invasive ventilation; OR, odds ratio; Reintubate 48, reintubation rates at 48 hours.

Table 3. Clinical outcomes in LVH patients based on type of LVH.

Clinical outcomes	Concentric LVH (n = 198)	Eccentric LVH (n = 66)	OR (95% CI)	p-value
NIV	82 (41.4)	16 (24.3)	2.2 (1.2–4.2)	0.01
MV	22 (11.1)	8 (12.1)	0.9 (0.4–2.2)	0.83
NIV days	1.0 ± 1.9	0.6 ± 1.4	–	0.01
MV days	0.9 ± 3.1	0.5 ± 1.5	–	0.80
NIV failure	11 (5.6)	1 (1.5)	3.8 (0.5–30.2)	0.31
Reintubation 48	3 (1.5)	2 (3.0)	0.5 (0.1–3.0)	0.60
ICU LOS (days)	1.6 ± 3.7	1.0 ± 2.6	–	0.12
Total LOS (days)	5.5 ± 4.6	5.1 ± 5.8	–	0.20
In-hospital mortality	4 (2.0)	3 (4.6)	0.4 (0.1–2.0)	0.37

Represented as: mean ± standard deviation, number (percentage) or odds ratio (95% confidence interval).

CI, confidence interval; ICU, intensive care unit; LOS, length of stay; LVH, left ventricular hypertrophy; MV, mechanical ventilation; NIV, non-invasive ventilation; OR, odds ratio; Reintubate 48, reintubation rates at 48 hours.

Concentric vs. eccentric LVH

In the total LVH cohort of 264 patients, concentric LVH (RWT > 0.42) was noted in 198 (75%) patients. Detailed baseline characteristics of patients with concentric and eccentric LVH are presented in Supplementary Table 1. Mean LVMI was not different between groups 139.5 ± 37.9 (concentric) vs. 146.0 ± 43.4 g/m² (eccentric) (p = 0.25). Presence of concentric LVH was predictive of higher need for NIV [82 (41.4%) vs. 16 (24.2%) patients, p = 0.01] and longer duration of NIV (1.0 ± 1.9 vs. 0.6 ± 1.4 days, p = 0.01). All other outcomes were not statistically different between the concentric and eccentric LVH groups as depicted in Table 3. In a multivariate regression analysis, as demonstrated in Table 4, concentric LVH was strongly associated with need for NIV [OR 2.5 (95% CI 1.3–5.1), p = 0.006] and duration of NIV [Est. 0.32 (SE 0.13), p = 0.02]

Table 4. Multivariable regression analysis of clinical outcomes based on type of LVH.

Categorical outcomes	OR (95% CI)	p-value
NIV	2.5 (1.3–5.1)	0.006
MV	0.8 (0.3–2.0)	0.55
NIV failure	3.2 (0.6–61.6)	0.22
Reintubation 48	0.4 (0.1–3.3)	0.34
In-hospital mortality	0.4 (0.1–2.5)	0.32
Continuous outcomes	Est. (SE)	p-value
NIV days	0.32 (0.13)	0.02
MV days	0.08 (0.21)	0.70
ICU LOS	0.17 (0.26)	0.52
Total LOS	−0.15 (0.36)	0.67

Represented as: Odds ratio (95% confidence interval) or estimate (standard error).

CI, confidence interval; Est., estimate; ICU, intensive care unit; LOS, length of stay; LVH, left ventricular hypertrophy; MV, mechanical ventilation; NIV, non-invasive ventilation; OR, odds ratio; Reintubate 48, reintubation rates at 48 hours; SE, standard error.

LVH without CKD, hypertension, and DM-2

In the cohort of AECOPD patients without CKD, hypertension, and DM-2, we had a total of 100 patients, with 31 (31%) demonstrating LVH. Patients with LVH had a longer duration of NIV use, MV use, ICU LOS, and total hospital LOS, but these were not statistically significant between groups (all p > 0.05). Detailed baseline variables and clinical outcomes are presented in Supplementary Tables 2 and 3.

Discussion

The major findings of our study are as follows: (a) patients with AECOPD with co-existing LVH form a high-risk cohort with greater co-morbidities, higher BNP level, and higher use of cardiovascular medications; (b) patients presenting with AECOPD with co-existing LVH have worse cardiac status as evidenced by lower LVEF, greater diastolic dysfunction, higher BNP, and larger LA diameter without significant differences in RV function; (c) LVH did not influence in-hospital outcomes in patients admitted with AECOPD in our patients; (d) after excluding alternate contributors to LVH such as hypertension, DM-2, and CKD, LVH was not predictive of worse in-hospital outcomes in AECOPD; and (e) concentric LVH was independently associated with increased need for NIV [OR 2.5 (95% CI 1.3–5.1), p = 0.006] and duration of NIV [Est. 0.32 (SE 0.13), p = 0.02)].

Our study is the first to evaluate the influence of LVH in patients presenting with AECOPD. Despite knowledge about the association of LVH with COPD since the 1970s (5), only recently has there been an interest in its role in COPD outcomes (2,3,6,7,10). The change in pleural and intra-thoracic pressures associated with COPD contributes to increased LV afterload and LV wall stress, resulting in subsequent hypertrophy (2,6). This increase in LV wall stress, independent of ventricular interdependence, results in concentric LV remodeling and contributes to a compensatory hypertrophy. LVH is second only to age as a predictor of adverse cardiovascular outcomes and is related to the development of sub-endocardial ischemia, myocardial fibrosis, left atrial dilatation, atrial fibrillation, diastolic dysfunction, heart failure with preserved ejection fraction, and sudden cardiac death (2,3,6,14,15). Smith et al. demonstrated that an increase of 0.71 L in pulmonary residual volume was associated with 7.2 g increase in LV mass on cardiac magnetic resonance (CMR) (2). Additionally, the release of systemic cytokines and stimulation of the RAAS is hypothesized to play a role in the chronic low-grade systemic inflammation in COPD (1,16). During an acute exacerbation, these inflammatory mediators are released in higher amounts contributing to worse in-hospital outcomes (16). Inhibitors of the RAAS that also work to attenuate myocardial hypertrophy such as ACE-i and ARB have been hypothesized to play a role in improving outcomes in COPD via multiple mechanisms (1,3). The combination of ACE-i/ARB with statin therapy has shown decreased mortality in COPD; however, statin therapy alone has not demonstrated benefit (16–18). Despite conventional theories of questionable harm in COPD, beta-blocker agents have been shown to be beneficial by virtue of their action on the myocardium (3,19). In our study, patients with LVH had a significantly higher use of ACE-i/ARBs, beta-blockers, and statins, which could have potentially neutralized the deleterious effects of LVH in AECOPD patients, decreasing the differences between the two groups. Consistent with existing literature (2,6,7,10,20), we demonstrated that LVH in COPD is independent of RV function, even during an acute exacerbation. Smoking, which is closely associated with LVH (3), was similar between both groups in our cohort increasing the generalizability of our data to an AECOPD cohort. Traditional measures of COPD severity such as Global Initiative of Chronic Obstructive Lung Disease class, systolic pulmonary artery pressures, and RV mass or function did not influence LVH, alluding to this as an independent marker of severity (2,6,7,10,20).

Using CMR, Neilan et al. (6) demonstrated increased myocardial extracellular volume using T1 measurements (a surrogate for myocardial fibrosis) and increased concentric remodeling and concentric LVH in patients with COPD. This was associated with decreased passive left atrial function. Concentric LVH is noted more frequently in these patients, irrespective of etiology of the COPD (21), pointing to pathophysiological mechanisms rather than genetic predisposition as the mechanism of action. In a large retrospective UK database, Short et al. (10) demonstrated concentric LVH to have significantly worse long-term outcomes in COPD with a 38% increased risk of mortality [hazard ratio 1.4 (1.1–1.8)]. Even though cardiovascular studies have demonstrated increased mortality and worse outcomes with eccentric LVH (22), we demonstrated the opposite consistent with other studies in the COPD population (6,10,21). We hypothesize that these differences could be due to the more global LV stress noted in COPD, similar to other systemic conditions such as hypertension or CKD, and contributes to worse outcomes (23). More dedicated studies of LV geometry in COPD are warranted to evaluate and further define this relationship given the independent role of LVH in COPD outcomes. Since the role of LV geometry in AECOPD has not been evaluated previously, our conclusion that concentric LVH is an independent predictor of NIV use merits further investigation. It is possible that during an AECOPD, further worsening of LV volumes and filling pressures could contribute to a transient congestive heart failure (CHF) with preserved ejection fraction phenotype causing worsening of acute dyspnea. However, this assessment requires the use of CMR, which has shown to be superior to echocardiography to detect co-existing CHF in patients with stable COPD (7,24).

Our study has certain limitations such as its retrospective design and use of an administrative database that carry inherent informational and selection biases. We performed a single-center study and limited our outcomes to a single hospital stay, which could potentially influence the generalizability of our findings. The AECOPD cohort was not classified by morphological or physiological derangements, which has been shown to correlate with LV mass in recent literature (2). Additionally, complete PFTs were not uniformly available on our patients and we were unable to assess the effects of lung volumes in our population. We did not evaluate the cardiac rhythm in our patients and are unable to comment on the correlation of LVH with atrial or ventricular arrhythmias in these patients (25). Electrocardiographic criteria for LVH were not considered in our evaluation, which has been demonstrated to be inferior to echocardiographic studies (3,20). Despite its superior ability to evaluate LV geometry, CMR was not uniformly used in our population due to the associated costs and lack of clear clinical indications (2,26). Finally, with an intention to develop a cohort of AECOPD patients without alternate influencing factors on LVH, e.g., hypertension, DM-2, and CKD, we could have excluded patients with controlled hypertension who could have had a regression of their LVH (3).

The strengths of this study include evaluation of the unique association of LVH with AECOPD. All prior studies have evaluated patients with stable COPD and long-term outcomes (2,3); however, there is no literature on AECOPD. By evaluating patients with and without co-morbid risk factors that influence LVH in addition to COPD, we were able to show consistent outcomes across both groups thereby increasing the applicability of our study to an AECOPD patient population. Our cohorts did not have significant differences in RV function or obstructive sleep apnea, which are common confounders of cardiac function in this population (2,6,7,10,20). Additionally, the use of a relatively inexpensive and clinically relevant tool like echocardiography, using the linear M-mode method, increases the overall applicability of our study.

Future directions for research include assessing static and dynamic measures of hyperinflation and differentiation of various morphological and pathophysiological phenotypes of COPD, especially during an acute exacerbation. Evaluation of LVH morphology and strain patterns in patients with different LV geometries will help determine the intrinsic differences between these phenotypes (10). Considering the low number of strategies that have shown mortality benefit in COPD, targeting therapeutic strategies to address LVH regression is pertinent, given its benefit in patients with cardiovascular diseases (2,3,10). It is possible that LVH with normal LV wall stress could be a physiological adaptation to enhance exercise capacity. However, this needs more research with the use of more sophisticated cardiac imaging such as CMR, 3-dimensional echocardiography, and echocardiographic strain imaging (2,5,14).

In summary, we highlight the role of LVH in patients presenting with AECOPD. We have established that these patients form a high-risk population both inclusive and independent of other co-morbidities such as CKD, DM-2, and hypertension. In our hypothesis-generating study, despite numerical trends toward worse outcomes, we were unable to demonstrate statistically significant differences between patients with and without LVH in AECOPD. In the group of patients with LVH, patients with concentric LVH did worse than those with eccentric morphology. Further dedicated prospective studies of outcomes and therapies are warranted to evaluate LVH as a modifiable risk factor in patients with COPD.

Declaration of interest

All authors report no financial disclosures or conflicts of interest with relevance to the current manuscript.