2022 Volume 86 Issue 7 Pages 1102-1112
Background: Many patients with dilated cardiomyopathy (DCM) progress to heart failure (HF), although some demonstrate left ventricular (LV) reverse remodeling (LVRR), which is associated with better outcomes. The pulmonary artery diameter (PAD) to ascending aortic diameter (AoD) ratio has been used as a prognostic predictor in patients with HF, although this tool’s usefulness in predicting LVRR remains unknown.
Methods and Results: Data from a prospective observational study of 211 patients diagnosed in 2000–2020 with DCM were retrospectively analyzed. Sixty-nine patients with New York Heart Association class I or II HF were included. LVRR was observed in 23 patients (33.3%). The mean LV ejection fraction (29%) and LV end-diastolic dimension (64.5 mm) were similar in patients with and without LVRR. The PAD/AoD ratio was significantly lower in patients with LVRR than those without (81.4% vs. 92.4%, respectively; P=0.003). The optimal PAD/AoD cut-off value for detecting LVRR was 0.9 according to the receiver operating characteristic curve analysis. Multivariate analysis identified a PAD/AoD ratio ≥0.9 as an independent predictor of presence/absence of LVRR. Cardiac events were significantly more common in patients with a PAD/AoD ratio ≥0.9 than those with a ratio <0.9, after a median follow up of 2.5 years (log-rank, P=0.007).
Conclusions: The PAD/AoD ratio can predict LVRR in patients with DCM.
Dilated cardiomyopathy (DCM) is a heterogeneous cardiac disease of variable etiology and one of the most common causes of heart failure (HF), cardiac death, and heart transplantation.1 Despite many patients progressing to severe HF, which leads to poor outcomes, 25–45% of patients with DCM demonstrate left ventricular (LV) reverse remodeling (LVRR), which is defined as an improvement in LV contractility and a reduction in LV volume.2–4 These phenomena might indicate a favorable response to drug therapies such as β-blockers. For example, Bristow et al5 previously demonstrated the dose-dependent effects of β-blockers on LVRR. Moreover, an increase in LV ejection fraction (LVEF) is associated with increased cardiac functional capacity and cardiac index, as well as decreased pulmonary capillary pressure, leading to better clinical outcomes.6–9 Patients with markedly improved LV systolic function have better outcomes, and several previous studies have attempted to determine the predictive factors for LVRR in patients with DCM.10–14 Patients who will develop LVRR should be identified because this influences decisions and timing regarding certain aggressive therapies and non-pharmacologic interventions, such as implantable cardioverter defibrillators, cardiac resynchronization therapy, LV-assist devices (LVADs), and heart transplantation.3 The ratio of main pulmonary arterial (PA) diameter (PAD) to ascending aortic diameter (AoD), assessed by computed tomography (CT), aids in evaluating dynamic cardiac volume status, and represents an indirect predictive measure for respiratory diseases, including chronic obstructive pulmonary disease,15 idiopathic pulmonary fibrosis,16 chronic thromboembolic pulmonary hypertension,17 and pulmonary hypertension due to pulmonary embolism.18 Studies also have investigated the prognostic value of the PAD/AoD ratio in patients with HF.19–21 Chimura et al19 evaluated the clinical effect of the ratio in patients with DCM with severe HF (New York Heart Association [NYHA] functional class III or IV). They showed patients with advanced HF and DCM had a high ratio, because of a lower AoD. Meanwhile, PAD was comparable between the high and low PAD/AoD ratio groups. They concluded assessment of AoD and PAD may have important clinical implications in determining whether patients with DCM are in an advanced stage of HF with a poorer prognosis.19 However, to our knowledge, no reports have studied the association between the PAD/AoD ratio and LVRR in patients with DCM with relatively stable HF (NYHA functional class I or II). Moreover, no reports have studied the association between the PAD/AoD ratio and LVRR in patients with DCM. This study thus aimed to clarify the association between the PAD/AoD ratio and LVRR in patients with DCM and stable HF.
We retrospectively analyzed data from a prospective observational study that investigated patients with suspected DCM and stable HF, and who were hospitalized at our institution for a definite diagnosis of cardiomyopathy between January and June 2020. All patients underwent 12-lead electrocardiography (ECG), laboratory measurements, echocardiography, coronary angiography, right heart catheterization, endomyocardial biopsy (EMB), and CT within 1 week to exclude secondary cardiomyopathy. CT was indicated to exclude the possibility of secondary cardiomyopathy caused by systemic diseases, such as collagen diseases, vasculitis, and autoimmune disease, as well as intercurrent diseases, such as malignant tumors and inflammatory diseases. This is important because DCM is diagnosed by exclusion; in other words, other causes of cardiomyopathy should be ruled out before diagnosing DCM.
A total of 211 patients diagnosed with DCM were ultimately enrolled. Patients who underwent cardiac resynchronization therapy until evaluation of LVRR (n=9) and those who did not undergo follow-up echocardiography (n=69) were excluded. Seven patients with clinically unstable conditions (acute NYHA class III or IV HF) were also excluded, as were 57 patients who did not undergo baseline CT. Sixty-nine patients with NYHA functional class I or II HF were ultimately included.
DCM was defined as the presence of LV dilation (LV end-diastolic dimension [LVEDD] >55 mm or an indexed LVEDD >33 mm/m2 [men] or 32 mm/m2 [women]) and LVEF <50% in the absence of significant coronary artery disease defined as luminal stenosis >50%, significant valvular heart disease, severe systemic arterial hypertension, persistent supraventricular tachyarrhythmia, or secondary cardiac muscle disease caused by any known systemic condition, as determined by EMB.22,23
Medical history, laboratory tests, echocardiography, 12-lead ECG, right heart catheterization, coronary angiography, EMB, and CT data were examined at baseline, and HF symptoms at baseline were evaluated following the NYHA classification. A history of HF was defined as requiring hospitalization for worsening HF before the study period. Other medical histories were adjudicated using clinical records. Optimal medical therapy was recommended following current clinical practice guidelines24–26 to reach all patients’ recommended target doses.
The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki, as reflected in prior approval from the ethics committee of Nagoya University Hospital (approval number: 2017-0031). Informed consent was obtained from all patients, and patient records were anonymized prior to analysis.
Echocardiography and EMBExperienced sonographers performed M-mode, 2-dimensional transthoracic echocardiography, and Doppler recordings, following American Society of Echocardiography guidelines.27 LVEF was assessed using the Teichholz formula. EMB was performed from the right ventricular side of the interventricular septum.28 Three to five EMB specimens were evaluated, and the collagen volume fraction (ratio of collagen-specific staining to total myocardial area in each biopsy sample) was calculated as an index for interstitial collagen using automated image analysis software (BZ 9000; Keyence Co. Ltd., Osaka, Japan). The myocardial fibrosis severity was then classified into 4 grades (none, mild, moderate, or severe) following visual assessment by a pathologist and cardiologist, in a blinded manner. Right heart catheterization data were collected simultaneously.
Measurement of PAD and AoD by CTAll CT measurements used non-enhanced images with 5-mm slice thickness, according to a standard protocol.29 We determined the PA bifurcation point in the axial, coronal, and sagittal views on plain CT. Transverse axial PAD was then measured manually at the level of PA bifurcation, and the minimum AoD19,30,31 was measured at the same level (Figure 1). Two cardiologists blinded to the patients’ clinical information performed the measurements. A third cardiologist re-measured measurements where there was a >5% deviation between the 2 cardiologists, and these measurements were used in the analysis.
Measurement of pulmonary arterial diameter (PAD) and ascending aortic diameter (AoD) using computed tomography. Transverse axial main PAD and minimum AoD were both measured at the level of pulmonary artery bifurcation.
Follow-up echocardiography was performed 1.0±0.5 years after baseline, and LVRR was investigated based on an absolute increase in LVEF ≥10%, final value >35%, and decrease in LVEDD ≥10% compared with baseline.9,12,32 We also evaluated the clinical score for predicting LVRR (LVRR predictive score) in patients with DCM using 5 predictors: history of hypertension (1 point), no family history of DCM (2 points), symptom duration <90 days (1 point), LVEF >35% (2 points), and QRS duration <116 ms (1 point).11 Cardiac events included a composite of hospitalization due to worsening HF, lethal arrhythmia (hospitalization for hemodynamically unstable sustained ventricular tachycardia or ventricular fibrillation), LVAD implantation, and cardiac death. We collected outcome data using medical records or via telephone calls to patients or their families.
Statistical AnalysisContinuous variables are shown as mean±standard deviation for normally distributed data and as medians for non-normally distributed data. Characteristics were compared using Kruskal-Wallis or chi-squared tests. Categorical variables are shown as numbers (percentages) and were compared using Fisher’s exact or chi-squared tests. A receiver operating characteristic (ROC) curve analysis was performed, and the area under the ROC curve was determined to investigate LVRR predictability. Univariate regression analyses were used to estimate LVRR predictors, and odds ratios and 95% confidence intervals are presented with the logistic regression analysis. Because of the number of patients with LVRR, we used 2 or 3 covariates in multivariate analyses to avoid model overfitting. Kaplan-Meier survival curves and log-rank statistics were used to assess the prognostic value of the event-free rate. Statistical significance was defined as P<0.05. All statistical analyses were performed using SPSS version 21.0 (IBM Corp., Armonk, NY, USA).
Table 1 summarizes patients’ background data. LVRR was detected in 23 patients (33.3%). Patients with LVRR had a higher prevalence of hypertension and lower prevalence of HF history. All patients presented NYHA class I or II HF. There were no significant differences in medications or laboratory parameters between patients with and without LVRR. ECG showed a wide QRS duration and higher prevalence of complete left bundle branch block in patients without LVRR. For baseline echocardiography parameters in all patients, the mean LVEF and LVEDD were 29.8% and 64.5 mm, respectively. The initial LVEF was similar between the 2 groups, with no significant difference in initial LVEDD between patients with and without LVRR. The median LVRR predictive score in patients with LVRR was 6 points, significantly higher than the 4 points in patients without LVRR. Right heart catheterization data showed patients with LVRR had a significantly lower PA wedge pressure (PAWP) and lower systolic and mean PA pressures. Cardiac output and cardiac index were similar in patients with and without LVRR. Notably, the myocardial fibrosis severity was significantly higher in patients without LVRR than those with LVRR. CT images showed the mean AoD was significantly greater in patients with LVRR than without (34.7 mm vs. 30.6 mm, respectively, P=0.001). Although there was no significant difference in PAD between the 2 groups, the mean PAD/AoD ratio was significantly lower in patients with LVRR (81.4% vs. 92.4%, respectively, P=0.003). None of the patients had ascending aortic dilation or aneurysm, or significant aortic regurgitation.
All (n=69) |
LVRR (+) (n=23) |
LVRR (−) (n=46) |
P value | |
---|---|---|---|---|
Age, years, mean (SD) | 50.0 (13.3) | 49.4 (10.8) | 50.3 (14.4) | 0.78 |
Female, n (%) | 22 (31.9) | 5 (21.7) | 17 (37.0) | 0.20 |
Body mass index, kg/m2, mean (SD) | 24.3 (4.8) | 24.3 (5.0) | 24.2 (4.8) | 0.95 |
NYHA class, n (%) | 0.60 | |||
I | 42 (60.9) | 13 (56.5) | 29 (63.0) | |
II | 27 (39.1) | 10 (43.5) | 17 (37.0) | |
Hypertension, n (%) | 18 (26.1) | 10 (43.5) | 8 (17.4) | 0.02 |
Diabetes mellitus, n (%) | 9 (13.0) | 3 (13.0) | 6 (13.0) | 1 |
Dyslipidemia, n (%) | 20 (29.0) | 8 (34.8) | 12 (26.1) | 0.45 |
Current smoker, n (%) | 24 (34.8) | 9 (39.1) | 15 (32.6) | 0.59 |
COPD, n (%) | 5 (7.2) | 2 (8.7) | 3 (6.5) | 0.74 |
Atrial fibrillation, n (%) | 6 (8.7) | 3 (13.0) | 3 (6.5) | 0.37 |
Family history, n (%) | 12 (17.4) | 5 (29.4) | 7 (13.5) | 0.13 |
History of heart failure, n (%) | 31 (44.9) | 11 (64.7) | 30 (38.5) | 0.06 |
Symptom duration, days, median (IQR) | 122 (50–264) | 52 (35–107) | 152 (91–516) | 0.001 |
Medication, n (%) | ||||
ACE-I/ARB | 62 (89.9) | 19 (82.6) | 43 (93.5) | 0.16 |
MRA | 43 (62.3) | 18 (78.3) | 25 (54.3) | 0.053 |
β-blocker | 63 (91.3) | 21 (91.3) | 42 (91.3) | >0.95 |
Diuretic | 51 (73.9) | 20 (87) | 31 (67.4) | 0.08 |
Statin | 9 (13.0) | 4 (17.4) | 5 (10.9) | 0.45 |
Laboratory data, mean (SD) or median (IQR) | ||||
eGFR, mL/min/1.73 m2 | 68.4 (22.7) | 62.2 (19.7) | 71.4 (23.7) | 0.11 |
Sodium, mEq/L | 140.1 (2.1) | 140.6 (2.2) | 140 (2.0) | 0.15 |
Hemoglobin, g/dL | 14.2 (3.7) | 14.3 (2.2) | 14.1 (1.8) | 0.66 |
Troponin T, ng/mL | 0.013 (0.07–0.027) | 0.015 (0.01–0.028) | 0.013 (0.06–0.024) | 0.25 |
BNP, pg/mL | 131 (58–318) | 134 (100–359) | 128 (55–314) | 0.65 |
Electrocardiography | ||||
QRS duration, ms, mean (SD) | 116.9 (24.3) | 109.7 (17.0) | 120.6 (26.6) | 0.044 |
CLBBB, n (%) | 11 (15.9) | 1 (4.3) | 10 (21.7) | 0.06 |
Echocardiography | ||||
LAD, mm, mean (SD) | 40.6 (7.9) | 41.1 (7.5) | 40.3 (8.2) | 0.72 |
LV end-diastolic dimension, mm, mean (SD) | 64.5 (7.1) | 65.1 (6.5) | 64.2 (7.5) | 0.65 |
LV end-systolic dimension, mm, mean (SD) | 55.3 (8.1) | 56.5 (6.1) | 54.8 (9.0) | 0.42 |
IVST, mm, mean (SD) | 8.4 (1.4) | 8.8 (1.3) | 8.2 (1.4) | 0.12 |
PWT, mm, mean (SD) | 8.6 (1.6) | 9.2 (1.5) | 8.3 (1.6) | 0.04 |
Moderate or severe MR, n (%) | 23 (30.4) | 7 (30.4) | 14 (30.4) | >0.95 |
LVEF, %, mean (SD) | 29.8 (10.4) | 27.4 (8.3) | 31.0 (11.2) | 0.18 |
EF improvement, %, median (IQR) | 11 (1.2–23.3) | 27.2 (21.2–31.9) | 3.2 (0.05–11.1) | <0.001 |
LVRR predictive score, median (IQR) | 5 (3–6) | 6 (5–7) | 4 (3–5) | <0.001 |
Right heart catheterization, mean (SD) | ||||
Heart rate, beats/min | 76.1 (14.9) | 78.7 (11.6) | 74.8 (16.2) | 0.31 |
Systolic blood pressure, mmHg | 116 (22.1) | 117.8 (20.9) | 115 (22.8) | 0.62 |
RAP, mmHg | 5.2 (3.0) | 3.9 (2.2) | 5.9 (3.1) | 0.006 |
PAWP, mmHg | 12.6 (6.2) | 10.3 (4.5) | 13.7 (6.7) | 0.017 |
Systolic PA pressure, mmHg | 28 (8.9) | 24.5 (6.0) | 29.8 (9.6) | 0.006 |
Mean PA pressure, mmHg | 19.1 (7.1) | 16.3 (4.9) | 20.5 (7.6) | 0.008 |
Cardiac output, L/min | 4.6 (1.1) | 4.6 (1.1) | 4.6 (1.1) | 0.99 |
Cardiac index, L/min/m2 | 2.5 (0.9) | 2.5 (0.5) | 2.5 (1.0) | 0.86 |
Myocardial fibrosis EMB specimen, none/mild/moderate/severe, n |
7/31/24/6 | 6/11/5/1 | 1/20/19/5 | 0.006* |
Computed tomography, mean (SD) | ||||
AoD, mm | 32 (5.0) | 34.7 (4.3) | 30.6 (4.8) | 0.001 |
PAD, mm | 27.8 (3.7) | 28 (3.0) | 27.7 (4.0) | 0.8 |
PAD/AoD ratio | 88.7 (17.0) | 81.4 (10.7) | 92.4 (18.4) | 0.003 |
*P value for myocardial fibrosis in EMB specimens was obtained by comparing the number of patients classified as none and mild with the number classified as moderate and severe using the Mann-Whitney U-test. ACE-I, angiotensin-converting enzyme inhibitor; AoD, ascending aortic diameter; ARB, angiotensin II receptor blocker; BNP, B-type natriuretic peptide; CLBBB, complete left bundle branch block; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate; EMB, endomyocardial biopsy; IQR, interquartile range; IVST, interventricular septal thickness; LAD, left atrial dimension; LV, left ventricle; LVEF, left ventricular ejection fraction; LVRR, left ventricular reverse remodeling; MR, mitral regurgitation; MRA, mineralocorticoid receptor antagonist; NYHA, New York Heart Association; PAD, pulmonary arterial diameter; PAWP, pulmonary artery wedge pressure; PWT, posterior LV wall thickness; RAP, right atrial pressure; SD, standard deviation.
A significant correlation existed between AoD and systolic blood pressure (r=0.321, P=0.008) and between PAD and mean PA pressure (r=0.301, P=0.012) (Figure 2). Supplementary Figure shows the relationships between PAD/AoD ratio and systolic blood pressure, mean PA pressure, and other parameters, including PAWP, systolic PA pressure, and cardiac index.
Correlation between vessel diameter and hemodynamic parameters. (A) Ascending aortic diameter and systolic blood pressure (r=0.321, P=0.008). (B) Pulmonary artery diameter and mean pulmonary artery pressure (r=0.301, P=0.012).
ROC curve analysis indicated a PAD/AoD cut-off value of 0.9 for predicting LVRR (Figure 3). The area under the ROC curve was 0.669, with 87% sensitivity and 50% specificity.
Receiver operating characteristic curve for predicting left ventricular reverse remodeling with pulmonary arterial diameter/ascending aortic diameter ratio. AUC, area under the curve.
Table 2 shows the differences in clinical parameters between patients with low (<0.9) and high (≥0.9) PAD/AoD ratios. The former had a significantly higher prevalence of hypertension and a shorter history of HF before the study. The latter also showed significantly less use of diuretics and a lower B-type natriuretic peptide (BNP) concentration. No significant difference existed between the 2 groups in β-blocker dose used (2.5 mg [1.25–5.0 mg] in the LVRR group vs. 3.13 mg [2.5–10.0 mg] in the non-LVRR group; P=0.311). Echocardiography showed a low PAD/AoD ratio was significantly associated with a smaller LVEDD and LV end-systolic dimension, higher LVEF at baseline, and lower prevalence of moderate or severe mitral regurgitation (MR). There were significantly more patients with LVRR in the low vs. high PAD/AoD ratio group (46.5% vs. 11.5%, respectively; P=0.003). The median LVRR predictive score was higher in the low vs. high PAD/AoD ratio group (5 and 4 points, respectively), although the difference was not significant. Right heart catheterization data indicated PAWP and systolic and mean PA pressures were significantly lower in the low PAD/AoD group. Severe myocardial fibrosis was detected in the high PAD/AoD ratio group. The AoD was significantly larger and PAD was significantly smaller in the low vs. high PAD/AoD ratio group.
Low PAD/AoD ratio group (PAD/AoD of <0.9, n=43) |
High PAD/AoD ratio group (PAD/AoD of ≥0.9, n=26) |
P value | |
---|---|---|---|
Age, years, mean (SD) | 51.0 (12.2) | 48.3 (15.0) | 0.41 |
Female, n (%) | 12 (27.9) | 10 (38.5) | 0.36 |
Body mass index, kg/m2, mean (SD) | 24.9 (4.9) | 23.2 (4.4) | 0.15 |
NYHA class, n (%) | 0.35 | ||
I | 28 (65.1) | 14 (53.8) | |
II | 15 (34.9) | 12 (46.2) | |
Hypertension, n (%) | 15 (34.9) | 3 (11.5) | 0.03 |
Diabetes mellitus, n (%) | 4 (9.3) | 5 (19.2) | 0.24 |
Dyslipidemia, n (%) | 16 (37.2) | 4 (15.4) | 0.053 |
Current smoker, n (%) | 16 (37.2) | 8 (30.8) | 0.57 |
COPD, n (%) | 3 (7) | 2 (7.7) | 0.91 |
Atrial fibrillation, n (%) | 4 (9.3) | 2 (7.7) | 0.82 |
Family history, n (%) | 6 (14) | 6 (23.1) | 0.33 |
History of heart failure, n (%) | 15 (34.9) | 16 (61.5) | 0.03 |
Symptom duration, days, median (IQR) | 91 (39–222) | 164 (82–570) | 0.04 |
Medications, n (%) | |||
ACE-I/ARB | 38 (88.4) | 24 (92.3) | 0.6 |
MRA | 25 (58.1) | 18 (69.2) | 0.36 |
β-blocker | 38 (88.4) | 25 (96.2) | 0.27 |
Diuretic | 28 (65.1) | 23 (88.5) | 0.03 |
Statin | 5 (11.6) | 4 (15.4) | 0.65 |
Laboratory data, mean (SD) or median (IQR) | |||
eGFR, mL/min/1.73 m2 | 67.9 (23.1) | 69.1 (22.6) | 0.83 |
Sodium, mEq/L | 140.5 (2.1) | 139.4 (1.9) | 0.03 |
Hemoglobin, g/dL | 14.3 (1.9) | 14.1 (1.8) | 0.34 |
Troponin T, ng/mL | 0.01 (0.007–0.022) | 0.015 (0.012–0.03) | 0.16 |
BNP, pg/mL | 106 (53–220) | 271 (126–818) | <0.001 |
Electrocardiography | |||
QRS duration, ms, mean (SD) | 115.1 (21.6) | 120.0 (28.3) | 0.42 |
CLBBB, n (%) | 7 (16.3) | 4 (15.4) | 0.92 |
Echocardiography | |||
LAD, mm, mean (SD) | 39.2 (8.4) | 42.8 (6.6) | 0.07 |
LV end-diastolic dimension, mm, mean (SD) | 62.3 (6.8) | 68.2 (6.1) | 0.001 |
LV end-systolic dimension, mm, mean (SD) | 52.8 (7.5) | 59.6 (7.4) | <0.001 |
IVST, mm, mean (SD) | 8.6 (1.4) | 8.1 (1.3) | 0.18 |
PWT, mm, mean (SD) | 8.8 (1.6) | 8.3 (1.7) | 0.17 |
Moderate or severe MR, n (%) | 9 (20.9) | 12 (46.2) | 0.027 |
LVEF, %, mean (SD) | 31.8 (10.2) | 26.4 (10.2) | 0.04 |
EF improvement, %, median (IQR) | 15.9 (3.9–27.1) | 3.3 (−2.2 to 12) | 0.006 |
LVRR, n (%) | 20 (46.5) | 3 (11.5) | 0.003 |
LVRR predictive score, median (IQR) | 5 (3–6) | 4 (3–5) | 0.172 |
Right heart catheterization, mean (SD) | |||
Heart rate, beats/min | 77.2 (14.1) | 74.2 (16.2) | 0.41 |
Systolic blood pressure, mmHg | 118.9 (19.9) | 111.9 (25.0) | 0.21 |
RAP, mmHg | 4.7 (3.2) | 6.2 (2.3) | 0.043 |
PAWP, mmHg | 11.1 (5.6) | 15 (6.6) | 0.02 |
Systolic PA pressure, mmHg | 25.3 (6.1) | 32.5 (10.9) | 0.004 |
Mean PA pressure, mmHg | 17.0 (5.5) | 22.7 (8.0) | 0.002 |
Cardiac output, L/min | 4.7 (1.0) | 4.4 (1.2) | 0.36 |
Cardiac index, L/min/m2 | 2.7 (0.8) | 2.2 (1.0) | 0.049 |
Myocardial fibrosis of EMB specimen, none/mild/moderate/severe, n |
7/20/14/1 | 0/11/10/5 | 0.01* |
Computed tomography, mean (SD) | |||
AoD, mm | 33.7 (4.5) | 29.2 (4.6) | <0.001 |
PAD, mm | 26.4 (3.4) | 30.1 (3.0) | <0.001 |
PAD/AoD ratio | 79.0 (8.0) | 104.9 (15.5) | <0.001 |
*P value for myocardial fibrosis in EMB specimens was obtained by comparing the number of patients classified as none and mild with the number classified as moderate and severe using the Mann-Whitney U-test. Abbreviations as in Table 1.
Table 3 shows the univariate and multivariate logistic regression analysis results for predicting LVRR. The univariate analysis identified hypertension, LVRR predictive score, right atrial pressure (RAP), PAWP, systolic PA pressure, myocardial fibrosis in EMB specimens, AoD, and low PAD/AoD ratio as significant predictors of LVRR. Additionally, using multivariate analysis-adjusted LVRR predictive scores and significant determinants identified in the univariate analysis, low PAD/AoD ratio was identified as an independent predictor of LVRR.
Variable | OR | 95% CI | P value |
---|---|---|---|
Univariate analysis | |||
Age | 0.995 | 0.958–1.033 | 0.80 |
Female | 0.474 | 0.149–1.508 | 0.21 |
Body mass index | 1.004 | 0.904–1.115 | 0.95 |
Hypertension | 3.654 | 1.189–11.231 | 0.02 |
Diabetes mellitus | 1.00 | 0.226–4.420 | >0.95 |
Dyslipidemia | 1.511 | 0.512–4.456 | 0.45 |
Current smoker | 1.329 | 0.470–3.758 | 0.59 |
Atrial fibrillation | 2.15 | 0.398–11.605 | 0.37 |
History of heart failure | 0.701 | 0.253–1.941 | 0.49 |
NYHA class | 1.312 | 0.474–3.635 | 0.60 |
Family history | 0.145 | 0.017–1.199 | 0.07 |
Symptom duration | 0.996 | 0.991–1.001 | 0.09 |
ACE-I/ARB | 0.331 | 0.067–1.627 | 0.17 |
β-blocker | 1.00 | 0.169–5.908 | >0.95 |
MRA | 3.024 | 0.959–9.533 | 0.059 |
Diuretic | 3.226 | 0.827–12.582 | 0.09 |
eGFR | 0.980 | 0.956–1.005 | 0.12 |
BNP/10 | 0.994 | 0.979–1.010 | 0.47 |
QRS duration | 0.977 | 0.951–1.004 | 0.09 |
CLBBB | 0.164 | 0.020–1.367 | 0.095 |
LVEF | 0.966 | 0.180–1.017 | 0.18 |
Moderate or severe MR | 0.911 | 0.541–1.534 | 0.73 |
LVRR predictive score | 2.590 | 1.585–4.232 | <0.001 |
Heart rate | 1.018 | 0.984–1.053 | 0.31 |
Systolic blood pressure | 1.006 | 0.983–1.029 | 0.62 |
RAP | 0.755 | 0.611–0.934 | 0.009 |
PAWP | 0.864 | 0.751–0.994 | 0.04 |
Systolic PA pressure | 0.920 | 0.855–0.989 | 0.025 |
Myocardial fibrosis of EMB specimen | 0.345 | 0.159–0.751 | 0.007 |
AoD | 1.222 | 1.072–1.394 | 0.003 |
PAD | 1.018 | 0.888–1.169 | 0.79 |
PAD/AoD ratio | 0.948 | 0.908–0.990 | 0.016 |
Low PAD/AoD ratio group | 6.667 | 1.738–25.565 | 0.006 |
Multivariate analysis | |||
LVRR predictive score | 2.507 | 1.52–4.136 | <0.001 |
Low PAD/AoD ratio group | 6.346 | 1.416–28.45 | 0.016 |
LVRR predictive score | 4.638 | 1.963–10.954 | <0.001 |
Low PAD/AoD ratio group | 83.560 | 4.515–1,546.474 | 0.003 |
MRA | 18.79 | 1.57–225.044 | 0.021 |
LVRR predictive score | 3.494 | 1.684–7.248 | 0.001 |
Low PAD/AoD ratio group | 39.075 | 3.083–495.23 | 0.005 |
Diuretic | 4.088 | 0.635–26.316 | 0.138 |
LVRR predictive score | 2.625 | 1.533–4.497 | <0.001 |
Low PAD/AoD ratio group | 7.664 | 1.59–36.942 | 0.011 |
CLBBB | 0.117 | 0.01–1.3 | 0.081 |
LVRR predictive score | 2.426 | 1.453–4.051 | 0.001 |
Low PAD/AoD ratio group | 5.276 | 1.143–24.361 | 0.033 |
RAP | 0.85 | 0.674–1.072 | 0.170 |
LVRR predictive score | 3.152 | 1.68–5.913 | <0.001 |
Low PAD/AoD ratio group | 5.522 | 1.134–26.886 | 0.034 |
PAWP | 0.857 | 0.749–0.981 | 0.025 |
LVRR predictive score | 2.658 | 1.542–4.583 | <0.001 |
Low PAD/AoD ratio group | 4.777 | 1.031–22.14 | 0.046 |
Systolic PA pressure | 0.922 | 0.836–1.016 | 0.102 |
LVRR predictive score | 2.533 | 1.478–4.342 | 0.001 |
Low PAD/AoD ratio group | 6.372 | 1.324–30.664 | 0.021 |
Myocardial fibrosis of EMB specimen | 0.741 | 0.277–1.979 | 0.55 |
CI, confidence interval; OR, odds ratio. Other abbreviations as in Table 1.
Composite cardiac events occurred in 17 patients (21.7%) after a median follow up of 2.5 (interquartile range: 1.5–5.0) years. Cardiac events were significantly less frequent in patients with LVRR than those without LVRR (4.3% vs. 30.4%, respectively; P=0.013) (Table 4). Kaplan-Meier survival curves showed significantly fewer composite cardiac events in patients with LVRR (log-rank, P=0.033) (Figure 4A). Patients with DCM with a PAD/AoD ratio <0.9 also had a significantly lower event-free rate than those with a PAD/AoD ratio ≥0.9 (log-rank, P=0.006) (Figure 4B).
Event | All (n=69) |
LVRR (+) (n=23) |
LVRR (−) (n=46) |
P value |
---|---|---|---|---|
Total events, n (%) | 15 (21.7) | 1 (4.3) | 14 (30.4) | 0.013 |
Hospitalization for worsening heart failure | 10 (14.5) | 1 (4.3) | 9 (19.6) | 0.09 |
Lethal arrhythmia | 6 (8.7) | 0 (0) | 6 (13) | 0.07 |
LVAD implantation | 1 (1.4) | 0 (0) | 1 (2.2) | 0.48 |
Cardiac death | 4 (5.8) | 0 (0) | 4 (8.7) | 0.15 |
LVAD, left ventricular assist device; LVRR, left ventricular reverse remodeling.
Kaplan-Meier survival curves for (A) patients with and without left ventricular reverse remodeling (LVRR) and (B) patients with low and high pulmonary arterial diameter (PAD)/ascending aortic diameter (AoD) ratios.
This study showed that among patients with stable DCM, the PAD/AoD ratio measured by CT was significantly lower in patients with LVRR than those without LVRR. The ratio predicted LVRR with a cut-off value of 0.9, with a low ratio associated with fewer cardiac events. To our knowledge, this study provides the first evidence for using the PAD/AoD ratio as a predictor of LVRR in patients with DCM.
This study found that a lower PAD/AoD ratio was associated with a higher possibility of LVRR. The difference in PAD/AoD ratio was driven mainly by a significant difference in AoD, similar to the findings by Chimura et al,19 who showed patients with advanced HF (NYHA functional class III/IV) and DCM had a high PAD/AoD ratio, owing to a lower AoD. Regarding the mechanism, aortic stiffness is related to LVRR in patients with DCM.33 Thus, patients without LVRR may have demonstrated increased aortic stiffness, resulting in decreased AoD and a consequent increase in PAD/AoD ratio.
AoD is positively correlated with age and hypertension.34,35 Notably, there was no significant difference in age between the 2 groups, suggesting the larger AoD in the low PAD/AoD group might be attributable to the higher proportion of patients with hypertension; however, there was no significant difference in systolic blood pressure between the 2 groups, suggesting potentially better blood pressure control among patients in the low PAD/AoD ratio group. A history of hypertension is a predictor of LVRR in patients with DCM,36 suggesting a possible relationship between high prevalence of hypertension and good prognoses in patients with a low PAD/AoD ratio. Target doses of anti-hypertensive agents, such as β-blockers, may also result in favorable outcomes;5,12,37 however, there was no significant difference in the β-blocker dose between this study’s 2 groups; thus, β-blockers’ dose-dependent effects likely did not influence LVRR.
Patients with a high PAD/AoD ratio also had a lower cardiac index, associated with poorer outcomes in patients with HF. This lower index potentially affects chronic aortic remodeling and may lead to a small-diameter descending aorta.
Myocardial fibrosis severity evaluated by myocardial biopsy is an independent predictor of LVRR.37 In this study, the PAD/AoD ratio was significantly higher in patients without LVRR than those with LVRR, and LVRR was less likely to occur in the presence of myocardial fibrosis.38 We observed a higher RAP, PAP, and PAWP in the high PAD/AoD group (patients without LVRR). This indicates backing up of blood, subsequently indicating decreased ventricular compliance and myocardial fibrosis. Indeed, an EMB of the right ventricle in the high PAD/AoD group showed myocardial fibrosis. This mechanism supports use of the PAD/AoD ratio for predicting LVRR in patients with stable DCM. These results are consistent with those found by Chimura et al.19 The PAD/AoD ratio use also excludes the effects of body surface area from both corrected PAD and AoD.39
Additionally, echocardiography showed LV enlargement and more severe MR in patients with a high PAD/AoD ratio. An increased cardiac chamber size might reflect chronic volume and pressure overload, resulting in cardiac remodeling and myocardial fibrosis.40,41
An elevated PAD/AoD ratio can correlate with mean PA pressure,20,42 and is an independent predictor of pulmonary hypertension and mortality in patients with HF.19,43 PA enlargement may reflect lung parenchyma destruction, hypoxia, centralization of blood flow, and blood volume loss, and may be a marker of increased PA pressure.44 There was, however, no significant difference in PAD between this study’s 2 groups. This apparent discrepancy may owe to differences in populations among studies. We investigated patients with relatively stable NYHA functional class I or II HF, who had not developed pulmonary volume loading and pressure loading. This study also found patients in the high PAD/AoD group used more diuretics and had a higher BNP concentration.
CT is not essential in the diagnosis of DCM, although it was necessary to exclude the possibility of secondary cardiomyopathy associated with systemic disease. CT was also necessary for accurate measurement of PAD and AoD to predict LVRR; however, because of the risk of radiation exposure, some patients did not consent to CT, and were therefore excluded.
Several studies have investigated other clinical factors predicting LVRR in patients with DCM.3,45,46 A clinical score developed by Kimura et al11 for predicting LVRR in patients with DCM showed good predictability based on five clinical indicators: hypertension history; family history; symptom duration; LVEF; and QRS duration. This score has better predictability for LVRR in patients with DCM compared with late gadolinium enhancement on cardiac magnetic resonance imaging or fibrosis severity according to EMB.11 The present study showed the PAD/AoD ratio still had significant predictability, even after adjusting for this clinical predictive score.
Accurate assessment of intracardiac pressure via right heart catheterization is useful for diagnosing pulmonary hypertension and determining treatment strategies. Meanwhile, EMB can reveal the degree of myocardial fibrosis, which is associated with reversibility of ventricular function.46 This study revealed that intracardiac pressure and myocardial fibrosis had significant predictability. Additionally, genetic investigations of EMB specimens examining the molecular pathogenesis of DCM also demonstrated good predictability for LVRR.23,47 However, the risk of invasive manipulation, difficulty in precise assessment of EMB, lack of access to techniques for evaluating genetic variants, and requirement for expensive techniques suggest alternative methods may be preferable.
Echocardiography might be the gold-standard for evaluating LVRR, and some echocardiographic parameters may help predict LVRR;46 however, measuring these parameters and estimating pulmonary hypertension by echocardiography are often technically difficult and susceptible to bias, including interobserver variability and limited visibility in some patients.
Conversely, although CT carries some risk of radiation exposure, it is less invasive than right heart catheterization and EMB, and more accurate and reproducible than echocardiography. CT may also reflect aortic remodeling due to chronic cardiac output reduction, as well as the effects of temporary intracardiac pressure and blood pressure on vessel diameter. Measuring PAD and AoD using CT may, thus, be useful for early prediction of LVRR and cardiac events. Measuring and interpreting the PAD/AoD ratio using CT is therefore relatively easy and may contribute to its high prognostic value in patients with DCM. Additional to right heart catheterization and echocardiography, CT may thus be a useful additional technique for predicting LVRR in patients with early-stage DCM.
This study had several limitations. First, it was a retrospective analysis of data from a single-center prospective observational study. The sample size was relatively small, and few patients had cardiovascular events. It is therefore possible that not all confounding factors were excluded. A future large-scale study could determine the PAD/AoD ratio’s comprehensive importance as a predictor of LVRR in patients with DCM. Second, data on symptom duration were based on patients’ self-reporting, which is potentially inaccurate. Third, patients who underwent thoracic CT during their hospital stay were included; some baseline differences may have occurred between patients who did and did not undergo CT, leading to biases when generalizing the results to all patients with DCM. Furthermore, CT data were not evaluated by ECG-gated axial chest CT because the scans were not intended specifically to measure PAD or AoD, and PAD and AoD showed small changes during the cardiac cycle, potentially affecting the results.48 Fourth, this study assessed LVEF using the Teichholz formula. Although the biplane method of disks may be preferable for investigating LVEF, data obtained using this method were not available for all patients. Finally, although some studies have reported the relationship between LVRR and use of sacubitril/valsartan49 or ivabradine,50 this study did not include patients taking these medications, although these medications were not indicated for HF therapy in Japan during the study period.
We showed that the PAD/AoD ratio is a significant predictor of LVRR, reflecting intracardiac pressure and myocardial fibrosis in patients with DCM. Such patients with a high PAD/AoD ratio have significantly worse long-term outcomes than those with a lower ratio.
T. Murohara is one of the Senior Advisory Editors of Circulation Journal.
T. Okumura received research grants from Ono Pharmaceutical Co. Ltd., Bayer Pharmaceutical Co. Ltd., Daiichi-Sankyo Pharma Inc., and Amgen Astellas BioPharma K. K. outside the submitted work. T. Okumura received honoraria from Ono Pharmaceutical Co. Ltd., Otsuka Pharmaceutical Co. Ltd., Novartis Pharma K. K., and Medtronic Japan Co. Ltd. T. Murohara received lecture fees from Bayer Pharmaceutical Co. Ltd., Daiichi-Sankyo Co. Ltd., Sumitomo Dainippon Pharma Co. Ltd., Kowa Co. Ltd., MSD K. K., Mitsubishi Tanabe Pharma Co., Nippon Boehringer Ingelheim Co. Ltd., Novartis Pharma K. K., Pfizer Japan Inc., Sanofi-Aventis K. K., and Takeda Pharmaceutical Co. Ltd. T. Murohara received an unrestricted research grant for the Department of Cardiology, Nagoya University Graduate School of Medicine from Astellas Pharma Inc., Daiichi-Sankyo Co. Ltd., Sumitomo Dainippon Pharma Co. Ltd., Kowa Co. Ltd., MSD K. K., Mitsubishi Tanabe Pharma Co., Nippon Boehringer Ingelheim Co. Ltd., Novartis Pharma K. K., Otsuka Pharma Ltd., Pfizer Japan Inc., Sanofi-Aventis K. K., Takeda Pharmaceutical Co. Ltd., and Teijin Pharma Ltd. The remaining authors have no conflicts of interest relevant to this article to disclose. This research received no grant from any funding agency in the public, commercial, or not-for-profit sectors.
The study was approved by the ethics committee of Nagoya University Hospital (approval number: 2017-0031).
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-21-0786