The left ventricular (LV) scar size detected by cardiac magnetic resonance (CMR) imaging in ischemic cardiomyopathy (IC) has been correlated with mortality. However, the associations among myocardial fibrosis, ventricular geometry, and physiologic measures of myocardial performance remain to be defined. A retrospective analysis of patients with stable chronic IC (LV ejection fraction ≤50%) who underwent CMR imaging from 2004 to 2010 and had plasma B-type natriuretic peptide (BNP) measured within 14 days of the CMR study was undertaken. A total of 38 patients met the criteria (mean age 66 ± 10 years; 31 men [82%]). The duration of IC was 67 ± 69 months. The CMR characteristics included LV dilation (LV end-diastolic dimension 62 ± 8 mm) and severe systolic dysfunction (LV ejection fraction 28 ± 11%). The average quantitated myocardial fibrosis was 20 ± 12% of the LV mass. When stratified by fibrotic mass, increased myocardial scar size was associated with increased LV cavity size (p = 0.007), lower LV ejection fraction (p = 0.04), and higher BNP (p = 0.013). In comparison, when stratified by median BNP (475 pg/ml), an elevated BNP level was associated, not only with LV size, function, and degree of fibrosis, but also with increased meridional wall stress (p = 0.002) and worse New York Heart Association functional class (p = 0.006). In conclusion, in chronic IC, quantitated myocardial fibrosis is associated with CMR structural and functional LV abnormalities. Elevated BNP levels are related to high-risk structural and functional CMR abnormalities and wall stress and functional status. Myocardial fibrosis appears to be related to plasma BNP through the processes of ventricular remodeling.
Myocardial fibrosis in patients with ischemic cardiomyopathy (IC) is thought to promote myocardial remodeling, to serve as a substrate for ventricular arrhythmias, and to portend an increased risk of death. However, associations among the fibrotic burden, cardiac structure and function, and cardiac biomarkers are not well defined and reflect a gap in our understanding of chronic IC. The present study was undertaken to determine the relation among myocardial fibrosis burden, cardiac magnetic resonance (CMR) parameters of cardiac structure and function, and the cardiac biomarker B-type natriuretic peptide (BNP) in patients with stable chronic IC.
Methods
A retrospective analysis of patients with IC who underwent CMR from January 2004 to April 2010 and had a plasma BNP level measured within 14 days of the CMR was performed with approval of the Mayo institutional review board and included only those patients who provided consent for research analysis as required by Minnesota Statute 144.335/CRF 21 (Part 50).
The inclusion criteria were age ≥18 years, diagnosis of IC, and left ventricular (LV) ejection fraction of ≤50%. The patients were identified by a keyword search of the internal CMR records for the study period. The resulting 266 patients were cross-referenced with the Mayo Clinical Laboratory Medicine database for BNP measurements within 14 days of the CMR study. Heart failure diagnoses were validated by physician review of the medical records using the Framingham criteria, a history of heart failure-related hospitalization, or a clinical diagnosis of heart failure by a Mayo Clinic cardiologist. Evidence of coronary artery disease was confirmed by the results of previous noninvasive stress testing or previous coronary angiography. Patients with a primary diagnosis of nonischemic cardiomyopathy, hypertrophic cardiomyopathy, infiltrative cardiomyopathy, primary valvular heart disease greater than mild to moderate, or congenital heart disease were excluded.
The CMR studies were performed using a 1.5T system (Signa Twinspeed, GE Healthcare, Waukesha, Wisconsin). After the initial localizing scans, short-axis, cine, steady-state, free-precession images were acquired from the base to apex. Additional 4-chamber, 3-chamber, and 2-chamber cine, steady-state, free-precession images were prescribed from the short-axis images. Delayed enhancement images were taken 10 minutes after intravenous injection of 0.2 mmol/kg gadolinium contrast (Omniscan, GE Healthcare). The optimal inversion time was obtained by performing a cine inversion recovery sequence (GE Healthcare) and selecting the inversion time with optimal myocardial nulling, typically 180 to 240 ms for most patients.
The CMR-derived parameters, according to the standard CMR definitions, included LV end-diastolic dimension, LV end-diastolic volume, LV end-systolic volume, LV ejection fraction, and LV mass (Mass Analysis Plus, MEDIS Medical Imaging Systems, Leiden, The Netherlands). The papillary muscles were included in the LV mass calculations but excluded from the LV volume calculations.
The diastolic sphericity index was defined as the end-diastolic LV short-axis dimension divided by end-diastolic long-axis dimension in the 4-chamber view. A normal sphericity index is ≤0.6. Meridional end-systolic LV wall stress was calculated by the equation [1.33 (systolic blood pressure) × (LV systolic cavity area/LV systolic myocardial area in the short-axis view at midventricular level) kdyne/cm 2 ] (1.33 is the conversion constant from mm Hg to dynes/cm 2 . The LV end-systolic cavity area was determined by tracing the endocardial border cross-sectional area in the short-axis view at the midventricular level. The LV end-systolic myocardial area was determined by subtraction of the traced end-systolic total LV cross-sectional area (epicardial border) from the traced end-systolic LV cavity area (endocardial border) in the short-axis view at the mid-ventricular level. A normal LV end-systolic wall stress was considered ≤60 kdyne/cm 2 .
Myocardial delayed enhancement (MDE) by CMR was used to identify the presence and extent of myocardial fibrosis. The fibrotic tissue volume was manually traced and multiplied by the myocardial specific gravity (1.05 g/cm 3 ) to determine the fibrotic mass. Myocardial fibrosis was further subclassified by location (mid-myocardial, subendocardial, subepicardial, right ventricular insertion, or transmural). The derived CMR data measurements and calculations, including those for the diastolic sphericity index, meridional end-systolic wall stress, and fibrotic mass, were performed independently by 2 of us (D.H. and J.G.) in a blinded manner. The derived CMR data were reported as an average of the 2 observer data points.
Echocardiographic data to enhance characterization of the patient cohort were obtained from the transthoracic echocardiographic studies obtained clinically within 4 weeks of CMR. The relative wall thickness, a dimensionless measure used to describe LV geometry, was defined as 2× posterior wall thickness (mm) divided by the LV end-diastolic internal dimension. Normal LV geometry was defined as a relative wall thickness of <0.44 and LV mass index of <110 g/m 2 . Mitral valve regurgitation was described qualitatively (1, mild; 2, mild-moderate; 3, moderate; 4, moderate-severe; 5, severe). The mitral inflow pulse Doppler E velocity to medial annular tissue e′ velocity ratio (E/e′ ratio) was used to estimate the LV filling pressure. The left atrial volume index and right ventricular systolic pressure were calculated using standard Doppler-echocardiographic techniques.
Descriptive statistics were used to present the demographic, clinical, laboratory, medication, and imaging data. Categorical variables are expressed as frequencies (percentages) and continuous variables as the mean ± SD for normally distributed values or the median (minimum to maximum) for non-normally distributed values. Primary comparisons were analyzed between the continuous variables using nonparametric tests (Wilcoxon, 2 sample) and between nominal variables by Fisher’s exact tests (2 tailed). Multiple regression modeling was used to test for independent associations between continuous variables (myocardial fibrosis and BNP) after adjustment for covariates. The Kaplan-Meier method was used to construct survival curves. Log-rank tests were used to detect differences in survival between the groups. Statistical analyses, including primary comparisons, multiple regression modeling, and Kaplan-Meier survival analysis, were performed using the JMP, version 8, package (SAS Institute, Cary, North Carolina). Statistical significance was accepted for p <0.05. Interobserver agreement was assessed by interclass correlation coefficients (consistency type, 2 observer model) with 95% confidence intervals (CIs) using MedCalc, version 11.5.1.0, software.
Results
A total of 38 patients met the clinical validation and inclusion/exclusion criteria and were included in the present study. The baseline demographic and clinical characteristics are listed in Table 1 . The CMR and echocardiographic parameters are listed in Table 2 . Of the 38 subjects, 35 had complete gadolinium-enhanced CMR studies. The cohort members had evidence of subendocardial (20 [57%]) and/or transmural (30 [86%]) MDE. Additionally, mid-MDE was seen in 4 subjects (11%) and right ventricular insertion-delayed enhancement was present in 3 (9%) of the patients with IC. No subepicardial MDE was seen. The interobserver agreement for derived CMR parameters demonstrated an interclass correlation coefficient of 0.8081 (95% CI 0.6608 to 0.8954) for diastolic sphericity, 0.8683 (95% CI 0.7611 to 0.9293) for LV meridional stress, and 0.7452 (95% CI 0.5570 to 0.8689) for fibrotic mass.
| Variable | Value |
|---|---|
| Age (years) | 66 ± 10 |
| Men | 31 (82%) |
| Body mass index (kg/m 2 ) | 29 ± 8 |
| Systolic blood pressure (mm Hg) | 118 ± 16 |
| Diastolic blood pressure (mm Hg) | 72 ± 13 |
| Heart rate (beats/min) | 78 ± 13 |
| Diabetes mellitus | 9 (24%) |
| Systemic hypertension | 24 (63%) |
| Hyperlipidemia | 19 (50%) |
| Smoking (ever) | 22 (58%) |
| Atrial fibrillation | 3 (8%) |
| Previous myocardial infarction | 33 (87%) |
| Peripheral arterial disease | 10 (26%) |
| New York Heart Association class | |
| I | 2 (5%) |
| II | 16 (42%) |
| III | 10 (26%) |
| IV | 10 (26%) |
| Duration of cardiomyopathy (mo) | 67 ± 69 |
| Left bundle branch block | 8 (21%) |
| Angiography | 34 (89%) |
| Percutaneous coronary intervention | 16 (42%) |
| Coronary artery bypass grafting | 13 (34%) |
| Hemoglobin (g/dl) | 13.7 ± 1.6 |
| Creatinine clearance (ml/min) | 58.9 ± 3.3 |
| Serum sodium (mmol/L) | 139 ± 3.5 |
| Serum potassium (mmol/L) | 4.3 ± 0.5 |
| Total cholesterol (mg/dl) | 143 ± 50 |
| Medications | |
| β Blocker | 30 (79%) |
| Angiotensin-converting enzyme or angiotensin receptor blocker | 27 (71%) |
| Antiplatelet therapy | 28 (74%) |
| Statin | 26 (68%) |
| Diuretics | 19 (50%) |
| Digoxin | 5 (13%) |
| Aldosterone inhibitor | 8 (21%) |
| Nitrates | 5 (13%) |
| Allopurinol | 1 (3%) |
| Parameter | Value |
|---|---|
| Left ventricular end-diastolic dimension (mm) | 62 ± 8 |
| Left ventricular ejection fraction (%) | 28 ± 11 |
| Left ventricular end-diastolic volume index (ml/m 2 ) | 122 ± 41 |
| Left ventricular mass index (g/m 2 ) | 82 ± 19 |
| Fibrotic mass index (g/m 2 ) | 17 ± 10 (median 16) |
| Percent left ventricular fibrosis (%) | 20 ± 12 |
| Diastolic sphericity index | 0.61 ± 0.07 |
| Meridional end-systolic wall stress (kdyne/cm 2 ) | 182 ± 65 |
| Echocardiographic parameters | |
| Relative wall thickness | 0.35 ±0.11 |
| Mitral valve regurgitation grade 1–5 | 1.34 ± 1.2 |
| E/e′ ratio | 26 ± 12 |
| Left atrial volume index (cm 2 /m 2 ) | 45 ± 13 |
| Right ventricular systolic pressure (mm Hg) | 46 ±16 |
The median fibrotic mass index (16 g/m 2 ) was used to stratify the cohort members ( Table 3 ). Statistically significant differences were demonstrated in systolic blood pressure (126 ± 13 vs 109 ± 14 mm Hg, p = 0.004), aspirin use (10 [58%] vs 16 [89%], p = 0.04), statin use (9 [53%] vs 15 [83%], p = 0.05), and nitrate use (5 [30%] vs 0 [0%], p = 0.004) between the cohort members with an elevated versus a low fibrotic mass. No difference was seen by scar burden in demographic characteristics, duration of cardiomyopathy (p = 0.181), or New York Heart Association functional status (p = 0.214).
| Variable | Fibrosis (g/m 2 ) | p Value | ||
|---|---|---|---|---|
| ≤61 (n = 17) | >16 (n = 17) | Univariate | Multivariate ⁎ | |
| Age (years) | 65 ± 10 | 65 ± 9 | NS | — |
| Cardiomyopathy duration (mo) | 86 ± 77 | 50 ± 59 | NS | NS |
| New York Heart Association class | NS | — | ||
| I | 0 (0%) | 2 (11%) | ||
| II | 9 (53%) | 6 (33%) | ||
| III | 3 (18%) | 6 (33%) | ||
| IV | 5 (29%) | 4 (22%) | ||
| B-type natriuretic peptide (pg/ml) | 689 ± 825 | 1,119 ± 909 | 0.013 | |
| Left ventricular end-diastolic dimension (mm) | 61 ± 8 | 66 ± 6 | 0.06 | NS |
| Left ventricular end-diastolic volume index (ml/m 2 ) | 107 ± 34 | 142 ± 40 | 0.04 | 0.007 |
| Left ventricular ejection fraction (%) | 32 ± 14 | 25 ± 8 | 0.007 | 0.041 |
| Left ventricular mass index (g/m 2 ) | 77 ± 16 | 89 ± 21 | NS | NS |
| Diastolic sphericity index | 0.59 ± 0.08 | 0.63 ± 0.06 | NS | NS |
| Meridional end-systolic wall stress (kdyne/cm 2 ) | 169 ± 56 | 203 ± 71 | NS | NS |
| Relative wall thickness | 0.36 ± 0.06 | 0.33 ± 0.09 | NS | NS |
| Mitral valve regurgitation grade 1–5 | 1.5 ± 1.3 | 1.3 ± 1.2 | NS | NS |
| E/e′ ratio | 26 ± 15 | 25 ± 10 | NS | NS |
| Left atrial volume index (cm 2 /m 2 ) | 45 ± 13 | 46 ± 14 | NS | NS |
| Right ventricular systolic pressure (mm Hg) | 44 ± 16 | 48 ± 16 | NS | 0.039 |
⁎ Adjusted for age, gender, and New York Heart Association class.
Also, the median BNP value (475 pg/ml) was used for cohort stratification ( Table 4 ). BNP levels greater than the median were associated with worse New York Heart Association functional class (p = 0.006), an increased heart rate (85 ± 13 vs 72 ± 9 beats/min, p = 0.003), and shorter cardiomyopathy duration (34 ± 23 vs 105 ± 81 months, p = 0.0008) in an absence of significant differences in co-morbidities, medication use, or renal function.
