Diabetic patients with coronary artery disease (CAD) are more likely to develop heart failure (HF) than nondiabetic patients, but the mechanism responsible is unclear. Evidence suggests that infarct size and accompanying remodeling may not explain this difference. We used cardiac magnetic resonance (CMR) imaging to compare degree of left ventricular (LV) myocardial scar and remodeling in diabetic and nondiabetic patients with CAD. We evaluated 85 patients (39 diabetic, 46 nondiabetic) who underwent coronary angiography showing obstructive CAD and CMR imaging within 6 months of each other. Myocardial scar was measured by late gadolinium enhancement on CMR imaging and was graded according to spatial and transmural extents on a semiquantitative scale. More diabetic than nondiabetic patients had HF (69% vs 43%, p <0.03); however, groups did not differ in total scar burden (0.94 ± 0.60 vs 1.17 ± 0.74, p = NS), spatial extent of scar, or extent of transmural scar. Diabetes remained an independent predictor of HF after adjustment for CAD and other variables. LV ejection fraction (36 ± 12% vs 37 ± 14%, p = NS) and end-diastolic volume (215 ± 56 vs 217 ± 76 ml, p = NS) were similar for diabetic and nondiabetic patients, respectively. In conclusion, although diabetic patients with CAD had a higher prevalence of HF than nondiabetic patients, there was no difference in myocardial scar, LV volume, or LV ejection fraction. These findings support the theory that mechanisms other than extent of myocardial injury and negative remodeling play a significant role in the development of HF in diabetic patients with CAD.
The aim of this study was to use cardiac magnetic resonance (CMR) imaging to compare degree of left ventricular (LV) myocardial scar and remodeling in diabetic and nondiabetic patients with coronary artery disease (CAD). We hypothesized that degree of myocardial scar and remodeling would be similar between the 2 groups, despite a higher prevalence of heart failure (HF) in diabetic patients.
Methods
We retrospectively identified consecutive patients from a CMR database at the Manhattan Veteran’s Affairs Hospital from January 2006 through June 2009 who underwent coronary angiography and CMR imaging within 6 months of each other. Patients were eligible only if they were found to have had ≥70% stenosis in ≥1 major coronary artery (or ≥50% stenosis of the left main coronary artery) and ≥1 of the following: myocardial scar on CMR imaging, Q-wave infarction on electrocardiogram, LV wall motion abnormality on echocardiogram or LV angiogram, or presence of ischemia at stress testing. Patients were excluded if they had (1) documented myocardial infarction with increased troponin levels within 1 week of CMR imaging, (2) any coronary event from the time of angiogram to CMR studies, (3) any previous coronary artery bypass surgery, or (4) evidence of infiltrative or nonischemic cardiomyopathy (including cardiomyopathy judged to have primary valvular pathology) based on clinical and/or CMR data. Institutional review board approval was granted for this study.
All clinical definitions are based on data recorded at or before the time of CMR imaging with the exception of estimated glomerular filtration rate (eGFR), which was based on serum creatinine levels obtained within 24 hours before coronary angiography. A clinical diagnosis of HF was determined by chart review and assigned to patients with documented signs and/or symptoms of HF adapted from the Framingham criteria. Diagnosis of diabetes was determined by documentation in a patient’s medical record by a clinical provider (i.e., physician, physician’s assistant, or nurse practitioner) or if the patient was taking hypoglycemic agents (or insulin). Similarly, diagnoses of dyslipidemia and hypertension were determined by documentation in a patient’s medical record by a clinical provider or, for dyslipidemia, a patient was taking lipid-modifying agents. In addition, based on documentation by clinical providers, the following parameters were recorded: body mass index, tobacco use (former or current use), previous myocardial infarction, and renal function (eGFR). Estimated GFR was calculated using the following equation: eGFR (milliliters per minute per 1.73 m 2 ) = 186 × (serum creatinine) −1.154 × (age) −0.203 × (0.742 for women) × (1.212 for African-Americans).
Patients underwent imaging using a 1.5-T MR system (Avanto, Siemens, Erlangen, Germany) in conjunction with electrocardiographic gating. Imaging was performed using the institutional standard protocol consisting of scout images to identify cardiac axes, cine images for evaluation of segmental wall motion, and contrast-enhanced T1-weighted inversion recovery late gadolinium enhancement imaging with inversion time set to null normal myocardial signal intensity for evaluation of myocardial scarring. Imaging was performed in standard 2-, 3-, and 4-chamber long-axis views. In addition, a short-axis series (base to apex) was acquired every 10 mm to cover the entire LV volume. Patients received intravenous gadolinium-diethylenetriamine penta-acetic acid (0.15 mmol/kg) or gadolinium-benzyl oxy propionic tetra-acetate (0.1 mmol/kg) contrast depending on baseline renal function, and delayed imaging was performed 10 to 15 minutes after contrast administration.
Myocardial scar was measured by presence of late gadolinium enhancement on CMR images for each patient using a 17-segment model of the left ventricle and was graded according to transmural extent on a semiquantitative scale (0 = none, 1 = 1% to 25%, 2 = 26% to 50%, 3 = 51% to 75%, 4 = 76% to 100%). Total scar burden was measured as mean grade per segment. Spatial extent of scarring was recorded as total number of segments with presence of any scar. Extent of transmural scar was recorded as the number of segments with grade 3 or 4 scar. Measurements of LV end-diastolic volume and LV ejection fraction by CMR imaging were also recorded for each patient.
Coronary angiography was performed according to clinical practice. Coronary lesions were graded by visual estimation of percent diameter stenosis and were recorded in the Veterans Affairs’ cardiac catheterization database (CART-CL). Any vessel that had undergone previous percutaneous intervention was considered to have had severe stenosis.
Fisher’s exact test was used for categorical variables. Exact Mann–Whitney test was used for continuous variables. Logistic regression was performed to evaluate the relation between HF and clinical, CMR, and angiographic variables. Stepwise variable regression was performed to identify significant independent predictors of HF. All reported p values are 2-sided and results were declared statistically significant at a p value ≤0.05. SAS 9.0 (SAS Institute, Cary, North Carolina) was used for all computations.
Results
Eighty-five patients met inclusion and exclusion criteria and represent the cohort for further study (39 diabetic patients and 46 nondiabetic patients). The cohort included patients referred for CMR imaging for clinical indication of viability (83 patients), LV function (1 patient), and evaluation of possible right atrial mass (1 patient). Baseline characteristics are listed in Table 1 . Prevalence of clinical HF was higher in diabetic patients compared to nondiabetic patients (69% vs 43%, p <0.03).
| Variable | Diabetes Mellitus | p Value | |
|---|---|---|---|
| Yes | No | ||
| (n = 39) | (n = 46) | ||
| Age (years) | 68 ± 9 | 68 ± 10 | 0.98 |
| Body mass index (kg/m 2 ) | 30 ± 6 | 29 ± 7 | 0.35 |
| Men | 38 (97%) | 46 (100%) | 0.46 |
| Hypertension | 37 (95%) | 35 (76%) | 0.03 |
| Dyslipidemia | 38 (97%) | 34 (74%) | <0.01 |
| Smoker | 27 (69%) | 32 (70%) | 1.0 |
| Estimated glomerular filtration rate (ml/min) | 71 ± 25 | 74 ± 20 | 0.63 |
| Previous myocardial infarction | 23 (59%) | 25 (54%) | 0.83 |
| Q-wave infarction on electrocardiogram | 19 (49%) | 26 (57%) | 0.48 |
| Number of coronary arteries narrowed | 0.03 | ||
| Left main coronary artery or 3 | 22 (56%) | 20 (43%) | |
| 2 | 14 (36%) | 11 (24%) | |
| 1 | 2 (5%) | 15 (33%) | |
| ≥1 total coronary occlusion | 22 (56%) | 31 (67%) | 0.30 |
Seventy-seven patients (91%) had evidence of scar on CMR images—100% of those with a clinical history of myocardial infarction and 78% of those without a history of myocardial infarction. There was no significant difference in degree of LV myocardial scar, LV ejection fraction, or LV end-diastolic volume between diabetic and nondiabetic patients ( Table 2 ). Clinical, angiographic, and CMR variables were analyzed for association with HF, and based on stepwise multivariate regression analysis, only diabetes and LV ejection fraction remained significant independent predictors of HF ( Table 3 ).
| Variable | Diabetes Mellitus | p Value | |
|---|---|---|---|
| Yes | No | ||
| (n = 39) | (n = 46) | ||
| Total scar burden ⁎ | 0.94 ± 0.60 | 1.17 ± 0.74 | 0.13 |
| Spatial extent of scar † | 5.5 ± 2.9 | 6.5 ± 4.2 | 0.22 |
| Extent of transmural scar † | 3.7 ± 2.9 | 4.5 ± 3.1 | 0.23 |
| Left ventricular ejection fraction (%) | 36 ± 12 | 37 ± 14 | 0.66 |
| Left ventricular end-diastolic volume (ml) | 215 ± 56 | 217 ± 76 | 0.85 |
| Left ventricular end-diastolic volume index (ml/m 2 ) | 106 ± 29 | 109 ± 37 | 0.96 |
⁎ Semiquantitative score as described in Methods.
| Univariate | Multivariate | |
|---|---|---|
| Age | 0.52 | — |
| Body mass index | 0.09 | — |
| Diabetes | 0.02 | 0.02 |
| Hypertension | 1.0 | — |
| Hyperlipidemia | 0.77 | — |
| Smoker | 0.64 | — |
| Glomerular filtration rate | 0.07 | — |
| Coronary disease burden | 0.12 | — |
| Left ventricular end-diastolic volume | 0.001 | 0.70 |
| Left ventricular ejection fraction | <0.0001 | <0.0001 |
| Spatial extent of scar | 0.03 | 0.86 |
| Transmural extent of scar | 0.03 | 0.31 |
| Total scar burden | 0.03 | 0.73 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
