Assessment of Early Phenotypes in Diabetic Heart Disease by Cardiac Magnetic Resonance

Assessment of early phenotypes in diabetic heart disease by cardiac magnetic resonance imaging and magnetic resonance spectroscopy

Levelt E1,2, MBBS MRCP; Mahmod M1, MBBS MRCP Dphil, Ntusi NAB BSc (Hons) MBCHB FCP(SA) DPhil 1, Ariga R 1MBBS, BSc (Hons)MRCP , Upton R3 BSc (Hons), MSc; Piechnick SK1, PhD; Francis JM1, DCR(R) DNM; Schneider J1, PhD; Karamitsos TD1, MD PhD; Leeson P3, PhD, FRCP; Holloway CJ1,2, FRACP DPhil; Clarke K2, PhD; Neubauer S 1, MD FRCP FACC FMedSci

Institutions (All):

1.University of Oxford Centre for Clinical Magnetic Resonance Research, University of Oxford, United Kingdom.

2. Department of Physiology, Anatomy and Genetics, University of Oxford, United Kingdom.

3. Oxford Cardiovascular Clinical Research Facility, University of Oxford, United Kingdom.

Background: The causes of diabetic cardiomyopathy are multifactorial and not fully elucidated. Cardiovascular magnetic resonance (CMR) imaging and spectrocopy (MRS) provide a non-invasive assessment of the functional, structural and metabolic status of the heart.

Objective: The aim of this study was to assess the earliest manifestations of diabetic cardiomyopathy using multiparametric CMR and MRS in patients with stable, uncomplicated type 2 diabetes, having a short duration of disease.

Methods: 20 patients (9 female, mean age 54 ± 1.65 years) with early-onset (median duration 1.5 [IQR: 0.5-2] years) type 2 diabetes and 12 healthy volunteers with moderately elevated body mass index (BMI) (5 female, mean age 52 ± 2.3 years) and 15 healthy volunteers with normal BMI (6 female, mean age 53± 3.5 years) were studied. Patients were either drug naive for diabetic therapy or on treatment with metformin monotherapy, HBA1c ≥ 6.4 and ≤ 8.8%, with no history of coronary artery disease or uncontrolled hypertension. Myocardial lipid content and PCr/ATP ratios were quantified using 1H- and 31P MRS, respectively. CMR included cine, tagging and native T1 mapping at 3.0 T. LV diastology was characterised using echocardiography.

Results: Diabetic patients were well-matched with both control groups (Table 1) for age and gender; they were weight matched with the elevated BMI control group. Myocardial energetics were impaired in diabetics when compared to controls with elevated and normal BMI (PCr/ATP ratio: 1.52 ± 0.07 vs. 1.99 ± 0.10 vs. 2.06± 0.12, p< 0.001) and myocardial triglyceride content was increased (1.09 ± 0.15 vs. 0.53 ± 0.07 vs. 0.51±0.07 % respectively, p=0.002), despite the relatively short disease duration. A modest correlation between myocardial energetics and myocardial triglyceride content was noted (R=-0.35, p=0.03). HBA1c also correlated inversely with PCr/ATP ratio (R=-0.49, p=0.004) and myocardial triglyceride content (R=-0.53, p=0.004) and showed a positive correlation with left ventricular wall thickness (R=0.46, p=0.009) and diastolic dysfunction (R=-0.44, p=0.014). Peak systolic circumferential strain was reduced in diabetics when compared to controls with elevated and normal BMI (-15.4 ± 0.9 vs -19 ± 0.7 vs 19±0.6 , p=0.002), indicating subtle LV dysfunction. Left ventricular wall thickness was increased (10.5±0.4 vs. 10.0±0.3 vs. 8.4±0.5 mm, p=0.03) and diastolic function was impaired in diabetic patients (mitral in-flow E/A ratio=0.89 ± 0.06 vs. 1.13±0.13 vs. 1.24±0.1, p=0.01), respectively ,when compared to elevated and normal BMI controls. Furthermore, despite the metabolic abnormalities observed in diabetics, there was no difference in native T1 values (as a measure of myocardial fibrosis) between diabetic patients and elevated and normal BMI controls (1192±6.5 vs. 1184±7 vs.1198±12 respectively, p=0.58).

Conclusions: Abnormal myocardial energy metabolism, cardiac steatosis, reduced LV strain and diastolic dysfunction are present in uncomplicated type 2 diabetes patients with a short duration of disease.HBA1c is an independent predictor of early cardiac involvement in diabetic patients. CMR is a sensitive, non-invasive tool for assessment of myocardial pathophysiology, and it is helpful in the comprehensive phenotyping and staging of myocardial involvement in diabetes.