Myocardial injury in systemic lupus erythematosus according to cardiac magnetic resonance tissue characterization: clinical and echocardiographic features

Objectives: To determine the prevalence of myocardial injury (MInj) in systemic lupus erythematosus (SLE) according to cardiac magnetic resonance (CMR) criteria. To compare clinical and echocardiographic features of patients with and without MInj and identify predictors of myocardial tissue characteristics according to CMR. Methods: SLE inpatients underwent CMR screening for MInj based on the Lake Louise Criteria (LLC). Tissue characteristics included inflammation (increased T2-weighted signal or early gadolinium enhancement ratio (EGEr)) and necrosis or fibrosis (late gadolinium enhancement (LGE)). Echocardiographic parameters included left (left ventricular ejection fraction (LVEF)) and right ventricular function (tricuspid annular plane systolic excursion (TAPSE)), global longitudinal strain (GLS), wall motion score (WMSi) and left ventricular internal diameter index (LVIDi). Variables were compared with regards to the presence/absence of CMR criteria. Logistic regression identified variables predictive of CMR tissue characteristics. Results: A hundred and six SLE patients were screened of whom 49 patients were included. Fifty-seven patients were excluded due to intolerance of or contraindication to CMR (27/57 due to renal impairment). Twenty-three patients had CMR evidence of MInj, of which 60.9% was subclinical. Inflammation occurred in 16/23 and necrosis/fibrosis in 12/23 patients. Patients with any evidence of MInj were more frequently anti-dsDNA positive (p = 0.026) and patients fulfilling LLC for myocarditis had higher SLE disease activity (p = 0.022). The LVIDi (p = 0.005), LVEF (p = 0.005) and TAPSE (p = 0.011) were more abnormal in patients with an increased EGEr, whereas WMSi (p = 0.002) and GLS (0.020) were more impaired in patients with LGE. On multivariable logistic regression analyses, TAPSE predicted inflammation (OR: 0.045, p = 0.006, CI: 0.005–0.415) and GLS predicted necrosis/fibrosis (OR: 1.329, p = 0.031, CI: 1.026–1.722). A model including lymphocyte count, TAPSE and LVIDi predicted an increased EGEr on CMR (receiver operating characteristic-curve analyses: area under the curve: 0.901, p ' 0.001, sensitivity: 88.9%, specificity: 76.3%). Conclusions: CMR evidence of MInj frequently occurs in SLE and is often subclinical. The utility of CMR in SLE is limited by a high exclusion rate, mainly due to renal involvement. Models including echocardiographic parameters (TAPSE, LVIDi and GLS) are predictive of CMR myocardial injury. Echocardiography can be used as a cost-effective screening tool with a high negative predictive value, in particular when CMR is contraindicated or unavailable.