P607Myocardial injury in systemic lupus erythematosus defined by cardiac magnetic resonance imaging: clinical and echocardiographic characteristics

Abstract

Abstract Background Lupus myocarditis (LM) occurs in 5–10% of patients with systemic lupus erythematosus (SLE). Subclinical myocardial inflammation is detected in 37% at post mortem. Echocardiographic strain analyses (speckle tracking [STE]) supports subclinical myocardial dysfunction in SLE. Tissue characterization by cardiac magnetic resonance (CMR) identifies myocardial inflammation, necrosis/fibrosis, detecting clinical and subclinical myocardial injury (MIN). It is the non-invasive diagnostic modality of choice for myocarditis (all types) based on the Lake Louise criteria (LLC). The utility of CMR is limited by cost (in resource limited settings) as well as intolerance of / contra-indications to CMR. Purpose Determine prevalence of MIN in SLE (as per LLC). Compare clinical and echocardiographic features of patients with and without MIN. Identify echocardiographic predictors of MIN. Methods A prospective crossectional study was done at Tygerberg Hospital, Western Cape, South Africa. Adult inpatients, fulfilling the 2012 classification criteria for SLE were screened for inclusion. Echocardiography (echo) included strain analyses (segmental and global [GLS]) through STE and regional function assessment (wall motion score (WMS)). Patients were grouped according to CMR evidence of MIN (absent LLC [AC]; single criterion [SC]; fulfilling LLC). Clinical, laboratory and echo parameters were compared between groups. Logistic regression and ROC were used to determine predictors of MIN. Results 49/106 SLE patients screened were included (Figure 1). The median age was 27 years (IQR: 22–35) and 88% were female. 46.9% of patients had MIN (≥1 criterion): 12.2% fulfilled LLC and 34.7% had a SC. Demographic features, cardiac risk factors (including antiphospholipid syndrome) and renal disease were similar among groups. Compared to the AC group, SLE disease activity was higher in patients fulfilling LLC (p=0.022), but not in the SC group (p=0.813). A clinical and echo diagnosis of LM was made in all patients fulfilling LLC (p<0.001), in 17.6% of patients in the SC group (p=0.026) vs none in the AC group. Anti-DsDNA (p=0.054) and anti-B2GP1 (p=0.081) were more frequently positive in the SC vs AC group. The WMS was higher in LLC and SC groups (p=0.006; p=0.083) with mid and basal strain more impaired in patients with MIN (p=0.043; p=0.047). LVID and mid STE score (number of segments with impaired STE) combined was the best predictor of MIN (OR: 2.1; 95% CI: 1.2–3.5; p=0.008). The predictive model had an AUC of 0.791 (PPV: 81.8%; NPV: 86.4%). Figure 1 Conclusion CMR is limited by a high exclusion rate in SLE, mainly due to renal disease. CMR evidence of MIN is common in SLE, even in the absence of clinical myocardial dysfunction or high lupus activity. Impaired echo regional and global function occurs more frequently in patients with MIN. STE combined with LVID predicts the presence of MIN detected by CMR and has potential as a cost effective screening tool.