Data Collection and Definitions

Data on demographics, symptoms, results of electrocardiography, cardiac imaging studies, and biopsies were collected from medical records. All baseline data were defined as obtained before initiation of disease‐modifying immunosuppressive therapy. In addition, Holter electrocardiography or device interrogation was reviewed to gather information on arrhythmias. The diagnostic workup has been previously described. ³ In brief, echocardiographic calculation of left ventricular (LV) systolic function was performed using the biplane Simpsons method. Four variables were considered to assess the diastolic function of LV: left atrial maximum volume index (normal value, <34 mL/m²), annular e′ velocity (normal value, septal e′>7 cm/s and lateral e′>10 cm/s), average E/e′ ratio (normal value, <14), and peak tricuspid regurgitation velocity (normal value, <2.8 m/s). ¹⁴ The presence of LV diastolic dysfunction was confirmed when ≥2 variables were not within the normal range. Right ventricular (RV) dysfunction was present when at least 1 of the following parameters was pathologic: tricuspid annular plane systolic excursion (TAPSE) (normal value, >17 mm); Doppler tissue imaging‐derived systolic S′ velocity of the tricuspid annulus (normal value, >9.5 cm/s); and fractional area change (normal value, >35%). ¹⁵

A 1.5‐T magnetic resonance imaging scanner (Gyroscan Intera or Achieva; Philips Healthcare, Best, the Netherlands; and Siemens Aera, Erlangen, Germany) was used for cardiac magnetic resonance (CMR) imaging, with a cardiac‐dedicated phased‐array coil, under ECG gating, and breath holding in line with standard recommendations. ¹⁶ Both LV and RV volumes and ejection fractions were calculated as previously described. ³ To assess for the presence of late gadolinium‐enhancing lesions, inverse recovery sequences were obtained 10 to 15 minutes after an intravenous injection of 0.2 mmol/kg gadolinium.