Relationships between cardiac structural and functional assessment by cardiac MRI and hemoglobin in end-stage renal disease

Abstract

Left ventricular mass index (LVMi) is an established surrogate endpoint for cardiovascular outcomes in hemodialysis patients [1]. Maintenance of hemoglobin between 100 and 120 g/l was associated with LVMi regression measured by echocardiography [2]. Cardiac magnetic resonance imaging (CMR) is the reference standard for assessing LV structure and function. Accordingly, we examined the relationships between hemoglobin, iron storage indices and erythropoietin stimulating agents (ESA) with changes in biventricular structure and function measured by CMR in hemodialysis patients. Details of this prospective two-center study were previously published [3]. Sixty-seven adult long-term (> 90 days) recipients of hemodialysis were included after obtaining consent, and the study was approved by the research ethics boards. Exclusion criteria included serious comorbidity with life expectancy < 1 year, planned kidney transplant, pregnancy, and contraindications to CMR. Thirty-seven patients converted to in-center nocturnal hemodialysis (7–8 h/session, three times per week) and 30 patients remained on conventional hemodialysis (3–4 h/session, three times per week). All aspects of hemodialysis care conformed to the guidelines and did not differ between the 2 groups. Each hemodialysis session aimed to achieve target dry weight where participants have minimal signs or symptoms of hypovolemia or hypervolemia, as tolerated. Laboratory investigations and CMR (1.5 T or 3 T scanner) were performed at baseline and at end of study (EOS). Blinded cardiac imagers reviewed CMR and performed post-processing analysis using a commercial software(Cvi42). The sum of the differences between the endocardial and epicardial borders was calculated and multiplied by the myocardial specific density (1.05 g/cm3) to calculate the left and right ventricular mass (LVM and RVM). CMR measurements included left and right ventricular ejection fraction (LVEF and RVEF), left and right ventricular end-systolic volume (LVESV and RVESV), and left and right ventricular end-diastolic volume (LVEDV and RVEDV), which were all indexed to body surface area as calculated by the Mosteller formula. LV remodeling was assessed using the LVM/LVEDV ratio. Temporal changes were calculated by subtracting each parameter at 1-year by its baseline value. Non-parametric Spearman correlation coefficients (r) were calculated for cross-sectional associations at baseline and for temporal changes. No adjustment was made for multiple testing. At 1-year follow-up, 57 participants underwent repeat CMR; 10 patients did not undergo follow-up CMR due to renal transplant (n = 2) or serious morbidity or death (n = 8). Characteristics of the study participants and CMR measurements at baseline and EOS were previously published [3]. The median age was 55 years (43–61) and 45% were female; 50.7% had arteriovenous fistula, 41.8% had central venous catheter and 7.5% had arteriovenous graft. The mean ± SD * Andrew T Yan [email protected]