Abstract
BACKGROUND: Limited data are now available to evaluate the relationship between serum magnesium level, anemia and mortality in the dialysis population. METHODS: Using data from the Japanese Dialysis Outcomes and Practice Patterns Study (J-DOPPS) phases 5 and 6, we analyzed the association between serum magnesium (s-Mg) levels and the erythropoiesis-stimulating agents resistance index (ERI) as the primary outcome. To estimate the longitudinal relationship, a mixed-effect model was used with ERI at each 4-month period as the dependent variable and quintiles of s-Mg at the previous 4-month period as the independent variable. We also examined incidence of infectious events, and the all-cause and cardiovascular disease (CVD)-related deaths as secondary outcomes by Cox regression with quintiles of s-Mg at baseline. RESULTS: Of the 4776 participants in J-DOPPS, 1650 were included in the analysis. The median of s-Mg at baseline was 2.5 mg/dL. A significant linear association of s-Mg with ERI (P for trend <.001) was revealed. Low and high s-Mg levels were not associated with the clinical outcomes of interest, except for the highest quintile of s-Mg being significantly associated with lower incidence of all-cause mortality and CVD-related deaths compared with the middle (reference) quintile. CONCLUSIONS: We observed that lower s-Mg levels subsequently induced higher ERI and that mild higher s-Mg levels were possibly associated with good rather than poor outcomes in Japanese hemodialysis patients. Adjustment of s-Mg levels may be proposed as a new strategy at a low cost and risk to reduce the risk of premature mortality.