Abstract Background and Aims Association between hypo- and/or hypermagnesaemia and cardiovascular (CV) outcomes or mortality showed conflicting results in chronic kidney disease (CKD) and was mostly conducted on total magnesium (tMg) levels and not ionized magnesium (iMg). Thus, the objectives of the present study were to (i) describe the serum iMg concentration in patients at various CKD stages, (ii) measure the correlation between iMg and tMg concentrations, (iii) identify their associated factors and (iv) determine whether serum tMg and/or iMg concentrations are associated with adverse outcomes, such as major adverse cardiovascular events (MACE) and death before kidney replacement therapy (KRT, defined as the initiation of chronic dialysis or kidney transplantation) in CKD patients, after adjustment for confounding factors. Method CKD-REIN is a French prospective cohort of patients with stage 3 to 5 CKD (estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m²) recruited from 40 nephrology facilities, neither on maintenance dialysis nor transplanted, and followed up for 5 years. Baseline iMg and tMg serum concentrations were centrally measured. Patients were compared by subgroups according to serum tMg tertiles (Tertile 1 (T1): ≤0.73 mmol/L, T2: 0.74-0.81 mmol/L, T3: >0.81 mmol/L). Multivariate linear regressions were used to identify factors associated with iMg and tMg serum concentrations. Adjusted cause-specific Cox proportional hazard models were used to test the association between the risk of first MACE, all-cause mortality, and CV mortality with serum iMg and tMg concentrations at baseline. Results Of the 2, 419 included patients, 66% were men, median age was 68, and the mean eGFR was 34.8 mL/min/1.73 m². Mean serum iMg and tMg concentrations were respectively 0.48 mmol/L and 0.77 mmol/L. They were strongly correlated (r = 0.89, P < .001) and were independently associated with eGFR. Overall, iMg and tMg concentrations were associated with the same factors. The adjusted HR [95% CI] for MACE associated with the baseline serum tMg level was 1.27 [0.95; 1.69] for patients in T1, and 1.56 [1.18; 2.06] for patients in T3, relative to patients in T2. The adjusted HR [95% CI] for MACE associated with baseline serum iMg was 0.93 [0.71; 1.23] for patients in T1, and 1.14 [0.88; 1.48] for patients in T3, relative to patients in T2. The risks of all-cause death (HR [95% CI] = 1.48 [1.11; 1.97]) and CV mortality (HR [95% CI] = 1.98 [1.27; 3.10]) were also increased in T3 of tMg compared with T2. The adjusted risks for mortality (all-cause or CV) associated with the baseline serum iMg level were not significantly different from one tertile to another. Conclusion Our analysis of a large cohort of patients with moderate-to-advanced CKD demonstrated that individuals with higher serum tMg concentrations, although still within the normal range, had a greater likelihood of MACE and mortality. Total magnesium is routinely assayed in hospital laboratories and might be a useful variable to monitor in CKD patients. We failed to demonstrate the added value of iMg when studying its association with CV outcomes and mortality.