Dose-response association between early postoperative serum magnesium and in-hospital mortality after cardiac surgery: a retrospective cohort study in MIMIC-IV with external replication in the eICU-CRD
Year: 2026
Type: article
Abstract: Background: Hypomagnesemia is common following cardiac surgery and is associated with arrhythmias and adverse outcomes. In postoperative cardiac intensive care unit (ICU) patients, cardiopulmonary bypass (CPB), hemodilution, perioperative electrolyte administration, and acute changes in renal function may alter early serum magnesium concentrations and their clinical interpretation. However, the dose-response relationship between early postoperative serum magnesium levels and mortality in this setting remains unclear. We investigated this association and sought to characterize its dose-response pattern. Methods: This retrospective cohort study included adult cardiac surgery patients admitted to the ICU in the Medical Information Mart for Intensive Care IV (MIMIC-IV; version 3.1; 2008-2022) who had an ICU stay of at least 24 hours and a serum magnesium measurement within 24 hours of ICU admission. The exposure was the first recorded serum magnesium concentration, modeled continuously for the principal dose-response analysis and secondarily categorized into quartiles. Baseline covariates were obtained from ICD-coded diagnoses and electronic health record data. The primary outcome was in-hospital mortality, ascertained from hospital discharge records in MIMIC-IV and unit or hospital discharge status in the eICU Collaborative Research Database (eICU-CRD). Multivariable logistic regression, restricted cubic spline (RCS) analysis, and piecewise regression were performed, with the threshold estimated from the data. The same eligibility criteria were applied in the eICU-CRD, and the replication analysis used a harmonized covariate set. Results: The MIMIC-IV cohort included 7,783 patients {median age 68 years [interquartile range (IQR), 60-75 years]; 27.9% women; 70.8% underwent coronary artery bypass grafting (CABG); median Sequential Organ Failure Assessment (SOFA) score 7}, with 130 in-hospital deaths (1.7%). RCS analysis showed an overall association (P=0.003) compatible with an L-shaped pattern, although the formal test for nonlinearity was not significant. Piecewise regression with a threshold of 2.5 mg/dL showed that each 1 mg/dL increase below this threshold was associated with lower odds of mortality [odds ratio (OR) =0.35; 95% confidence interval (CI): 0.18-0.68; P=0.002], whereas no association was observed above the threshold (OR =0.97; P=0.89). In the fully adjusted model, compared with Q2 (2.1-2.3 mg/dL), Q3 (2.4-2.6 mg/dL) was associated with lower odds of mortality (OR =0.50; 95% CI: 0.26-0.95; P=0.03). Sensitivity analyses were broadly consistent, and exploratory subgroup estimates were directionally similar overall. The eICU-CRD cohort included 3,929 patients (median age 68 years; 31.1% women; median SOFA score 8), with 67 deaths (1.71%); the Q3 estimate was directionally consistent but imprecise (OR =0.64; 95% CI: 0.29-1.36). Conclusions: Early postoperative serum magnesium was associated with in-hospital mortality after cardiac surgery, with a hypothesis-generating inflection near 2.5 mg/dL and an association concentrated in the lower magnesium range. Because this is an observational analysis, the findings do not establish a treatment target or demonstrate that increasing magnesium to a specific concentration improves outcomes. Prospective studies are needed to validate this pattern and assess its relevance to postoperative risk stratification.
Source: Journal of Thoracic Disease
Authors: Jinjing Wu, Wen Wen, Peirong Lin, Sheng Wang
Institutions: Capital Medical University, Beijing Anzhen Hospital
Language: English
Subfield: Nutrition and Dietetics
Field: Nursing
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7222623604