Post-sepsis low magnesium status and newly recorded diabetes: a propensity score–matched cohort study

Abstract

Background Whether low magnesium status after sepsis is associated with subsequent diabetes remains uncertain. We examined the association between post-sepsis magnesium status and newly recorded diabetes among adult sepsis survivors. Methods We conducted a retrospective multicenter cohort study using the TriNetX Global Collaborative Network. Adults with a qualifying magnesium measurement between 2010 and 2023 and a history of sepsis were classified as having low magnesium (< 1.80 mg/dL) or normal magnesium (1.80–2.20 mg/dL). Patients with diabetes or hemoglobin A1c (HbA1c) ≥ 6.5% before the index date were excluded, and a 90-day landmark was applied. After 1:1 propensity score matching, 15,201 patients remained in each group. The primary outcome was a diabetes composite outcome, defined as newly recorded type 2 diabetes mellitus (T2DM) or HbA1c ≥ 6.5%. Secondary outcomes included the individual composite components, all-cause mortality, other diabetes diagnoses, and HbA1c ≥ 9.0%. Outcomes were assessed from day 90 through day 730 after the index date. Results The diabetes composite occurred in 3.31% of the low-magnesium group and 2.38% of controls (absolute risk difference, 0.93 percentage points; hazard ratio [HR], 1.43; 95% confidence interval [CI], 1.25–1.63). Associations were consistent for newly recorded T2DM (HR, 1.38; 95% CI, 1.19–1.59) and HbA1c ≥ 6.5% (HR, 1.51; 95% CI, 1.21–1.89). Low magnesium was also associated with higher all-cause mortality (HR, 1.22; 95% CI, 1.12–1.33) and other diabetes diagnoses (HR, 1.78; 95% CI, 1.12–2.84), whereas HbA1c ≥ 9.0% was not significantly increased. The association with the primary outcome remained generally consistent across several sensitivity analyses and when severe magnesium deficiency (< 1.5 mg/dL) defined the exposure (HR, 1.53; 95% CI, 1.26–1.85). Conclusions Among adult sepsis survivors without recognized diabetes, low magnesium status was associated with a modestly higher subsequent risk of newly recorded diabetes and mortality. These findings support magnesium as a potential post-sepsis risk marker but do not establish causality.

The authors' abstract, as published at the source. Frontiers in Endocrinology, 2026 · DOI ↗