Serum Sodium Dynamics and Acute Kidney Injury in Pregnancy Associated Sepsis: Insights from Group-Based Trajectory Modeling

Abstract

Background

Pregnancy-associated sepsis poses significant maternal, fetal and neonatal risks, with acute kidney injury (AKI) being a critical complication. The dynamic relationship between serum sodium trajectories and AKI in this population remains unclear.

Methods

In this retrospective cohort study, 138 pregnant patients with sepsis were analyzed using the MIMIC-IV 3.1 database. Three serum sodium trajectory groups were identified via group-based trajectory modeling (GBTM). AKI was defined per KDIGO criteria. Analyses employed logistic regression, inverse probability weighting, multivariable adjustments and cubic spline models.

Results

Three distinct trajectories emerged: Group 1 (low initial sodium that subsequently increased, n = 34), Group 2 (stable sodium levels, n = 83), and Group 3 (high sodium levels throughout, n = 21). Groups 1 and 3 exhibited higher AKI incidence (ORs: 4.04 [95%CI: 1.63-9.96] and 3.97 [95%CI: 1.33-11.87], respectively; both p < 0.05), prolonged ICU stay hours(72, 118 vs 47, p < 0.001), and elevated SOFA scores (p = 0.01) compared to Group 2. Cubic spline analysis revealed a U-shaped risk relationship, with AKI incidence rising at sodium levels >145 mmol/L (p-value for overall was 0.037 and for nonlinear was 0.021).

Conclusions

Dynamic sodium trajectories, particularly low initial sodium that subsequently increased, independently predict AKI and adverse outcomes in pregnancy-associated sepsis.