Predictive value of hyponatremia for short-term mortality in supratentorial spontaneous intracerebral hemorrhage: a single center study

Abstract

Background: Hyponatremia is a common electrolyte disturbance in patients with neurological disease; however, its predictive role for outcome in patients with supratentorial spontaneous intracerebral hemorrhage (sICH) is controversial. This study aims to explore the association between hyponatremia within 7 days after bleeding and 90-day mortality in patients with supratentorial sICH.

Methods: A retrospective analysis was conducted at our institution. Patients with sICH meeting the inclusion criteria were enrolled in this study. Multivariate regression analyses were performed to determine the predictive value of hyponatremia (serum sodium <135 mmol/L) for 90-day mortality and functional outcome. Subgroup analysis was performed based on the degree and duration of hyponatremia and therapeutic strategies. The Spearman correlation test was performed to explore the relationship between hyponatremia severity and duration with variables in a multivariate regression model. Kaplan-Meier curve was depicted to reveal the relationship between hyponatremia and mortality. The receiver operating characteristic (ROC) curve was plotted to show the diagnostic effect of the minimum concentration of serum sodium (sodium_min) on 90-day mortality.

Results: A total of 960 patients were enrolled, 19.6% (188) of whom were patients with hyponatremia and 26.0% (250) had 90-day mortality. The incidence of hyponatremia was roughly 2.5 times in non-survivors compared with survivors (34.8% vs. 14.2%). Multivariate regression analysis revealed that hyponatremia was the independent predictor of 90-day mortality (OR 2.763, 95%CI 1.836-4.157) and adverse outcome (OR 3.579, 95%CI 2.332-6.780). Subgroup analysis indicated an increased trend in mortality risk with both duration (more or less than 48 h) and severity of hyponatremia (mild, moderate, and severe) and confirmed the predictive value of hyponatremia for mortality in patients undergoing surgical intervention (external ventricular drainage, craniotomy, and decompressive craniectomy; all p < 0.05). The Spearman correlation test indicated no moderate or strong relationship between hyponatremia severity and duration with other variables in the multivariate model (all |r_s| < 0.4). The ROC curve suggested the moderate diagnostic performance of sodium_min for mortality in both general patients and subgroups of therapeutic method patients (AUC from 0.6475 to 0.7384).

Conclusion: Hyponatremia occurring in the first 7 days after bleeding is an independent predictor of 90-day morality and adverse outcome. Rigorous electrolyte scrutiny in patients treated surgically is required.

Keywords: 90-day; functional outcome; hyponatremia; intracranial hemorrhage; mortality.

Figure 1

The flow diagram of patient selection.

Figure 2

Subgroup analysis to explore the predictive value of hyponatremia in different conditions. EVD, external ventricular drainage; DC, decompressive craniectomy. *Model adjusted for age, Graeb score, hemorrhage volume, diabetes mellitus, GCS score, and ICH score. ^Model adjusted for age, Graeb score, diabetes mellitus, GCS score, and ICH score.

Figure 3

Kaplan-Meier curve indicates the relationship between hyponatremia and mortality in the whole cohort and subgroups with different therapeutic strategies. (A) the whole cohort. (B) Conservative treatment group. (C) EVD group. (D) Craniotomy group. (E) DC group.

Figure 4

ROC curve demonstrates the moderate diagnostic value of sodium_min in mortality in the whole cohort and subgroups with different therapeutic strategies.

Figure 5

Correlations between degrees of hyponatremia and factors in multivariable models. The clockwise and counterclockwise graphs demonstrate positive and negative correlations, respectively. *, **, and *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively. (A) Presents the correlations in the whole cohort. (B) The group with adverse outcomes. (C) Shows patients who had died by the 90-day follow-up. EVD external ventricular drainage; DC decompressive craniectomy.

Figure 6

Correlations between duration of hyponatremia and factors in multivariable models. The clockwise and counterclockwise graphs demonstrate positive and negative correlations, respectively. *, **, and *** indicate p < 0.05, p < 0.01, and p < 0.001, respectively. (A) Presents the correlations in the whole cohort. (B) The group with adverse outcomes. (C) Shows patients who had died by the 90-day follow-up. EVD external ventricular drainage; DC decompressive craniectomy.

Figure 7

Distribution of the modified Rankin Scale (mRS) score at 90-day follow-up for patients with and without hyponatremia.