Association Between Early Hyponatremia and Clinical Outcomes in Critically Ill Patients: A Retrospective Cohort Study.

Abstract

Hyponatremia, frequently encountered in intensive care (ICU) settings, plays a critical role in shaping patient outcomes. Despite its prevalence, contemporary research into its newly classified severity categories and their implications on mortality, renal function, and length of stay remains limited. This study aims to fill this gap by examining the impact of hyponatremia severity on these critical outcomes. A retrospective analysis of ICU patients aged >18 years who were admitted between March 2019 and December 2022 was conducted at Hamamatsu University Hospital, Shizuoka, Japan. Patients who were readmitted or had incomplete data were excluded. Hyponatremia was categorized as mild (130-135 mmol/L), moderate (125-129 mmol/L), or severe (<125 mmol/L), following the criteria set by the European Society of Intensive Care Medicine. This classification utilized thelowest sodium concentration within 24 hours of ICU admission. The outcomes were in-hospital mortality, ICU mortality, newly implemented renal replacement therapy (RRT), and length of hospital and ICU stay. Outcomes were analyzed using multivariable logistic and linear regression models, adjusting for relevant covariates including age, sex, Acute Physiology and Chronic Health Evaluation (APACHE) III scores, and the use of mechanical ventilation. Of the 3,538 patients analyzed, 1,072 (30.3%) experienced hyponatremia: 894 (25.3%) mild, 144 (4.1%) moderate, and 34 (1.0%) severe. Multivariable analysis revealed no significant association between hyponatremia severity and in-hospital mortality rates across normonatremia (3.8%), mild (5.2%), moderate (11.8%), and severe (23.5%) groups, nor with ICU mortality. However, compared to normonatremia, moderate and severe hyponatremia were associated with increased RRT initiation (odds ratios = 3.83 and 6.36, respectively) and prolonged hospital stay (mean difference = 7.06 and 9.66 days, respectively), and ICU stays (mean difference, 1.02 and 2.70 days, respectively). Mild hyponatremia was not significantly associated with RRT or length of stay. Moderate-to-severe hyponatremia did not influence mortality but was associated with increased RRT initiation and prolonged hospital and ICU stay. By contrast, mild hyponatremia was not associated with any clinical outcome.Further research is required to determine if correcting hyponatremia directly improves ICU patient outcomes, given the observational nature of the study.