In this study, we investigated the severity and impact of hypernatremia on mortality of neurocritically ill patients, depending on the degree of hypernatremia. Major findings of this study were as follows. First, poor clinical outcomes were more common in patients with hypernatremia compared with those without hypernatremia in overall and propensity score-matched population. Second, patients with mild hypernatremia in the matched population showed the best survival rate. Finally, multivariable analysis revealed that moderate and severe hypernatremia, use of vasopressors and GCS on ICU admission were significantly associated with in-hospital mortality and 28-day mortality in neurocritically ill patients. Specifically, mild hypernatremia was not associated with poor clinical outcomes in this study.
In neurocritically ill patients, hyperosmolar therapy, including mannitol, glycerin or hypertonic saline, are frequently used to control intracranial hypertension9,16. Hypernatremia may occur due to the use of these agents4,9,10,16. Mild hypernatremia is the goal of serum sodium when hypertonic saline is used to lower ICP continuously16. Neurocritically ill patients are frequently subjected to hyperosmolar therapy. Hyperosmolar therapy and associated acute kidney injury can aggravate hypernatremia11,17,18. In addition, hypothalamic dysfunction due to brain injury can contribute to hypernatremia11,19. Therefore, hypernatremia occurs easily in patients with severe neurological disease compared with those manifesting benign disease. Eventually, it is not easy to determine whether hypernatremia itself is associated with a poor prognosis, or patients with hypernatremia show poor prognosis because of their neurocritical illness. Therefore, a propensity score matching method was used to adjust for this confounder in this study. In brief, moderate and severe hypernatremia were significantly associated with poor clinical outcomes in neurocritically ill patients.
The prevalence of ICU-acquired hypernatremia was about 5.7% to 50.7% in previous studies1-3. Hypernatremia and its associated hyperosmolar conditions lead to metabolic derangement and organ dysfunction, including abnormal hepatic gluconeogenesis, decreased lactate clearance, increased insulin resistance of peripheral tissues, cardiac dysfunction, muscle cramps and rhabdomyolysis1,3,5. Therefore, hypernatremia itself may be associated with multiple complications, prolonged ICU stay, or even death1,6,8,11,20,21. In addition, hypernatremia leads to increased cellular dehydration and decreased cerebral edema, which are often the therapeutic goals in neurocritical care. However, these homeostatic changes can injure myelin and even cause neuronal death. Thus, hypernatremia leads to additional secondary brain injury11,22. Eventually, hypernatremia is also associated with poor clinical prognosis in neurocritically ill patients3,4,11,18,19,23,24.
In this study, neurocritically ill patients with mild hypernatremia did not manifest worse outcomes compared to those without hypernatremia. However, patients with moderate or severe hypernatremia had worse prognosis than those without hypernatremia. In neurocritically ill patients with mild hypernatremia, favorable clinical outcomes may be associated with the ICP-lowering effect induced by mild hypernatremia with fewer complications than those detected during moderate and severe hypernatremia. Therefore, hyperosmolar agents may not be reduced or discontinued to maintain normal level of serum sodium in case of mild hypernatremia during treatment of patients with intracranial hypertension.
This study has several limitations. First, this was a retrospective review of medical records and the data extracted from Clinical Data Warehouse. The nonrandomized nature of registry data may have resulted in selection bias. Laboratory studies, including serum sodium levels, were not protocol-based. Second, hypernatremia was easily induced with hypertonic saline. Although, a small number of patients used hypertonic saline in this study, they could not be identified from Clinical Data Warehouse due to technical challenges. Third, the distribution of neurosurgical diseases differed from that of the general neurosurgical ICU, and the proportion of patients with brain tumors was particularly high. Although this study still provides valuable insight, prospective large-scale studies are needed to confirm the safety and effectiveness of mild hypernatremia in neurocritically ill patients to arrive at evidence-based conclusions.