Modeling the Neurologic and Cognitive Effects of Hyponatremia

Acute hyponatremia causes swelling of brain cells and, if of a sufficient degree, increased intracranial pressure. This is a consequence of the relative impermeability of the blood-brain barrier to sodium ions¹; this principle underlies the clinical utility of hypertonic NaCl infusion in the treatment of the cerebral edema accompanying traumatic brain injury or stroke. That acute hyponatremic brain swelling should cause neurologic symptoms is unsurprising. Brain swelling is absent in chronic hyponatremia, in contrast, and the basis for the neurologic and cognitive symptoms in this context is less clear.

There is an emerging view that even mild-to-moderate degrees of chronic hyponatremia cause neurologic symptoms and that reversal of hyponatremia ameliorates them.² Data in support of this attractive hypothesis are few. Renneboog et al. tested gait and coordination in 16 patients with syndrome of inappropriate antidiuresis (SIAD) both on and off their chronic therapy (with mean plasma sodium concentrations of 138 and 128 mEq/L, respectively).³ They reported that the total traveled way, an index of balance performance and gait instability, was abnormal under hyponatremic conditions but normalized after correction of hyponatremia. All participants were outpatients at the time of evaluation; therefore, there was no confounding by acute medical comorbidity. To eliminate an additional potential source of bias, and in support of reversibility, half were initially tested on treatment, and half were tested in the untreated state.

Additional evidence comes from the Study of Ascending Levels of Tolvaptan in Hyponatremia-I (SALT-I) and SALT-II trials.⁴ These studies tested the effect of the vasopressin V₂-receptor antagonist, tolvaptan, on plasma sodium concentration in hyponatremic participants with an initial plasma sodium concentration <135 mEq/L. One secondary end point was a 12-Item Short Form Health Survey (SF-12) score. The SF-12 is a proprietary instrument that quantifies study participants’ self-reported well-being using 12 general questions addressing mood and health-related limitations on activity.⁵ After 30 days of tolvaptan treatment, during which the mean plasma sodium concentration increased from 129 to 136 mEq/L, the mental component of the SF-12 score significantly improved (P=0.02), whereas the physical component was unaffected. This metric is subjective (i.e., self-reported) and emphasizes feelings of depression and anxiety; it does not directly assess cognitive performance or neurologic deficit. In addition, there was no control for the order of treatment; all participants were initially tested when hyponatremic and then tested again after tolvaptan treatment. Nonetheless, these findings in a moderately large randomized controlled trial are consistent with a reversible neuropsychiatric effect of even mild hyponatremia.

Furthermore, in population-based studies, chronic hyponatremia associates with bone fracture risk^6–8; however, whether this is a consequence of a neurologic phenotype predisposing to falls⁹ or of a bone phenotype predisposing to fracture¹⁰ (or perhaps a combination of both) remains unclear.

The important study by Fujisawa and colleagues in this issue¹¹ uses a rat model of SIAD to assess the effects of hyponatremia on multiple domains of neurocognitive function at the behavioral, cellular, and molecular levels. Importantly, it also addresses reversibility of a number of these findings. In so doing, it brings a new level of integration and mechanistic sophistication to studies addressing neurologic sequelae of chronic hyponatremia.

To generate hyponatremia, the authors used a model of continuous 1-deamino-8 d-arginine vasopressin infusion at one of two doses via osmotic mini-pump to achieve either moderate (121 mEq/L) or severe (111 mEq/L) hyponatremia. In these chronic studies, the absence of cerebral edema was confirmed via an imaging-based approach and contrasted with the evident cerebral edema accompanying a model of acute hyponatremia.

Gait was investigated in detail. A prior rodent study failed to detect an effect of hyponatremia on gait¹²; however, Fujisawa et al. used a more sensitive assessment in which the rat traverses a glass plate and its footprints are recorded. Software then calculates print dimensions and the time and distance between successive footfalls. Hyponatremic rats exhibited a smaller stride and wider stance, consistent with an ataxic gait.

The authors also found that chronic hyponatremia increased anxiety levels in the open field test. In this test, rats are placed on a flat open surface that is marked with a grid and enclosed by walls. Grid crossings serve as an index of general locomotor activity and willingness to explore,¹³ whereas other variables (e.g., proportion of time spent in the center of the field) have been interpreted as indices of anxiety.¹⁴

Recognition memory was tested using the novel-object recognition test. In this test, a rat is presented with two similar objects during an initial session, and then one of the two objects is replaced by a new object during a second session. Because rodents typically spend more time investigating an unfamiliar object, the amount of time spent exploring the new object is an index of recognition memory.¹⁵ Recognition memory was impaired in the moderately and severely hyponatremic rats.

Associative memory was also tested using the contextual fear conditioning test. In associative memory, a previous experience is recalled by thinking of something that is linked with it, therefore invoking the association. In the specific case of contextual fear conditioning, an aversive stimulus (e.g., foot shock) is paired with an innocuous stimulus or cue (placement in a conditioning cage).¹⁶ On presenting the cue, freezing behavior is observed in anticipation of the shock. The latency or delay before the freezing response is an index of associative memory, and this was impaired in the hyponatremic groups.

Because vasopressin itself can influence memory and anxiety, the authors used an additional—and cleverly designed—control group consisting of vasopressin-treated rats made normonatremic. This was achieved by feeding 1-deamino-8 d-arginine vasopressin–treated rats with a high-salt liquid diet. The normonatremic vasopressin-treated rats did not differ from control rats in terms of anxiety, novel-object recognition, or associative memory, indicating that the hyponatremia itself—and not the exogenous vasopressin analog—was the basis for the aberrant behavior.

Importantly, with respect to human hyponatremia, correction of the chronic hyponatremia in the rat models with tolvaptan alleviated the increased anxiety and restored the impaired recognition and associative memory. Unfortunately, these studies were not bona fide correction studies (i.e., with measurements obtained first in the hyponatremic state and then again in the corrected normonatremic state); rather, the authors showed that corrected rats were no different from control normonatremic rats in these important phenotypes.

The authors went on to address the mechanism of memory impairment in chronic hyponatremia and show an adverse effect on hippocampal long-term potentiation. In the hippocampus, information from short-term memory is consolidated into long-term memory. Within this tissue, the phenomenon of long-term potentiation is believed to be a main neural mechanism for memory storage.¹⁷ Therefore, an effect of hyponatremia on this process is a neurophysiologic correlate of the hyponatremia-induced memory deficits. Because elevated extracellular glutamate impairs long-term potentiation, the authors tested for and observed, using microdialysis, that chronic hyponatremia increased extracellular glutamate levels in the hippocampus. The authors speculate on potential molecular mechanisms leading to the increase and demonstrate decreased glutamate uptake by astrocytes in cultured cells. Moreover, ataxic gait is also seen in both humans and mice with mutated variants of the glutamate transporter glutamate aspartate transporter, encoded by the SLC1A3 gene, providing additional circumstantial support for the authors’ glutamate hypothesis of hyponatremic hippocampal effects.

In aggregate, this set of investigations represents a milestone in our understanding of the neurocognitive effects of chronic hyponatremia. Clearly, the best way to treat overt symptomatology is to treat the underlying hyponatremia. However, this study and ones like it have the potential to more precisely define the threshold at which mild hyponatremia becomes symptomatic, through the use of sophisticated or invasive experimental approaches unavailable in the clinic.

It is important to emphasize that there are insufficient data to permit extrapolation of these key findings to clinical care; it should not be concluded that active correction of hyponatremia is now warranted in the absence of clinical symptoms, based solely on inferred neurologic impairment. Current treatment recommendations for hyponatremia distinguish between the symptomatic and asymptomatic patient. As the tools used to detect neurologic impairment become more sensitive, refinement of existing recommendations may be required. Until then, better data on the functional significance of human chronic hyponatremia and the threshold at which symptomatology is likely to emerge would be welcome.

Disclosures

None.