Background
Both hyponatremia and hypernatremia are important risk factors for high morbidity and mortality in several clinical settings [
1,
2]. However, reports have conflicted about the dependent and independent relationships of sodium with mortality. A relationship is plausible because both hyponatremia and hypernatremia have adverse consequences via their effects on the brain, heart, and other organs [
3,
4]. However, dysnatremia might be only a marker of comorbidities, with deaths attributable mainly to the severity of the diseases and not to dysnatremia itself [
5]. A recent review of hyponatremia cases concluded that deaths were associated with the underlying illness, rather than with the severity of hyponatremia [
6]. Given the nature of this issue, it is unclear whether the correction of sodium can reduce the mortality risk. Most randomized controlled trials of drugs for sodium correction have not demonstrated a survival advantage [
7]. Although these results have not excluded the need for sodium correction, it is necessary to address whether sodium correction alone can reduce mortality. A well-designed observational study could be helpful because randomized controlled trials for this issue are not feasible.
Continuous renal replacement therapy (CRRT) is increasingly used in severe acute kidney injury (AKI), because it can easily control biochemical imbalances due to AKI [
8]. The presence of AKI or comorbidities hinders the effectiveness of conventional fluid or drug management, but CRRT can overcome these difficulties and achieve normonatremia. Here, we first addressed how sodium levels predict mortality in a prospective cohort of patients receiving CRRT, in whom the control of dysnatremia was extremely difficult with conventional treatment alone. Whether sodium correction is independently associated with reduced mortality in comorbid conditions was also addressed.
Discussion
The present study enrolled patients with specific characteristics (extremely severe AKI and high mortality and morbidity risks); thus, CRRT was the preferred method for correction of imbalances in biochemical parameters, including dysnatremia. The results report for the first time a correlation between baseline sodium level and mortality risk in a group undergoing CRRT. However, the change in sodium, as a marker of sodium correction, was not associated with mortality risk. These analyses were helpful in determining whether the correction of dysnatremia independently reduces mortality in the presence of several comorbidities.
Dysnatremia is associated with various negative consequences, such as decreased brain function [
3], compromised cardiac contractility [
4], increased insulin resistance [
14], abnormalities in neuromuscular function [
15], creation of an inflammatory milieu [
16], and aggravation of interstitial edema [
17]. Based on these pathologic conditions, it is plausible that dysnatremia contributes to mortality risk. The present study demonstrated that dysnatremia predicted mortality in a CRRT cohort. For patients undergoing CRRT, any of the negative conditions driven by dysnatremia could have contributed to the high mortality.
Similar to the present results with CRRT, dysnatremia has been associated with high mortality in other several clinical settings, even with normal kidney function [
18]. Most studies adjusted for various comorbidities in their analyses, thus suggesting an independent association between sodium and mortality. However, this methodological approach is not sufficient to draw a final conclusion. Other important data are needed, including whether each death is directly related to dysnatremia, and whether the reduction or treatment of dysnatremia itself reduces mortality. A previous analysis of cases with severe hyponatremia indicated that deaths were primarily caused by underlying illness and not by hyponatremia itself [
6]. Furthermore, two thirds of those patients died, although their sodium levels returned to normal or near normal ranges. No studies have investigated the effects of sodium correction on mortality by using randomization of participants and controlling for confounders.
In addition to the correlation issue between dysnatremia and mortality, it remains unresolved whether the correction of sodium could reduce mortality, particularly in the presence of several comorbidities. A study of this issue is difficult to conduct. Thus, the present study might be helpful, despite its observational design, in understanding the correlation. In contrast to the baseline level, subsequent sodium levels did not predict mortality irrespective of timeframe (24 and 72 h), possibly because of case-mix results, such as the group with baseline dysnatremia later having normonatremia. Furthermore, the change in sodium level, including the appropriate correction of dysnatremia, was not correlated with mortality. These results suggested that sodium level itself might accurately predict mortality in association with comorbidities or other related factors; however, the correction of sodium alone did not contribute to a survival benefit, particularly under the strong influence of comorbidities.
The association with mortality may differ for hyponatremia and hypernatremia. In the present cohort, mild hypernatremia was not associated with mortality in contrast to the results of mild hyponatremia. The discrepancy between the effects of hypo- and hypernatremia has been reported in other observational studies [
19‐
21]. Each association with mortality could be sensitive to the disease state of study. However, although the baseline characteristics differ across studies, hyponatremia and hypernatremia should be considered separately in research, as well as in clinical practice.
Although the present results are informative, this study has some limitations. First, the observational study design limits the drawing of conclusions, because of causality. However, randomized trials of dysnatremia treatment to determine survival benefit are not feasible, particularly in the presence of comorbidities; thus, the present results might be helpful in raising awareness of the correlation between sodium correction and outcomes. Second, the present cohort of patients who underwent CRRT is a particular subgroup, thus hindering the applicability of the conclusions to other populations. Furthermore, a single therapeutic intervention (i.e., sodium correction) might not show a survival benefit under conditions of high comorbidity rates of CRRT patients [
22]. Third, sodium levels corrected for glucose may not be accurate, because the correction factor may be altered by patient characteristics or high glucose states [
23,
24]. Lastly, we did not retrieve important variables (e.g., correction rate of dysnatremia and neurologic outcomes), which could affect the overall outcome.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
SSH designed the study, collected the data, analyzed and interpreted the results, and drafted the manuscript. EB collected the data. DKK and YSK designed the study and collected the data. JSH conceived the study and interpreted the data. KWJ conceived the study, analyzed the results, interpreted the data, and reviewed the manuscript. All of the authors read and approved the final manuscript.