Sodium and Its Impact on Outcome After Aneurysmal Subarachnoid Hemorrhage in Patients With and Without Delayed Cerebral Ischemia

OBJECTIVES: To perform a detailed examination of sodium levels, hyponatremia and sodium fluctuations, and their association with delayed cerebral ischemia (DCI) and poor outcome after aneurysmal subarachnoid hemorrhage (aSAH). DESIGN: An observational cohort study from a prospective SAH Registry. SETTING: Tertiary referral center focused on SAH treatment in the Amsterdam metropolitan area. PATIENTS: A total of 964 adult patients with confirmed aSAH were included between 2011 and 2021. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: A total of 277 (29%) developed DCI. Hyponatremia occurred significantly more often in DCI patients compared with no-DCI patients (77% vs. 48%). Sodium levels, hyponatremia, hypernatremia, and sodium fluctuations did not predict DCI. However, higher sodium levels were significantly associated with poor outcome in DCI patients (DCI onset –7, DCI +0, +1, +2, +4, +5, +8, +9 d), and in no-DCI patients (postbleed day 6–10 and 12–14). Also, hypernatremia and greater sodium fluctuations were significantly associated with poor outcome in both DCI and no-DCI patients. CONCLUSIONS: Sodium levels, hyponatremia, and sodium fluctuations were not associated with the occurrence of DCI. However, higher sodium levels, hypernatremia, and greater sodium fluctuations were associated with poor outcome after aSAH irrespective of the presence of DCI. Therefore, sodium levels, even with mild changes in levels, warrant close attention.

To predict DCI, countless biomarkers related to the multifactorial pathophysiology of DCI have been investigated (10,11).Essentially, any event that decreases cerebral perfusion pressure (CPP) and cerebral blood flow (CBF) could contribute to the development of DCI (12)(13)(14).As volemic status is closely related to CPP and CBF, inducing hypervolemia was recommended in the past with the assumption of increasing CPP and CBF, but studies failed to show beneficial effects (15).However, studies did show that fluid restriction was associated with a higher occurrence rate of DCI (16)(17)(18).Therefore, current guidelines recommend euvolemia and fluid management guided by fluid balance (8,9,19).Because euvolemia is not clearly defined and subject to interpretation, this has often been translated to maintaining a positive fluid balance in clinical practice.However, recent studies showed that higher fluid intake was also associated with DCI and fluid restriction without compromising cardiac preload and CBF was possible (20)(21)(22), which conflicts with the knowledge on which guidelines are based (16)(17)(18).
The main objective is to examine the relation between sodium, dysnatremia, sodium fluctuations, and DCI and poor outcome at 6 months after aSAH, while taking the day of DCI onset into account.

Patient Population
A single-center cohort study from a prospective SAH Registry was performed.This registry included consecutive SAH patients admitted to the Amsterdam UMC, a tertiary SAH referral center in the Amsterdam metropolitan area.Adult patients (≥ 18 yr) admitted between December 2011 and December 2021 were included if SAH was confirmed by noncontrast computer tomography/ lumbar puncture due to a ruptured aneurysm confirmed by CT-angiography and/or digital subtraction angiography, with at least one sodium measurement in blood in the first 14 days after ictus.Patients with non-aSAH, death within 3 days after ictus and unavailable DCI status were excluded.This study was conducted in accordance with the Helsinki Declaration.Ethical approval by the institutional review board of the Amsterdam UMC was not necessary as this study did not fall under the board's guidelines as human subjects research.

Data Collection
The following data were used from the prospective SAH Registry: age, sex, hypertension, hypercholesterolemia, cardiovascular disease, World Federation of Neurosurgical Societies scale (WFNS) at treatment center admission (54), modified Fisher grade (55), aneurysm location, treatment modality, DCI defined according to Vergouwen et al (56), day of DCI onset, complications during admission (rebleeding, hydrocephalus, meningitis, and seizures), and clinical outcome at 6 months assessed with the modified Rankin Scale (mRS) (57)(58)(59).Sodium levels were extracted retrospectively from the electronic patient records.Definitions of the variables are described in Supplementary materials (http://links.lww.com/CCM/H480).

Sodium
As DCI occurs mainly between days 3-14 after ictus, sodium levels during the first 14 days after ictus

KEY POINTS
Question: To determine the relation between sodium levels, occurrence of hyponatremia, hypernatremia, fluctuations of sodium levels and delayed cerebral ischemia (DCI), and poor outcome after aneurysmal subarachnoid hemorrhage (aSAH).
Findings: In this single-tertiary center cohort study, daily sodium levels, hyponatremia, hypernatremia, and fluctuations of sodium levels were not predictive of DCI.Higher sodium levels, hypernatremia, and greater fluctuations of sodium levels were significantly associated with poor outcome in both patients with and without DCI.
Meaning: Sodium levels, even with mild changes, appear to have a negative impact on clinical outcome after aSAH.
were collected if available (postbleed day [PBD] 0-14) (60).If sodium was measured multiple times a day, the lowest sodium level was used for analysis.Hyponatremia was defined as a sodium level less than 135 mmol/L.Hypernatremia was defined as a sodium level greater than 145 mmol/L.Sodium fluctuation was defined as the difference between the highest and lowest absolute sodium level per mmol/L.
In DCI patients, sodium levels were centered around the day of DCI onset.The occurrence of dysnatremia and sodium fluctuations in DCI patients were evaluated in three distinct time intervals; between ictus and DCI onset (before DCI onset), between DCI onset and PBD 14 (after DCI onset), and during admission (PBD 0-14).This was done to gain a better understanding in whether hyponatremia and sodium fluctuations precede or follow DCI onset.

Patient Management
Patients were treated according to our institutional protocol based on (inter)national guidelines (8,19).A detailed description is available in the Supplementary materials (http://links.lww.com/CCM/H480).

Statistical Analysis
The first main outcome was the occurrence of DCI during admission.The second main outcome was poor outcome (mRS 4-6) at 6 months after aSAH.The normality of the data was tested using the Shapiro-Wilk test (W ≤ 0.90 was considered skewed).Data were presented as absolute numbers with percentages (%), median with interquartile range or mean with sd.The chi-square, Fisher exact-test, Mann-Whitney U, or independent T-test were used accordingly for analyzing group differences.
Sodium levels and fluctuations were graphically visualized and compared during PBD 0-14 in subgroups based on DCI status, WFNS (WFNS 1-2 and 3-5) and location of ruptured aneurysm (anterior communicating artery [ACOM] and no ACOM).For DCI, we compared the sodium levels and fluctuations before DCI onset and after DCI onset of DCI patients with sodium levels and fluctuations during admission of no-DCI patients.
The predictive ability of sodium-related variables for DCI was assessed with logistic regression analyses by calculating odds ratios (OR) and 95% CIs.Daily sodium levels before DCI onset in DCI patients and daily sodium levels during admission in no-DCI patients were used as the independent variable.Adjusted OR with 95% CI were calculated using multivariate logistic regression analyses with adjustments for known predictors of DCI (age, sex, WFNS, and modified Fisher grade) (61).Additionally, to evaluate whether the presence of DCI is predictive of sodium-related variables, logistic or linear regression analyses were repeated accordingly with DCI as the independent variable.In linear regression analyses, unadjusted and adjusted regression coefficients with 95% CI (Beta and aBeta, respectively) were calculated.All analyses were repeated for poor outcome versus good outcome.IBM SPSS Statistics (version 28) and GraphPad Prism (version 9.1.0)were used for statistical analyses and graphic presentations, respectively.Statistical significance was set at two-sided p value of less than 0.05.

Sodium and DCI
Before DCI onset, sodium levels in patients who would develop DCI were not significantly different from patients who did not develop DCI (Fig. 1A; and Table S2A, http://links.lww.com/CCM/H480).Multivariate logistic regression analysis showed that daily sodium levels were not predictive of the development of DCI (Table S4, http://links.lww.com/CCM/H480).After the onset of DCI, sodium levels were significantly lower in DCI patients than no-DCI patients (Fig. 1B; and Table S2B, http://links.lww.com/CCM/H480).Multivariate linear regression analysis showed that the presence of DCI was significantly predictive of lower sodium levels on PBD 3-14 (Table S5, http://links.lww.com/CCM/H480).
Before and after the onset of DCI, sodium fluctuations were not significantly different in DCI patients from no-DCI patients (Fig. S2, A and B, http://links.lww.com/CCM/H480).Sodium fluctuations were not predictive of the development of DCI in multivariate logistic regression analysis (Table 1).The presence of DCI was significantly predictive of smaller sodium fluctuations in multivariate linear regression analysis (aBeta -1.25; 95% CI, -2.11 to -0.39; p = 0.005) (Table S6, http://links.lww.com/CCM/H480).
Within the no-DCI group, patients with poor outcome had significantly lower sodium levels on PBD 0 and 1, but significantly higher sodium levels on PBD 5-10 and PBD 12-14 than patients with good outcome (Fig. 2B).In multivariate logistic regression analysis, higher sodium levels on PBD 6-10 and PBD 12-14 were significantly associated with poor outcome in no-DCI patients (Table S7B, http://links.lww.com/CCM/H480).Within DCI and no-DCI patients, not hyponatremia but hypernatremia (before DCI onset, after DCI onset and during admission) was significantly associated with poor outcome in multivariate logistic regression analysis (Table S8, A and B, http://links.lww.com/CCM/H480).

DISCUSSION
This study in a large cohort of aSAH patients shows that sodium levels, hyponatremia and fluctuations of sodium levels are not predictive of the development of DCI.However, higher sodium levels, hypernatremia and greater sodium fluctuations were associated with poor clinical outcome, irrespective of the occurrence of DCI.
Since 1985, the importance of volemic status and dysnatremia has been emphasized as a possible risk factor for DCI (16)(17)(18), leading to the recommendation of maintaining euvolemia and preventing/treating hyponatremia to prevent DCI (8,9,19).Thus, the use of fludrocortisone or hypertonic saline was deemed reasonable, although evidence is weak (8,19,62).Although hyponatremia is one of the most common electrolyte disturbances after aSAH, guidelines vary widely on management of dysnatremia and often lack recommendations on electrolyte monitoring (8,9,(63)(64)(65)(66), perhaps resulting in practice variation across hospitals worldwide.Despite these recommendations, studies still find that hyponatremia occurs more often in DCI patients (34,35), but with conflicting associations between hyponatremia, sodium fluctuations, and DCI (31,34,38,(40)(41)(42)(43)(44).Significantly lower sodium levels, within the hyponatremia (30) and normal range (39), and natriuresis were observed more in DCI patients than no-DCI patients (30,32).Similarly, evidence regarding poor outcome and sodium is conflicting; some find significant associations between poor outcome, hyponatremia (35,47), hypernatremia (31,37,46,48,49) or both (50), whereas others do not find any significant associations (34,42).Furthermore, a recent randomized trial showed that continuous infusion of 20% hypertonic saline for a minimum of 48 hours did not have a significant effect on clinical outcome than standard care after moderate to severe traumatic brain injury (TBI) (67).This is despite the observation that patients in the intervention group had higher blood osmolality and sodium levels than the control group.Hereby suggesting a lack of effect of sodium on clinical outcome.However, only moderate to severe TBI patients were included and sodium levels were measured only the first 5 days, while we demonstrate that sodium levels start to differ around day 4 after aSAH between patients with poor and good outcome.Additionally, the power of the trial was limited, the CIs for the findings were wide and the pathophysiological mechanisms triggered by aneurysmal rupture is essentially different from TBI, which makes it difficult to extrapolate these results to aSAH.Interestingly, recent studies consistently show an association between sodium fluctuations and poor outcome (42,43,49,53), and patients with poor outcome have significantly higher mean sodium levels than good outcome, which is in line with our findings (52,53).In previous reports, sodium and free water disorders were more often observed in patients with a ruptured ACOM aneurysm; however, we did not find differences between these patients (68).Comparison of different studies is virtually impossible due to heterogeneity in the definition of hyponatremia and sodium fluctuations, methodology, frequency/timing of sampling, without regards to the day of DCI onset and the timing of clinical outcome assessment (11,68).Similarly, there is significant variation in fluid management across hospitals (21,69,70).Of note, it is well known that fluid balance does not reflect volemic status adequately (71) and higher fluid input, besides fluid restriction, is also associated with DCI (20,22) and poor outcome (21).Vergouw et al (20) even showed that significant reduction of fluid input was possible while maintaining adequate cardiac preload and thus CBF.This suggests that current approaches for assessing and maintaining optimal volemic status in aSAH need to be further optimized.
We show that lower sodium levels, including hyponatremia, hypernatremia, and greater sodium fluctuations are not predictive of the development of DCI.Rather, significantly lower sodium levels are observed in patients after DCI occurs compared to patients without DCI, and the presence of DCI seems to be predictive of developing hyponatremia, lower sodium levels, and smaller sodium fluctuations.This can be partially explained as DCI patients received hypertension induction, which is standard treatment of DCI in our institution.Hypertension can cause natriuresis, resulting in lower sodium levels, while noradrenalin is associated with reduced natriuresis (72).However, enhanced salt loss due to cerebral salt wasting or a more pronounced syndrome of inappropriate antidiuretic hormone (8,(25)(26)(27)(28)(29) due to DCI, can further contribute to lower sodium levels.Of note, from our data it is not possible to differentiate the underlying mechanisms of reduced sodium levels.Interestingly, we find that higher sodium levels, hypernatremia, and greater sodium fluctuations (independently) are associated with poor outcome in DCI and no-DCI patients.To the best of our knowledge, this has not been demonstrated before in this manner.Although, sodium levels, including changes, seem to be the result of DCI, our results suggest that changes or any deviation from the physiologic state of sodium levels of the patient could have a negative impact on the clinical outcome.However, the influence of other medication, (excessive) fluid intake or cardiac dysfunction, which could affect the sodium/water balance and volemic status, could not be ruled out except for hypertonic saline use.Since relatively great sodium fluctuations can occur within the clinically accepted normal range of sodium, a single measurement of sodium may fail to capture these relevant changes.Therefore, the focus should perhaps shift to sodium fluctuations, or even osmolality.Minimizing sodium fluctuations to maintain a stable sodium trajectory and osmolality might be of more importance than sodium levels and correcting hyponatremia.To minimize sodium fluctuations, a narrower range of "normal" sodium levels could be considered with the upper limit being lowered as patients with poor outcome had higher sodium levels than patients with good outcome.This should be considered carefully as the management of sodium may have limited influence on outcome due to the multifactorial and not-fully comprehended nature of the pathophysiology triggered by aSAH (10,73).
Because previous studies have shown that fluid restriction and excessive fluid intake are risk factors for DCI and poor outcome, and our findings suggest that sodium is associated with poor outcome, future research should focus on sodium fluctuations while taking volemic status, underlying etiology of dysnatremia, medication and cardiac function into account (8,(16)(17)(18)74).Monitoring sodium levels more than once a day in patients will provide a more accurate reflection of the sodium trajectory after aSAH.Furthermore, recording fluid balances prospectively may improve its reliability.If the importance of sodium and fluid balance is established, it might create incentive to record fluid balances accurately on the wards despite its labor-intensive nature or new ways to assess this.This study shows a robust signal and could serve as a prequel to future trials with a rigorous preselection of patients targeting the underlying causes of dysnatremia while controlling for confounders such as hypertension or hypotension and DCI.This could fill the gaps in knowledge and future directions (74).
Strengths of our study include the prospective registry use with standardized/structured outcome assessment, large sample size, and taking the day of DCI onset into account.Limitations are the retrospective collection of sodium levels resulting in missing and variation in available sodium levels between patients (ascertainment bias), conducting multiple statistical analyses (type 1 error), an underestimation or overestimation of the occurrence of hyponatremia and imprecise determination of the timing of hyponatremia occurrence (75).Oral/IV fluid intake, medications affecting sodium/water balance, and fluid balances were not included as it is often improperly recorded and unreliable when collected retrospectively (76)(77)(78)(79)(80). Lastly, patients who died within 3 days were excluded, which could affect the generalizability of our findings.

CONCLUSIONS
Hyponatremia and sodium fluctuations still occurred frequently in our aSAH cohort despite our institutional protocol, but was not associated with DCI.The changes in sodium levels observed in DCI patients are likely the result of DCI and its therapeutic management.However, disorders in sodium levels, including mild changes, appear to have a negative impact on clinical outcome at 6 months in aSAH patients with and without DCI.

Figure 1 .
Figure1.Daily sodium levels in patients with and without delayed cerebral ischemia (DCI).A, In patients with DCI, only sodium levels before the onset of DCI were used for analyzing group differences.For example, if a patient developed DCI on postbleed day (PBD) 4, only sodium levels on PBD 0, 1, 2, and 3 were included in the analysis.B, In patients with DCI, only sodium levels after the onset of DCI were used for analyzing group differences.For example, if a patient developed DCI on PBD 4, only sodium levels on PBD 5, 6, and so forth were included in the analysis.Days are presented as PBD with PBD 0 as the day of aneurysmal subarachnoid hemorrhage (aSAH) ictus.The number of available sodium measurements (n = x) per day are noted under the x-axis.*p < 0.05, **p < 0.001.

Figure 2 .
Figure 2. Daily sodium levels in: A, delayed cerebral ischemia (DCI) patients with poor and good outcome at 6 months after aneurysmal subarachnoid hemorrhage (aSAH).Days are centered around the day of DCI onset (DCI 0).B, Daily sodium levels in patients without DCI with poor and good outcome at 6 months after aSAH.Days are presented as postbleed days (PBD) with PBD 0 as the day of aSAH ictus.The number of available sodium measurements (n = x) per day are noted under the x-axis.*p < 0.05, **p < 0.001.

Figure 3 .
Figure 3. Median (interquartile range) sodium fluctuations (A) in delayed cerebral ischemia (DCI) patients with poor and good outcome, and (B) in patients without DCI with poor and good outcome.Sodium fluctuations were defined as the difference between the minimum and maximum sodium level in the concerning time interval.The number of available sodium fluctuation (n = x) per interval are noted under the x-axis.*p < 0.05, **p < 0.001.

TABLE 1 .
a Adjusted for age, sex, WFNS, and modified Fisher grade.For the multivariate logistic regression six patients were excluded due to unknown WFNS grade (n = 5) and unknown modified Fisher grade (n = 1).b with DCI were scored as hypernatremia if they had a sodium level > 145 mmol/L before the onset of DCI.Patients without DCI were scored as hypernatremia if they had a sodium level > 145 mmol/L at any time point during PBD 0-14.d Univariate logistic regression analysis with sodium fluctuations as the independent variable.This analysis was carried out in 929 patients of whom 261 (28%) eventually developed DCI..
Supplemental digital content is available for this article.Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal's website (http://journals.lww.com/ccmjournal).Drs Verbaan and Müller contributed equally to this work.Dr. Labib had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.Drs.Labib, Tjerkstra, Vandertop, Verbaan, and Muller were involved in study concept, design, acquisition, analysis, or interpretation of data.Drs.Labib and Tjerkstra were involved in drafting of the article.Drs.Coert, Post, Vandertop, Verbaan, and Muller critical were involved in revision of the article for important intellectual content.Dr. Labib disclosed work for hire.The remaining authors have disclosed that they do not have any potential conflicts of interest.The need for ethical approval for this study was waived by the institutional review board of the Amsterdam UMC, no review was needed and thus no waiver number was needed as it did not fall under the board's guidelines as human subjects research.
For information regarding this article, E-mail: h.labib@amsterdamumc.nlThis study involves human participation.This study was conducted in accordance with the Helsinki Declaration.