Abstract

Background. Potassium and sodium are inextricably linked to the maintenance of cell potential and electrolyte homeostasis. Few studies have examined their joint relationship with outcomes after stroke. The objective of this study is to ascertain whether combined low nadir serum sodium and potassium levels are correlated with an unfavorable outcome in ischemic stroke. Methods. Retrospective cohort study of 2,920 consecutive patients with first-ever ischemic stroke. Serum sodium and potassium levels were measured at 10 consecutive time points over 10 days poststroke. The Youden Index was performed to identify patients with low nadir sodium (<140 mmol/L) and potassium (<3.65 mmol/L) levels, defined as a NaK score of 2. Multivariable logistic regression and Cox proportional hazard analysis were used to evaluate the independent relationship of sodium and potassium levels with clinical outcomes at three months. Results. A total of 2,920 patients with ischemic stroke were analyzed (mean age years; 60.19% male), of whom 740 (25.3%) with both low nadir sodium and potassium levels had a quintuple 3-month case fatality compared to other patients (10.6% vs. 2.1%). Multivariable analyses identified as an independent predictor of 3-month death (adjusted odds ratio (OR) 2.23; 95% confidence interval (CI) 1.17-4.53; ) and an unfavorable shift in the distribution of scores on the modified Rankin scale (adjusted OR 1.51, 95% CI 1.12-2.04; ). Conclusions. Low sodium and potassium levels are common after ischemic stroke and are independent predictors of subsequent death.

1. Introduction

Electrolyte disturbances, such as hyponatremia and hypokalemia, are common in acute illnesses including stroke and other forms of cardiovascular disease [1, 2]. Hyponatremia is associated with increased mortality after stroke, both ischemic and intracerebral hemorrhage (ICH) [1, 3], while hypokalemia has a similar predictive significance regardless of age, stroke severity, and history of hypertension [4]. Moreover, recent research suggests that serum potassium levels have a near-linear association with the incidence of stroke [5]. However, it is uncertain whether a low sodium-to-potassium excretion ratio predicts cardiovascular events in community-dwelling adults better than either sodium or potassium alone [6].

Compared with admission parameters, recent studies indicate that a nadir value (e.g., albumin and hemoglobin levels), which might better reflect the severity of the acute event, has a stronger correlation with clinical outcomes. For example, low nadir albumin levels, evaluated based on the lowest point during hospitalization rather than at admission, were associated with an increased risk of acute worsening of renal function [79]. The trend of the nadir levels of serum sodium and potassium during hospitalization tends to represent their average levels, which in turn reflect the comprehensive impact during hospitalization, including the acute phase of ischemic stroke and other potential influencing factors, thereby endowing them with predictive value. Additionally, since sodium and potassium levels are partly affected by diet and medications, nadir levels were more suitable for evaluating electrolyte disturbance rather than a single measurement at admission [1, 10].

We selected low nadir sodium and potassium levels as the potential biomarkers [3, 11] to examine their joint effects on the relationship with outcomes after stroke [6]. We hypothesized that low nadir serum sodium and potassium levels, even in the normal range, might act synergistically to increase the risk of poor outcomes after ischemic stroke.

2. Methods

The data that support the findings of this study are available upon reasonable request from the corresponding author. The local ethics committee approved the study protocol.

2.1. Study Design and Participants

We analyzed the database from the Chengdu Stroke Registry between January 2009 and April 2016. This is an ongoing database that contains information on consecutive patients with stroke admitted to West China Hospital since January 2002 [12]. The registry is approved by the Biomedical Research Ethics Committee of West China Hospital, Sichuan University; this committee approved the protocol for this study [13]. Written informed consent was obtained from all participants or their legal guardians. Patients with first-ever ischemic stroke were consecutively recruited: they had to be adults (≥18 years); diagnosed with first-ever ischemic stroke according to the World Health Organization criteria, with brain computed tomography (CT) or magnetic resonance imaging (MRI) scans; within 24 hours of the presumed symptom onset. Patients underwent clinical assessments by certified neurologists at baseline and three months poststroke.

2.2. Procedures

Information about baseline demographic characteristics was obtained predominantly through in-person interviews with patients and family members. In-hospital details, including clinical features and diagnosis, were obtained through a combination of medical records and interviews with patients or their families. Follow-up details were obtained primarily through telephone interviews. Medical history variables included the existence of any of the following: hypertension, diabetes mellitus, hyperlipidemia, stroke, and heart disease (including any history of atrial fibrillation/heart attack/myocardial infarction, angina, coronary heart disease, or valvular heart disease), either self-reported or diagnosed before ischemic stroke onset in-hospital. Comorbidities were defined if the patient had coronary heart disease, congestive heart failure, cancer, leukocythemia, chronic pulmonary disease, diabetes mellitus, hepatic insufficiency, or renal insufficiency. For this study, information about acute reperfusion and electrolyte-related therapies during hospitalization was collected, such as mannitol, saline solution, and potassium intake. Stroke severity on admission was assessed on the National Institutes of Health Stroke Scale (NIHSS, where scores range from 0 to 42, with higher scores indicating greater neurological deficit [14]). The key clinical outcomes were death and a composite of death/disability at three months, defined by scores of 3 to 6 on the modified Rankin scale (mRS) [15]. Follow-up was 89% (2605/2920).

Participants admitted for the first time with a diagnosis of ischemic stroke were included in the study. Exclusion criteria from the study were as follows: presence of brain tumor or other stroke mimic, hypertension treated with diuretics, absence of informed consent, onset of hours, presence of do-not-treat or do-not-resuscitate orders, and absence of sodium or potassium value.

We collected serum sodium and potassium levels over a period of 10 consecutive days, which conformed to the median hospitalization stay for stroke patients of 10 days (7-14 days). Nadir sodium and potassium levels were defined as the lowest value during hospitalization. According to the Health Industry Standard of China (WS/T 404.3-2012; reference interval for common clinical biochemistry tests/part 3: serum potassium, sodium, and chloride), hyponatremia was defined as sodium  mmol/L and hypokalemia as potassium  mmol/L.

2.3. Statistical Analysis

Categorical variables are presented as counts (%), and the continuous or discrete variables are presented as mean (standard deviation (SD)) or median (interquartile range (IQR)). Student’s -test, the test, ANOVA, Mann–Whitney test, Fisher’s exact test, and Kruskal-Wallis test were used for univariate analysis among scales with relevant variables as appropriate. Associations of clinical characteristics with death were analyzed using logistic regression or Cox proportional hazard models, whereas associations of clinical characteristics with mRS were analyzed using shift analysis of the full range of scores on the mRS. The hazard ratio (HR) was calculated using Cox proportional hazard modeling with adjustment for potential confounders. Data are reported as odds ratios (OR) and 95% confidence intervals (CI). Where appropriate, adjusted ORs are reported. Age, sex, NIHSS, albumin, serum creatinine, random blood glucose, systolic blood pressure, pneumonia, comorbidities, and reperfusion and electrolyte-related therapies during hospitalization (e.g., saline, potassium intake, and mannitol), identified as confounding variables in the total cohort, were adjusted in all multivariable models. Two-sided values are reported, with considered statistically significant in all tests unless another threshold was given. All statistical analyses were performed in R Core Team (2017) (R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL: https://www.R-project.org/).

3. Results

A total of 3,101 patients with acute ischemic stroke were screened, but 181 patients were excluded for the following reasons: presence of brain tumor or other stroke mimic, hypertension treated with diuretics, absence of informed consent, onset of hours, presence of do-not-treat or do-not-resuscitate orders, and absence of sodium or potassium values (Supplementary Figure 1). Thus, 2,920 patients were included in the final analysis, with 1,754 (60.2%) male with a mean age of years. The baseline sodium and potassium levels were  mmol/L and  mmol/L, respectively, and the mean sodium and potassium levels during hospitalization were  mmol/L and  mmol/L, respectively (Table 1).

Neither hyponatremia (adjusted odds ratio (aOR) 1.46, 95% confidence interval (CI) 0.96-2.19; ) nor hypokalemia (aOR 1.33, 95% CI 0.84-2.10; ) in isolation exhibited an association with 3-month mortality. Furthermore, the combined presence of hyponatremia and hypokalemia did not demonstrate a synergistic effect in elevating the odds of this outcome (aOR 1.31, 95% CI 0.69-2.43; , Supplementary Table 1).

Figure 1 illustrates that utilizing the Youden Index with different cut-off values for serum sodium (140 mmol/L) and serum potassium (3.65 mmol/L) enhances the predictive performance for 3-month mortality. Subsequently, patients were categorized into three groups based on the nadir NaK scale (NaK) from 0 to 2, high nadir sodium (≥140 mmol/L) and potassium (≥3.65 mmol/L) were defined as 0, low nadir sodium (<140 mmol/L) or potassium levels (<3.65 mmol/L) were defined as 1, and both low nadir sodium and potassium levels were defined as 2 (Table 1). There were 1556 (53.3%) patients with low nadir sodium and 1272 (43.6%) with low nadir potassium levels. In comparison with patients having a NaK score of 0, those with were less likely to be male (54.1% vs. 62.2%), had higher NIHSS scores (7 vs. 3), higher systolic blood pressure (144.6 vs. 141.9 mmHg), lower hemoglobin (132.8 vs. 136.6 g/L), lower albumin (39.8 vs. 40.9 mmol/L), higher blood glucose (8.50 vs. 6.68 mmol/L), higher INR (1.08 vs. 1.05), a higher frequency of complication such as respiratory (31.2% vs. 11.8%) and urinary tract (4.7% vs. 1.0%) infections, and a higher frequency of 3-month mortality (10.6% vs. 2.1%, Table 1).

The 10-day course of mean sodium and potassium levels was plotted based on the nadir NaK scale. Sodium levels in patients with were significantly higher than those with or 2 (; Figure 2(a)). As depicted in Figure 2(b), patients with and 2 exhibited significantly lower baseline potassium levels (3.81 and 3.50 mmol/L, respectively) compared to 4.01 mmol/L in patients with . By the end of the 10 days, potassium levels had risen to 4.10 and 4.04 mmol/L in patients with and 2, respectively.

We conducted additional investigations into the association between the NaK scale and 3-month outcomes using multivariate models (Table 2). Compared with patients with , those with exhibited an increased risk of 3-month mortality with an adjusted odds ratio (aOR) of 2.23 (95% CI 1.17-4.53; ). Supplementary Figure 2 graphically represents the correlation between the nadir NaK scale and neurological recovery on the mRS: 45% of patients with exhibited poor 3-month mRS scores of 3-6, compared to 33.3% of patients with and 18.7% of patients with (). The multivariable shift analysis showed a significant favorable shift in the distribution of scores on the mRS with low nadir sodium and potassium () levels (aOR for the shift towards higher mRS scores 1.51, 95% CI 1.12-2.04; ).

Subgroup analyses were conducted to explore associations, revealing no heterogeneity based on age, sex, level of consciousness, stroke severity, and the use of specific therapies (Supplementary Figure 3). Kaplan–Meier survival curves over 90 days are presented in Supplementary Figure 4. Overall, 10.6% () of patients with a NaK score of 2 versus 2.1% () of patients with a NaK score of 0 were dead, indicating an approximately 4-fold difference in mortality (log-rank ). Further analysis using the Cox proportional hazards model demonstrated a significantly higher risk of death over time ( vs. : adjusted hazard ratio (HR) 3.01, 95% CI 1.41-6.46; ; vs. : adjusted HR 1.93, 95% CI 0.86-4.32; ).

4. Discussion

In this study, the first to examine the joint association of nadir sodium and potassium with outcomes in ischemic stroke, we have demonstrated that low nadir sodium and potassium levels were correlated with elevated case fatality and disability.

Recent research has suggested that serum potassium was linearly associated with an increased incidence of stroke, even within the normal range [5]. We defined the cut points of <140 mmol/L and <3.65 mmol/L, selected using the Youden Index, to determine low sodium and potassium, respectively, and to define a nadir NaK scale. The combination of nadir serum sodium and potassium enhanced the strength of association in the predictive model, supporting the hypothesis that they are independent predictors of adverse clinical outcomes. Previous studies have typically examined these biochemical variables separately, but O’Donnell et al. proposed that the combination of moderate sodium with high potassium intakes is associated with reduced mortality and serious cardiovascular events in a population-based study [6]. Another research has reported that the urinary sodium-to-potassium secretion ratio is associated with cardiovascular events or mortality in international community-dwelling participants [16, 17].

Potential detrimental mechanisms associated with low sodium levels in stroke could involve an inappropriate secretion of antidiuretic hormone (ADH) triggered by the overstimulation of the neurohumoral axis or by cerebral hypoperfusion in elderly patients with baroreceptor failure [18, 19]. Another possible mechanism leading to hyponatremia is cerebral salt wasting (CSW). Factors such as increased blood volume, sympathetic stimulus, or elevated angiotensin II promote the release of natriuretic peptide. This, in turn, results in a reduction in the activity of the renin-angiotensin-aldosterone system, affecting the distal tubules of the kidneys, leading to increased excretion of urinary sodium and ultimately causing hyponatremia [20]. Besides, secondary adrenal insufficiency due to pituitary ischemia or hemorrhage can lead to hyponatremia or hypokalemia. Cortisol is primarily secreted by the adrenal glands and plays a crucial role in facilitating free water excretion in the kidney. Insufficient cortisol levels result in the inability to effectively excrete a free water load, leading to the development of dilutional hyponatremia [21]. Conversely, potassium might act directly or indirectly through sodium allostasis upon vascular smooth-muscle cell contraction to induce hypertension [22].

In clinical practice, some stroke patients, especially elderly individuals, may experience alterations in serum sodium and potassium levels due to the use of medications such as thiazides or furosemide. Thiazides and thiazide-like agents are common diuretics that can lead to low sodium or low potassium levels. They act solely in the distal tubules and do not interfere with the ability of ADH, which is crucial for the development of hyponatremia [23, 24]. Additionally, when used in combination with potassium-sparing diuretics such as amiloride, they may exacerbate hyponatremia [25]. Furthermore, furosemide has less of an impact on sodium and potassium compared to thiazide diuretics because it impairs both renal concentrating and diluting mechanisms [23, 24].

Hyponatremia and hypokalemia are prevalent electrolyte abnormalities encountered in clinical practice. The multitude of underlying etiologies necessitates a meticulous differential diagnosis, taking into account various comorbidities, concurrent medications, clinical examination findings, and various laboratory measurements. This comprehensive approach is crucial for guiding decisions regarding management [2630].

Studies indicate a frequency of hyponatremia ranging from 10 to 30% across various serious conditions that require hospital admission [31, 32]. Questions have been raised as to whether restoration of sodium to normal levels improves outcomes after acute stroke [30]. On the one hand, reversing hyponatremia, irrespective of direct or indirect causes, could offer benefits, but on the other hand, there are potential harms related to arrhythmias.

The strengths of our study encompass a sizable sample size, the incorporation of repeated biochemical measures, and comprehensive details on various management strategies employed during hospitalization, such as the infusion of saline solutions, potassium intake, and mannitol usage. However, inherent limitations exist in observational studies like ours, characterized by a single-center cohort design and challenges related to residual confounding. These limitations constrain our ability to make definitive causal inferences.

5. Conclusion

Low nadir sodium and potassium levels are common in patients with ischemic stroke, and their combination is predictive of adverse 3-month outcomes.

Abbreviations

CIs:Confidence intervals
CT:Computed tomography
CSW:Cerebral salt wasting
GCS:Glasgow Coma Scale
HR:Hazard ratio
ICH:Intracerebral hemorrhage
IQR:Interquartile range
NaK:NaK scale, low nadir sodium and potassium scale
MRI:Magnetic resonance imaging
mRS:Modified Rankin scale score
NIHSS:National Institutes of Health Stroke Scale
OR:Odds ratio
SD:Standard deviation
SIADH:Syndrome of inappropriate antidiuretic hormone secretion
s-potassium:Serum potassium
s-sodium:Serum sodium
YI:Youden Index.

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding authors upon reasonable request.

Ethical Approval

All methods were performed in accordance with the Declaration of Helsinki and must have been approved by the Biomedical Research Ethics Committee and the Committee on Human Research of West China Hospital, Sichuan University (2013 [124]).

Informed consent was obtained from participants or their guardians.

Conflicts of Interest

The authors report no conflicts of interest concerning this study.

Authors’ Contributions

ML, SZ, ZF, and TC designed and conducted the study. ZF, TC, PL, ZZ, and ZW collected the data and constructed the database. ZF, TC, PL, ZW, YW, QS, and CY analyzed the data. ML, SZ, ZF, TC, PL, and ZZ wrote the paper. All authors reviewed the manuscript. Zijuan Feng and Ting Chen contributed equally to this work.

Acknowledgments

This study was funded by grants from the National Natural Science Foundation of China (82371210 and 92049115), the National Key Research and Development Program of China of the Ministry of Science and Technology of China (2018YFC1312300-303), and the 1.3.5 project for disciplines of excellence, West China Hospital, Sichuan University (ZYG D18009). The names of all copyediting or translation services employed in the preparation of our manuscript are as follows: Microsoft Office, Google Translate, DeepL, NetEase Youdao Translate, ChatGPT, and Grammarly.

Supplementary Materials

Supplementary Table 1: multivariable logistic regression analysis of the association between hyponatremia and hypokalemia and 3-month death. Supplementary Figure 1: flow chart illustrating the process of inclusion and exclusion of our study participants. Supplementary Figure 2: dichotomized analysis of the modified Rankin Scale (mRS) distribution at 3-month for all patients according to the classification of the nadir NaK scale. Supplementary Figure 3: effects of low nadir s-sodium and s-potassium levels on 3-month death in ischemic stroke patients by age, sex, state of consciousness, severity, and therapies. Supplementary Figure 4: Kaplan-Meier survival curve from day 0 to day 90 in patients according to the classification of the nadir NaK scale. (Supplementary Materials)