Impact of Hyponatremia on COPD Exacerbation Prognosis

The most common electrolyte disorder among hospitalized patients, hyponatremia is a predictor of poor prognosis in various diseases. The aim of this study was to establish the prevalence of hyponatremia in patients admitted for acute exacerbation of chronic obstructive pulmonary disease (AECOPD), as well as its association with poor clinical progress. Prospective observational study carried out from 1 October 2016 to 1 October 2018 in the following hospitals: Salnés in Vilagarcía de Arousa, Arquitecto Marcide in Ferrol, and the University Hospital Complex of Santiago de Compostela, Galicia, Spain, on patients admitted for AECOPD. Patient baseline treatment was identified, including hyponatremia-inducing drugs. Poor progress was defined as follows: prolonged stay, death during hospitalization, or readmission within one month after the index episode discharge. 602 patients were enrolled, 65 cases of hyponatremia (10.8%) were recorded, all of a mild nature (mean 131.6; SD 2.67). Of all the patients, 362 (60%) showed poor progress: 18 (3%) died at admission; 327 (54.3%) had a prolonged stay; and 91 (15.1%) were readmitted within one month after discharge. Patients with hyponatremia had a more frequent history of atrial fibrillation (AF) (p 0.005), pleural effusion (p 0.01), and prolonged stay (p 0.01). The factors independently associated with poor progress were hyponatremia, pneumonia, and not receiving home O2 treatment prior to admission. Hyponatremia is relatively frequent in patients admitted for AECOPD, and it has important prognostic implications, even when mild in nature.


Introduction
Hyponatremia is the most common electrolyte disorder among hospitalized patients, particularly among the elderly and women, and in the presence of comorbidities [1][2][3][4][5][6]. Its prevalence varies from 8% to 51%, and it often occurs as an acute disease complication, as a chronic disease decompensation, or as a consequence of interventions during patient treatment [1,5,6]. Hyponatremia is a predictor of poor prognosis in various diseases, including heart failure, coronary syndrome, liver cirrhosis, chronic kidney disease, and stroke [3,5,[7][8][9]. It is featured on clinical assessment scales, such as APACHE

Statistical Analysis
The data obtained through statistical analysis are expressed as mean values ± standard deviation (SD) in continuous variables, and as frequencies and percentages in categorical variables. Continuous variables were compared using the Student's t-test or the Wilcoxon test; for categorical variables, the chi-square test and the Fisher's exact test were used. Variables independently related to hyponatremia were identified by logistic regression, including those with p ≤ 0.05 in the univariate analysis. Variables associated with p < 0.05 were considered statistically significant. All analyses were completed with SPSS 15.

Results
A total of 602 patients were enrolled in the study; patient mean age was 73.8 years (SD 10.6) and 86% were male. 65 cases of hyponatremia (10.8%) were recorded, 53 of a mild nature (mean 132.6; SD 1.12), and 12 of a moderate nature (mean 126; SD 2.28). Of all the patients, 362 (60%) featured poor progress: 18 (3%) died at admission; 327 (54.3%) had a prolonged stay; and 91 (15.1%) were readmitted within one month after discharge. The most frequent cause of death was respiratory (53%). Baseline characteristics of the patients are shown in Table 1. Patients with worse prognosis were more frequently assigned GOLD Severe and Very Severe grades (p < 0.0001), had worse baseline dyspnea (p < 0.0001), higher percentage of flu vaccination, and higher AF (p 0.002), CHF (p 0.001), and CKD (0.002) frequency. Regular treatments are shown in Table 2. Of the patients, 47.7% received potentially hyponatremia-inducing treatments. Patients with poor progress were more frequently treated with short-acting bronchodilators (SABA (short-acting β2 agonists) (p 0.003); SAMA (short-acting anticholinergics) (p 0.04), chronic oral corticosteroids (p 0.005), theophylline (p 0.03), azithromycin (p 0.02), acetylcysteine (p 0.02), phosphodiesterase inhibitors (p 0.01), and home O 2 (p 0.008), although these results were not maintained upon the multivariate analysis.
The most frequent cause of AECOPD had a viral origin, although that was more common in the patients with poor progress (p < 0.0001); they also demonstrated a higher heart rate (0.02), higher leukocytosis (p 0.04), higher neutrophilia (p 0.02), more cases of hyponatremia (p 0.005), and troponin elevation (p 0.01); higher urea, potassium, and PCO 2 levels (p 0.001, p 0.01 and 0.001, respectively), higher frequency of pneumonia episodes (p 0.001), admission to the ICU (p 0.01), mechanical ventilation (p 0.02), and non-invasive mechanical ventilation (p 0.003). Patients with favorable progress had lower PO 2 (p 0.009) ( Table 3). The history of the patients with hyponatremia featured a higher frequency of AF (p 0.005), pleural effusion (p 0.01), higher leukocytosis (p 0.02), prolonged stay (p 0.01), and poor progress (p 0.01) ( Table 4), and the relationship between hyponatremia and AF, pleural effusion and prolonged stay was maintained after the multivariate analysis ( Table 5). The factors independently associated with poor progress were hyponatremia, pneumonia, and not receiving home O 2 treatment prior to admission (Table 6).

Discussion
In our study, the prevalence of hyponatremia in patients admitted for AECOPD is 10.8% higher than that of patients with community-acquired pneumonia as reported by Cuesta et al. [1], but lower than that of patients hospitalized for different causes, including AECOPD [2][3][4]23,24]. In addition to the low prevalence of hyponatremia, the absence of moderate and severe cases also seems noteworthy, despite the high percentage of patients with risk factors to develop it (heart failure, chronic kidney disease, cancer, and certain treatments).
In our study, factors related to poor progress were hyponatremia, pneumonia, and not being home oxygen users. We also observed some differential characteristics between patients with hyponatremia and those with normal sodium levels: the former demonstrate higher prevalence of AF, chest X-ray with pleural effusion, and prolonged stay, and these associations were independent. The higher prevalence of AF in patients with hyponatremia could be explained by its arrhythmogenic capacity, appearing as a result of a series of changes in the electrical activity and properties of the sinus node and pulmonary veins [25,26]. Recent studies show pleural effusion on chest X-ray in 33%-50% of the patients admitted for acute heart failure, and pleural effusion associated with a high rate of adverse events after hospital discharge (readmission, new heart failure episode, or death) [27,28]. In our case, there was a history of heart failure in 30% of the patients, and pleural effusion was recorded in 12%. Although no relationship between pleural effusion and poor progress was found, almost 22% of the patients with hyponatremia also had pleural effusion. Compensatory mechanisms counteracting the low cardiac output include stimulation of the renin-angiotensin-aldosterone system, sympathetic nervous system and release of vasopressin, which contribute to an increase in blood volume and, therefore, to hyponatremia and pleural effusion [29].
Results of the previous studies with the same sodium level cutoff to define hyponatremia (< 135 mEq/L) are similar to ours. Chalela et al. associate hyponatremia with longer hospital stay and higher hospital mortality in patients admitted for AECOPD [2]. Al Mawed et al. report that hyponatremia is independently associated with hospital mortality and increase in dependence or need for care after discharge, regardless of the diagnosis leading to admission, including AECOPD [24]. The risk of readmission within the first 30 days after discharge due to an episode of heart failure was significantly higher in patients with persistent hyponatremia, regardless of the heart failure severity [30].
Pneumonia is an independent factor of poor prognosis in patients with COPD and has been associated with longer average stay and lower survival [31,32]. Apart from that, hyponatremia occurs in 8%-31% of adults with pneumonia, it can be mild and asymptomatic [33], and it is associated with adverse outcomes, such as higher hospital mortality, the need for intensive care, and longer average stay [1,34,35]. Causality of hyponatremia and pneumonia is uncertain, although the syndrome of inappropriate antidiuretic hormone (ADH) secretion seems to play an important role in this association [1]. Other contributing factors could include the loss of volume and sodium through lungs and skin as a result of hyperventilation and temperature in patients with pneumonia, or adrenal suppression due to corticosteroid treatment-even inhaled-frequent in patients with COPD [36].
In our study, continuous home oxygen therapy (HOT) reduces the risk of poor prognosis. Although HOT is known to increase survival, reduce admission rates and improve the quality of life in patients with COPD and chronic respiratory failure [37,38], the relationship between HOT and patient progress following an AECOPD episode is not clearly defined. Recent literature shows apparently contradictory results in the relationship between HOT and prognosis after AECOPD: mortality, admission rate and average stay increase according to some studies [39][40][41]; and patients with HOT would show higher survival rates after admission to ICU and lower admission numbers according to other authors [42,43].
Better progress of patients requiring HOT could be explained by stricter medical follow-up, allowing for base treatment optimization, health education strengthening, and therapy compliance improvement, consequently impacting progress in a beneficial manner [44][45][46]. It could also allow for early identification of exacerbations, optimizing early management [47][48][49] Patients with HOT have been observed to have lower acidosis during exacerbations -which is an indicator of the lower severity of acute episodes-than those not requiring HOT [42]. Other pathophysiological mechanisms -not sufficiently clarified yet-have also been identified, such as the impact on certain mitochondrial genes by HOT in patients with COPD, which could explain these differences in outcome [50]. This study has several limitations. The cause of hyponatremia was not established. The fact that we only had mild and moderate cases of hyponatremia did not allow us to verify the impact of sodium levels on clinical results. Treatments aimed at correcting hyponatremia were not evaluated. The observational nature prevents the study from establishing causality and excluding residual confounding factors.
In conclusion, it seems that hyponatremia is relatively frequent in patients admitted for AECOPD, and it has prognostic implications. Therefore, patients with hyponatremia should be carefully monitored.