If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
1 These authors contributed equally to this article.
Corinna Giuliani
Footnotes
1 These authors contributed equally to this article.
Affiliations
Endocrine Unit, “Center for Research, Transfer and High Education on Chronic, Inflammatory, Degenerative and Neoplastic Disorders for the Development of Novel Therapies” (DENOThe), Department of Experimental and Clinical Biomedical Sciences “Mario Serio”, University of Florence, Careggi Hospital, Italy
Andrology Unit, “Center for Research, Transfer and High Education on Chronic, Inflammatory, Degenerative and Neoplastic Disorders for the Development of Novel Therapies” (DENOThe), Department of Experimental and Clinical Biomedical Sciences “Mario Serio”, University of Florence, Careggi Hospital, Italy
Endocrine Unit, “Center for Research, Transfer and High Education on Chronic, Inflammatory, Degenerative and Neoplastic Disorders for the Development of Novel Therapies” (DENOThe), Department of Experimental and Clinical Biomedical Sciences “Mario Serio”, University of Florence, Careggi Hospital, Italy
Requests for reprints should be addressed to Alessandro Peri, MD, PhD, Endocrine Unit, Department of Experimental and Clinical Biomedical Sciences, “Mario Serio”, University of Florence, Viale Pieraccini, 6, Florence 50139, Italy.
Endocrine Unit, “Center for Research, Transfer and High Education on Chronic, Inflammatory, Degenerative and Neoplastic Disorders for the Development of Novel Therapies” (DENOThe), Department of Experimental and Clinical Biomedical Sciences “Mario Serio”, University of Florence, Careggi Hospital, Italy
Hyponatremia is the most common electrolyte abnormality observed in clinical practice. Several studies have demonstrated that hyponatremia is associated with an increased length of hospital stay and of hospital resource utilization. To clarify the impact of hyponatremia on the length of hospitalization and costs, we performed a meta-analysis based on published studies that compared hospital length of stay and cost between patients with and without hyponatremia.
Methods
An extensive Medline, Embase, and Cochrane search was performed to retrieve all studies published up to April 1, 2015 using the following words: “hyponatremia” or “hyponatraemia” AND “hospitalization” or “hospitalisation.” A meta-analysis was performed including all studies comparing duration of hospitalization and hospital readmission rate in subjects with and without hyponatremia.
Results
Of 444 retrieved articles, 46 studies satisfied the inclusion criteria, encompassing a total of 3,940,042 patients; among these, 757,763 (19.2%) were hyponatremic. Across all studies, hyponatremia was associated with a significantly longer duration of hospitalization (3.30 [2.90-3.71; 95% CIs] mean days; P < .000). Similar results were obtained when patients with associated morbidities were analyzed separately. Furthermore, hyponatremic patients had a higher risk of readmission after the first hospitalization (odds ratio 1.32 [1.18-1.48; 95% CIs]; P < .000). A meta-regression analysis showed that the hyponatremia-related length of hospital stay was higher in males (Slope = 0.09 [0.05-0.12; 95% CIs]; P = .000 and Intercept = −1.36 [−3.03-0.32; 95% CIs]; P = .11) and in elderly patients (Slope = 0.002 [0.001-0.003; 95% CIs]; P < .000 and Intercept = 0.89 [0.83-0.97; 95% CIs]; P < .001). A negative association between serum [Na+] cutoff and duration of hospitalization was detected. No association between duration of hospitalization, serum [Na+], and associated morbidities was observed. Finally, when only US studies (n = 8) were considered, hyponatremia was associated with up to around $3000 higher hospital costs/patient when compared with the cost of normonatremic subjects.
Conclusions
This meta-analysis confirms that hyponatremia is associated with a prolonged hospital length of stay and higher risk of readmission. These observations suggest that hyponatremia may represent one important determinant of the hospitalization costs.
Patients with hyponatremia have an increased length of stay in the hospital compared with patients without hyponatremia.
•
Patients with hyponatremia have a higher risk of readmission to the hospital.
•
Hyponatremia may represent one important determinant of the hospitalization costs.
Hyponatremia, defined as a serum sodium concentration ([Na+]) < 136 mEq/L, is the most common electrolyte abnormality observed in clinical practice and occurs in up to 30% of hospitalized patients in its mild form (serum [Na+] 130-135 mEq/L).
Clinical and experimental evidence have demonstrated that hyponatremia represents a serious problem that is associated with significant morbidity and mortality. It is well known that acute severe hyponatremia may have severe neurological consequences due to cerebral edema, and that it can be lethal if not diagnosed and effectively treated.
Recent studies have demonstrated that chronic and mild hyponatremia also may negatively affect health status, with deleterious effects that affect, particularly, the central nervous system and the bone, causing gait instability, attention deficits, falls, osteoporosis, and fractures.
Lower serum sodium is associated with increased short-term mortality in hospitalized patients with worsening heart failure: results from the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbations of Chronic Heart Failure (OPTIME-CHF) Study.
We have recently confirmed in an extensive meta-analysis that hyponatremia is significantly associated with an increased risk of overall mortality in hospitalized patients.
We have also observed a rather small difference in serum [Na+] (mean 4.8 mEq/L) in patients who died, compared with survivors, suggesting that mild hyponatremia may also lead to death. However, whether hyponatremia is a direct cause of death or is simply a marker of severity of an underlying pathological condition (such as heart failure, cancer, cirrhosis) that is associated with death, is not clear, so far.
Another critical point to clarify is whether the correction of hyponatremia could revert or reduce the increased risk of death associated with this electrolyte disorder. For this reason we recently performed a further meta-analysis that compared the mortality rate in patients with or without improvement of hyponatremia. As hypothesized, we found that any improvement of serum [Na+] was associated with a significant reduced risk of overall mortality.
Taken together, these data suggest that hyponatremia, even when mild and chronic, represents also an economic and social burden because of the associated disability and mortality. Therefore, it is not surprising that hyponatremia is associated with an increased length of hospital stay and in hospital resource utilization and costs.
and could be associated not only with the increased mortality and morbidity associated with this condition, but also with a prolonged length of stay in the hospital.
To clarify the impact of hyponatremia on length of hospital stay and costs, here we performed an original meta-analysis based on published studies that compared hospital length of stay, risk of hospital readmission, and costs between patients with and without hyponatremia.
Methods
This meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) checklist (see Appendix 1, available online) (http://www.prisma-statement.org/).
Eligibility Criteria
All studies specifically addressing the duration of hospitalization, cost, and readmission risk in subjects with and without hyponatremia were included in the analysis.
Information Source and Search Strategy
An extensive Medline, Embase, and Cochrane search was performed including the following words: (“hyponatraemia”[All Fields] OR “hyponatremia”[MeSH Terms] OR “hyponatremia”[All Fields]) AND (“hospitalisation”[All Fields] OR “hospitalization”[MeSH Terms] OR “hospitalization”[All Fields]). The search (up to November 1, 2015) was restricted to English-language articles and human studies. A hand-searched bibliography of retrieved papers for additional references was performed. Details of the literature search process are outlined in the flow chart. The identification of relevant abstracts, the selection of studies based on the criteria described above, and the subsequent data extraction were performed independently by 2 of the authors (GP, CG), and conflicts resolved by a third investigator (GC). Full-text articles and meeting abstracts were included. The quality of studies was assessed using the Cochrane criteria.
The meta-analysis was performed including all studies, comparing duration of hospitalization and hospital readmission rate in subjects with and without hyponatremia (see Figure 1 and Table 1). Patients who died during hospitalization were not excluded, and in this case, length of hospitalization was considered the time between admission and death. Studies that did not specifically address the association between duration of hospitalization or hospitalization readmission rate in subjects with and without hyponatremia were excluded from the analysis (see Appendix 2, available online, and Table 2).
OPTIMIZE-HF Investigators and Coordinators Relationship between admission serum sodium concentration and clinical outcomes in patients hospitalized for heart failure: an analysis from the OPTIMIZE-HF registry.
Prospective cohort study in a sample of Medicare patients to identify patient characteristics associated with adverse events that are present early in a hospital state. Hyponatremia was an independent risk factor of readmission within 30 days but there were no data about rate of readmission in patients with hyponatremia vs patients without hyponatremia.
Retrospective cohort of 964,263 adults undergoing major surgery and observed for 30 days to determine whether preoperative hyponatremia is a predictor of 30-day perioperative morbidity and mortality. Hyponatremia was associated with prolonged median lengths of stay by approximately 1 day but there were no specific data available for meta-analysis.
Retrospective analysis of data from the EVEREST trial to identify risk factors predicting cardiovascular rehospitalization and mortality. Hyponatremia was an independent risk factor for re-hospitalization and mortality but there were no data about rate of readmission in patients with hyponatremia vs patients without hyponatremia.
The principal outcome of this analysis was to compare the duration of the hospitalization and costs in subjects with or without hyponatremia at admission. A secondary outcome included the risk of hospitalization readmission in patients with or without hyponatremia. The quality of the studies was assessed using the Cochrane criteria.
Heterogeneity on duration of hospitalization in patients with or without hyponatremia was assessed by using I2 statistics. Even when a low heterogeneity was detected, a random-effects model was applied, because the validity of tests of heterogeneity can be limited with a small number of component studies. To estimate possible publication or disclosure bias, we used funnel plots, the Begg-adjusted rank correlation test, and Egger's test. A meta-regression analysis was performed to test the effect of age, serum [Na+] cutoff, and several associated morbidities at enrollment on duration of hospitalization. In addition, a linear regression analysis model, which weighted each study for the number of subjects enrolled, was performed to verify the independent effect of hyponatremia on duration of hospitalization after the adjustment for age and sex and serum [Na+] cutoff.
All data were calculated using Comprehensive Meta-analysis Version 2 (Biostat, Englewood, NJ). Logistic multivariate analysis was performed on SPSS (Statistical Package for the Social Sciences; Chicago, Ill.) for Windows 20.1.
Results
Of 444 retrieved articles, 358 were excluded for different reasons. The flow of the meta-analysis is summarized in Figure 1, and the characteristics of the studies included in the meta-analysis are summarized in Table 1 (see references
OPTIMIZE-HF Investigators and Coordinators Relationship between admission serum sodium concentration and clinical outcomes in patients hospitalized for heart failure: an analysis from the OPTIMIZE-HF registry.
). Eighty-six full-text articles were considered potentially eligible for the meta-analysis. However, 37 of them were excluded because they did not include information about serum [Na+] (n = 4) or about duration of hospitalization (n = 33) (see Appendix 2, available online). Three additional studies were excluded because they met inclusion criteria but did not provide data for meta-analysis (Table 2).
The 46 selected studies were grouped according to different pathological conditions (Figure 1 and Table 1). Among them, 5 studies evaluated pediatric patients. Eleven studies reported data on the effect of hyponatremia for combined mixed diseases, which could not be grouped separately. An additional 9 studies retrospectively investigated the effect of hyponatremia in hospitalized series of subjects. In these studies, a major diagnosis was not specified.
The risk of hospital readmission in subjects with and without hyponatremia was specified in 8 of the aforementioned studies.
Overall, 3,940,042 patients and 757,763 hyponatremic subjects (19.2%) were included in the meta-analysis. Hyponatremia was defined according to varying cutoff definitions in the included studies (Table 1). When the same study reported data according to different serum [Na+] thresholds, a weighted arithmetic mean was considered. I2 in trials assessing the duration of hospitalization in subjects with and without hyponatremia was 99.4 (P < .000). Funnel plot and Begg-adjusted rank correlation test (Kendall's τ: 0.18; P = .09) suggested no major publication bias.
When all 45 studies reporting data on duration of hospitalization were considered, hyponatremia was significantly associated with a significantly longer duration of hospitalization (3.30 [2.90-3.71; 95% CIs] days; P < .000). Similar results were obtained when patients with specific diseases or series of hospitalized patients were analyzed separately (Figure 2, panels A-G). Similarly, hyponatremic patients had a higher risk of hospitalization readmission (mean follow-up 72 days after first hospital discharge) when compared with those with normal [Na+] at the first hospital admission (Figure 3). Sub-analysis for risk of hospitalization readmission in specific diseases was not performed because of insufficient data.
Figure 2Mean duration of hospitalization (95% confidence interval; days) in patients with or without (no) hyponatremia in hospitalized series of subjects (A) or according to the presence of heart failure (B), heart surgery (C), liver diseases (D), pulmonary diseases (E), pediatric patients (F), or combined disease (G). CI = confidence interval.
Figure 2Mean duration of hospitalization (95% confidence interval; days) in patients with or without (no) hyponatremia in hospitalized series of subjects (A) or according to the presence of heart failure (B), heart surgery (C), liver diseases (D), pulmonary diseases (E), pediatric patients (F), or combined disease (G). CI = confidence interval.
Figure 2Mean duration of hospitalization (95% confidence interval; days) in patients with or without (no) hyponatremia in hospitalized series of subjects (A) or according to the presence of heart failure (B), heart surgery (C), liver diseases (D), pulmonary diseases (E), pediatric patients (F), or combined disease (G). CI = confidence interval.
Figure 2Mean duration of hospitalization (95% confidence interval; days) in patients with or without (no) hyponatremia in hospitalized series of subjects (A) or according to the presence of heart failure (B), heart surgery (C), liver diseases (D), pulmonary diseases (E), pediatric patients (F), or combined disease (G). CI = confidence interval.
A meta-regression analysis showed that the hyponatremia-related duration of hospitalization was higher in males (Slope = 0.09 [0.05-0.12; 95% CIs]; P = .000 and Intercept = −13.6 [−3.03-0.32; 95% CIs]; P = .11). In addition, a mild association between age and longer hospitalization was also observed (Slope = 0.002 [0.001-0.003; 95% CIs]; P < .0001 and Intercept = 0.89 [0.83-0.97; 95% CIs]; P < .000). Finally, a negative association between serum [Na+] cutoff and duration of hospitalization was detected (Figure 4). The latter association was confirmed in a multivariate regression model adjusting for age and sex (adj r = −0.210; P < .000). Conversely, no association between duration of hospitalization, serum [Na+], and several associated morbidities such as history of heart failure, cardiovascular diseases, diabetes mellitus, hypertension, cirrhosis, or chronic renal insufficiency at enrollment was observed (not shown).
Figure 4Weighted relationship between serum [Na+] cutoff definition and duration of hospitalization.
The present meta-analysis, which derives from the study of a large number of patients, confirms that hyponatremia represents a frequent condition observed in up to 20% of hospitalized patients. In addition, our data suggest that hyponatremia is associated with a prolonged hospital length of stay, and higher risk of readmission. The data were similar when European studies were compared with those performed in North America. Length of hospital stay is a strong determinant of the cost of hospital care.
Hence, taken together, these observations indicate that hyponatremia may represent one important determinant of the medical costs associated with hospitalization. Accordingly, our data showed that when only US studies were considered, hyponatremia was associated with an increase of hospital costs of up to $3000 when compared with normonatremia. A consensus panel of expert physicians suggested that the annual costs/hospitalized patient directly attributable to hyponatremia (cut off for serum [Na+] < 130 mEql/L) in the US ranges from $1528 to $3441.
More recently, longitudinal data from a large US managed care claim database (Waltham, Mass.) indicated that an even higher 1-year mean inpatient cost could be related to hyponatremic subjects ($10,636).
Despite these data, hyponatremia is still underdiagnosed and often poorly considered if not ignored. Accordingly, we recently reported that an improvement of serum [Na+] of any degree was obtained in only 53.2% of hospitalized patients.
In agreement with these data, it has been shown that only a small percentage of patients with hyponatremia receive an International Classification of Diseases, 9th Revision, Clinical Modification diagnosis code for hyponatremia.
In particular, among 2,134,363 admittances from all departments of internal medicine of the Spanish National Health System, only 1.5% coded hyponatremia as a primary or a secondary diagnosis.
As reported above, such a percentage is very far from the true prevalence of this condition in hospitalized patients. An analysis of the reasons behind these data is beyond the aim of the present study. However, it is interesting to note that in the vast majority of cases (81.7%) hyponatremia was codified as a secondary diagnosis according to the International Classification of Diseases, 9th Revision, Clinical Modification.
Accordingly, whether hyponatremia represents an independent risk factor for mortality and morbidity or it is simply associated with other conditions that significantly raise these risks remains to be elucidated. Current data cannot clarify this point; however, it is interesting to note that hyponatremia-related increased length of hospital stay was confirmed even after the adjustment for age and sex, and it was also independent of the type of related diseases. Similarly, a retrospective analysis of 1046 patients hospitalized for heart failure, who participated in the CHF Benchmark Project, a large collaborative quality improvement project coordinated by the University Health System Consortium in the US, confirmed that hyponatremia was an independent predictor of the duration of hospitalization.
In addition, more recently, data from the Minimum Basic Data Set of discharged patients from all departments of internal medicine of the Spanish National Health System (NHS) between 2007 and 2010 showed that hyponatremia was independently associated with a longer hospitalization, higher costs, and was considered a powerful predictor of hospital readmission even after the adjustment of associated morbidities, as detected by the Charlson index.
Another interesting finding of the present study derives from the meta-regression analysis, which indicates that the duration of hospitalization was even longer in patients with lower serum [Na+] at enrollment. Interestingly, in a recent meta-analysis we reported that the favorable effects of hyponatremia correction were inversely related to hyponatremia at enrollment.
These findings suggest that patients with a more severe hyponatremia deserve an even more thoughtful work-up, because the outcome of patients who were correctly treated for this condition may be more favorable compared with those who did not receive an appropriate treatment.
The possible pathogenetic mechanisms underlying hyponatremia-related morbidity and mortality are not completely understood. Hyponatremia-induced oxidative stress has been suggested to play an important role.
However, a growing body of evidence also indicates that several inflammatory mediators such as interleukin-1 and interleukin-6 can induce hyponatremia through excessive vasopressin release,
Hypothalamic-pituitary-adrenal axis activation and stimulation of systemic vasopressin secretion by recombinant interleukin-6 in humans: potential implications for the syndrome of inappropriate vasopressin secretion.
These data further confirm that a cause–effect relationship between hyponatremia, hospital length of stay, and readmission rate cannot be extrapolated from the studies that have been analyzed here. Accordingly, it should be recognized that the data were adjusted only for age and sex, whereas the prevalence of associated morbidities was not considered as a possible confounder, because they were not adequately reported in a sufficient number of studies. However, it should be recognized that a large survey about hospitalized patients confirmed an association between hyponatremia and longer hospitalization even after the adjustment of possible confounders including age, sex, race, admission service, and associated morbidities, as derived from the Deyo-Charlson Comorbidity Index score.
A further limitation of the present study is represented by incomplete reports of the data on hospital duration and readmission rate in trials only marginally designed for the assessment of these endpoints. The prolonged hospital length of stay could be affected by the economic system of reimbursement in hospitals located in different countries. Admittedly, this issue may have an influence on the behavior of health professionals. In addition, the statistical analysis showed the presence of heterogeneity.
Finally, even if statistical analyses did not suggest any relevant publication bias, the possibility of selective reporting cannot be excluded. In any case, the economic impact of hyponatremia in terms of prolonged hospitalization suggests that diagnostic and therapeutic algorithms used in clinical practice should be implemented. These considerations highlight the need for additional, well-designed studies of clinical outcomes with effective therapies in patients with hyponatremia.
THE ECONOMIC BURDEN OF HYPONATREMIA: SYSTEMATIC REVIEW AND META-ANALYSIS
1
Abstract
Structured summary
2
Background: Hyponatremia is the most common electrolyte abnormality observed in clinical practice. Several studies have demonstrated that hyponatremia is associated with an increased length of hospital stay and of hospital resource utilization. To clarify the impact of hyponatremia on the length of hospitalization and costs, we performed a meta-analysis based on published studies that compared hospital length of stay between patients with or without hyponatremia. Methods: An extensive Medline, Embase, and Cochrane search was performed to retrieve all studies published up to April 1, 2015, using the following words: “hyponatremia” or “hyponatraemia” AND “hospitalization.” A meta-analysis was performed including all studies comparing duration of hospitalization and hospital readmission rate in subjects with or without hyponatremia. Findings: Of 444 retrieved articles, 46 studies satisfied inclusion criteria, encompassing a total of 3,940,042 patients, of whom 757,763 (19.2%) were hyponatremic. Across all studies, hyponatremia was associated with a significantly longer duration of hospitalization (3.30 [2.90-3.71; 95% CIs] mean days; P < .0001). Similar results were obtained when patients with associated morbidities were analyzed separately. Furthermore, hyponatremic patients had a higher risk of hospitalization at the first hospital admission (OR = 1.32 [1.18; 1.48; 95% CIs]; P < .0001). A meta-regression analysis showed that the hyponatremia-related length of hospital stay was higher in males (Slope [S] = 0.09 [0.05-0.12; 95% CIs]; P = .0001 and Intercept [I] = −1.36 [−3.03-0.32; 95% CIs]; P = .11) and in elderly patients (S = 0.002 [0.001-0.003; 95% CIs]; P < .0001 and I = 0.89 [0.83-0.97; 95% CIs]; P < .001). A negative association between serum [Na+] cutoff and duration of hospitalization was detected. No association between duration of hospitalization, serum [Na+] and associated morbidities was observed. Finally, when only US studies (n = 8) were considered, hyponatremia was associated with up to $3000 higher hospital costs/patients when compared with the cost of normonatremic subjects. Interpretation: This meta-analysis confirms that hyponatremia is associated with a prolonged hospital length of stay and higher risk of readmission. These observations indicate that hyponatremia represents one important determinant of the hospitalization costs.
2
Introduction
Rationale
3
In the US, the direct medical costs of hyponatremia were estimated to range between $1.6 billion and $3.6 billion and could be associated not only with the increased mortality and morbidity, but also with a prolonged length of stay in hospital.
4-5
Objectives
4
To clarify the impact of hyponatremia on length of hospital stay and costs, we performed a meta-analysis based on published studies that compared hospital length of stay and cost between patients with or without hyponatremia.
5
Methods
Protocol and registration
5
All studies specifically addressing the duration of hospitalization, cost and readmission risk in subjects with or without hyponatremia were included in the analysis.
6
Eligibility criteria
6
Studies not including information about serum [Na+] or about duration of hospitalization were excluded from the analysis (see Table 2).
6
Information sources
7
PubMed from 1965 – November 1, 2015 EMBASE from 1974 – November 1, 2015 Cochrane from 1967 – November 1, 2015. The principal source of information was derived from published articles.
6
Search
8
An extensive Medline, Embase, and Cochrane search was performed including the following words: “hyponatraemia” [All Fields] OR “hyponatremia” [MeSH Terms] OR “hyponatremia” [All Fields]) AND (“hospitalisation” [All Fields] OR “hospitalization” [MeSH Terms] OR “hospitalization” [All Fields].
6
Study selection
9
We did not employ a search software. We hand-searched bibliographies of retrieved papers for additional references Details of the literature search process are outlined in the flow chart (Figure 1).
6
Data collection process
10
Data extraction were performed independently by 2 of the authors (GP, CG), and conflicts resolved by a third investigator (GC). The credentials of all investigators are indicated in the author list.
6
Data items
11
The principal outcome of this analysis was to compare the duration of the hospitalization and cost in subjects with or without hyponatremia at admission. A secondary outcome included the risk of hospitalization readmission in patients with or without hyponatremia.
7
Risk of bias in individual studies
12
Quality of the studies was assessed using the Cochrane criteria.
7
Summary measures
13
The identification of relevant studies was performed independently by 2 of the authors (GP, CG), and conflicts resolved by a third investigator (GC). Mean duration of hospitalization and cost with 95% confidence interval (CI) was calculated for in subjects with or without hyponatremia at admission. In addition, odds ratios (95% CI) of hospitalization readmission in patients with or without hyponatremia were also evaluated.
6-7
Synthesis of results
14
Heterogeneity on duration of hospitalization in patients with or without hyponatremia was assessed by using I2 statistics. For a more conservative approach, results of random effect models were presented.
7
Risk of bias across studies
15
The Begg-adjusted rank correlation test was used to test the presence of possible bias across studies.
7
Additional analyses
16
A meta-regression analysis was performed to test the effect of age, serum [Na+] cutoff and several associated morbidities at enrollment on duration of hospitalization. In addition, a linear regression analysis model, weighting each study for the number of subjects enrolled, was performed to verify the independent effect of hyponatremia on duration of hospitalization after the adjustment for age and sex and serum [Na+] cutoff.
7
Results
Study selection
17
Of 444 retrieved articles, 358 articles were excluded for different reasons. The flow of the meta-analysis is summarized in Figure 1, and the characteristics of the trials included in the meta-analysis are summarized in Table 1. Among the 46 selected studies, 3, 10, 4, 4, 5 studies evaluated the effect of hyponatremia in subjects undergoing heart surgery, in those with heart failure (HF), end-stage liver diseases, pulmonary diseases, and in pediatric patients, respectively. In addition, another 11 studies reported data on the effect of hyponatremia for combined mixed diseases, which could not be grouped separately (Table 1). In addition, 9 studies retrospectively investigated the effect of hyponatremia in a hospitalized series of subjects. In these studies, a major diagnosis was not specified.
7-8
Study characteristics
18
Overall 3,940,042 patients and 757,763 hyponatremic subjects (19.2%) were included in the meta-analysis. Hyponatremia was defined according to varying cutoff definitions in the included studies (Table 1). When the same study reported data according to different serum [Na+] thresholds, a pondered mean was considered. I2 in trials assessing the duration of hospitalization in subjects with or without hyponatremia was 99.4 (P < .0001).
7-8
Risk of bias within studies
19
For a more conservative approach, results of random effect models were presented.
7-8
Results of individual studies
20
We included 7 forest plots evaluating duration of hospitalization on patients with specific diseases or series of hospitalized patients. In addition, one plot on risk of hospitalization readmission and one on hospitalization cost were also included. Finally, a forest plot to examine effect age and [Na+] in predicting duration of hospitalization (meta-regression analysis) was also evaluated.
7-8
Synthesis of results
21
Present results of each meta-analysis were done, including confidence intervals and measures of consistency.
7-9
Risk of bias across studies
22
Funnel plot and Begg-adjusted rank correlation test (Kendall's τ: 0.18; P = .09) suggested no major publication bias.
8
Additional analysis
23
A meta-regression analysis showed that the hyponatremia-related duration of hospitalization was higher in males (S = 0.09 [0.05; 0.12; 95% CIs]; P = .0001 and I = −13.6 [−3.03; 0.32; 95% CIs]; P = .11). In addition, a mild association between age and longer hospitalization was also observed (S = 0.002 [0.001; 0.003; 95% CIs]; P < .0001 and I = 0.89 [0.83; 0.97; 95% CIs]; P < .0001). Finally, a negative association between serum [Na+] cut-off and duration of hospitalization was detected (Figure 4). The latter association was confirmed in a multivariate regression model adjusting for age and sex (adj r = −0.210; P < .0001). Conversely, no association between duration of hospitalization, serum [Na+] and several associated morbidities such as history of heart failure, cardiovascular diseases, diabetes mellitus, hypertension, cirrhosis or chronic renal insufficiency at enrolment was observed (not shown).
9
Discussion
Summary of evidence
24
The present meta-analysis, deriving from the study of a huge number of patients, confirms that hyponatremia represents a frequent condition observed in up to 20% of hospitalized patients. In addition, our data show that hyponatremia is associated with a prolonged hospital length of stay, and higher risk of readmission. The data were similar when European studies were compared with those performed in North America.
9-10
Limitations
25
Several limitations should be recognized. First of all, it should be recognized that the data were adjusted only for age and sex, whereas the prevalence of associated morbidities was not considered as possible confounders, because they were not reported adequately in a sufficient number of studies. Hence, potential unmeasured confounders may have caused residual confounding effects, but the measured factors that are correlated with such confounders should have mitigated this bias. However, meta-analysis is particularly useful when there is a variety of reports with low statistical power; in this situation, pooling of data can improve power and provide a more convincing result. A further limitation of the present study is represented by incomplete reporting of the data on hospital duration and readmission rate in trials only marginally designed for the assessment of these endpoints. In particular, all the reviewed studies analyzed were observational surveys and none of them was originally designed to address hyponatremia-related hospitalization outcome or medical burden. In addition, the statistical analysis showed the presence of heterogeneity.
Finally, even if statistical analyses did not suggest any relevant publication bias, the possibility of selective reporting cannot be excluded.
11-12
Conclusions
26
Taken together, present data indicate that hyponatremia represents one important determinant of the hospitalization medical costs. The economic impact of hyponatremia in terms of prolonged hospitalization suggests new therapeutic options with the need to develop and implement new diagnostic and therapeutic algorithms in clinical practice.
12-13
Funding
Funding
27
No separate funding was necessary for the undertaking of this systematic review and meta-analysis.
Lower serum sodium is associated with increased short-term mortality in hospitalized patients with worsening heart failure: results from the Outcomes of a Prospective Trial of Intravenous Milrinone for Exacerbations of Chronic Heart Failure (OPTIME-CHF) Study.
(Version 5.0.1 [updated September 2008])Higgins J.P.T. Green S. Cochrane Handbook for Systematic Reviews of Interventions. The Cochrane Collaboration,
London2008 (Available at:)
Relationship between admission serum sodium concentration and clinical outcomes in patients hospitalized for heart failure: an analysis from the OPTIMIZE-HF registry.
Hypothalamic-pituitary-adrenal axis activation and stimulation of systemic vasopressin secretion by recombinant interleukin-6 in humans: potential implications for the syndrome of inappropriate vasopressin secretion.
Conflict of Interest: AP has served as a consultant and member of advisory boards for Otsuka. GLC has served as a consultant on advisory boards for Astellas, LEO Pharma, Sanofi, Merck Sharp and Dohme, DOC Generici, Takeda, UCB Pharma, and MerckSerono, and has received research and educational grants from Bayer HealthCare, Takeda, Gilead Sciences, Merck Sharp and Dohme, Otsuka and LEO Pharma. The authors report no other conflicts of interest in this work.
Authorship: All authors had access to the data. Study concept and design: GC; Acquisition of data: CG, GP; Analysis and interpretation of data: GC, AP, GLC; Drafting of the manuscript: GC, AP, CG; Critical revision of the manuscript for important intellectual content: GF, MM, AS, GLC; Statistical analysis: GC; Study supervision: AP.
30315-1&id=gr1.jpg){kind=link}
30315-1&id=gr2ab.jpg){kind=link}
30315-1&id=gr2cd.jpg){kind=link}
30315-1&id=gr2ef.jpg){kind=link}
30315-1&id=gr2g.jpg){kind=link}
30315-1&id=gr3.jpg){kind=link}
30315-1&id=gr4.jpg){kind=link}
30315-1&id=gr5.jpg){kind=link}