Skip Navigation



NDT Advance Access published online on October 11, 2007
Nephrology Dialysis Transplantation, doi:10.1093/ndt/gfm376
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
23/1/136    most recent
gfm376v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Liamis, G.
Right arrow Articles by Elisaf, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liamis, G.
Right arrow Articles by Elisaf, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2007]. Published by Oxford University Press on behalf of ERA-EDTA. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Clinical and laboratory characteristics of hypernatraemia in an internal medicine clinic

George Liamis1, Vasilis Tsimihodimos1, Michalis Doumas1, Athanasia Spyrou1, Eleni Bairaktari2 and Moses Elisaf1

1Department of Internal Medicine and 2Laboratory of Clinical Chemistry, Medical School, University of Ioannina, Ioannina, Greece

Correspondence and offprint requests to: Moses Elisaf, MD, FRSH, Professor of Medicine, Department of Internal Medicine, Medical School, University of Ioannina, 45110 Ioannina, Greece. Email: egepi{at}cc.uoi.gr



   Abstract
Top
 Abstract
Introduction
 Subjects and methods
 Results
 Discussion
 References
 
Background. Hypernatraemia is a frequent electrolyte disorder in hospitalized patients that has been mainly studied in an entire hospital population. The aim of this study was to determine the incidence, clinical characteristics, concomitant electrolyte abnormalities and outcome of hypernatraemia in an internal medicine clinic. Also, we sought to identify differences between patients who were admitted with hypernatraemia and those who developed hypernatraemia during hospitalization.

Methods. We prospectively studied patients who either on admission to our internal medicine clinic or during their hospitalization were found to have hypernatraemia (sodium concentration greater than 148 meq/l, 148 mmol/l). One hundred and thirteen patients out of 9158 patients at risk had hypernatraemia (incidence 1.2%). Of those, fifty patients had hypernatraemia on admission, whereas 63 had hospital-acquired hypernatraemia.

Results. Patients who developed hypernatraemia before hospital admission had a much lower mortality rate than patients with hospital-acquired hypernatraemia (28% vs 47.6%, P = 0.03), despite the fact that they had a higher peak serum sodium concentration (160.4 ± 9.9 vs 154.4 ± 2.4 meq/l, P = 0.000). Furthermore, they did not differ in either age or the frequency of concomitant electrolyte abnormalities in comparison with patients who developed hypernatraemia during hospitalization. There were two main subgroups of patients with hospital-acquired hypernatraemia. A total of 26 Patients (41%) exhibited a biochemical profile consistent with extracellular volume depletion, whereas 32 patients (51%) with euvolaemia. On the contrary, the majority of patients (82%) who were hypernatraemic on admission had hypovolaemic hypernatraemia. The construction of the receiver operating characteristics (ROC) plots revealed that the urea to creatinine ratio was the best predictor of the extracellular volume status. Indeed, a urea to creatinine value of 57 could differentiate between the groups with euvolaemic or hypovolaemic hypernatraemia with a sensitivity of 96.5% and a specificity of 100%.

Conclusion. The incidence of hypernatraemia in the present study was 1.2% with a high mortality rate mainly in patients with hospital-acquired hypernatraemia. There were two main profiles of hospital-acquired hypernatraemia, one consistent with extracellular volume depletion and another with euvolaemia. On the contrary, the majority of hypernatraemic patients on admission exhibited hypovolaemia. Almost half of our hypernatraemic patients had at least one additional electrolyte disturbance.

Keywords: electrolyte abnormalities; hypernatraemia



   Introduction
Top
 Abstract
 Introduction
Subjects and methods
 Results
 Discussion
 References
 
Hypernatraemia is a common electrolyte disorder with an incidence ranging from <1% to >3%, while the reported mortality rates range from ~40% to >60% [1–4]. Hypernatraemia has been studied mainly in elderly or mentally handicapped patients as well as in an entire hospital population [1,2,5–7]. However, to our knowledge no studies have been conducted including exclusively hypernatremic patients hospitalized in an internal medicine clinic.

In the current prospective study of hypernatremic patients (either on admission or during their hospitalization to our internal medicine clinic), we determined the incidence, clinical characteristics and outcomes of hypernatraemia and described concomitant electrolyte abnormalities encountered in these patients and sought to identify differences between patients who were admitted with hypernatraemia and those who developed hypernatraemia in hospital.



   Subjects and methods
Top
 Abstract
 Introduction
 Subjects and methods
Results
 Discussion
 References
 
Patients
Over a period of 2.5 years (starting on 5 February 1999), we prospectively studied non-selected, consecutive, adult patients (over 18-years of age) who either on admission to our clinic or during their hospitalization were found to have hypernatraemia. The study took place in the internal medicine clinic (60 beds) at the University Hospital of Ioannina (600 beds). To be eligible, patients had to have a serum sodium concentration [Na+] >148 meq/l (148 mmol/l), [which represents the upper normal value for our laboratory (145 meq/l) plus 1 SD (2.8 meq/l)] that was verified by repeat measurement to exclude laboratory error. In hyperglycaemic patients the corrected serum sodium concentration was evaluated. In this setting, the corrected [Na+] was calculated by increasing [Na+] by 1.6 meq/l for every 100 mg/dl (5.6 mmol/l) increment in the serum glucose above normal, while the correction factor 2.4 meq/l was used for serum glucose concentration >400 mg/dl (22.2 mmol/l) [8]. Patients with corrected [Na+] <149 meq/l were excluded from the study. In all cases a detailed medical history was obtained with special attention to the determination of a recent use of drugs that are associated with hypernatraemia (lactulose, sodium bicarbonate, lithium, dexamethasone e.g) and to the identification of potential causes of central or nephrogenic diabetes insipidus while each patient underwent a complete physical examination with special attention to orthostatic changes in pulse rate and blood pressure, jugular venous pressure, skin turgor, moisture in the axillae and hydration of mucous membranes. Orthostatic hypotension and orthostatic change in pulse rate were defined as the reduction in systolic blood pressure of at least 20 mmHg and the increase in pulse rate of at least 10% after 2 min in the upright position compared with the supine position, respectively. Furthermore, special attention was paid to determine the duration as well as the presence of symptoms of hypernatraemia. Prior to any therapeutic intervention, venous blood was obtained for the determination of serum glucose, urea, creatinine, uric acid, sodium, potassium, chloride, calcium, magnesium, phosphorus, total protein, albumin, and haematocrit. Arterial blood was obtained for blood gas measurements. Also, a fresh urine specimen was tested for osmolality (Uosm), glucose, urea, creatinine, uric acid, sodium, potassium, chloride, calcium, magnesium, phosphorus, and protein. Standard formulas were used for the determination of the fractional excretion (FE) of electrolytes, urea, and uric acid (UA).

Classification of hypernatraemia
Hypovolaemic hypernatraemia was diagnosed in patients with fractional excretion of sodium <0.5% and serum urea to creatinine ratio >40. In patients receiving diuretics considered to be still acting, only the urea to creatinine ratio was used for the diagnosis of hypovolaemia. Edematous hypernatraemia was diagnosed in patients with obvious oedema defined as the presence of >0.5 cm of pressure-induced dependent oedema or ascites, while the rest of the patients were included in the subgroup of euvovolaemic hypernatraemia.

Therapeutic interventions
In patients with symptomatic hypernatraemia that was developed over a period of <48 h, a rapid correction of serum sodium concentration (1–2 meq/l/h) was initially performed, while a slower rate of serum sodium reduction (<0.5 meq/l/h) was attained in patients with hypernatraemia of longer or unknown duration. In all patients, though, the targeted fall in the serum sodium concentration was up to 12 meq/l per day. Finally, the goal of treatment was to reduce the serum sodium concentration to 145 meq/l. Patients with severe hypovolaemia, as evidenced by hypotension and oliguria, were initially treated with isotonic (0.9%) saline that was followed by 0.45% saline administration, while milder volume deficit was treated with 0.45% saline. Moreover, 5% dextrose in water intravenously (or by drinking water in subjects able to take fluids orally) either alone or in combination with furosemide was employed in euvolaemic and hypervolaemic hypernatraemia, respectively. The amount of water necessary to correct hypernatraemia was estimated using the following equation: water requirement (litres) = TBW x {([Na+]s)/140) – 1}; where {Delta} [Na+]s represents the patient's serum sodium concentration, expressed in meq/l and TBW represents the patient's estimated total body water, expressed in litres. The total body water (TBW) was estimated as 60 and 50% of lean body weight in men and women, respectively, while in water-depleted hypernatraemic patients (namely subjects without hypervolaemic hypernatraemia) we used lower values (50% of lean body weight in men and 40% in women).

Given their diverse and non-specific nature, the symptoms of hypernatraemia (lethargy, weakness, irritability, seizures and coma), were ascribed to high serum sodium levels after excluding other possible causes. To exclude the possibility of the development of cerebral oedema during the correction of hypernatraemia, fundoscopy and assessment of the mental status were carried out on a regular basis. If the findings of physical examination were non-diagnostic, a brain computerized tomography (CT) was performed.

Analytical methods
Laboratory determinations were carried out by automated chemical analysis in our laboratory using an Olympus AU 600 analyser (Olympus Diagnostica, Hamburg, Germany), as previously described [9]. Arterial blood pH and PCO2 were determined using a pH blood gas analyser, and serum bicarbonate was calculated from blood pH and carbon dioxide tension according to the Henderson–Hasselbalch equation with an acidity exponent of 6.10 and solubility coefficient of 0.0301. Serum and urine osmolality was assayed using a vapour pressure osmometer.

Statistical analysis
The results were expressed as means ± SD. Comparisons between individual groups were performed using Student's t-test. Correlations between laboratory parameters were assessed with linear regression analysis. P values <0.05 were considered to indicate statistical significance. The ability of the various serum and urine parameters to discriminate between the two subgroups (hypovolaemic and euvolaemic) of hospital-acquired hypernatraemia was evaluated with the calculation of the areas under (AUC) the receiver-operating characteristics (ROC) plots and the optimal cut-off point was determined using the maximization of the Youden index [10].



   Results
Top
 Abstract
 Introduction
 Subjects and methods
 Results
Discussion
 References
 
Patient characteristics
One hundred and thirteen patients fulfilled the inclusion criterion of the serum sodium concentration. Of those, fifty (20 male, 30 female) were hypernatraemic on hospital admission and 63 (44 male, 29 female) during the course of their hospitalization. In both subgroups of patients, there were no statistically significant differences between males and females in all biochemical parameters measured (data not shown). Furthermore, the age of patients who developed hypernatraemia before hospital admission was not significantly different from that of the patients with hospital-acquired hypernatraemia (76.3 ± 12.2 vs 73.9 ± 14.4 years, P = NS). The period between the diagnosis of hospital-acquired hypernatraemia and the previous normal serum sodium laboratory results was 1.8 ± 0.9 (range 1–3) days. Taking into consideration that 9100 patients were admitted to our clinic during the 2.5-year study period, the incidence of hypernatraemia on admission was 0.5%. The incidence of hospital–acquired hypernatraemia in the 9108 patients at risk (58 patients admitted before the start of the study and 9050 patients admitted without hypernatraemia) was 0.7%. Finally, the overall incidence of hypernatraemia of the study population was 1.2%.

Compared with the patients who developed hypernatraemia during hospitalization, patients with hypernatraemia on hospital admission had higher serum concentrations of sodium, urea, creatinine, glucose, uric acid, phosphorus and magnesium. Also, they exhibited lower fractional excretion of sodium, potassium, urea and uric acid (Table 1).


View this table:
[in this window]
[in a new window]

  		Table 1. Laboratory characteristics in patients with hypernatraemia

 
In the majority of hypernatraemic patients (in both subgroups) more than one condition contributed to the development of hypernatraemia. Specifically, in patients with hypernatraemia on admission the most common of these causes were: febrile illnesses (temperature >37.8°C, n = 42, 84%; mainly pulmonary infections), uncontrolled diabetes mellitus [including both type 1 and type 2 diabetic patients with serum glucose levels > 180 mg/dl (10mmol/l), n = 20, 40%], gastrointestinal losses as osmotic diarrhoea after lactulose administration (n = 4, 8%), furosemide administration (n = 2, 4%), hypercalcaemia (n = 1, 2%), primary hypodipsia (a patient with frank psychosis denying the intake of fluids, 2%), central diabetes insipidus (a patient with small cell lung cancer and brain metastases exhibiting polyuria, polydipsia and urine osmolality of 265 mosmol/Kg, 2%) and high environment temperature (n = 18, 36%). Additionally, in almost all cases the water intake was markedly diminished because of the patients’ altered mental status. Moreover, contributing factors and causes of in-hospital hypernatraemia included febrile states (n = 45, 72%; mainly pulmonary infections), uncontrolled diabetes mellitus (n = 19, 30%), mannitol (n = 13, 21%) or furosemide (n = 6, 9%) administration, hypercalcaemia (n = 2, 3%) and gastrointestinal losses as osmotic diarrhoea after the administration of lactulose (10%). Finally, there were no statistically significant differences between the two subgroups of patients regarding the incidence of the two most frequent causes of hypernatraemia, namely febrile illnesses and uncontrolled diabetes mellitus.

Forty-one patients (82%) in whom hypernatraemia was present on admission had hypovolaemia. Of those, 34 subjects had also pre-renal acute renal failure (average creatinine level 2.6 ± 0.9 mg/dl, 230 ± 80µmol/l) that was proved reversible in all not fatal cases. Finally, 32 patients (64%) who were hypernatraemic before hospitalization had symptoms (mainly recent change in consciousness) that could be directly attributed to high serum sodium concentration.

Five (8%) of the patients with hospital-acquired hypernatraemia had marked oedema. These patients had an underlying medical condition that predisposed to the development of oedema [heart failure (n = 3) or hepatic cirrhosis (n = 2)] and had received normal saline (plus potassium chloride if hypokalaemia was also present) in the face of an acute insult such as septic shock, pancreatitis or gastrointestinal haemorrhage. In this patient group oedematous hypernatraemia was diagnosed. These patients exhibited also high serum urea concentration (150 ± 50 mg/dl, 53.6 ± 17.9 mmol/l) as well as low serum albumin level (2.8 ± 0.3 mg/dl). The remaining 58 patients had either hypovolaemic (n = 26) or euvolaemic (n = 32) hypernatraemia. It should be mentioned that the biochemical profiles of patients with hypovolaemic hypernatraemia (either on admission or hospital-acquired) were actually similar (data not shown). Table 2 depicts the comparison of laboratory parameters between the main subgroups of patients with hospital-acquired hypernatraemia. Patients with hypovolaemia had higher serum concentrations of urea and urea/creatinine ratio, but lower fractional excretion of sodium, chloride and urea than patients with euvolaemia, while there were no statistically significant differences regarding the serum sodium levels. Furthermore, (as judged by clinical assessment of the patients’ cardiovascular condition and/or by a clinical and biochemical estimate of the extracellular volume status), the average daily volume of administered infusate (normal saline ± potassium chloride) during the day preceding the development of hypernatraemia was 750 ± 200 ml vs 1500 ± 250 ml (P = 0.001) in patients with hypovolaemia and euvolaemia, respectively. Moreover, in all cases the electrolyte-free water intake was <1 l/day during the day preceding the onset of hypernatraemia. It should be noted that these patients had impaired ability to defend against hypertonicity through thirst and drinking and were dependent on prescribed fluid administration for the maintenance of water balance. In patients with euvolaemic hospital-acquired hypernatraemia, serum levels of urea [50.5 ± 22.1 mg/dl (18 ± 7.9 mmol/l) vs 60.1 ± 26.3 mg/dl (21.5 ± 9.4 mmol/l)], creatinine [1.4 ± 0.7 mg/dl, (124 ± 62 µmol/l) vs 1.25 ± 0.5 mg/dl (111 ± 44 µmol/l)], uric acid [6 ± 3.1 mg/dl (357 ± 184 µmol/l) vs 5.6 ± 1.8 mg/dl (333 ± 107 µmol/l)] and urea/creatinine ratio (49.5 ± 22 vs 50.5 ± 23) were not statistically different compared with the last laboratory results preceding the occurrence of hypernatraemia. On the contrary, patients with hypovolaemia had higher serum concentrations of urea [125 ± 48 mg/dl (44.6 ± 17.8 mmol/l) vs 76.1 ± 31.3 mg/dl (27.2 ± 11.2 mmol/l), P = 0.03], creatinine [1.5 ± 0.6 mg/dl, (133 ± 53 µmol/l) vs 1.4 ± 0.7 mg/dl (124 ± 62 µmol/l), NS], uric acid (7.5 ± 3.4 mg/dl (446 ± 202 µmol/l) vs 6.9 ± 3.2 mg/dl (410 ± 190 µmol/L), NS} and urea/creatinine ratio (81.7 ± 19.2 vs 57 ± 22, P = 0.001) in comparison with the previous laboratory results with normal serum sodium levels. It should be noted that no differences regarding serum levels of sodium, urea, creatinine, uric acid and urea/creatinine ratio were observed between the two groups in the laboratory tests performed before the diagnosis of hypernatraemia. Finally, 21 patients (33%) with hospital-acquired hypernatraemia exhibited symptoms that could directly be ascribed to hypernatraemia itself.


View this table:
[in this window]
[in a new window]

  		Table 2. Laboratory characteristics in patients with hospital-acquired hypernatraemia according to volume status

 
The determination of the areas under the ROC curves in the entire study population (including patients on diuretics) revealed that the ratio of urea to creatinine had the best discriminative value between the patients with the hypovolaemic hypernatraemia and those with the euvolaemic one (AUC value 0.985). The calculation of the Youden index showed that a urea to creatinine value of 57 could differentiate between the two groups with a sensitivity of 96.5% and a specificity of 100% (Figure 1).


Figure 1
View larger version (9K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Fig. 1. ROC plot of the ability of urea to creatinine ratio to descriminate between hypovolemic and euvolemic hypernatremia.

 
Concomitant electrolyte disturbances
Twenty-seven patients (54%) who were hypernatraemic prior to admission had at least one additional electrolyte disturbance. As shown in Table 3, the main electrolyte disorders in these patients were hypermagnesaemia (44%) and hyperphosphataemia (28%), while a coexistent pre-renal acute renal failure was also observed in all [the mean serum creatinine concentration was 2,8 ± 0,8 mg/dl (248 ± 71 µmmol/l); range 1.5–3.7 mg/dl (133 ± 327 µmmol/l)]. Moreover, 31 (49.2%) of the patients with hospital-acquired hypernatraemia had one or more additional electrolyte abnormalities. Hypophosphataemia (range 1.5–2.4 mg/dl, 0.48–0.77 mmol/l) was the most frequent electrolyte disorder observed in 16 patients (25.4%) and was accompanied by inappropriate phosphaturia (Formula >20%) in all. These hypophosphataemic patients had uncontrolled diabetes mellitus (n = 12) and/or were receiving either mannitol (n = 8) or furosemide (n = 5). Finally, it should be noted that in the two subgroups of patients, the serum sodium level was not correlated with the serum concentrations of all the other electrolytes studied (data not shown).


View this table:
[in this window]
[in a new window]

  		Table 3. Electrolyte abnormalities in patients with hypernatraemia on admission and hospital-acquired hypernatraemia

 
Outcome
Fourteen patients in whom hypernatraemia was present at admission died (mortality 28%). Patients who died were older than the rest of hypernatraemic patients on admission (81.3 ± 14.3 vs 74.3 ± 10.9 years, P = 0.07), while there were no statistically significant differences in the serum sodium concentration (160 ± 10 vs 160 ± 10 meq/l). As far as the fatal cases are concerned, 10 patients died several days after the restoration of the serum sodium levels, while four patients died before their hypernatraemia was corrected (the mean serum sodium concentration on the day of death was 151 ± 1; range 149–152 meq/l). However, in all fatal cases the rate of decrement in the serum sodium concentration was the desirable level (see patients and methods, page 6). Finally, although it has been proposed that hyperglycaemia may adversely affect the outcome in patients presenting with hypernatraemia, in our cohort the patients with inadequately controlled diabetes did not exhibit higher mortality rate (data not shown).

Thirty patients of a total of 63 patients with hospital-acquired hypernatraemia died (mortality 47.6%). The mean age (75.7 ± 13.9 vs 72.3 ± 14.9 years) as well as the peak serum sodium concentration (154 ± 6 vs 154 ± 6 meq/l) were not significantly different among fatal and non-fatal cases. Seventeen patients died with normal serum sodium levels, while 13 patients died before the complete correction of their hypernatraemia (the mean serum sodium concentration on the day of death was 153 ± 2; range 149–157 meq/l). It should be noted, however, that in all these patients the rate of decrease in the serum sodium concentration did not deviate from the chosen one. Overall, 44 patients out of 113 hypernatraemic patients died, giving a total mortality of 38.9%. Only in 17 cases (15%) did the fatal outcome antedate the complete restoration of hypernatraemia. It should be mentioned, however, that in all fatal cases, hypernatraemia occurred in the setting of serious underlying disease mainly sepsis or/and stroke. Finally, patients who developed hypernatraemia before hospital admission had a much lower mortality rate than patients with hospital-acquired hypernatraemia (28% vs 47.6%, P = 0.03), despite the fact that they had a higher peak serum sodium concentration (160 ± 10 vs 154 ± 2 meq/l, P = 0.000). It must be noted that none of our patients developed clinical or radiographic (n = 8) findings of cerebral oedema during the correction of hypernatraemia.



   Discussion
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
References
 
Hypernatraemia is a relatively common electrolyte disorder that has been studied mainly in elderly or mentally handicapped patients as well as in the entire hospital population [1,2,5–7]. Its incidence varies considerably (from <1% to >3%) depending on the population at risk and the definition of hypernatraemia used in the various studies [1,2,5,6]. In accordance with the data of the literature [6] the incidence of hypernatraemia in the present study was 1.2%. Previous studies showed that hypernatraemia develops usually after hospital admission [1,6,11]. On the contrary, in our study (that has restricted its analysis to patients hospitalized in an internal medicine ward), the incidence of the hospital-acquired hypernatraemia was only slightly higher than the incidence of hypernatraemia developed before hospitalization.

Previous studies in hospitalized patients showed considerable variation in the etiology of hypernatraemia reflecting largely inclusion of entire hospital populations of heterogeneous nature [12]. Therefore, these studies are not comparable with the present one. In the current study, the hypernatraemia on admission as well as the hospital-acquired hypernatraemia, in the vast majority of cases, had also heterogenous and multifactorial causes. It should be noted that there were no statistically significant differences between the two subgroups of patients regarding the incidence of the two most frequent causes of hypernatraemia, namely febrile illnesses and uncontrolled diabetes mellitus (though the mean serum glucose levels were considerably higher in patients who developed hypernatraemia before hospital admission).

It is known that net water loss, either in the absence of a sodium deficit (pure water loss) or in its presence (hypotonic fluid loss), accounts for the majority of cases of hypernatraemia [13]. Our study, however, clearly showed that the vast majority of hypernatraemic patients on admission (82%) exhibited hypovolaemic hypernatraemia. Thus, for all practical purposes, hypernatraemia on admission is essentially usually a hypovolaemic one.

Few patients (8%) with hospital-acquired hypernatraemia, in the present study, exhibited oedematous hypernatraemia. All these patients were severely ill and exhibited hypoalbuminaemia and azotaemia, hence they showed the same clinical and laboratory characteristics as the patients described by T. Kahn [14]. Excluding the oedematous patients, the majority of patients who developed hypernatraemia during hospitalization consisted mainly of patients with febrile states, uncontrolled diabetes mellitus, stroke, gastrointestinal losses or diuretics administration. In these settings, normal saline ± potassium chloride was administered intravenously to correct the hypovolaemia and the hypokalaemia, if present. Moreover, mannitol solution was also given in patients with cerebral oedema due to stroke. Consequently, in patients unable to drink water, hypernatraemia resulted from insensible losses in combination with the ongoing renal or extrarenal hypotonic losses coupled with the administration of isotonic or hypertonic solution. In fact, the addition of potassium chloride significantly increases the osmolality of administered fluids (1 L of 0.9% sodium chloride plus potassium chloride (two ampules containing 13.5 meq/l of potassium each) results in an osmolality of 360 mmol/lH2O) [15,16]. Moreover, in the subgroup of patients with hypovolaemic hospital-acquired hypernatraemia the infusates were prescribed in inadequate amounts and as a result patients developed not only hypernatraemia but also marked extracellular volume depletion. Additionally, a deterioration of the renal function of pre-renal origin was evident. It is of interest that these two subgroups of patients (hypovolaemic and euvolaemic) did not feature statistically significant differences in parameters used in assessing extracellular volume status, namely serum urea, creatinine, and uric acid concentrations and urea/creatinine ratio in the last laboratory tests performed before the diagnosis of hypernatraemia depriving us of an index that could predict the impending extracellular volume depletion. Consequently serial measurements of the serum sodium concentration are required in order to avoid not only the development of hypernatraemia but also the deterioration of renal function due to hypovolaemia. Additionally, our study showed that in all hypernatraemic patients the urea to creatinine ratio readily discriminates between two most common profiles of hypernatraemia, one consistent with extracellular volume depletion and another with euvolaemia. For practical purposes we suggest that a value of this ratio >57 may represent a useful tool that can guide the therapeutic approach in patients with elevated serum sodium concentrations, without the need of measurement of urine electrolyte values. This is of great importance in patients receiving diuretics or other medications (such as mannitol) that do not allow the reliable interpretation of the urine biochemical parameters used in the determination of the extracellular volume status. However, in the absence of diuretic use, the fractional excretion of sodium may represent a more sensitive index of extracellular volume status compared to the urea to creatinine ratio that is also subjected to several limitations. Indeed, this later ratio can be misinterpreted in numerous clinical situations such as upper gastrointestinal bleeding, increased protein catabolism, severe liver damage, low protein intake etc.

Hypernatraemia
Hypenatremia always represents a hyperosmolar state, so central nervous system (CNS) symptoms and signs (lethargy, weakness, irritability, hyperflexia, spasticity, seizures, and coma) are prominent. However, it is often difficult to attribute these manifestations directly to high serum sodium concentration given that the majority of hypernatraemic patients have also severe underlying disorders or pre-existing neurological dysfunctions. In the present cohort symptoms were ascribed to high serum sodium levels after carefully excluding other possible causes. Additionally, the resolution of these symptoms after the correction of high serum sodium levels (while the underlying condition remained relatively stable) was also strongly suggestive of the relationship between hypernatraemia and the above mentioned symptoms. It is worth reminding that almost 30% of our hypernatraemic patients had poorly controlled diabetes mellitus. There is evidence that hyperglycaemic patients with hypertonicity are symptomatic only if hypernatraemia is present. It has been reported that neurological symptoms may be absent in cases of severe gradually developing hyperglycaemia. Moreover, in cases of hyperglycaemic hyperosmolar syndrome altered mental status is predicted best by serum sodium levels; serum glucose levels alone are considered a poor indicator [17].

It is known that the treatment of hypernatraemia is directed toward correcting the cause of the fluid loss and replacing water and, as needed, electrolytes. The present study showed that the vast majority of hypernatremic patients on admission as well as a considerable percentage of patients with hospital-acquired hypernatraemia exhibited hypovolaemia. In such cases the strategy in fluid administration aims first at restoring blood pressure and vital organ perfusion and thus 0.9% saline should be initially administered. Furthermore, saline infusion will eventually lower tonicity, as its osmolarity is hypotonic to the hypernatraemic patients’ serum. In some cases, however, characterized by ongoing hypotonic fluid losses the administration of normal saline could be lead in to an increase in the serum sodium concentration. This increment of the serum sodium values can be attributed to the fact that the administered solution (in spite of being hypotonic as compared with patients’ serum) is relatively hypertonic compared with the ongoing hypotonic fluid losses [9]. However, it should be emphasized that when the patients’ haemodynamic status is sufficiently compromised the expansion of effective vascular volume is initially desired than the correction of hypernatraemia. In a second stage, 0.45% saline can be used to replace the free water loss. Half-isotonic saline or quarter-isotonic saline can be administrated, as an initial approach, in hypernatraemic patients with milder volume deficit. Furthermore, it is known that oedematous hypernatraemia is characterized by an excess of both sodium and water but the excess of sodium is proportionately greater. Therefore, its treatment consists of providing electrolyte free water intravenously (IV) or orally to diminish hyperosmolality along with IV furosemide administration in order to remove the excess sodium. As far as patients with euvolaemic hypernatraemia are concerned water losses far exceed solute losses and as a result the mainstay of therapy is 5% dextrose. Finally, it should be underlined that extreme care including serial measurements of the serum sodium concentration as well as the reassessment of the patients’ clinical status at regular intervals are of paramount importance to avoid rapid correction of hypernatraemia, which increase the risk of iatrogenic cerebral oedema, with possible catastrophic consequences.

Almost half of hypernatraemic patients in the present study had at least one concurrent electrolyte disorder. Moreover, patients with hypernataemia which preceded hospitalization did not differ concerning the frequency of concomitant electrolyte abnormalities as a whole in comparison with patients who developed hypernatraemia during hospitalization. However, hypermagnesaemia, hyperphosphataemia and hyperkalaemia were more frequent observed in patients who were hypernatraemic prior to admission and hypophosphataemia in patients with hospital-acquired hypernatraemia. The former can be mainly ascribed to the reduction of renal function while the latter was mainly attributed to inappropriate urinary phosphate loss due to osmotic diuresis (uncontrolled diabetes mellitus) and/or to administration of mannitol or diuretics. Other electrolytes disorders (hypokalemia and hypomagnesaemia) frequently observed had also multifactorial etiology. Low potassium or magnesium dietary intake in these severely ill patients in combination with increased renal (due to hyperglycaemia and/or mannitol or furosemide administration) or extrarenal (gastrointestinal) losses of these substances explains the majority of cases [18]. It should be mentioned, however, that hypokalaemia impairing the kidney concentrating ability might have played a role in the development of hypernatraemia per se [19]. Finally, taking into consideration the absence of a statistically significant correlation between serum sodium and other electrolytes levels, these disturbances should be ascribed mainly to the underlying cause of hypernatraemia rather than to hypernatraemia per se.

Our prospective series of hypernatraemia confirms previously described epidemiologic features of the syndrome, including the high mortality rates as well as the higher mortality in hospital-acquired hypernatraemia [1–4]. It is known that hypernatraemia, in adults, takes place in the setting of serious underlying disease, so the quantification of the influence of hypernatraemia on adverse outcome is difficult. In fact, it has been reported that hypernatraemia contributed to the death of 16% of patients [19,20]. This study, however, did not provide details on the way through hypernatraemia contributed to the fatal outcome [19,20]. On the contrary, in our study most deaths occurred several days after the correction of hypernatraemia (achieving the desired rate of decrement) and were possibly due to the progression of severe underlying disease. Indeed, the higher mortality rates observed in patients with hospital-acquired hypernatraemia could potentially be attributed to the more severe underlying conditions. In only 15% of our patients, who still were hypernatraemic when they died, hypernatraemia, theoretically, could have contributed to the fatal outcome. Thus, we believe that the present study provides a reliable estimate of the contribution of hypernatraemia regarding the mortality. Additionally, we attribute the relatively low contribution of hypernatraemia in mortality observed in our series to a scrupulously disciplined approach to management. Indeed, in patients with symptomatic hypernatraemia that was developed over a period of <48 h a rapid correction of serum sodium concentration (1–2 meq/l/h) was initially performed, while a slower rate of serum sodium reduction (<0.5 meq/l/h) was attained in patients with hypernatraemia of longer or unknown duration. We took special pains (performing serial measurements of the serum sodium concentration and reassessing the fluid prescription at regular intervals in the light of laboratory values and the patient's clinical status) to limit the decrease in the serum sodium levels up to 12 meq/l/per day. Thus, our experience emphasizes the general principle that the outcome of hypernatraemia depends mainly on the underlying disease provided that prudence is exercised in its management. Finally, it appears that patients’ age and serum sodium levels do not significantly affect the prognosis of hypernatraemia.

In conclusion, the incidence of hypernatraemia in the present study was 1.2% with a mortality rate of about 39%. Patients who developed hypernatraemia before hospital admission had a much lower mortality rate than patients with hospital-acquired hypernatraemia. There were two main profiles of hospital-acquired hypernatraemia, one consistent with extracellular volume depletion and another with euvolaemia. On the contrary, almost all hypernatraemic patients on admission exhibited hypovolaemia. Almost half of our hypernatraemic patients had at least one additional electrolyte disturbance. Hypermagnesaemia and hyperphosphataemia were the main electrolyte disorders in patients who were hypernatraemic prior to admission and hypophosphataemia in patients with hospital-acquired hypernatraemia.

Conflict of interest statement. None declared.



   References
Top
 Abstract
 Introduction
 Subjects and methods
 Results
 Discussion
 References
 

  1. Palevsky P, Bhagrath R, Greenberg A. Hypernatraemia in hospitalized patients. Ann Inter Med (1996) 124:197–203.[Abstract/Free Full Text]
  2. Himmelstein DU, Jones AA, Woolhandler S. Hypernatremic dehydration in nursing home patients: an indicator of neglect. J Am Geriatr Soc (1983) 31:466–471.[Web of Science][Medline]
  3. Ross EJ, Christie SBM. Hypernatraemia. Medicine (Baltimore) (1969) 48:441–473.[Medline]
  4. Macaulay D, Watson M. Hypernatraemia as a cause of brain damage. Arch Dis Child (1967) 42:485–491.[Free Full Text]
  5. Borra SI, Beredo R, Kleinfeld M. Hypernatraemia in the aging: Causes, manifestasions, and outcome. J Natl Med Assoc (1995) 87:220–224.[Medline]
  6. Snyder NA, Feigal DW, Arieff AI. Hypernatraemia in elderly patients: A heterogenous, morbid, and iatrogenic entity. Ann Intern Med (1987) 107:309–319.[Abstract/Free Full Text]
  7. MacDonald NJ, McConnell KN, Stephen MR, et al. Hypernatremic dehydration in patients in a large hospital fort the mentally handicapped. Br Med J (1989) 299:1426–1429.[Abstract/Free Full Text]
  8. Hillier TA, Abbott RD, Barrett EJ. Hyponatremia: evaluating the correction factor for hyperglycemia. Am J Med (1999) 106:399–403.[CrossRef][Web of Science][Medline]
  9. Liamis G, Kalogirou M, Saugos V, Elisaf M. Therapeutic approach in patients with dysnatraemias. Nephrol Dial Transplant (2006) 21:1564–1569.[Abstract/Free Full Text]
  10. Youden WJ. Index for rating diagnostic tests. Cancer (1950) 3:32–35.[CrossRef][Web of Science][Medline]
  11. Long CA, Marin P, Bayer AJ, Shetty HG, Pathy MS. Hypernatraemia in an adult in-patient population. Postgrad Med J (1991) 67:643–645.[Abstract/Free Full Text]
  12. Palevsky PM. Hypernatraemia. Semin Nephrol (1998) 18:20–30.[Web of Science][Medline]
  13. Gennarri FJ, Kassirer JP. Osmotic diuresis. N Engl J Med (1974) 291:714–720.[Web of Science][Medline]
  14. Kahn T. Hypernatraemia with edema. Arch Intern Med (1999) 159:93–98.[Abstract/Free Full Text]
  15. Liamis G, Kalaitzidis R, Elisaf M. Hypernatraemia during correction of hypercalcemia. Nephron (2000) 86:358.[CrossRef][Web of Science][Medline]
  16. Milionis HJ, Liamis G, Elisaf MS. Hypernatraemia in hospitalized patients: a sequel of inadvertent fluid administration. Arch Intern Med (2000) 160:1541–1542.[Free Full Text]
  17. Milionis HJ, Liamis G, Elisaf MS. Plasma sodium changes in the hyperglycaemic state: Clinical aspects of pathophysiology and management. Int J Diabetes Met (2001) 9:60–66.
  18. Elisaf MS, Milionis HJ, Siamopoulos KC. Electrolyte abnormalities in elderly patients admitted to a general medical ward. Geriatr Nephrol Urol (1997) 7:73–79.[CrossRef][Medline]
  19. Marples D, Frokiaer J, Dorup J, Knepper MA, Nielsen S. Hypokalemia-induced downregulation of aquaporin-2 water channel expression in rat kidney medulla and cortex. J Clin Invest (1996) 97:1960–1968.[Web of Science][Medline]
  20. Moritz M. Hypernatraemia in hospitalized patients. Ann Inter Med (1996) 125:860.[Free Full Text]
Received for publication: 16. 1.07
Accepted in revised form: 22. 5.07


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 NDT PlusHome page
G. Liamis, H. J. Milionis, and M. Elisaf
A review of drug-induced hypernatraemia
NDT Plus, October 1, 2009; 2(5): 339 - 346.
[Abstract] [Full Text] [PDF]

Home page
 Nephrol Dial TransplantHome page
T. Kahn
Hypernatraemia
Nephrol. Dial. Transplant., December 1, 2008; 23(12): 4080 - 4081.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
23/1/136    most recent
gfm376v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Liamis, G.
Right arrow Articles by Elisaf, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Liamis, G.
Right arrow Articles by Elisaf, M.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?