A case of hyponatremia in a previously healthy 76-year-old woman is described with discussion revolving around key questions:
Nephrol @ point care 2015; 1(1): e2 - e11
Article Type: QUESTIONS @ POINT OF CARE
- • Accepted on 18/03/2015
- • Available online on 26/05/2015
This article is available as full text PDF.
A 76-year-old astute, alert woman residing in a nursing home had been in excellent health until 2 months previously when her visiting general practitioner (GP) prescribed a thiazide diuretic (hydrochlorothiazide 25 mg/day) to treat her mild hypertension. Her blood pressure before therapy was 165/85 mm Hg. The GP also instructed the patient and the nursing staff to reduce “as much as possible” her dietary salt intake. After 6 weeks of therapy, her blood pressure had fallen to 135/80 mm Hg, but the nursing personnel and her visiting daughter noted some change in her mental status. She was less alert and had difficulties in concentrating – symptoms that were not usual for her. Her appetite had decreased over the last week, and she ate only 1 or 2 slices of bread with butter and jam but drank her usual number of 5 to 7 cups of tea per day. The nursing home staff also observed increasing gait instability, and she had sustained a fall, fortunately without any fractures, 2 weeks before admission to the hospital. Further, she had no history of cardiac or liver disease. Besides the hydrochlorothiazide, 25 mg/day, her medications also included aspirin, 80 mg/day.
On physical examination in the admission room, there was a normal skin turgor, no clear orthostatic fall in blood pressure or acceleration of the pulse. Her blood pressure was 125/80 mm Hg. Neurological examination revealed that she was fully oriented, with no focal findings, but the woman had an obvious gait disturbance that necessitated a walker for ambulation. The rest of the physical examination was normal. Her body weight is 58 kg, which, according to the patient, was unchanged over the last weeks. The patient weighed herself once a week, and in the nursing home notes no mention of body weight was found. An electrocardiogram (EKG) was compatible with mild hypokalemia. The lab results are given in
Laboratory test results for patient
|CKD-EPI = Chronic Kidney Disease Epidemiology Collaboration formula; eGFR = estimated glomerular filtration rate.|
|Creatinine, mg/dL (µmol)||1.4 (123 µmol/L)||-|
|Blood Urea, mmol/L||6 mmol/L (36.4 mg/dl)||-|
|Uric acid, mg/dL||4.9 (291.4 µmol/L)|
|Fractional excretion uric acid, %||9|
|eGFR (CKD-EPI creatinine), ml/min per 1.73 m²||39|
Initial laboratory tests showed a serum sodium level of 116 mmol/L, potassium of 2.8 mmol/L, chloride of 74 mmol/L, bicarbonate 30 mmol/L, an increased serum creatinine (1.4 mg/dl or 123 µmol/L)) and blood urea, and a plasma osmolality of 244 mOsm/kg water. The plasma uric acid is 4.9 mg/dl. The blood glucose was normal, and there were no further biological abnormalities. The calculated fractional excretion of uric acid was 9% and the estimated glomerular filtration rate (eGFR), calculated with the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula was 39 ml/min per 1.73 m².
In summary, this lady who had started a therapy 2 months earlier with 25 mg of thiazides daily presented with a profound hypotonic hyponatremia, moderate hypokalemia and mild metabolic alkalosis. On admission her eGFR was down to 39 ml/min. Previous values for electrolytes and/or kidney function were not available.
Given the absence of liver or heart disease and the clinically evaluated normal volume status, she was considered to have euvolemic hyponatremia. On the other hand, the continuous intake of a thiazide diuretic, the blood pressure of 125/80 mm Hg on admission, which was clearly below her previous blood pressure, and the increased serum creatinine and indeed blood urea values could also have been compatible with a mildly decreased circulating volume. A number of questions should be asked in such situation:
Is hyponatremia really hypotonic?
A decrease in serum Na (i.e., hyponatremia) generally, but not always, reflects a state of hypotonicity – i.e., both the sodium concentration and the effective osmolality are low. According to most authors, a serum sodium of 116 mOsm/L is a profound hyponatremia.
The presence of an increased concentration of solutes that do not cross the cell membrane (such as glucose, mannitol or glycine) can lead to the development of translocational hyponatremia as a result of the movement of water from the cells to the extracellular space. Such patients can have normal, or even high, serum osmolality. Another setting in which hyponatremia is associated with a normal serum tonicity occurs in the presence of high levels of lipids or proteins. This is designated as pseudohyponatremia and is the result of the increased proportion of serum volume taken up by these substances. The serum osmolality remains normal in pseudohyponatremia and can be used to eliminate this diagnosis (1).
The serum sodium in this patient was measured by ion selective electrode, and none of the causes that could lead to the presence of either hypertonic or isotonic hyponatremia were present, so that a true hypotonic hyponatremia was diagnosed. In addition, the serum osmolality was 244 mOsm/kg water, a value which always reflects a hypotonic state.
Is there an emergency present on admission necessitating acute therapeutic action to improve the electrolyte disturbances?
There was no evidence that in the case presented here, a hemodynamic or neurological emergency was present since there were no symptoms that could be related to acute hyponatremia – i.e., hyponatremia that should have developed over the last 48 hours. The mild disorientation and the fall 2 weeks before admission rather pointed to chronic hyponatremia.
A more important concern was the hypokalemia with mild but significant signs of hypokalemia on the EKG (see below paragraph 6 -what is the role of hypokalemia).
What is the best approach to obtaining a correct diagnosis?
A number of expert algorithms (2, 3) have been published to approach the problem of hyponatremia, but it is only very recently that evidence-based guidelines on this topic have become available (4). Although no fundamental differences in the initial approach to hyponatremia exist between the algorithms and the guidelines, some of the differences are worthwhile to consider briefly in the discussion of this case.
In the traditional diagnostic algorithm workup for hyponatremia, the evaluation of the status of extracellular fluid volume (ECFV) is critical because it best allows a differentiation among the causes of hypervolemic, hypovolemic and euvolemic hyponatremia, thus determining the treatment strategy (2).
However, the practice guideline experts, based on the notoriously difficult clinical evaluation of volemic status, particularly in elderly patients (5, 6), recommend starting with an analysis of the urinary sodium concentration (4). Determination of the sodium (Na) concentration from spot urine (urinary Na excretion [U-Na]) or fractional urine Na excretion (FE-Na) is diagnostically useful and considered the reference standard to differentiate decreased effective arterial blood volume (EABV) in hypovolemic or hypervolemic disorders (U-Na <30 mmol/L) from euvolemic hyponatremia as in the syndrome of inappropriate antidiuresis (SIAD) (U-Na >30 mmol/L) (7). Because not all patients with SIAD have elevated circulating levels of arginine vasopressin (AVP), the term SIAD was proposed as a more accurate description of this condition than the previously used term
Fenske et al (8) studied a total of 86 consecutive hyponatremic patients (serum Na <130 mmol/L) and classified them based on their history, clinical evaluation, osmolality and saline response to isotonic saline, into a SIAD and non-SIAD group. A total of 31 patients (36%) had a diagnosis of SIAD, and 55 (64%) were classified as non-SIAD. There were 57 patients (68%) who were on diuretics (15 in the SIAD group, 42 in the non-SIAD group). In the absence of diuretic therapy, SIAD was accurately diagnosed using U-Na (area under the receiver operating characteristic curve 0.96; 95% confidence interval [95% CI], 0.92-1.02). However, in patients on diuretics, the diagnosis was unreliable (area under the curve 0.85; 95% CI, 0.73-0.97). There, the FE of urinary uric acid (FE-UA) performed best compared with all other markers tested (area under the curve 0.96; 95% CI, 0.92-1.12), resulting in a positive predictive value of 100% if a cutoff value for FE-UA of 12% was used. In contrast to those for sodium and urea, the transport mechanisms of urate are localized exclusively in the proximal tubule. Therefore, a direct interaction with common diuretics is not to be expected. However, changes in ECFV are important factors modulating urate excretion. In healthy euvolemic subjects, FE-UA is approximately 10% (9). Contraction of ECFV decreases FE-UA, and expansion of ECFV enhances FE-UA, an effect independent of the increase in urinary flow (10). Although the mechanisms of either effects are unknown, one may speculate that urate reabsorption is indirectly coupled to Na transport by an electroneutral anion exchanger (11), and therefore, an increased proximal Na reabsorption explains the decreased fractional urate excretion in volume-depleted disorders.
Following the guidelines, a urinary analysis had already been performed in the admission room of the hospital in our study case (
A high urine osmolality was also found, which, in the presence of plasma hypotonicity, usually indicates a high concentration of vasopressin within the serum; in these cases, it is unlikely that fluid restriction will correct the hyponatremia, unless fluid intake is restricted considerably (17). The response of patients to fluid restriction can be predicted using a quantification of water and electrolyte balance – electrolyte free water clearance (CeH2O) (18) in the equation:
Approaches to raising serum tonicity with fluid restriction
|U/P electrolyte ratio||Insensible water losses||Expected net water loss||Recommended water consumption|
|These approaches assume that urine sodium and potassium losses are replaced, that a patient has an average body surface area of 1.73 m2 and eats a normal diet, and are based on a calculation for the period during which the next 1 L of urine is excreted. Reproduced from (18).|
|U/P = urine to plasma ratio.|
|>1.0||800 mL||800 mL||0 mL|
|0.5-1.0||800 mL||800-1,300 mL||Up to 500 mL|
|< 0.5||800 mL||1,300-1,500 mL||Up to 1,000 mL|
What is the best approach to raise serum Na?
Approaches to raising serum Na by fluid restriction can be guided by the patient’s urine to plasma (U/P) electrolyte ratio as summarized in
The patient was also not taking drugs that are known to be associated with the SIAD (
Drugs that may cause syndrome of inappropriate antidiuretic hormone secretion (SIAD)
|Reproduced from (19).|
|Antidepressent agents (selective serotonin reuptake inhibitors, tricyclic antidepressants)|
|Nonsteroidal anti-inflammatory drugs|
|Desmopressin (DDAVP, Ferring, Kiel, Germany), vasopressin|
The urinary sodium level (>20 mEq/L), urinary osmolality (>100 mOsm/kg) and relatively low serum uric acid levels are findings that also can support the diagnosis of SIAD. However, a SIAD diagnosis is an exclusion diagnosis and cannot be considered in a patient on thiazide therapy.
With regard to the age of the patient, changes in the renal concentrating mechanism brought about by aging have been extensively studied, but the effect of aging on the renal diluting process has received much less attention. When given a water load, healthy elderly persons can usually readily dilute their urine to <100 mOsm/kg, but the rate of free water excretion is slower than in younger controls (16). This decrement is further enhanced if they are receiving thiazide diuretics or nonsteroidal antiinflammatory drugs, both of which are commonly used in this population.
The subtle impairment in the excretion of water may be due to age-related reductions in GFR because creatinine clearance was substantially lower in the older cohort (16). A decrease in the expression of the Na-K-2Cl cotransporter in the ascending limb of the loop of Henle and the Na-Cl cotransporter in the distal tubule has been reported in aging rodents (20). These changes would result in increased delivery of solute to more distal sites of the nephron, limiting free water clearance. Whether such down-regulation occurs in humans is not known, but, if present, it could impair both maximal concentrating and diluting abilities. Finally, an age-related decrement in the percentage of body water content makes older persons more prone to dysnatremias because smaller disturbances in water balance will cause greater changes in the serum sodium concentration. Nonetheless, most elderly persons have well-preserved urinary diluting ability, and the development of hyponatremia is likely to supervene only when additional pharmacological or pathological processes are operant, as they frequently are with advancing age.
What are the risk factors, symptoms and pathophysiology of thiazide-induced hyponatremia?
In a study of 223 patients with chronic thiazide-induced hyponatremia (mean serum Na+ 116 mmol/L), symptoms reported included malaise/lethargy, dizzy spells, vomiting, falls, headaches and seizures (12). In a study of adult outpatients newly treated for hypertension over a 10-year interval, it was found that over 3 in 10 patients developed hyponatremia among those receiving thiazide diuretics. The overall relative risk of hyponatremia was approximately 60% higher in patients exposed to thiazide diuretics compared with alternative therapy, and appeared to be similar regardless of patient age or sex. The increased risk of hyponatremia began early after starting treatment and persisted for at least a decade (21). This risk is particularly high with a high dose of thiazide (22). Also, elderly and very elderly women seem to be at an especially high risk of developing hyponatremia (22, 23). Of note, many elderly subjects are commonly treated with selective serotonin reuptake inhibitors for depression or anxiety, and the concomitant use of diuretics and selective serotonin reuptake inhibitors also increases the risk of hyponatremia (24).
Although the exact underlying pathophysiological mechanisms of diuretic-induced hyponatremia are unclear, thiazides may induce salt and water depletion, stimulate excessive thirst and inappropriate ADH secretion and hypokalemia, and may induce renal-diluting disturbances particularly in elderly patients, and may provoke ADH-independent water retention (25).
The effect most responsible for thiazide-induced hyponatremia is the induced reduction of free water clearance by the kidneys. Thiazide diuretics interfere with maximum dilution of urine by blocking sodium chloride cotransport in the distal tubule, the main diluting site in the kidney. Excretion of sodium is increased, while excretion of free water is diminished (25).
Inappropriate secretion of vasopressin (antidiuretic hormone; SIAD) has also been implicated in the pathogenesis of thiazide-induced hyponatremia (13, 14), and a subset of patients with rapidly developing hyponatremia with thiazides share characteristic laboratory findings seen in SIAD, with a low serum uric acid level and increased uric acid clearance (26, 27). In some clinical settings of thiazide-induced hyponatremia, release of vasopressin is thought to be due to decreased EABV. Notwithstanding, at least in some patients, the degree of decreased EABV does not seem to be large enough to cause the release of vasopressin (28). Acutely reducing EABV by as much as 7% in healthy adults has little effect on vasopressin levels; a 10-15% reduction in EABV is required to double plasma vasopressin levels (29). In a series of patients with thiazide-induced hyponatremia reported by Sonnenblick et al (30), 4 patients had their plasma vasopressin level measured, and in 3 of them, it was either low or below the level of detection. Ghose (31) reported measurements of plasma AVP in 6 patients with diuretic-induced hyponatremia. In 3 of these patients, it was ~0.4 pg/mL, while it was ~1 pg/mL in the other 3. Urine osmolality was 300 mOsm/kg water or less in 3 of 5 patients in whom urine osmolality was reported. Oh et al reported 2 patients with what these authors called “trickle-down hyponatraemia.” Both of these patients had a low rate of excretion of osmoles, urine osmolality of close to 325 mOsm/kg water and undetectable levels of vasopressin in their plasma (32). Thaler et al (33) reported a patient with hyponatremia and low dietary solute intake in whom urine osmolality was 81 mOsm/ kg water and the vasopressin level in plasma was below the limits of detection.
Recently, copeptin, a surrogate marker for vasopressin, has been used in the study of hyponatremia. Copeptin is an amino acid glycopeptide that is released from vasopressin’s precursor peptide in equimolar quantities with vasopressin. In contrast to vasopressin, copeptin is stable and easier to measure, and appears to be a reliable surrogate of vasopressin activity. Fenske et al (8, 34) measured copeptin levels along with other serum and urine parameters in 106 patients admitted to the hospital with plasma sodium levels less than 130 mmol/L. Patients with primary polydipsia had copeptin levels below controls, indicative of suppressed vasopressin. Patients with hypovolemic and hypervolemic hyponatremia, classically described as states of decreased EABV, had the highest copeptin values, followed by those patients with SIAD. Patients with diuretic-induced hyponatremia had levels lower than those for SIAD, essentially comparable with those of controls. In sum, vasopressin may play a role in the hyponatremia associated with thiazides, but the cause of vasopressin secretion and the relative contribution of vasopressin independent of the other effects of thiazides to decrease water excretion are unclear.
There are two additional factors that may reduce electrolyte-free water excretion sufficiently and lead to the development of hyponatremia even in the absence of vasopressin effects: the volume of filtrate that is delivered to the distal nephron and the volume of water that is reabsorbed in the inner medullary collecting duct through its residual water permeability (RWP) (35, 36).
The volume of distal delivery of filtrate will be reduced if the GFR is decreased or if the fractional reabsorption of NaCl in the proximal convoluted tubule is increased. The latter will occur in response to decreased EABV. This can be due to a total body deficit of NaCl – e.g., by diuretic use in a patient who consumes little salt as was the case in our patient. To put this in a quantitative perspective, consider the case described above where the patient had a low baseline GFR of ±50 l/day. The use of diuretics and a low-salt diet led to a sodium deficit and a mild reduction in EABV. If she were now to reabsorb 90% of her GFR (which may be even lower because of the mild reduction in her EABV) in the proximal convoluted tubule instead of 83%, less than 5 l/day will be delivered distally, this would be the maximum urine volume she could excrete. This volume exceeds the usual daily intake of water, but hyponatremia can still develop in such a patient because there is water reabsorption by RWP along the inner medullary collecting duct even in the absence of vasopressin action, as was the case in the patient described above. Even a relatively small decrease in EABV leads to sympathetic activation: β-adrenergic stimulation activates the renin-angiotensin-aldosterone system, both of which will activate proximal tubular reabsorption of sodium and water and therefore further reduce distal delivery of filtrate.
There are other examples of patients in whom a low distal delivery of filtrate and water reabsorption by RWP may be important factors in the pathophysiology of hyponatremia. RWP is water reabsorbed from the inner medullary collecting duct because it is constitutively somewhat permeable to water (35, 36).
One group of patients where these mechanism may play a role in their hyponatremia is those who exercise and markedly reduce their dietary intake to lose weight (33). These patients consume a large volume of water, a diet low in salt and protein, and their exercise causes a loss of NaCl in sweat. They thus have some reduction in EABV (likely only mild since they continue to exercise), increased reabsorption of NaCl along the proximal convoluted tubule, and a modest reduction in distal delivery of filtrate. Therefore, for these patients to develop hyponatremia, in addition to a large water intake, they will need a large volume of water reabsorption in the inner medullary collecting duct through RWP. Perhaps a larger proportion of the potential urine undergoes retrograde flux, and this may present a greater opportunity to reabsorb more of the distal delivery in the inner medullary collecting duct.
Beer potomania (37) is another condition in which in some patients a low distal delivery of filtrate and water reabsorption by RWP may be important factors in the pathophysiology of their hyponatremia. In this setting, there is a very large intake of electrolyte-free fluid in the form of beer. In addition, since each liter of urine will have perhaps 10 mmol of Na, a deficit of Na will develop over a number of days if the intake of NaCl is low, and hence the volume of distal delivery of filtrate will be decreased. A low intake of protein contributes by decreasing the number of osmoles delivered to the inner medullary collecting duct, and hence, more water is reabsorbed by RWP.
What is the role of hypokalemia in the pathophysiology of hyponatremia?
Hypokalemia induced by thiazides is an independent predictive factor for the development of hyponatremia (15, 38). Because intracellular and extracellular osmolality are always equal, loss of either sodium or potassium, unless accompanied by loss of water, will result in hypotonicity. Although it is intuitively evident why changes in body sodium and water levels should determine serum sodium concentration, the role of potassium is less obvious, but as is illustrated in the case presented here, it is nevertheless very important. Edelman et al (39) showed that serum sodium concentration is a function not only of total exchangeable sodium and total-body water, but also of total exchangeable potassium. The primary mechanism is that potassium depletion results in a shift of sodium into the cell with a commensurate exit of potassium from the cell into extracellular fluid. The reverse occurs during potassium repletion and explains Laragh’s (40) observation that oral potassium chloride administration resulted in an increase in serum sodium levels in hyponatremic patients in the absence of administered sodium. A similar observation was reported by Fichman et al (14) in patients with diuretic-induced hyponatremia and hypokalemia. This effect of potassium repletion to increase serum sodium concentration may be enhanced by the entry of chloride into the cell along with the potassium, which renders the cell hypertonic and draws water from the extracellular fluid. Potassium entry may also be accompanied by the movement of hydrogen ions from the intracellular to extracellular space, where they are buffered and thereby made osmotically inactive. This would decrease effective extracellular tonicity, again causing water to move into the cells, increasing the extracellular concentration of sodium. Whichever mechanism is dominant, the important observation is that potassium depletion could be associated with hyponatremia, and potassium repletion results in an increase in serum sodium concentration (41).
A recent study investigated differences in risk of hyponatremia between chlorthalidone and hydrochlorothiazide, adjusted for daily dose. It appeared that hyponatremia was more common with chlorthalidone than with hydrochlorothiazide at an equal dose per day (adjusted odds ratio was 2.09; 95% CI, 1.13-3.88, for 12.5 mg/day; and 1.72; 95% CI, 1.15-2.57, for 25 mg/day). This study suggests the need for a lower dose of chlorthalidone than hydrochlorothiazide to achieve similar blood pressure reduction which likely compensates for the increased risk of hyponatremia at an equal dose of the diuretics (42).
What are the options for therapy to raise the plasma sodium in this patient?
The major cause of the hyponatremia in the case described here was a deficit of NaCl, potassium and a mild positive balance of water. Although some water restriction could be advised, the main goal was to administer NaCl and KCl. It should be clear that in cases of hypovolemic hypotonic hyponatremia, there is no place for therapies that are currently used in the treatment of chronic euvolemic or hypervolemic hyponatremia, including therapies interfering with the actions of vasopressin in the kidney.
In the study case, KCl was given together with isotonic NaCl (1 L 0.9% NaCl + 30 mEq KCl per L) at a rate of 250 ml/hour for 4 hours, with the goal of raising the plasma Na to an upper limit of 4-6 mmol/L per 24 hours initially because the patient had many risk factors for osmotic demyelination (i.e., she was an elderly woman with potassium depletion). Frequent monitoring of the plasma electrolytes, every 2 hours in the beginning and later every 6 hours after start of therapy, was needed. As soon as water diuresis developed, the infusion rate was slowed. In the case of a too rapid increase in plasma sodium, small doses of desmopressin could be administered at regular intervals to control the speed of plasma sodium increase. This intervention was, however, not needed in this case.
After 48 hours of treatment, the plasma sodium was 128 mmol/L, and the patient was put on an oral water restriction and oral KCl solution. The thiazides were stopped during the whole duration of the hospitalization. After 10 days, she could be discharged from the hospital with a plasma sodium of 134 mmol/L and a plasma potassium of 3.8 mmol/L. Her blood pressure was 145/85 mm Hg. Her primary care physician was advised not to prescribe a thiazide again in case of further development of hypertension. The patient was advised to eat a more balanced diet, and the nursing home was asked providing a diet containing at least 5-8 g of salt per day.
Potassium depletion poses a vexing challenge in managing hyponatremia. Failure to consider the effect of potassium replacement on the level of serum sodium has caused many cases of osmotic demyelination (38, 43), and prudent management requires that the clinician first focuses on potassium replacement. Considering that 1 mmol of retained potassium affects serum sodium as much as 1 mEq of retained sodium, even partial correction of potassium depletion can cause an excessive rise in serum sodium without sodium administration. Depending on clinical circumstances, potassium can be administered orally, intravenously (i.v.) or by both routes. It should be recalled that potassium depletion predisposes to osmotic demyelination, and it frequently coexists with additional risk factors for this complication. Retention of only 3 mEq/kg of potassium is sufficient to raise serum sodium by as much as the daily threshold of 6 mmol/L (for total body water of 50% body weight). The oral route for K+ administration is safer if feasible and if bowel sounds are present, because this avoids local vascular damage from high local potassium concentrations in i.v. fluids. The oral route also enables the administration of KCl without a large volume of i.v. fluid. It is simply impossible to calculate the negative balance of K+ at the outset. Frequent monitoring of EKGs is essential until the plasma K+ rises to a safe level of ±3.5 mmol/L.
During therapy with isotonic saline and KCl repletion, a substantial water diuresis can occur because vasopressin release is inhibited. Understanding this pathophysiology has clinical implications for the management of patients with hyponatremia. Consider if the patient in our case example had presented to the emergency department. She is found to have hyponatremia which is thought to be due to stimulation of vasopressin release due to decreased EABV owing to her intake of a thiazide diuretic, and hence is given isotonic saline to re-expand her EABV. A relatively small volume of saline (especially, if were given as a bolus) may be sufficient to reduce the fractional reabsorption of filtrate in the proximal convoluted tubule and increase its distal delivery. If the fractional reabsorption in proximal convoluted tubule is decreased to say 83% of the GFR of 40 l/day, distal delivery of filtrate will increase to ~7 l/day. This exceeds water reabsorption by RWP causing water diuresis. Because of her small muscle mass, even a modest water diuresis may be large enough to cause a rapid rise in plasma Na and hence the risk of osmotic demyelination syndrome especially if she is malnourished or K- depleted (38, 43). This combination occurs particularly in elderly patients like the patient described above who is taking thiazide diuretics and consuming little salt (tea and toast diet), in whom contraction of the ECFV is likely to be a cause of vasopressin release. As mentioned above, these patients also demonstrate an inability to excrete water because of a diminished distal delivery of filtrate (“trickle-down” hyponatremia) (44). Consequently, an infusion of saline can cause water diuresis and a larger than expected rise in the serum Na because of the patient’s small muscle mass. If the patient is hypokalemic, then there are 2 other possible reasons for a rapid rise in the serum Na. First, hypertonic KCl will raise the serum Na. Second, if the patient has an ileus and the intestinal luminal fluid contains Na+, a rise in the arterial plasma K might trigger an occult infusion of saline by increasing gastrointestinal motility. Parenthetically, it is possible that Na+ might have entered the lumen of the gastrointestinal tract by diffusion (45).
Frequent monitoring of inputs and outputs is thus needed, and the serum Na and K should be measured at frequent intervals (every 2-4 hours). If water diuresis were to occur, DDAVP must be given to stop the loss of water in the urine and prevent a rise of PNa that exceeds the daily upper limit of 4-5 mmol/L. Water diuresis indicates that the EABV is restored sufficiently to increase the distal delivery of filtrate. If the patient is still hyponatremic despite the onset of water diuresis, the plan for therapy is to allow a daily negative balance of water to achieve the desired rise in the serum Na that day. One can give DDAVP to limit the urine output to ensure that the rise in serum Na does not exceed the daily maximum limit.
Two additional points should be made: First, one should not try to increase the ECFV to one that is equal to individuals who eat a higher amount of salt, because they need an expanded ECFV to excrete their higher salt load. Second, if a water diuresis does not occur after the EABV has been expanded and, after discontinuation of eventually administered DDAVP, one should look for other causes for the release of vasopressin (
Summary of all causes that can be associated with unintentional overcorrection of hyponatraemia
|Cause of hyponatremia||Mechanism of escape from antidiuresis|
|Reproduced from (46).|
|SIAD = syndrome of the inappropriate secretion of antidiuretic hormone; SSRI = selective serotonin reuptake inhibitor.|
|Hypovolemia||Volume repletion reverses baroreceptor-mediated vasopressin secretion|
|Beer potomania, tea and toast diet||Increased solute intake enhances delivery of glomerular filtrate to distal diluting sites|
|Thiazide diuretics||Discontinuation of diuretic restores diluting function of the distal tubule|
|SSRI||Discontinuation of antidepressant eliminates drug-induced SIAD|
|Desmopressin||Discontinuation of synthetic vasopressin eliminates antidiuretic state|
|Hypopituitarism||Cortisol replacement restores ability to suppress vasopressin secretion|
|Addison disease||Volume and cortisol replacement|
|Hypoxemia||Correction of hypoxemia eliminates non-osmotic stimulus for vasopressin|
|Nausea, surgery, pain, or stress||Spontaneous resolution of SIAD|
Patients with a previous episode of thiazide-induced hyponatremia demonstrate increased susceptibility to a recurrence. When compared with both elderly and young controls, patients with a prior history had lower basal urine osmolality and demonstrated a greater fall in serum Na after rechallenge with a single dose of diuretics. Interestingly, although both control groups lost weight after receiving the diuretic, the patients who developed hyponatremia gained weight (47). Taken together with the frequent lack of clinical evidence for volume depletion, these data suggest a role for abnormal thirst and water intake in individuals who develop thiazide-induced hyponatremia (3).
How should the concomitant hypokalemia be treated, and is it an additional danger in the treatment of the hyponatremia?
SIAD has been implicated in the pathogenesis of thiazide-induced hyponatremia for many years (13, 14). Many patients with hyponatremia resulting from thiazides appear euvolemic. In 1971, Fichman et al (14) studied 10 patients with thiazide-induced hyponatremia who appeared euvolemic and had significant hypokalemia with evidence of increased ADH activity in plasma. They postulated that potassium depletion increased the sensitivity for release of ADH, but other factors such as polydipsia and sodium depletion were present. In contrast, Ashraf et al (26) studied patients with thiazide-induced hyponatremia who had low free water clearance after rechallenge with a thiazide and with undetectable ADH levels. The subset of patients with rapidly developing hyponatremia with thiazides shared characteristic laboratory findings seen in SIAD, with a low serum uric acid level and increased uric acid clearance (27, 30). Vasopressin may be stimulated by nausea that often accompanies hyponatremia, and it is difficult to measure the hormone accurately.
What are the lessons from this case for the management of chronic “asymptomatic” or mildly symptomatic hyponatremic patients
Many patients with long-standing hyponatremia, even when severe (sodium <125 mmol/L), appear by most clinical criteria to be essentially asymptomatic, probably as a consequence of the restoration of brain cell volume brought about by the exit of intracellular electrolytes and organic osmolytes. The loss of these solutes, although critical to the cell volume adaptive process, leaves the brain with a decreased amount of various substances, such as glutamine, a major neurotransmitter, that are important for normal neuronal function (48).
However, as mentioned above, in patients with chronic thiazide-induced hyponatremia (mean serum Na of 116 mmol/L), symptoms varying in severity were reported (12). Thus, although these patients may appear to be “asymptomatic,” or at least may be presenting with minimal symptomatology, more careful studies have led to the question, Does chronic asymptomatic hyponatremia really exist? (49). In this regard, Renneboog and colleagues administered a battery of visual and auditory tests to 16 patients with chronic hyponatremia (mean age 63 years; mean serum sodium concentration 128 mEq/L) (50). Hyponatremia was associated with an increase in error rate and latency time that was highly significant compared with patients who had a normal serum sodium concentration. Furthermore, these investigators reported significant disturbances in gait in 12 “asymptomatic” hyponatremic patients with a mean serum sodium of 128 mEq/L that were worse than those observed in patients with blood alcohol levels of 0.05%; these gait abnormalities corrected when the serum sodium levels returned to normal.
In a case-control study of 122 hyponatremic patients (mean serum sodium 126 mEq/L; mean age 72 years), these investigators found that the gait disturbance associated with hyponatremia culminated in an increase in risk for falls by an odds ratio of 67.4 (95% CI, 7.48-607.4; p = 0.001).
A subsequent case-control study of 530 patients with a mean age of 81 years also found that the presence of hyponatremia (mean serum Na 131 mEq/L) was associated with a fourfold greater risk of presenting with a fracture compared with age-matched normonatremic controls (51). A subsequent study also found this association of hyponatremia with large bone fractures in elderly patients (52).
A more recent prospective, population-based study of 5,208 elderly patients, 399 of whom were hyponatremic (mean serum sodium concentration 133 mEq/L), found a significant increase in nonvertebral fractures in the hyponatremic cohort (hazard ratio 1.39; 95% CI, 1.11-1.73) (53).
The propensity for fractures in elderly hyponatremic patients may not relate solely to gait disturbance but may also be enhanced by a direct effect on bone mineralization. Verbalis and colleagues reported a significant decrease in bone mineralization in rats when their serum sodium concentration was decreased to 110 mEq/L (54). Recently, evidence has also been accumulating to suggest that chronic “asymptomatic” hyponatremia may also have long-term consequences pertaining to bone metabolism, with several studies reporting an association between hyponatremia and fracture occurrence (55) or osteoporosis (54). More importantly, adults with mild hyponatremia (mean serum Na 133 mEq/L) displayed a significantly increased risk for osteoporosis at the hip (odds ratio 2.85; 95% CI, 1.03-7.86) and femoral neck (odds ratio 2.87; 95% CI, 1.41-5.81). These observations may be related to stimulation of osteoclastic activity and enhanced bone resorption in the setting of a low serum sodium concentration (56). Commensurate with the above discussion, our patient had disturbed gait and had sustained a fall, fortunately without fracture, 2 weeks before admission.
Thus, the rationale for initiating a therapeutic intervention to increase the plasma sodium concentration in even “asymptomatic” chronic hyponatremia appears compelling. As a disorder whose pathogenesis revolves around the retention of water and the kidney’s reduced ability to excrete it, the cornerstone of treatment of chronic hyponatremia has been restriction of water intake. This approach has the virtue of addressing the underlying mechanism responsible and is very attractive for its lack of any associated cost. However, experience has revealed that adherence to significant water restriction is problematic and that such restriction is poorly tolerated over time. Although a decrement in tonicity should in itself suppress thirst, a large portion of fluid intake is not driven by thirst but is rather determined by habit and other factors. Despite the absence of any scientific support, limitation of water intake is often strongly encouraged. Furthermore, water restriction is not always effective, particularly when the diluting defect is severe.
As was explained above, according to the elegant analysis of Furst and colleagues (18), when the sum of the concentration of urinary sodium plus potassium is greater than the serum sodium concentration, no electrolyte-free water is excreted and therefore almost no amount of water restriction will result in an increase in the serum sodium concentration. Only when the diluting defect is mild and this ratio is 0.5 will a tolerable restriction of approximately 1 l/day be of any therapeutic benefit.
Thus, the response to water intake restriction is variably effective and is often insufficient to adequately correct significant hyponatremia. In addition, patients with hypotonic hyponatremia like patients with SIAD continue drinking despite the serum hypotonicity that should suppress the thirst sensation. Elegant studies (57) have shown that the act of drinking suppressed thirst in both SIAD and controls but did not suppress plasma AVP concentrations in SIAD compared with controls. It appears thus that a downward resetting of the osmotic threshold for thirst in SIAD is present but that thirst responds to osmotic stimulation and is suppressed by drinking around the lowered set point. In addition, drinking does not completely suppress plasma AVP in SIAD.
A detailed discussion of the therapeutic options for treating chronic hyponatremia is beyond the scope of this paper, but extensive and very instructive reviews on this topic are available in the literature (1, 3, 19, 58-59-60-61-62-63-64).
Would this patient be approached differently in Europe and the United States?
One of the major differences between the European Best Practice guidelines for hyponatremia (4) and the expert panel recommendations (3) is the recommendation of the European guidelines against the use of vaptans in patients with expanded ECFV and the nonrecommendation of their use in patients with SIAD, while the expert panel recommends these vasopressin-blocking drugs (vaptans) for managing oligosymptomatic/moderately symptomatic hyponatremia associated with SIAD or heart failure. The major argument of the European guideline authors for not recommending the vaptans in these patients was that, compared with placebo, vasopressin receptor antagonists did not reduce the number of deaths, that there was always the risk of overcorrecting the hyponatremia and finally that there were recent signs of potential hepatotoxicity of these drugs (4).
Relevant for this discussion is that in addition to assessing changes in serum Na, the SALT-1 and SALT-2 trials also investigated changes in self-assessed health status following tolvaptan treatment. In the post hoc analysis of patients with SIAD, treatment with tolvaptan was reported to increase scores on the 12-Item Short Form (SF-12) General Health Survey (65). A significantly greater change in Physical Component Summary (PCS) score from baseline was also noted in patients with SIAD treated with tolvaptan compared with those who received placebo (3.64 vs. -0.16, respectively, p = 0.019). These findings suggest that patients who received treatment with tolvaptan perceived an improvement in physical functioning (65). Interestingly, a tendency toward an improvement in Mental Component Summary (MCS) score from baseline was also reported (5.47 change from baseline with tolvaptan vs. -0.45 change from baseline with placebo, p = 0.051). This increase in MCS score was of comparable magnitude to that observed in the full analysis of the SALT trials and approached, but failed to reach, statistical significance. The authors attributed this to the reduced patient numbers available within the SIAD subgroup analysis reducing the power of the analysis to detect differences between the treatment groups (65).
Questions remain about the optimum degree of correction of hyponatremia in specific patient populations; however, there is a growing body of evidence to suggest that prompt effective treatment of hyponatremia decreases morbidity and length of hospital stay.