The difficulty that nonnephrologists sometimes have with the differential diagnosis of hyponatremic patients often results from misinterpreting the significance of measured and calculated serum osmolalities, effective serum osmolalities (tonicities), and the influence of various normal (eg, serum urea nitrogen) and abnormal (eg, ethanol) solutes. Among the more commonly held misconceptions are that high serum urea or alcohol levels will, by analogy with glucose, cause hyponatremia, and that a normal (or elevated) measured serum osmolality in a hyponatremic patient excludes the possibility of hypotonicity. This article describes typical and deliberately comparative data of the serum levels of sodium, glucose, urea nitrogen, and mannitol and/or ethanol (if present); calculated and measured osmolality; effective osmolality; and the potential risk of hypotonicity-induced cerebral edema for each of 6 prototypical hyponatremic states. This provides a helpful educational tool for untangling these interrelationships and for clarifying the differences among various hyponatremic conditions.
Nonnephrologists, as well as house staff and other trainees, frequently have difficulty with the differential diagnosis in hyponatremic patients and with assessing their relative risk for developing cerebral edema. This difficulty often derives from misinterpreting (1) the measured and calculated serum osmolalities, (2) the effective serum osmolalities (tonicities), and (3) the influences of other normal and abnormal solutes. The confusion relates not only to the concepts but to the details of calculation and assessment. Some of the most commonly held misconceptions are shown in Table 1.
This article will clarify the interrelationships of serum sodium, serum glucose, serum urea nitrogen (SUN), and serum mannitol and/or ethanol (if present) concentrations; calculated and measured serum osmolality; effective serum osmolality; and potential risk of cerebral edema (induced by hypotonicity) in various clinic settings. The focal point for the discussion is Table 2, which provides typical and deliberately comparative laboratory data for each condition.
True hyponatremia is a reduction in serum sodium concentration to below the lower limit of normal, which results from a decrease in sodium content per unit volume of water. In true hyponatremia (Table 3), total body sodium and total body water levels may be decreased, normal, or increased, whereas in pseudohyponatremia these variables are unchanged from normal.
Pseudohyponatremia is a method-dependent, artifactual reduction of serum sodium concentration resulting from displacement of a portion of the water phase of the plasma by lipid (eg, in severe hypertriglyceridemia) or by protein (eg, in multiple myeloma).1-6 This problem (which may be observed with flame emission spectrometry) can be avoided by using a sodium-selective electrode without dilution.
Calculated serum osmolality (Osmc) is the osmolality estimated by using various equations that include the sum of the contributions of sodium, urea, and glucose.7-9(For the purposes of this discussion, we neglect the differences between osmolality and osmolarity. The small error caused by failing to use the correction factor for serum water content7,10 in these conversions is clinically unimportant.) We have used the equation commonly applied at the bedside: Osmc=2[Na]+SUN/2.8+GLU/18, where [Na] indicates serum sodium concentration, and GLU, serum glucose level. (Other, more accurate equations require somewhat more calculation.7,8) Dividing the SUN level (in milligrams per deciliter) by 2.8 converts the units to millimoles per liter, because nitrogen has an atomic weight of approximately 14 and a molecule of urea contains 2 nitrogen atoms. Similarly, dividing the serum glucose concentration (in milligrams per deciliter) by 18 converts the units to millimoles per liter, because the molecular weight of glucose is approximately 180.
Measured serum osmolality (Osmm) is the osmolality directly measured by some type of osmometer, most commonly with the technique of freezing-point depression. (Some laboratories use vapor pressure osmometry, which may fail to detect certain volatile solutes, including ethanol.11,12)
The osmolal gap is the difference between the measured and the calculated osmolality (Osmm−Osmc).7-9,13,14 According to Dorwart and Chalmers,7 the average (± SD) osmolal gap in normal individuals with the use of molar units has a negative value (−9.0±6.4 mmol/L). Although a range of normal values has not been defined, positive values in excess of 10 mmol/kg H2O are considered to be clearly abnormal. To simplify the comparisons in Table 2, we assume that the average normal osmolal gap in each case is 0 mmol/kg H2O (a value within the normal range), and that values greater than 10 mmol/kg H2O are abnormally high.
Effective osmolality, a term synonymous with tonicity, is that portion of total osmolality that has the potential to induce transmembrane water movement.9,12,14 Substances (such as urea and ethanol) that easily cross cell membranes and contribute to measured osmolality, but not to tonicity, are called ineffective solutes.9,11,12,14 In contrast, effective solutes (such as sodium and mannitol) are confined largely to the extracellular fluid (ECF) compartment and contribute to both measured osmolality and tonicity. Effective osmolality can be estimated from clinical data, but its value is not actually measured in clinical laboratories. For ease of comparison, we have calculated the values of effective osmolality cited in the table by subtracting the contributions of urea (and ethanol, if present) from the measured osmolality.
The risk of cerebral edema refers to the immediate likelihood that, before treatment, water will move from the ECF into the intracellular fluid compartment of the brain, thereby producing the osmotic type of cerebral edema. When the effective osmolality of the ECF is low in comparison to that of the intracellular fluid, there is an increased risk of cerebral edema. In Table 2, the relative risk in various abnormalities is compared with the absence of risk in a normal individual. When the effective osmolality of the ECF is low in comparison with that of the intracellular fluid, there is an increased risk of cerebral edema.
Description of prototypic conditions
Table 2 is constructed to illustrate typical values for normal individuals and for patients with each of 7 pathological conditions, 6 of which are associated with hyponatremia. Data are provided for serum sodium, serum glucose, SUN, serum mannitol or ethanol (if either is present), Osmc, Osmm, osmolal gap, and effective osmolality, as well as the risk of cerebral edema. For illustrative consistency, the normal (typical) values of the variables are chosen as follows: serum sodium, 140 mmol/L; serum glucose, 5 mmol/L (90 mg/dL); SUN, 5 mmol/L (14 mg/dL); Osmc, 290 mmol/kg H2O; Osmm, 290 mmol/kg H2O; osmolal gap, 0 mmol/kg H2O; and effective osmolality, 285 mmol/kg H2O. All of the hyponatremic states are arbitrarily given a serum sodium level of 120 mmol/L.
The Osmc=(2)(140)+5+5=290 mmol/kg H2O. Because, as mentioned above, the normal osmolal gap is arbitrarily set at 0 mmol/kg H2O, the Osmm is also 290 mmol/kg H2O. The effective osmolality is 290 mmol/kg H2O minus the 5-mmol/kg H2O contribution of urea, ie, 285 mmol/kg H2O. Because the effective osmolality (or tonicity) is normal, there is no risk of cerebral edema.
The Osmc is (2)(120)+5+5=250 mmol/kg H2O, a value that is lower than normal; the Osmm is also 250 mmol/kg H2O, so the osmolal gap is 0 mmol/kg H2O. In this case, the effective osmolality is 250−5, ie, 245 mmol/kg H2O, an abnormally low value that indicates increased susceptibility to cerebral edema. Patients who are hypo-osmolal must also be hypotonic, but even patients who are either normo-osmolal or hyperosmolal may still be hypotonic (see the examples given below of true hyponatremia together with either a high SUN level or a high serum ethanol level).
In a patient with profound hypertriglyceridemia, as discussed above, the serum sodium concentration (depending on the method of analysis; see "Definitions of Terms" section, above) might be reduced to 120 mmol/L.3-5 If the SUN and serum glucose levels are normal, this must result in a subnormal Osmc; in this example, the Osmc will be (2)(120)+ 5+5=250 mmol/kg H2O. An osmometer measures solutes only in the water phase of serum, and therefore the measurement is not substantially influenced by the hyperlipemia.4 Because the Osmm will be 290 mmol/kg H2O, the osmolal gap will be 290−250=40 mmol/kg H2O, an abnormally high value. Thus, lipid- or protein-induced pseudohyponatremia must be associated with an elevated osmolal gap.2 Therefore, the absence of an increased osmolal gap rules out pseudohyponatremia.
Hyponatremia Caused by Severe Hyperglycemia
The hyponatremic response to hyperglycemia is variable and determined by several factors,15,16 most of which are not known or measurable during the usual clinical circumstances. Therefore, we used the correction factor of Katz,17 whereby the serum sodium level decreases by 1.6 mmol/L for each 5.6-mmol/L (100-mg/dL) increase in glucose level.17 Use of this formulation in Table 2 predicts that elevation of the serum glucose level has induced a translocation of water from the intracellular fluid to the ECF sufficient to reduce the serum sodium level to approximately 120 mmol/L (a reduction from 140 mmol/L of [1.6][12.5]=20 mmol/L) if the baseline serum glucose level was 5.6 mmol/L (100 mg/dL). The Osmc will be (2)(120)+ 75+5=320 mmol/kg H2O. It should be noted that this is an elevated value (see Table 1, misconception No. 5). Because the measured osmolality includes the contribution of glucose, and glucose is also a variable in the equation to calculate serum osmolality, the Osmm increases in parallel with the Osmc; thus, the osmolal gap is 0. Nevertheless, the effective osmolality (320−5=315 mOsm/kg H2O) is abnormally increased. This situation contrasts with that in lipid-induced hyponatremia, and the hypertonicity rules out an increased risk of osmotic cerebral edema. (In fact, acute hyperglycemia can increase the risk of cerebral dehydration. Other outcomes are possible, however, depending on clinical circumstances.18)
Hyponatremia Caused by Retention of Mannitol
If it is assumed that mannitol (molecular weight, approximately 182) exerts the same effect on a normal serum sodium level of 140 mmol/L as does glucose, a serum mannitol concentration of 75 mmol/L (1365 mg/dL) might be expected to be associated with a serum sodium level of approximately 120 mmol/L.19-21 As in the case of hyperglycemia, we are neglecting the effect of any osmotic diuresis on the serum sodium concentration. The Osmc, which does not take mannitol into account, will be (2)(120)+5+5=250 mmol/kg H2O, an abnormally low value. (Note that this contrasts with the situation in hyperglycemia-induced hyponatremia, wherein Osmc is increased.) The Osmm will be (2)(120)+ 75+5+5=325 mmol/kg H2O; thus, the osmolal gap is 325−250=75 mmol/kg H2O, a considerably elevated level, again contrasting with the situation obtaining in hyperglycemia wherein the osmolal gap is normal. The effective osmolality is 325−5=320 mmol/kg H2O; this high value, indicative of hypertonicity, places the patient at increased risk for cerebral dehydration, rather than cerebral edema.22,23
True Hyponatremia Together With Elevated SUN Level
We have found that this situation is one of the least well understood conditions regarding the interrelationships between serum sodium, osmolality and osmolal gap, and tonicity. A patient with true hyponatremia (eg, a patient with the syndrome of inappropriate secretion of antidiuretic hormone whose serum sodium level is 120 mmol/L) might develop incidental acute renal failure. If the SUN level were to rise to 45 mmol/L (126 mg/dL), the Osmc would be (2)(120)+5+45=290 mmol/kg H2O, a normal value. The Osmm would also be 290 mmol/kg H2O because urea's contribution is not only measured in Osmm but also calculated in Osmc. Thus, the osmolal gap is 0 mmol/kg H2O. The important point, frequently not appreciated, is that finding a normal Osmm (or even an increased Osmm) does not preclude an increased risk of cerebral edema from hypotonicity. Because urea is not an effective solute (see "Definitions of Terms" section, above), the effective osmolality is 290−45=245 mmol/kg H2O, an abnormally low value. Thus, such a patient has the same increased risk of cerebral edema that he or she would have at the same serum sodium level if the SUN level were within the normal range.
True Hyponatremia Together With High Blood Alcohol Level
Instead of developing acute renal failure, suppose a patient with the syndrome of inappropriate secretion of antidiuretic hormone and with a serum sodium level of 120 mmol/L has ingested an excessive amount of ethanol. When ethanol (molecular weight, approximately 46) is present in serum, Osmm increases without a change in tonicity.10,12 If the serum ethanol concentration is 40 mmol/L (184 mg/dL) in this patient, the Osmc will be (2)(120)+ 5+5=250 mmol/kg H2O.12 (As is the case with urea, the presence of ethanol in the ECF does not by itself affect the serum sodium concentration, because ethanol is not an effective solute.) Neglecting the correction factor for serum water content,7,10 the Osmm will be (2)(120)+5+5+40=290 mmol/kg H2O. Thus, the osmolal gap will be 290−250=40 mmol/kg H2O, an abnormally elevated value.
This situation contrasts with that of the azotemic hyponatremic patient discussed above, because ethanol is not a variable in the equation used herein to calculate serum osmolality. The effective osmolality is 290−(40+5)=245 mmol/kg H2O, an abnormally low value, which is indicative of an increased risk of cerebral edema. Thus, in our examples, both the patient with a high SUN level and the patient with a high alcohol level are hypotonic, but only the ethanol abuser has an increased osmolal gap.
This example is provided for contrast. If the patient has a serum sodium concentration of 160 mmol/L, the Osmc is (2)(160)+5+5=330 mmol/kg H2O; the Osmm will be the same, and therefore, the osmolal gap is 0 mmol/kg H2O. The effective osmolality will be 330−5=325 mmol/kg H2O, which is abnormally increased. Thus, a hypernatremic patient is always both hyperosmolal and hypertonic22,23; there is no such thing as "pseudohypernatremia."
Accepted for publication June 12, 1998.
We thank Alan L. Bisno, MD, Gabriel N. Contreras, MD, Guido O. Perez, MD, and Richard A. Preston, MD, for their criticisms and suggestions, and Felice R. Dominguez for manuscript preparation.
Corresponding author: James R. Oster, MD, Medical Service (111), Veterans Affairs Medical Center, 1201 NW 16th St, Miami, FL 33125.
EF A simple and precise method of determining true sodium, potassium, and chloride concentrations in hyperlipemia. J Lab Clin Med.
1976;88683- 688Google Scholar
DD Misleading hyponatremia due to hyperlipemia: a method-dependent error. Ann Intern Med.
1981;95707- 708Google ScholarCrossref
JP The displacement of serum water by the lipids of hyperlipemic serum: a new method for the rapid determination of serum water. J Clin Invest.
1955;341483- 1488Google ScholarCrossref
OS Misleading reductions of serum sodium and chloride associated with hyperproteinemia in patients with multiple myeloma. Proc Soc Exp Biol Med.
1962;110145- 148Google ScholarCrossref
L Comparison of methods for calculating serum osmolality from chemical concentrations, and the prognostic value of such calculations. Clin Chem.
1975;21190- 194Google Scholar
JP Calculated vs measured plasma osmolalities revisited. Clin Chem.
1984;301703- 1705Google Scholar
G Are osmolality measurements clinically useful? Ann Clin Biochem.
1987;24566- 571Google ScholarCrossref
AG Osmometry in the evaluation of alcohol intoxication. Clin Toxicol.
1972;5343- 363Google ScholarCrossref
RP The missing gap: a pitfall in the diagnosis of alcohol intoxication by osmometry. Arch Intern Med.
1986;1461843- 1844Google ScholarCrossref
JA Effect of alcohols and selected solvents on serum osmolality measurements. J Toxicol Clin Toxicol.
1983;20115- 132Google ScholarCrossref
RL The variable hyponatremic response to hyperglycemia. West J Med.
1985;14249- 53Google Scholar
AH What determines the degree of hyponatremia in hyperglycemia? Clin Chem.
1988;34641- 642Google Scholar
MA Hyperglycemia-induced hyponatremia: calculation of expected serum sodium depression. N Engl J Med.
1973;289843- 844Google ScholarCrossref
CM Correlates of brain edema in uncontrolled IDDM. Diabetes.
1992;41627- 632Google ScholarCrossref
ML Hyponatremia after mannitol administration in the presence of renal failure. Anesth Analg.
1981;60165- 167Google ScholarCrossref
P Studies on the significance of decreased serum sodium levels following the clinical use of mannitol. J Urol.
1965;94470- 474Google Scholar
TD Hyperosmolality with hyponatremia, caused by inappropriate administration of mannitol. Am J Med.
1967;42648- 650Google ScholarCrossref