Clinical Assessment of Extracellular Fluid Volume in Hyponatremia

HSAIO-MIN CHUNG, M.D. Assessment of the status of extracellular fluid volume is important in
RUDIGER KLUGE, M.D. evaluating the cause and selecting appropriate therapy for hyponatre-
ROBERT W. SCHRIER, M.D. mic disorders. Since the sensitivity and specificity of clinical assess-
ROBERT J. ANDERSON, M.D. ment of extracellular fluid volume status in hyponatremic states remain
Denver, Coloracjo unknown, 58 non-edematous patients with serum sodium less than
130 meq/liter were prospectively evaluated. Patients were judged to
be either normovolemic (no response of serum sodium to saline
infusion) or hypovolemic (saline infuslon significantly corrected hypon-
atremia). Hypovolemlc patients had significantly higher plasma renin
activity (5.0 f 1.5 versus 2.5 f 0.5 rig/ml per three hours, p <O-05)
and norepinephrine (1,054 f 252 versus 519 f 55 pg/ml, p <0.05)
concentrations than did normovolemic patients. Clinical assessment
correctly identified only 47 percent of hypovolemic patients and 48
percent of normovolemic patients. Thus, clinical assessment was of
limited sensitivity and specificity in identlfylng extracellular fluid vol-
ume status in these hyponatremic patients. However, the concentra-
tion of sodium in a spot urine sample clearly separated hypovolemic
(mean Uua = 18.4 f 3.1 meq/liter) from normovolemlc (mean UNa =
72 f 3.7 meq/liter, p <O.OOl) hyponatremic patients.

Hyponatremia is a commonly encountered electrolyte disturbance seen

in diverse clinical settings [1,2]. In 1962, Leaf [3] first proposed that
clinical assessment of extracellular fluid volume status provides a useful
means of evaluating the cause and of selecting the appropriate therapy in
hyponatremic patients [3]. Subsequently, it has become common prac-
tice to classify hyponatremic patients on clinical grounds as having an
expanded, contracted, or normal extracellular fluid volume [1,4-61.
However, the sensitivity and specificity clinical assessment of extracellu-
lar fluid volume status in evaluating hyponatremic states remain un-
known. We prospectively evaluated 58 non-edematous patients with
hyponatremia. Our purpose was to ascertain the utility of clinical assess-
ment in determining if extracellular fluid volume expansion corrects
hyponatremia.

From the Department of Medicine, Veterans Ad- PATIENTS AND METHODS

ministration Medical Center, and Department of Methods of patient selection have been described in detail previously [ 11.
Medicine, University of Colorado Health Sci- Fifty-eight of 196 consecutively encountered hospitalized patients who
ences Center, Denver, Colorado. Requests for
gave informed consent and had a serum sodium concentration of less than
reprints should be addressed to Dr. Robert J.
Anderson, Department of Medicine, Veterans 130 meq/llter constituted the study population. These patlents met the
Administration Medical Center, 1055 Clermont following criteria: blood glucose level of less than 200 mg/dl, blood creatl-
Street, Denver, Colorado 80220. Manuscript nine level of less than 4.0 mg/dl, absence of edema and ascltes, absence
submitted March 27, 1987, and accepted July 1, of therapy for hyponatremla at the time of initial evaluation, receipt of at
1987. least 150 mM of isotonic sodium chloride per day for a mlnlmum of two

November 1987 The American Journal of Medlclne Volume 83 905

ASSESSMENT OF HYPONATREMIA-CHUNG ET AL

sessment as to whether the patient had a contracted or

Saline Saline
Responders Nonresponders
normal extracellular fluid volume state. We believed that
normal saline administration would correct the hyponatre-
mic state in patients with a contracted extracellular fluid
Or T volume state and would have little effect on plasma sodium
Systolic ,20 concentration in patients with a normal state. A saline
Blood responder was defined as a patient who had a sustained
increase in plasma sodium concentration of greater than 5
Pressure , ,.
meq/liter following saline administration.
(mm Hd Data were analyzed by two-tailed unpaired Student t
tests. A p value of less than 0.05 was considered signifi-
cant. All data are expressed as the mean f 1 SEM.

RESULTS
Diastolic
Of the 58 patients, 29 were believed on clinical grounds
Blood 7o to be extracellular fluid volume depleted and therefore
Pressure saline responsive. All patients predicted to be saline
(mm Hg) 6o
responders met at least two of the following six criteria:
(1) a history or clinical setting compatible with extracellu-
lar fluid volume loss; (2) a decrease in skin turgor or
axillaty moisture, dry mucous membranes, or presence of
a subjective sensation of thirst; (3) a greater than O&kg
Pulse 110 weight loss; (4) at least a 10 percent decrease in ottho-
static (supine to two-minute upright) systolic blood pres-
Rate sure; (5) at least a IO percent increase in otthostetic
(beats/min) 9. pulse rate; and (6) a blood urea nitrogen to creatinine ratio
SupineUpright SupineUpright greater than 20. A saline response was observed in only
seven of 29 (24 percent) patients predicted to be saline
Figure 1. Effect of two minutes upright posture on systok responsive. The remaining 29 patients were believed to
ic (upper panel) and diastolic (middle panel) blood pres- have a normal or slightly expanded extracellular fluid
sures and pulse rates (lower panel). Saline responders are
denoted by the open bars, saline nonresponders by
hatched bars.
Saline Saline
Responders Nonresponders

6.0 T

II
consecutive days, and measurement of serial plasma sodi-
um concentrations following saline therapy. More than 75 Plasma
percent of the 58 patients received greater than 300 meq of P<.O5
0.9 percent sodium chloride. At initial evaluation, an exten- Renin
sive chart review with special attention to recent medica- Activity
tions, clinical events, body weights, intake and output, (ng/ml/hr)
chest radiographs, laboratory data, and serial vital signs and
other available hemodynamic measurements was made.
Each patient underwent a complete physical examination
by one of us. This examination included special attention to
cardiac parameters, jugular venous pressure, orthostatic
changes in pulse and blood pressure, skin turgor, moisture
in the axillae, and hydration of mucous membranes. Plasma 1000
At the initial evaluation of patients, spot urine samples Norepinephrine
P<.O5
and lo-minute supine blood samples for measuring plasma
(pcdml)
renin activity and norepinephrine concentrations were ob- 500
tained [I]. Plasma sodium and creatinine concentrations,
blood urea nitrogen, plasma renin activity, and urinary sodi-
um concentration were measured as described previously
[Il. Figure 2. Supine plasma renin activity (upper panel) and
After chart review and physical examination, two of us norepinephrine concentrations (lower panel) in saline re-
thoroughly discussed the available data and made an as- sponders and nonresponders.

906 November 1967 The American Journal of Medicine Volume 83

 ASSESSMENT OF HYPONATREMIA-CHUNG ET AL

volume status and were predicted to be saline nonrespon-

140
sive. Twenty-one (72 percent) of these patients were e
saline unresponsive. Thus, clinical assessment was more
8
accurate in predicting no saline response than in predict- 120
1
ing saline response. The hyponatremia of 15 of the 58
hyponatremic patients was correctable by saline adminis-
tration. Seven of these 15 (47 percent) were predicted to
be saline responsive. Conversely, 43 patients did not
loo-
i
respond to saline. Twenty-one of these patients (48 per-
!
80 -
cent) were predicted to be saline unresponsive.
In an effort to determine why clinical assessment was Spot UNa tt
of limited utility in determining response to saline, we 60 - t
compared supine and upright pulses and blood pressures
(mEq/L) 0
in saline responsive and nonresponsive patients (Figure
l
1). There were no significant differences in either supine 40 - tl
or upright pulse and blood pressure when saline respon-
8
sive and nonresponsive patients were compared. By con-
20 -
trast, plasma renin activity and blood concentrations of :
1
norepinephrine were significantly higher in saline re-
sponders than in nonresponders (Figure 2). Blood nitro- I
O-
gen (23 f 4 mg/dl in responders and 13 f 3 mg/dl in Saline Saline
nonresponders, p <0.05) and serum creatinine (1.24 f Responders Nonresponders
0.25 mg/dl in responders and 0.8 f 0.10 mg/dl in
nonresponders, p <O-05) were also different in saline Figure 3. Spot urinary sodium concentrations in saline
responders and nonresponders. All individual values as well
responders and nonresponders. In addition, spot urinary
as the mean f 1 SEM are denoted.
sodium concentrations were higher in saline nonre-
sponders (72.7 f 3.7 meq/liter) than in saline respond-
ers (18.4 f 3.1 meq/liter, p <O.OOl). A spot urinary
sodium concentration of less than 30 meq/liter was 100 siveness. However, the cost and delay involved in obtain-
percent specific and 80 percent sensitive for predicting ing the concentrations of these plasma hormones make
saline response (Figure 3). this approach primarily of research interest. The mean
blood urea nitrogen and serum creatinine concentrations
COMMENTS differentiated the saline-responsive and saline-nonre-
The current results indicate that clinical assessment does sponsive patients. This suggests that the hyponatremia of
not predict serum sodium response to a saline challenge saline-responsive patients is mediated by hypovolemic-
in non-edematous, hyponatremic patients. Thus, clinical induced factors such as non-osmotic release of arginine
evaluation was associated with 50 percent false-negative vasopressin and intrarenal factors.
and 50 percent false-positive results in incriminating ex- In contrast to clinical assessment, the kidney was very
tracellular fluid volume depletion as a primary factor in sensitive in determining saline responsiveness. The 43
mediating hyponatremia in non-edematous patients. It is saline nonresponsive patients had urinary sodium concen-
noteworthy that only 50 percent of patients whose hypon- trations greater than 30 meq/liter (mean: 72 f 4 meq/
atremia responded to saline were identified by a history liter). On the other hand, 12 of 15 saline-responsive
and physical examination and chart review by two nephro- patients exhibited urinary sodium concentrations less than
logists. This may have been due, at least in part, to water 30 meq/liter. It should be noted, however, that patients
retention resulting in modest but incomplete correction of with volume depletion due to renal causes of salt wasting
hypovolemia. Such water retention could obscure any (i.e., primary adrenal insufficiency and diuretic use) may
clinical evidence of extracellular fluid volume depletion have saline-responsive hyponatremia with relatively high
[7,8]. Alternatively, it is well known that the hemodynamic urinary sodium concentrations.
response to extracellular fluid volume depletion is depen- In summary, the current results indicate that hypovole-
dent on the rate, magnitude, and source of extracellular mic hyponatremic states, as documented by elevated
fluid volume loss [9-l 11. It is possible that only mild plasma renin and norepinephrine concentrations and re-
degrees of extracellular fluid volume depletion of chronic sponse of hyponatremia to saline administration, are quite
duration occurred in our saline response patients [IO, 1 I]. difficult to detect by history and physical examination. A
Plasma renin activity and plasma norepinephrine con- urinary sodium concentration below 30 meq/liter, how-
centrations showed a good correlation with saline respon- ever, suggests a saline-responsive hyponatremic patient.

November 1987 The American Journal of Medicine Volume 83 907

ASSESSMENT OF HYPONATREMIA-CHUNG ET AL

REFERENCES
1. Anderson RJ, Chung HM, Kluge R, Schrier RW: Hyponatre- Schrier RW, ed. Renal and electrolyte disorders, 3rd ed.
mia: a prospective analysis of its epidemiology and the Boston: Little, Brown, 1986; l-77.
pathogenetic role of vasopressin. Ann Intern Med 1985; 7. Gross PA, Anderson RJ: Effects of DDAVP and AVP on
102: 165-168. sodium and water balance in conscious rat. Am J Physiol
2. Baran D, Hutchinson TA: The outcome of hyponatremia in a 1982; 243: R512-R519.
general hospital population. Clin Nephrol 1984; 22: 72- 8. Gross PA, Travis V, Horwitz L, Schrier RW, Anderson RJ:
76. Effects of DDAVP-induced water retention on arterial
3. Leaf A: The clinical and physiologic significance of the pressure and systemic hemodynamics in the rat. Am J
serum sodium concentration. N Engl J Med 1962; 267: Physiol 1982; 243: H934-H940.
24-30,77-83. 9. Elkington JR, Danowski TS, Winkler AW: Hemodynamic
4. Humes HD, Narins RG, Brenner BM: Disorders of Water changes in salt depletion and dehydration. J Clin Invest
balance. Hosp Pratt [Off] 1979; 14: 133-145. 1946; 25: 120-129.
5. Levi M, Berl T: Water metabolism. In: Gonick HC, ed. Current 10. McCance RA: Experimental sodium chloride deficiency in
nephrology, vol 5. New York: John Wiley 8 Sons, 1982; man. Proc R Sot Lond (Biol) 1936; 119: 245-268.
l-86. 11. Shapiro JS, Anderson RJ: Sodium-depletion states. Con-
6. Schrier RW, Berl T: Disorders of water metabolism. In: temp Nephrol 1987; 16: 245-276.

908 November 1987 The American Journal of Medicine Volume 83