Central Pontine Myelinolysis and Pontine Lesions After Rapid Correction of Hyponatremia: A Prospective Magnetic Resonance Imaging Study John E. Brunner, MD,# Janice M. Redmond, MB,? Allan M. Haggar, MD,$ Davida F. Kruger, MSN,” and Stanton B. Elias, MD1-
The rate at which profound hyponatremia should be corrected is the focus of a recent clinical debate. We prospectively studied neurological outcomes with serial magnetic resonance imaging in 13 hyponatremic subjects with serum sodium concentrations of less than 115 mmol/L (mean initial serum sodium concentration, 103.7;range, 93-113 mmoy L). All subjects were corrected to mildly hyponatremic levels at 24 hours and ultimately underwent an increase in serum sodium concentration of 2 5 mmol/L without development of hypernatremia. Magnetic resonance imaging revealed the development of pontine lesions in 3 patients. The correction rate of hyponatremia over the first 24 hours was significantly faster in patients with pontine lesions (mean 2 SD, 1.25 f 0.4 mmol/(L . hr) versus 0.74 & 0.3 mmoY (L hi); p < 0.05). Initial sodium concentration was also significantly lower in the pontine lesion group (97.3 2 6.7 vs 105.6 & 5.2 mmoUL, p < 0.05). We conclude that the correction rate of hyponatremia plays a significant role in the pathogenesis of pontine lesions in individuals with profound hyponatremia who undergo large increases in sodium concentration as a result of severe initial hyponatremia.
-
Brunner JE, Redmond JM, Haggar AM, Kiuger DF, Elias SB. Central pontine myelinolysis and pontine lesions after rapid correction of hyponatremia: a prospective magnetic resonance imaging study. Ann Neurol 1990;27:61-66
Patients with hyponatremia may present with a spectrum of acute central nervous system dysfunction, including stupor, seizures, or coma. Prompt recognition and therapy usually result in the complete resolution of these disturbances. If either recognition or therapy is delayed, permanent brain damage due to hyponatremia itself can result El-31. For this reason some investigators have advocated rapid correction of hyponatremia at 1.5 to 2.0 mmol/(L * hr) 14, 5). Others recommend that correction of hyponatremia be undertaken slowly, at a rate of less than 0.7 m o l l (L * hr) to avoid the cerebral demyelinating disorder central pontine myelinolysis (CPM) I6- lo]. Development of this serious complication has been associated with rapid correction of hyponatremia and overcorrection to hypernatremia in animal models and patients studied retrospectively [G- 11). Others have suggested that the rate of correction is not as important in the development of CPM as an absolute increase in sodium concentration of 25 mmovL 111, 12). Patients with a sodium concentration of less than
115 mmoyl. are very likely to undergo a 25-mmolfl. change in sodium concentration, with the inevitable normalization of their serum sodium concentration. The question that remains unanswered is how fast this deficit should be reversed. This investigation was designed to determine whether the development of pontine lesions is associated with the correction rate of hyponatremia in patients with a serum sodium concentration of less than 115 mmol/L. Our study utilized serial magnetic resonance imaging (MRI) and neurological examinations to detect central nervous system abnormalities in patients after correction of profound hyponatremia.
From the Departments of *Internal Medicine, ?Neurology, and $Diagnostic Radiology, Henry Ford Hospital, Detroit, MI. Received for publication Feb 20, 1989, and in revised form Jun 5. Accepted for publication June 6, 1989.
Address correspondence to D r Brunner, Mercy Endocrine and Diabetes Care Center, 2238 Jefferson Ave, Toledo, OH 43624.
Methods Study Protocol Patients with serum sodium concentrations of less than 115 mmol/L. were identified through emergency room referral and the central laboratory computer. With one exception, the evaluation, treatment, and correction rate of hyponatremia were established by attending physicians who were
Copyright 0 1990 by the American Neurological Association
61
not participants in the study. Therapy of hyponatremia commenced on discovery in all cases. Most of the patients received multiple modes of therapy in an attempt to correct hyponatremia. The neurological condition was assessed by standard neurological and mental status examinations while the patients were hospitalized. An initial MRI scan was performed within the first week of therapy; this was followed by a second scan a week later and a third scan at approximately 1 month. Precise duplication of the timing of MRI was not possible for all cases. All subjects had serial MRI performed within 5 weeks of admission, with one exception in which serial high-resolution computed tomography (CT) was used until MRI was available at 12 weeks. T I - and T2-weighted spin echo images were obtained utilizing a super-conducting magnet operating at 1.5 Tesla. Hgh-resolution CT was performed in patients in whom lesions were identified by MRI. Protocol approval was obtained from the Henry Ford Hospital Human Research Committee.
Measurement of Serum Soditlm Concentration, Calculation of Changes, and Statistical Analysis Serum sodium concentrations were measured in the hospital chemistry laboratory with an ion-specific electrode 24 and 48 hours after the initial identification of hyponatremia Before 24 hours, sodium concentrations were measured at intervals determined by the clinical staff. Intervals of sodium measurements varied from once in 24 hours to hourly. The correction rate of hyponatremia was calculated by subtracting the initial low sodium concentration from the sodium concentration at 24 and 48 hours later, divided by 24 and 48 (hours), and expressed as mmol/(L . hr). Subjects with pontine lesions evident on MRI were compared to those without lesions. Subjects who had persistent alterations in m e n d status 72 hours or more after hyponatremia correction (including those subjects with pontine lesions) were compared with those who had normal mental status after hyponatremia correction. Continuous variables were analyzed by a two-sided Student’s t test. Categorical variables (i.e., alcoholism, sex, and the presence or absence of central nervous system dysfunction at discovery of hyponatremia) were analyzed by Fisher’s Exact test. All means are expressed as 2 one standard deviation.
Results Study Subjects Thirteen subjects (10 female, 3 male) were studied (Table 1). Hyponatremia was the admitting diagnosis in 10 subjects and developed in 3 patients while they were hospitalized. At the presentation of hyponatremia, 8 patients had new or increased central nervous system dysfunction, ranging from altered mental status (6 subjects) to coma (2 subjects). Hyponatremia was attributed tcl Addison’s disease and hypotonic fluid administration in 1 subject. The syndrome of inappropriate antidiuretic hormone secretion was demonstrated in 6 subjects: 2 with polydipsia, 2 without polydipsia, and 2 postoperatively. Diuretic administration was implicated in 4 subjects, 3 of whom were polydipsic. Beer potomania and bladder
62 Annals of Neurology
Vol 27
N o 1 January 1990
Table 1. Charactmitics of 13 Study Subjects Variable
Mean
SD
Age (yr) Initial “a} (mmofi) [Na] after 24 hr (mmoVL) [Na] after 48 hr (mmofi) Correction rate at 24 hr (mmol/[L . hr]) Correction rate at 48 hr (mmol/CL . hr]) Time to reach “a], 135 mmoyL (days) Maximum [Na] (mmol/L) Time to reach maximum CNa7 (days)
64.0 103.7 124.1 128.7 0.86
14.2 37-83 93-113 6.3 113-134 5.9 119-140 6.0 0.39 0.21-1.67
0.56
0.21 0.33-0.97
5.3
4.1
2-16
136.1 4.46
3.8 2.2
130-143 2-9
Range
SD = standard deviation.
irrigation with glycine during a transurethral resection of the prostate were identified as the cause of hyponatremia in the remaining 2 patients. One subject admitted to current alcohol abuse and 1 had a history of remote alcohol abuse. Neither had evidence of hepatic insufficiency. All patients eventually underwent a change of 25 mmol/L in sodium concentration at variable intervals. All subjects were mildly hyponatremic after 24 hours of treatment and none developed hypernatremia.
Patients with Pontine Lesions Three of 13 subjects were shown to have pontine lesions by MRI after correction of hyponatremia. None of these lesions were demonstrated by CT. PATIENT 1. A 65-year-old man with essential hypertension and a history of remote alcohol abuse was admitted for an elective transurethral resection of the prostate, in which a glycine solution (200 mosdkg) was used for irrigation. Preoperatively, the sodium concentration was 133 mmofi. Postoperatively, the patient failed to awaken and required mechanical ventilation, but no respiratory arrest occurred. The serum sodium concentration was 93 mmol/L. Intravenous furosemide and normal saline were started immediately and were subsequently changed to fluid restriction. After 8 hours of therapy, at a serum sodium concentration of 118 mmoYL, 2 brief generalized seizures occurred. At 24 hours serum sodium was 133 mmol/L. The patient was able to follow simple commands and move all extremities to painful stimuli; extraocular muscle movements were full and gag reflex was normal. Several days later he was extubated, and at that time he was found to be disoriented. Formal neuropsychological evaluation revealed impaired memory, problem-solving, and psychomotor skills. Electroencephalography revealed diffuse slow-wave activity consistent with a moderate cerebral disturbance. Auditory evoked potentials showed no pontine dysfunction. Somatosensory evoked potentials in the upper
A
B
D Patient I. T1 spin echo magnetic resonance images on days 4 (A), 7 (B), and 18 (C) and at 4 months (0).An evolving lowsignal abnomlity is seen within the pons that is most evident on dzy 18 and is barely detectable at 4 months.
extremities demonstrated only a mild peripheral neuropathy. At discharge from the hospital, 13 days after admission, he was alert but had an unsteady gait. Two months later a second neuropsychological evaluation demonstrated dramatic improvement. MRI studies were obtained after 4, 7, and 18 days and after 4 months (Figure). They demonstrated an evolving pontine lesion that was most evident on day 18 and barely detectable at 4 months.
2. A 37-year-old woman was admitted with a serum sodium concentration of 94 mmoVL. This case has been reported previously [ 131.) Sodium concentrations were
106 mmolfl. at 6 hours, 123 mmoYL at 24 hours, and 129 mmoVL at 48 hours. After correction the patient’s neurological status on examination appeared to improve, bur 3 days later she became disoriented and subsequently comatose with decorticate posturing. There was no specific evidence of a pontine lesion. Her neurological status gradually improved. After 30 months of follow-up, she had persistent mild impairment of cortical capacities that continued to prevent her from returning to work as a schoolteacher. Serial CT scans of the pons and brain at 1,2, and 12 weeks were normal, but an MRI obtained at 12 weeks demonstrated a large pontine defect consistent with CPM f13). MRI studies at 12 and 30 months demonstrate marked improvement in the lesions.
PATIENT
PATIENT 3. A 78-year-old woman was admitted with a complaint of urinary retention. She had been following a
Brunner et al: Pontine Lesions and Rapid Correction of Hyponatremia
63
Table 2. Subjects With Versus Subjects Without Pontine hsions 95% Confidence Limits Yes (n = 3) Variable ~~~~~~~
4 w (yr) Initial {Na] (mmovL) {Na] after 24 hr (mmol/L) [Na] after 48 hr (mmol/L) Correction rate at 24 hr (mmoV[L . hr]) Correction rate at 48 hr (mmoVEL . hr]) Time to change {Na] 25 m m o f i (hr) Maximum [Na] (mmofi) Time to reach maximum [Nal (days)
No (n = 10)
With Pontine Lesions
Mean
SD
Mew
SD
PValue"
60.0 97.3 127.3 131.3 1.25 0.7
21.0 6.7 5.1 7.8 0.4 0.20
65.0 105.6 123.10 127.9 0.74 0.50
13.0 5.2 6.10 5.60 0.32 0.20
0.62 0.04 0.31 0.4 1 0.04 0.16
- 26.4 - 16.3 - 4.4 - 5.4
25.7 136.0 6.3
20.1 5.6 3.8
55.3 136.2 3.9
28.5 3.60 1.4
0.13 0.94 0.10
- 69.0
Without Pontine Lesions
~
0.023
- 0.09 -6.1 - 0.6
16.4
- 0.3 12.8 12.2 0.997 0.49 9.8 5.6 5.4
"Two-tailed Student's t test.
self-prescribed weight loss regimen that consisted of no food and large amounts of water. Her medications included a combination diuretic (spironolactone, 25 mg, and hydrochlorothiazide, 25 mg). The initial serum sodium concentration was 105 mmovL. Results of initial general physical, neurological, and mental status examinations were normal. She was treated with intravenous 514 mmofi saline followed by 154 mmol/L saline and then fluid restriction. The sodium concentrations were: 109 mmofi at 3 hours, 121 mmofi at 5 hours, 126 mmovL at 24 hours, and 129 mmoV L after 48 hours. During therapy no seizures or hypoxic episodes were observed. Thirty-six hours after admission the patient was confused and had inappropriate behavior. At 48 hours she had progressed to flaccid quadriplegia with a right gaze preference. The clinical diagnosis of CPM was made and 4 mg of dexamethasone was administered intravenously every 6 hours for a week. After 72 hours she began to regain use of all her limbs and the eye movement abnormalities resolved. Electroencephalography demonstrated diffuse slowing without focal or lateralizing features. MRI was performed on days 4, 10, and 3 1 after admission. On the first scan the pons was normal. In the cerebral hemispheres scattered, small, nonsymmetrical white matter lesions were demonstrated, which remained unchanged throughout the study. These lesions in the cerebral hemispheres are consistent with T 2 enhancing lesions, which are frequently seen in elderly patients but have limited pathological significance. Subsequent scans demonstrated a transient, small hyperintense lesion within the pons with a surrounding area of decreased signal on TI images. The third scan demonstrated a small rightsided lesion of the pons near the midline and anterior to the fourth ventricle on T2-weighted imaging.
Patients Without Pontine Lesions In 10 subjects MRI did not demonstrate pontine lesions. Four of these 10 had persistent mental status alterations that lasted at least 3 days in 4 subjects, 1 of whom had memory deficits that persisted during 18 months of follow-up study. Two additional subjects
64 Annals of Neurology Vol 27 N o 1 January 1990
had prior neurological dysfunction that worsened at discovery of hyponatremia: 1 had a history of chronic dementia and the other had an old left parietal lobe infarct. After correction of hyponatremia these patients returned to baseline and neither had pontine abnormalities on MRI. Four patients had both normal neurological findings, mental status, and MRI of the brain after correction of hyponatremia.
Statistical AnaLysis Among the 3 patients with pontine lesions, the initial sodium concentration was significantly lower ( p = 0.04) and the correction rate of sodium at 24 hours significantly more rapid ( p = 0.04) than in patients without pontine abnormalities (Table 2). Before 24 hours, the mean correction rate of sodium for the 3 subjects in whom pontine lesions developed was 2.8 2 0.72 mmol/(L * hr) at an average of 5.7 hours. Among the 10 subjects without pontine lesions, 8 had sodium measurements between 5 and 11 hours of treatment. We calculate an early correction rate for this group at 1.40 & 0.80 mmol/(L . hr) at an average interval of 7.6 hours. No differences were detected in sex, presence or absence of central nervous system dysfunction at discovery of hyponatremia, or alcohol abuse. In the 7 subjects with central nervous system dysfunction that persisted for 72 hours or more despite correction, including those with and those without pontine lesions, an association was also seen with a lower initial sodium concentration and more rapid correction of hyponatremia at both 24 and 48 hours (Table 3). No discernible statistically significant differences were found in age, absolute sodium concentration at 24 and 48 hours, highest sodium concentration achieved, time to undergo a 2hnmolfl. change in sodium concentration, or time to reach highest sodium concentration. Likewise, for categorical variables no
Table 3. Subjects With Versus Subjects Without Persistent Central Nervous System Dysfunction Ajier Hyponatremia
95% Confidence Limits
Yes (n Variable
Mean
SD
Mean
SD
p Valueb
With CNS Dysfunction
Age (yr)
60.1 99.3 126.0 128.3 1.1
15.1 4.8 4.3 6.6 0.3
61.8 108.0 122.0 128.8 0.6
10.3 3.3 6.2 0.3
0.86 0.01 0.33 0.91 0.02
-21.1 - 14.9 - 4.8 - 10.0 0.07
0.7
0.2
0.4
0.1
0.04
0.05
34.6
30.0
60.0
17.4
0.21
- 62.9
12.1
135.4 5.3
3.4 2.7
134.5 3.3
2.6 1.3
0.65 0.19
- 3.6
- 1.3
5.4 5.3
Initial [Na) (mmol/L) “a) after 24 hr (mmol/L) [Na) after 48 hr (mmoVL) Correction rate at 24 hr (mmol/ EL . hrl) Correction rate at 48 hr (mmoV EL . hrl) Time to change “a), 25 mmoYL (hr) Maximum “a) (mmol/L) Time to reach maximum {Na)
= 7)”
No (n
= 4)
8.8
Without CNS Dysfunction 17.7 - 2.5
12.8 8.4 0.93 0.55
(days) “Two patients with known CNS dysfunction before hyponatremia discovery are not used in analysis. bTwo-tailed Student’s t test.
significant differences were seen in sex, presence or absence of central nervous system dysfunction at discovery of hyponatremia, or alcohol abuse. Discussion We have described 13 subjects with profound hyponatremia in whom a rise in sodium concentration of 25 mmol/L or more occurred. All subjects were mildly hyponatremic at 24 hours and no one had hypernatremia. Pontine lesions developed in 3 patients after the correction of hyponatremia. The correction rate of hyponatremia at 24 hours was significantly more rapid in those patients who had pontine lesions than in those who did not. Correction rates of hyponatremia at intervals before 24 hours were faster than the rates calculated at 24 hours; this reflects efforts to increase the sodium concentration by the treating physicians. In patients without pontine lesions, the rate was 1.45 mmol/(L . hr) at 7.3 hours. In those in whom pontine lesions developed, the rate before 24 hours was 2.8 mmol/(L hr) at 5.7 hours. A statistical comparison of the correction rates before 24 hours is not provided because serum sodium concentrations were not obtained at the same interval. Two experiences from this study indicate that early aggressive therapy may expose patients to the risk of CPM. In Patient 3 a brief period of rapid correction (12 mmoVL over 2 hours) was associated with clinical and radiographic evidence of a pontine lesion despite a total change in sodium concentration that was limited to 21 mmoVL at 24 hours and 24 mmoVL at 48 hours. In Patient 1, the treating physician’s intent was to rapidly correct one-half of the sodium concentration deficit. Slow correction by fluid restriction was to fol-
low. However, during fluid restriction the sodium concentration spontaneously and rapidly increased, presumably because of the patient’s own homeostatic mechanisms. The initial severity of hyponatremia was also associated with the development of pontine lesions. Severe hyponatremia has not been associated with demyelinating lesions in animals with uncorrected hyponatremia that has been sustained for as long as 10 days [7, 8 , 147. However, large increases in sodium concentration have been associated with demyelinating lesions in humans and animal models E6-8, l l , 123. Severely hyponatremic individuals (initial sodium concentration less than 110 mmollL) must inevitably undergo a large rise in sodium concentration. We believe this presents an added risk for development of pontine lesions if the hyponatremia is reversed too rapidly. The fact that CPM is not limited to patients with the greatest severity of hyponatremia does not negate the significance of this observation, as other factors, such as hepatic insufficiency and overcorrection to hypernatremia, may be additional risk factors for CPM 1 1 1 , 12, 157. Prospective animal models demonstrate that rapid correction of hyponatremia leads to more extensive demyelinating lesions in chronically hyponatremic animals than in animals with acute hyponatremia of similar severity that is corrected at an identical rate [14, 167. Ten of the 13 subjects in this study developed hyponatremia as outpatients and the duration of the hyponatremia cannot be determined. However, Patient l documents an episode of acute, unsustained, and symptomatic hyponatremia that was rapidly corrected ( 3 mmoV[L. hr} at 8 hours, 1.67 mmol/[L hr} at 24 hours) in which CPM developed.
-
Brunner et al: Pontine Lesions and Rapid Correction of Hyponatremia 65
Ayus and associates Ell] have noted that seizures and hypoxia occur frequently in patients in whom CPM subsequently develops after hyponatremia. They suggest that many clinical and pathological findings resemble delayed anoxic encephalopathy. However, the lesions of delayed anoxic encephalopathy occur predominantly in the white matter of the cerebral hemispheres and are associated with profound anoxia and hypotension 117-191. There have been no reports of delayed anoxic encephalopathy with pontine lesions resembling CPM. As in other reports, seizures occurred in this study after significant increases in sodium concentration had occurred {9}. The occurrence of seizures could alternatively be interpreted as sequellae to very rapid correction rates during the initial hours of therapy. In this study we prospectively used serial MRI in determining the presence or absence of pontine lesions after correction of hyponatremia. This is the likely explanation for the higher rate of pontine lesions compared with other studies C4, 20-231. Our experience utilizing MRI demonstrated that pontine lesions can occur without demonstrable clinical findings of pontine dysfunction, that pontine lesions may be delayed several days after correction of hyponatremia, that resolution or improvement of lesions may also occur, and that MRI is more useful than highresolution CT in the detection of pontine lesions associated with hyponatremia correction. The data from this study cannot absolutely resolve the question concerning the appropriate correction rate of hyponatremia because the sample size is small and its design is not randomized. However, our data support the following conclusions: (1) Rapid correction of hyponatremia is associated with clinical neurological dysfunction and radiographic evidence of pontine lesions. ( 2 ) Patients with hyponatremia less than 105 m m o n are at greater risk to develop the complications associated with rapid correction of hyponatremia than are patients with higher sodium concentrations. (3) Patients with acute hyponatremia are also at risk for complications associated with rapid correction of hyponatremia.
Financial support for this study was provided by the Fund for Henry Ford Hospital.
References 1. Arieff AI, Llach F, Shad GM. Neurological manifestations and morbidity of hyponatremia: correlation with brain water and electrolytes. Medicine 1976;55:12 1-129 2. Ashraf N, Locksley R, Arieff AI. Thiazide-induced hypona-
66 Annals of Neurology Vol 27 No 1 January 1990
tremia associated with death or neurologic damage in outpatients. Am J Med 1981;70:1163-1168 3. Arieff AI. Hyponatremia, convulsions, respiratory arrest, and permanent brain damage after elective surgery in healthy women. N Engl J Med 1986;314:1529-1535 4. Ayus JC, Oliver0 JJ, Frommer JP. Rapid correction of severe hyponatremia with intravenous hypertonic saline solution. Am J Med 1982;72:43-48 5. Ayus JC, KrothapalLi RK, Arieff AI. Changing concepts in treatment of severe symptomatic hyponatremia Am J Med 1985;78:897-902 6. Norenberg MD, Leslie KO, Robertson AS. Association between rise in serum sodium and central pontine myelinolysis. Ann Neurol 1982;11:128-139 7. Kleinschmidt-DeMasters BK, Norenberg MD. Rapid correction of hyponatremia causes demyelination: relation to central pontine myelinolysis. Science 1981;211:1068-1070 8. Laureno R.Central pontine myelinolysis following rapid correction of hyponatremia. Ann Neurol 1983;13:232-242 9. Stems RH,Riggs JE, Schochet SS. Osmotic demyelination syndrome following correction of hyponatremia. N Engl J Med 1986;3 14:1535-1542 10. Sterns RH. Severe symptomatic hyponatremia: treatment and outcome: a study of 64 cases. Ann Intern Med 1987;107:656664 11. Ayus JC, Krothapalli RK, Armstrong DL. Rapid correction of severe hyponatremia in the rat: histopathologicalchanges in the brain. Am J Physiol 1985;248:F711-F719 12. Ayus JC, Krothapalli RK, Arieff AI. Treatment of symptomatic hyponatremia and its relation to brain damage: a prospective study. N Engl J Med 1987;317:1190-1195 13. Brunner JE, Redmond JM, Haggar AM, Elias SB. Central pont h e myelinolysis after rapid correction of hyponatremia: a magnetic resonance imaging study. Ann Neurol 1988;23:389-391 14. Illowsky BP, Laureno R Encephalopathy and myelinolysis after rapid correction of hyponatremia Brain 1987;110:855-867 15. Adams RD, Victor M, Mancall EL. Central pontine myelinolysis: a hitherto undescribed disease occurring in alcoholic and malnourished patients. Arch Neurol Psychiatry 1959;81:154172 16. Norenberg MD, Papendick RE. Chronicity of hyponatremia as a factor in experimental myelinolysis. Ann Neurol 1984;15: 544-547 17. Ginsberg MD. Delayed neurological deterioration following hypoxia. Adv Neurol 1979;262 1-44 18. Plum F, Posner JB, Hain RF. Delayed neurological deterioration after anoxia. Arch Intern Med 1962;11018-25 19. Jennett B, Plum F. Persistent vegetative state after brain damage: a syndrome in search of a name. Lancet 1972;1:734-737 20. Decaux G, Unger J, Brimioulle S, Mockel J. Hyponatremia in the syndrome of inappropriate secretion of antidiuretic hormone: rapid correction with urea, sodium chloride, and water restriction therapy. JAMA 1982;247:471-474 21. Ashouri 0s. Severe diuretic-induced hyponatremia in the elderly: a series of eight patients. Arch Intern Med 1986;146: 1355-1357 22. Hantman D , Rossier B, Zohlman R,Schrier R. Rapid correction of hyponatremia in the syndrome of inappropriate secretion of antidiuretic hormone and alternative treatment of hypertonic saline. Ann Intern Med 1973;78:870-875 23. Worthley LIG, Thomas PD. Treatment of hyponatremic seizures with intravenous 29.2% saline. Br Med J 1986;292:168170
The rate at which profound hyponatremia should be corrected is the focus of a recent clinical debate. We prospectively studied neurological outcomes with serial magnetic resonance imaging in 13 hyponatremic subjects with serum sodium concentrations of less than 115 mmol/L (mean initial serum sodium concentration, 103.7;range, 93-113 mmoy L). All subjects were corrected to mildly hyponatremic levels at 24 hours and ultimately underwent an increase in serum sodium concentration of 2 5 mmol/L without development of hypernatremia. Magnetic resonance imaging revealed the development of pontine lesions in 3 patients. The correction rate of hyponatremia over the first 24 hours was significantly faster in patients with pontine lesions (mean 2 SD, 1.25 f 0.4 mmol/(L . hr) versus 0.74 & 0.3 mmoY (L hi); p < 0.05). Initial sodium concentration was also significantly lower in the pontine lesion group (97.3 2 6.7 vs 105.6 & 5.2 mmoUL, p < 0.05). We conclude that the correction rate of hyponatremia plays a significant role in the pathogenesis of pontine lesions in individuals with profound hyponatremia who undergo large increases in sodium concentration as a result of severe initial hyponatremia.
-
Brunner JE, Redmond JM, Haggar AM, Kiuger DF, Elias SB. Central pontine myelinolysis and pontine lesions after rapid correction of hyponatremia: a prospective magnetic resonance imaging study. Ann Neurol 1990;27:61-66
Patients with hyponatremia may present with a spectrum of acute central nervous system dysfunction, including stupor, seizures, or coma. Prompt recognition and therapy usually result in the complete resolution of these disturbances. If either recognition or therapy is delayed, permanent brain damage due to hyponatremia itself can result El-31. For this reason some investigators have advocated rapid correction of hyponatremia at 1.5 to 2.0 mmol/(L * hr) 14, 5). Others recommend that correction of hyponatremia be undertaken slowly, at a rate of less than 0.7 m o l l (L * hr) to avoid the cerebral demyelinating disorder central pontine myelinolysis (CPM) I6- lo]. Development of this serious complication has been associated with rapid correction of hyponatremia and overcorrection to hypernatremia in animal models and patients studied retrospectively [G- 11). Others have suggested that the rate of correction is not as important in the development of CPM as an absolute increase in sodium concentration of 25 mmovL 111, 12). Patients with a sodium concentration of less than
115 mmoyl. are very likely to undergo a 25-mmolfl. change in sodium concentration, with the inevitable normalization of their serum sodium concentration. The question that remains unanswered is how fast this deficit should be reversed. This investigation was designed to determine whether the development of pontine lesions is associated with the correction rate of hyponatremia in patients with a serum sodium concentration of less than 115 mmol/L. Our study utilized serial magnetic resonance imaging (MRI) and neurological examinations to detect central nervous system abnormalities in patients after correction of profound hyponatremia.
From the Departments of *Internal Medicine, ?Neurology, and $Diagnostic Radiology, Henry Ford Hospital, Detroit, MI. Received for publication Feb 20, 1989, and in revised form Jun 5. Accepted for publication June 6, 1989.
Address correspondence to D r Brunner, Mercy Endocrine and Diabetes Care Center, 2238 Jefferson Ave, Toledo, OH 43624.
Methods Study Protocol Patients with serum sodium concentrations of less than 115 mmol/L. were identified through emergency room referral and the central laboratory computer. With one exception, the evaluation, treatment, and correction rate of hyponatremia were established by attending physicians who were
Copyright 0 1990 by the American Neurological Association
61
not participants in the study. Therapy of hyponatremia commenced on discovery in all cases. Most of the patients received multiple modes of therapy in an attempt to correct hyponatremia. The neurological condition was assessed by standard neurological and mental status examinations while the patients were hospitalized. An initial MRI scan was performed within the first week of therapy; this was followed by a second scan a week later and a third scan at approximately 1 month. Precise duplication of the timing of MRI was not possible for all cases. All subjects had serial MRI performed within 5 weeks of admission, with one exception in which serial high-resolution computed tomography (CT) was used until MRI was available at 12 weeks. T I - and T2-weighted spin echo images were obtained utilizing a super-conducting magnet operating at 1.5 Tesla. Hgh-resolution CT was performed in patients in whom lesions were identified by MRI. Protocol approval was obtained from the Henry Ford Hospital Human Research Committee.
Measurement of Serum Soditlm Concentration, Calculation of Changes, and Statistical Analysis Serum sodium concentrations were measured in the hospital chemistry laboratory with an ion-specific electrode 24 and 48 hours after the initial identification of hyponatremia Before 24 hours, sodium concentrations were measured at intervals determined by the clinical staff. Intervals of sodium measurements varied from once in 24 hours to hourly. The correction rate of hyponatremia was calculated by subtracting the initial low sodium concentration from the sodium concentration at 24 and 48 hours later, divided by 24 and 48 (hours), and expressed as mmol/(L . hr). Subjects with pontine lesions evident on MRI were compared to those without lesions. Subjects who had persistent alterations in m e n d status 72 hours or more after hyponatremia correction (including those subjects with pontine lesions) were compared with those who had normal mental status after hyponatremia correction. Continuous variables were analyzed by a two-sided Student’s t test. Categorical variables (i.e., alcoholism, sex, and the presence or absence of central nervous system dysfunction at discovery of hyponatremia) were analyzed by Fisher’s Exact test. All means are expressed as 2 one standard deviation.
Results Study Subjects Thirteen subjects (10 female, 3 male) were studied (Table 1). Hyponatremia was the admitting diagnosis in 10 subjects and developed in 3 patients while they were hospitalized. At the presentation of hyponatremia, 8 patients had new or increased central nervous system dysfunction, ranging from altered mental status (6 subjects) to coma (2 subjects). Hyponatremia was attributed tcl Addison’s disease and hypotonic fluid administration in 1 subject. The syndrome of inappropriate antidiuretic hormone secretion was demonstrated in 6 subjects: 2 with polydipsia, 2 without polydipsia, and 2 postoperatively. Diuretic administration was implicated in 4 subjects, 3 of whom were polydipsic. Beer potomania and bladder
62 Annals of Neurology
Vol 27
N o 1 January 1990
Table 1. Charactmitics of 13 Study Subjects Variable
Mean
SD
Age (yr) Initial “a} (mmofi) [Na] after 24 hr (mmoVL) [Na] after 48 hr (mmofi) Correction rate at 24 hr (mmol/[L . hr]) Correction rate at 48 hr (mmol/CL . hr]) Time to reach “a], 135 mmoyL (days) Maximum [Na] (mmol/L) Time to reach maximum CNa7 (days)
64.0 103.7 124.1 128.7 0.86
14.2 37-83 93-113 6.3 113-134 5.9 119-140 6.0 0.39 0.21-1.67
0.56
0.21 0.33-0.97
5.3
4.1
2-16
136.1 4.46
3.8 2.2
130-143 2-9
Range
SD = standard deviation.
irrigation with glycine during a transurethral resection of the prostate were identified as the cause of hyponatremia in the remaining 2 patients. One subject admitted to current alcohol abuse and 1 had a history of remote alcohol abuse. Neither had evidence of hepatic insufficiency. All patients eventually underwent a change of 25 mmol/L in sodium concentration at variable intervals. All subjects were mildly hyponatremic after 24 hours of treatment and none developed hypernatremia.
Patients with Pontine Lesions Three of 13 subjects were shown to have pontine lesions by MRI after correction of hyponatremia. None of these lesions were demonstrated by CT. PATIENT 1. A 65-year-old man with essential hypertension and a history of remote alcohol abuse was admitted for an elective transurethral resection of the prostate, in which a glycine solution (200 mosdkg) was used for irrigation. Preoperatively, the sodium concentration was 133 mmofi. Postoperatively, the patient failed to awaken and required mechanical ventilation, but no respiratory arrest occurred. The serum sodium concentration was 93 mmol/L. Intravenous furosemide and normal saline were started immediately and were subsequently changed to fluid restriction. After 8 hours of therapy, at a serum sodium concentration of 118 mmoYL, 2 brief generalized seizures occurred. At 24 hours serum sodium was 133 mmol/L. The patient was able to follow simple commands and move all extremities to painful stimuli; extraocular muscle movements were full and gag reflex was normal. Several days later he was extubated, and at that time he was found to be disoriented. Formal neuropsychological evaluation revealed impaired memory, problem-solving, and psychomotor skills. Electroencephalography revealed diffuse slow-wave activity consistent with a moderate cerebral disturbance. Auditory evoked potentials showed no pontine dysfunction. Somatosensory evoked potentials in the upper
A
B
D Patient I. T1 spin echo magnetic resonance images on days 4 (A), 7 (B), and 18 (C) and at 4 months (0).An evolving lowsignal abnomlity is seen within the pons that is most evident on dzy 18 and is barely detectable at 4 months.
extremities demonstrated only a mild peripheral neuropathy. At discharge from the hospital, 13 days after admission, he was alert but had an unsteady gait. Two months later a second neuropsychological evaluation demonstrated dramatic improvement. MRI studies were obtained after 4, 7, and 18 days and after 4 months (Figure). They demonstrated an evolving pontine lesion that was most evident on day 18 and barely detectable at 4 months.
2. A 37-year-old woman was admitted with a serum sodium concentration of 94 mmoVL. This case has been reported previously [ 131.) Sodium concentrations were
106 mmolfl. at 6 hours, 123 mmoYL at 24 hours, and 129 mmoVL at 48 hours. After correction the patient’s neurological status on examination appeared to improve, bur 3 days later she became disoriented and subsequently comatose with decorticate posturing. There was no specific evidence of a pontine lesion. Her neurological status gradually improved. After 30 months of follow-up, she had persistent mild impairment of cortical capacities that continued to prevent her from returning to work as a schoolteacher. Serial CT scans of the pons and brain at 1,2, and 12 weeks were normal, but an MRI obtained at 12 weeks demonstrated a large pontine defect consistent with CPM f13). MRI studies at 12 and 30 months demonstrate marked improvement in the lesions.
PATIENT
PATIENT 3. A 78-year-old woman was admitted with a complaint of urinary retention. She had been following a
Brunner et al: Pontine Lesions and Rapid Correction of Hyponatremia
63
Table 2. Subjects With Versus Subjects Without Pontine hsions 95% Confidence Limits Yes (n = 3) Variable ~~~~~~~
4 w (yr) Initial {Na] (mmovL) {Na] after 24 hr (mmol/L) [Na] after 48 hr (mmol/L) Correction rate at 24 hr (mmoV[L . hr]) Correction rate at 48 hr (mmoVEL . hr]) Time to change {Na] 25 m m o f i (hr) Maximum [Na] (mmofi) Time to reach maximum [Nal (days)
No (n = 10)
With Pontine Lesions
Mean
SD
Mew
SD
PValue"
60.0 97.3 127.3 131.3 1.25 0.7
21.0 6.7 5.1 7.8 0.4 0.20
65.0 105.6 123.10 127.9 0.74 0.50
13.0 5.2 6.10 5.60 0.32 0.20
0.62 0.04 0.31 0.4 1 0.04 0.16
- 26.4 - 16.3 - 4.4 - 5.4
25.7 136.0 6.3
20.1 5.6 3.8
55.3 136.2 3.9
28.5 3.60 1.4
0.13 0.94 0.10
- 69.0
Without Pontine Lesions
~
0.023
- 0.09 -6.1 - 0.6
16.4
- 0.3 12.8 12.2 0.997 0.49 9.8 5.6 5.4
"Two-tailed Student's t test.
self-prescribed weight loss regimen that consisted of no food and large amounts of water. Her medications included a combination diuretic (spironolactone, 25 mg, and hydrochlorothiazide, 25 mg). The initial serum sodium concentration was 105 mmovL. Results of initial general physical, neurological, and mental status examinations were normal. She was treated with intravenous 514 mmofi saline followed by 154 mmol/L saline and then fluid restriction. The sodium concentrations were: 109 mmofi at 3 hours, 121 mmofi at 5 hours, 126 mmovL at 24 hours, and 129 mmoV L after 48 hours. During therapy no seizures or hypoxic episodes were observed. Thirty-six hours after admission the patient was confused and had inappropriate behavior. At 48 hours she had progressed to flaccid quadriplegia with a right gaze preference. The clinical diagnosis of CPM was made and 4 mg of dexamethasone was administered intravenously every 6 hours for a week. After 72 hours she began to regain use of all her limbs and the eye movement abnormalities resolved. Electroencephalography demonstrated diffuse slowing without focal or lateralizing features. MRI was performed on days 4, 10, and 3 1 after admission. On the first scan the pons was normal. In the cerebral hemispheres scattered, small, nonsymmetrical white matter lesions were demonstrated, which remained unchanged throughout the study. These lesions in the cerebral hemispheres are consistent with T 2 enhancing lesions, which are frequently seen in elderly patients but have limited pathological significance. Subsequent scans demonstrated a transient, small hyperintense lesion within the pons with a surrounding area of decreased signal on TI images. The third scan demonstrated a small rightsided lesion of the pons near the midline and anterior to the fourth ventricle on T2-weighted imaging.
Patients Without Pontine Lesions In 10 subjects MRI did not demonstrate pontine lesions. Four of these 10 had persistent mental status alterations that lasted at least 3 days in 4 subjects, 1 of whom had memory deficits that persisted during 18 months of follow-up study. Two additional subjects
64 Annals of Neurology Vol 27 N o 1 January 1990
had prior neurological dysfunction that worsened at discovery of hyponatremia: 1 had a history of chronic dementia and the other had an old left parietal lobe infarct. After correction of hyponatremia these patients returned to baseline and neither had pontine abnormalities on MRI. Four patients had both normal neurological findings, mental status, and MRI of the brain after correction of hyponatremia.
Statistical AnaLysis Among the 3 patients with pontine lesions, the initial sodium concentration was significantly lower ( p = 0.04) and the correction rate of sodium at 24 hours significantly more rapid ( p = 0.04) than in patients without pontine abnormalities (Table 2). Before 24 hours, the mean correction rate of sodium for the 3 subjects in whom pontine lesions developed was 2.8 2 0.72 mmol/(L * hr) at an average of 5.7 hours. Among the 10 subjects without pontine lesions, 8 had sodium measurements between 5 and 11 hours of treatment. We calculate an early correction rate for this group at 1.40 & 0.80 mmol/(L . hr) at an average interval of 7.6 hours. No differences were detected in sex, presence or absence of central nervous system dysfunction at discovery of hyponatremia, or alcohol abuse. In the 7 subjects with central nervous system dysfunction that persisted for 72 hours or more despite correction, including those with and those without pontine lesions, an association was also seen with a lower initial sodium concentration and more rapid correction of hyponatremia at both 24 and 48 hours (Table 3). No discernible statistically significant differences were found in age, absolute sodium concentration at 24 and 48 hours, highest sodium concentration achieved, time to undergo a 2hnmolfl. change in sodium concentration, or time to reach highest sodium concentration. Likewise, for categorical variables no
Table 3. Subjects With Versus Subjects Without Persistent Central Nervous System Dysfunction Ajier Hyponatremia
95% Confidence Limits
Yes (n Variable
Mean
SD
Mean
SD
p Valueb
With CNS Dysfunction
Age (yr)
60.1 99.3 126.0 128.3 1.1
15.1 4.8 4.3 6.6 0.3
61.8 108.0 122.0 128.8 0.6
10.3 3.3 6.2 0.3
0.86 0.01 0.33 0.91 0.02
-21.1 - 14.9 - 4.8 - 10.0 0.07
0.7
0.2
0.4
0.1
0.04
0.05
34.6
30.0
60.0
17.4
0.21
- 62.9
12.1
135.4 5.3
3.4 2.7
134.5 3.3
2.6 1.3
0.65 0.19
- 3.6
- 1.3
5.4 5.3
Initial [Na) (mmol/L) “a) after 24 hr (mmol/L) [Na) after 48 hr (mmoVL) Correction rate at 24 hr (mmol/ EL . hrl) Correction rate at 48 hr (mmoV EL . hrl) Time to change “a), 25 mmoYL (hr) Maximum “a) (mmol/L) Time to reach maximum {Na)
= 7)”
No (n
= 4)
8.8
Without CNS Dysfunction 17.7 - 2.5
12.8 8.4 0.93 0.55
(days) “Two patients with known CNS dysfunction before hyponatremia discovery are not used in analysis. bTwo-tailed Student’s t test.
significant differences were seen in sex, presence or absence of central nervous system dysfunction at discovery of hyponatremia, or alcohol abuse. Discussion We have described 13 subjects with profound hyponatremia in whom a rise in sodium concentration of 25 mmol/L or more occurred. All subjects were mildly hyponatremic at 24 hours and no one had hypernatremia. Pontine lesions developed in 3 patients after the correction of hyponatremia. The correction rate of hyponatremia at 24 hours was significantly more rapid in those patients who had pontine lesions than in those who did not. Correction rates of hyponatremia at intervals before 24 hours were faster than the rates calculated at 24 hours; this reflects efforts to increase the sodium concentration by the treating physicians. In patients without pontine lesions, the rate was 1.45 mmol/(L . hr) at 7.3 hours. In those in whom pontine lesions developed, the rate before 24 hours was 2.8 mmol/(L hr) at 5.7 hours. A statistical comparison of the correction rates before 24 hours is not provided because serum sodium concentrations were not obtained at the same interval. Two experiences from this study indicate that early aggressive therapy may expose patients to the risk of CPM. In Patient 3 a brief period of rapid correction (12 mmoVL over 2 hours) was associated with clinical and radiographic evidence of a pontine lesion despite a total change in sodium concentration that was limited to 21 mmoVL at 24 hours and 24 mmoVL at 48 hours. In Patient 1, the treating physician’s intent was to rapidly correct one-half of the sodium concentration deficit. Slow correction by fluid restriction was to fol-
low. However, during fluid restriction the sodium concentration spontaneously and rapidly increased, presumably because of the patient’s own homeostatic mechanisms. The initial severity of hyponatremia was also associated with the development of pontine lesions. Severe hyponatremia has not been associated with demyelinating lesions in animals with uncorrected hyponatremia that has been sustained for as long as 10 days [7, 8 , 147. However, large increases in sodium concentration have been associated with demyelinating lesions in humans and animal models E6-8, l l , 123. Severely hyponatremic individuals (initial sodium concentration less than 110 mmollL) must inevitably undergo a large rise in sodium concentration. We believe this presents an added risk for development of pontine lesions if the hyponatremia is reversed too rapidly. The fact that CPM is not limited to patients with the greatest severity of hyponatremia does not negate the significance of this observation, as other factors, such as hepatic insufficiency and overcorrection to hypernatremia, may be additional risk factors for CPM 1 1 1 , 12, 157. Prospective animal models demonstrate that rapid correction of hyponatremia leads to more extensive demyelinating lesions in chronically hyponatremic animals than in animals with acute hyponatremia of similar severity that is corrected at an identical rate [14, 167. Ten of the 13 subjects in this study developed hyponatremia as outpatients and the duration of the hyponatremia cannot be determined. However, Patient l documents an episode of acute, unsustained, and symptomatic hyponatremia that was rapidly corrected ( 3 mmoV[L. hr} at 8 hours, 1.67 mmol/[L hr} at 24 hours) in which CPM developed.
-
Brunner et al: Pontine Lesions and Rapid Correction of Hyponatremia 65
Ayus and associates Ell] have noted that seizures and hypoxia occur frequently in patients in whom CPM subsequently develops after hyponatremia. They suggest that many clinical and pathological findings resemble delayed anoxic encephalopathy. However, the lesions of delayed anoxic encephalopathy occur predominantly in the white matter of the cerebral hemispheres and are associated with profound anoxia and hypotension 117-191. There have been no reports of delayed anoxic encephalopathy with pontine lesions resembling CPM. As in other reports, seizures occurred in this study after significant increases in sodium concentration had occurred {9}. The occurrence of seizures could alternatively be interpreted as sequellae to very rapid correction rates during the initial hours of therapy. In this study we prospectively used serial MRI in determining the presence or absence of pontine lesions after correction of hyponatremia. This is the likely explanation for the higher rate of pontine lesions compared with other studies C4, 20-231. Our experience utilizing MRI demonstrated that pontine lesions can occur without demonstrable clinical findings of pontine dysfunction, that pontine lesions may be delayed several days after correction of hyponatremia, that resolution or improvement of lesions may also occur, and that MRI is more useful than highresolution CT in the detection of pontine lesions associated with hyponatremia correction. The data from this study cannot absolutely resolve the question concerning the appropriate correction rate of hyponatremia because the sample size is small and its design is not randomized. However, our data support the following conclusions: (1) Rapid correction of hyponatremia is associated with clinical neurological dysfunction and radiographic evidence of pontine lesions. ( 2 ) Patients with hyponatremia less than 105 m m o n are at greater risk to develop the complications associated with rapid correction of hyponatremia than are patients with higher sodium concentrations. (3) Patients with acute hyponatremia are also at risk for complications associated with rapid correction of hyponatremia.
Financial support for this study was provided by the Fund for Henry Ford Hospital.
References 1. Arieff AI, Llach F, Shad GM. Neurological manifestations and morbidity of hyponatremia: correlation with brain water and electrolytes. Medicine 1976;55:12 1-129 2. Ashraf N, Locksley R, Arieff AI. Thiazide-induced hypona-
66 Annals of Neurology Vol 27 No 1 January 1990
tremia associated with death or neurologic damage in outpatients. Am J Med 1981;70:1163-1168 3. Arieff AI. Hyponatremia, convulsions, respiratory arrest, and permanent brain damage after elective surgery in healthy women. N Engl J Med 1986;314:1529-1535 4. Ayus JC, Oliver0 JJ, Frommer JP. Rapid correction of severe hyponatremia with intravenous hypertonic saline solution. Am J Med 1982;72:43-48 5. Ayus JC, KrothapalLi RK, Arieff AI. Changing concepts in treatment of severe symptomatic hyponatremia Am J Med 1985;78:897-902 6. Norenberg MD, Leslie KO, Robertson AS. Association between rise in serum sodium and central pontine myelinolysis. Ann Neurol 1982;11:128-139 7. Kleinschmidt-DeMasters BK, Norenberg MD. Rapid correction of hyponatremia causes demyelination: relation to central pontine myelinolysis. Science 1981;211:1068-1070 8. Laureno R.Central pontine myelinolysis following rapid correction of hyponatremia. Ann Neurol 1983;13:232-242 9. Stems RH,Riggs JE, Schochet SS. Osmotic demyelination syndrome following correction of hyponatremia. N Engl J Med 1986;3 14:1535-1542 10. Sterns RH. Severe symptomatic hyponatremia: treatment and outcome: a study of 64 cases. Ann Intern Med 1987;107:656664 11. Ayus JC, Krothapalli RK, Armstrong DL. Rapid correction of severe hyponatremia in the rat: histopathologicalchanges in the brain. Am J Physiol 1985;248:F711-F719 12. Ayus JC, Krothapalli RK, Arieff AI. Treatment of symptomatic hyponatremia and its relation to brain damage: a prospective study. N Engl J Med 1987;317:1190-1195 13. Brunner JE, Redmond JM, Haggar AM, Elias SB. Central pont h e myelinolysis after rapid correction of hyponatremia: a magnetic resonance imaging study. Ann Neurol 1988;23:389-391 14. Illowsky BP, Laureno R Encephalopathy and myelinolysis after rapid correction of hyponatremia Brain 1987;110:855-867 15. Adams RD, Victor M, Mancall EL. Central pontine myelinolysis: a hitherto undescribed disease occurring in alcoholic and malnourished patients. Arch Neurol Psychiatry 1959;81:154172 16. Norenberg MD, Papendick RE. Chronicity of hyponatremia as a factor in experimental myelinolysis. Ann Neurol 1984;15: 544-547 17. Ginsberg MD. Delayed neurological deterioration following hypoxia. Adv Neurol 1979;262 1-44 18. Plum F, Posner JB, Hain RF. Delayed neurological deterioration after anoxia. Arch Intern Med 1962;11018-25 19. Jennett B, Plum F. Persistent vegetative state after brain damage: a syndrome in search of a name. Lancet 1972;1:734-737 20. Decaux G, Unger J, Brimioulle S, Mockel J. Hyponatremia in the syndrome of inappropriate secretion of antidiuretic hormone: rapid correction with urea, sodium chloride, and water restriction therapy. JAMA 1982;247:471-474 21. Ashouri 0s. Severe diuretic-induced hyponatremia in the elderly: a series of eight patients. Arch Intern Med 1986;146: 1355-1357 22. Hantman D , Rossier B, Zohlman R,Schrier R. Rapid correction of hyponatremia in the syndrome of inappropriate secretion of antidiuretic hormone and alternative treatment of hypertonic saline. Ann Intern Med 1973;78:870-875 23. Worthley LIG, Thomas PD. Treatment of hyponatremic seizures with intravenous 29.2% saline. Br Med J 1986;292:168170
