Introduction
Hyponatremia is one of the most common electrolyte disturbances, occurring in approximately 14%–42% of hospitalized patients, and it is associated with higher mortality (1,2). Although the higher risk of death associated with hyponatremia may reflect severity of related illnesses (e.g., congestive heart failure, cirrhosis, and malignancy), acute or severe hyponatremia can result in life-threatening cerebral edema (3). Among patients admitted with severe hyponatremia (sodium <120 mEq/L), in-hospital mortality ranges from 6% to 10% (3,4). Raising serum sodium by 4–6 mEq/L is sufficient to resolve cerebral edema; however, overly rapid correction of severe hyponatremia can result in osmotic demyelination syndrome and central pontine myelinolysis (4–6). Manifestations of osmotic demyelination syndrome can include encephalopathy, seizures, Parkinsonian-like movement disorders, and locked-in syndrome (7–9).
Recent United States guidelines recommend that correction rates not exceed 8 mEq/L for any 24-hour period in patients at high risk for osmotic demyelination (serum sodium ≤105 mEq/L, hypokalemia, malnutrition, or liver disease). European guidelines suggest limiting correction to ≤10 mEq/L in the first 24 hours and ≤8 mEq/L for any 24-hour period thereafter (10–12). Few studies have primarily examined risk factors of overly rapid correction or osmotic demyelination and have been limited by relatively small sample size, limited imaging data, and/or single centers (4,13,14). Reports of osmotic demyelination have mostly been limited to patient reports and small studies (7,15,16). Thus, it remains unclear how often rapid correction or osmotic demyelination occurs in patients presenting with severe hyponatremia.
Using data from seven hospitals in the Geisinger Health System, we examined incidence and risk factors of rapid correction and osmotic demyelination among patients presenting with severe hyponatremia. Our goal was to identify risk factors on the day of admission for rapid correction and osmotic demyelination, enabling clinicians to recognize high-risk patients and potentially prevent devastating neurologic consequences.
Materials and Methods
Study Design and Setting
We extracted data from seven hospitals (one academic and six nonacademic) in the Geisinger Health System, a fully integrated health care system serving central and northeastern Pennsylvania. We included adults ≥18 years of age admitted between January 1, 2001 and February 22, 2017 with an initial serum sodium <120 mEq/L. We excluded patients who had no serum sodium values within 12 hours of the 24- or 48-hour time points after admission and those with serum glucose >300 mg/dl on admission. The Geisinger Institutional Review Board reviewed and approved the research study.
Definition of Rapid Correction Outcomes
Serum sodium and other electrolytes were measured using an indirect ion-selective electrode method (Cobas; Roche Diagnostics). For each study participant, we calculated the estimated serum sodium at 24 hours [Na (24)] using the following formula: Na (24)= Naa + [(Nab − Naa) ×(24− Ta)/Tb − Ta)], where Naa and Ta are the closest serum sodium and time values before the 24-hour mark, respectively, and Nab and Tb are the closest serum sodium and time values after the 24-hour mark, respectively (13). For patients with only one serum sodium value within 12 hours of the 24-hour mark, we used the following formula to estimate serum Na (24): Na (24)= [(Naa − Na0)/Ta] ×24+ Na0 or Na (24)= [(Nab − Na0)/Tb] ×24+ Na0 depending on whether Naa or Nab was available.
The primary outcome was rapid correction at 24 hours, which was defined as the estimated rate of serum sodium correction >8 mEq/L at 24 hours. Secondary outcomes included alternative definitions of rapid correction: correction >8 mEq/L at any point during the first 24 hours, correction >10 mEq/L at 24 hours, or correction >18 mEq/L at 48 hours.
Definition of Osmotic Demyelination Syndrome Outcomes
To determine incidence of osmotic demyelination in the study population, we conducted chart reviews on patients with International Classification of Disease (ICD) 9 and 10 diagnostic codes or magnetic resonance imaging (MRI) of the brain at any time after the index serum sodium. MRI reports were manually reviewed by a nephrology fellow by searching for terms such as central pontine myelinolysis, central pontine gliosis, acute osmotic demyelination, and osmotic demyelination syndrome. A nephrology fellow and two attending nephrologists reviewed all charts with MRI evidence of osmotic demyelination to abstract additional details about presentation, hospital course, and outcome.
Other Variables of Interest
Additional data collected included demographics, comorbid conditions by ICD-9/10 codes (cirrhosis, chronic liver disease, nonalcoholic steatohepatitis, hepatic steatosis, fatty liver, alcohol abuse, malnutrition, central pontine myelinolysis, CKD, congestive heart failure, diabetes mellitus, depression, bipolar disorder, and schizophrenia), medications (thiazide and loop diuretics, aldosterone antagonists, selective serotonin reuptake inhibitors, 0.9% normal saline solution, 3% saline solution, vasopressin receptor antagonists, and intravenous or oral electrolyte repletion during admission, including potassium, magnesium, calcium, and phosphorus), clinical data (height, weight, and systolic and diastolic BP measurements), laboratory data on the day of admission (serum sodium, creatinine, potassium, phosphorus, serum osmolality, albumin, urinalysis, urine sodium, urine potassium, and urine osmolality), last outpatient serum sodium value <135 mEq/L, intensive care unit (ICU) stay during admission, and 30-day mortality.
Statistical Analyses
We examined differences between patients who did or did not experience correction >8 mEq/L at 24 hours using nonparametric Kruskal–Wallis tests for continuous variables and chi-squared tests for categorical variables. The final models for estimation of adjusted odds ratios (aORs) were developed on the basis of clinical rationale and forward selection using Akaike Information Criterion. Because urine sodium was not measured on all patients and inpatient management strategies were likely on the basis of characteristics assessed on presentation, we constructed three multivariable models: (1) model 1 included all significant unadjusted risk factors of rapid correction except for urine sodium or inpatient treatment factors, (2) model 2 included model 1 covariates and urine sodium (<30 or ≥30 mEq/L), and (3) model 3 included model 1 covariates and inpatient treatment factors. A P value of <0.05 was considered statistically significant for all comparisons without adjustment for multiple comparisons.
Results
Study Cohort Characteristics
A total of 1718 patients were admitted between January 1, 2001 and February 22, 2017 with severe hyponatremia on admission (sodium <120 mEq/L). After excluding 42 patients missing serum sodium values within 12 hours of the 24- or 48-hour time points after admission and 186 patients who had plasma glucose >300 mg/dl on admission, 1490 patients were included in the main analysis. The baseline characteristics are shown in Table 1. Median (interquartile range [IQR]) change in serum sodium was 6.8 mEq/L (IQR, 3.4–10.2) at 24 hours and 10.3 mEq/L (IQR, 6.5–14.8) at 48 hours (Figure 1). A total of 606 (41%) and 390 (26%) patients had correction >8 mEq/L and correction >10 mEq/L at 24 hours, respectively; 166 (12%) of 1346 patients with 48-hour sodium data had correction >18 mEq/L at 48 hours.
Table 1. -
Characteristics of adults admitted to Geisinger system hospitals with an initial serum
sodium <120 mEq/L by change in serum
sodium at 24 hours after admission
Characteristic |
Na+ Correction of ≤8 mEq/L at 24 h, n=884 |
Na+ Correction of >8 mEq/L at 24 h, n=606 |
Age, yr |
68 (15) |
63 (15) |
Women, n (%) |
460 (52) |
359 (59) |
Non-Hispanic white |
865 (98) |
594 (98) |
Smoking status, n (%)
|
|
|
Current smoker |
216 (26) |
225 (40) |
Former smoker |
287 (35) |
138 (24) |
Never smoker |
310 (37) |
186 (33) |
Unknown |
18 (2) |
18 (3) |
Body mass index, kg/m2
|
28 (8) |
26 (6) |
Systolic BP, mm Hg |
133 (29) |
136 (30) |
Diastolic BP, mm Hg |
71 (17) |
74 (18) |
Comorbidities, n (%)
|
|
|
Chronic liver disease |
72 (8) |
36 (6) |
CKD |
109 (12) |
57 (9) |
Nonalcoholic steatohepatits |
13 (2) |
11 (2) |
Hepatic steatosis |
30 (3) |
27 (5) |
Fatty liver |
66 (8) |
35 (6) |
Alcohol abuse |
140 (16) |
122 (20) |
Malnutrition |
304 (34) |
202 (33) |
Congestive heart failure |
164 (19) |
73 (12) |
Diabetes mellitus |
145 (16) |
83 (14) |
Depression |
141 (16) |
123 (20) |
Bipolar disorder |
41 (5) |
37 (6) |
Schizophrenia |
12 (1) |
22 (4) |
Epilepsy |
83 (9) |
79 (13) |
Seizure |
81 (9) |
80 (13) |
Stroke |
49 (6) |
32 (5) |
Dementia |
9 (1) |
6 (1) |
Cancer |
218 (25) |
115 (19) |
Charlson, n (%)comorbidity index
|
|
|
0 |
26 (3) |
45 (7) |
1 |
46 (5) |
60 (10) |
2 |
80 (9) |
86 (14) |
≥3 |
732 (83) |
415 (69) |
ICU stay during the first 24 h after hospital admission, n (%) |
187 (21) |
129 (21) |
Outpatient Na+ value <135 mEq/L, n (%) |
528 (73) |
294 (59) |
Admission laboratory values
|
|
|
Sodium, mEq/L, n=1490 |
117 (4) |
115 (5) |
Creatinine, mg/dl, n=1452 |
1.3 (1.5) |
1.3 (2.0) |
eGFR, ml/min/1.73m2, n=1452 |
72 (35) |
80 (36) |
Potassium, mEq/L, n=1489 |
4.3 (1.0) |
4.0 (1.0) |
Phosphorus, mg/dL, n=851 |
3.5 (1.9) |
3.3 (2.0) |
Magnesium, mg/dL, n=927 |
1.9 (0.4) |
1.8 (0.4) |
Osmolality, mOsm/kg, n=951 |
263 (94) |
258 (19) |
Albumin, g/dL, n=1280 |
3.3 (0.7) |
3.6 (0.6) |
Glucose, mg/dL, n=1489 |
127 (40) |
128 (40) |
Urine sodium, mEq/L, n=1141 |
43 (38) |
35 (31) |
Urine potassium, mEq/L, n=421 |
32 (17) |
27 (19) |
Urine osmolality, mOsm/kg, n=1083 |
369 (150) |
270 (149) |
Outpatient medications, n (%)
|
|
|
Thiazide diuretics |
64 (7) |
36 (6) |
Loop diuretics |
226 (26) |
76 (13) |
Aldosterone antagonists |
102 (12) |
25 (4) |
Selective serotonin reuptake inhibitors |
150 (17) |
113 (19) |
Antiseizure medications |
154 (17) |
121 (20) |
Antipsychotic medications |
98 (11) |
92 (15) |
Inpatient medications, n (%)
|
|
|
Hypertonic saline |
82 (9) |
104 (17) |
Electrolyte repletion |
240 (27) |
236 (39) |
Vaptans |
11 (1) |
7 (1) |
Mortality within 30 d of hospital admission, n (%) |
167 (19) |
46 (8) |
Values are presented as mean (SD) or number (%). ICU, intensive care unit; Na+, sodium.
Figure 1.: Distribution of sodium correction from baseline to 24 and 48 hours and degree of sodium rise above cutoff level in patients admitted to Geisinger with initial serum sodium <120 mEq/L.
Patients who experienced correction >8 mEq/L at 24 hours were more likely to be younger (63 versus 68 years old), be current smokers (40% versus 26%), have lower body mass index (26 versus 28 kg/m2), have a history of depression (20% versus 16%), have schizophrenia (4% versus 1%), and have seizures (13% versus 9%), and they were less likely to have prior hyponatremia (59% versus 73%), chronic liver disease (6% versus 8%), congestive heart failure (12% versus 19%), or cancer (19% versus 25%). Rapid correctors had lower mean values for initial serum sodium (115 versus 117 mEq/L), random urine sodium (35 versus 43 mEq/L), urine potassium (27 versus 32 mEq/L), and urine osmolality (270 versus 369 mOsm/kg).
Risk Factors of Rapid Correction at 24 Hours
In multivariable analyses not including urine sodium or inpatient treatment factors (model 1), being a woman was associated with higher risk of rapid correction at 24 hours (aOR, 1.49; 95% confidence interval [95% CI], 1.14 to 1.96), whereas treatment at an academic medical center (aOR, 0.70; 95% CI, 0.54 to 0.90), higher Charlson comorbidity index (aOR, 0.94; 95% CI, 0.89 to 0.99), older age (aOR, 0.98; 95% CI, 0.97 to 0.99), prior hyponatremia (aOR, 0.62; 95% CI, 0.48 to 0.81), higher baseline serum sodium at hospitalization (per 1-mEq/L higher serum sodium; aOR, 0.90; 95% CI, 0.87 to 0.93), and outpatient aldosterone antagonist use (aOR, 0.48; 95% CI, 0.28 to 0.82) were associated with lower risk of rapid correction at 24 hours (Table 2). In multivariable analyses including urine sodium (model 2), urine sodium <30 mEq/L was significantly associated with greater risk of rapid correction at 24 hours (aOR, 1.58; 95% CI, 1.17 to 2.13); other results were largely unchanged. In multivariable analyses including urine sodium and inpatient treatment factors (model 3), none of the inpatient treatment factors (electrolyte repletion, vaptan, or ICU stay) were associated with risk of rapid correction at 24 hours.
Table 2. -
Factors associated with
sodium correction >8 mEq/L at 24 hours in patients admitted to Geisinger system hospitals with an initial serum
sodium <120 mEq/L
Variables |
OR (95% CI) |
Unadjusted |
Model 1 |
Model 2 |
Model 3 |
Academic center |
0.65 (0.53 to 0.81)
a
|
0.70 (0.54 to 0.90)
a
|
0.70 (0.52 to 0.94)
b
|
0.70 (0.54 to 0.91)
a
|
Women |
1.34 (1.08 to 1.65)
a
|
1.49 (1.14 to 1.96)
a
|
1.73 (1.26 to 2.37)
a
|
1.50 (1.14 to 1.96)
a
|
Age, per 1 yr |
0.98 (0.97 to 0.99)
a
|
0.98 (0.97 to 0.99)
a
|
0.98 (0.97 to 0.99)
a
|
0.98 (0.97 to 0.99)
a
|
White |
1.03 (0.49 to 2.15) |
1.03 (0.37 to 2.92) |
0.74 (0.23 to 2.35) |
1.01 (0.36 to 2.88) |
Schizophrenia |
2.73 (1.34 to 5.57)
a
|
2.24 (1.03 to 4.89)
b
|
2.48 (1.02 to 6.02)
b
|
2.24 (1.03 to 4.88)
b
|
Congestive heart failure |
0.60 (0.45 to 0.81)
a
|
0.94 (0.66 to 1.33) |
1.08 (0.72 to 1.63) |
0.94 (0.67 to 1.34) |
Charlson comorbidity index |
0.86 (0.82 to 0.89)
a
|
0.94 (0.89 to 0.99)
b
|
0.93 (0.87 to 0.99)
b
|
0.94 (0.88 to 0.99)
b
|
Outpatient Na+ <135 |
0.52 (0.41 to 0.67)
a
|
0.62 (0.48 to 0.81)
a
|
0.62 (0.46 to 0.84)
a
|
0.62 (0.48 to 0.81)
a
|
Inpatient baseline Na+ at hospitalization |
0.88 (0.86 to 0.91)
a
|
0.90 (0.87 to 0.93)
a
|
0.90 (0.87 to 0.93)
a
|
0.90 (0.87 to 0.93)
a
|
Urine Na+ <30 mEq/L |
1.46 (1.15 to 1.85)
a
|
|
1.58 (1.17 to 2.13)
a
|
|
K+≥5 mEq/L |
0.87 (0.66 to 1.16) |
1.22 (0.88 to 1.71) |
1.14 (0.76 to 1.70) |
1.23 (0.87 to 1.74) |
K+<3.5 mEq/L |
1.81 (1.41 to 2.33)
a
|
1.32 (0.97 to 1.79) |
1.39 (0.97 to 1.97) |
1.24 (0.88 to 1.73) |
Loop diuretics (outpatient) |
0.42 (0.31 to 0.56)
a
|
0.71 (0.49 to 1.02) |
0.66 (0.43 to 1.01) |
0.71 (0.49 to 1.02) |
Aldosterone antagonists (outpatient) |
0.33 (0.21 to 0.52)
a
|
0.48 (0.28 to 0.82)
a
|
0.45 (0.23 to 0.84)
b
|
0.48 (0.28 to 0.83)
a
|
Hypertonic saline (inpatient) |
2.03 (1.49 to 2.76)
a
|
|
|
1.09 (0.73 to 1.64) |
Electrolyte repletion (inpatient) |
1.71 (1.37 to 2.13
a
|
|
|
1.14 (0.84 to 1.55) |
Vaptans (inpatient) |
0.93 (0.36 to 2.41) |
|
|
1.21 (0.40 to 3.63) |
ICU stay in the first 24 h after admission |
1.01 (0.78 to 1.30) |
|
|
1.07 (0.79 to 1.44) |
Model 1 included all significant unadjusted risk factors of rapid correction except for urine sodium or inpatient treatment factors. Model 2 included model 1 covariates and urine sodium (<30 or ≥30 mEq/L). Model 3 included model 1 covariates and inpatient treatment factors. OR, odds ratio; 95% CI, 95% confidence interval; Na+, sodium; K+, potassium; ICU, intensive care unit.
aSignificant at P=0.01.
bSignificant at P=0.05.
In sensitivity analyses using different definitions of rapid correction (>10 mEq/L at 24 hours, >18 mEq/L at 48 hours, or >8 mEq/L from baseline to any point during the initial 24 hours), results were largely consistent with a few exceptions (Supplemental Tables 1–3). Vaptan use was associated with higher risk (aOR, 3.92; 95% CI, 1.09 to 14.09) and ICU stay was associated with lower risk (aOR, 0.54; 95% CI, 0.31 to 0.96) of rapid correction >18 mEq/L at 48 hours. Treatment at an academic center was not associated with greater risk of correction >8 mEq/L from baseline to any point during the initial 24 hours (aOR, 1.03; 95% CI, 0.78 to 1.36).
Incidence and Risk Factors of Osmotic Demyelination
A total of 295 (20%) patients had brain MRI completed during follow-up, with nine patients (0.6%) showing radiologic evidence of osmotic demyelination; no patients had ICD diagnoses of central pontine myelinolysis. One patient already had osmotic demyelination on admission MRI. Patient characteristics, risk factors, sodium trends, treatments, and outcomes are shown in Figure 2 and Table 3. Of the eight (0.5%) patients who developed incident osmotic demyelination, seven (88%) had documented sodium correction >8 mEq/L during any 24-hour period before brain MRI. The one patient without documented evidence of rapid correction was noted to have serum sodium levels of 105 and 132 mEq/L in the month before index admission, but further details on timing were lacking. Important characteristics observed in the patients with incident osmotic demyelination included hypovolemia (75%), beer potomania (63%), outpatient thiazide diuretic use (25%), alcohol use disorder (50%), malnutrition (50%), and hypokalemia (63%). Three patients received 3% saline before rapid correction due to acute neurologic symptoms. Dextrose 5% water solution was given to three patients and desmopressin was given to one patient to slow the rate of sodium correction. Five patients with documented osmotic demyelination had recovery with no neurologic deficits, two patients died from unrelated causes, and two were lost to follow-up.
Figure 2.: Serum sodium trends during the first 24 and 48 hours of admission in patients with radiologic evidence of osmotic demyelination.
Table 3. -
Characteristics of patients admitted to Geisinger system hospitals with an initial serum
sodium <120 mEq/L and osmotic demyelination on magnetic resonance imaging
Patient |
Hospital |
Initial Serum; Urine Sodium, mEq/L |
Hyponatremia Etiology |
Osmotic Demyelination Risk Factors |
Initial Treatment |
Urine Output over the First 24 h, ml |
Correction >8 mEq/L before MRI (Maximum over 24 h); Actions taken to Slow Rise |
Neurologic Signs before MRI |
Nephrology Consult |
Timing of MRI after Initial Na+
|
Site(s) Involved |
Outcome |
Patients with incident osmotic demyelination occurring after admission
|
|
|
|
|
|
|
|
|
|
|
|
|
Patient 1: 39-yr-old man with alcoholism, presented with pneumonia and encephalopathy |
Academic center |
Serum 110; urine 46
a
|
Hypovolemia, beer potomania |
Hypokalemia, alcohol use disorder, malnutrition |
3% Saline |
3900 |
Yes (12 mEq/L), on day 1; D5W given |
Upper extremity spasticity, mutism, encephalopathy |
Yes |
18 d later |
Central pons |
Wheelchair bound 1 yr, no neurologic deficits at 4 yr
b
; alcohol cessation |
Patient 2: 52-yr-old woman with alcoholism, HTN, depression on sertraline presented with lethargy |
Transfer from OSH to academic center |
Serum 98; urine 25 |
Hypovolemia, beer potomania, thiazide |
Hypokalemia, malnutrition, alcohol use disorder |
0.9% Saline |
Not documented |
Yes (11 mEq/L), on day 1; D5W given |
Hyper-reflexia, ataxia, bilateral lower extremity weakness, confusion |
Yes |
7 d later |
Central pons |
No neurologic deficits at 3 mo
b
; alcohol cessation |
Patient 3: 52-yr-old woman with alcoholism, depression on mirtazapine presented with seizures and hypotension |
Nonacademic hospital |
Serum 107; urine 52
a
|
Hypovolemia, beer potomania |
Alcohol use disorder |
3% Saline, 0.9% saline |
4300 |
Yes (22 mEq/L), on day 1 |
Lower extremity hyporeflexia, recurrent seizures |
No |
3 d later |
Central pons |
No neurologic deficits at 2 yr
b
; ongoing alcohol abuse |
Patient 4: 58-yr-old woman with alcoholism, spinal stenosis, HTN, prior hyponatremia on salt tablets presented with seizure, inebriation |
Nonacademic hospital |
Serum 112; urine 114
c
|
Beer potomania, thiazide |
Alcohol use disorder |
0.9% Saline |
1100 in an 8-h period, then not documented |
Yes (15 mEq/L), on day 1 |
Ataxia, lower extremity hyporeflexia, seizure |
Yes |
14 mo later |
Central pons |
Gait dysfunction, recurrent episodes of severe hyponatremia and alcohol intoxication; died 4 yr later from sepsis and hepatic encephalopathy |
Patient 5: 38-yr-old man with alcoholism, HTN on thiazide, depression on fluoxetine presented with unsteadiness and acute pancreatitis |
Nonacademic hospital |
Serum 113 |
Hypovolemia, beer potomania |
Hypokalemia, alcohol use disorder |
0.9% Saline |
2300 |
Yes (16 mEq/L), on day 2 |
Decreased visual acuity, hyper-reflexia, ataxia |
No |
11 d later |
Central pons, bilateral frontal, parieto-occipital, cerebellum, basal ganglia, and external capsules |
No neurologic deficits at 6 mo
b
; ongoing alcohol abuse |
Patient 6: 59-yr-old woman with multiple sclerosis, RA, HTN on thiazide, bipolar disorder on quetiapine and mirtazapine presented with encephalopathy, hypotension, and blurred vision |
Academic center |
Serum 117; urine <20 |
Hypovolemia, thiazide |
Hypokalemia, malnutrition |
0.9% Saline |
3690 |
Yes (13 mEq/L), on day 1; D5W, desmopressin |
Aphasia, lower extremity weakness |
Yes |
124 d later |
Central pons, bilateral cerebral white matter, not seen on prior MRI before rapid correction |
Death at 1 yr from septic shock due to clostridium difficile colitis |
Patient 7: 36-yr-old woman with alcoholism presented with shortness of breath, severe anemia |
Transfer from OSH to academic center |
Serum 115; urine <10 |
Hypervolemic |
Hypokalemia, malnutrition, alcohol use disorder, end stage liver disease (MELD score 29) |
0.9% Saline, 3% saline |
515 |
Yes (9 mEq/L), on day 3 |
Seizure, generalized weakness |
Yes |
18 d later |
Central pons, bilateral thalamus, subinsular regions |
Lost to follow-up |
Patient 8: 69-yr-old woman with diffuse large B cell lymphoma, prior hyponatremia presented with shortness of breath, malignant pleural effusion |
Academic center |
Serum 118 |
Hypovolemia |
Prior hyponatremia |
0.9% Saline |
175 |
No (7 mEq/L) but sodium 105 and 132 mEq/L in prior month at OSH without documentation of timing |
Encephalopathy, seizure |
No |
14 d later |
Central pons, bilateral basal ganglia |
No neurologic deficits
b
|
Patient with osmotic demyelination occurring before hospitalization with severe hyponatremia
|
|
|
|
|
|
|
|
|
|
|
|
|
Patient 9: 32-yr-old man with depression, heavy alcohol use presented with 5 d of dysarthria and ataxia; also reported salt craving and high salt intake in the 2 wk before presentation |
Transfer from OSH to academic center |
Serum 118; urine <10 |
Hypovolemia, beer potomania |
Hypokalemia, alcohol use disorder, malnutrition |
0.9% Saline |
977 |
No (7 mEq/L); D5W
c
|
Ataxia, dysarthria, dysmetria, intention tremor, opsoclonus |
Yes |
<24 h later |
Central pons, cerebellum |
Lost to follow-up |
MRI, magnetic resonance imaging; Na+, sodium; D5W, dextrose 5% in water; HTN, hypertension; OSH, outside hospital; RA, rheumatoid arthritis; MELD, model for end stage liver disease.
aChecked after receiving 3% saline.
bPer follow-up progress notes.
cOn salt tablets as outpatient.
Discussion
In a large cohort of patients presenting with severe hyponatremia, we examined clinical and radiologic data to describe incidence and risk factors of rapid correction and osmotic demyelination. We found that 41% of patients experienced correction >8 mEq/L at 24 hours, that 12% had correction >18 mEq/L at 48 hours, and that 0.5% of patients had incident osmotic demyelination confirmed by MRI. We found a significant number of risk factors of rapid correction and osmotic demyelination, confirming previously described associations and identifying some novel risk factors. Risk was more than twofold higher among patients with schizophrenia, although none of the 22 patients with schizophrenia who rapidly corrected developed osmotic demyelination. Because primary polydipsia is common in patients with schizophrenia, it seems likely that hyponatremia may have developed acutely in these patients from water intoxication before chronic hyponatremia brain cell adaptation occurred (17). Patients who presented at an academic center had a 30% lower risk of rapid correction >8 mEq/L at 24 hours and a 61% lower risk of rapid correction >18 mEq/L at 48 hours. This suggests that improved, timely access to specialists, such as nephrologists or intensivists, may be helpful in managing patients presenting with severe hyponatremia, although carefully designed, prospective studies are needed to show this.
To our knowledge, our cohort is the largest to examine the prevalence of MRI-confirmed osmotic demyelination in patients presenting with severe hyponatremia. A prospective imaging study investigating neurologic outcomes using serial brain MRI in 13 patients with severe hyponatremia (serum sodium <115 mEq/L) found osmotic demyelination lesions in three patients, all of whom had a serum sodium increase well beyond 8 mEq/L at 24 hours (mean of 30 mEq/L per day) (18). In other retrospective studies of patients who were severely hyponatremic, osmotic demyelination occurred in 0.2%–2% of patients (4,13). In our study, we examined all brain MRI reports completed on patients presenting with severe hyponatremia. Interestingly, despite a 41% incidence of rapid correction, only eight patients experienced incident osmotic demyelination during hospitalization, and one presented with osmotic demyelination syndrome, which was confirmed by MRI on admission day. Seven of eight patients with incident osmotic demyelination experienced sodium correction >8 mEq/L over any 24-hour period before brain MRI, and one possibly had rapid correction during admission at an outside hospital. Consistent with other patient series, we observed considerable heterogeneity in neurologic symptoms ranging from encephalopathy to persistent seizures or primarily cerebellar symptoms (19). Of the six alcoholic patients who had osmotic demyelination, two were lost to follow-up, and four surprisingly had recovery of neurologic function according to follow-up chart documentation, with two of four maintaining long-term sobriety. Interestingly, long-term neurologic outcomes in our cohort seemed better than those of other prior reports (19,20). This could be, in part, due to differences in study design and cohort selection.
Our findings suggest an association between rapid correction and osmotic demyelination, consistent with experimental animal models and some but not all observational studies (16,18,21–23). We found that further chart review was required to ascertain rapid correction in some patients with incident osmotic demyelination, because outpatient serum sodium values from outside hospitals were available in some progress notes, and two patients had sodium correction >8 mEq/L over a 24-hour period after day 1. Osmotic demyelination can also occur in other settings, such as hyperosmolar hyperglycemia, hyperammonemia, hypoxia, severe liver disease, and chronic alcoholism, in the absence of documented rapid sodium correction (24–29). The exact mechanism of demyelination in the setting of alcohol abuse is not entirely clear but could be related to direct neurotoxicity of alcohol, malnutrition, or underlying liver disease (26).
In our cohort, common features among patients who developed osmotic demyelination included hypovolemia, beer potomania, malnutrition, and hypokalemia. Commonly, patients who are hypovolemic are treated with normal saline boluses in the emergency department, often empirically on the basis of clinical assessment of fluid status. Although correction of hypovolemia is essential for stabilization of fluid status and hemodynamics, such patients are at risk for brisk water diuresis on correction of their hypovolemia, which can then lead to rapid sodium correction and the potentially catastrophic effects of osmotic demyelination. Hypokalemia was common on presentation in patients with incident osmotic demyelination, similar to other studies (15,16,30). We also found that a random urine sodium <30 mEq/L was associated with higher risk of rapid correction. This directly correlates with previous findings that a urine sodium of <30 mEq/L is superior to clinical assessment of volume status in identifying hyponatremic patients who would correct with isotonic saline administration (4,31). However, in edematous states, such as heart failure or cirrhosis, urine sodium may not be useful to assess risk of rapid correction, because a low value could reflect a state of decreased effective arterial blood volume in the setting of diminished cardiac output or systemic vasodilation. We also found that most patients who developed osmotic demyelination received 0.9% saline rather than 3% saline empirically on the basis of clinical assessment of hypovolemia. These findings emphasize the importance of thoughtful initial fluid management of patients who are severely hyponatremic, being mindful of potential risk of rapid correction when euvolemia is restored.
Treatment of severe hyponatremia with hypertonic saline is necessary and indicated for patients who present with severe neurologic symptoms (32). Previous uncontrolled studies have suggested that, when used in a specific, regimented protocol, hypertonic saline can be safe and effective in reversing the symptoms of hyponatremic encephalopathy (32–34). However, the effect of hypertonic saline dosing on serum sodium increase can be unpredictable, even with use of traditional formulas for dose calculation (30). Other strategies, such as combined administration of desmopressin and 3% saline, have been suggested to raise serum sodium in a controlled fashion (35). Further studies are needed to determine optimal initial management strategies for severe hyponatremia. Reversing rapid correction of hyponatremia with desmopressin and/or hypotonic fluids in osmotic demyelination animal models has been shown to reduce mortality, but data in humans are limited (36).
Our study has several strengths and limitations. Major strengths include the large number of patients in the cohort with data on rapid correction and osmotic demyelination as defined by manual review of every brain MRI report after initial hospitalization for severe hyponatremia and selected chart review for those with radiologic evidence of osmotic demyelination. The results should be interpreted cautiously given the study’s retrospective design and our inability to conduct chart review on the entire cohort to determine etiology of hyponatremia. We did not have complete laboratory data (such as urine sodium) on all patients, and we did not quantify exact doses of electrolyte repletion. Although the Geisinger Health System includes most hospitals in its coverage area, some patients were transferred from outside hospitals and may have lacked data on serum sodium values or signs of osmotic demyelination before transfer. Our study population was limited to mostly white patients in central and northeast Pennsylvania, although we included data from six nonacademic hospitals and one academic center. We only assessed risk factors within the first 24 hours of admission and did not examine rapid correction that may have occurred during the later periods of hospitalization for all patients. As noted in our study, ICD codes have poor sensitivity for osmotic demyelination, and we may have missed patients with osmotic demyelination who did not undergo brain MRI. Therefore, the incidence of rapid correction of severe hyponatremia and osmotic demyelination may have been underestimated in our cohort. Further research with blinded evaluation of brain MRIs in patients with severe hyponatremia may be needed to detect patients with more subtle cases of osmotic demyelination.
In conclusion, sodium correction >8 mEq/L at 24 hours occurred in 41% of patients presenting with severe hyponatremia in a large integrated health system. Risk factors of rapid correction included being a woman, younger age, history of schizophrenia, lower initial serum sodium value, and urine sodium <30 mEq/L. Hypovolemia, beer potomania, malnutrition, and hypokalemia were common in patients presenting with severe hyponatremia who developed incident osmotic demyelination. Future efforts are needed to identify optimal strategies for reducing risk of rapid correction and osmotic demyelination in patients presenting with severe hyponatremia.
Disclosures
None.
Acknowledgments
A.R.C. is supported by National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases grant K23 DK106515-01.
A poster presentation of this work was presented at the American Society of Nephrology Kidney Week on November 2, 2017 in New Orleans, Louisiana.
References
1. Waikar SS, Mount DB, Curhan GC: Mortality after hospitalization with mild, moderate, and severe hyponatremia. Am J Med 122: 857–865, 200919699382
2. Upadhyay A, Jaber BL, Madias NE: Incidence and prevalence of hyponatremia. Am J Med 119[Suppl 1]: S30–S35, 200616843082
3. Chawla A, Sterns RH, Nigwekar SU, Cappuccio JD: Mortality and serum
sodium: Do patients die from or with hyponatremia? Clin J Am Soc Nephrol 6: 960–965, 201121441132
4. Vu T, Wong R, Hamblin PS, Zajac J, Grossmann M: Patients presenting with severe hypotonic hyponatremia: Etiological factors, assessment, and outcomes. Hosp Pract (1995) 37: 128–136, 2009
5. Sterns RH, Hix JK, Silver S: Treatment of hyponatremia. Curr Opin Nephrol Hypertens 19: 493–498, 201020539224
6. Sterns RH, Nigwekar SU, Hix JK: The treatment of hyponatremia. Semin Nephrol 29: 282–299, 2009
7. Martin RJ: Central pontine and extrapontine myelinolysis: The osmotic demyelination syndromes. J Neurol Neurosurg Psychiatry 75[Suppl 3]: iii22–iii28, 200415316041
8. Sterns RH: Disorders of plasma
sodium. N Engl J Med 372: 1269, 201525806924
9. de Souza A, Desai PK: More often striatal myelinolysis than pontine? A consecutive series of patients with osmotic demyelination syndrome. Neurol Res 34: 262–271, 201222450756
10. Hoorn EJ, Zietse R: Diagnosis and treatment of hyponatremia: Compilation of the guidelines. J Am Soc Nephrol 28: 1340–1349, 201728174217
11. Adrogué HJ, Madias NE: Diagnosis and treatment of hyponatremia. Am J Kidney Dis 64: 681–684, 201424996937
12. Spasovski G, Vanholder R, Allolio B, Annane D, Ball S, Bichet D, Decaux G, Fenske W, Hoorn EJ, Ichai C, Joannidis M, Soupart A, Zietse R, Haller M, van der Veer S, Van Biesen W, Nagler E; Hyponatraemia Guideline Development Group: Clinical practice guideline on diagnosis and treatment of hyponatraemia. Nephrol Dial Transplant 29[Suppl 2]: i1–i39, 2014
13. Geoghegan P, Harrison AM, Thongprayoon C, Kashyap R, Ahmed A, Dong Y, Rabinstein AA, Kashani KB, Gajic O:
Sodium correction practice and clinical outcomes in profound hyponatremia. Mayo Clin Proc 90: 1348–1355, 201526434962
14. Sterns RH, Cappuccio JD, Silver SM, Cohen EP: Neurologic sequelae after treatment of severe hyponatremia: A multicenter perspective. J Am Soc Nephrol 4: 1522–1530, 19948025225
15. Gharaibeh KA, Brewer JM, Agarwal M, Fülöp T: Risk factors, complication and measures to prevent or reverse catastrophic
sodium overcorrection in chronic hyponatremia. Am J Med Sci 349: 170–175, 201525163018
16. Heng AE, Vacher P, Aublet-Cuvelier B, Garcier JM, Sapin V, Deteix P, Souweine B: Centropontine myelinolysis after correction of hyponatremia: Role of associated hypokalemia. Clin Nephrol 67: 345–351, 200717598369
17. Poirier S, Legris G, Tremblay P, Michea R, Viau-Guay L, Mérette C, Bouchard RH, Maziade M, Roy MA: Schizophrenia patients with polydipsia and water intoxication are characterized by greater severity of psychotic illness and a more frequent history of alcohol abuse. Schizophr Res 118: 285–291, 201020096540
18. Brunner JE, Redmond JM, Haggar AM, Kruger DF, Elias SB: Central pontine myelinolysis and pontine lesions after rapid correction of hyponatremia: A prospective magnetic resonance imaging study. Ann Neurol 27: 61–66, 19902301929
19. Odier C, Nguyen DK, Panisset M: Central pontine and extrapontine myelinolysis: From epileptic and other manifestations to cognitive prognosis. J Neurol 257: 1176–1180, 201020148334
20. Kallakatta RN, Radhakrishnan A, Fayaz RK, Unnikrishnan JP, Kesavadas C, Sarma SP: Clinical and functional outcome and factors predicting prognosis in osmotic demyelination syndrome (central pontine and/or extrapontine myelinolysis) in 25 patients. J Neurol Neurosurg Psychiatry 82: 326–331, 201120826870
21. Laureno R: Experimental pontine and extrapontine myelinolysis. Trans Am Neurol Assoc 105: 354–358, 19807348981
22. Laureno R: Central pontine myelinolysis following rapid correction of hyponatremia. Ann Neurol 13: 232–242, 19836847135
23. Sterns RH, Riggs JE, Schochet SS Jr .: Osmotic demyelination syndrome following correction of hyponatremia. N Engl J Med 314: 1535–1542, 19863713747
24. Graff-Radford J, Fugate JE, Kaufmann TJ, Mandrekar JN, Rabinstein AA: Clinical and radiologic correlations of central pontine myelinolysis syndrome. Mayo Clin Proc 86: 1063–1067, 201121997578
25. Desai RA, Davies AL, Tachrount M, Kasti M, Laulund F, Golay X, Smith KJ: Cause and prevention of demyelination in a model multiple sclerosis lesion. Ann Neurol 79: 591–604, 201626814844
26. Mochizuki H, Masaki T, Miyakawa T, Nakane J, Yokoyama A, Nakamura Y, Okuyama K, Kamakura K, Motoyoshi K, Matsushita S, Higuchi S: Benign type of central pontine myelinolysis in alcoholism--clinical, neuroradiological and electrophysiological findings. J Neurol 250: 1077–1083, 200314504969
27. Tanneau RS, Henry A, Rouhart F, Bourbigot B, Garo B, Mocquard Y, Goas JY: High incidence of neurologic complications following rapid correction of severe hyponatremia in polydipsic patients. J Clin Psychiatry 55: 349–354, 19948071304
28. Adams RD, Victor M, Mancall EL: Central pontine myelinolysis: A hitherto undescribed disease occurring in alcoholic and malnourished patients. AMA Arch Neurol Psychiatry 81: 154–172, 195913616772
29. Crivellin C, Cagnin A, Manara R, Boccagni P, Cillo U, Feltracco P, Barbieri S, Ferrarese A, Germani G, Russo FP, Burra P, Senzolo M: Risk factors for central pontine and extrapontine myelinolysis after liver transplantation: A single-center study. Transplantation 99: 1257–1264, 201525427166
30. Mohmand HK, Issa D, Ahmad Z, Cappuccio JD, Kouides RW, Sterns RH: Hypertonic saline for hyponatremia: Risk of inadvertent overcorrection. Clin J Am Soc Nephrol 2: 1110–1117, 200717913972
31. Chung HM, Kluge R, Schrier RW, Anderson RJ: Clinical assessment of extracellular fluid volume in hyponatremia. Am J Med 83: 905–908, 19873674097
32. Achinger SG, Ayus JC: Treatment of hyponatremic encephalopathy in the critically ill. Crit Care Med 45: 1762–1771, 201728704229
33. Ayus JC, Caputo D, Bazerque F, Heguilen R, Gonzalez CD, Moritz ML: Treatment of hyponatremic encephalopathy with a 3%
sodium chloride protocol: A case series. Am J Kidney Dis 65: 435–442, 201525465163
34. Bhaskar E, Kumar B, Ramalakshmi S: Evaluation of a protocol for hypertonic saline administration in acute euvolemic symptomatic hyponatremia: A prospective observational trial. Indian J Crit Care Med 14: 170–174, 201021572746
35. Sterns RH, Hix JK, Silver S: Treating profound hyponatremia: A strategy for controlled correction. Am J Kidney Dis 56: 774–779, 201020709440
36. Gankam Kengne F, Soupart A, Pochet R, Brion JP, Decaux G: Re-induction of hyponatremia after rapid overcorrection of hyponatremia reduces mortality in rats. Kidney Int 76: 614–621, 200919606078