Original Contribution
Hyponatremia Is an Independent Predictor of In-Hospital
Mortality in Spontaneous Intracerebral Hemorrhage
Joji B. Kuramatsu, MD*; Tobias Bobinger, MD*; Bastian Volbers, MD; Dimitre Staykov, MD;
Hannes Lücking, MD; Stephan P. Kloska, MD; Martin Köhrmann, MD; Hagen B. Huttner, MD
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Background and Purpose—Hyponatremia is the most frequent electrolyte disturbance in critical care. Across various
disciplines, hyponatremia is associated with increased mortality and longer hospital stay, yet in intracerebral hemorrhage
(ICH) no data are available. This the first study that investigated the prevalence and clinical associations of hyponatremia
in patients with ICH.
Methods—This observational study included all consecutive spontaneous ICH patients (n=464) admitted during a 5-year
period to the Department of Neurology. Patient characteristics, in-hospital measures, mortality, and functional outcome
(90 days and 1 year) were analyzed to determine the effects of hyponatremia (Na <135 mEq/L). Multivariable regression
analyses were calculated for factors associated with hyponatremia and predictors of in-hospital mortality.
Results—The prevalence of hyponatremia on hospital admission was 15.6% (n=66). Normonatremia was achieved and
maintained in almost all hyponatremia patients <48 hours. In-hospital mortality was roughly doubled in hyponatremia
compared with nonhyponatremia patients (40.9%; n=27 versus 21.1%; n=75), translating into a 2.5-fold increased odds
ratio (P<0.001). Multivariable analyses identified hyponatremia as an independent predictor of in-hospital mortality (odds
ratio, 2.2; 95% confidence interval, 1.05–4.62; P=0.037). Within 90 days after ICH, hyponatremia patients surviving
hospital stay were also at greater risk of death (odds ratio, 4.8; 95% confidence interval, 2.1–10.6; P<0.001); thereafter,
mortality rates were similar.
Conclusions—Hyponatremia was identified as an independent predictor of in-hospital mortality with a fairly high prevalence
in spontaneous ICH patients. The presence of hyponatremia at hospital admission is related to an increased short-term
mortality in patients surviving acute care, possibly reflecting a preexisting condition that is linked to worse outcome
due to greater comorbidity. Correction of hyponatremia does not seem to compensate its influence on mortality, which
strongly warrants future research. (Stroke. 2014;45:00-00.)
Key Words: cerebral hemorrhage ◼ hyponatremia
H
yponatremia is the most frequently encountered electrolyte disturbance in patients in critical care and has been
associated with increased mortality and morbidity in various
diseases such as heart failure, cirrhosis, or chronic kidney
disease.1–4 Roughly, one third of these patients present with
preexisting hyponatremia, and there is also a high incidence
of hyponatremia developing during hospital stay.2,3,5 In general, there are no large prospective randomized controlled
trials investigating the clinical impact of hyponatremia. For
neurointensive care patients, only few studies are available
focussing on postoperative neurosurgical patients, traumatic
brain injury, and subarachnoid hemorrhage supporting the
associations of hyponatremia with increased mortality, longer
hospital stay, and raised complications.6–9
Pathophysiological mechanisms that link hyponatremia to
outcome in neurointensive care include the development of
cerebral edema, seizures, and delayed cerebral infarctions.
In these patients, hyponatremia may be caused by the syndrome of inappropriate antidiuretic hormone secretion and
cerebral salt wasting syndrome.6,8,10 These aspects conclude
that hyponatremia may rather be developing than preexisting, possibly considering hyponatremia as bystander just
reflecting the disease severity. Beyond these aspects, recent
research has strengthened our understanding of hyponatremia
itself as a potentially preexisting marker influencing outcome
negatively.11,12
For ischemic or hemorrhagic stroke (intracerebral hemorrhage, ICH), there are no data available. This study is the first
Received November 12, 2013; accepted March 10, 2014.
From the Departments of Neurology (J.B.K., T.B., B.V., D.S., M.K., H.B.H.) and Neuroradiology (H.L., S.P.K.), University of Erlangen-Nuremberg,
Erlangen, Germany.
Guest Editor for this article was Eric E. Smith, MD, MPH.
*Drs Kuramatsu and Bobinger contributed equally.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.
113.004136/-/DC1.
Correspondence to Joji B. Kuramatsu, MD, Departments of Neurology, University of Erlangen-Nuremberg, Schwabachanlage 6, 91054 Erlangen,
Germany. E-mail [email protected]
© 2014 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org
DOI: 10.1161/STROKEAHA.113.004136
1
2 Stroke May 2014
to investigate the prevalence of hyponatremia, its associations
with clinical characteristics, and its influence on mortality and
functional outcome in patients with spontaneous ICH.
Methods
Patient Selection
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
The study was approved by the local ethics committee of
­Friedrich-Alexander University Erlangen-Nuremberg, Germany. All
consecutive spontaneous ICH patients admitted to the Department
of Neurology during a 5-year period (2006–2010) were retrospectively analyzed from our prospective institutional database (n=464).
Secondary ICH etiologies were excluded, that is, ICH related to
anticoagulation, trauma, tumor, arteriovenous malformations,
or ICH after acute thrombolysis or coagulopathy (platelet count
<50.000/μL; International normalized ratio >1.4). Moreover, we
excluded 13 patients that received larger fluid resuscitation (volume >1000 cm3) or osmotherapy (mannitol or hypertonic saline
infusions) before initial laboratory data were available (ie, hospital
transferrals with missing initial laboratory data) as well as 29 patients that were lost to follow-up. Of all analyzed patients 37.4%
(n=158) were transferred from peripheral hospitals, and the remaining 264 patients were direct admissions. Of all transferred patients,
exclusively initial laboratory data from peripheral hospitals were
used. Hence, 422 patients of central European descent remained for
final analysis.
Parameter Acquisition
Data of all patients was retrieved from our institutional prospective
database and reviewed for demographics, previous medical history,
imaging, and in-hospital parameters (Glasgow Coma Scale, National
Institutes of Health Stroke Scale, ICH score, blood pressure, mechanical ventilation, length of stay, external ventricular drainage for hydrocephalus, diagnosis of pneumonia, urinary tract infection, and sepsis
according to established criteria).13–15 Initial laboratory parameters
were selected from our laboratory database. Anemia on admission
was evaluated and defined as previously described.16 Hyponatremia
was defined as Na <135 mEq/L5 and was corrected using 0.9% sodium infusions or hypertonic saline infusions (NaCl, 3%) targeting
an increase of ≤8 mg/dL per 24 hours.
Outcome Measures
Mortality was assessed during hospital stay and followed up until 1
year after hospital discharge. Functional outcome was analyzed at 90
days and 1 year using the modified Rankin Scale (mRS).17 Outcome
measures were assessed by a standardized mailed questionnaire or
semiquantitative phone interview (conducted whenever this questionnaire did not return <2–4 weeks) with patients, relatives, or primary
care physicians. The interviews were performed by 2 scale-trained
physicians certified for data collection on disability and quality of
life (n=313; 74.2% by questionnaires and n=109; 25.8% by telephone
interviews). Favorable outcome was defined as mRS=0 to 3 versus
mRS=4 to 6.
Imaging
Diagnosis was made either by CT (SOMATOM Volume Zoom or
SOMATOM Definition AS+; Siemens, Erlangen, Germany) or MRI
(Sonata; 1.5 T; Siemens). Two neuroradiologists blinded to clinical
data reviewed the scans independently, and in cases of discrepancies,
a second consensus analysis was made. Hematoma growth (volume
>33%) was determined on follow-up imaging, which was performed
<24±6 hours.18 Parenchymal ICH volume was calculated by the formula of ellipsoids (ABC/2), and different imaging modalities were
compared by a validated conversion model.19,20 The Graeb score summation was used to score the extent of ventricular involvement.21
Cerebral perihemorrhagic edema evolution was measured and quantified semiautomatically as described previously.22
Statistical Analysis
Statistical analysis was performed with SPSS version 19.0 (SPSS
Inc). The Kolmogorov–Smirnov and Shapiro–Wilk test was applied
to determine the distribution of the data. The latter are presented as
mean±SD (compared using the Student t test) or as median and interquartile range (compared using the Mann–Whitney U test) as appropriate. The Pearson χ2 and Fisher exact tests were used to compare
frequency distributions of categorized variables between hyponatremia and nonhyponatremia patients. The significance level, which was
set at α=0.05, was corrected for multiple comparisons (type I error)
by the Bonferroni method. Statistical tests were 2-sided. The relationship between hyponatremia and mortality was calculated using
a log-rank test and visualized by Kaplan–Meier curve. Stepwise forward inclusion multivariable analyses were calculated to investigate
(1) factors associated with hyponatremia (for more details, see Table I
in the online-only Data Supplement) and (2) predictors for in-hospital
mortality. Parameters reaching a statistical trend in univariable analysis (ie, P<0.1) were included into the multivariable models.
Results
Prevalence of Hyponatremia, Associated Factors,
and Course of Sodium Levels
The prevalence of hyponatremia (Na <135 mEq/L) on hospital admission was 15.6% (n=66 of 422 ICH patients). Patient
transferal status (transferred versus direct admissions) did not
influence the prevalence of hyponatremia (transferred 15.2%
[n=24/158] versus direct admission 15.9% [n=42/264]),
and there was no association of serum sodium levels with
time from symptom onset until admission (Figure I in the
­online-only Data Supplement). Evaluated baseline characteristics and previous medical history are presented in Table 1.
Hyponatremia patients were showing statistical trends to a
worse preadmission mRS (P=0.081), increased use of thiazide diuretics (P=0.095), and more frequent preexisting liver
disease (P=0.005). However, all of these parameters did not
reveal a significant association with hyponatremia on logistic
regression modeling.
Table 2 shows the comparison of clinical parameters
between patients with and without hyponatremia documenting associations for hyponatremia with reduced osmolarity, decreased hemoglobin and hematocrit levels, as well
as increased ICH score (all P<0.001). The assessment of
radiological parameters revealed trends (after statistical correction) for an increased rate of intraventricular hemorrhage
(IVH; hyponatremia 72.6% versus nonhyponatremia 55.6%;
P=0.010) with higher Graeb scores (P=0.008). We hypothesized that IVH could lead to progressive hyponatremia over
time due to the compression of midline cerebral structures
including the pituitary gland; however, increasing time to
presentation was not associated with sodium level in patients
with or without IVH, and higher Graeb score was not associated with hyponatremia (P>0.05; Figure II in the online-only
Data Supplement). To rule out that the increased occurrence
of IVH among the group of hyponatremia patients reflects
a major reason for the overall increased mortality of hyponatremia patients, Kaplan–Meier survival curves revealed
that hyponatremia remained significantly associated with
mortality despite the association of IVH with mortality
(Figure III in the online-only Data Supplement). Focussing
on specific complications that are supposed to be related to
Kuramatsu et al Hyponatremia in Spontaneous ICH 3
Table 1. Demographics and Previous Medical History
Spontaneous ICH (n=422)
Age*, y
Sex†
Hyponatremia
(n=66)
Nonhyponatremia
(n=356)
P Value
69.8±13.6
69.6±11.7
0.928
31 (47.0%)
159 (44.7%)
0.729
1 (0–2)
0.081
Previous medical history
Pre-mRS‡
BMI‡
1 (0–3)
27.6 (24.1–30.8)
24.4 (23.1–30.4)
0.231
288 (80.9%)
0.13
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Hypertension†
48 (72.7%)
Diabetes mellitus†
15 (22.7%)
72 (20.2%)
0.647
Hypercholesterolemia†
17 (25.8%)
106 (29.8%)
0.507
Atrial fibrillation†
10 (15.2%)
44 (12.4%)
0.532
Previous stroke†
13 (19.7%)
84 (23.6%)
0.488
Cardiac event†
12 (18.2%)
51 (14.3%)
0.42
Alcohol abuse†
12 (18.2%)
67 (18.8%)
0.92
Smoking†
21 (31.8%)
124 (34.8%)
0.639
3 (4.6%)
14 (3.9%)
0.733
History of seizures†
History of liver disease†
History of renal
insufficiency†
Heart failure†
Use of thiazide diuretics†
Hypothyreodism†
6 (9.1%)
6 (1.7%)
0.005§
11 (16.7%)
53 (14.9%)
0.708
0.61
4 (6.1%)
28 (7.9%)
19 (28.8%)
70 (19.7%)
0.095
7 (10.6%)
25 (7.0%)
0.312
BMI indicates body mass index; ICH, intracerebral hemorrhage; and mRS,
modified Rankin Scale.
*Mean±SD.
†n (%).
‡Median (interquartile range; 25th–75th percentile).
§Not significant after Bonferroni correction.
hyponatremia, that is, edema formation, seizures, and infections, no significant differences could be elucidated between
the compared groups.
The 14-day course of median serum sodium levels is shown
in Figure 1. In >50% of patients, hyponatremia was corrected
to normal range levels <24 hours. Sodium levels remained significantly different within the first 48 hours after admission.
In essence, preexisting hyponatremia was corrected early, and
thereafter median sodium levels remained normonatremic
during hospital stay both in patients with and without initial
hyponatremia.
Mortality
Graphical analysis (Kaplan–Meier survival curve from day 0
until day 90 [ie, <90 d]) is shown in Figure 2. The median
length of hospital stay was 11 days (95% confidence interval
[CI], 1–28 days), and mortality difference at this time point
is depicted by the dotted line. Hence, all-cause in-hospital
mortality was roughly doubled (40.9% [n=27 in hyponatremia] versus 21.1% [n=75 in nonhyponatremia]; P<0.001). The
evaluation of short-term mortality rates at 90 days revealed that
overall 59.1% (n=39) of patients with hyponatremia versus
27.8% (n=99) of nonhyponatremia had died, translating into
≈4-fold increased 90-day mortality (odds ratio [OR], 3.8; 95%
CI, 2.2–6.5; P<0.001). Of note, hyponatremia patients received
significantly more often do-not-treat/do-not-resuscitate orders
(18.2% versus 8.9%; P=0.025). To rule out that increased
d­o-not-treat/do-not-resuscitate orders among the group of
hyponatremia patients represented a major reason for the overall increased mortality in hyponatremia patients, all patients
that received these orders were excluded for subanalyses. The
presence of hyponatremia still remained significantly associated with mortality (OR, 2.8; 95% CI, 1.5–5.5; P=0.0015; for
Kaplan–Meier survival curves after exclusion of those patients
with withdrawal/withhold orders, see Figure IV in the onlineonly Data Supplement). Though withdrawal/withhold orders
were more frequent in hyponatremia patients, there was no significant impact of these orders on increased in-hospital mortality of hyponatremia patients.
Functional Outcome
Short- and long-term functional outcome for the entire
cohort is shown in Figure 3. The proportion of patients with
a favorable functional outcome was different among patients
with and without hyponatremia (90 days: P=0.072; 1 year:
P=0.033). Although functional outcome appeared to be driven
mainly by in-hospital mortality, the fraction of hospital stay–
surviving patients who died thereafter but before 90 days was
significantly higher in hyponatremia patients (18.2% versus
6.7%; OR, 4.8; 95% CI, 2.1–10.6; P<0.001). In contrast, the
rate of 3-month survivors who died during 1-year follow-up
was essentially the same in both groups (4.5% versus 6.5%;
OR, 1.3; 95% CI, 0.4–4.5; P=0.723). This highlights that
the significant difference in long-term functional outcome is
mainly attributed to a higher in-hospital mortality and an early
postdischarge (<90 days) mortality.
Hyponatremia Is Independently Associated With
In-Hospital Mortality
In multivariable-adjusted analyses, hyponatremia was independently associated with increased ICH score (P=0.03) and
with the presence of anemia at hospital admission (P=0.005;
see Table I in the online-only Data Supplement). Multivariable
analyses of predictors of in-hospital mortality showed that
hyponatremia was independently associated with mortality
(OR, 2.2; 95% CI, 1.05–4.6; P=0.04) controlling for age, ICH
score, ICH volume, and IVH (Table 3). Adding preexisting
liver disease to the model did not influence the independent
association of hyponatremia with in-hospital mortality (data
not shown).
Discussion
This study represents the first analysis of hyponatremia in ICH
patients and revealed that hyponatremia is a condition frequently associated with increased in-hospital and s­ hort-term
mortality. Moreover, hyponatremia was identified as previously unrecognized independent predictor of in-hospital mortality. Some aspects emerge from the data.
The prevalence of hyponatremia in an ambulatory hospital care setting ranges between 11% to 21% and increases to
28.2% in acutely hospitalized patients.12,23 In neurointensive
care, as shown here for ICH, 15.6% of patients presented with
hyponatremia on admission, mirroring previous investigations
of elderly community-dwelling populations.24 Existing data on
4 Stroke May 2014
Table 2. Neurological Admission Status, Neuroradiological, Laboratory Data, and In-Hospital
Measures for all ICH Patients With Hyponatremia vs Nonhyponatremia
Spontaneous ICH (n=422)
Hyponatremia (n=66)
Nonhyponatremia (n=356)
P Value
241 (93.8–811.8)
216 (99.8–629.8)
0.476
Admission status
Symptom onset until arrival*, min
Intubated on arrival†
25 (37.9%)
114 (32.0%)
0.354
NIHSS*
19 (11–32)
16 (6–26)
0.002§
Glasgow Coma Scale*
10 (3–14)
13 (5–15)
0.010§
3 (1–4)
2 (1–3)
0.001
ICH score*
MAP‡, mm Hg
111±28.6
116 ±27.4
0.313
Neuroradiological data
Basalganglia hemorrhage†
ICH volume*, cm3
IVH†
Graeb score*
Relative perihemorrhagic edema‡
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Midline shift*, mm
39 (59.1%)
23.4 (7.93–60.68)
220 (61.8%)
0.682
17.7 (5.83–50.13)
0.169
48 (72.7%)
198 (55.6%)
0.010§
4 (0–7)
1 (0–5)
0.008§
1.89±1.68
1.83±1.49
3 (0–6)
2 (0–7)
0.84
0.481
Laboratory values on admission
Sodium*, mg/dL
131.2 (129.8–133.0)
139.0 (137.0–141.0)
<0.001
Osmolarity*, mosmol/kg
276.9 (270.6–280.6)
291.3 (287.5–295.6)
<0.001
9.19 (7.29–12.50)
9.30 (6.96–11.65)
0.706
Leucocytes*, 109/L
Hemoglobin‡, mmol/L
7.96±1.47
8.58±1.17
<0.001
36.91±6.65
40.14±5.48
<0.001
Thrombocytes*, 109/L
221 (173–274)
232 (192–289)
0.234
Creatinine*, μmol/L
0.89 (0.71–1.20)
0.92 (0.78–1.12)
0.559
Hematocrit‡, %
Urea*, mmol/L
12.14 (8.21–16.96)
12.14 (9.64–14.99)
0.552
C-reactive protein*, nmol/L
5.5 (1.5–12.5)
3.1 (1.0–8.0)
0.165
Glucose*, mmol/L
135 (105–175)
134 (112–168)
0.664
HbA1c*, %
5.8 (5.3–6.1)
5.8 (5.5–6.3)
0.211
1.07 (0.99–1.13)
1.01 (0.97–1.07)
0.783
International normalized ratio*
AST*, U/L
28 (23–40)
29 (25–37)
0.663
ALT*, U/L
20 (15–29)
22 (17–30)
0.525
GGT*, U/L
28 (17–44)
31 (20–61)
0.268
11 (6.0–19.2)
0.252
In-hospital measures
Length of stay in hospital*, d
10 (2.8–18.1)
Mechanical ventilation†
32 (48.5%)
151 (42.4%)
0.361
Tracheotomy†
21 (31.8%)
126 (35.4%)
0.577
PEG†
14 (21.2%)
61 (17.1%)
0.427
Pneumonia†
35 (53.0%)
203 (57.0%)
0.549
Sepsis†
13 (19.7%)
42 (11.8%)
0.079
7 (10.6%)
46 (12.9%)
0.603
External ventricular drainage†
19 (28.8%)
96 (27.0%)
0.76
Seizures during hospital stay†
6 (9.1%)
30 (8.4%)
0.862
Urinary tract infection†
ALT indicates alanine aminotransferase; AST, aspartate aminotransferase; GGT, gamma-glutamyl transferase; ICH, intracerebral
hemorrhage; IVH, intraventricular hemorrhage; MAP, mean arterial pressure; NIHSS, National Institutes of Health Stroke Scale; and
PEG, percutaneous endoscopic gastrostomy.
*Median (interquartile range; 25th–75th percentile).
†n (%).
‡Mean±SD.
§Not significant after Bonferroni correction.
elderly patients propose chronically increased plasma vasopressin (arginine vasopressin, AVP) concentrations to account
for hyponatremia, which is further complicated by impaired
water excretory capacity, more frequent comorbidities, and
polymedication.23,25 In line with existing data on subarachnoid
hemorrhage or traumatic brain injury in which hyponatremia
Kuramatsu et al Hyponatremia in Spontaneous ICH 5
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Figure 1. Graphical depiction of median serum sodium levels over a 14-day period and corresponding interquartile range (25th–75th percentile) for all patients with intracerebral hemorrhage with and without hyponatremia (HN).
was linked to disease severity,6 hyponatremia patients showed
higher ICH scores and were more frequently anemic on
admission, which has recently been independently associated
with poorer functional long-term outcome.16 Because there
are no data from experimental or human studies reporting on
hyperacute (minutes to hours) development of hyponatremia
after neurological injury, we think that hyponatremia at hospital presentation rather reflects a preexisting condition (Figures
I and II in the online-only Data Supplement).
Hyponatremia has gained wide recognition as a factor
being associated with a negative prognosis in a variety of
conditions.1,3,26,27 In patients with decompensated heart failure, hyponatremia was present in 38% and went along with
increased mortality, frequent hospital readmissions, and worse
quality of life.1,3,4 A postulated mechanism of hyponatremia
influencing outcome also in neurointensive care could be its
relation to an increased nonosmotic release of AVP caused by
the overstimulation of the neurohumoral axis or by baroreceptor failure in elderly patients leading to cerebral hypoperfusion.11,28 After severe neurotrauma such as ICH, increased
intracranial pressures and cerebral hypoperfusion may even
aggravate secondary brain injury. Assuming that these mechanisms play a role in patients with ICH, increased AVP levels
themselves could contribute to a reduced tolerance to cerebral
hypoxia.29 Potentially supporting this association in ICH, previous studies on the neurohormone copeptin, the C-terminal
portion of provasopressin, which is cosecreted with AVP
from the hypothalamus,30 report that increased copeptin was
a marker of poorer outcome, which was shown in subarachnoid hemorrhage, ischemic stroke, and also in patients with
Figure 2. Kaplan–Meier survival curve from
day 0 until day 90 (ie, <90 d) in patients with
and without hyponatremia (HN). ICH indicates
intracerebral hemorrhage. Length of stay (LOS)
is presented as horizontal bold line with its
corresponding 95% confidence interval (length
of horizontal line); the median LOS is presented as vertical dotted line.
6 Stroke May 2014
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Figure 3. Dichotomized analysis of the modified Rankin Scale (mRS) distribution at 90 d and 1 y for all patients with and without hyponatremia (HN). Favorable (mRS=0–3 vs mRS=4–6) functional outcome is shown as bold line. In-hospital mortality (IH-M) and mortality at 90
d and 1 y are shown as dotted lines.
ICH.31–33 Despite these hypothetical pathophysiological considerations of hyponatremia-mediated effects on outcome, it
remains unresolved whether hyponatremia simply represents a
bystander of disease, that is, preexisting comorbidity or polymedication reflecting poorer overall status, which itself may
be responsible for worse outcome. This aspect is even more
important in light of the findings presented here, because all
hyponatremia patients presented with hyponatremia on hospital admission rather than developed it during clinical course.
Finally, up to now (and including this study), there is no
evidence across various disciplines and diseases that correcting hyponatremia could result in improved clinical end
Table 3. Multivariable Model for Predictors of In-Hospital
Mortality
Multivariable Parameters
Mortality Predictors,
OR (95% CI)
P Value (P<0.05)
Age
1.033 (1.003–1.065)
0.0290
<0.0001
ICH Score
1.924 (1.391–2.661)
ICH volume
1.015 (1.006–1.024)
0.0014
IVH (Graeb score)
1.137 (1.032–1.253)
0.0095
Ventilation
0.520 (0.249–1.086)
0.0819
Diagnosis of diabetes
mellitus
0.666 (0.315–1.406)
0.2865
Anemia
1.242 (0.643–2.397)
0.5190
Hyponatremia
2.199 (1.048–4.617)
0.0372
CI indicates confidence interval; ICH, intracerebral hemorrhage; IVH,
intraventricular hemorrhage; and OR, odds ratio.
points.1,11,12 If AVP-mediated hyponatremia would be relevant
in ICH, it has to be mentioned that a randomized controlled
trial investigating treatment with vasopressin antagonist
tolvaptan, on the one hand, revealed significantly increased
sodium levels but, on the other hand, was not able to show
effects on mortality.34 From a neurological perspective beyond
pontine myelinolysis, the correction of hyponatremia could
even be harmful because cerebral adaption mechanisms may
not work properly after neurological injury (ie, hypoxia),
resulting in demyelinizations even at normal serum sodium
levels.11,12 Vice versa, recent studies on cerebral edema evolution report on continuous hypertonic saline infusions with
upper limit sodium levels targeted to be safe and possibly even
beneficial in ICH.35
Our study has several shortcomings, mainly its retrospective, single-center design that left this study unable to answer
why hyponatremia patients showed increased in-hospital mortality. We have to assume that a precise analysis of preexisting
comorbidities and polymedication possibly related to hyponatremia and outcome may have not been sufficiently achieved.
This was because logistic regression analyses for these parameters did not reveal significant associations with hyponatremia and mortality. Moreover, the assessment of exact volume
status at hospital admission (eu-, hypo-, or hypervolemic)
and urine analysis was not realized because of the study
design. Hypovolemic hyponatremia from post-ICH dehydration seems to be unlikely, because there was no correlation
between sodium level and time from ICH symptom onset to
admission (Figure I in the online-only Data Supplement). In
Kuramatsu et al Hyponatremia in Spontaneous ICH 7
addition, the mailed questionnaires may have been answered
wrongly with respect to the time point of follow-up and validity of mRS estimation,36 and results may have been biased
by patients lost to follow-up. Finally, correction for multiple
comparisons (Bonferroni) may have decreased the chances of
detecting false-negative but true associations.
In summary, hyponatremia is a common electrolyte abnormality in patients with spontaneous ICH and an independent
predictor of in-hospital mortality. Hyponatremia seems to be
a preexisting condition rather than developing acutely, being
independently associated with anemia on admission. Correction
of hyponatremia is appropriate although the mechanism of the
association between hyponatremia and mortality is unclear.
Acknowledgments
We thank Friederike Eichhorn for helping with clinical data acquisition, and Dr Inken Martin for valuable discussions.
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Disclosures
Dr Kuramatsu has received travel grants and speaker’s honoraria from
Otsuka. The other authors have no conflicts to report. All authors have
read the article, agree with the contents, and approved the final version of the article.
References
1. Kovesdy CP, Lott EH, Lu JL, Malakauskas SM, Ma JZ, Molnar MZ, et al.
Hyponatremia, hypernatremia, and mortality in patients with chronic
kidney disease with and without congestive heart failure. Circulation.
2012;125:677–684.
2. Funk GC, Lindner G, Druml W, Metnitz B, Schwarz C, Bauer P, et al.
Incidence and prognosis of dysnatremias present on ICU admission.
Intensive Care Med. 2010;36:304–311.
3.Wald R, Jaber BL, Price LL, Upadhyay A, Madias NE. Impact of
hospital-associated hyponatremia on selected outcomes. Arch Intern
­
Med. 2010;170:294–302.
4. Klein L, O’Connor CM, Leimberger JD, Gattis-Stough W, Pina IL, Felker
GM, et al. Lower serum sodium is associated with increased short-term mortality in hospitalized patients with worsening heart failure: Results from the
outcomes of a prospective trial of intravenous milrinone for exacerbations of
chronic heart failure (optime-chf) study. Circulation. 2005;111:2454–2460
5. DeVita MV, Gardenswartz MH, Konecky A, Zabetakis PM. Incidence
and etiology of hyponatremia in an intensive care unit. Clin Nephrol.
1990;34:163–166.
6. Rahman M, Friedman WA. Hyponatremia in neurosurgical patients:
clinical guidelines development. Neurosurgery. 2009;65:925–935, discussion 935.
7. Moro N, Katayama Y, Igarashi T, Mori T, Kawamata T, Kojima J.
Hyponatremia in patients with traumatic brain injury: incidence, mechanism, and response to sodium supplementation or retention therapy with
hydrocortisone. Surg Neurol. 2007;68:387–393.
8. Hasan D, Wijdicks EF, Vermeulen M. Hyponatremia is associated with
cerebral ischemia in patients with aneurysmal subarachnoid hemorrhage.
Ann Neurol. 1990;27:106–108.
9. Lehmann L, Bendel S, Uehlinger DE, Takala J, Schafer M, Reinert M,
et al. Randomized, double-blind trial of the effect of fluid composition
on electrolyte, acid-base, and fluid homeostasis in patients early after
subarachnoid hemorrhage. Neurocrit Care. 2013;18:5–12.
10. Diringer M, Ladenson PW, Borel C, Hart GK, Kirsch JR, Hanley DF.
Sodium and water regulation in a patient with cerebral salt wasting. Arch
Neurol. 1989;46:928–930.
11. Schrier RW, Sharma S, Shchekochikhin D. Hyponatraemia: more than
just a marker of disease severity? Nat Rev Nephrol. 2013;9:124.
12. Hoorn EJ, Zietse R. Hyponatremia and mortality: moving beyond associations. Am J Kidney Dis. 2013;62:139–149.
13.Tejerina E, Esteban A, Fernández-Segoviano P, Frutos-Vivar F,
Aramburu J, Ballesteros D, et al. Accuracy of clinical definitions of
ventilator-associated pneumonia: comparison with autopsy findings.
­
J Crit Care. 2010;25:62–68.
14. Muckart DJ, Bhagwanjee S. American College of Chest Physicians/
Society of Critical Care Medicine Consensus Conference definitions of
the systemic inflammatory response syndrome and allied disorders in
relation to critically injured patients. Crit Care Med. 1997;25:1789–1795.
15. Wilson ML, Gaido L. Laboratory diagnosis of urinary tract infections in
adult patients. Clin Infect Dis. 2004;38:1150–1158.
16.Kuramatsu JB, Gerner S, Lucking H, Kloska SP, Schellinger PD,
Kohrmann M, et al. Anemia is an independent prognostic factor in
intracerebral hemorrhage: an observational cohort study. Crit Care.
2013;17:R148.
17. Rankin J. Cerebral vascular accidents in patients over the age of 60. I.
General considerations. Scott Med J. 1957;2:127–136.
18. Brott T, Broderick J, Kothari R, Barsan W, Tomsick T, Sauerbeck L, et al.
Early hemorrhage growth in patients with intracerebral hemorrhage.
Stroke. 1997;28:1–5.
19.Kothari RU, Brott T, Broderick JP, Barsan WG, Sauerbeck LR,
Zuccarello M, et al. The ABCs of measuring intracerebral hemorrhage
volumes. Stroke. 1996;27:1304–1305.
20. Burgess RE, Warach S, Schaewe TJ, Copenhaver BR, Alger JR, Vespa
P, et al. Development and validation of a simple conversion model for
comparison of intracerebral hemorrhage volumes measured on CT and
gradient recalled echo MRI. Stroke. 2008;39:2017–2020.
21. Graeb DA, Robertson WD, Lapointe JS, Nugent RA, Harrison PB.
Computed tomographic diagnosis of intraventricular hemorrhage.
Etiology and prognosis. Radiology. 1982;143:91–96.
22. Volbers B, Wagner I, Willfarth W, Doerfler A, Schwab S, Staykov D.
Intraventricular fibrinolysis does not increase perihemorrhagic edema
after intracerebral hemorrhage. Stroke. 2013;44:362–366.
23. Upadhyay A, Jaber BL, Madias NE. Incidence and prevalence of hyponatremia. Am J Med. 2006;119(7 suppl 1):S30–S35.
24. Miller M, Morley JE, Rubenstein LZ. Hyponatremia in a nursing home
population. J Am Geriatr Soc. 1995;43:1410–1413.
25. Chawla A, Sterns RH, Nigwekar SU, Cappuccio JD. Mortality and
serum sodium: do patients die from or with hyponatremia? Clin J Am
Soc Nephrol. 2011;6:960–965.
26.Ginès P, Schrier RW. Renal failure in cirrhosis. N Engl J Med.
2009;361:1279–1290.
27. Hoorn EJ, Lindemans J, Zietse R. Acute and concomitant deterioration
of hyponatremia and renal dysfunction associated with heart and liver
failure. Clin Nephrol. 2006;65:248–255.
28. Brimioulle S, Orellana-Jimenez C, Aminian A, Vincent JL. Hyponatremia
in neurological patients: cerebral salt wasting versus inappropriate antidiuretic hormone secretion. Intensive Care Med. 2008;34:125–131.
29. Wallace BK, Foroutan S, O’Donnell ME. Ischemia-induced stimulation of Na-K-Cl cotransport in cerebral microvascular endothelial cells
involves AMP kinase. Am J Physiol Cell Physiol. 2011;301:C316–C326.
30. Yalta K, Yalta T, Sivri N, Yetkin E. Copeptin and cardiovascular disease:
a review of a novel neurohormone. Int J Cardiol. 2013;167:1750–1759.
31. Fung C, De Marchis GM, Katan M, Seiler M, Arnold M, Gralla J, et al.
Copeptin as a marker for severity and prognosis of aneurysmal subarachnoid hemorrhage. PLoS One. 2013;8:e53191.
32. Dong X, Tao DB, Wang YX, Cao H, Xu YS, Wang QY. Plasma copeptin
levels in Chinese patients with acute ischemic stroke: a preliminary
study. Neurol Sci. 2013;34:1591–1595.
33. Zhang A, Li J, Li X, Song L, Li H. The prognostic value of copeptin for
acute intracerebral hemorrhage patients. Exp Ther Med. 2013;5:467–470.
34. Konstam MA, Gheorghiade M, Burnett JC Jr, Grinfeld L, Maggioni AP,
Swedberg K, et al; Efficacy of Vasopressin Antagonism in Heart Failure
Outcome Study With Tolvaptan (EVEREST) Investigators. Effects of
oral tolvaptan in patients hospitalized for worsening heart failure: the
EVEREST Outcome Trial. JAMA. 2007;297:1319–1331.
35. Wagner I, Hauer EM, Staykov D, Volbers B, Dörfler A, Schwab S, et al.
Effects of continuous hypertonic saline infusion on perihemorrhagic
edema evolution. Stroke. 2011;42:1540–1545.
36. Banks JL, Marotta CA. Outcomes validity and reliability of the modified
Rankin scale: implications for stroke clinical trials: a literature review
and synthesis. Stroke. 2007;38:1091–1096.
Hyponatremia Is an Independent Predictor of In-Hospital Mortality in Spontaneous
Intracerebral Hemorrhage
Joji B. Kuramatsu, Tobias Bobinger, Bastian Volbers, Dimitre Staykov, Hannes Lücking,
Stephan P. Kloska, Martin Köhrmann and Hagen B. Huttner
Downloaded from http://stroke.ahajournals.org/ by guest on June 18, 2017
Stroke. published online April 8, 2014;
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2014 American Heart Association, Inc. All rights reserved.
Print ISSN: 0039-2499. Online ISSN: 1524-4628
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://stroke.ahajournals.org/content/early/2014/04/08/STROKEAHA.113.004136
Data Supplement (unedited) at:
http://stroke.ahajournals.org/content/suppl/2014/04/08/STROKEAHA.113.004136.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click
Request Permissions in the middle column of the Web page under Services. Further information about this
process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Stroke is online at:
http://stroke.ahajournals.org//subscriptions/
SUPPLEMENTAL MATERIAL
Supplemental Title: Analyses for the associations of HN
Supplemental methods
Step-wise forward inclusion multivariable analysis was calculated to investigate factors that
were independently associated with HN. Parameters reaching a statistical trend in univariable
analysis (i.e. p<0.1) were included into the multivariable model. Moreover, graphical analyses
were calculated for associations of HN. Two linear regression models were calculated with
Serum sodium levels as dependent variable and time from symptom onset to admission as
explanatory variable. The regression coefficient (R²) is shown within the figures and
delineates the predictive association of the explanatory variables (0 = poor association; 1 =
excellent association). Further, three Kaplan-Meier survival curves were calculated using a
log-rank test for analysis of mortality over 90 days moreover including the statistical analysis
of HN with in-hospital mortality presented as: Length of stay [LOS] as horizontal bold line
with its corresponding 95% confidence interval (length of horizontal line), the median LOS is
presented as vertical dotted line.
Supplemental Table I
multivariable - parameters
Multivariable model for factors independently associated with HN.
Associations with HN
p-value
Odds ratio (95% CI)
(p<0.05)
ICH Score
1.511 (1.046-2.183)
0.0278
ICH Volume
0.996 (0.985-1.007)
0.4663
IVH (Graeb Score)
0.993 (0.885-1.113)
0.9008
Pre-morbid mRS
1.173 (0.894-1.540)
0.2486
Diagnosis of Diabetes
0.624 (0.232-1.678)
0.3500
Diagnosis of Liver disease
5.852 (0.625-54.805)
0.1216
Use of Thiazide diuretics
1.782 (0.276-11.522)
0.5438
Anemia
3.151 (1.406-7.063)
0.0053
Multivariable logistic regression model of parameters associated with HN. All parameters that
showed a statistical trend (i.e. p < 0.1) in univariable testing were included into this stepwise
forward regression model. Significant parameters after adjustment are expressed in bold.
Supplemental Figure I Association of serum sodium levels with time from symptom onset
until admission.
Serum sodium levels at hospital admission are graphically analyzed by linear regression
analysis to determine the potential influence of poor per oral fluid intake at ictus over time
(time from symptom onset until admission).
Supplemental Figure II The influence of symptom onset until admission on the development
of HN in regards to the presence of IVH.
Serum sodium levels at hospital admission are graphically analyzed by linear regression
analysis to determine the influence of IVH on HN over time (time from symptom onset until
admission). Patients with present IVH are displayed as black diamonds; Patients without IVH
are displayed as grey circles.
Supplemental Figure III Survival analysis for patients with and without IVH dichotomized
according to HN status.
Kaplan-Meier survival curve over 90 days in patients with and without IVH dichotomized
according to HN status; Length of stay [LOS] is presented as horizontal bold line with its
corresponding 95% confidence interval (length of horizontal line), the median LOS is
presented as vertical dotted line.
Supplemental Figure IV Survival analysis for patients with and without HN excluding
patients that received “withdrawal/withhold of care” orders.
Kaplan-Meier survival curve over 90 days in patients with and without HN excluding patients
with “withdrawal/withhold of care” orders ; Length of stay [LOS] is presented as horizontal
bold line with its corresponding 95% confidence interval (length of horizontal line), the
median LOS is presented as vertical dotted line.