Influence of Raised Plasma Osmolality on Clinical Outcome
After Acute Stroke
Ajay Bhalla, MRCP; Suki Sankaralingam, MSc; Ruth Dundas, MSc; R. Swaminathan, FRCPath;
Charles D.A. Wolfe, MD; Anthony G. Rudd, FRCP
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Background and Purpose—Abnormal physiological parameters after acute stroke may induce early neurological
deterioration. Studies of the effect of dehydration on stroke outcome are limited. We examined the association of raised
plasma osmolality on stroke outcome at 3 months and the change of plasma osmolality with hydration during the first
week after stroke.
Methods—Acute stroke patients had their plasma osmolality measured at admission and at days 1, 3, and 7. Maximum
plasma osmolality and the area under curve (AUC) were also calculated during the first week. Patients were stratified
according to how they were hydrated: orally, intravenously, or both. Outcome included survival at 3 months after stroke.
Logistic regression was performed to examine the association between raised plasma osmolality (⬎296 mOsm/kg) and
survival, adjusting for stroke severity. Linear regression was performed to examine the pattern of plasma osmolality
across hydration groups.
Results—One hundred sixty-seven patients were included. Mean admission (300 mOsm/kg, SD 11.4), maximum (308.1
mOsm/kg, SD 17.1), and AUC (298.3 mOsm/kg, SD 11.7) plasma osmolality were significantly higher in those who
died compared with survivors (293.1 mOsm/kg [SD 8.2], 297.7 mOsm/kg [SD 8.7], and 291.7 mOsm/kg [SD 8.1],
respectively; P⬍0.0001). Admission plasma osmolality ⬎296 mOsm/kg was significantly associated with mortality
(OR 2.4, 95% CI 1.0 to 5.9). In patients hydrated intravenously, there was no significant fall in plasma osmolality
compared with patients hydrated orally (P⫽0.68).
Conclusions—Raised plasma osmolality on admission is associated with stroke mortality, after correcting for case mix.
Correction of dehydration after stroke requires a more systematic approach. Trials are required to determine whether
correcting dehydration after stroke improves outcome. (Stroke. 2000;31:2043-2048.)
Key Words: stroke outcome 䡲 cerebrovascular disorders 䡲 medical management
T
here is evidence that organized management in stroke
units improves survival and reduces dependency.1 The
precise factors that influence this benefit are unknown and
remain under investigation. One particular theme that has
emerged from stroke unit trials is that there appear to be
differences in the way acute physiology (such as temperature,
blood pressure, and hydration) are managed between these
units and nonstroke units.2 How this improves survival is
unclear, but parenteral fluids may have reduced the occurrence of dehydration and maintained systemic blood pressure
after acute stroke.3,4 Studies of electrolyte imbalance after
stroke are not extensive, and it remains unclear whether initial
hydration status influences mortality or functional recovery.5,6 Plasma osmolality may be a more useful marker of
hydration status compared with conventional biochemical
measurements such as plasma sodium or urea.7 The aim of
this study was to measure the effect of dehydration status of
acute stroke patients on outcome and to examine the pattern
of change of plasma osmolality with hydration.
Subjects and Methods
All patients with a clinical diagnosis of stroke (WHO criteria8),
admitted to St Thomas’ Hospital from November 1, 1998, to
November 1, 1999, were identified prospectively. Patients with
subarachnoid hemorrhage were excluded. The criterion for inclusion
for the study was that stroke onset could be accurately determined
and that blood samples were taken within 24 hours of stroke onset.
Informed written consent was obtained from the patients, their
caregivers, or both. The study was approved by the Guy’s and St
Thomas’ Hospital Ethical Committee.
Information collected at initial assessment included the following:
(1) sociodemographic characteristics (age, sex, and Barthel Index
score prior to stroke).9 (2) Case severity (clinical state at the time of
admission to hospital included level of consciousness [Glasgow
Coma scale] and the presence of dysphagia, dysphasia, and incontinence). (3) Comorbidity (a history of hypertension, myocardial
infarction, diabetes mellitus, atrial fibrillation, and previous cerebro-
Received February 18, 2000; final revision received May 23, 2000; accepted June 6, 2000.
From the Department of Public Health Sciences, Guy’s, King’s and St Thomas’ Hospital School of Medicine (A.B., R.D., C.D.A.W.), and the
Departments of Chemical Pathology (S.S., R.S.) and Elderly Care (A.G.R.), Guy’s and St Thomas’ Hospital, London, England.
Correspondence to Dr Ajay Bhalla, Department of Public Health Sciences, Guy’s, King’s and St Thomas’ Hospital School of Medicine, Capital House,
42 Weston St, London, SE1 7EH UK. E-mail [email protected]
© 2000 American Heart Association, Inc.
Stroke is available at http://www.strokeaha.org
2043
2044
Stroke
September 2000
vascular disease was identified through hospital and general practitioner records; use of diuretics prior to admission was also recorded).
Stroke subtype was determined according to the Bamford
Classification.10
Every patient had a standardized swallowing assessment within 24
hours of admission by a speech therapist or by experienced nursing
staff to ascertain the presence of dysphagia. The swallow test was not
attempted in a number of circumstances in which the patients were
drowsy, did not open their eyes to speech, or were drooling. Five
milliliters water was given to the patient, while 2 fingers were placed
above and below the larynx to feel the swallow reflex. If coughing or
choking occurred after 2 minutes or there was an absent swallow,
this indicated a failed test. If this procedure was successful, the test
was repeated with one-third-filled glass of water, observing for the
same signs for a failed test.
Analysis
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Blood samples were taken within 24 hours of stroke onset (admission) day 1 and 3 and 7 days after stroke. All patients had plasma
osmolality, serum sodium, and urea measured at these time points.
Plasma glucose was also measured within 24 hours of stroke onset.
Patients were recumbent for at least 30 minutes before samples were
taken. Samples were collected in lithium heparinized containers and
transferred to the chemical pathology laboratory for further processing within 30 minutes. Plasma samples were frozen at 4°C and then
centrifuged. The osmolality of the plasma was determined with the
depression of freezing point method (Advanced Micro, Advanced
Instruments). Calculated plasma osmolality was estimated by using
the equation 2⫻(sodium⫹potassium)⫹glucose⫹urea. Serial plasma
urea and electrolytes were determined by the routine analysis Vitros
950 (Johnson & Johnson).
Whether the patient was hydrated orally, by intravenous fluids, or
by both methods was recorded from bedside fluid balance charts
during the first week. All staff, apart from the principal investigator
(A.B.), were unaware of plasma osmolality results during the first
week of admission.
Outcome
Patients were assessed during face-to-face interviews if alive at 3
months after stroke. Outcome assessment included death and dependency at 3 months (Barthel Index of 20, independent; Barthel Index
⬍20, dependent).10
Statistical Analysis
Analysis of serial measurements of plasma osmolality during the first
week was summarized in 2 ways: by calculating the area under the
curve (AUC) and also the maximum value for each patient.11
Univariate associations of plasma osmolality on admission between
survivors and dead patients were analyzed with the Student t test.
Multiple logistic regression was undertaken to examine the association between admission plasma osmolality and mortality, adjusting
for age, sex, stroke severity, Barthel Index before stroke, plasma
glucose, diabetes mellitus, and stroke subtype (classified as cerebral
infarction or primary intracerebral hemorrhage). Plasma osmolality
was considered as a dichotomous variable: ⱕ296 mOsm/kg and
⬎296 mOsm/kg, the upper limit of normal being 296 mOsm/kg.
Comparison of admission, maximum, AUC, and day 7 plasma
osmolality across the 3 hydration groups were analyzed by using
linear regression with 95% confidence intervals for differences,
adjusting for age, sex, and stroke severity. Linear regression was also
used to determine the association between age, serum sodium, serum
urea, and plasma osmolality. Samples of 5 patients in each hydration
group were selected at random (1 in 5 random sample) and plotted to
determine the pattern of plasma osmolality variation. To assess the
agreement between measured plasma osmolality and calculated
plasma osmolality, the mean difference of both these measures and
the standard deviations of these differences were estimated. As the
data were normally distributed, we would expect most of the
TABLE 1. Demographic and Case Mix Data of Stroke Patients
at Admission
Variable
Patients, n
(%)
n⫽167
Age, mean⫾SD, y
73.2⫾12.2
Male, n
80 (48)
Barthel index before stroke
⬍20
20 (12)
20
147 (88)
Comorbidity
Hypertension
81 (49)
Diabetes mellitus
17 (10)
Myocardial infarction
20 (12)
Atrial fibrillation
17 (10)
Cerebrovascular disease
26 (16)
Severity
Dysphagia
68 (41)
Dysphasia
88 (53)
Glasgow Coma Scale ⬍13
40 (24)
Incontinence
85 (51)
Stroke subtype
TACS
31 (19)
PACS
56 (33)
POCS
14 (8)
LACS
48 (29)
PICH
16 (10)
Unclassified
2 (1)
TACS indicates total anterior circulatory stroke; PACS, partial anterior
circulatory stroke; POCS, posterior circulatory stroke; LACS, lacunar circulatory
stroke; and PICH, primary intracerebral hemorrhage.
differences to lie within ⫾2 SD. The limits of agreement were
calculated from (1) mean difference minus 2 SD and (2) mean
difference plus 2 SD.28
Results
One hundred eighty-four consecutive patients with a diagnosis of stroke (WHO criteria8) were admitted to St Thomas’
Hospital, of whom 5 with a diagnosis of subarachnoid
hemorrhage were excluded and 12 were admitted 24 hours
after stroke onset. One hundred sixty-seven patients were
therefore eligible for the study. Demographic and clinical
data are shown in Table 1. The mean plasma osmolality on
admission was 295.2 mOsm/kg (SD 9.8, range 274 to 340
mOsm/kg). The mean serum sodium was 140.1 mmol/L (SD
2.9, range 131 to 148 mmol/L). The mean serum urea was 7.1
(SD 3.5, range 1.7 to 29.1 mmol/L). Fifty-five (33%) patients
were hyperglycemic on admission (glucose ⬎7 mmol/L).
Thirty-nine patients (26%) had stress hyperglycemia (glucose
⬎7 mmol/L in nondiabetics) on admission. Diabetic patients
had significantly raised plasma osmolality (300 mOsm/kg)
compared with nondiabetic patients (294 mOsm/kg; 95% CI
Bhalla et al
Plasma Osmolality and Stroke Outcome
2045
TABLE 2. Relationship of Mean Admission, Maximum, and AUC Plasma
Osmolality (pOsm) on 3-Month Stroke Mortality and Disability
Patients
Variable
Dead (n⫽50)
Alive (n⫽117)
95% CI
P
pOsm, mOsm/kg
Admission
300 (11.4)
293.1 (8.2)
⫺10 to ⫺3.9
⬍0.0001
Maximum*
308.1 (17.1)
297.7 (8.7)
⫺14.6 to ⫺6.3
⬍0.0001
298.3 (11.7)
291.7 (8.1)
⫺9.9 to ⫺3.2
Barthel ⬍20
(Dependent, n⫽65)
Barthel ⫽20
(Independent, n⫽52)
AUC*
0.0001
pOsm, mOsm/kg
Admission
294.8 (8.0)
290.9 (7.9)
Maximum
299 (9.2)
295.9 (7.8)
⫺0.1 to 6.2
0.06
292.7 (7.5)
290.4 (8.7)
⫺0.6 to 5.3
0.12
AUC
0.9 to 6.8
0.001
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Values are given as n (SD).
*Eleven patients who did not survive during the first week were excluded from the analysis.
⫺11 to ⫺1.4, P⫽0.01). Patients with stress hyperglycemia
had significantly raised plasma osmolality (298 mOsm/kg)
compared with normoglycemic patients (293 mOsm/kg; 95%
CI ⫺8 to ⫺1.4, P⫽0.006). Forty-four patients who were
taking a diuretic before admission (295.5 mOsm/kg) had the
same osmolality as 123 patients who were not taking diuretics
(295.1 mOsm/kg; 95% CI ⫺3.8 to 2.9, P⫽0.8). There was no
significant association with the use of diuretics before admission and raised plasma osmolality (P⫽0.68). Similarly, there
were no differences in (AUC) plasma osmolality in 30
patients (292.2 mOsm/kg) who continued diuretic therapy in
the first week after stroke compared with 125 patients (293.5
mOsm/kg) who did not (95% CI ⫺2.9 to 4.7, P⫽0.64).
Serum sodium was significantly associated with plasma
osmolality (coefficient⫽1.7 per mmol/L increase in serum
sodium, 95% CI 1.2 to 2.1; P⬍0.001). Serum urea was
significantly associated with plasma osmolality (coefficient⫽1.4 per mmol/L increase in serum urea, 95% CI 1.1 to
1.8; P⬍0.001). Age was also significantly associated with
plasma osmolality (coefficient⫽0.15 per year increase in age,
95% CI 0.03 to 0.27; P⫽0.02). There was no difference in
plasma osmolality levels between patients with cerebral
infarction (295.4 mOsm/kg, SD 10) and primary intracerebral
hemorrhage (292.5 mOsm/kg; 95% CI ⫺1.8 to 7.6, P⫽0.23).
Outcome
One hundred seventeen patients (70.1%) were still alive at 3
months after their stroke. Eleven (6.6%) died within 1 week
of admission to hospital. Mean admission and maximum and
AUC plasma osmolality were all significantly higher in those
who died than in those who survived. In patients who
survived, 52 (44.4%) achieved a Barthel Index score of 20 3
months after their stroke. Only mean admission plasma
osmolality was significantly higher in patients who were
dependent after their stroke (Table 2). There was no significant difference in survivors between diabetic and nondiabetic
patients (P⫽0.28). There was also no significant difference in
survivors between those who used diuretics before admission
and those who did not (P⫽0.4).
In the multiple logistic regression model, plasma osmolality ⬎296 mOsm/kg on admission showed a significant
association with stroke mortality at 3 months independent of
age, sex, stroke severity, Barthel Index before stroke, and
stroke subtype. (Table 3). From the regression equation
between measured and calculated plasma osmolality, a measured plasma osmolality ⬎296 mOsm/kg is equated to a
calculated plasma of ⬎294 mOsm/kg. After controlling for
the identical factors in the logistic regression model, the OR
of calculated plasma osmolality ⬎294 mOsm/kg associated
with stroke mortality at 3 months was 1.5 (95% CI 0.58 to
3.7, P⫽0.4). To assess the agreement between measured
plasma osmolality and calculated plasma osmolality, the
mean differences in both measurements were 0.97 mOsm/kg
TABLE 3. Multiple Logistic Regression Analysis Showing
Association Between Plasma Osmolality and Death
at 3 Months*
OR
95% CI
P
Plasma osmolality ⬎296
mOsm/kg
2.4
1.0–5.9
0.05
Glucose†
2.2
0.7–6.7
0.17
Diabetes mellitus
1.5
0.33–6.9
0.6
Age ⱖ65 y
1.9
0.48–7.26
0.36
Sex, female
1.14
0.43–2.59
0.91
Dysphasia
1.5
0.53–4.6
0.5
Dysphagia
8.7
2.82–26.9
0.01
Glasgow Coma Scale ⬍13
2.34
0.79–6.93
0.13
Incontinence
1.52
0.38–6.16
0.6
Barthel Index ⬍20
2.34
0.65–8.44
0.2
Primary intracerebral hemorrhage
2.42
0.58–10.1
0.22
Variable
*Adjusted for age, sex, stroke severity, Barthel Index before stroke, plasma
glucose, diabetes, and stroke subtype.
†Continuous variable.
2046
Stroke
September 2000
TABLE 4. Relationship of Mean, Maximum, AUC, and Day 7 Plasma Osmolality on Hydration During
First Week
Plasma Osmolality, mOsm/kg
Admission
Oral
(n⫽70)
292.4 (8.3)
95% CI for difference from oral
Maximum
296.6 (8)
Intravenous/Oral
(n⫽39)
Intravenous
(n⫽47)
Unadjusted
P
Adjusted
P*
294.7 (8.8)
299.1 (11.2)
⬍0.001
0.97
⫺4.8 to 3.8
⫺6.5 to 5.9
⬍0.001
0.25
⬍0.001
0.13
0.01
0.16
298.5 (9.1)
306.8 (16.4)
⫺6 to 4.6
⫺3.2 to 12.5
95% CI for difference from oral
AUC
291.4 (7.3)
Day 7
291.6 (8.4)
95% CI for difference from oral
291.1 (8.9)
⫺6.3 to 2.3
95% CI for difference from oral
298 (11.3)
⫺3.5 to 8.9
288.7 (26.1)
299.4 (16.8)
⫺13.3 to 2.5
⫺9 to 12.7
Comparison of plasma osmolality (pOsm) within hydration groups was performed by using linear regression, with oral group as the
baseline. Values are given as n (SD).
*Adjusted for age, sex, and stroke severity.
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
(SD 7.13). The limits of agreement were ⫺13.12 mOsm/kg
and 15.23 mOsm/kg. Therefore, the calculated plasma osmolality may be 13 mOsm/kg below or 15 mOsm/kg above the
measured plasma osmolality.
Hydration
Of 156 patients who survived the first week after stroke, 70
(45%) were hydrated orally, 47 (30%) were hydrated intravenously, and 39 (25%) were initially hydrated intravenously
(median 2 days, range 1 to 5 days) and then orally. Patients
who were only hydrated intravenously throughout the first
week were more likely to be dysphagic (P⬍0.001), dysphasic
(P⬍0.001), and incontinent (P⬍0.001), and have a Glasgow
coma scale ⬍13 (P⬍0.001). There were significant differences in plasma osmolality across the 3 hydration groups.
Patients who were hydrated intravenously throughout the first
week had significantly higher admission, peak, and AUC
plasma osmolality levels than orally hydrated patients; however, after adjusting for age, sex, and stroke severity, these
differences were not significant (Table 4). Examples of serial
data for the 3 groups of hydration are shown in the Figure. In
the 47 patients who were hydrated by intravenous fluids only,
there was no significant fall in plasma osmolality from
admission to day 7 compared with the 70 patients who were
hydrated orally (P⫽0.68).
Discussion
This is the first study to date to examine the relationship
between plasma osmolality and stroke mortality in a cohort of
consecutively admitted patients. In this study, we have found
an association between raised plasma osmolality during the
first week after stroke and mortality and functional outcome.
Dehydration is a common phenomenon after stroke, particularly in the elderly. In the early phase of stroke, dehydration may be caused by decreased oral intake of water due to
disturbed consciousness or dysphagia. Electrolyte disturbances such as hypernatremia or hyponatremia, resulting
from the syndrome of inappropriate antidiuretic hormone, can
lead to complications such as seizures or death.13,14 Dehydration acutely after stroke may worsen the ischemic process,
owing to an increase in blood viscosity and a decrease in
blood pressure.12 Studies have shown that hemodilution can
affect cerebral blood flow and cerebral hemodynamics by
influencing plasma viscosity,15,16 but no beneficial effect on
stroke outcome has been established.25
A significant relationship between raised plasma osmolality and mortality has been shown in long-stay community
patients and acutely ill patients, which suggests that dehydration may have influenced survival.17,18 Studies involving
stroke patients are limited. Joynt et al5 demonstrated no
significant difference in plasma osmolality on admission
between stroke patients and control subjects. O’Neill et al,6
who studied only 15 patients, found no significant differences
in plasma osmolality between survivors and dead stroke
patients.
In this study, admission mean plasma osmolality in stroke
patients showed a wide variation, indicating a large variance
in water homeostasis, which implies that some patients were
overhydrated and others underhydrated. The upper limit of
normal for measured plasma osmolality in this study was 296
mOsm/kg. Previous reports27 suggest that a plasma osmolality ⬎296 mOsm/kg is considered indicative of a hyperosmolar state. Hypernatremia was evident, with 1 patient
having a plasma sodium of 160 mmol/L. Diabetic patients
and those with stress hyperglycemia were also prone to raised
plasma osmolality levels after stroke.26
The association between raised plasma osmolality and
reduced survival may have reflected stroke severity, stroke
subtype, increasing age, or coexisting infection. There is
evidence19 that lesions around the hypothalamus and, in
particular, intracranial hemorrhage may alter plasma osmolality through vasopressin release. However, after adjustment
for age, sex, subtype, and stroke severity, we demonstrated
that raised plasma osmolality, defined as ⬎296 mOsm/kg on
admission, was independently related to mortality. This
implies that the association may be a direct one. We were not
able to measure or control for infective complications after
stroke. Because we did not exclude patients with cardiac or
renal disease, it is difficult to say whether admission plasma
osmolality levels were a true indicator of hydration status on
admission.
Bhalla et al
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
A, Mean plasma osmolality in examples of 5 stroke patients who
were hydrated orally. B, Mean plasma osmolality in examples of
5 stroke patients who were hydrated intravenously. C, Mean
plasma osmolality in examples of 5 stroke patients who were
hydrated intravenously and then orally.
To assess the agreement between measured plasma osmolality and calculated plasma osmolality, we used a wellvalidated approach, as suggested by Bland and Altman.28 The
limits of agreement between both of these measures were
⫺13.12 mOsm/kg and 15.23 mOsm/kg. This implied that the
calculated plasma osmolality may be 13.12 mOsm/kg below
or 15 mOsm/kg above the measured plasma osmolality.
These values are unacceptable for clinical purposes, and
therefore the agreement between both measures is not satisfactory. This is further reinforced by the fact that the
corresponding OR for the calculated plasma osmolality
(⬎294 mOsm/kg) failed to be significantly associated with
stroke mortality at 3 months. The added disadvantage of
Plasma Osmolality and Stroke Outcome
2047
using calculated plasma osmolality rather than measured
plasma osmolality is that one is estimating plasma osmolality
from at least 4 variables (sodium, potassium, urea, and
glucose), each with its own measurement error which may
contribute to the overall measurement error of the calculated
plasma osmolality.
From this study, it appeared that stroke patients were
initially hydrated appropriately, with patients with high
plasma osmolality levels at admission being hydrated intravenously. The explanation of higher maximum and AUC
plasma osmolality levels in patients hydrated intravenously
throughout the first week may have been explained by
differences in stroke severity. Patients who were intravenously hydrated tended to have severe strokes. There is still
controversy over whether dysphagia leads to significant
dehydration, and therefore these findings cannot be explained
by a simple division of intravenously hydrated and orally
hydrated groups reflecting poor and good outcome.20 One
may have expected a fall in plasma osmolality from admission to day 7 in individuals hydrated intravenously throughout the first week compared with orally hydrated individuals,
assuming that individuals who were receiving intravenous
fluids were receiving larger volumes of fluid; however, no
significant fall was demonstrated. Measurement of oral fluid
intake was not achieved owing to practical difficulty in
obtaining accurate volume measurements. These findings are
different from those in a study by O’Neill and colleagues, 6 in
which significant falls in plasma osmolality were demonstrated in patients who were hydrated intravenously rather
than orally.
Fluid balance, particularly in elderly patients, is notoriously haphazard and difficult to manage because the clinical
indicators of dehydration can be subtle.21 A decline in thirst
perception is a consequence of aging, and it has been shown
that high levels of plasma osmolality can lead to a diminished
subjective awareness of thirst.22 Cortical dysfunction after
cerebrovascular disease has also been implicated in causing
hypodipsia.23 When a patient is dysphagic and is thus
dependent on intravenous fluids, clinicians have to use
clinical and biochemical evidence for dehydration to adequately rehydrate patients. Evidence suggests that independent elderly individuals often drink what is offered to them
and then communicate their requests for further replenishment, unlike semidependent individuals, who often go
unnoticed.24
In this study, we have demonstrated that the present method
of hydration with fluids may be inappropriate. Although patients
are being hydrated intravenously, clinicians must be mindful that
insufficient fluid replacement may be taking place. Frequent
measurement of serum urea and electrolytes as well as plasma
osmolality, which gives valuable supportive evidence of water
homeostasis, must be considered.
In this study we have demonstrated that high plasma
osmolality levels in the acute phase of stroke are associated with excessive mortality rates. This may enable
identification of stroke patients who may benefit from
fluid replacement in a more systematic fashion. Further
work is required to determine the scale of water homeostasis and stroke subtype. Fluid intervention trials are re-
2048
Stroke
September 2000
quired to test the hypothesis that plasma osmolality levels
after acute stroke are indicators of water balance and that
improving plasma osmolality levels in the acute phase will
also improve clinical outcome.
Acknowledgments
This study was supported by the Research and Development Cerebrovascular Disease Program, London, and the Stroke Association.
References
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
1. Stroke Unit Trialists Collaboration. Collaborative systemic review of the
randomised trials of organised inpatient (stroke unit) care after stroke.
BMJ. 1997;314:1151–1159.
2. Inredavik B, Bakke RPT, Slordahl SA, Rokseth R, Haheim LL. Treatment
in a combined acute and rehabilitation stroke unit: which aspects are most
important? Stroke. 1999;30:917–923.
3. Ronning OM, Guldvog B. Stroke unit versus medical wards, II: neurological deficits and activities of daily living: a quasi-randomized controlled trial. Stroke. 1998;29:586 –590.
4. Sulter G, De Keyser J. From stroke unit care to stroke care unit. J Neurol
Sci. 1999;162:1–5.
5. Joynt RJ, Feibel JH, Sladek CM. Antidiuretic hormone levels in stroke
patients. Ann Neurol. 1980;9:182–184.
6. O’Neill PA, Davies I, Fullerton KJ, Bennett D. Fluid balance in elderly
patients following acute stroke. Age Ageing. 1992;21:280 –285.
7. O’Neill PA. Aging and salt and water balance. Rev Clin Gerontol.
1996;6:305–313.
8. WHO Special Report. Stroke–1989: recommendations on stroke prevention, diagnosis, and therapy. Stroke. 1989;20 1407–1431.
9. Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md
State Med J. 1965;14:61– 65.
10. Bamford JM, Sandercock PAG, Dennis MS, Burn J, Warlow CP. Classification and natural history of clinically identifiable subtypes of cerebral
infarction. Lancet. 1991;337:1521–1526.
11. Mathews JNS, Altman DG, Campbell MJ, Royston P. Analysis of serial
measurements in medical research. BMJ. 1990;300:230 –235.
12. Yamaguchi T, Minematsu K, Hasegawa Y. General care in acute stroke.
Cerebrovasc Dis. 1997;7(suppl 3):12–17.
13. Goldberg M, Handler JS. Hyponatraemia and renal wasting of sodium in
patients with malfunction of the central nervous system. N Engl J Med.
1960;263:1037–1043.
14. Bartter FC, Schwartz WB. The syndrome of inappropriate secretion of
antidiuretic hormone. Am J Med. 1967;42:790 – 806.
15. Harrison MJG. The influence of haematocrit in the cerebral circulation.
Cerebrovasc Brain Metab Rev. 1989;1:55– 67.
16. Heros RC, Kurusoe K. Hemodilution for cerebral ischemia. Stroke. 1989;
20:423– 427.
17. McLean KA, O’Neill PA, Davies I. Plasma osmolality: a predictor of
outcome in acutely ill elderly patients? Age Ageing. 1990;19(suppl 2):7.
18. O’Neill PA, Faragher EB, Davies I, Wears R, McLean KA, Fairweather
DS. Reduced survival with increasing plasma osmolality in elderly continuing care patients. Age Ageing. 1990;19:68 –71.
19. Nussey SS, Ang VTY, Jenkins JS. Chronic hypernatraemia and hypothermia
following subarachnoid haemorrhage. Postgrad Med J. 1986;62:476–471.
20. Gordon C, Langton Hewer R, Wade DT. Dysphagia in acute stroke. BMJ.
1987;295:411– 414.
21. Gross CR, Lindquist RD, Woolley AC, Granieri R, Allard K, Webster B.
Clinical indicators of dehydration severity in elderly patients. J Emerg
Med. 1992;10:267–274.
22. Philips PA, Rolls BJ, Ledingham JGG, Forsling ML, Crowe MJ, Wollner
L. Reduced thirst after water deprivation in healthy elderly men. N Engl
J Med. 1984;311:753–759.
23. Miller PD, Kerbs RA, Neal BJ, McIntyre DO. Hypodipsia in geriatric
patients. Am J Med. 1982;73:344 –346.
24. Gaspar PM. What determines how much patients drink? Geriatr Nurs.
1988 July/August;9:218 –221.
25. Asplund K, Israelsson K, Schampi. I. Haemodilution in acute ischaemic
stroke (Cochrane review). In: The Cochrane Library, issue 1, 1999.
Oxford, UK:Update Software.
26. Asplund K. Eriksson S. Hagg E. Lithner F. Strand T. Wester PO. Hyperosmolar non-ketotic coma in diabetic stroke patients. Acta Med Scand.
1982;212:407– 411.
27. Wears R, Davies I, Fairweather DS. Serum osmolality in elderly long stay
patients. Clin Sci. 1986;71(suppl 15):16.
28. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet. 1986;1:
307–310.
Influence of Raised Plasma Osmolality on Clinical Outcome After Acute Stroke
Ajay Bhalla, Suki Sankaralingam, Ruth Dundas, R. Swaminathan, Charles D. A. Wolfe and
Anthony G. Rudd
Downloaded from http://stroke.ahajournals.org/ by guest on June 16, 2017
Stroke. 2000;31:2043-2048
doi: 10.1161/01.STR.31.9.2043
Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2000 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/31/9/2043
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/