Journal of the Neurological Sciences 179 (2000) 65–69

www.elsevier.com / locate / jns

Continuous pulse oximetry in acute hemiparetic stroke

a, a b b a
Geert Sulter *, Jan Willem Elting , Roy Stewart , Adri den Arend , Jacques De Keyser
a
Department of Neurology, Academisch Ziekenhuis Groningen, Postbus 30.001, 9700 RB Groningen, The Netherlands
b
Northern Centre for Healthcare Research, Rijks Universiteit Groningen, Groningen, The Netherlands

Received 8 March 2000; received in revised form 18 May 2000; accepted 10 July 2000

Abstract

Background and purpose: Hypoxemia can adversely affect ischemic brain tissue in laboratory animals. The aim of this study was to
assess the value of early continuous monitoring with pulse oximetry in detecting arterial oxygen desaturations in patients with acute
hemiparetic stroke, and the effects of oxygen administration. Methods: Over a period of 6 months 49 consecutive patients with acute
hemiparetic stroke of #12 h duration were monitored for the first 48 h with pulse oximetry. Patients in whom arterial oxygen saturation
(SaO 2 ) fell beneath 96% for a period longer than 5 min were treated with oxygen administered via nasal prongs or oxygen mask. Results:
Thirty-one patients (63.3%) developed arterial oxygen desaturations. Of these patients 28 could effectively be treated with oxygen up to a
flow-rate of 5 l / min. Only 3 patients required higher oxygen concentrations from 6 to 10 l / min. No acute adverse effects of oxygen
treatment were observed. All patients with a history of cardiac and pulmonary disease developed drops in SaO 2 . The occurrence of arterial
oxygen desaturations was related to stroke severity (P50.024), the presence of dysphagia (P50.047), and older age (P50.037).
Conclusion: Patients with acute hemiparetic stroke frequently develop arterial oxygen desaturations. Continuous monitoring with pulse
oximetry in this group of patients is a simple and useful method to detect drops in SaO 2 and to titrate oxygen administration.  2000
Elsevier Science B.V. All rights reserved.

1. Introduction formation and lipid peroxidation, and hence aggravate

Patients with acute hemiparetic stroke are prone to
develop abnormal respiratory function and arterial hypox-
emia. Causes of respiratory dysfunction that have been
identified are hypoventilation due to impairment of move-
ments of the chest wall and diaphragm [1,2], atelectasis,
Cheyne-Stokes respiration, and aspiration pneumonia [3–
5]. In laboratory animals, it has been shown that hypox-
emia is poorly tolerated in areas of focal cerebral ischemia
and aggravates brain damage [6]. On the other hand, in
vitro studies suggest that reoxygenation with pure oxygen
after cerebral hypoxia-ischemia may enhance free radical
*Corresponding author. Tel.: 1-31-50-3612-430; fax: 1-31-50-3611-
707.
E-mail address: g.a.sulter@neuro.azg.nl (G. Sulter).
brain damage [7]. Thus, in the acute phase of stroke,
inadequate brain oxygenation may worsen ischemic brain
damage and blind administration of oxygen could theoret-
ically be harmful because it may induce free radical
formation in the ischemic area, especially after reperfusion.
In some centres, little attention is paid to the occurrence of
hypoxemia in stroke patients, while in others oxygen is
given without monitoring arterial oxygen saturation.
Arterial oxygen saturation (SaO 2 ) is relatively easy to
monitor by using pulse oximetry. It may serve as an
indicator for initiating, monitoring and adjusting oxygen
treatment [5]. Current pulse oximetry devices use two
wavelengths to measure the absorption of oxygenated and
desaturated haemoglobin. Problems with the accuracy of
the readings of a pulse oximeter may occur in jaundice or
in patients with poor peripheral circulation. In this study

0022-510X / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved.
PII: S0022-510X( 00 )00378-6
66 G. Sulter et al. / Journal of the Neurological Sciences 179 (2000) 65 – 69

we assessed the value of continuous monitoring with pulse
oximetry in detecting and correcting SaO 2 desaturations in
patients with acute hemiparetic stroke.
2. Materials and methods
We prospectively studied consecutive conscious patients
with acute hemiparetic stroke admitted to the stroke care
unit of the neurological department of the University
Hospital Groningen [8], over a period of 6 months from
September 1997 to February 1998. This project was part of
a protocol routinely used in our Stroke Care Unit. A local
independent human subjects research committee authorized
the study and decided that consent was not necessary. We
only included patients who were admitted within 12 h after
stroke onset, and in whom a CT of the brain was
compatible with ischemic stroke or primary intracerebral
hemorrhage. Sedative drugs were discontinued on admis-
sion. Stroke severity was measured using the National
Institutes of Health Stroke Scale (NIHSS). Severe stroke
was defined as a score of .13 on the NIHSS [9]. All
patients underwent a chest X-ray, 12-lead electrocar-
diogram, routine blood analysis, and a swallowing test for
detection of dysphagia [10,11]. The swallowing test con-
sisted of the administration with a spoon of small aliquots
of cold water to the patient sitting in the upright position.
Suction equipment was available. The test was considered
positive if the throat was not totally cleared from water,
which was recognised by coughing or ‘slurring’ of the
voice when the patient was asked to speak after each
swallow. A positive test led to immediate precautions to
prevent aspiration (nil per os and nasogastric tube feeding).
Patients developing desaturations and fever at the same
time were examined routinely to diagnose pneumonia.
SaO 2 was continuously monitored for at least 48 h after
admission, using a pulse oximeter connected to a Marquet-
te Eagle 4000 patient monitor. Respiratory rate and rectal
temperature were simultaneously monitored, but were not
recorded for this study.
In healthy elderly people arterial oxygen saturation is
higher than 95% [12]. Therefore, when SaO 2 dropped
below 96% for a period longer than 5 min, oxygen therapy
was started following a predetermined protocol as shown
in Table 1. In patients requiring .5 l / min of oxygen,
blood gasses were determined. After the 48 h monitoring
period patients requiring oxygen were classified into 3
different grades of oxygen desaturation severity on the
basis of the extent of oxygen therapy that was required
(Table 1). The decision to continue oxygen therapy after
the mandatory 48 h of monitoring was based on the
clinical judgement of the treating physician and was not
part of the study.
Commercial software (Microsoft Access and Excel,
SPSS) was used for management of the data and statistical
analysis. Comparisons between oxygen treated and non-
oxygen treated groups were made with the use of logistic
regression in the case of continuous data or with the use of
Chi-square test in the case of categorical data. Correlation
between dysphagia and initial stroke severity was defined
using Pearson’s analyses. In order to determine indepen-
dent risk factors for the occurrence of arterial oxygen
desaturations, logistic regression was performed on the
following variables: gender, age, dichotomised initial
stroke severity (NHISS# or .13), and ischemic stroke
versus primary intracerebral haemorrhage. We performed a
second logistic regression with the same factors, except
that initial stroke severity was replaced with a positive
dysphagia test as a variable to rule out the interference of a
strong relationship between dysphagia and initial stroke

Table 1
Treatment protocol and grading of severity of arterial oxygen desaturations
Event Action Grading of severity
SaO 2 96–100% No oxygen treatment No grading
SaO 2 ,96% – assessment that airways are free from Grade 1
obstructions and positioning of the patient in
the most comfortable way
– start oxygen treatment via nasal prong,
flow-rate 1–2 l / min
– chronic obstructive lung disease: start
ipratropium / salbutamol inhalation therapy
– heart failure: start appropriate treatment
– pneumonia or bronchitis: antibiotics
Oxygen treatment flow-rate Raise flow-rate up to 5 l / min, target SaO 2 Grade 2
1–2 l / min and SaO 2 ,96% 98–99%
Oxygen flow-rate 5 l / min – analysis of arterial blood gasses Grade 3
and SaO 2 ,96% – positioning of low flow oxygen mask,
flow-rate up to a maximum of 10 l / min
 G. Sulter et al. / Journal of the Neurological Sciences 179 (2000) 65 – 69 67

Table 3
Cause of arterial oxygen desaturations in function of grades of severity
Cause Grade 1 Grade 2 Grade 3
(n519) (n510) (n53)
Left ventricular failure 1 3 2
Pulmonary disease 3 4 1
Cheyne Stokes respiration 2 0 0
No apparent cause 13 3 0

sedative drugs, that may influence respiration. At time of

Fig. 1. Treatment of desaturation in an acute stroke patient with a history
of Chronic Obstructive Pulmonary Disease.
severity on the logistic regression model when both
variables would be included.
3. Results
In total 49 patients (21 males and 28 females) with a
mean (6S.D.; range) age of 70 (611; 45–89) years were
included in this study. Thirty-one of the 49 patients
(63.3%) developed a decline in SaO 2 below 96% during
the first 48 h after admission, and they were treated with
oxygen following the protocol listed in Table 1. A
representative case of a patient developing arterial oxygen
desaturation is shown in Fig. 1. Patients requiring oxygen
were older and had a higher median score on the NIHSS.
Mean age in the group requiring oxygen was 73 years
(S.D.610; range 45–89), in the group without desatura-
tions mean age was 65 (S.D.612; range 46–88).
None of the patients received medication, such as
admission on the stroke care unit none of the patients had
already received oxygen therapy. None were found to be
icteric or anemic.
Logistic regression analyses showed that the risk of
developing arterial oxygen desaturations in the acute phase
of stroke was strongly correlated to a positive dysphagia
test or initial stroke severity, and to a lesser degree to age.
In addition, all patients with a history of cardiac or
pulmonary disease developed desaturations (Table 2).
Only 5 (16%) patients had low admission SaO 2 levels;
in 26 (84%) oxygen desaturation occurred several hours
(from 1 up to 44 h) after admission in the stroke care unit.
Except for two patients with Cheyne Stokes respiration,
patients with oxygen desaturation severity grades 1 and 2
showed no clinical signs of dyspnea before or during
treatment. Initially the 3 patients in the severity grade 3
group showed no clinical signs of dyspnea when the drop
in SaO 2 was detected. During oxygen treatment all 3
developed signs of dyspnea. Patients with a clinical
diagnosis of pneumonia were treated with antibiotics. The
causes responsible for the decline in SaO 2 levels are
summarised in Table 3.
Twenty-eight of the 31 hypoxemic patients could be
effectively treated by oxygen administered via a nasal
prong up to a flow-rate of 5 l / min (severity grades 1 and
2). In these patients, SaO 2 levels normalised within
minutes after initiation of treatment. Three patients did not

Table 2
Multivariate risk factor analysis for arterial oxygen desaturations
Variable Patients requiring Patients with normal SaO 2 P value
oxygen levels
(n531) (n518)
NIHSS.13 a 13 1 0.024
Increasing age † N.A. N.A. 0.037
Positive dysphagia test a 13 1 0.047
Male sex 13 8 0.571
Ischemic stroke 27 14 0.637
(versus PICH ‡ )
Pulmonary disease 5 0 NA
Cardiac disease 4 0 NA
a † ‡
Analysis with either one of the two variables included in the logistic regression. Continuous data. Primary Intracerebral Hemorrhage. NA5not
applicable.
68 G. Sulter et al. / Journal of the Neurological Sciences 179 (2000) 65 – 69
respond adequately to low flow oxygen treatment (severity
grade 3), and these were given oxygen up to 10 l / min via
an oxygen mask. No acute adverse events attributable to
oxygen therapy occurred in the patients with severity
grades 1 and 2, and no CO 2 retention was measured in the
arterial blood gas analysis performed in the three patients
with severity grade 3. No arterial blood gas analysis was
performed in severity grade 1 and 2 patients.
4. Discussion
There have been previous reports on the occurrence of
respiratory dysfunction in stroke patients [1,2,4,13,14].
Our study is the first that used continuous pulse oximetry
during the acute phase, to detect arterial oxygen desatura-
tions and to monitor the effects of oxygen therapy. More
than half of the patients with acute hemiparetic stroke
developed a decline in SaO 2 , despite that in the majority
there were no clinical signs of dyspnea. Only 2 patients
showed a Cheyne Stokes respiration, and 3 developed
dyspnea. It is also important to mention that the decline in
SaO 2 often occurred several hours after admission, and
that this would have gone unnoticed without continuous
monitoring.
Oropharyngeal dysphagia is common in patients with
stroke. With old age swallowing problems are more likely,
due to alterations in oral transit time, pharyngeal stage of
swallowing and esophageal peristalsis. In addition, in
stroke patients poor oral control, failure of laryngeal
adduction and delayed triggering of the swallow reflex
causes a high risk of aspiration [15]. Although the bedside
dysphagia test is not sensitive enough to detect all patients
who will aspirate, it is a useful method to demonstrate the
presence of a clinically significant swallowing problem.
We found that a positive dysphagia test on admission is a
strong predictor for the development of arterial oxygen
desaturations. A possible explanation for this observation
is that patients with swallowing problems are at high risk
for aspiration and pneumonia. However, only 3 of the 14
patients with a positive dysphagia test actually developed a
clinically apparent pneumonia, suggesting that other fac-
tors may be involved. We also found a strong correlation
between a positive dysphagia test and stroke severity, and
between admission stroke severity (NIHSS.13) and the
development of arterial oxygen desaturations. This would
indicate that hypoventilation due to impairment of move-
ments of the chest wall and diaphragm may be an
important mechanism. Alternatively, mild neurogenic pul-
monary edema could also play a role, because this has
been described in the course of sudden neurologic injury,
including stroke [16–18]. Sleep apnea, a syndrome often
found in hemispheric stroke patients, is also a cause of
arterial oxygen desaturations [19]. In this study variables
such as history of sleep apnea, body weight, obesity and
sleep during the monitoring phase were not recorded, so it
remains unclear if sleep apnea is a risk factor for develop-
ing desaturations. Other factors associated with the de-
velopment of arterial oxygen desaturations were old age
and a history of cardiac or pulmonary disease.
It was not the purpose of our study to demonstrate that
arterial oxygen desaturations adversely affects clinical
outcome following acute stroke. To demonstrate such an
effect would require a large trial to compare the outcome
of hypoxemic stroke patients randomly assigned to receive
oxygen or sham treatment. We feel that it would be
unethical to perform such a trial and to withhold the
administration of oxygen in hypoxemic patients with focal
brain ischemia. The additional oxygen may accelerate the
re-establishment of normal oxidative metabolism and
prevent secondary brain damage On the other hand, the
blind administration of oxygen could induce free radical
formation in the ischemic area [20]. Preliminary results of
a randomised trial, comparing blind administration of
supplemental oxygen versus no supplemental oxygen in
stroke patients suggested a worse outcome for those having
received oxygen [21]. Continuous monitoring with pulse
oximetry allows on-demand administration of oxygen and
selection of patients, who are really in need for oxygen
therapy and prevents over-oxygenation.
In the majority of the patients arterial oxygen saturation
could effectively and quickly be restored with low flow
oxygen therapy (#5 l / min) with a nasal prong as route of
administration. Nasal prongs prevent rebreathing and
therefore there is no danger of hypercapnia when low flow
oxygen is administered. They are comfortable for long
periods, and allow oxygen to be continued during talking
and eating. Only three patients required higher concen-
trations of oxygen administered through a facial mask, and
this intervention was also effective to regain normal
oxygen saturation. We did not observe acute side effects
related to the oxygen therapy.
In conclusion, continuous monitoring with pulse oxi-
metry in acute stroke patients is a simple and useful
method to detect unwanted drops in SaO 2 and it allows
careful titration of oxygen administration.
References
[1] Fluck DC. Chest movements in hemiplegia. Clin Sci 1966;31:383–
8.
[2] Houston JG, Morris AD, Grosset DG, Lees KR, McMillan N, Bone
I. Ultrasonic evaluation of movement of the diaphragm after acute
cerebral infarction. J Neurol Neurosurg Psychiatry 1995;58:738–41.
[3] Nachtmann A, Siebler M, Rose G, Sitzer M, Steinmetz H. Cheyne-
Stokes respiration in ischemic stroke. Neurology 1995;45:820–1.
[4] Walshaw MJ, Pearson MG. Hypoxia in patients with acute hemip-
legia. Brit Med J 1984;288:15–7.
[5] Bateman NT, Leach RM. ABC of oxygen. Brit Med J
1998;317:798–801.
[6] Back T, Kohno K, Hossman KA. Cortical negative DC deflections
following middle cerebral artery occlusion and KCl-induced spread-
 G. Sulter et al. / Journal of the Neurological Sciences 179 (2000) 65 – 69
ing depression: effect on blood flow, tissue oxygenation, and
electroencephalogram. J Cereb Blood Flow Metab 1994;14:12–9.
[7] Bagenholm R, Hagberg H, Kjellmer I. Impact of reoxygenation with
oxygen and air on the extent of the brain damage after hypoxia-
ischaemia in neonatal rats. Acta Paediatr 1996;85:1228–31.
[8] Sulter G, De Keyser J. From stroke unit care to stroke care unit. J
Neurol Sci 1999;162:1–5.
[9] Lai SM, Duncan PW, Keighley J. Prediction of functional outcome
after stroke. Stroke 1998;29:1838–42.
[10] DePippo KL, Holas MA, Reding MJ. Validation of the 3-oz water
swallow test for aspiration following stroke. Arch Neurol
1992;49:1259–61.
[11] Kidd D, Lawson J, Nesbitt R, MacMahon J. Aspiration in acute
stroke: a clinical study with videofluoroscopy. Q J Med
1993;86:825–9.
[12] Ogburn-Russell, Johnsen JE. Oxygen saturation levels in the well
elderly: altitude makes a difference. J Gerontol Nurs 1990;16:26–
30.
[13] McMahon SM, Heyman A. The mechanics of breathing and
stabilization of ventilation in patients with unilateral cerebral
infarction. Stroke 1974;5:518–27.
69
[14] Elizabeth J, Singarayar J, Ellul J, Barer D, Lye M. Arterial oxygen
saturation and posture in acute stroke. Age Ageing 1993;22:269–72.
[15] Veis SL, Logemann JA. Swallowing disorders in persons with
cerebrovascular accident. Arch Phys Med Rehabil 1988;66:372–5.
[16] Smith WS, Matthay MA. Evidence for a hydrostatic mechanism in
human neurogenic pulmonary edema. Chest 1997;111:1326–33.
[17] Carlson RW, Schaeffer RJ, Michaels SG, Weil MH. Pulmonary
edema following intracranial hemorrhage. Chest 1979;75:731–4.
[18] Simon RP. Neurogenic pulmonary edema. Neurol Clin
1993;11(2):309–23.
[19] Kapen, Park A, Goldberg J, Wynter J, Park A. The incidence and
severity of obstructive sleep apnea in ischemic cerebrovascular
disease. Neurology 1991;41(Suppl 1):125.
[20] Miyamato O, Auer RN. Hypoxia, hyperoxia, ischemia, and brain
necrosis. Neurology 2000;54:362–71.
[21] Rønning OM, Guldvog B. Does breathing supplemental O2 affect
outcome of acute stroke? A randomised trial. Cerebrovasc Dis
1999;9(Suppl 1):127.