Age and Ageing Advance Access originally published online on September 1, 2006
Age and Ageing 2006 35(6):627-629; doi:10.1093/ageing/afl099
© The Author 2006. Published by Oxford University Press on behalf of the British Geriatrics Society. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
Is overnight tube feeding associated with hypoxia in stroke?
SIR—Dysphagia is found in about 45% of patients who have
had a stroke [
1], and nasogastric (NG) or percutaneous endoscopic
gastrostomy (PEG) tubes are often used for enteral nutrition.
The insertion of an NG or PEG tube in stroke patients has been
associated with a modest drop in arterial oxygen saturation
(SaO
2) [
2]. In a study of NG tube-fed patients with chronic
obstructive airways disease, there was a small drop in SaO
2 associated with feeding [
3]. Reduced SaO
2 during swallowing
and oral feeding has also been demonstrated in stroke [
4–6].
However, it is not known whether tube feeding in stroke patients
leads to significant episodes of arterial desaturation and whether
there is a case for routine pulse oximetry or supplemental oxygen
during tube feeding. A study that compared SaO
2 during a 30
min period of tube feeding with orally fed stroke controls found
a very small but statistically significant change in SaO
2 in
tube-fed patients [
7].
The detection of hypoxia following a stroke is of great importance because a reduced SaO2 is thought to be associated with worse outcomes after a stroke [8, 9]. Stroke patients have been shown to have lower baseline oxygen saturations compared to controls and are at risk of further hypoxia [6, 8, 10]. Several factors such as weakness of respiratory muscles [11], altered central regulation of respiration [12], sleep apnoea [13], aspiration [4, 6], chest infections [10, 14] and pulmonary emboli [10, 14] contribute to post-stroke hypoxia, which can be intermittent and often worse at night. Overnight pulse oximetry during sleep in stroke patients shows an SaO2 about 1% less than that of controls [15]. Body posture can also influence SaO2, and a 1% difference in SaO2 between the supine and upright positions has been demonstrated [16]. Other studies, however, have had more inconsistent results, and no difference has been shown between the supine position and lying on the right or left side [8] or in oximetry readings between the hemiparetic and non-paretic side [17]. Non-stroke factors such as cardiovascular or respiratory problems and body mass index can also influence SaO2. This study was designed to determine whether overnight tube feeding is associated with hypoxia in stroke patients who are not hypoxic at baseline and not on supplemental oxygen.
 |
Methods
|
|---|
This was a prospective, observational study of consecutive dysphagic
acute stroke patients fed by NG or PEG tubes in a stroke unit.
Patients acted as their own controls, and repeated oximetry
readings were taken under the same conditions with and without
tube feeding. Brain imaging was performed, and both ischaemic
and haemorrhagic strokes were eligible. The stroke was classified
according to the Oxfordshire Community Stroke Project classification
system as a total anterior circulation stroke (TACS), partial
anterior circulation stroke (PACS), lacunar stroke (LACS) or
posterior circulation stroke (POCS) [
18]. The modified Rankin
scale was used as a measure of disability [
19,
20]. Demographic
data and previous medical problems were recorded. All patients
were assessed by speech and language therapists, and patients
who were unsafe for oral feeding had NG tubes initially and
some went on to have PEG tubes. Patients who were terminally
ill or already on supplemental oxygen were excluded. Patients
who were confused or restless and who tended to displace their
feeding tubes were also excluded. The study had the approval
of the local ethics committee, and consent was written, witnessed
verbal or with permission from next of kin in a few cases.
Recordings were taken twice for each patient on two consecutive nights with a finger probe. The hand on which the sensor was placed was usually the paretic side [17] or randomly chosen in those without hemiparesis. The probe was repositioned once during the recording to reduce the likelihood of probe burns. The readings were taken in the same, semi-recumbent position on both occasions. On one night, the tube was kept in situ during the recording but not used for feeding. Monitoring was commenced at midnight and continued till 8 am. On the other night, tube feeding was commenced at midnight at a rate of 75–100 ml/h with continuous monitoring. To reduce bias, the night on which the patient received tube feeding was randomly selected by a randomisation method comprising sealed envelopes.
The instrument used was a Ohmeda 3700 pulse oximeter (Datex-Ohmeda, Hertfordshire, UK) with a download software package that allowed automatic recording of SaO2 levels (with mean, median and standard deviations), pulse rate, a graphical representation of SaO2 levels and percentage of readings with an SaO2 level of <93%. Artefacts were identified by concurrent anomalies in the graphical trace and pulse recording and were excluded from the recording. Hypoxia was defined as an SaO2 level of <90%. Because the mean duration of actual recordings was 
6 h, results for T90 (the time the subjects spent with an SaO2 <90%) [15] were corrected to a standard 6 h period with the formula: T90c (in minutes) = (T90 in minutes/actual recording time in minutes) x 360 [15]. The minimum sample size needed to detect a change in the SaO2 of 1% or more at a significance level of 5% and power of 90% was calculated to be 16 based on data from an earlier study [7]. Data were assessed for normality and the paired sample t-test used in all instances using the statistical package SPSS.
|
Results
|
|---|
Of 165 patients admitted to the stroke unit during January to
September 2005,
32 were fed by tube for >7 days. There were
several exclusions because of pre-existing hypoxia, terminal
prognosis and frequent NG tube displacement. Sixteen tube-fed
patients consented to enter the study during this period. Baseline
characteristics are summarised in Table
1. The mean time to
assessment from stroke onset was 25.19 days.
Results of pulse oximetry are summarised in Table 2. The mean duration of recording was 6.4 h (range 5–8 h). The mean SaO2 for patients not fed was 94.9% (SD 1.11) and that of patients during tube feeding was 94.3% (SD 1.03). The mean percentage of readings with SaO2 <93% was 14.2 without feeding and 19.9 during feeding. The mean duration of time spent with SaO2 <90% (corrected to 6 h) was 8.0 min for patients when not fed and 18.8 min for patients during tube feeding. Four patients spent >30 min with SaO2 <90%. There seemed to be a tendency to hypoxia in the readings during tube feeding, but the difference was not statistically significant regardless of which of these three parameters were used for comparison.
|
Discussion
|
|---|
There are several factors that can potentially contribute to
hypoxia after a stroke, and there are reasons to suspect that
tube feeding might be one of them. Reduced SaO
2 can adversely
affect stroke outcomes though we do not know what constitutes
a clinically significant degree of hypoxia very early after
a stroke. It is normal practice to monitor oxygen saturations
after a stroke and prescribe oxygen in patients with hypoxia.
The aim of this study was simply to make an observation of oxygen
saturations in overnight tube-fed patients to look for an association
of feeding with hypoxia. Patients acted as their own controls,
and the only variable that was different in the circumstances
of the two paired readings was the infusion of the feed through
the feeding tube. We believe that this method had the greatest
chance of eliminating the large number of confounding factors
that can influence SaO
2 after a stroke.
Even though the sample studied was small, it was adequate to detect a difference of 1% or greater in SaO2 in the two sets of readings. No patients were hypoxic at baseline because patients with pre-existing hypoxia who required supplemental oxygen were excluded. Although there seemed to be a trend towards a lower SaO2 in the tube-fed readings and a greater period of time spent with SaO2 readings <90%, this difference was small, not statistically significant and possibly because of chance alone. The night-time reduction in SaO2 detected in our patients is in keeping with the demonstration of nocturnal hypoxia in an earlier study of acute stroke patients [15]. In this study, patients who were normoxaemic at baseline dropped their oxygen saturations by a mean of 1% as compared with non-stroke controls and about 25% had periods of hypoxia (SaO2 <90%) lasting half an hour or more [15].
In this small study, we have shown that tube feeding does not cause large (>1%) falls in oxygen saturation. Feeding per se is not an indication for routine pulse oximetry or supplemental oxygen. There was, however, a trend towards lower SaO2 in patients fed over night. A larger study would be needed to conclusively determine whether tube feeding contributes to the hypoxic burden in stroke.
|
Key points
|
|---|
- Hypoxia during swallowing, oral feeding and feeding tube placement has been demonstrated in stroke patients.
- It is not known whether overnight tube feeding in stroke patients causes episodes of hypoxia.
- In this small study, we have shown that tube feeding does not cause large (>1%) falls in oxygen saturation.
|
Conflicts of interest
|
|---|
None.
Dipankar Dutta*,
Tracy Wood,
Rhys Thomas and
Muhammad Asrar ul Haq
Gloucestershire Royal Hospital, Great Western Road, Gloucester GL1 3NN, UK
* To whom correspondence should be addressed Tel: (+44) 08454 226321; Fax: (+44) 08454 226979; Email: dipankar.dutta{at}glos.nhs.uk
|
Acknowledgements
|
|---|
Funding No external funding.
|
References
|
|---|
- Gordon C, Langton-Hewer R, Wade DT. Dysphagia in acute stroke. Br Med J 1987; 295: 411–4.[ISI][Medline]
- Rowat A, Wardlaw J, Dennis M. Changes in arterial oxygen saturation before and after enteral feeding tube insertion in dysphagic stroke patients. Age Ageing 2004; 33: 42–5.[Abstract/Free Full Text]
- Brandstetter RD, Zakkay Y, Gutherz P, Goldberg RJ. Effect of nasogastric feedings on arterial oxygen tension in patients with symptomatic chronic obstructive pulmonary disease. Heart Lung 1988; 17: 170–2.[ISI][Medline]
- Zaidi NH, Smith HA, King SC, Park C, O’Neil PA, Connolly MJ. Oxygen desaturation on swallowing as a potential marker of aspiration in acute stroke. Age Ageing 1995; 24: 267–70.[Abstract/Free Full Text]
- Collins MJ, Bakheit AM. Does pulse oximetry reliably detect aspiration in dysphagic stroke patients? Stroke 1997; 28: 1773–5.[Abstract/Free Full Text]
- Rowat AM, Wardlaw JM, Dennis MS, Warlow CP. Does feeding alter arterial oxygen saturation in patients with acute stroke? Stroke 2000; 31: 2134–40.[Abstract/Free Full Text]
- Dutta D, Bannerjee M, Chambers T. Is tube feeding associated with altered arterial oxygen saturation in stroke patients? Age Ageing 2004; 33: 493–6.[Abstract/Free Full Text]
- Roffe C. Hypoxaemia and stroke. Rev Clin Gerontol 2001; 11: 323–5.[CrossRef]
- Bhalla A, Wolfe SDA, Rudd AG. Management of acute physiological parameters after stroke. QJM 2001; 94: 167–72.[Abstract/Free Full Text]
- Davenport RJ, Dennis MS, Wellwood I, Warlow CP. Complications after acute stroke. Stroke 1996; 27: 415–20.[Abstract/Free Full Text]
- 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.[Abstract]
- Nachtmann A, Siebler Rose M, Sitzer G, Steinmetz M, Cheyne-Stokes H. Respiration in ischemic stroke. Neurology 1995; 45: 820–1.[Abstract/Free Full Text]
- Harbison JA, Gibson GJ. Snoring, sleep apnea, and stroke: chicken or scrambled egg? QJM 2000; 93: 647–54.[Free Full Text]
- Langhorne P, Stott DJ, Robertson L et al. Medical complications after stroke. Stroke 2000; 31: 1223–9.[Abstract/Free Full Text]
- Roffe C, Sills S, Halim M, Wilde K, Allen M, Jones P. Unexpected nocturnal hypoxia in patients with acute stroke. Stroke 2003; 34: 2641–5.[Abstract/Free Full Text]
- Elisabeth J, Singarayar J, Ellul J, Barer D, Lye M. Arterial oxygen saturation and posture in acute stroke. Age Ageing 1993; 22: 269–72.[Abstract/Free Full Text]
- Roffe C, Sills S, Wilde K, Crome P. Effect of hemiparetic stroke on pulse oximetry readings on the affected side. Stroke 2001; 32: 1808–10.[Abstract/Free Full Text]
- Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. Classification and natural history of clinically identifiable subtypes of cerebral infarction. Lancet 1991; 337: 1521–6.[CrossRef][ISI][Medline]
- Rankin J. Cerebral vascular accidents in people over the age of 60. Scott Med J 1957; 2: 200–15.[Medline]
- Warlow CP, Dennis MS, van Gijn J et al. Stroke – A Practical Guide to Management, 2nd edition. Oxford: Blackwell, 2001.
CiteULike 
Connotea 
Del.icio.us What's this?
Top
Methods
Results