Q J Med 2002; 95: 741-747
© 2002 Association of Physicians
Sleep-disordered breathing following acute stroke
From the Department of Medicine, University of Newcastle upon Tyne and Department of Respiratory Medicine, Freeman Hospital, Newcastle upon Tyne, UK
Received 17 April 2002 Accepted for publication 21 July 2002.
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Background: The reported prevalence of sleep apnoea following stroke varies between 44% and 72%, but its course and relations are unclear.
Aim: To determine the prevalence and course of sleep-disordered breathing in acute stroke in-patients, and its relation to age, stroke subtype, pre-stroke handicap and post-stroke outcome.
Design: Prospective uncontrolled observational study.
Methods: Paired respiratory sleep studies were performed at week 2 and week 6–9 following stroke. Pre- and post-stroke handicap (modified Rankin Score) and week 2 and week 6–9 disability (Barthel Score) and impairment (Scandinavian Neurological Stroke Score) were assessed. Pre-stroke sleepiness was determined by Epworth Score.
Results: There were 68 patients in the week 2 study, and 50 in the week 6–8 study. Mean week 2 Apnoea Hypopnoea Index (AHI) was 30; 64 patients (94%) had an AHI 
10. Mean AHI was higher in subjects with lacunar vs. cortical strokes (44 vs. 28, p<0.05), in subjects aged 65 years (32 vs. 21, p<0.05) and in those with greater pre-stroke handicap (modified Rankin Score 2 vs. <2) (41 vs. 27, p<0.05). In 50 paired studies, mean AHI fell from 31 to 24 (p<0.01) and the proportion with AHI 10 fell from 96% to 72%. Pre-stroke sleepiness was associated with post-stroke neurological impairment (r=-0.325, p<0.05) and disability (r=-0.377, p<0.05).
Discussion: Sleep-disordered breathing improves in the first 6–9 weeks following stroke, but remains highly prevalent. Worse sleep-disordered breathing was observed following lacunar stroke, and in older subjects or those with pre-stroke handicap.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Several studies have shown a high prevalence of sleep-disordered breathing (SDB) in patients following stroke,1 but its natural history is unclear and its relation to post-stroke disability or mortality uncertain. Most studies have examined populations with an unusually high proportion of younger (
65 years) patients, and the proportion with severe strokes has been relatively small, perhaps reflecting the difficulty of performing detailed sleep studies in patients with more severe impairment. Studies in younger patients or in milder strokes may underestimate the prevalence or severity of sleep apnoea in stroke patients in general. The aims of the present study were to determine the prevalence and course of sleep-disordered breathing in patients of typical age and severity following stroke and to assess any relations with pre-stroke characteristics, stroke type, and post stroke impairment, handicap, disability and mortality.
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Patients
A total of 262 consecutive patients admitted to the stroke service at Freeman Hospital, who were still in-patients one week after stroke onset, were screened for inclusion in the study. The Freeman Hospital Acute Stroke service admits >90% of all strokes from the city of Newcastle upon Tyne (population
300 000). We excluded subjects with cognitive impairment or agitation, which would have rendered accurate sleep study impracticable. Subjects with co-morbid disease likely significantly to affect the results of sleep studies, e.g. pneumonia or pulmonary oedema, those living >5 miles from the hospital and patients requiring hypnotic or sedative medication were also excluded. In all, 194 were excluded for reasons outlined in Figure 1
. The remaining 68 previously independent, self-caring patients (Modified Rankin Score 2), who continued to require inpatient rehabilitation at 7 days post stroke, were included in the study.
|
Sleep studies and assessment scales
Sleep studies were performed during the second week of admission and again at 6–9 weeks post stroke using the ResMed Autoset II plus system in diagnostic mode. This equipment records airflow by nasal cannulae, arterial oxygen saturation (SaO2) by a finger oximeter probe, chest movement by a thoracic resistance band, and body position. It was chosen because of patient tolerability and ease of placement/replacement of sensors, and because it permitted easy handling and pressure area care in hemiparetic patients. Several studies have validated this device in comparison with polysomnography,2–4 and it has been used previously to diagnose SDB in stroke patients.5 The results were expressed as Apnoea Hypopnoea Index (AHI), the frequency of apnoeas and hypopnoeas per hour of study. An apnoea was scored if the 2-s moving average ventilation dropped below 25% of the recent average (time constant 100 s) for at least 10 consecutive seconds. A hypopnoea was defined as a 50%–75% decrease in nasal ventilation for at least 10 s. Median and nadir values of nocturnal arterial oxygen saturation were also recorded.
Severity of handicap (Modified Rankin Handicap Scale, MRS),6,7 disability (Barthel Disability Score)8 and impairment (Scandinavian Neurological Stroke Scale, SNSS)9 were assessed by experienced stroke research medical and nursing staff without knowledge of the sleep study results. Strokes were categorized according to the Oxfordshire Community Stroke Project (OCSP) classification,10 based on clinical history, examination and brain CT/MR findings. The severity of sleepiness pre-stroke was assessed using the Epworth Sleepiness score (ESS)11 applied retrospectively12 in 43 of the 50 subjects who completed two sleep studies; in 27 of these the information came from the patients themselves and in the remaining 16 patients from next of kin. An ESS >10 was taken to indicate excessive daytime sleepiness.13
Approval was obtained from the local Ethics Committee and written, informed consent was obtained from the patients or their next of kin prior to inclusion in the study. Procedures followed were in accordance with institutional guidelines.
Statistical methods
Results from week 2 and week 6–9 studies were compared using paired t-tests. The Mann-Whitney U test was used for comparisons of subgroups determined by age, OCSP stroke type, level of premorbid handicap and mortality. Relations between continuous variables were assessed using correlation coefficients and multiple regression analysis. Statistical calculations were performed using a proprietary, computerized statistics package (SPSS 9.0.0, SPSS 1999).
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Patients
Sixty-eight patients (mean age 73 years) underwent an initial sleep study (mean 10 days post-stroke) and, in fifty of these, repeat studies were performed successfully at week 6–9 (mean 45 days post-stroke). Repeat studies could not be successfully performed in eight patients (4 technical failures, 2 declined, 1 medical complications of stroke, 1 moved from region). Seven patients died and three had recurrent strokes before a second study. In all, 10 of the 68 patients died within 3 months of the initial study (Figure 1
).
Initial studies
Patient characteristics and results of initial sleep studies are detailed in Table 1. Of the total of 68 patients, 64 (94%) had an AHI 10. SDB was more common in older patients (65 years, n=56, mean AHI 32, 95%CI 28–37) vs. younger patients (<65 years, n=12, mean AHI 21, 95%CI 12–29) (p<0.05) and in those with greater pre-stroke handicap (Modified Rankin 2 vs.<2) (AHI 41 vs. 27, p<0.05).
|
AHI correlated significantly with pre-stroke Modified Rankin Score (r=0.313, p<0.01), but not with admission SNSS (r=0.07, p=0.57) or week 2 Barthel score (r=-0.09, p=0.43).
Patients with lacunar infarcts, as defined by OCSP criteria, had worse sleep-disordered breathing than patients with anterior circulation cortical strokes (mean AHI 44 vs. 28, p<0.05), even though neurological impairment at admission was significantly less in the lacunar group (SNSS 44 vs. 27, p<0.05). The difference in AHI between those who died within 3 months (n=10) and survivors was not significant, (mean AHI 33 vs. 25, p=0.58). As expected, subjects who died had greater neurological impairment (mean admission SNSS 18 vs. 44, p<0.01). Although they were older (mean 78 (95%CI 68–88) years vs. 72 (95%CI 69–75) years) and had greater pre-stroke handicap (mean Modified Rankin score 1.1 (95%CI 0.5–1.8) vs. 0.6 (95%CI 0.4–0.8)), neither difference was significant.
The average median nocturnal arterial oxygen saturation was 96%, and the mean nadir value was 85%. Subjects with median oxygen saturation <95% had higher mortality (5/16) than non-hypoxic subjects (4/49); OR 5.3 (1.2–23.2). However, multiple logistic regression analysis of the factors associated with mortality showed only admission SNSS to be significant. Age, pre-stroke handicap, AHI and nocturnal oxygen saturation were not independently associated with mortality.
Two of the three patients who suffered further strokes after their initial studies underwent repeat studies two weeks post recurrent stroke. Both showed worse SDB in the repeat study (AHI 10 vs. 54 and 49 vs. 101).
Multiple regression analysis was used to evaluate the factors associated independently with AHI at week 2. Of age, pre-stroke Modified Rankin score, admission SNSS and Barthel score at week 2, only pre-stroke Modified Rankin score was a significant predictor of AHI (p<0.05).
Paired studies
Fifty patients underwent repeat study at week 6–9 (Table 2). In 39, SDB improved, in 10 it worsened, and in one it remained unchanged. There was no difference in mortality or neurological recovery between patients in whom SDB improved or deteriorated. Overall, the AHI fell significantly between the two studies (mean 31±17 vs. 24±16, p<0.01). Correspondingly, fewer patients had an AHI 10 on repeat study compared to week 2 (74% vs. 96%, p<0.01). The change in AHI did not correlate with changes in neurological impairment, disability or handicap scores. However, as with the first sleep studies, AHI in the second study correlated significantly with pre-stroke Modified Rankin score. Median and nadir SaO2 values showed no significant differences between the two studies (Table 2).
|
Sleepiness and stroke severity
Pre-stroke Epworth score (available in 43 individuals) correlated significantly with SNSS at week 6–9 (r=-0.325, p<0.05) and with Barthel scores at week 2 (r=-0.357, p<0.05) and week 6–9 (r=-0.377, p<0.05).
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
The prevalence of sleep-disordered breathing (AHI
10) in our population was higher than reported in most previous studies. Sleep-disordered breathing was worse in older subjects and in those with lacunar strokes and improved during the first few weeks following stroke. AHI in both sleep studies post-stroke correlated with the severity of pre-stroke handicap. Post-stroke AHI was not related to 3-month mortality.
Previous studies
Several studies have reported a high prevalence of SDB following stroke, varying between 43% and 72%.14–19 Most previous researchers have used AHI 10 as indicating SDB and, while this figure is arbitrary, we have used it for purposes of comparison. A notable feature of many studies has been the atypically young age of the stroke populations studied, often averaging 65 years or less. Since younger patients tend to suffer less severe strokes and have a better prognosis, we aimed to study the prevalence and progress of sleep-disordered breathing post stroke in a more typical older population.
Two previous studies have shown a possible association between SDB or nocturnal hypoxaemia post stroke and mortality. Good et al.20 studying 47 patients a median of 13 days post stroke showed a relation between mortality and time spent with SaO2<90%, but not with mean SaO2, desaturation index or AHI. Dyken et al.,15 studying 24 patients, found a higher AHI in the five patients who died in the subsequent 4 years than in those who survived, but the difference was not significant, and none of the deaths were due to stroke.
Methods and technical aspects
We chose a limited sleep study system for convenience and acceptability in dependent and potentially agitated patients. Prior to the main study we performed a brief feasibility study of the use of a portable polysomnography (PSG) system. Of 12 subjects studied, four were unable to sleep wearing the PSG system. All four had suffered severe, total anterior circulation strokes, three were dysphasic and the fourth was mildly confused. Previous investigators have also reported that cognitively impaired older subjects tolerate polysomnography poorly.21 As we intended to study more elderly and severely impaired subjects than those studied previously, we opted for a less complicated and hopefully better tolerated system.
The Autoset system and other simplified sleep systems have been used in other studies of SDB post stroke,5,18,19 but they have the disadvantage that sleep duration and quality can not be assessed. As AHI determined by these systems is expressed in relation to study time rather than sleep time, its value may, if anything, tend to underestimate the severity of SDB. Another potential disadvantage is that the system does not allow confident distinction between obstructive and central sleep-breathing events (particularly hypopnoeas), and therefore we did not attempt to classify the events recorded in this way.
Cross-sectional study
There are a number of possible reasons for the higher prevalence of SDB in our subjects compared to previous studies. The prevalence increases with age in the ‘healthy’ population and the average age of our patients (73 years) was higher than in most other studies. It is, however, similar to that of the population admitted to our stroke unit and to epidemiological studies of acute stroke. Our patients appear to have been more handicapped before their stroke than those studied by previous groups. We recruited only patients who were independent for personal activities of daily living, but those with a history of cerebrovascular disease were not excluded. Although we did not recruit patients with a pre-stroke modified Rankin score >2, the study subjects often had pre-existing minor handicap, and this was associated with a greater frequency of SDB. Also, the prevalence of previous stroke or TIA was lower than in the typical population suffering stroke in Newcastle upon Tyne (20% vs. 29%). Therefore, while the average pre-stroke handicap exhibited by our study population was greater than in previous studies, it was still less than that found in our actual practice. Hence, the severity of SDB amongst all stroke patients may be even higher than reported in this study.
We did not study a control group but a recent study of healthy subjects from the same community with mean age 68 years reported a prevalence of SDB (AHI 10) of 30%.22 This probably represents the best estimate of the prevalence of respiratory sleep disorders in the population from which our subjects were drawn.
The only pre-stroke factor related significantly to stroke severity was sleepiness. This association has not previously been described, but a recent study has shown that sleepiness is related to cardiovascular mortality.23 Sleepiness has many possible causes and it might be a surrogate marker for some other factor such as pre-stroke handicap.
We found that patients with lacunar strokes had worse SDB than those with anterior circulation cortical strokes. A similar finding was recently reported by Lawrence et al.24 At first sight this is surprising, given that the level of neurological impairment following cortical strokes was greater than in our patients with lacunar strokes. Obstructive sleep apnoea is, however, now recognized to be related independently to systemic hypertension25 and it is accompanied by considerable variation in nocturnal blood pressure,26 reduced cerebral blood flow during apnoeas27 and recurrent episodes of hypoxia. These factors might be more likely to contribute to the development of lacunar rather than cortical stroke, through effects on small vessel perfusion. An alternative explanation is that lacunar stroke is often associated with diffuse white-matter ischaemia (leukoaraiosis), and this may be a greater risk factor for SDB than isolated cortical infarcts. White-matter disease might also be relevant to the increasing prevalence of sleep-disordered breathing in the older population in general, and might account for the association we observed between pre-morbid handicap and SDB following stroke. Davies and colleagues28 recently reported that subjects with typical Obstructive Sleep Apnoea Syndrome (OSAS) have no more cerebral white-matter disease than did controls, which argues against an association between symptomatic SDB and white-matter disease. However, their subjects were appreciably younger and, while our patients may have had frequent nocturnal apnoeas, most did not display the typical features of OSAS. It remains possible that white-matter disease contributes to SDB in older subjects, particularly those who have suffered a stroke.
Nocturnal hypoxaemia, defined as a median oxygen saturation <95%, was associated with higher mortality at 3 months, but the patients who died had had more severe strokes and in multiple regression analysis hypoxaemia was not an independent risk factor for mortality. However, the number of deaths was relatively small and the study was not adequately powered to evaluate such an association.
Longitudinal study
Our findings show some differences from those of two other recently published longitudinal studies of respiratory sleep disorders following stroke. Parra et al.19 reported a small decline in mean AHI between 48 h and 3 months following stroke but changes in the proportions of patients with either mild (AHI 10) or severe (AHI >30) SDB were not significant. The average age of the population studied (72 years) was similar to ours, but only patients with first strokes or transient ischaemic attacks were included; post-stroke disability was considerably less than in our patients, suggesting that pre-stroke handicap was probably also less. Lawrence and colleagues24 measured the desaturation index (DI) at 1 week and 3 months following stroke. They found no difference between the two studies, but they did not report values of AHI.
Conclusions
The prevalence of sleep-disordered breathing following stroke in our patients was higher than in most previous studies, probably because our patients were more typical of the stroke population in terms of age and severity of disability. On average, SDB improved in the first few weeks following stroke, but a high prevalence persisted after 6–9 weeks. Sleep-disordered breathing following stroke was associated with pre-stroke handicap but not with stroke severity or extent of recovery.
|
Acknowledgments |
|---|
Dr Joseph Harbison held a National Health Service (Northern & Yorkshire), Research and Development Clinical Research Fellowship. We gratefully acknowledge the generosity of ResMed UK Ltd, in providing diagnostic equipment for this study and Dr Neil Rajan for his assistance.
|
Notes |
|---|
Address correspondence to Professor G.J. Gibson, Department of Respiratory Medicine, Freeman Hospital, High Heaton, Newcastle upon Tyne NE7 7DN. e-mail: g.j.gibson{at}ncl.ac.uk
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Harbison JA, Gibson GJ. Snoring, sleep apnoea and stroke: chicken or scrambled egg? Q J Med2000; 93:647–54.
2. Bradley PA, Mortimore IL, Douglas NJ. Comparison of polysomnography with ResCare Autoset in the diagnosis of the sleep apnoea/hypopnoea syndrome. Thorax1995; 50:1201–3.[Abstract]
3. Kiely JL, Delahunty C, Matthews S, McNicholas WT. Comparison of a limited computerized diagnostic system (ResCare Autoset) with polysomnography in the diagnosis of obstructive of obstructive sleep apnoea syndrome. Eur Respir J1996; 9:2360–4.[Abstract]
4. Fleury B, Rakotonanahary D, Hausser-Hauw C, Lebau B, Guimmenault C. A laboratory Validation Study of the Daignostic Mode of the AutoSet system for sleep related respiratory disorders. Sleep1996; 19:503–5.
5. Bassetti C, Milanova M, Gugger M. Sleep apnea in acute stroke: diagnosis and treatment with an intelligent CPAP. Cerebrovasc Dis 1999; 9 (Suppl. 1):1–128.
6. Rankin J. Cerebral vascular accidents in patients over the age of 60: II. Prognosis. Scot Med J1957; 2:200–15.[Medline]
7. Farrell B, Godwin J, Richards S, Warlow C. The United Kingdom transient ischaemic attack (UK-TIA) aspirin trial: final results. J Neurol Neurosurg Psychiatry1991; 54:1044–54.[Abstract]
8. Mahoney RI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J1965; 14:61–3.[Medline]
9. Scandinavian Stroke Study Group. Multicenter trial of hemodilution in ischemic stroke: background and study protocol. Stroke1985; 16:885–90.
10. Mead GH, Wardlaw JM, Dennis M, Lewis SC, Warlow CP. Relationship between pattern of intracranial artery abnormalities on transcranial doppler and Oxfordshire Community Stroke Project clinical classification of ischemic stroke. Stroke2000; 31:714–19.
11. Johns MW. Sensitivity and specificity of the multiple sleep latency test (MSLT), the maintenance of wakefulness test and the Epworth sleepiness scale: failure of the MSLT as a gold standard. J Sleep Research2000; 9:5–11.[ISI][Medline]
12. Zimmermann C, Kohler D, Schonhofer B. Value of retrospective assessment of the Epworth Sleepiness Scale after long-term CPAP therapy in obstructive sleep apnoea disorder. Pneumologie2000; 54:572–4.[Medline]
13. Briones B, Adams N, Strauss M, Rosenberg C, Whalen C, Carskadon M, Roebuck T, Winters M, Redline S. Relationship between sleepiness and general health status. Sleep1996; 19:583–8.[ISI][Medline]
14. Mohsenin V, Valor R. Sleep apnea in patients with hemispheric stroke. Arch Phys Med Rehabil1995; 76:71–6.[ISI][Medline]
15. Dyken ME, Somers VK, Yamada T, Ren ZY, Zimmerman MB. Investigating the relationship between stroke and obstructive sleep apnoea. Stroke1996; 27:401–7.
16. Bassetti C, Aldrich MS. Sleep apnea in acute cerebrovascular diseases. final report on 128 patients. Sleep1999; 22:217–23.[ISI][Medline]
17. Wessendorf TE, Teschler H, Wang YM, Konietzko N, Thilmann AF. Sleep-disordered breathing among patients with first-ever stroke. J Neurol2000; 247:41–7.[ISI][Medline]
18. Sandberg O, Franklin KA, Bucht G, Gustafson Y. Sleep apnea, delirium, depressed mood, cognition, and ADL ability after stroke. J Am Geriatr Soc2001; 49:391–7.[ISI][Medline]
19. Parra O, Arboix A, Bechich S, Garcia-Eroles L, Monserrat J, Lopez JA, Ballester E, Guerra JM, Sopena JJ. Time course of sleep-related breathing disorders in first-ever stroke or transient ischaemic attack. Am J Respir Crit Care Med2000; 161:375–80.
20. Good DC, Henkle JQ, Gelber D, Welsh J, Verhulst S. Sleep-disordered breathing and poor functional outcome after stroke. Stroke1996; 27:252–9.
21. Ancoli-Israel S, Clopton P, Klauber MR, Fell R, Mason W. Use of wrist activity for monitoring sleep/wake in demented nursing-home patients. Sleep1997; 20:24–7.[ISI][Medline]
22. Davies DP, Oates J, Close P, Rodgers H, James OFW, Gibson GJ. Is obstructive sleep apnoea associated with stroke? Eur Respir J 1997; 10 (Suppl. 25):135.
23. Whitney CW, Enright PL, Newman AB, Bonekat W, Foley D, Quan SF. Correlates of daytime sleepiness in 4578 elderly persons: the Cardiovascular Health Study. Sleep1998; 21:27–36.[ISI][Medline]
24. Lawrence E, Dundas R, Higgens S, Howard R, Rudd A, Wolfe C, Williams A. The natural history and associations of sleep disordered breathing in first ever stroke. Int J Clin Pract2001; 9:584–9.
25. Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med2000; 342:1378–84.
26. Davies RJ, Crosby J, Vardi-Visy K, Clarke M, Stradling JR. Non-invasive beat to beat arterial blood pressure during non-REM sleep in obstructive sleep apnoea and snoring. Thorax1994; 49:335–9.[Abstract]
27. Netzer N, Werner P, Jochums I, Lehmen M, Strohl KP. Blood flow of the middle cerebral artery with sleep-disordered breathing. Correlation with obstructive hypopneas. Stroke1998; 29:87–93.
28. Davies CW, Crosby JH, Mullins RL, et al. Case control study of cerebrovascular damage defined by magnetic resonance imaging in patients with OSA and normal matched control subjects. Sleep2001; 24:715–20.[ISI][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. L. Brown, L. D. Lisabeth, M. J. Zupancic, M. Concannon, C. Martin, and R. D. Chervin High Prevalence of Supine Sleep in Ischemic Stroke Patients Stroke, September 1, 2008; 39(9): 2511 - 2514. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. McArdle, R. L. Riha, M. Vennelle, E. L. Coleman, M. S. Dennis, C. P. Warlow, and N. J. Douglas Sleep-Disordered Breathing as a Risk Factor for Cerebrovascular Disease: A Case-Control Study in Patients With Transient Ischemic Attacks Stroke, December 1, 2003; 34(12): 2916 - 2921. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||