Abstract
Background and purpose
Whilst sleep disturbances are associated with stroke, their association with stroke severity is less certain. In the INTERSTROKE study, the association of pre‐morbid sleep disturbance with stroke severity and functional outcome following stroke was evaluated.
Methods
INTERSTROKE is an international case–control study of first acute stroke. This analysis included cases who completed a standardized questionnaire concerning nine symptoms of sleep disturbance (sleep onset latency, duration, quality, nocturnal awakening, napping duration, whether a nap was planned, snoring, snorting and breathing cessation) in the month prior to stroke (n = 2361). Two indices were derived representing sleep disturbance (range 0–9) and obstructive sleep apnoea (range 0–3) symptoms. Logistic regression was used to estimate the magnitude of association between symptoms and stroke severity defined by the modified Rankin Score.
Results
The mean age of participants was 62.9 years, and 42% were female. On multivariable analysis, there was a graded association between increasing number of sleep disturbance symptoms and initially severe stroke (2–3, odds ratio [OR] 1.44, 95% confidence interval [CI] 1.07–1.94; 4–5, OR 1.66, 95% CI 1.23–2.25; >5, OR 2.58, 95% CI 1.83–3.66). Having >5 sleep disturbance symptoms was associated with significantly increased odds of functional deterioration at 1 month (OR 1.54, 95% CI 1.01–2.34). A higher obstructive sleep apnoea score was also associated with significantly increased odds of initially severe stroke (2–3, OR 1.48; 95% CI 1.20–1.83) but not functional deterioration at 1 month (OR 1.19, 95% CI 0.93–1.52).
Conclusions
Sleep disturbance symptoms were common and associated with an increased odds of severe stroke and functional deterioration. Interventions to modify sleep disturbance may help prevent disabling stroke/improve functional outcomes and should be the subject of future research.
Keywords: disability evaluation; sleep; sleep apnea, obstructive; sleep initiation and maintenance disorders; stroke
INTRODUCTION
Severe stroke is associated with increased mortality, morbidity, healthcare cost and disability‐adjusted life‐years [1, 2]. Identifying modifiable determinants of increased stroke severity is an important element in the effort towards reducing the global burden of stroke. Whilst symptoms of sleep disturbance have been associated with increased risk of stroke, their association with functional outcomes has not been extensively studied [3]. Previous research has found significant associations between pre‐morbid sleep disorders, extremes of sleep duration, napping and stroke severity. Results have not been consistent across studies, however, and research has been limited by sample size, incomplete measurement of the spectrum of sleep disturbance, and outcome studied (frequently mortality alone) [4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15]. There are reasons to suspect an association between pre‐morbid sleep disturbance and both stroke severity and subsequent recovery, however. Chronic sleep disturbances, particularly those related to obstructive sleep apnoea (OSA), have been linked with causal aetiologies known to be associated with more severe stroke (e.g., atrial fibrillation, loss of nocturnal blood pressure dipping) [16, 17, 18]. Sleep disturbances may also affect recovery after stroke, either directly, via intermittent hypoxia and neuronal apoptosis, or indirectly, with reduced daytime activity and increased delirium incidence, reducing the intensity of rehabilitation [4, 19].
In a subgroup of stroke patients in the INTERSTROKE study, the association of pre‐admission symptoms of sleep disturbance with both acute stroke severity and functional outcome (mortality and disability) at 1‐month follow‐up was evaluated.
METHODS AND MATERIALS
Population
INTERSTROKE was a large case–control study that recruited participants with acute stroke and age/sex‐matched controls from 32 countries between 2007 and 2015 [20]. Eligible stroke patients were those admitted to hospital with a clinical diagnosis of first acute stroke, as defined by the World Health Organization, with computed tomography or magnetic resonance imaging available or planned within 1 week of presentation, recruited within 72 h of hospital admission. A standardized questionnaire was administered by trained research staff with clinical measurements (e.g., vital signs, anthropometric measurements) recorded at the time of interview and/or from the participant's notes. A valid proxy respondent (spouse or first‐degree relative living in the same home or aware of the participant's medical history and therapies) was employed for those unable to complete the questionnaire due to severe stroke and/or aphasia. For the current analysis, a subset of participants (stroke cases) were included who completed a supplementary questionnaire which incorporated questions about sleep practices (n = 2361). This questionnaire was introduced in July 2012 to consecutive participants in participating centres.
Sleep disturbance
Participants (or proxy respondent) were administered a series of standardized questions about sleep in the month prior to stroke, which have been described previously [3]. In brief, nocturnal sleep duration, self‐reported sleep quality, nocturnal awakening, sleep onset latency (SOL) and daytime sleeping habits were ascertained, in addition to symptoms suggestive of OSA, including snoring, snorting and breathing cessation during sleep. Two cumulative sleep disturbance scores were derived, nominally an OSA score and Sleep Disturbance Symptom Burden. The derived OSA score ranged from 0 to 3, with one point awarded for either endorsing or answering ‘don't know’ to questions about snoring, snorting or breathing cessation. The derived Sleep Disturbance Symptom Burden ranged from 0 to 9, where higher scores represented a greater number of sleep disturbance symptoms (extremes of sleep duration, poor or fair quality sleep, waking more than once at night, SOL, unplanned napping and napping for more than 1 h, and OSA symptoms).
Outcome measures
For the current analyses, the primary outcome measure was severe stroke at initial interview, with severe stroke defined as a modified Rankin Score (mRS) of more than 3 (unable to walk or attend bodily functions without assistance +/− further disability or death) [21]. Secondary outcomes included mortality and change in level of disability from initial assessment to 1‐month follow‐up (measured with mRS), categorized into three groups: improved, deteriorated and unchanged (Appendix S1, Section 1).
Covariates
Covariates for this analysis were selected a priori, guided by the development of a directed acyclic graph, informed by knowledge of prior literature and biological inference of the association between sleep impairment and stroke severity (Appendix S1, Section 2). Potential confounders included age, sex, mRS prior to stroke (categorized as 0, 1 or >1 given the low percentage of participants with mRS >1 prior to stroke), geographical region, marital status (separated/not currently married or currently married or living with partner), prior OSA diagnosis, atrial fibrillation, evidence of previous stroke on imaging, statin prescription pre‐admission, heart failure pre‐admission and hypertension (history of hypertension or adjusted blood pressure >140/90 at admission).
Statistical analysis
Demographic characteristics were described by means and standard deviations (SD) for continuous variables, and numbers/proportions for categorical variables. Differences in distributions were analysed using the Wilcoxon rank sum test or Pearson's chi‐squared test, as appropriate. Missing data were not imputed.
Univariate and multivariable logistic regression were used to determine the association between all sleep disturbance variables and initial stroke severity. Additive models were constructed, with model 1 being unadjusted and model 2 adjusting for age, sex, geographical region, marital status, mRS at baseline and OSA diagnosis. Model 3 (primary model) adjusted for model 2 variables in addition to atrial fibrillation, hypertension, previous stroke on imaging, statin therapy and heart failure history. The associations of both OSA score and the Sleep Disturbance Symptom Burden with secondary outcomes were determined using both univariate and multivariable regression (binary logistic regression for mortality and multinomial for functional change). The multivariable model adjusted for both age and initial mRS at presentation (numeric variables), based on previous research [22].
Covariates that were suspected to modify the association of sleep disturbance with initial stroke severity were also selected a priori, and included stroke subtype (ischaemic or intracranial haemorrhage, ICH) and whether participants awoke with symptoms. The Wald likelihood ratio test was used to test for statistical interactions. Interaction between sleep disturbance variables and geographical region was also tested post hoc, based on our initial analysis. Similarly, when looking at stroke mortality, interactions between OSA score and both body mass index and stroke subtype were tested post hoc. Where a significant interaction was found, subgroup analysis was undertaken.
Sensitivity analyses involved completing analyses that excluded those who had received thrombolytic therapy and those with an mRS >0 prior to stroke. Statistical significance was defined as a two‐tailed p value of ≤0.05. All statistical analyses were undertaken using R statistical software version 1.3.959 [23].
RESULTS
Demographic characteristics
Overall, 2361 participants with responses to questions about sleep practices in the month prior to stroke were included, including 1898 with ischaemic stroke, 458 with ICH and five with an undetermined stroke subtype. The median time from presentation to initial interview was 2 days (interquartile range 1–2 days), with severe stroke reported in 972 participants (41%). At 1 month, 1272 patients (54%) had improved, 697 (30%) were unchanged and 373 (16%) had deteriorated, with 177 (7%) deaths. The mRS at 1 month was unavailable for 16 participants (<1%).
Pre‐admission sleep disturbance symptoms were common and occurred with a higher frequency in those with a severe stroke at presentation (89% of those with severe stroke vs. 75% of those with mild to moderate stroke; p < 0.001). Short sleep duration, impaired sleep quality, nocturnal awakening, napping and potential OSA symptoms were significantly more common in those with an mRS >3 at presentation, as were elevated Sleep Disturbance Symptom Burden and OSA scores (Table 1, Figure 1), with demographic and disease factors varying by stroke severity (Table 1).
TABLE 1.
Distribution of demographic, risk factor and sleep variables by initial stroke severity.
| Characteristic | Mild to moderate, N = 1389 (59%) a | Severe, N = 972 (41%) a | p value b |
|---|---|---|---|
| Demographic/risk factors | |||
| Stroke type | |||
| Ischaemic | 1233 (89%) | 665 (69%) | <0.001 |
| ICH | 156 (11%) | 302 (31%) | |
| Days from hospital attendance to mRS measure | 1.90 (2.14) | 1.95 (1.95) | 0.38 |
| Age, years | 63 (14) | 63 (14) | 0.50 |
| Sex | |||
| Female | 540 (39%) | 455 (47%) | <0.001 |
| Male | 849 (61%) | 517 (53%) | |
| mRS prior to stroke | |||
| 0 | 1108 (80%) | 731 (75%) | 0.005 |
| 1 | 210 (15%) | 162 (17%) | |
| >1 | 71 (5.1%) | 79 (8.1%) | |
| Region | |||
| Western Europe/North America/Australasia | 491 (35%) | 114 (12%) | <0.001 |
| Eastern/Central Europe/Middle East | 272 (20%) | 119 (12%) | |
| Africa | 92 (6.6%) | 117 (12%) | |
| South Asia | 133 (9.6%) | 384 (40%) | |
| China | 226 (16%) | 186 (19%) | |
| Southeast Asia | 78 (5.6%) | 5 (0.5%) | |
| South America | 97 (7.0%) | 47 (4.8%) | |
| Marital status | |||
| Currently married or living with partner | 1007 (72%) | 699 (72%) | 0.75 |
| Separated/not currently married | 382 (28%) | 273 (28%) | |
| History of atrial fibrillation/flutter | 173 (12%) | 162 (17%) | 0.004 |
| Diagnosis of OSA | 29 (2.1%) | 29 (3.0%) | 0.17 |
| Evidence of previous stroke on CT (not consistent with current presentation) | 108 (7.8%) | 53 (5.5%) | 0.028 |
| Statin pre‐admission | 261 (19%) | 101 (10%) | <0.001 |
| History of hypertension or adjusted BP > 140/90 at admission | 1014 (73%) | 776 (80%) | <0.001 |
| Previous history of heart failure | 48 (3.5%) | 19 (2.0%) | 0.031 |
| Sleep symptoms | |||
| Sleep duration | |||
| 6–7 | 566 (41%) | 326 (34%) | <0.001 |
| <6 | 171 (12%) | 226 (23%) | |
| >7 | 652 (47%) | 420 (43%) | |
| Sleep quality | |||
| Good | 833 (60%) | 443 (46%) | <0.001 |
| Fair or bad | 556 (40%) | 527 (54%) | |
| Sleep onset latency | 359 (26%) | 378 (39%) | <0.001 |
| Frequency of nocturnal awakening | |||
| Waking once or less | 881 (63%) | 479 (49%) | <0.001 |
| Waking more than once | 508 (37%) | 491 (51%) | |
| Napping, planned or unplanned | |||
| No nap | 754 (54%) | 480 (49%) | 0.021 |
| Unplanned nap | 228 (16%) | 198 (20%) | |
| Planned nap | 407 (29%) | 294 (30%) | |
| Napping duration | |||
| 1 h nap | 446 (32%) | 268 (28%) | <0.001 |
| >1 h nap | 186 (13%) | 224 (23%) | |
| Snoring | |||
| Never | 518 (37%) | 266 (27%) | <0.001 |
| Don't know | 118 (8.5%) | 75 (7.7%) | |
| Does | 752 (54%) | 631 (65%) | |
| Snorting | |||
| Never | 988 (71%) | 619 (64%) | <0.001 |
| Don't know | 163 (12%) | 144 (15%) | |
| Does | 238 (17%) | 209 (22%) | |
| Breathing cessation | |||
| Never | 947 (68%) | 528 (54%) | <0.001 |
| Don't know | 162 (12%) | 137 (14%) | |
| Does | 280 (20%) | 307 (32%) | |
| OSA score | |||
| 0–1 | 929 (67%) | 517 (53%) | <0.001 |
| 2–3 | 459 (33%) | 455 (47%) | |
| Sleep Disturbance Symptom Number | |||
| 0–1 | 294 (21%) | 107 (11%) | <0.001 |
| 2–3 | 516 (37%) | 283 (29%) | |
| 4–5 | 422 (30%) | 331 (34%) | |
| >5 | 156 (11%) | 248 (26%) | |
Abbreviations: BP, blood pressure; CT, computed tomography; ICH, intracranial haemorrhage; mRS, modified Rankin Score; OSA, obstructive sleep apnoea.
n (%); mean (SD).
Pearson's chi‐squared test; Wilcoxon rank sum test.
FIGURE 1.
Modified Rankin Scale at onset and 1 month by sleep disturbance indices.
Individual sleep symptoms, sleep disturbance indices and initial stroke severity
On multivariable analyses, sleeping <6 h versus 6–7 h at night (OR 1.53, 95% CI 1.14–2.04), SOL (OR 1.50, 95% CI 1.21–1.85), impaired sleep quality (OR 1.39, 95% CI 1.14–1.68), waking more than once (OR 1.44, 95% CI 1.18–1.75), snoring (OR 1.56, 95% CI 1.26–1.93) and snorting (OR 1.58, 95% CI 1.24–2.00) during sleep were associated with a significantly increased odds of severe stroke. Unplanned napping, napping for >1 h and breathing cessation during sleep were associated with significantly increased odds of severe stroke in univariate analysis, but not with multivariable adjustment (Table 2). Long sleep duration (>7 h at night), planned napping and napping for ≤1 h had no significant association with stroke severity, in any adjustment model (Table 2).
TABLE 2.
Association between sleep disturbance symptoms/indices and the odds of initially severe stroke.
| Symptom assessed (reference level) | Model | Potential risk factor levels assessed | ||
|---|---|---|---|---|
| Sleep duration (6–7 h) | <6 h | >7 h | ||
| Model 1 | 2.29 (1.80–2.92) | 1.12 (0.93–1.34) | ||
| Model 2 | 1.57 (1.18–2.09) | 1.17 (0.94–1.44) | ||
| Model 3 | 1.53 (1.14–2.04) | 1.17 (0.94–1.45) | ||
| Nocturnal awakening (waking once or less) | Waking more than once | |||
|---|---|---|---|---|
| Model 1 | 1.78 (1.51–2.10) | |||
| Model 2 | 1.47 (1.21–1.78) | |||
| Model 3 | 1.44 (1.18–1.75) | |||
| Sleep onset latency (no) | Yes | |||
|---|---|---|---|---|
| Model 1 | 1.83 (1.54–2.19) | |||
| Model 2 | 1.55 (1.26–1.91) | |||
| Model 3 | 1.50 (1.21–1.85) | |||
| Sleep quality (good) | Fair or bad | |||
|---|---|---|---|---|
| Model 1 | 1.78 (1.51–2.10) | |||
| Model 2 | 1.42 (1.17–1.71) | |||
| Model 3 | 1.39 (1.14–1.68) | |||
| Napping plan (no nap) | Planned | Unplanned | ||
|---|---|---|---|---|
| Model 1 | 1.13 (0.94–1.37) | 1.36 (1.09–1.70) | ||
| Model 2 | 0.96 (0.77–1.19) | 1.18 (0.90–1.54) | ||
| Model 3 | 0.97 (0.78–1.20) | 1.19 (0.91–1.55) | ||
| Napping duration (no nap) | 1 hour | >1 hour | ||
|---|---|---|---|---|
| Model 1 | 0.95 (0.78–1.15) | 1.90 (1.52–2.38) | ||
| Model 2 | 0.96 (0.77–1.19) | 1.21 (0.93–1.59) | ||
| Model 3 | 0.96 (0.78–1.20) | 1.24 (0.94–1.64) | ||
| Snoring (never) | Don't know | Does | ||
|---|---|---|---|---|
| Model 1 | 1.24 (0.89–1.71) | 1.63 (1.36–1.96) | ||
| Model 2 | 1.23 (0.84–1.78) | 1.60 (1.29–1.97) | ||
| Model 3 | 1.19 (0.81–1.73) | 1.56 (1.26–1.93) | ||
| Snorting (never) | Don't know | Does | ||
|---|---|---|---|---|
| Model 1 | 1.52 (1.18–1.95) | 1.97 (1.62–2.39) | ||
| Model 2 | 1.36 (1.01–1.83) | 1.60 (1.26–2.02) | ||
| Model 3 | 1.32 (0.98–1.77) | 1.58 (1.24–2.00) | ||
| Breathing cessation (never) | Don't know | Does | ||
|---|---|---|---|---|
| Model 1 | 1.41 (1.10–1.80) | 1.40 (1.13–1.73) | ||
| Model 2 | 1.21 (0.91–1.62) | 1.12 (0.87–1.45) | ||
| Model 3 | 1.20 (0.89–1.60) | 1.10 (0.85–1.42) | ||
| OSA score (0–1) | 2–3 | |||
|---|---|---|---|---|
| Model 1 | 1.78 (1.51–2.11) | |||
| Model 2 | 1.51 (1.23–1.87) | |||
| Model 3 | 1.48 (1.20–1.83) | |||
| Sleep disturbance symptoms (0–1) | 2–3 | 4–5 | >5 | |
|---|---|---|---|---|
| Model 1 | 1.51 (1.16–1.97) | 2.16 (1.66–2.81) | 4.37 (3.25–5.90) | |
| Model 2 | 1.47 (1.10–1.98) | 1.70 (1.26–2.29) | 2.71 (1.93–3.81) | |
| Model 3 | 1.44 (1.07–1.94) | 1.66 (1.23–2.25) | 2.58 (1.83–3.66) |
Note: Model 1: univariate (unconditional). Model 2: adjusted for age, sex, mRS at baseline, region, marital status and OSA diagnosis (unconditional). Model 3: adjusted for age, sex, mRS at baseline, region, marital status, history of atrial fibrillation/flutter, OSA diagnosis, previous stroke on imaging, statin pre‐admission, history of hypertension or adjusted BP >140/90 at admission and heart failure pre‐admission (unconditional).
Abbreviations: BP, blood pressure; mRS, modified Rankin Score; OSA, obstructive sleep apnoea.
On multivariable analyses, a higher burden of sleep disturbance symptoms was associated with significantly increased odds of initially severe stroke (>5, OR 2.58, 95% CI 1.83–3.66; 4–5, OR 1.66, 95% CI 1.23–2.25; 2–3, OR 1.44, 95% CI 1.07–1.94 vs. 0–1) (Table 2, Figure 2). An OSA score of 2–3 was also associated with significantly increased odds of initially severe stroke (2–3, OR 1.48, 95% CI 1.20–1.83 vs. 0–1) (Table 2). No significant interaction was found between either Sleep Disturbance Symptom Burden or OSA score by stroke subtype (p = 0.21 and p = 0.17), ‘wake up’ status (p = 0.11 and p = 0.98) or region (p = 0.33 and p = 0.36).
FIGURE 2.
Odds of severe stroke at initial assessment for Sleep Disturbance Symptom Burden categories.
Sleep disturbance indices and outcomes at 1 month
Reporting >5 sleep disturbance symptoms versus 0–1 was significantly associated with functional deterioration by 1 month (OR 1.84, 95% CI 1.3–2.6), but not mortality (OR 0.70, 95% CI 0.40–1.25), following adjustment for age and initial mRS score. An OSA score of 2–3 versus 0–1 was not significantly associated with functional deterioration by 1 month (OR 1.19, 95% CI 0.93–1.52) but was associated with significantly lower odds of mortality (OR 0.54, 95% CI 0.38–0.76) following adjustment for age and initial mRS (Appendix S1, Section 3).
In relation to stroke mortality, there was a significant interaction between OSA score and stroke subtype (p < 0.001), with a stronger association between OSA score and stroke mortality found in patients with ICH (OR 0.24, 95% CI 0.13–0.45) compared with ischaemic stroke (OR 0.78, 95% CI 0.51–1.17). A further interaction analysis by ICH subtype did not reveal difference by lobar versus non‐lobar ICH (p = 0.76), with no significant interaction between OSA score and body mass index (p = 0.48).
Sensitivity analysis
When those who had received thrombolytic therapy were excluded, the association between initially severe stroke and both a higher burden of sleep disturbance symptoms (>5, OR 2.72, 95% CI 1.89–3.94; 4–5, OR 1.67, 95% CI 1.21–2.30; 2–3, OR 1.45, 95% CI 1.06–1.99 vs. 0–1) and OSA score (2–3, OR 1.47, 95% CI 1.18–1.84) remained significant. This was also the case when those with an mRS >0 prior to stroke were excluded, with both a higher burden of sleep disturbance symptoms (>5, OR 2.56, 95% CI 1.72–3.84; 4–5, OR 1.68, 95% CI 1.19–2.37; 2–3, OR 1.54, 95% CI 1.10–2.15 vs. 0–1) and OSA score (2–3, OR 1.41, 95% CI 1.11–1.79) remaining significantly associated with an increased odds of initially severe stroke.
DISCUSSION
In a large, international cohort of cases of acute first stroke, it was found that pre‐admission sleep disturbance was significantly associated with an increased odds of severe stroke and a lower odds of functional improvement at 1 month. A higher OSA score was associated with increased initial stroke severity, having a paradoxical association with mortality (lower at 1 month) conditional on initial severity.
Our results are supported by their consistency with previous prospective cohort studies that have reported significant associations between extremes of sleep duration, napping and insomnia with increased stroke severity or stroke mortality [12, 13, 14]. In addition, in small, retrospective, observational studies, pre‐stroke insomnia and restless leg symptoms have also been associated with higher levels of disability after stroke [5, 15]. The association of sleep disturbances and stroke severity may be related to a variety of mechanisms. Other risk factors known to be associated with increased stroke severity (e.g., atrial fibrillation) may be more prevalent in patients with sleep disturbances [17]. There are some indications of this in our results, as the majority of estimates were attenuated when adjusting for disease risk factors. It can also be expected that pre‐admission sleep disturbance will result in higher risk of sleep impairments during hospitalization. This may increase the risk of delirium and disrupt adherence to rehabilitative regimens, with post‐stroke insomnia previously associated with increased disability at follow‐up [15, 24, 25]. In addition, findings may relate to the effect of key confounders, as sleep disturbance may result from other factors that could negatively impact recovery or mortality (e.g., frailty, comorbidity, age, geographical region and gender) [26, 27, 28]. However, the relationship between sleep and these factors may be bidirectional (e.g., poorly controlled comorbidities may adversely affect sleep, whilst poor sleep may also exacerbate comorbidities) and a significant association between symptoms of insomnia and initial stroke severity was also found following adjustment for many of these factors, although a measure of frailty other than pre‐admission mRS could not be included.
Our findings contribute to an improved understanding of the association between sleep disturbance and stroke severity, given that a broader spectrum of sleep domains than prior studies was measured and both individual and cumulative symptoms were examined. Impaired sleep quality, SOL, nocturnal awakening and short sleep duration all demonstrated a significant association following adjustment, with cumulative symptoms of disturbance demonstrating a graded association with severity. Long sleep duration and short (≤1 h) and unplanned napping were not associated with an increased odds of initially severe stroke, in any adjustment model, with prolonged (>1 h) and unplanned napping being significant in univariate analysis, alone. These findings may inform the development of interventions to modify sleep impairments for prevention of severe stroke, by identifying the symptoms with the strongest association.
Obstructive sleep apnoea is established as a risk factor for stroke [29]. As such, it is an important target of primary prevention studies, with some indications that treatment may reduce the risk of stroke, although further work is needed given the recent findings of the SAVE trial [30, 31, 32]. OSA is also common following stroke, and, in this setting, it has been associated with increased risk of disability and mortality [33, 34, 35]. The impact of treatment post‐stroke is unclear, however, and the subject of an ongoing randomized controlled trial [36]. In this analysis, pre‐morbid OSA symptoms were examined in a cohort of patients diagnosed with stroke. Pre‐stroke OSA has been studied less frequently in relation to stroke severity, with previous observational research indicating some potentially protective effects in ischaemic stroke and subarachnoid haemorrhage patients [37, 38]. The biological explanation proposed in this context is that of ischaemic preconditioning. Ischaemic preconditioning, induced by nocturnal hypoxia, may result in neovascularization, rendering the brain more tolerant of subsequent ischaemia, and potentially reducing oedema in ICH [37, 39, 40, 41]. However, OSA has also been associated with causal aetiologies known to be associated with more severe stroke (e.g., atrial fibrillation, loss of nocturnal blood pressure dipping) [16, 17, 18]. In addition, sustained OSA symptoms following stroke may negatively impact recovery, via recurrent hypoxic injury, reduced cerebrovascular reactivity or sleep fragmentation affecting rehabilitation engagement [19, 35, 42, 43, 44]. In this study, it was found that a higher OSA score, representative of cumulative, potential OSA symptoms, was associated with initially increased severity but lower mortality after adjusting for this initially increased severity. These apparently divergent findings may have occurred due to chance; however, an alternative explanation may stem from careful consideration of these contrasting causal pathways. Higher stroke severity may be a consequence of aetiology and the impact of OSA on recovery, with an otherwise severe stroke potentially protected from mortality by ischaemic preconditioning. Future randomized controlled trials looking at the impact of OSA treatment on incident stroke should also consider looking at stroke severity as a secondary outcome.
Limitations
This study has some limitations. First, sleep disturbance measures were self‐reported or proxy reported, and prior symptoms were ascertained following incident stroke, being subject to recall and misclassification bias. However, subjects were recruited within 72 h of admission, with questionnaires completed within a median time of 2 days. In addition, questions applied to a short time period (1 month) prior to stroke. Secondly, initial severity was not determined at the exact time of stroke onset, and mRS was used to measure stroke severity rather than the National Institutes of Health Stroke Scale [45]. An advantage of using mRS compared to the National Institutes of Health Stroke Scale is the lower proportion of missing data, missing in only 16 participants at 1‐month follow‐up. Finally, residual confounding cannot be ruled out, although an effort was made to adjust for all available, significant, sleep associated, stroke severity risk factors. This approach probably biased results towards the null, as sleep may have a bidirectional relationship with some of these risk factors, with potential adjustment down the causal pathway. One significant variable that was not available, however, was the use of hypnotic medications, which may impact stroke recovery [46, 47], and the lack of significance for long sleep duration, in contrast to other nocturnal sleep symptoms (potentially more likely to be prescribed hypnotics), may also indicate their importance. In addition, established OSA diagnosis was low and unlikely an accurate reflection of OSA in the population.
CONCLUSION
Whilst previous research indicates that sleep disturbance may represent a risk factor for stroke incidence, our findings indicate that sleep disturbance may also be a risk factor for severe stroke. Sleep disturbance represents a potential target for both primary and secondary prevention in this context. Public health interventions targeting improvements in sleep disturbance may reduce both stroke incidence and severity, whilst screening for sleep disturbance symptoms at stroke onset, and further research on optimal management, should be considered a priority given the indication that they could negatively impact recovery.
AUTHOR CONTRIBUTIONS
Christine E. Mc Carthy: Writing – original draft; formal analysis; writing – review and editing; visualization. Salim Yusuf: Investigation; writing – review and editing; conceptualization; funding acquisition. Conor Judge: Formal analysis; writing – review and editing. John Ferguson: Formal analysis; writing – review and editing. Graeme Hankey: Investigation; writing – review and editing. Shahram Oveis Gharan: Investigation; writing – review and editing. Albertino Damasceno: Investigation; writing – review and editing. Helle Klingenberg Iversen: Investigation; writing – review and editing. Annika Rosengren: Investigation; writing – review and editing. Okechukwu Ogah: Investigation; writing – review and editing. Luísa Avezum: Writing – review and editing. Patricio Lopez‐Jaramillo: Investigation; writing – review and editing. Denis Xavier: Investigation; writing – review and editing. Xingyu Wang: Investigation; writing – review and editing. Sumathy Rangarajan: Investigation; writing – review and editing. Martin J. O'Donnell: Investigation; conceptualization; funding acquisition; writing – original draft; writing – review and editing; formal analysis.
FUNDING INFORMATION
Sources of funding for the INTERSTROKE study include the Canadian Institutes of Health Research, Heart and Stroke Foundation of Canada, Canadian Stroke Network, Swedish Research Council, Swedish Heart and Lung Foundation, the Health and Medical Care Committee of the Regional Executive Board, Region Västra Götaland (Sweden), and unrestricted grants from several pharmaceutical companies with major contributions from AstraZeneca, Boehringer Ingelheim (Canada), Pfizer (Canada), MSD, Chest, Heart and Stroke Scotland, and the Stroke Association, with support from the UK Stroke Research Network. The Department of Neurology at the University Duisburg‐Essen received research grants from the German Research Council, German Ministry of Education and Research, European Union, NIH, Bertelsmann Foundation and Heinz‐Nixdorf Foundation.
CONFLICT OF INTEREST STATEMENT
The authors have no conflicts of interest to disclose.
Supporting information
Appendix S1.
Mc Carthy CE, Yusuf S, Judge C, et al. Pre‐morbid sleep disturbance and its association with stroke severity: results from the international INTERSTROKE study. Eur J Neurol. 2024;31:e16193. doi: 10.1111/ene.16193
DATA AVAILABILITY STATEMENT
Information on the design and rationale of INTERSTROKE has been published previously [20]. Individual participant data, or other documents, will not be made available at this time.
REFERENCES
- 1. Katan M, Luft A. Global burden of stroke. Semin Neurol. 2018;38(2):208‐211. doi: 10.1055/s-0038-1649503 [DOI] [PubMed] [Google Scholar]
- 2. Haupenthal D, Kuramatsu JB, Volbers B, et al. Disability‐adjusted life‐years associated with intracerebral hemorrhage and secondary injury. JAMA Netw Open. 2021;4(7):e2115859. doi: 10.1001/jamanetworkopen.2021.15859 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mc Carthy CE, Yusuf S, Judge C, et al. Sleep patterns and the risk of acute stroke: results from the INTERSTROKE international case−control study. Neurology. 2023;100(21):e2191‐e2203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Fulk GD, Boyne P, Hauger M, et al. The impact of sleep disorders on functional recovery and participation following stroke: a systematic review and meta‐analysis. Neurorehabil Neural Repair. 2020;34(11):1050‐1061. doi: 10.1177/1545968320962501 [DOI] [PubMed] [Google Scholar]
- 5. Medeiros CAM, de Bruin PFC, Paiva TR, Coutinho WM, Ponte RP, de Bruin VMS. Clinical outcome after acute ischaemic stroke: the influence of restless legs syndrome: restless legs syndrome and cerebrovascular disease. Eur J Neurol. 2011;18(1):144‐149. doi: 10.1111/j.1468-1331.2010.03099.x [DOI] [PubMed] [Google Scholar]
- 6. Kawachi T, Wada K, Nakamura K, et al. Sleep duration and the risk of mortality from stroke in Japan: the Takayama cohort study. J Epidemiol. 2016;26(3):123‐130. doi: 10.2188/jea.JE20140272 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Pan A, De Silva DA, Yuan JM, Koh WP. Sleep duration and risk of stroke mortality among Chinese adults: Singapore Chinese health study. Stroke. 2014;45(6):1620‐1625. doi: 10.1161/STROKEAHA.114.005181 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kakizaki M, Kuriyama S, Nakaya N, et al. Long sleep duration and cause‐specific mortality according to physical function and self‐rated health: the Ohsaki cohort study. J Sleep Res. 2013;22(2):209‐216. doi: 10.1111/j.1365-2869.2012.01053.x [DOI] [PubMed] [Google Scholar]
- 9. Kim Y, Wilkens LR, Schembre SM, Henderson BE, Kolonel LN, Goodman MT. Insufficient and excessive amounts of sleep increase the risk of premature death from cardiovascular and other diseases: the multiethnic cohort study. Prev Med. 2013;57(4):377‐385. doi: 10.1016/j.ypmed.2013.06.017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Ikehara S, Iso H, Date C, et al. Association of sleep duration with mortality from cardiovascular disease and other causes for Japanese men and women: the JACC study. Sleep. 2009;32(3):295‐301. doi: 10.1093/sleep/32.3.295 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Cai H, Shu XO, Xiang YB, et al. Sleep duration and mortality: a prospective study of 113,138 middle‐aged and elderly Chinese men and women. Sleep. 2015;38(4):529‐536. doi: 10.5665/sleep.4564 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Wang H, Sun J, Sun M, Liu N, Wang M. Relationship of sleep duration with the risk of stroke incidence and stroke mortality: an updated systematic review and dose−response meta‐analysis of prospective cohort studies. Sleep Med. 2021;1:267‐278. doi: 10.1016/j.sleep.2021.11.001 [DOI] [PubMed] [Google Scholar]
- 13. Helbig AK, Stöckl D, Heier M, Ladwig KH, Meisinger C. Symptoms of insomnia and sleep duration and their association with incident strokes: findings from the population‐based MONICA/KORA Augsburg cohort study. PloS One. 2015;10(7):e0134480. doi: 10.1371/journal.pone.0134480 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Tanabe N, Iso H, Seki N, et al. Daytime napping and mortality, with a special reference to cardiovascular disease: the JACC study. Int J Epidemiol. 2010;39(1):233‐243. doi: 10.1093/ije/dyp327 [DOI] [PubMed] [Google Scholar]
- 15. Leppävuori A, Pohjasvaara T, Vataja R, Kaste M, Erkinjuntti T. Insomnia in ischemic stroke patients. Cerebrovasc Dis. 2002;14(2):90‐97. doi: 10.1159/000064737 [DOI] [PubMed] [Google Scholar]
- 16. Gangwisch JE. A review of evidence for the link between sleep duration and hypertension. Am J Hypertens. 2014;27(10):1235‐1242. doi: 10.1093/ajh/hpu071 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Chokesuwattanaskul R, Thongprayoon C, Sharma K, Congrete S, Tanawuttiwat T, Cheungpasitporn W. Associations of sleep quality with incident atrial fibrillation: a meta‐analysis. Intern Med J. 2018;48(8):964‐972. doi: 10.1111/imj.13764 [DOI] [PubMed] [Google Scholar]
- 18. Lo EH, Albers GW, Dichgans M, et al. Circadian biology and stroke. Stroke. 2021;52(6):2180‐2190. doi: 10.1161/STROKEAHA.120.031742 [DOI] [PubMed] [Google Scholar]
- 19. Ferdinand P, Roffe C. Hypoxia after stroke: a review of experimental and clinical evidence. Exp Transl Stroke Med. 2016;8(1):9. doi: 10.1186/s13231-016-0023-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. O'Donnell M, Xavier D, Diener C, et al. Rationale and design of INTERSTROKE: a global case−control study of risk factors for stroke. Neuroepidemiology. 2010;35(1):36‐44. doi: 10.1159/000306058 [DOI] [PubMed] [Google Scholar]
- 21. de Haan R, Limburg M, Bossuyt P, van der Meulen J, Aaronson N. The clinical meaning of Rankin ‘handicap’ grades after stroke. Stroke. 1995;26(11):2027‐2030. doi: 10.1161/01.STR.26.11.2027 [DOI] [PubMed] [Google Scholar]
- 22. Zhang MY, Mlynash M, Sainani KL, Albers GW, Lansberg MG. Ordinal prediction model of 90‐day modified Rankin Scale in ischemic stroke. Front Neurol. 2021;12:727171. doi: 10.3389/fneur.2021.727171 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. R Core Team . R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria; 2013. http://www.R‐project.org/
- 24. Tang WK, Grace Lau C, Mok V, Ungvari GS, Wong KS. Insomnia and health‐related quality of life in stroke. Top Stroke Rehabil. 2015;22(3):201‐207. doi: 10.1179/1074935714Z.0000000026 [DOI] [PubMed] [Google Scholar]
- 25. Glozier N, Moullaali TJ, Sivertsen B, et al. The course and impact of poststroke insomnia in stroke survivors aged 18 to 65 years: results from the Psychosocial Outcomes in StrokE (POISE) study. Cerebrovasc Dis Extra. 2017;7(1):9‐20. doi: 10.1159/000455751 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Mc Carthy CE. Sleep disturbance, sleep disorders and co‐morbidities in the care of the older person. Med Sci. 2021;9(2):31. doi: 10.3390/medsci9020031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Zeng LN, Zong QQ, Yang Y, et al. Gender difference in the prevalence of insomnia: a meta‐analysis of observational studies. Front Psych. 2020;11:577429. doi: 10.3389/fpsyt.2020.577429 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Soldatos CR, Allaert FA, Ohta T, Dikeos DG. How do individuals sleep around the world? Results from a single‐day survey in ten countries. Sleep Med. 2005;6(1):5‐13. doi: 10.1016/j.sleep.2004.10.006 [DOI] [PubMed] [Google Scholar]
- 29. Wang X, Ouyang Y, Wang Z, Zhao G, Liu L, Bi Y. Obstructive sleep apnea and risk of cardiovascular disease and all‐cause mortality: a meta‐analysis of prospective cohort studies. Int J Cardiol. 2013;169(3):207‐214. doi: 10.1016/j.ijcard.2013.08.088 [DOI] [PubMed] [Google Scholar]
- 30. Li J, Zheng D, Loffler KA, et al. Sleep duration and risk of cardiovascular events: the SAVE study. Int J Stroke. 2020;15(8):858‐865. doi: 10.1177/1747493020904913 [DOI] [PubMed] [Google Scholar]
- 31. Lin HJ, Yeh JH, Hsieh MT, Hsu CY. Continuous positive airway pressure with good adherence can reduce risk of stroke in patients with moderate to severe obstructive sleep apnea: an updated systematic review and meta‐analysis. Sleep Med Rev. 2020;54:101354. doi: 10.1016/j.smrv.2020.101354 [DOI] [PubMed] [Google Scholar]
- 32. Peker Y, Thunström E, Glantz H, Eulenburg C. Effect of obstructive sleep apnea and CPAP treatment on cardiovascular outcomes in acute coronary syndrome in the RICCADSA trial. J Clin Med. 2020;9(12):4051. doi: 10.3390/jcm9124051 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Xie W, Zheng F, Song X. Obstructive sleep apnea and serious adverse outcomes in patients with cardiovascular or cerebrovascular disease: a PRISMA‐compliant systematic review and meta‐analysis. Medicine (Baltimore). 2014;93(29):e336. doi: 10.1097/MD.0000000000000336 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Nair R, Radhakrishnan K, Chatterjee A, Gorthi SP, Prabhu VA. Sleep apnea‐predictor of functional outcome in acute ischemic stroke. J Stroke Cerebrovasc Dis. 2019;28(3):807‐814. doi: 10.1016/j.jstrokecerebrovasdis.2018.11.030 [DOI] [PubMed] [Google Scholar]
- 35. Turkington PM, Allgar V, Bamford J, Wanklyn P, Elliott MW. Effect of upper airway obstruction in acute stroke on functional outcome at 6 months. Thorax. 2004;59(5):367‐371. doi: 10.1136/thx.2003.005348 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Brown DL, Durkalski V, Durmer JS, et al. Sleep for stroke management and recovery trial (sleep SMART): rationale and methods. Int J Stroke. 2020;15(8):923‐929. doi: 10.1177/1747493020903979 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Zhang L, Meng R, Shang S, et al. Obstructive sleep apnea before ischemic stroke: clinical relevance to infarction volume and neurological recovery. J Stroke Cerebrovasc Dis. 2019;28(8):2132‐2139. doi: 10.1016/j.jstrokecerebrovasdis.2019.04.008 [DOI] [PubMed] [Google Scholar]
- 38. Kaculini C, Wallace DJ, Haywood AE, et al. Protective effects of obstructive sleep apnea on outcomes after subarachnoid hemorrhage: a nationwide analysis. Neurosurgery. 2020;87(5):1008‐1015. doi: 10.1093/neuros/nyaa242 [DOI] [PubMed] [Google Scholar]
- 39. Festic N, Alejos D, Bansal V, et al. Sleep apnea in patients hospitalized with acute ischemic stroke: underrecognition and associated clinical outcomes. J Clin Sleep Med. 2018;14(1):75‐80. doi: 10.5664/jcsm.6884 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Lapow JM, Dicpinigaitis AJ, Pammal RS, et al. Obstructive sleep apnea confers lower mortality risk in acute ischemic stroke patients treated with endovascular thrombectomy: National Inpatient Sample Analysis 2010–2018. J NeuroIntervent Surg. 2021;14(12):1195‐1199. doi: 10.1136/neurintsurg-2021-018161 [DOI] [PubMed] [Google Scholar]
- 41. He Y, Karabiyikoglu M, Hua Y, Keep RF, Xi G. Ischemic preconditioning attenuates brain edema after experimental intracerebral hemorrhage. Transl Stroke Res. 2012;3(S1):180‐187. doi: 10.1007/s12975-012-0171-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Fei W, Jiao W, Feng X, Chen X, Wang Y. Intermittent hypoxia mimicking obstructive sleep apnea aggravates early brain injury following ICH via neuroinflammation and apoptosis. Mol Med Rep. 2021;24(5):824. doi: 10.3892/mmr.2021.12464 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Gudberg C, Johansen‐Berg H. Sleep and motor learning: implications for physical rehabilitation after stroke. Front Neurol. 2015;6:6. doi: 10.3389/fneur.2015.00241 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Ponsaing LB, Lindberg U, Rostrup E, Iversen HK, Larsson HBW, Jennum P. Impaired cerebrovascular reactivity in obstructive sleep apnea: a case−control study. Sleep Med. 2018;43:7‐13. doi: 10.1016/j.sleep.2017.10.010 [DOI] [PubMed] [Google Scholar]
- 45. Broderick JP, Adeoye O, Elm J. Evolution of the modified Rankin Scale and its use in future stroke trials. Stroke. 2017;48(7):2007‐2012. doi: 10.1161/STROKEAHA.117.017866 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46. Frank B, Fulton RL, Lees KR, Sanders RD, for the VISTA Collaborators . Impact of benzodiazepines on functional outcome and occurrence of pneumonia in stroke: evidence from VISTA. Int J Stroke. 2014;9(7):890‐894. doi: 10.1111/ijs.12148 [DOI] [PubMed] [Google Scholar]
- 47. Jennum P, Baandrup L, Iversen HK, Ibsen R, Kjellberg J. Mortality and use of psychotropic medication in patients with stroke: a population‐wide, register‐based study. BMJ Open. 2016;6(3):e010662. doi: 10.1136/bmjopen-2015-010662 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Appendix S1.
Data Availability Statement
Information on the design and rationale of INTERSTROKE has been published previously [20]. Individual participant data, or other documents, will not be made available at this time.