J Neurol DOI 10.1007/s00415-015-7699-2
ORIGINAL COMMUNICATION
Obstructive sleep apnea is independently associated with arterial stiffness in ischemic stroke patients Chung-Yao Chen1,4 • Chia-Ling Chen3,5 • Chung-Chieh Yu2,4
Received: 6 January 2015 / Revised: 23 February 2015 / Accepted: 5 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Obstructive sleep apnea (OSA) is a predictor of all-cause mortality and recurrent vascular events following stroke. However, few studies have investigated the pathophysiology of OSA in ischemic stroke patients. Whether OSA independently increases arterial stiffness in ischemic stroke patients is determined by measuring the carotid– femoral pulse wave velocity (PWV) and via the central augmentation index (AIx). This cross-sectional study consecutively recruited 127 subacute ischemic stroke patients who were admitted to a teaching hospital for inpatient rehabilitation (median age, 61.3 years; IQR 53.6–72.7 years). Vascular measurements were performed following polysomnography. Multivariate linear regression analysis was performed to determine the relationship between arterial stiffness and OSA. Patients with severe OSA were significantly older, had significantly higher PWV and mean blood pressure, and a significantly higher risk of hypertension than those with non-severe OSA. The significant bivariate correlation between AIx@75 and the desaturation
& Chung-Yao Chen [email protected] 1
Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital, No. 222, Maijin Rd., Anle District, Keelung 204, Taiwan
2
Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, Chang Gung Memorial Hospital, Keelung, Taiwan
3
Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital, Taoyuan, Taiwan
4
School of Medicine, College of Medicine, Chang Gung University, Taoyuan, Taiwan
5
Graduate Institute of Early Intervention, College of Medicine, Chang Gung University, Taoyuan, Taiwan
index (DI) (Spearman’s q = 0.182, P = 0.040) became insignificant by multivariate regression analysis. The PWV was significantly correlated with the apnea–hypopnea index (AHI) (Pearson’s r = 0.350, P = 0.000) and DI (Spearman’s q = 0.347, P = 0.000). The correlation between PWV and OSA parameters, including presence of severe OSA, AHI and DI, remained significant by multivariate regression analysis with age, systolic blood pressure, diabetic mellitus, hypertension and the Barthel index as potential confounders. Arterial stiffness is independently associated with OSA, and PWV can be applied as an intermediate endpoint in further intervention trials of ischemic stroke patients with OSA. Keywords Obstructive sleep apnea Ischemic stroke Arterial stiffness Pulse wave velocity Augmentation index Obstructive sleep apnea (OSA) is a negative predictor of all-cause mortality and recurrent vascular events following stroke or transient ischemic attack (TIA) [1]. The primary hypothesized mechanisms of OSA for stroke incidence and prognosis are hypertension (HTN), endothelial damage, atherogenesis, variability in cerebral blood flow, oxygen desaturation, proinflammatory changes and cardiac arrhythmias [1, 2]. However, results of studies focusing on pathophysiology of OSA in ischemic stroke patients were inconclusive [2]. As far as endothelial dysfunction [3] and cardiac arrhythmia [4] is concerned, studies enrolled stroke patients with sleep-disordered breathing (SDB) and found only atrial fibrillation was associated with severe SDB in acute ischemic stroke patients. As the high prevalence of central sleep apnea (CSA) in acute stroke patients [5], these findings cannot apply to post-acute stroke patients with OSA.
123
J Neurol
For inflammation, findings in ischemic stroke patients with OSA have been controversial. Increased inflammatory biomarker levels, such as C-reactive protein (CRP), in acute ischemic stroke patients with OSA have been reported [6], consistent with findings for typical middle-aged OSA patients without stroke. In contrast, Chen et al. [7] identified adaptive antioxidative capacity and decreased CRP levels in subacute ischemic stroke patients with severe OSA. Therefore, inflammation seems to play only a small role in the pathogenesis of OSA on mortality and recurrent vascular events in ischemic stroke patients. Ischemic stroke patients with OSA had increased prevalence of HTN [7, 8]. Furthermore, OSA is found to be an independent risk factor for carotid atherosclerosis, and may outweigh the effect of HTN in ischemic stroke patients [8]. Arterial stiffness (AS), a composite indicator of arterial health, is strongly associated with atherosclerosis [9]. Stiffening of large artery can be accessed via pulse wave velocity (PWV) and the central augmentation index (AIx). The PWV measurement is the most robust and reproducible method to assess AS and considerable evidence suggests that carotid–femoral PWV, the ‘gold standard’ measurement for AS, is a strong predictor of fatal and nonfatal cardiovascular (CV) events [9]. In contrast to PWV, a direct measure of AS, the AIx is indirect and surrogate measures of AS [9]. Most studies of non-stroke OSA participants assessed AS using the PWV, and suggested that OSA is associated with increased PWV magnitude [10]. However, few studies involved patients with ischemic stroke. Cereda et al. [3], who recently examined 37 subacute ischemic stroke patients, demonstrated increased AIx in patients with moderate–severe SDB. However, PWV was not measured and patients with CSA were not excluded, rendering findings not perfectly applicable to OSA pathophysiology for ischemic stroke patients. This study investigates whether OSA independently increases AS in ischemic stroke patients using carotid– femoral PWV and the central AIx. By conducting measurements in the post-acute phase, the effects of acute stroke on the hemodynamic stability and the severity of OSA are avoid, making the findings not confounded for the mechanisms for increased mortality and recurrent vascular events in ischemic stroke patients with OSA.
Materials and methods Participants Post-acute ischemic stroke patients admitted to the Rehabilitation Ward of a teaching hospital were consecutively recruited for this prospective study. Ischemic stroke was diagnosed based on a full clinical assessment with detailed
123
neurological examinations and neuroimaging studies. Exclusion criteria were evidence of overt congestive cardiac failure, second- or third-degree atrio-ventricular block, persistent atrial fibrillation or flutter, advanced renal disease (chronic kidney disease stage 3 or higher), severely compromised consciousness, history of intracranial hemorrhage or malignance, and such unstable medical and neurological conditions as severe infection or uncontrolled diabetic mellitus (DM). Patients with CSA were also excluded. Study protocol was approved from the local ethics committee, and all patients or their next-of-kin when a patient’s communication was impaired gave informed consent. Clinical evaluation A comprehensive history, including the prevalence of risk factors for stroke (i.e., smoking, HTN, dyslipidemia, hypertriglyceridemia, DM, cardiac arrhythmia, previous ischemic stroke) and demographic data were taken at admission. Scores on the National Institutes of Health Stroke Scale [11], which were taken from relevant medical reports, represented initial stroke severity. Stroke subtype was categorized according to the Trial of Org 10172 in Acute Stroke Treatment criteria [12]. Hypertension was diagnosed if systolic and diastolic blood pressure levels were 140 and/or 90 mmHg or more on at least two occasions in the ward, or the history of hypertension with current antihypertensive drug treatment. Neck circumference, body mass index, Barthel index (BI) [13] and Epworth Sleepiness Scale (ESS) [14] were determined on the same day as PSG examination. The BI is widely used to measure the functional outcomes in stroke patients, and the ESS was used to evaluate the propensity for sleep. Polysomnography The PSG was via an Embla N7000 (Somnologica, Iceland). It was conducted at the sleep center from 10:00 pm to 5:00 am. At least 5 h of recorded time was required to validate the sleep study. It included an electrooculogram, a chin and bilateral anterior tibial surface electromyogram, an electrocardiogram, six electroencephalography channels (F3A1, F4-A2, C3-A1, C4-A2, O1-A1, and O2-A2), nasal and oral airflow sensors (nasal pressure cannula and oronasal thermistor), thoracic and abdominal movements sensors (inductance plethysmography), and an oxyhemoglobin saturation detector (finger pulse oximetry). Diagnosis was according to the American Academy of Sleep Medicine Task Force recommendations [15]. Sleep onset latency, sleep efficiency, and the percentage of total sleep time spent in slow-wave sleep and rapid eye movement sleep
J Neurol
were recorded. Apnea was defined as the cessation of airflow for at least 10 s; hypopnea was defined as a reduction of [50 % in airflow for C10 s with either an arousal or oxygen desaturation C3 %; and the oxyhemoglobin desaturation index (DI) was defined as the number of times per hour of sleep that the blood’s oxygen saturation level drops by C3 % from baseline. Each OSA diagnosis was made when [50 % of respiratory events were of obstructive or mixed types. Severe OSA was categorized as [30 apneas and/or hypopneas per sleep hour [apnea–hypopnea index (AHI) [30 events h-1]. Additionally, CSA was classified when C50 % of respiratory events were the central type. The periodic limb movements during sleep (PLMS) were scored using the criteria set by the International Restless Legs Syndrome Study Group [16]. The PLMS index was defined as the number of limb movements per hour during sleep. One author, a board-certified somnologist, visually scored the PSG recording. Pulse wave velocity and pressure waveform analysis by applanation tonometry All vascular measurements were performed in a quiet temperature-controlled (21–23 °C) examination room immediately after the patients woke and before breakfast. Patients refrained from smoking, drinking alcohol or caffeinated beverages for at least 12 h prior to their examination. Each patient’s BP was recorded using a validated oscillometric method (Vital Signs Monitor 300 Series; Welch Allyn, Inc., Skaneateles Falls, NY, USA) in the non-paretic arm 5 min before measurements for pulse wave calibration. The technique of pulse wave analysis using the commercially available SphygmoCorÒ CPV System (AtCor Medical, Sydney, Australia), which determines both AIx and PWV, has been described elsewhere [17–19]. The applanation probe was positioned on the radial artery of the non-paretic arm. The quality of the recording can be obtained using the build-in operator index. Data were discarded when the operator index was less than 90 %. The radial signal was then used to generate a corresponding central (ascending aortic) waveform using the transfer function in the SphygmoCorÒ software (version 9.0), as described previously [20]. This transfer function has been prospectively validated for assessing ascending aortic BP [21], and the system has good repeatability [22]. AIx, a measure of systemic AS [23], was calculated as the difference between the second and first systolic peaks, and expressed as a percentage of pulse pressure. Because the AIx is associated with heart rate, an index normalized for a heart rate of 75 bpm (AIx@75) was used, as suggested by Wilkinson et al. [24]. The aortic PWV was measured by sequentially recording electrocardiography-gated carotid and femoral artery waveforms of
the non-paretic side using the same device, as described [22]. The applanation tonometry would be performed on the contralateral carotid artery, if there was a significant unilateral carotid stenosis. Statistical analyses Statistical analyses were performed using SPSS version 20 (SPSS; Chicago, IL, USA). Normality of the data distribution was examined. Non-normally distributed variables were log-transformed (PWV, AHI, total cholesterol and high-density lipoprotein) to achieve normality, followed by analysis with parametric tests. Geometric means were then calculated and are reported here. Participants were classified into severe and non-severe OSA groups. Categorical variables were compared using a Chi-square test or Fisher exact test, and continuous variables were compared using the unpaired t test or the Mann–Whitney U test to assess the significance of intergroup differences. Relationships between continuous variables were, depending on whether data were normally distributed, evaluated by Pearson or Spearman correlation analysis. Potential covariates were included when they were related to AS parameters (PWV and AIx@75) at P \ 0.2 into the multivariate linear regression model and were then removed by the backward stepwise selection procedure to determine the strength of the association between OSA severity and AS. Parameters of OSA (presence of severe OSA, AHI and DI) were used separately in the model to avoid unintended statistical interaction. Because BP parameters [systolic BP (SBP), diastolic BP and mean BP] were highly correlated, only SBP, the parameter most strongly correlated with the PWV, was included in multivariate regression analysis to avoid multiple collinearity. Values of P \ 0.05 were considered significant.
Results Thirty-five women and 92 men (median age, 61.3; IQR 53.6–72.7) were enrolled. Median time interval between index stroke and PSG study was 2.5 months (IQR 1.2–4.7). Participants in the severe OSA group were predominantly male, significantly older, had significantly higher PWV and mean BP, thicker neck circumference and significantly higher risk of HTN than those in the non-severe OSA group (Table 1). Sleep parameters of both groups were shown in Table 2. Patients with severe OSA spent a small, albeit significantly lower, percentage of time in slow-wave sleep. Notably, AIx@75 was significantly correlated with DI (Spearman’s q = 0.182, P = 0.040), age (Spearman’s q = 0.255, P = 0.004) but not AHI (Spearman’s
123
J Neurol Table 1 Epidemiological, clinical variables and laboratory parameters stratified by AHI [/ B30/h
AHI B30 n = 46
AHI [30 n = 81
P value
Age (years)
56.1 (14.9)
65.4 (11.4)
0.000*
Women (n)
18 (39.1 %)
17 (21.0 %)
0.028*
Stroke etiology (n)
0.753
Large artery
24 (52.2 %)
Cardioembolism
2 (4.3 %)
38 (46.9 %) 5 (6.2 %)
Lacune
8 (17.4 %)
10 (12.3 %)
Other
1 (2.2 %)
1 (1.2 %)
Undetermined
11 (23.9 %)
27 (33.3 %)
Recurrent stroke (n)
8 (17.4 %)
25 (30.9 %)
0.096
BMIa (kg/m2)
24.1 (21.4–26.5)
24.5 (21.3–27.2)
0.740
Neck circ. (cm) HIHSSa (pt)
37.6 (3.1) 11.0 (6.0–15.0)
39.2 (3.4) 8.5 (5.0–11.0)
0.011* 0.147
Barthel index (pt)
42.0 (21.6)
39.1 (19.4)
0.451
a
ESS (pt)
7.0 (4.0–12.0)
9.0 (6.0–13.5)
0.307
Interval of PSGa,c (months)
2.1 (1.2–4.5)
2.8 (1.2–4.8)
0.374
SBP (mmHg)
125.8 (12.9)
130.5 (14.9)
0.076
DBP (mmHg)
78.1 (8.5)
81.5 (10.7)
0.061
MBP (mmHg)
94.0 (8.9)
97.9 (11.2)
0.046*
Risk factor (n) Hypertension
31 (67.4 %)
70 (86.4 %)
0.011*
Diabetes
20 (43.5 %)
37 (45.7 %)
0.811
Smoking
21 (46.7 %)
39 (48.1 %)
0.873
Dyslipidemia
42 (91.3 %)
66 (81.5 %)
0.136
Total cholesterolb (mg/dL)
174.8
173.3
0.867
HDLb (mg/dL)
39.7
38.2
0.457
LDLa (mg/dL)
111.0 (80.0–141.5)
108.5 (92.0–136.0)
0.783
8.16 (1.80) 25.9 (8.6)
9.61 (2.05) 28.1 (7.3)
0.000** 0.122
PWV (m/s) AIx@75 (%)
Values are means (SD), unless indicated otherwise PWV pulse wave velocity, AIx@75 augmentation index adjusted for heart rate, NIHSS National Institutes of Health Stroke Scale, ESS Epworth Sleepiness Scale, BMI body mass index, Neck circ. neck circumference, HDL high-density lipoprotein, LDL low-density lipoprotein, SBP systolic blood pressure, DBP diastolic blood pressure, MBP mean blood pressure * P \ 0.05; ** P \ 0.01 a
Medium (IQR)
b
Geometric mean
c
Time interval from stroke onset to the date of PSG study
q = 0.148, P = 0.098) nor SBP (Spearman’s q = 0.157, P = 0.078). Stepwise multivariate regression demonstrated that only age was significantly correlated with AIx@75 (P = 0.001) using AIx@75 as the dependent variable, and DI, age, and SBP as the independent variables. Several variables were significantly associated with the PWV level (Table 3). The PWV was significantly higher in participants with HTN than those without and a trend existed that patients with DM had a high PWV (P = 0.058). The PWV was positively correlated with age, AHI, DI, SBP,
123
diastolic BP, mean BP and negatively correlated with BI. The correlation between PWV and parameters of OSA, including the presence of severe OSA, AHI and DI, remained statistically significant in multivariate stepwise regression analysis using age, SBP, DM, HTN and BI as potential confounders (Table 4). The findings that unstandardized regression coefficient of PWV on the presence of severe OSA was 0.685 (P = 0.039) indicated that the presence of severe OSA would increase risk-adjusted PWV of 0.685 m/s in ischemic stroke patients.
J Neurol Table 2 Sleep parameters stratified by AHI [/B30/h
AHI B30 n = 46
AHI [30 n = 81
P value
AHIb (events/h)
13.8
49.6
0.000**
DIa (events/h)
9.8 (5.8–16.5)
46.7 (30.3–59.5)
0.000**
PLMS indexa (movements/h)
1.05 (0–5.35)
0.60 (0–5.5)
0.655
Total sleep time (min)
365.0 (38.6)
354.2 (36.0)
0.116
Sleep onset latencya (min)
18.5 (10.5–34.0)
15.8 (8.3–26.9)
0.212
Sleep efficiencya (%)
74.9 (60.3–87.1)
72.6 (61.2–82.8)
0.500
REM sleep (%)
12.6 (8.0)
10.7 (7.2)
0.176
Slow-wave sleep (%)
19.4 (8.0)
15.1 (11.7)
0.016*
Values are means (SD), unless indicated otherwise AHI apnea–hypopnea index, DI desaturation index, PLMS periodic limb movements during sleep, REM rapid eye movement * P \ 0.05, ** P \ 0.01 a
Medium (IQR)
b
Geometric mean
Discussion To the best of our knowledge, this is the first study to demonstrate that OSA is independently associated with increased PWV, a robust signature of increased AS, in subacute ischemic stroke patients after adjusting common covariates. In contrast to previous findings that post-stroke patients with SDB had a significantly increased AIx [3], this study demonstrates that OSA is not independently associated with AIx@75. Different study designs, including methods of AIx assessment and cut-off value for OSA, keep the results of these two studies from direct comparison. As AIx is dependent on heart rate through ejection duration [25], lack of adjustment for heart rate in their study likely accounts for these different findings. Further, that study recruited patients with TIA or minor stroke and did not exclude patients with CSA, another possible reason for different findings. Contrary to considerable evidences supporting the prognostic values of PWV in stroke patients [26, 27], the prognostic value of AIx in ischemic stroke patients is unclear. Gasecki et al. [27] found AIx had no predictive power for functional outcome after adjusting for covariates in acute ischemic stroke patients. Increased AIx has even been reported to be paradoxically associated with lower in-hospital mortality in ischemic stroke patients [28]. As AIx depends both on arterial elasticity and ventricular contractility [9], the analytical results obtained by this study do not suggest that AIx is a surrogate marker of the impact of OSA on AS in ischemic stroke patients. Given the lack of solid evidence on which antihypertensive medications work best in reducing AS [10], this study did not control for drug class-specific effects.
Obstructive sleep apnea was independently associated with increased PWV after adjusting for age, DM, HTN, SBP and BI in stroke patients. Age and BP, the most wellestablished determinants of PWV found by a systemic review study (91 and 90 % of studies, respectively) [29], were included in the regression analysis to strengthen the independent relationship between OSA and PWV. The presence of severe OSA even plays a more important role than SBP for increased AS. Furthermore, an independent dose–response relationship existed between OSA severity and PWV; this is in agreement with the analytical results reported in recent review, showing that for patients without stroke, OSA is independently associated with AS, either as a dichotomous or continuous variable [10]. Accordingly, this study suggests that PWV is a better surrogate marker of the impact of OSA on AS in ischemic stroke patients. Excessive vessel shear stress due to intrathoracic pressure swings, HTN due to sympathetic activation, oxidative stress, systemic inflammation and endothelial dysfunction have been identified as trigger mechanisms of OSA-induced increased AS in patients without stroke [30]. Since the detrimental impact of OSA on endothelium function [3], inflammation and oxidative stress [7] was not significant in stroke patients, sympathetic activation and intrathoracic pressure swings seem to be the primary factors contributing to increased AS in subacute ischemic stroke patients with OSA. The American Stroke Association recently added the new recommendation in its latest guidelines that sleep study and treatment with continuous positive airway pressure (CPAP) might be considered for ischemic stroke patients, while it acknowledged the large amount of conflicting evidence from single randomized trial or nonrandomized studies [31]. As increased PWV in stroke
123
J Neurol Table 3 Relationships between PWV and covariates (univariate analyses)
Pulse wave velocity (m/s)
Male
t Test P value
Presence of covariate
Absence of covariate
8.98c
8.56c
c
8.10c
0.017*
0.205
Hypertension
9.06
Diabetes
9.47 ± 1.90
8.77 ± 2.17
0.058
Dyslipidemia
9.10 ± 2.10
9.02 ± 1.95
0.885
Hypertriglyceridemia
8.85 ± 1.83
9.18 ± 2.18
0.433
Statin usage
8.90 ± 1.91
9.20 ± 2.18
0.423
Smoking
9.14 ± 1.91
9.05 ± 2.24
0.826
Stroke recurrence
9.45 ± 1.92
8.96 ± 2.12
0.244
Correlation coefficient
P value
Ageb Neck circ.a
0.539 –0.032
0.000** 0.722
BMIa
–0.072
0.422
AHIa
0.350
0.000**
DIb
0.347
0.000**
PLMS index
b
0.079
0.375
SBPa
0.238
0.007**
DBPa
0.193
0.030*
MBPa
0.230
0.009**
cholesterola
–0.022
0.813
logHDLa
–0.029
0.769
LDL
a
–0.024
0.802
NIHSSb
–0.075
0.491
Barthel indexa
–0.266
0.003**
ESSb a
Pearson correlation with PWV
b
Spearman correlation with PWV
c
Geometric mean
0.078
0.379
* P \ 0.05; ** P \ 0.01 NIHSS National Institute of Health Stroke Scale, BMI body mass index, Neck circ. neck circumference, HDL high-density lipoprotein, LDL low-density lipoprotein, AHI apnea–hypopnea index, DI desaturation index, PWV pulse wave velocity, ESS Epworth Sleepiness Scale, SBP systolic blood pressure, DBP diastolic blood pressure, MBP mean blood pressure, PLMS periodic limb movements during sleep
patients was found to be associated with increased risk for poor functional outcomes [18] and mortality [26], increased PWV due to OSA per se seems to be a convincing reason why OSA is a negative predictor of all-cause mortality following stroke or TIA [1]. In their recent metaanalysis of 17 longitudinal studies, Vlachopoulos et al. [32] reported that an increase in aortic PWV by 1 m/s corresponded to a risk factor-adjusted risk increase of 14, 15, and 15 % in total CV events, CV mortality, and all-cause mortality, respectively. The difference of 0.685 m/s in PWV between patients with or without severe OSA offers quantitative evidence that underlines the importance of effective treatment of OSA in ischemic stroke patients. By using the same measuring device in current study, Dariusz et al. [18] found that carotid–femoral PWV C9.4 m/s was
123
an independent predictor for poor functional outcome in ischemic stroke patients. The mean carotid–femoral PWV of 9.6 m/s in the severe OSA group of this study further support that ischemic stroke patients with severe OSA are at great risk of poor functional outcome. Further study on the effects of such treatments as CPAP using PWV as an intermediate endpoint is suggested as the competence of CPAP treatment to alleviate the serious adverse outcomes after ischemic stroke remains controversial [1]. This study has several important limitations. First, as PWV and AIx cannot be measured precisely when heart rate was irregular, study findings cannot be applied to patients with arrhythmia, an important cause of stroke. Second, referral bias cannot be ruled out as patients with TIA or mild stroke receiving outpatient rehabilitation were not
J Neurol Table 4 Multivariate backward stepwise regression analysis of the influence of different variables on pulse wave velocity
Adjusted R2
Model 1 0.353
Variables
b
Severe OSA
Model 2 0.329
Model 3 0.335
b
P
0.164
0.039*
0.000**
0.481
0.000**
0.131
0.077
0.1
0.036*
–0.132 0.137
0.083 0.069
0.160
P
DI 0.434
P
0.180
0.021*
0.485
0.000**
0.151
0.041*
0.039*
AHI Age
b
Hypertension DM Barthel index SBP
b standardized regression coefficient. Parameters of OSA including presence of severe OSA, AHI and DI were used separately in model 1–3 OSA obstructive sleep apnea, AHI apnea–hypopnea index, DI desaturation index, DM diabetic mellitus, SBP systolic blood pressure * P \ 0.05; ** P \ 0.01
enrolled and severe stroke patients with impaired consciousness were excluded. Finally, the causative relationship between OSA and AS was not established in this cross-sectional study.
Summary In conclusion, increased AS is independently associated with OSA and is likely the main mechanism for increased mortality and poor functional outcome in ischemic stroke patients with OSA. Due to the poor compliance of stroke patients for CPAP therapy, further studies should focus on optimal ‘‘de-stiffening’’ strategies, such as positional devices, oral appliance combined with different categories of anti-HTN drugs to ameliorate intrathoracic pressure swings and sympathetic activation in stroke patients with OSA. Acknowledgments Ted Knoy is appreciated for his editorial assistance. The authors would like to thank the Chang Gung Medical Research Council for financially supporting this research under Contract No. CMRPG2E0011. Conflicts of interest On behalf of all authors, the corresponding author states that there is no conflict of interest. Ethical standard Study protocol was approved from the local ethics committee, and all patients or their next-of-kin when a patient’s communication was impaired gave informed consent.
References 1. Birkbak J, Clark AJ, Rod NH (2014) The effect of sleep disordered breathing on the outcome of stroke and transient ischemic attack: a systematic review. J Clin Sleep Med 10(1):103–108
2. Ferre A, Ribo M, Rodriguez-Luna D, Romero O, Sampol G, Molina CA, Alvarez-Sabin J (2013) Strokes and their relationship with sleep and sleep disorders. Neurologia 28(2):103–118 3. Cereda CW, Tamisier R, Manconi M, Andreotti J, Frangi J, Pifferini V, Bassetti CL (2013) Endothelial dysfunction and arterial stiffness in ischemic stroke: the role of sleep-disordered breathing. Stroke 44(4):1175–1178. doi:10.1161/STROKEAHA. 111.000112 4. Shibazaki K, Kimura K, Uemura J, Sakai K, Fujii S, Sakamoto Y, Aoki J (2013) Atrial fibrillation is associated with severe sleepdisordered breathing in patients with ischaemic stroke and transient ischaemic attack. Eur J Neurol 20(2):266–270 5. Parra O, Arboix A, Bechich S, Garcia-Eroles L, Montserrat JM, Lopez JA, Ballester E, Guerra JM, Sopena JJ (2000) Time course of sleep-related breathing disorders in first-ever stroke or transient ischemic attack. Am J Respir Crit Care Med 161(2): 375–380 6. Dziewas R, Ritter M, Kruger L, Berger S, Langer C, Kraus J, Dittrich R, Schabitz WR, Ringelstein EB, Young P (2007) C-reactive protein and fibrinogen in acute stroke patients with and without sleep apnea. Cerebrovasc Dis 24(5):412–417 7. Chen JY, Chen CL, Yu CC, Chen TT, Tseng ST, Ho CH (2014) Association of inflammation and oxidative stress with obstructive sleep apnea in ischemic stroke patients. Sleep Med. doi:10.1016/j. sleep.2014.07.027 8. Dziewas R, Ritter M, Usta N, Boentert M, Hor H, Dittrich R, Schabitz WR, Ringelstein EB, Young P (2007) Atherosclerosis and obstructive sleep apnea in patients with ischemic stroke. Cerebrovasc Dis 24(1):122–126 9. Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, Pannier B, Vlachopoulos C, Wilkinson I, StruijkerBoudier H, On behalf of the European Network for Non-invasive Investigation of Large A (2006) Expert consensus document on arterial stiffness: methodological issues and clinical applications. Eur Heart J 27(21):2588–2605 10. Phillips CL, Butlin M, Wong KK, Avolio AP (2013) Is obstructive sleep apnoea causally related to arterial stiffness? A critical review of the experimental evidence. Sleep Med Rev 17(1):7–18 11. Brott T, Adams HP, Olinger CP, Marler JR, Barsan WG, Biller J, Spilker J, Holleran R, Eberle R, Hertzberg V (1989)
123
J Neurol
12.
13.
14. 15.
16.
17.
18.
19.
20.
21.
22.
Measurements of acute cerebral infarction: a clinical examination scale. Stroke 20(7):864–870 Adams HP, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE (1993) Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 24(1):35–41 Shah S, Vanclay F, Cooper B (1989) Improving the sensitivity of the Barthel index for stroke rehabilitation. J Clin Epidemiol 42(8):703–709. doi:10.1016/0895-4356(89)90065-6 Johns MW (1991) A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 14(6):540–545 American Academy of Sleep Medicine Task Force (1999) Sleeprelated breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep 22(5):667–689 Zucconi M, Ferri R, Allen R, Baier PC, Bruni O, Chokroverty S, Ferini-Strambi L, Fulda S, Garcia-Borreguero D, Hening WA, Hirshkowitz M, Ho¨gl B, Hornyak M, King M, Montagna P, Parrino L, Plazzi G, Terzano MG (2006) The official World Association of Sleep Medicine (WASM) standards for recording and scoring periodic leg movements in sleep (PLMS) and wakefulness (PLMW) developed in collaboration with a task force from the International Restless Legs Syndrome Study Group (IRLSSG). Sleep Med 7(2):175–183. doi:10.1016/j.sleep. 2006.01.001 O’Rourke MF, Blazek JV, Morreels CL, Krovetz LJ (1968) Pressure wave transmission along the human aorta: changes with age and in arterial degenerative disease. Circ Res 23(4):567–579 Dariusz G, Agnieszka R, Mariusz K, Marlena K, Pierre B, Walenty N, Stephane L, Krzysztof N (2012) Aortic stiffness predicts functional outcome in patients after ischemic stroke. Stroke 43:543–544 O’Rourke MF, Adji A (2005) An updated clinical primer on large artery mechanics: implications of pulse waveform analysis and arterial tonometry. Curr Opin Cardiol 20(4):275–281 Karamanoglu M, O’Rourke MF, Avolio AP, Kelly RP (1993) An analysis of the relationship between central aortic and peripheral upper limb pressure waves in man. Eur Heart J 14(2):160–167 Pauca AL, O’Rourke MF, Kon ND (2001) Prospective evaluation of a method for estimating ascending aortic pressure from the radial artery pressure waveform. Hypertension 38(4):932–937 Wilkinson IB, Fuchs SA, Jansen IM, Spratt JC, Murray GD, Cockcroft JR, Webb DJ (1998) Reproducibility of pulse wave velocity and augmentation index measured by pulse wave analysis. J Hypertens 16(12):2079–2084
123
23. Safar ME, London GM (2000) Therapeutic studies and arterial stiffness in hypertension: recommendations of the European Society of Hypertension. J Hypertens 18(11):1527–1535 24. Wilkinson IB, MacCallum H, Flint L, Cockcroft JR, Newby DE, Webb DJ (2000) The influence of heart rate on augmentation index and central arterial pressure in humans. J Physiol 525(1):263–270 25. Adji A, O’Rourke MF, Namasivayam M (2011) Arterial stiffness, its assessment, prognostic value, and implications for treatment. Am J Hypertens 24(1):5–17 26. Kim J, Song TJ, Song D, Lee KJ, Kim EH, Lee HS, Nam CM, Nam HS, Kim YD, Heo JH (2014) Brachial-ankle pulse wave velocity is a strong predictor for mortality in patients with acute stroke. Hypertension 64(2):240–246 27. Gasecki D, Rojek A, Kwarciany M, Kowalczyk K, Boutouyrie P, Nyka W, Laurent S, Narkiewicz K (2012) Pulse wave velocity is associated with early clinical outcome after ischemic stroke. Atherosclerosis 225(2):348–352 28. Tziomalos K, Bouziana SD, Spanou M, Giampatzis V, Papadopoulou M, Kazantzidou P, Kostaki S, Dourliou V, Savopoulos C, Hatzitolios AI (2014) Increased augmentation index is paradoxically associated with lower in-hospital mortality in patients with acute ischemic stroke. Atherosclerosis 236(1): 150–153 29. Cecelja M, Chowienczyk P (2009) Dissociation of aortic pulse wave velocity with risk factors for cardiovascular disease other than hypertension: a systematic review. Hypertension 54(6): 1328–1336 30. Kohler M, Stradling JR (2010) Mechanisms of vascular damage in obstructive sleep apnea. Nat Rev Cardiol 7:677–685 31. Kernan WN, Ovbiagele B, Black HR, Bravata DM, Chimowitz MI, Ezekowitz MD, Fang MC, Fisher M, Furie KL, Heck DV, Johnston SC, Kasner SE, Kittner SJ, Mitchell PH, Rich MW, Richardson D, Schwamm LH, Wilson JA (2014) Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. doi:10.1161/STR.0000000000000024 32. Vlachopoulos C, Aznaouridis K, Stefanadis C (2010) Prediction of cardiovascular events and all-cause mortality with arterial stiffness. a systematic review and meta-analysis. J Am Coll Cardiol 55(13):1318–1327. doi:10.1016/j.jacc.2009.10.061
ORIGINAL COMMUNICATION
Obstructive sleep apnea is independently associated with arterial stiffness in ischemic stroke patients Chung-Yao Chen1,4 • Chia-Ling Chen3,5 • Chung-Chieh Yu2,4
Received: 6 January 2015 / Revised: 23 February 2015 / Accepted: 5 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Obstructive sleep apnea (OSA) is a predictor of all-cause mortality and recurrent vascular events following stroke. However, few studies have investigated the pathophysiology of OSA in ischemic stroke patients. Whether OSA independently increases arterial stiffness in ischemic stroke patients is determined by measuring the carotid– femoral pulse wave velocity (PWV) and via the central augmentation index (AIx). This cross-sectional study consecutively recruited 127 subacute ischemic stroke patients who were admitted to a teaching hospital for inpatient rehabilitation (median age, 61.3 years; IQR 53.6–72.7 years). Vascular measurements were performed following polysomnography. Multivariate linear regression analysis was performed to determine the relationship between arterial stiffness and OSA. Patients with severe OSA were significantly older, had significantly higher PWV and mean blood pressure, and a significantly higher risk of hypertension than those with non-severe OSA. The significant bivariate correlation between AIx@75 and the desaturation
& Chung-Yao Chen [email protected] 1
Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital, No. 222, Maijin Rd., Anle District, Keelung 204, Taiwan
2
Division of Pulmonary, Critical Care and Sleep Medicine, Department of Internal Medicine, Chang Gung Memorial Hospital, Keelung, Taiwan
3
Department of Physical Medicine and Rehabilitation, Chang Gung Memorial Hospital, Taoyuan, Taiwan
4
School of Medicine, College of Medicine, Chang Gung University, Taoyuan, Taiwan
5
Graduate Institute of Early Intervention, College of Medicine, Chang Gung University, Taoyuan, Taiwan
index (DI) (Spearman’s q = 0.182, P = 0.040) became insignificant by multivariate regression analysis. The PWV was significantly correlated with the apnea–hypopnea index (AHI) (Pearson’s r = 0.350, P = 0.000) and DI (Spearman’s q = 0.347, P = 0.000). The correlation between PWV and OSA parameters, including presence of severe OSA, AHI and DI, remained significant by multivariate regression analysis with age, systolic blood pressure, diabetic mellitus, hypertension and the Barthel index as potential confounders. Arterial stiffness is independently associated with OSA, and PWV can be applied as an intermediate endpoint in further intervention trials of ischemic stroke patients with OSA. Keywords Obstructive sleep apnea Ischemic stroke Arterial stiffness Pulse wave velocity Augmentation index Obstructive sleep apnea (OSA) is a negative predictor of all-cause mortality and recurrent vascular events following stroke or transient ischemic attack (TIA) [1]. The primary hypothesized mechanisms of OSA for stroke incidence and prognosis are hypertension (HTN), endothelial damage, atherogenesis, variability in cerebral blood flow, oxygen desaturation, proinflammatory changes and cardiac arrhythmias [1, 2]. However, results of studies focusing on pathophysiology of OSA in ischemic stroke patients were inconclusive [2]. As far as endothelial dysfunction [3] and cardiac arrhythmia [4] is concerned, studies enrolled stroke patients with sleep-disordered breathing (SDB) and found only atrial fibrillation was associated with severe SDB in acute ischemic stroke patients. As the high prevalence of central sleep apnea (CSA) in acute stroke patients [5], these findings cannot apply to post-acute stroke patients with OSA.
123
J Neurol
For inflammation, findings in ischemic stroke patients with OSA have been controversial. Increased inflammatory biomarker levels, such as C-reactive protein (CRP), in acute ischemic stroke patients with OSA have been reported [6], consistent with findings for typical middle-aged OSA patients without stroke. In contrast, Chen et al. [7] identified adaptive antioxidative capacity and decreased CRP levels in subacute ischemic stroke patients with severe OSA. Therefore, inflammation seems to play only a small role in the pathogenesis of OSA on mortality and recurrent vascular events in ischemic stroke patients. Ischemic stroke patients with OSA had increased prevalence of HTN [7, 8]. Furthermore, OSA is found to be an independent risk factor for carotid atherosclerosis, and may outweigh the effect of HTN in ischemic stroke patients [8]. Arterial stiffness (AS), a composite indicator of arterial health, is strongly associated with atherosclerosis [9]. Stiffening of large artery can be accessed via pulse wave velocity (PWV) and the central augmentation index (AIx). The PWV measurement is the most robust and reproducible method to assess AS and considerable evidence suggests that carotid–femoral PWV, the ‘gold standard’ measurement for AS, is a strong predictor of fatal and nonfatal cardiovascular (CV) events [9]. In contrast to PWV, a direct measure of AS, the AIx is indirect and surrogate measures of AS [9]. Most studies of non-stroke OSA participants assessed AS using the PWV, and suggested that OSA is associated with increased PWV magnitude [10]. However, few studies involved patients with ischemic stroke. Cereda et al. [3], who recently examined 37 subacute ischemic stroke patients, demonstrated increased AIx in patients with moderate–severe SDB. However, PWV was not measured and patients with CSA were not excluded, rendering findings not perfectly applicable to OSA pathophysiology for ischemic stroke patients. This study investigates whether OSA independently increases AS in ischemic stroke patients using carotid– femoral PWV and the central AIx. By conducting measurements in the post-acute phase, the effects of acute stroke on the hemodynamic stability and the severity of OSA are avoid, making the findings not confounded for the mechanisms for increased mortality and recurrent vascular events in ischemic stroke patients with OSA.
Materials and methods Participants Post-acute ischemic stroke patients admitted to the Rehabilitation Ward of a teaching hospital were consecutively recruited for this prospective study. Ischemic stroke was diagnosed based on a full clinical assessment with detailed
123
neurological examinations and neuroimaging studies. Exclusion criteria were evidence of overt congestive cardiac failure, second- or third-degree atrio-ventricular block, persistent atrial fibrillation or flutter, advanced renal disease (chronic kidney disease stage 3 or higher), severely compromised consciousness, history of intracranial hemorrhage or malignance, and such unstable medical and neurological conditions as severe infection or uncontrolled diabetic mellitus (DM). Patients with CSA were also excluded. Study protocol was approved from the local ethics committee, and all patients or their next-of-kin when a patient’s communication was impaired gave informed consent. Clinical evaluation A comprehensive history, including the prevalence of risk factors for stroke (i.e., smoking, HTN, dyslipidemia, hypertriglyceridemia, DM, cardiac arrhythmia, previous ischemic stroke) and demographic data were taken at admission. Scores on the National Institutes of Health Stroke Scale [11], which were taken from relevant medical reports, represented initial stroke severity. Stroke subtype was categorized according to the Trial of Org 10172 in Acute Stroke Treatment criteria [12]. Hypertension was diagnosed if systolic and diastolic blood pressure levels were 140 and/or 90 mmHg or more on at least two occasions in the ward, or the history of hypertension with current antihypertensive drug treatment. Neck circumference, body mass index, Barthel index (BI) [13] and Epworth Sleepiness Scale (ESS) [14] were determined on the same day as PSG examination. The BI is widely used to measure the functional outcomes in stroke patients, and the ESS was used to evaluate the propensity for sleep. Polysomnography The PSG was via an Embla N7000 (Somnologica, Iceland). It was conducted at the sleep center from 10:00 pm to 5:00 am. At least 5 h of recorded time was required to validate the sleep study. It included an electrooculogram, a chin and bilateral anterior tibial surface electromyogram, an electrocardiogram, six electroencephalography channels (F3A1, F4-A2, C3-A1, C4-A2, O1-A1, and O2-A2), nasal and oral airflow sensors (nasal pressure cannula and oronasal thermistor), thoracic and abdominal movements sensors (inductance plethysmography), and an oxyhemoglobin saturation detector (finger pulse oximetry). Diagnosis was according to the American Academy of Sleep Medicine Task Force recommendations [15]. Sleep onset latency, sleep efficiency, and the percentage of total sleep time spent in slow-wave sleep and rapid eye movement sleep
J Neurol
were recorded. Apnea was defined as the cessation of airflow for at least 10 s; hypopnea was defined as a reduction of [50 % in airflow for C10 s with either an arousal or oxygen desaturation C3 %; and the oxyhemoglobin desaturation index (DI) was defined as the number of times per hour of sleep that the blood’s oxygen saturation level drops by C3 % from baseline. Each OSA diagnosis was made when [50 % of respiratory events were of obstructive or mixed types. Severe OSA was categorized as [30 apneas and/or hypopneas per sleep hour [apnea–hypopnea index (AHI) [30 events h-1]. Additionally, CSA was classified when C50 % of respiratory events were the central type. The periodic limb movements during sleep (PLMS) were scored using the criteria set by the International Restless Legs Syndrome Study Group [16]. The PLMS index was defined as the number of limb movements per hour during sleep. One author, a board-certified somnologist, visually scored the PSG recording. Pulse wave velocity and pressure waveform analysis by applanation tonometry All vascular measurements were performed in a quiet temperature-controlled (21–23 °C) examination room immediately after the patients woke and before breakfast. Patients refrained from smoking, drinking alcohol or caffeinated beverages for at least 12 h prior to their examination. Each patient’s BP was recorded using a validated oscillometric method (Vital Signs Monitor 300 Series; Welch Allyn, Inc., Skaneateles Falls, NY, USA) in the non-paretic arm 5 min before measurements for pulse wave calibration. The technique of pulse wave analysis using the commercially available SphygmoCorÒ CPV System (AtCor Medical, Sydney, Australia), which determines both AIx and PWV, has been described elsewhere [17–19]. The applanation probe was positioned on the radial artery of the non-paretic arm. The quality of the recording can be obtained using the build-in operator index. Data were discarded when the operator index was less than 90 %. The radial signal was then used to generate a corresponding central (ascending aortic) waveform using the transfer function in the SphygmoCorÒ software (version 9.0), as described previously [20]. This transfer function has been prospectively validated for assessing ascending aortic BP [21], and the system has good repeatability [22]. AIx, a measure of systemic AS [23], was calculated as the difference between the second and first systolic peaks, and expressed as a percentage of pulse pressure. Because the AIx is associated with heart rate, an index normalized for a heart rate of 75 bpm (AIx@75) was used, as suggested by Wilkinson et al. [24]. The aortic PWV was measured by sequentially recording electrocardiography-gated carotid and femoral artery waveforms of
the non-paretic side using the same device, as described [22]. The applanation tonometry would be performed on the contralateral carotid artery, if there was a significant unilateral carotid stenosis. Statistical analyses Statistical analyses were performed using SPSS version 20 (SPSS; Chicago, IL, USA). Normality of the data distribution was examined. Non-normally distributed variables were log-transformed (PWV, AHI, total cholesterol and high-density lipoprotein) to achieve normality, followed by analysis with parametric tests. Geometric means were then calculated and are reported here. Participants were classified into severe and non-severe OSA groups. Categorical variables were compared using a Chi-square test or Fisher exact test, and continuous variables were compared using the unpaired t test or the Mann–Whitney U test to assess the significance of intergroup differences. Relationships between continuous variables were, depending on whether data were normally distributed, evaluated by Pearson or Spearman correlation analysis. Potential covariates were included when they were related to AS parameters (PWV and AIx@75) at P \ 0.2 into the multivariate linear regression model and were then removed by the backward stepwise selection procedure to determine the strength of the association between OSA severity and AS. Parameters of OSA (presence of severe OSA, AHI and DI) were used separately in the model to avoid unintended statistical interaction. Because BP parameters [systolic BP (SBP), diastolic BP and mean BP] were highly correlated, only SBP, the parameter most strongly correlated with the PWV, was included in multivariate regression analysis to avoid multiple collinearity. Values of P \ 0.05 were considered significant.
Results Thirty-five women and 92 men (median age, 61.3; IQR 53.6–72.7) were enrolled. Median time interval between index stroke and PSG study was 2.5 months (IQR 1.2–4.7). Participants in the severe OSA group were predominantly male, significantly older, had significantly higher PWV and mean BP, thicker neck circumference and significantly higher risk of HTN than those in the non-severe OSA group (Table 1). Sleep parameters of both groups were shown in Table 2. Patients with severe OSA spent a small, albeit significantly lower, percentage of time in slow-wave sleep. Notably, AIx@75 was significantly correlated with DI (Spearman’s q = 0.182, P = 0.040), age (Spearman’s q = 0.255, P = 0.004) but not AHI (Spearman’s
123
J Neurol Table 1 Epidemiological, clinical variables and laboratory parameters stratified by AHI [/ B30/h
AHI B30 n = 46
AHI [30 n = 81
P value
Age (years)
56.1 (14.9)
65.4 (11.4)
0.000*
Women (n)
18 (39.1 %)
17 (21.0 %)
0.028*
Stroke etiology (n)
0.753
Large artery
24 (52.2 %)
Cardioembolism
2 (4.3 %)
38 (46.9 %) 5 (6.2 %)
Lacune
8 (17.4 %)
10 (12.3 %)
Other
1 (2.2 %)
1 (1.2 %)
Undetermined
11 (23.9 %)
27 (33.3 %)
Recurrent stroke (n)
8 (17.4 %)
25 (30.9 %)
0.096
BMIa (kg/m2)
24.1 (21.4–26.5)
24.5 (21.3–27.2)
0.740
Neck circ. (cm) HIHSSa (pt)
37.6 (3.1) 11.0 (6.0–15.0)
39.2 (3.4) 8.5 (5.0–11.0)
0.011* 0.147
Barthel index (pt)
42.0 (21.6)
39.1 (19.4)
0.451
a
ESS (pt)
7.0 (4.0–12.0)
9.0 (6.0–13.5)
0.307
Interval of PSGa,c (months)
2.1 (1.2–4.5)
2.8 (1.2–4.8)
0.374
SBP (mmHg)
125.8 (12.9)
130.5 (14.9)
0.076
DBP (mmHg)
78.1 (8.5)
81.5 (10.7)
0.061
MBP (mmHg)
94.0 (8.9)
97.9 (11.2)
0.046*
Risk factor (n) Hypertension
31 (67.4 %)
70 (86.4 %)
0.011*
Diabetes
20 (43.5 %)
37 (45.7 %)
0.811
Smoking
21 (46.7 %)
39 (48.1 %)
0.873
Dyslipidemia
42 (91.3 %)
66 (81.5 %)
0.136
Total cholesterolb (mg/dL)
174.8
173.3
0.867
HDLb (mg/dL)
39.7
38.2
0.457
LDLa (mg/dL)
111.0 (80.0–141.5)
108.5 (92.0–136.0)
0.783
8.16 (1.80) 25.9 (8.6)
9.61 (2.05) 28.1 (7.3)
0.000** 0.122
PWV (m/s) AIx@75 (%)
Values are means (SD), unless indicated otherwise PWV pulse wave velocity, AIx@75 augmentation index adjusted for heart rate, NIHSS National Institutes of Health Stroke Scale, ESS Epworth Sleepiness Scale, BMI body mass index, Neck circ. neck circumference, HDL high-density lipoprotein, LDL low-density lipoprotein, SBP systolic blood pressure, DBP diastolic blood pressure, MBP mean blood pressure * P \ 0.05; ** P \ 0.01 a
Medium (IQR)
b
Geometric mean
c
Time interval from stroke onset to the date of PSG study
q = 0.148, P = 0.098) nor SBP (Spearman’s q = 0.157, P = 0.078). Stepwise multivariate regression demonstrated that only age was significantly correlated with AIx@75 (P = 0.001) using AIx@75 as the dependent variable, and DI, age, and SBP as the independent variables. Several variables were significantly associated with the PWV level (Table 3). The PWV was significantly higher in participants with HTN than those without and a trend existed that patients with DM had a high PWV (P = 0.058). The PWV was positively correlated with age, AHI, DI, SBP,
123
diastolic BP, mean BP and negatively correlated with BI. The correlation between PWV and parameters of OSA, including the presence of severe OSA, AHI and DI, remained statistically significant in multivariate stepwise regression analysis using age, SBP, DM, HTN and BI as potential confounders (Table 4). The findings that unstandardized regression coefficient of PWV on the presence of severe OSA was 0.685 (P = 0.039) indicated that the presence of severe OSA would increase risk-adjusted PWV of 0.685 m/s in ischemic stroke patients.
J Neurol Table 2 Sleep parameters stratified by AHI [/B30/h
AHI B30 n = 46
AHI [30 n = 81
P value
AHIb (events/h)
13.8
49.6
0.000**
DIa (events/h)
9.8 (5.8–16.5)
46.7 (30.3–59.5)
0.000**
PLMS indexa (movements/h)
1.05 (0–5.35)
0.60 (0–5.5)
0.655
Total sleep time (min)
365.0 (38.6)
354.2 (36.0)
0.116
Sleep onset latencya (min)
18.5 (10.5–34.0)
15.8 (8.3–26.9)
0.212
Sleep efficiencya (%)
74.9 (60.3–87.1)
72.6 (61.2–82.8)
0.500
REM sleep (%)
12.6 (8.0)
10.7 (7.2)
0.176
Slow-wave sleep (%)
19.4 (8.0)
15.1 (11.7)
0.016*
Values are means (SD), unless indicated otherwise AHI apnea–hypopnea index, DI desaturation index, PLMS periodic limb movements during sleep, REM rapid eye movement * P \ 0.05, ** P \ 0.01 a
Medium (IQR)
b
Geometric mean
Discussion To the best of our knowledge, this is the first study to demonstrate that OSA is independently associated with increased PWV, a robust signature of increased AS, in subacute ischemic stroke patients after adjusting common covariates. In contrast to previous findings that post-stroke patients with SDB had a significantly increased AIx [3], this study demonstrates that OSA is not independently associated with AIx@75. Different study designs, including methods of AIx assessment and cut-off value for OSA, keep the results of these two studies from direct comparison. As AIx is dependent on heart rate through ejection duration [25], lack of adjustment for heart rate in their study likely accounts for these different findings. Further, that study recruited patients with TIA or minor stroke and did not exclude patients with CSA, another possible reason for different findings. Contrary to considerable evidences supporting the prognostic values of PWV in stroke patients [26, 27], the prognostic value of AIx in ischemic stroke patients is unclear. Gasecki et al. [27] found AIx had no predictive power for functional outcome after adjusting for covariates in acute ischemic stroke patients. Increased AIx has even been reported to be paradoxically associated with lower in-hospital mortality in ischemic stroke patients [28]. As AIx depends both on arterial elasticity and ventricular contractility [9], the analytical results obtained by this study do not suggest that AIx is a surrogate marker of the impact of OSA on AS in ischemic stroke patients. Given the lack of solid evidence on which antihypertensive medications work best in reducing AS [10], this study did not control for drug class-specific effects.
Obstructive sleep apnea was independently associated with increased PWV after adjusting for age, DM, HTN, SBP and BI in stroke patients. Age and BP, the most wellestablished determinants of PWV found by a systemic review study (91 and 90 % of studies, respectively) [29], were included in the regression analysis to strengthen the independent relationship between OSA and PWV. The presence of severe OSA even plays a more important role than SBP for increased AS. Furthermore, an independent dose–response relationship existed between OSA severity and PWV; this is in agreement with the analytical results reported in recent review, showing that for patients without stroke, OSA is independently associated with AS, either as a dichotomous or continuous variable [10]. Accordingly, this study suggests that PWV is a better surrogate marker of the impact of OSA on AS in ischemic stroke patients. Excessive vessel shear stress due to intrathoracic pressure swings, HTN due to sympathetic activation, oxidative stress, systemic inflammation and endothelial dysfunction have been identified as trigger mechanisms of OSA-induced increased AS in patients without stroke [30]. Since the detrimental impact of OSA on endothelium function [3], inflammation and oxidative stress [7] was not significant in stroke patients, sympathetic activation and intrathoracic pressure swings seem to be the primary factors contributing to increased AS in subacute ischemic stroke patients with OSA. The American Stroke Association recently added the new recommendation in its latest guidelines that sleep study and treatment with continuous positive airway pressure (CPAP) might be considered for ischemic stroke patients, while it acknowledged the large amount of conflicting evidence from single randomized trial or nonrandomized studies [31]. As increased PWV in stroke
123
J Neurol Table 3 Relationships between PWV and covariates (univariate analyses)
Pulse wave velocity (m/s)
Male
t Test P value
Presence of covariate
Absence of covariate
8.98c
8.56c
c
8.10c
0.017*
0.205
Hypertension
9.06
Diabetes
9.47 ± 1.90
8.77 ± 2.17
0.058
Dyslipidemia
9.10 ± 2.10
9.02 ± 1.95
0.885
Hypertriglyceridemia
8.85 ± 1.83
9.18 ± 2.18
0.433
Statin usage
8.90 ± 1.91
9.20 ± 2.18
0.423
Smoking
9.14 ± 1.91
9.05 ± 2.24
0.826
Stroke recurrence
9.45 ± 1.92
8.96 ± 2.12
0.244
Correlation coefficient
P value
Ageb Neck circ.a
0.539 –0.032
0.000** 0.722
BMIa
–0.072
0.422
AHIa
0.350
0.000**
DIb
0.347
0.000**
PLMS index
b
0.079
0.375
SBPa
0.238
0.007**
DBPa
0.193
0.030*
MBPa
0.230
0.009**
cholesterola
–0.022
0.813
logHDLa
–0.029
0.769
LDL
a
–0.024
0.802
NIHSSb
–0.075
0.491
Barthel indexa
–0.266
0.003**
ESSb a
Pearson correlation with PWV
b
Spearman correlation with PWV
c
Geometric mean
0.078
0.379
* P \ 0.05; ** P \ 0.01 NIHSS National Institute of Health Stroke Scale, BMI body mass index, Neck circ. neck circumference, HDL high-density lipoprotein, LDL low-density lipoprotein, AHI apnea–hypopnea index, DI desaturation index, PWV pulse wave velocity, ESS Epworth Sleepiness Scale, SBP systolic blood pressure, DBP diastolic blood pressure, MBP mean blood pressure, PLMS periodic limb movements during sleep
patients was found to be associated with increased risk for poor functional outcomes [18] and mortality [26], increased PWV due to OSA per se seems to be a convincing reason why OSA is a negative predictor of all-cause mortality following stroke or TIA [1]. In their recent metaanalysis of 17 longitudinal studies, Vlachopoulos et al. [32] reported that an increase in aortic PWV by 1 m/s corresponded to a risk factor-adjusted risk increase of 14, 15, and 15 % in total CV events, CV mortality, and all-cause mortality, respectively. The difference of 0.685 m/s in PWV between patients with or without severe OSA offers quantitative evidence that underlines the importance of effective treatment of OSA in ischemic stroke patients. By using the same measuring device in current study, Dariusz et al. [18] found that carotid–femoral PWV C9.4 m/s was
123
an independent predictor for poor functional outcome in ischemic stroke patients. The mean carotid–femoral PWV of 9.6 m/s in the severe OSA group of this study further support that ischemic stroke patients with severe OSA are at great risk of poor functional outcome. Further study on the effects of such treatments as CPAP using PWV as an intermediate endpoint is suggested as the competence of CPAP treatment to alleviate the serious adverse outcomes after ischemic stroke remains controversial [1]. This study has several important limitations. First, as PWV and AIx cannot be measured precisely when heart rate was irregular, study findings cannot be applied to patients with arrhythmia, an important cause of stroke. Second, referral bias cannot be ruled out as patients with TIA or mild stroke receiving outpatient rehabilitation were not
J Neurol Table 4 Multivariate backward stepwise regression analysis of the influence of different variables on pulse wave velocity
Adjusted R2
Model 1 0.353
Variables
b
Severe OSA
Model 2 0.329
Model 3 0.335
b
P
0.164
0.039*
0.000**
0.481
0.000**
0.131
0.077
0.1
0.036*
–0.132 0.137
0.083 0.069
0.160
P
DI 0.434
P
0.180
0.021*
0.485
0.000**
0.151
0.041*
0.039*
AHI Age
b
Hypertension DM Barthel index SBP
b standardized regression coefficient. Parameters of OSA including presence of severe OSA, AHI and DI were used separately in model 1–3 OSA obstructive sleep apnea, AHI apnea–hypopnea index, DI desaturation index, DM diabetic mellitus, SBP systolic blood pressure * P \ 0.05; ** P \ 0.01
enrolled and severe stroke patients with impaired consciousness were excluded. Finally, the causative relationship between OSA and AS was not established in this cross-sectional study.
Summary In conclusion, increased AS is independently associated with OSA and is likely the main mechanism for increased mortality and poor functional outcome in ischemic stroke patients with OSA. Due to the poor compliance of stroke patients for CPAP therapy, further studies should focus on optimal ‘‘de-stiffening’’ strategies, such as positional devices, oral appliance combined with different categories of anti-HTN drugs to ameliorate intrathoracic pressure swings and sympathetic activation in stroke patients with OSA. Acknowledgments Ted Knoy is appreciated for his editorial assistance. The authors would like to thank the Chang Gung Medical Research Council for financially supporting this research under Contract No. CMRPG2E0011. Conflicts of interest On behalf of all authors, the corresponding author states that there is no conflict of interest. Ethical standard Study protocol was approved from the local ethics committee, and all patients or their next-of-kin when a patient’s communication was impaired gave informed consent.
References 1. Birkbak J, Clark AJ, Rod NH (2014) The effect of sleep disordered breathing on the outcome of stroke and transient ischemic attack: a systematic review. J Clin Sleep Med 10(1):103–108
2. Ferre A, Ribo M, Rodriguez-Luna D, Romero O, Sampol G, Molina CA, Alvarez-Sabin J (2013) Strokes and their relationship with sleep and sleep disorders. Neurologia 28(2):103–118 3. Cereda CW, Tamisier R, Manconi M, Andreotti J, Frangi J, Pifferini V, Bassetti CL (2013) Endothelial dysfunction and arterial stiffness in ischemic stroke: the role of sleep-disordered breathing. Stroke 44(4):1175–1178. doi:10.1161/STROKEAHA. 111.000112 4. Shibazaki K, Kimura K, Uemura J, Sakai K, Fujii S, Sakamoto Y, Aoki J (2013) Atrial fibrillation is associated with severe sleepdisordered breathing in patients with ischaemic stroke and transient ischaemic attack. Eur J Neurol 20(2):266–270 5. Parra O, Arboix A, Bechich S, Garcia-Eroles L, Montserrat JM, Lopez JA, Ballester E, Guerra JM, Sopena JJ (2000) Time course of sleep-related breathing disorders in first-ever stroke or transient ischemic attack. Am J Respir Crit Care Med 161(2): 375–380 6. Dziewas R, Ritter M, Kruger L, Berger S, Langer C, Kraus J, Dittrich R, Schabitz WR, Ringelstein EB, Young P (2007) C-reactive protein and fibrinogen in acute stroke patients with and without sleep apnea. Cerebrovasc Dis 24(5):412–417 7. Chen JY, Chen CL, Yu CC, Chen TT, Tseng ST, Ho CH (2014) Association of inflammation and oxidative stress with obstructive sleep apnea in ischemic stroke patients. Sleep Med. doi:10.1016/j. sleep.2014.07.027 8. Dziewas R, Ritter M, Usta N, Boentert M, Hor H, Dittrich R, Schabitz WR, Ringelstein EB, Young P (2007) Atherosclerosis and obstructive sleep apnea in patients with ischemic stroke. Cerebrovasc Dis 24(1):122–126 9. Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, Pannier B, Vlachopoulos C, Wilkinson I, StruijkerBoudier H, On behalf of the European Network for Non-invasive Investigation of Large A (2006) Expert consensus document on arterial stiffness: methodological issues and clinical applications. Eur Heart J 27(21):2588–2605 10. Phillips CL, Butlin M, Wong KK, Avolio AP (2013) Is obstructive sleep apnoea causally related to arterial stiffness? A critical review of the experimental evidence. Sleep Med Rev 17(1):7–18 11. Brott T, Adams HP, Olinger CP, Marler JR, Barsan WG, Biller J, Spilker J, Holleran R, Eberle R, Hertzberg V (1989)
123
J Neurol
12.
13.
14. 15.
16.
17.
18.
19.
20.
21.
22.
Measurements of acute cerebral infarction: a clinical examination scale. Stroke 20(7):864–870 Adams HP, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE (1993) Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 24(1):35–41 Shah S, Vanclay F, Cooper B (1989) Improving the sensitivity of the Barthel index for stroke rehabilitation. J Clin Epidemiol 42(8):703–709. doi:10.1016/0895-4356(89)90065-6 Johns MW (1991) A new method for measuring daytime sleepiness: the Epworth sleepiness scale. Sleep 14(6):540–545 American Academy of Sleep Medicine Task Force (1999) Sleeprelated breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep 22(5):667–689 Zucconi M, Ferri R, Allen R, Baier PC, Bruni O, Chokroverty S, Ferini-Strambi L, Fulda S, Garcia-Borreguero D, Hening WA, Hirshkowitz M, Ho¨gl B, Hornyak M, King M, Montagna P, Parrino L, Plazzi G, Terzano MG (2006) The official World Association of Sleep Medicine (WASM) standards for recording and scoring periodic leg movements in sleep (PLMS) and wakefulness (PLMW) developed in collaboration with a task force from the International Restless Legs Syndrome Study Group (IRLSSG). Sleep Med 7(2):175–183. doi:10.1016/j.sleep. 2006.01.001 O’Rourke MF, Blazek JV, Morreels CL, Krovetz LJ (1968) Pressure wave transmission along the human aorta: changes with age and in arterial degenerative disease. Circ Res 23(4):567–579 Dariusz G, Agnieszka R, Mariusz K, Marlena K, Pierre B, Walenty N, Stephane L, Krzysztof N (2012) Aortic stiffness predicts functional outcome in patients after ischemic stroke. Stroke 43:543–544 O’Rourke MF, Adji A (2005) An updated clinical primer on large artery mechanics: implications of pulse waveform analysis and arterial tonometry. Curr Opin Cardiol 20(4):275–281 Karamanoglu M, O’Rourke MF, Avolio AP, Kelly RP (1993) An analysis of the relationship between central aortic and peripheral upper limb pressure waves in man. Eur Heart J 14(2):160–167 Pauca AL, O’Rourke MF, Kon ND (2001) Prospective evaluation of a method for estimating ascending aortic pressure from the radial artery pressure waveform. Hypertension 38(4):932–937 Wilkinson IB, Fuchs SA, Jansen IM, Spratt JC, Murray GD, Cockcroft JR, Webb DJ (1998) Reproducibility of pulse wave velocity and augmentation index measured by pulse wave analysis. J Hypertens 16(12):2079–2084
123
23. Safar ME, London GM (2000) Therapeutic studies and arterial stiffness in hypertension: recommendations of the European Society of Hypertension. J Hypertens 18(11):1527–1535 24. Wilkinson IB, MacCallum H, Flint L, Cockcroft JR, Newby DE, Webb DJ (2000) The influence of heart rate on augmentation index and central arterial pressure in humans. J Physiol 525(1):263–270 25. Adji A, O’Rourke MF, Namasivayam M (2011) Arterial stiffness, its assessment, prognostic value, and implications for treatment. Am J Hypertens 24(1):5–17 26. Kim J, Song TJ, Song D, Lee KJ, Kim EH, Lee HS, Nam CM, Nam HS, Kim YD, Heo JH (2014) Brachial-ankle pulse wave velocity is a strong predictor for mortality in patients with acute stroke. Hypertension 64(2):240–246 27. Gasecki D, Rojek A, Kwarciany M, Kowalczyk K, Boutouyrie P, Nyka W, Laurent S, Narkiewicz K (2012) Pulse wave velocity is associated with early clinical outcome after ischemic stroke. Atherosclerosis 225(2):348–352 28. Tziomalos K, Bouziana SD, Spanou M, Giampatzis V, Papadopoulou M, Kazantzidou P, Kostaki S, Dourliou V, Savopoulos C, Hatzitolios AI (2014) Increased augmentation index is paradoxically associated with lower in-hospital mortality in patients with acute ischemic stroke. Atherosclerosis 236(1): 150–153 29. Cecelja M, Chowienczyk P (2009) Dissociation of aortic pulse wave velocity with risk factors for cardiovascular disease other than hypertension: a systematic review. Hypertension 54(6): 1328–1336 30. Kohler M, Stradling JR (2010) Mechanisms of vascular damage in obstructive sleep apnea. Nat Rev Cardiol 7:677–685 31. Kernan WN, Ovbiagele B, Black HR, Bravata DM, Chimowitz MI, Ezekowitz MD, Fang MC, Fisher M, Furie KL, Heck DV, Johnston SC, Kasner SE, Kittner SJ, Mitchell PH, Rich MW, Richardson D, Schwamm LH, Wilson JA (2014) Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. doi:10.1161/STR.0000000000000024 32. Vlachopoulos C, Aznaouridis K, Stefanadis C (2010) Prediction of cardiovascular events and all-cause mortality with arterial stiffness. a systematic review and meta-analysis. J Am Coll Cardiol 55(13):1318–1327. doi:10.1016/j.jacc.2009.10.061
