Echocardiographic Findings in Ischemic Stroke Patients with Obstructive Sleep Apnea

Abstract

Background

Obstructive sleep apnea (OSA) has been associated with cardiac abnormalities. Whether any cardiac dysfunction is present in ischemic stroke patients with OSA is not known. The purpose of this study was to compare echocardiographic findings in ischemic stroke patients with and without OSA.

Methods

Nocturnal polysomnography was performed on 28 ischemic stroke subjects within 7 days of symptom onset. OSA was defined as an apnea-hypopnea index of ≥ 10. Echocardiographic variables were compared between the OSA and non OSA groups using Wilcoxon signed-rank, chi-square, or Fisher’s exact tests.

Results

The 14 (50%) subjects with OSA had comparable cardiac function and structure to those without OSA (n=14). Left ventricular (LV) mass index, LV ejection fraction, LV diastolic function, left atrial area, and right ventricular systolic function were not different between groups. Ischemic stroke subjects, regardless of their OSA status, had LV diastolic dysfunction with preserved systolic function.

Conclusions

Subjects with and without OSA, based on polysomnography in the first 7 days after stroke, have comparable right and left ventricular function.

INTRODUCTION

Stroke is the leading cause of adult disability and the third leading cause of death in the United States.[1] Among patients with ischemic stroke, the prevalence of sleep-disordered breathing exceeds 60%.[2] Untreated obstructive sleep apnea (OSA) has been recognized as a novel and independent risk factor for the development of ischemic stroke.[3, 4] Although the pathogenesis of stroke in sleep apnea patients is not completely understood, it is hypothesized that hemodynamic disturbance, platelet activation, impaired cerebral autoregulation, increased carotid intima-media thickness, inflammation, and vascular endothelial dysfunction may contribute to the development of stroke in OSA patients.[5–7] While the presence of OSA is associated with poor functional outcome and death after stroke, greater severities of OSA also predict poorer stroke recovery and functional outcome.[8–10]

Previous studies have shown that OSA may be associated with structural and functional cardiac changes that predispose OSA patients to heart failure and atrial fibrillation,[11–15] possibly due to the repetitive episodes of hypoxemia and increased afterload associated with apneic sleep [16]. Magalang et al. reported that continuous positive airway pressure (CPAP) reduces right ventricular (RV) volume [17], and Oliveira et al. observed improved left ventricular (LV) diastolic function and left atrial (LA) emptying with the use of CPAP [18]. This supports a causal role of OSA in cardiac dysfunction. Abnormal cardiac structure and function have also been reported in patients with ischemic stroke. An elevated LA volume index was observed in approximately 75% of patients with first ischemic stroke, and the LA volume index was independently associated with stroke and was predictive of post-stroke survival.[19–21] Left ventricular systolic dysfunction has been shown to be independently associated with an increased risk of ischemic stroke, although any systolic dysfunction is only present in approximately a quarter of ischemic stroke patients.[22] Whether cardiac dysfunction is present in stroke patients with sleep apnea has not been studied.

The purpose of this study was to compare cardiac structure and function (measured by transthoracic echocardiograms) in acute ischemic stroke patients with and without OSA. We hypothesized that ischemic stroke patients with OSA have poorer right and left ventricular function than those without OSA, as found in the non-stroke population. Any structural and functional ventricular abnormalities already evident in acute stroke patients with sleep apnea but not in those without it would support the notion that OSA predates stroke.

METHODS

Patients

We performed a cross-sectional study using subjects enrolled in a pilot randomized controlled trial of CPAP for stroke patients (NCT00282815).[23, 24] Patients on the Inpatient Neurology Service at the University of Michigan were eligible if they were over 18 years of age and had an ischemic stroke within 7 days of the planned overnight polysomnogram. Exclusion criteria were decompensated congestive heart failure, myocardial infarction or respiratory arrest within the prior 3 months, severe pneumonia, prior exposure to CPAP, bullous emphysema, previous pneumothorax, acute sinus or ear infection, or hypertension refractory to treatment. Subjects were eligible for the current analysis if they completed the polysomnogram and had a transthoracic echocardiogram performed as part of routine clinical care. No subject had a previous history of sleep apnea based on review of medical records.

Baseline demographic characteristics including age, gender, self-reported height and weight, and medical history were obtained from patients’ clinical and research records, supplemented by interview. Stroke severity was measured by the National Institutes of Health Stroke Scale performed at the time of enrollment by certified personnel. Informed consent was obtained from the subjects or a surrogate. The study was approved by the University of Michigan’s Institutional Review Board.

Polysomnography

Subjects underwent full, standard overnight polysomnography within 7 days of symptom onset, either in their hospital room during their admission or immediately thereafter in the Sleep Research Laboratory of the University of Michigan General Clinical Research Center (GCRC). Sleep apnea severity was measured by the apnea-hypopnea index (AHI), calculated as the total number of apneas and hypopneas per hour of sleep. An apnea was defined as complete cessation of breathing for ≥ 10 seconds. An hypopnea was defined as a 30% or greater reduction of nasal pressure or thermocouple-monitored airflow, for at least 10 seconds, when followed by an oxygen desaturation of ≥ 4%, arousal, or awakening. These criteria resemble, though they do not duplicate, current[25] criteria published after this study was initiated. For this analysis, OSA was defined as an AHI ≥ 10, based on a median split. Seven patients (with AHI>20) underwent a split night protocol where CPAP or sham CPAP titration occurred in the second half of the night.

Echocardiography

Transthoracic 2-dimensional echocardiography was performed for clinical purposes during the acute stroke hospitalization. All studies were performed at the University of Michigan by experienced sonographers and reviewed by a cardiologist, with advanced training in echocardiography, who was masked to sleep and clinical data.

Right ventricular systolic pressure and function, LA enlargement, LV mass index, LV ejection fraction, and diastolic function were assessed. Left ventricular mass index was calculated using a formula (LV mass/ body surface area) based on measurements of left ventricular dimensions in the parasternal long-axis view at the end of diastole.[26] Left atrial area was calculated in standard apical 4-chamber view. The normal value for LA area has been reported to be ≤ 20 cm².[26] Diastolic function was assessed using standard mitral valve inflow velocity ratio and tissue-Doppler annular velocities. The RV systolic pressure was determined by adding the peak pressure gradient of tricuspid regurgitation flow and right atrial pressure gradient as assessed by inferior vena cava dimension and respiratory variation.

Statistical Analysis

The results are reported as medians with interquartile ranges or as numbers and percentages. Continuous variables were compared between groups of ischemic stroke patients with and without OSA using Wilcoxon signed-rank tests. Categorical variables were compared using chi-square or Fisher’s exact tests. Statistical significance was defined as P<0.05 using 2-sided tests.

RESULTS

Twenty-eight subjects were available for analysis. Echocardiograms were performed a median of 1 day (IQR: 1, 2) after the stroke, and polysomnography was performed a median of 3 days (2, 4) after the stroke. Only one echocardiogram was performed subsequent to active CPAP titration. Baseline information for the 14 OSA and 14 non-OSA patients are found in Table 1. No differences between groups existed in demographics or comorbidities. OSA patients had a higher median AHI than non-OSA patients (39 vs 2, P<0.01).

Table 1.

Comparison of clinical characteristics of ischemic stroke subjects with and without sleep apnea.

Patient Characteristics	OSA (N=14) N (%) or median (IQR)	No OSA (N=14) N (%) or median (IQR)	P value
Male	8 (57)	7 (50)	0.70
Age, years	69 (61, 80)	70 (63, 79)	0.68
Body mass index, kg/m2	31 (29, 33)	33 (26, 38)*	0.65
NIH Stroke Scale	5 (3, 10)	7 (4, 13)	0.78
Hypertension	9 (64)	11 (79)	0.40
Hyperlipidemia	9 (64)	5 (36)	0.13
Diabetes mellitus	6 (43)	5 (36)	0.70
Coronary artery disease	1 (7)	0 (0)	1.0
Congestive heart failure	1 (6)	0 (0)	1.0
Prior stroke	3 (21)	2 (14)	1.0
Atrial fibrillation	3 (21)	3 (21)	1.0
Current smoker	2 (14)	0 (0)	0.48
Sleep Study Characteristics
AHI, events/hour	39 (28, 74)	2 (1, 8)	< 0.01
Baseline oxygen saturation, %	95 (93, 97)	95 (94, 96)*	0.84
Lowest oxygen saturation, %	88 (79, 91)	90 (85, 91)*	0.42

^*N=13

NIH Stroke Scale National Institutes of Health Stroke Scale

IQR Interquartile range; AHI Apnea Hypopnea Index

All subjects had qualitatively normal RV systolic function and normal or nearly normal RV systolic pressure. Their median LV ejection fraction was 65% (IQR: 58–71). Left ventricular diastolic dysfunction of any degree, however, was frequent (85%). The LA was enlarged (calculated LA area > 20 cm²) in 36% of stroke subjects.

Ischemic stroke subjects with OSA had comparable cardiac function and structure to those without OSA. Specifically, LV mass index, LV ejection fraction, and LV diastolic function were not different between the two groups (Table 2). A mild to moderate degree of diastolic dysfunction was similarly present in 80% of subjects with or without OSA (P=1.0). The LA area was mildly to moderately enlarged (LA area of 20–40 cm²) in 36% of stroke subjects with and without OSA (P=1.0). Because transthoracic echocardiograms were performed for clinical purposes only, we could adequately estimate the RV systolic pressure in only 6 OSA and 7 non-OSA stroke subjects. All 6 OSA subjects had a normal RV systolic pressure while 85.7% (6/7) of stroke patients without OSA had a normal RV systolic pressure.

Table 2.

Comparison of echocardiographic characteristics in ischemic stroke subjects with and without OSA.

Echocardiographic Measures	Ischemic Stroke with OSA (N=14)	Ischemic Stroke without OSA (N=14)	P value
Right Ventricle
RV systolic pressure, n (%)*			1.0
Normal, < 38 mmHg	6 (100 %)	6 (85.7 %)
Mild elevation, 38–50 mmHg	0 (0%)	1 (14.3 %)
RV systolic function, qualitatively
Normal, n (%)	14 (100%)	14 (100%)	N/A
Left Ventricle
LV mass index, g/m2, median (IQR)	98 (90, 110)	120 (89, 132)	0.32
LV ejection fraction, median (IQR)	63% (58, 68)	68% (63, 71)	0.41
Diastolic function, n (%)†			1.00‡
Normal	1 (10%)	2 (20 %)
Mild dysfunction	8 (80%)	8 (80 %)
Moderate dysfunction	1 (10%)	0 (0%)
Left Atrium
LA area, n (%)			1.0‡
Normal, ≤ 20 cm2	9 (64%)	9 (64%)
Mild enlargement, >20–30 cm2	3 (21%)	5 (36%)
Moderate enlargement, >30–40 cm2	2 (14%)	0 (0%)

RV right ventricle; LV left ventricle; LA left atrium; IQR Interquartile range

^*RV systolic pressure could be estimated in 6 OSA subjects and 7 non-OSA subjects

^†Diastolic function was assessed in 10 OSA subjects and 10 non-OSA subjects

^‡P value corresponds to a comparison of normal vs. any abnormal measurements

DISCUSSION

This cross-sectional polysomnography and echocardiography study suggests that subjects with and without OSA have comparable right and left ventricular function when studied within the first 7 days of acute ischemic stroke. OSA was not associated with worse RV systolic pressure or function, LV ejection fraction or diastolic function, LV mass index, or LA enlargement. Over 80% of the stroke patients, irrespective of their OSA status, had mild LV diastolic dysfunction. This is not surprising given the association between diastolic dysfunction and stroke risk factors such as older age, systemic hypertension, coronary artery disease, and diabetes.[27] These risk factors were similarly prevalent in our OSA and non OSA groups. Limited data from Ambrosi et al. also showed very high prevalence of diastolic dysfunction in stroke patients, although this finding is not commonly reported.[28]

The lack of an association between sleep apnea and echocardiographic abnormalities in our study contrasts with observations from the general OSA population without stroke, who are reported to have impairment of LV systolic function, LV diastolic function, and RV function, along with increased LA size.[11–15, 29] Several possible explanations exist for the lack of association between OSA and cardiac abnormalities in our sample of stroke patients. First, in many cases, OSA may have emerged after the stroke. New OSA may not have had time to impact cardiac function. However, OSA is known to be an independent risk factor for stroke, which makes its preexistence before the stroke likely in many cases. Second, despite mild pre-existing OSA, perhaps not severe enough to affect cardiac structure or function, stroke may exacerbate OSA in some stroke patients. But no infarction location has been associated with post-stroke OSA.[30] Third, OSA may affect stroke patients differently from the general population. After all, OSA in stroke patients is not associated with excessive daytime somnolence or body mass index, in contrast to the general population, which supports the possibility of differential effects.[31] Fourth, other contributors to cardiac dysfunction, such as older age, hypertension, and other comorbidities, may have overshadowed the effects of OSA, which would be less likely in a younger, healthier population than this stroke group. Fifth, those who are willing to enroll in a CPAP clinical trial may not be representative of all stroke patients. However, the high frequency of comorbid conditions and the very high frequency of diastolic dysfunction argue against biased selection of unusually healthy subjects. Diastolic dysfunction with preserved ejection fraction is common in elderly patients with coexisting cardiovascular conditions[27] and in stroke patients,[28] which also suggests that our study group is representative of the general stroke population. Because left ventricular structure and function findings in stroke patients are conflicting, our LV findings are not inconsistent with the literature.[20, 22, 32] Most importantly, the limited sample size may have decreased our ability to detect small differences. However, previous research did detect echocardiographic differences between similar numbers of OSA and non OSA patients without stroke.[13] Post-hoc power calculations suggest that, with an alpha of 0.05 and our sample size of 14 in each group, we had 80% power to detect a difference in LV ejection fraction of 8.8 and LV mass index of 39.6.

The strengths of this study include performance of overnight polysomnography in all study subjects within 7 days of the ischemic stroke. In addition, all echocardiograms were assessed by a masked cardiologist specializing in echocardiography. However, as all of the echocardiograms in this study were performed for clinical purposes only, we were unable to estimate RV systolic pressure in most study subjects.

Ventricular systolic and diastolic dysfunction is often present in patients with sleep apnea. These cardiac changes may be associated with the known increased cardiovascular morbidity and mortality in OSA patients. Despite the hemodynamic stresses elicited by OSA, it was not associated with abnormal cardiac function or structure in our ischemic stroke subjects. The absence of worse cardiac function and structure in ischemic stroke patients with OSA is interesting and may suggest a unique group of sleep apneics that warrants further investigation.