Central Sleep Apnea is Uncommon after Stroke

Abstract

Objective/Background:

Stroke is often considered a risk factor for central sleep apnea (CSA). The goal of this study was to determine the prevalence and clinical correlates of CSA in patients with ischemic stroke.

Patients/Methods:

In this analysis, 1,346 participants in the Brain Attack Surveillance in Corpus Christi (BASIC) project underwent a home sleep apnea test shortly after ischemic stroke. Respiratory events during sleep were classified as central apneas, obstructive apneas, or hypopneas. Central apnea index (CAI) was defined as number of central apneas divided by recording time. CSA was defined as CAI ≥ 5/hour with at least 50% of all scored respiratory events classified as central apneas. Demographics and co-morbidities were ascertained from the medical record.

Results:

Median CAI was 0/hour. Nineteen participants (1.4%) met criteria for CSA. Participants with CSA were more likely to be male, and had lower prevalence of obesity than participants without CSA. There was no association between CSA and other co-morbidities.

Conclusions:

CSA was uncommon in this large cohort of patients with recent ischemic stroke.

1. Introduction

Sleep apnea is the most common form of sleep-disordered breathing, its estimated prevalence approaching 20% among adults in industrialized countries¹. In obstructive sleep apnea, the intermittent cessation of airflow during sleep occurs secondary to a narrowing or occlusion of the upper airway. In contrast, central sleep apnea (CSA) is defined as airflow cessation due to lack of respiratory effort. Based on early case series demonstrating Cheyne-Stokes Breathing after stroke ^2–4, many clinicians consider stroke to be a risk factor for CSA ⁵, though recent evidence suggests that the prevalence of CSA after stroke may not be as high as previously thought⁶. The goal of this population-based cohort study was to assess the prevalence of CSA in patients with recent ischemic stroke, and to identify subgroups who may be at increased risk for CSA.

2. Materials and Methods

Participants were recruited from the Brain Attack Surveillance in Corpus Christi (BASIC) project in Nueces County, Texas, a population-based study of adults ≥45 years of age with ischemic stroke. Exclusion criteria were supplemental oxygen, mechanical or positive pressure ventilation, and pregnancy. Methods have been described in detail elsewhere ⁷. Demographic information, NIH stroke scale (NIHSS), and co-morbidities were abstracted from medical records. All participants underwent home sleep apnea testing (HSAT, ApneaLink Plus, Resmed Inc., San Diego, CA) during hospitalization or after discharge from the hospital (median time from stroke recognition to HSAT 13 days, interquartile range 6 to 21 days). Apneas were defined as a decrease in nasal pressure of ≥80% from baseline for ≥10 seconds. Central apneas were differentiated from obstructive apneas based on interruption of respiratory effort during ≥80% of an apneic episode. Hypopneas were defined by a decrease in nasal pressure of ≥30% compared to baseline, lasting for ≥10 seconds and associated with a ≥4% drop in oxygen saturation. If oximetry data were missing, hypopneas were defined as a reduction in nasal pressure of ≥50% for ≥10 seconds. Automated scoring was followed by review by a registered polysomnographic technologist ⁸. The respiratory event index (REI) was defined as the sum of all apneas and hypopneas divided by hours of recording. The central apnea index (CAI) was calculated as number of central apneas divided by recording time. Central apneas detected with ApneaLink Plus® have been shown to be very highly correlated with central apneas detected on gold-standard, in-lab polysomnography (r=0.94, p<0.001)⁹. CSA was defined as CAI ≥5/hour and CAI/REI >0.50 (i.e. more than 50% of all scored respiratory events were central apneas)¹⁰. We use the term CSA though it is not possible to confirm that respiratory events occurred during sleep. HSATs with <2 hours of valid data were excluded. Chi-square, Fisher exact, and t-tests were used to compare subgroups. Written informed consent was provided by all participants or a surrogate. The University of Michigan and Corpus Christi hospital systems’ Institutional Review Boards approved this project.

3. Results

Among 1,346 participants, the median time from stroke recognition to HSAT was 13 days (Interquartile range: 6, 21). The median CAI in the entire analytic cohort was 0/h (IQR: 0, 2) (table 1). Nineteen participants (1.4%) met criteria for CSA. Participants with CSA were more likely to be male (p=0.0215), to have a normal-to-overweight rather than obese body habitus (p=0.0161) and to have severe sleep apnea based on an overall REI > 30/hour (p=0.0053) compared to participants without CSA. Neither stroke severity (NIHSS) (p=0.23), heart failure (p=0.19), nor any other medical condition was associated with CSA (all other p-values >0.40).

Table 1:

Characteristics of individuals with ischemic stroke with or without CSA

		Entire cohort	Participants without vs. with CSA
	N	CAI (median, IQR)	CAI<5/h and/or CAI<50% (n,%)	CAI≥5/h and CAI≥50% (n, %)	p-value
Age (years)					0.2344
I quartile (<57)	305	0 (0,1)	304 (23)	1 (5)
II quartile (57 to <65)	360	0 (0,1)	353 (27)	7 (37)
III quartile (65 to <74)	331	1 (0,2)	325 (24)	6 (32)
IV quartile (≥74)	354	0 (0,2)	349 (26)	5 (26)
Gender					0.0215
Male	711	1 (0,3)	696 (52)	15 (79)
Female	639	0 (0,1)	635 (48)	4 (21)
Race/Ethnicity					0.4604
Non-Hispanic White	435	0 (0,2)	430 (32)	5 (26)
Mexican American	832	0 (0,2)	820 (62)	12 (63)
African American	83	0 (0,1)	81 (6)	2 (11)
Body Mass Index (kg/m2)					0.0161
Low/normal	290	0 (0,1)	282 (21)	8 (42)
Overweight	462	0 (0,2)	454 (34)	8 (42)
Obese	598	0 (0,1)	595 (45)	3 (16)
NIH Stroke Scale					0.2317
I quartile (<1)	190	0 (0,1)	189 (14)	1 (6)
II quartile (1 to <3)	358	0 (0,2)	356 (27)	2 (11)
III quartile (3 to <6)	390	0 (0,2)	383 (29)	7 (39)
IV quartile (≥6)	405	1 (0,2)	397 (30)	8 (44)
Hypertension					0.5257
No	214	0 (0,1)	210 (16)	4 (21)
Yes	1134	0 (0,2)	1119 (84)	15 (79)
Diabetes mellitus					0.5994
No	629	0 (0,2)	619 (47)	10 (53)
Yes	719	0 (0,2)	710 (53)	9 (47)
Atrial Fibrillation					0.4382
No	1206	0 (0,2)	1190 (90)	16 (84)
Yes	140	1 (0,3)	137 (10)	3 (16)
Coronary artery disease or myocardial infarction					0.7247
No	943	0 (0,1)	929 (70)	14 (74)
Yes	403	1 (0,2)	399 (30)	5 (26)
Chronic heart failure					0.1873
No	1239	0 (0,2)	1223 (92)	16 (84)
Yes	107	0 (0,1)	104 (8)	3 (16)
Respiratory event index					0.0053
0–15	695	0 (0,1)	690 (52)	5 (26)
>15–30	379	1 (0,2)	375 (28)	4 (21)
30+	276	3 (1,8)	266 (20)	10 (53)
Central apnea index – median (IQR)	1346	0 (0,2)	0 (0, 2)	17 (9,24)	<.0001
Obstructive apnea index – median (IQR)	1346	4 (2,11)	4 (2, 11)	5 (3, 12)	0.7089
Oxygen saturation <90 (% of recording time) – median (IQR)	1346	1 (0,5)	2 (0, 5)	1 (0,3)	0.7439

4. Discussion

In this large population-based cohort, 1.4% of participants met criteria for CSA after ischemic stroke. These results challenge conventional thought that post-stroke CSA is common, a concept advanced by early small case series focused on likely non-representative samples ^2–4. Our results align with recent evidence suggesting CSA is uncommon in ischemic stroke ^{6, 11–16}, and extend the existing literature through examination of CSA prevalence in different subgroups within a large, ethnically diverse population. CSA was more common in males and in participants with normal or overweight as opposed to obese body habitus, mirroring previous reports in non-stroke populations ¹⁷. We could not show that higher NIHSS scores or chronic heart failure were significantly associated with CSA, although associations were in the expected direction and may have failed to reach statistical significance due to the small overall number of cases with CSA, or because the majority of participants in this study had mild strokes with an NIHSS scores of six or smaller.

It is conceivable that the prevalence of CSA would be higher among participants with severe strokes, as severe strokes are more likely to cause profound neuromuscular weakness and/or involvement of brainstem respiratory centers, which could increase the risk of CSA. Of note, we did not include infarction location (brainstem versus supratentorial) as previous work by our group¹⁸ and others¹¹ did not suggest brainstem involvement to be associated with higher CAIs. Furthermore, there is some evidence suggesting that CAI after stroke is higher in the acute phase (i.e. within 48–72 hours) compared to chronic stroke ¹¹. In our cohort, most HSATs were performed one to three weeks after stroke onset, which could explain the lower CAI.

The strengths of this study include a large sample size, population-based design, recruitment from a diverse patient population, and the use of an objective diagnostic HSAT within a short time window after ischemic stroke. The limited number of recording channels which do not include electroencephalography to differentiate sleep from wakefulness, and lack of attended polysomnography may lead to some level of misestimation of sleep apnea. In accordance with current HSAT scoring guidelines ¹⁰, apneas but not hypopneas were differentiated as central versus obstructive. This may have led to underestimation of the proportion of central respiratory events. Of note though, all HSATs were reviewed by a registered polysomnography technologist to correct automated scoring as needed, which likely improved the sensitivity of the test ⁸, and central apneas detected with our testing device and gold-standard polysomnography show high correlation ⁹.

5. Conclusion

The CSA is low in this large and diverse cohort of patients with ischemic stroke, even in the presence of additional CSA risk factors. These findings suggest that HSATs, which may not always discriminate well between central and obstructive events, may be an acceptable alternative to in-lab polysomnography in patients with recent stroke for whom in-lab polysomnography is not available, not feasible or cost-prohibitive.

Highlights:

• The prevalence of central sleep apnea after ischemic stroke is low

• Central sleep apnea is not associated with stroke severity or heart failure

• Home sleep testing may be appropriate in patients with recent ischemic stroke

ABBREVIATIONS

CAI

Central Apnea Index

CSA

Central Sleep Apnea

HSAT

Home Sleep Apnea Test

NIHSS

National Institutes of Health Stroke Scale