Sleep Apnea-Related Hypoxemia Predicts Post-Stroke Disability Better Than Apnea-Hypopnea Index

Article information

Sleep Med Res. 2025;16(2):109-114
Publication date (electronic) : 2025 June 26
doi : https://doi.org/10.17241/smr.2025.02845
1Department of Cardiology, Hanoi Medical University, Hanoi, Vietnam
2Stroke Center, Bach Mai Hospital, Hanoi, Vietnam
3Department of Internal Medicine, VNU-University of Medicine and Pharmacy, Hanoi, Vietnam
4Respiratory Center, Bach Mai Hospital, Hanoi, Vietnam
5Department of Internal Medicine, Hanoi Medical University, Hanoi, Vietnam
6Vietnam National Heart Institute, Bach Mai Hospital, Hanoi, Vietnam
Corresponding Author Giap Van Vu, MD, PhD Respiratory Center, Bach Mai Hospital, Department of Internal Medicine, Hanoi Medical University, Hanoi 116177, Vietnam Tel +84-989335356 Fax +84-243-8523798 E-mail vuvangiaphmu@gmail.com
Received 2025 April 16; Revised 2025 May 6; Accepted 2025 May 28.

Abstract

Background and Objective

Sleep apnea affects two-thirds of acute ischemic stroke (AIS) patients, but its prognostic utility remains debated. This study evaluates nocturnal hypoxemia metrics and traditional sleep apneic indices as predictors of post-stroke functional outcomes.

Methods

This prospective cohort study enrolled patients with AIS at the Stroke Center of Bach Mai Hospital between April and October 2024. Participants underwent overnight cardiorespiratory polygraphy. Functional outcomes (modified Rankin Scale [mRS]) were evaluated at 3 months. Multivariate logistic regression analysis identified predictors of unfavorable outcomes (mRS >1).

Results

Of the 125 AIS patients enrolled (median National Institutes of Health Stroke Scale [NIHSS] score 3, interquartile range 2–4), 66.4% were diagnosed with sleep apnea (apnea-hypopnea index [AHI] >10). At 3-month follow-up, univariate analysis revealed that younger age, absence of diabetes, lower baseline NIHSS, absence of early neurological deterioration (END), lower AHI, reduced oxygen desaturation index (ODI), higher mean SpO2, and shorter percentage of time with SpO2 <90% were significantly associated with favorable functional outcomes (all p<0.05). In multivariate logistic regression, NIHSS, END, diabetes, ODI, percentage of time with SpO2 <90%, and mean SpO2 were independent predictors of unfavorable outcomes (all p<0.05). Notably, AHI did not show a significant association with outcomes (p>0.05).

Conclusions

NIHSS, END, diabetes, and nocturnal hypoxemia—not AHI—independently predict post-stroke disability. Hypoxemia duration (ODI, % time SpO2 <90%) offers superior prognostic value, advocating for sleep apnea management strategies targeting hypoxia mitigation.

INTRODUCTION

Stroke remains one of the leading causes of death and disability worldwide, with ischemic strokes accounting for 76% of all stroke cases [1]. Sleep apnea syndrome, characterized by recurrent episodes of apnea (complete cessation of breathing) and/or hypopnea (partial reduction in airflow) during sleep, is strongly associated with excessive daytime sleepiness and cardiovascular comorbidities [2].

In 2010, the American Heart Association (AHA) launched the “Life’s Simple 7” initiative to address modifiable risk factors for cardiovascular diseases, including stroke. Over a decade later, mounting evidence on the role of sleep-disordered breathing in stroke pathogenesis prompted the American Stroke Association (ASA) to update its guidelines to Life’s Essential 8 in 2022, formally recognizing sleep apnea as an independent, modifiable risk factor for stroke [3].

Optimizing post-stroke outcomes requires identifying modifiable predictors of recovery. While early neurological deterioration (END) and baseline National Institutes of Health Stroke Scale (NIHSS) scores are established prognostic markers, emerging evidence highlights the role of sleep apnea—a condition affecting 50%–70% of stroke survivors globally—as a critical yet understudied contributor to poor recovery [4]. Nocturnal hypoxemia, a consequence of apnea, may exacerbate cerebral injury through intermittent hypoxia-reoxygenation cycles, yet its impact on functional outcomes in Southeast Asian populations remains poorly characterized.

Prior studies have linked hypoxemia to adverse cardiovascular events in stroke patients [5]. However, its distinct role in functional recovery remains poorly characterized, especially in the acute phase. Most research has evaluated sleep apnea in subacute or chronic phases (e.g., 30–45 days post-stroke) [6] or prioritized apnea-hypopnea index (AHI), neglecting hypoxemia’s direct impact on penumbral viability. While emerging evidence from randomized trials demonstrates that diagnosis and interventions initiated during the acute post-stroke period improve penumbral salvage rates and functional recovery, likely by mitigating hypoxemia-mediated secondary injury in metabolically vulnerable regions [7].

This prospective cohort study, conducted at a tertiary stroke center in Hanoi, Vietnam, aimed to: 1) investigate the association between acute-phase hypoxemia burden (Oxygen Desaturation Index [ODI], % time SpO2 <90%) and 3-month functional outcomes (modified Rankin Scale [mRS] >1) and 2) compare the prognostic utility of hypoxemia metrics with traditional AHI.

We hypothesized that hypoxemia burden, rather than AHI, would independently predict unfavorable functional outcomes, offering actionable insights for optimizing post-stroke care in resource-limited settings.

METHODS

Study Population

This prospective cohort study enrolled consecutive patients with acute ischemic stroke admitted to the Stroke Center of Bach Mai Hospital (Hanoi, Vietnam) between April and October 2024. Inclusion criteria were: age ≥18 years; admission within 7 days of symptom onset; and completion of cardiorespiratory polygraphy (Home Sleep Apnea Test [HSAT] type 3) during hospitalization.

Patients were excluded if they presented with altered consciousness (Glasgow Coma Scale [GCS] <12) or had insufficient sleep study recordings (<4 hours of interpretable data). Additional exclusion criteria included active pneumonia, heart failure (NYHA Class III/IV), or acute sinusitis, as these conditions may muddle respiratory measurements. To avoid confounding by existing management, we also excluded patients with prior obstructive sleep apnea (OSA) diagnosis treated with positive airway pressure therapy and subjects taking sedatives or opioids during the study period due to possible respiratory effects (Fig. 1).

Fig. 1.

Flow diagram of participant enrollment, follow-up, and analysis.

Data Collection

Demographical data, vascular risk factors, and the time of symptoms’ onset were recorded. This information was given by the patients themselves or their relatives. Wake-up stroke was defined when the patient was normal before sleep and woke up with neurological deficits in the morning. All patients were assessed with a neurologic examination that included the GCS and the NIHSS at admission. END was defined as an increase in the NIHSS score by ≥2 points within the first 72 hours of admission, not attributable to hemorrhagic transformation, recurrent stroke, or systemic complications (e.g., infection, metabolic derangements). Body mass index (BMI) was also registered.

Routine laboratory evaluations—including complete blood count, biochemical profile, and renal function tests—were performed upon admission. Brain imaging, either computed tomography or magnetic resonance imaging (MRI), was conducted to confirm ischemic stroke. Stroke subtypes were classified using the Trial of Org 10172 in Acute Stroke Treatment criteria [8].

HSAT: HSAT was performed with a Type 3 portable sleep study device (NOX T3S, NOX MEDICAL) recording airflow via nasal cannula, respiratory effort via thoracoabdominal belts, and oxygen saturation and heart rate via pulse oximetry. All recordings were performed at the bedside within 7 days after stroke onset. Study scores were manually scored by a board-certified sleep specialist, blinded to MRI results, according to the 2023 AASM version 3 criteria.

Sleep apnea assessment and respiratory event definitions

For sleep apnea assessment, only studies with ≥4 hours of recorded airflow, respiratory effort, and oxygen saturation were analyzed. Apnea was defined as a ≥90% decrease in airflow lasting >10 seconds and classified as obstructive when respiratory effort was present (via thoracic/abdominal belts) or central when respiratory effort was absent. Hypopnea was defined as a >30% decrease in airflow for ≥10 seconds associated with a >3% reduction in oxygen saturation. Sleep apnea was classified as obstructive or central based on the predominant type of apnea events. Specifically, OSA was diagnosed when more than 50% of apneas were obstructive, whereas central sleep apnea (CSA) was defined as present when more than 50% of apneas were central in nature, in accordance with the International Classification of Sleep Disorders, 3rd Edition. ODI was defined as the number of times per hour of recording that blood oxygen saturation (SpO2) dropped by ≥3% from baseline. Percentage time SpO2 <90% was calculated as the proportion of total recording time during which the patient’s oxygen saturation (SpO2) fell below 90%, measured via overnight cardiorespiratory polygraphy.

Standardized clinical management during study enrollment

All patients received guideline-directed medical therapy according to the AHA/ASA recommendations, including antiplatelet/anticoagulant therapy, statins, and blood pressure management. Patients with moderate-to-severe OSA (AHI >15) were referred to board-certified sleep medicine pulmonologists for evaluation and treatment initiation (e.g., continuous positive airway pressure [CPAP] therapy, positional therapy). Only one patient received CPAP therapy, and ten patients agreed to a trial of mechanical ventilation. Patients with residual neurological deficits eligible for rehabilitation (e.g., motor impairment, dysphagia, aphasia) were referred to undergo structured physical, occupational, or speech therapy tailored to their functional status.

Primary Outcome Assessment mRS at 3 Months

The degrees of functional recovery were assessed by the mRS at 3 months after stroke onset. It ranges from 0 (no symptoms) to 6 (death). The mRS scores were assessed through in-person follow-up visits or telephone interviews conducted by certified physicians. Assessors were blinded to cardiorespiratory polygraphy results to ensure objectivity.

Statistical Analysis

Continuous variables are expressed as means±standard deviations or medians with interquartile ranges (IQR). Categorical variables are presented as frequencies and percentages. For univariate comparisons, Student’s t-test, Mann-Whitney U-test, chi-square test, or Fisher’s exact test were applied appropriately based on data distribution and variable type.

Backward stepwise logistic regression was performed to identify independent predictors of favorable functional outcomes (defined as a mRS score of 0 or 1) at 3 months post-stroke. Variables with a p-value <0.1 in univariate analysis were included in the multivariate model. Statistical significance was set at a two-tailed p-value <0.05. All analyses were conducted using SPSS statistical software, version 25.0 (IBM Corp.).

Ethic Statements

The study was conducted in accordance with the Declaration of Helsinki, and informed consent was obtained from all subjects and their legal guardian(s). The protocol and questionnaire were reviewed and approved by the Institutional Review Board at the Bach Mai Hospital, Hanoi, Vietnam (IRB-VN01019), decision No. 982/BM-HDDD. To protect privacy, any personal information that could be used to identify a specific individual has been removed. Data were collected anonymously and managed in compliance with data protection legislation.

RESULTS

Baseline Characteristics

A total of 125 patients with a median NIHSS score of 3 (IQR, 2–4) were enrolled. The cohort was predominantly male (68.8%, n=86), with a mean age of 60.6±11.5 years. Sleep apnea was highly prevalent in this population, with a median AHI of 14.7 (IQR, 8.2–27.8). A total of 66.4% of patients (n=83) had sleep apnea, defined as an AHI >10. Predominant OSA was observed in 90.4% of cases, while predominant CSA was rare (n=8, 10%). The median nocturnal SpO2 was 93.8% (IQR, 92.5%–95%).

Patients were stratified by AHI severity (≤10 vs. >10) (Table 1). Compared to those with AHI ≤10, patients in the AHI >10 group had a significantly higher BMI (24.6±3.3 vs. 23.0±2.6 kg/m2, p=0.008), a greater prevalence of diabetes (32.5% vs. 11.9%, p=0.016), and a higher incidence of END (20.7% vs. 4.9%, p=0.032). Baseline NIHSS scores and age did not differ significantly between the two groups (p>0.05).

Characteristics of apneic (AHI >10) and nonapneic patients

Functional Outcomes at 3 Months

At follow-up, 70.4% of patients (n=88) achieved a favorable functional outcome, defined as a mRS score ≤1. Univariate analysis identified several predictors of poor outcome (mRS >1), as presented in Table 2. Patients with poor outcomes were older (64.3±11.1 vs. 59.0±11.4 years, p=0.018), more likely to have diabetes (40.5% vs. 19.3%, p=0.023), and had higher baseline NIHSS scores (median 4 [IQR, 3–6.5] vs. 3 [2–4], p<0.001). END occurred more frequently in this group (32.4% vs. 8.1%, p=0.002), and they were also more likely to have received reperfusion therapy (16.2% vs. 3.4%, p=0.019). Regarding sleep parameters, patients with poor outcomes had significantly higher AHI (37 [13.4–43.0] vs. 12.8 [7.5–22.5] events/hr, p=0.002), higher ODI (22 [10.8–37.9] vs. 10.6 [5.7–20.8] events/hr, p=0.004), and spent a greater percentage of time with SpO2 below 90% (3.3% [0.3%–21.5%] vs. 0.9% [0%–3.8%], p=0.008).

Univariate analysis identified several factors significantly associated with mRS >1

In multivariate analysis, there are six variables independently predicted poor functional outcomes (mRS >1) (Table 3).

Multivariate predictors of unfavorable functional outcomes (mRS >1) in separate models incorporating percentage time SpO₂ <90% and mean SpO₂

Receiver Operating Characteristic Curve Analysis of Hypoxemia Parameters

Receiver operating characteristic (ROC) curves were generated to evaluate the discriminatory power of nocturnal oxygenation metrics for predicting unfavorable functional outcomes (mRS >1), as shown in Fig. 2. The ODI demonstrated moderate predictive accuracy with an area under the curve (AUC) of 0.665 (95% confidence interval [CI]: 0.559–0.770, p=0.004). Similarly, the percentage of time spent with SpO2 below 90% showed moderate discrimination, with an AUC of 0.648 (95% CI: 0.534–0.763, p=0.009). In contrast, mean nocturnal SpO2 had poor discriminatory capacity, with an AUC of 0.356 (95% CI: 0.243–0.469, p=0.011).

Fig. 2.

Receiver operating characteristic curve analysis of nocturnal hypoxemia parameters for predicting unfavorable functional outcomes (modified Rankin Scale >1). ODI, oxygen desaturation index.

DISCUSSION

To the best of our knowledge, our study is the first to demonstrate the association between hypoxemia during the acute phase of ischemic stroke (assessed within the first 7 days post-onset) and unfavorable functional outcomes (mRS >1). While prior studies focused on subacute-phase assessments or relied on the AHI [9-11], our findings highlight the critical role of hypoxemia burden (ODI and % time SpO2 <90%) in the acute phase—a time when the ischemic penumbra is most vulnerable and salvageable.

Prior research has linked hypoxemia in acute ischemic stroke to adverse cardiovascular events. Our findings extend this evidence by highlighting its distinct role in functional recovery [5,12], challenging the conventional reliance on the AHI as the primary metric for evaluating sleep apnea in acute stroke care. Although sleep apnea (AHI >10) was prevalent (66.4%) in our study, AHI lacked independent prognostic value. Instead, hypoxemia burden—measured by ODI and time with SpO2 <90%—better predicts functional outcomes by quantifying cumulative ischemic damage from prolonged desaturation. These metrics reflect neuronal vulnerability more accurately than AHI’s event counts, underscoring their clinical relevance.

The limited prognostic value of AHI likely stems from its focus on respiratory event frequency rather than hypoxemia severity. For example, patients with low AHI but prolonged hypoxemia (e.g., % time SpO2 <90%) may face higher risks than those with high AHI but brief desaturations. These observations advocate for a paradigm shift in sleep apnea management, prioritizing hypoxemia mitigation (e.g., optimizing CPAP to reduce nocturnal desaturations) over AHI reduction.

Intermittent hypoxia during apneic episodes exacerbates ischemic injury in the vulnerable penumbral region, ultimately hindering functional recovery [13,14]. When cerebral hypoxia occurs, the brain activates compensatory mechanisms, such as vasodilation and increased cerebral blood flow, to preserve oxygen delivery [15]. However, in patients with OSA, these adaptive responses are compromised by impaired cerebral autoregulation—a phenomenon well-documented in prior studies [16-18]. Nocturnal hypoxemia exacerbates cerebral hemodynamic instability through recurrent hypoxia-reoxygenation cycles, which intensify oxidative stress, impair endothelial function, and drive neuroinflammatory cascades [19,20]. During apneic events, hypoxia synergizes with compromised cerebral autoregulation to precipitate significant reductions in cerebral perfusion. This hypoperfusion disproportionately impacts the ischemic penumbra—a region of critically perfused, salvageable tissue bordering the infarct core—where marginal oxygen supply heightens vulnerability to secondary injury. Over time, the cumulative burden of hypoxemia and hemodynamic dysregulation may overwhelm compensatory mechanisms, converting the penumbra into irreversibly infarcted tissue.

A previous study reported associations between ODI, AHI, and severe desaturations (SpO2 ≤85%) with unfavorable functional outcomes but found no link to milder desaturations (SpO2 ≤90%) [6]. This discrepancy may stem from differences in timing of sleep apnea assessments: the referenced study evaluated sleep apnea within 30–45 days post-stroke, whereas our study assessed sleep apnea during the acute phase (within 7 days of stroke onset). Early hypoxemia may disproportionately harm the ischemic penumbra, while later assessments capture chronic sleep apnea effects.

END has long been recognized as a robust predictor of poor functional outcomes in ischemic stroke [21]. In our study, END demonstrated a strong and independent association with unfavorable outcomes (mRS >1), consistent with prior literature. Patients experiencing END had over a 10-fold increased odds of disability or dependency at 3 months (odds ratio=10.5, p<0.001).

Our study has several limitations. First, the modest sample size (n=125) and relatively few unfavorable outcomes (n=37) restricted subgroup analyses and may have reduced statistical power to detect subtler associations. Second, the use of cardiorespiratory polygraphy instead of polysomnography precluded sleep-stage differentiation, potentially underestimating hypoxemia severity and AHI due to reliance on total recording time rather than true sleep time. Because arousal-based hypopneas could not be identified using HSAT, the AHI in our study may be underestimated. This limitation could partially explain why AHI was not a significant predictor of outcome, while oxygen desaturation metrics were. Third, unmeasured confounders—such as CPAP adherence, rehabilitation intensity, socioeconomic status, or medication compliance—could influence outcomes. Finally, the 3-month follow-up may not capture long-term recovery or cardiovascular complications.

In conclusion, while sleep apnea is prevalent among ischemic stroke patients, our findings suggest that oxygenation-related parameters rather than AHI alone are more strongly associated with unfavorable outcomes. Addressing nocturnal hypoxia may be a key therapeutic target for improving stroke recovery.

Notes

Availability of Data and Material

The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.

Author Contributions

Conceptualization: Tho Dinh Phung, Giap Van Vu, Hoai Thi Thu Nguyen. Data curation: Tho Dinh Phung, Dung Viet Nguyen. Formal analysis: Tho Dinh Phung, Dung Viet Nguyen. Funding acquisition: Giap Van Vu, Hoai Thi Thu Nguyen. Investigation: Tho Dinh Phung, Giap Van Vu, Hoai Thi Thu Nguyen. Methodology: all authors. Project administration: Giap Van Vu. Resources: Tho Dinh Phung, Hoai Thi Thu Nguyen. Software: Tho Dinh Phung, Dung Viet Nguyen. Supervision: Giap Van Vu. Validation: Hoai Thi Thu Nguyen. Visualization: Tho Dinh Phung. Writing—original draft: Tho Dinh Phung, Giap Van Vu, Hoai Thi Thu Nguyen. Writing—review & editing: Tho Dinh Phung, Giap Van Vu, Hoai Thi Thu Nguyen.

Conflicts of Interest

The authors have no potential conflicts of interest to disclose.

Funding Statement

None

Acknowledgements

None

References

1. Ton MD, Dao PV, Nguyen DT, Nguyen TH, Tran CC, Duong HQ, et al. Sex disparity in stroke outcomes in a multicenter prospective stroke registry in Vietnam. Int J Stroke 2023;18:1102–11.
2. Cumpston E, Chen P. Sleep Apnea Syndrome. In: StatPearls [Internet]. Treasure Island: StatPearls Publishing; 2023. [accessed 2024 January 10]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK564431/.
3. Lloyd-Jones DM, Allen NB, Anderson CAM, Black T, Brewer LC, Foraker RE, et al. Life’s essential 8: updating and enhancing the American Heart Association’s construct of cardiovascular health: a presidential advisory from the American Heart Association. Circulation 2022;146:e18–43.
4. Hasan F, Gordon C, Wu D, Huang HC, Yuliana LT, Susatia B, et al. Dynamic prevalence of sleep disorders following stroke or transient ischemic attack: systematic review and meta-analysis. Stroke 2021;52:655–63.
5. Korostovtseva L, Bochkarev M, Amelina V, Nikishkina U, Osipenko S, Vasilieva A, et al. Sleep-disordered breathing and prognosis after ischemic stroke: it is not apnea-hypopnea index that matters. Diagnostics (Basel) 2023;13:2246.
6. Brown DL, Shafie-Khorassani F, Kim S, Chervin RD, Case E, Yadollahi A, et al. Do apneas and hypopneas best reflect risk for poor outcomes after stroke? Sleep Med 2019;63:14–7.
7. Duss SB, Brill AK, Baillieul S, Horvath T, Zubler F, Flügel D, et al. Effect of early sleep apnoea treatment with adaptive servo-ventilation in acute stroke patients on cerebral lesion evolution and neurological outcomes: study protocol for a multicentre, randomized controlled, rater-blinded, clinical trial (eSATIS: early Sleep Apnoea Treatment in Stroke). Trials 2021;22:83.
8. Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, et al. 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 1993;24:35–41.
9. Lisabeth LD, Sánchez BN, Lim D, Chervin RD, Case E, Morgenstern LB, et al. Sleep-disordered breathing and poststroke outcomes. Ann Neurol 2019;86:241–50.
10. Camilo MR, Schnitman SV, Sander HH, Eckeli AL, Fernandes RM, Leite JP, et al. Sleep-disordered breathing among acute ischemic stroke patients in Brazil. Sleep Med 2016;19:8–12.
11. Menon D, Sukumaran S, Varma R, Radhakrishnan A. Impact of obstructive sleep apnea on neurological recovery after ischemic stroke: a prospective study. Acta Neurol Scand 2017;136:419–26.
12. Huhtakangas JK, Saaresranta T, Vähänikkilä H, Huhtakangas J. Nocturnal hypoxemia and central apneas increase mortality, but not recurrent ischemic events after ischemic stroke. Sleep Med 2022;97:1–9.
13. Kim TJ, Ko SB, Jeong HG, Kim CK, Kim Y, Nam K, et al. Nocturnal desaturation is associated with neurological deterioration following ischemic stroke: a retrospective observational study. J Clin Sleep Med 2017;13:1273–79.
14. Tsivgoulis G, Alexandrov AV, Katsanos AH, Barlinn K, Mikulik R, Lambadiari V, et al. Noninvasive ventilatory correction in patients with acute ischemic stroke: a systematic review and meta-analysis. Stroke 2017;48:2285–8.
15. Hoiland RL, Bain AR, Rieger MG, Bailey DM, Ainslie PN. Hypoxemia, oxygen content, and the regulation of cerebral blood flow. Am J Physiol Regul Integr Comp Physiol 2016;310:R398–413.
16. Durgan DJ, Bryan RM Jr. Cerebrovascular consequences of obstructive sleep apnea. J Am Heart Assoc 2012;1:e000091.
17. Phillips SA, Olson EB, Morgan BJ, Lombard JH. Chronic intermittent hypoxia impairs endothelium-dependent dilation in rat cerebral and skeletal muscle resistance arteries. Am J Physiol Heart Circ Physiol 2004;286:H388–93.
18. Reichmuth KJ, Dopp JM, Barczi SR, Skatrud JB, Wojdyla P, Hayes D Jr, et al. Impaired vascular regulation in patients with obstructive sleep apnea: effects of continuous positive airway pressure treatment. Am J Respir Crit Care Med 2009;180:1143–50.
19. Al Ahmad A, Gassmann M, Ogunshola OO. Involvement of oxidative stress in hypoxia-induced blood-brain barrier breakdown. Microvasc Res 2012;84:222–5.
20. Liu X, Ma Y, Ouyang R, Zeng Z, Zhan Z, Lu H, et al. The relationship between inflammation and neurocognitive dysfunction in obstructive sleep apnea syndrome. J Neuroinflammation 2020;17:229.
21. Thanvi B, Treadwell S, Robinson T. Early neurological deterioration in acute ischaemic stroke: predictors, mechanisms and management. Postgrad Med J 2008;84:412–7.

Article information Continued

Fig. 1.

Flow diagram of participant enrollment, follow-up, and analysis.

Fig. 2.

Receiver operating characteristic curve analysis of nocturnal hypoxemia parameters for predicting unfavorable functional outcomes (modified Rankin Scale >1). ODI, oxygen desaturation index.

Table 1.

Characteristics of apneic (AHI >10) and nonapneic patients

Variables All patients (n=125) AHI ≤10 (n=42) AHI >10 (n=83) p-value
Age (yr) 60.6±11.5 58.3±11.7 61.7±11.4 0.113
Male 86 (68.8) 31 (73.8) 55 (66.3) 0.815
BMI (kg/m2) 24.0±3.2 23.0±2.6 24.6±3.3 0.008
Hypertension 100 (80) 33 (78.6) 67 (80.4) 0.815
Diabetes 32 (25.6) 5 (11.9) 27 (32.5) 0.016
AF 6 (4.8) 1 (2.4) 5 (6.0) 0.633
NIHSS 3 (2–4) 3 (1.75–5) 3 (2–4) 0.746
END 19 (15.2) 2 (4.9) 17 (20.7) 0.032

Values are presented as mean±standard deviation, number (%), or median (interquartile range).

AHI, apnea-hypopnea index; BMI, body mass index; AF, atrial fibrillation; NIHSS, National Institutes of Health Stroke Scale; END, early neurological deterioration.

Table 2.

Univariate analysis identified several factors significantly associated with mRS >1

Variables mRS ≤1 (n=88) mRS >1 (n=37) p-value
Age (yr) 59.0±11.4 64.3±11.1 0.018
Diabetes 17 (19.3) 15 (40.5) 0.023
BMI (kg/m2) 24.1±3.0 23.8±3.5 0.682
NIHSS 3 (2–4) 4 (3–6.5) <0.001
END 7 (8.1) 12 (32.4) 0.002
Reperfusion treatment 3 (3.4) 6 (16.2) 0.019
Cardioembolism 5 (5.7) 4 (10.8) 0.449
Large artery artherosclerosis 19 (21.8) 9 (24.3) 0.816
AHI (events/hr) 12.8 (7.5–22.5) 37 (13.4–43.0) 0.002
ODI (events/hr) 10.6 (5.7–20.8) 22 (10.8–37.9) 0.004
Mean SpO2 (%) 94 (93–95.1) 93 (91.3–94.5) 0.011
% time SpO2 <90% 0.9 (0–3.8) 3.3 (0.3–21.5) 0.008

Values are presented as mean±standard deviation, number (%), or median (interquartile range).

mRS, modified Rankin Scale; BMI, body mass index; NIHSS, National Institutes of Health Stroke Scale; END, early neurological deterioration; AHI, apnea-hypopnea index; ODI, oxygen desaturation index.

Table 3.

Multivariate predictors of unfavorable functional outcomes (mRS >1) in separate models incorporating percentage time SpO₂ <90% and mean SpO₂

Variables Model 1 (OR, 95% CI) Model 2 (OR, 95% CI)
Diabetes 3.3 (1.11–10.10) 3.8 (1.22–11.92)
NIHSS 1.6 (1.26–2.07) 1.5 (1.25–2.00)
END 11.5 (3.16–42.15) 11.8 (3.1–44.99)
ODI 1.1 (1.01–1.18) 1.1 (1.01–1.18)
Mean SpO2 0.8 (0.64–0.94) -
% time SpO2 <90% - 1.1 (1.01–1.06)

Model 1: independent variables: ages, diabetes, NIHSS, END, reperfusion treatment, AHI, ODI, mean SpO2. Model 2: independent variables: ages, diabetes, NIHSS, END, reperfusion treatment, AHI, ODI, % time SpO2 <90%.

NIHSS, National Institutes of Health Stroke Scale; END, early neurological deterioration; ODI, oxygen desaturation index; mRS, modified Rankin Scale; OR, odds ratio; CI, confidence interval; AHI, apnea-hypopnea index; -, not applicable.