Obstructive Sleep Apnea in the ICU: Elevated Prevalence, Diagnostic Challenges, and Treatment Limitations

1. Introduction

Obstructive sleep apnea (OSA) is a highly prevalent but often underdiagnosed disorder characterized by recurrent episodes of upper airway collapse during sleep, resulting in intermittent hypoxia, hypercapnia, and sleep fragmentation [1]. The pathophysiology of OSA involves a complex mix of anatomical, neuromuscular, and ventilator control abnormalities, all of which synergistically contribute to repetitive airway obstruction [2,3]. During these obstructive events, pharyngeal airway collapse leads to airway cessation despite ongoing respiratory effort, resulting in large negative intrathoracic pressure swings, intermittent hypoxemia, and hypercapnia. The sum of these physiological disturbances triggers cortical arousals, leading to sleep fragmentation. These apneic events lead to activation of the sympathetic nervous system (SNS), oxidative stress, systemic inflammation, and endothelial dysfunction, all of which contribute to a higher risk of cardiovascular, metabolic, and neuropsychiatric comorbidities [4,5].

The global burden of OSA is substantial, with an estimated 936 million adults from ages 30 to 69 years old worldwide diagnosed with mild to severe OSA, and in which 425 million have moderate to severe disease, according to several modeling studies [6,7,8]. In the United States alone, data from the Wisconsin Sleep Cohort Study estimates that 13% of men and 6% of women have moderate to severe (AHI ≥ 15 events/ hour), with prevalence of OSA positively increasing with age and body mass index [9]. However, it is estimated that 80-90% of individuals with moderate to severe OSA remain underdiagnosed, revealing a substantial gap in public health. The underdiagnosis and undertreatment of OSA not only worsens individual health outcomes but also amplifies societal costs and public safety risks. Addressing this burden requires greater public awareness, standardized screenings, and integrated care strategies that particularly cater towards high-risk and underserved populations [7,10].

Concerningly, OSA is associated with a wide range of deleterious health outcomes. Cardiovascular complications have been well documented in association with OSA, including systemic hypertension (HTN), arrhythmia, coronary artery disease, heart failure, and stroke [11,12,13]. The Sleep Heart Health Study and other large cohort studies have shown that untreated OSA increases the risk of cardiovascular events and mortality [14]. Metabolically, OSA is also linked to insulin resistance, dyslipidemia, and poor glycemic control, serving as an independent risk factor for type 2 diabetes mellitus (T2DM) [12]. Neurologically, OSA has been associated with impaired attention, memory deficits, executive dysfunction, and mood disorders, all of which can negatively impact quality of life. OSA is further linked to increased perioperative complications, postoperative respiratory events, and higher mortality, especially in patients with undiagnosed OSA [15,16]. During the COVID-19 pandemic, OSA was seen to worsen hospitalized patient outcomes in regard to increased incidence of respiratory failure and heart failure [17,18]. Despite these risks, treatment with continuous positive airway pressure (CPAP) can improve patient quality of life and may mitigate some consequences of cardiovascular and neurocognitive adverse events, though evidence for cardiovascular event reduction is mixed, possibly due to the heterogenous nature of OSA and variable patient compliance to treatment adherence [11,19]. Recent research has emphasized the importance of identifying pathophysiological traits to guide personalized treatment strategies and targeted therapies to better or further improve patient outcomes and alleviate the public health burden of OSA [20,21,22]. This review examines the high prevalence of OSA in the ICU, focusing on challenges in accurate diagnosis, increased complications related to comorbidities, and limitations of current treatment strategies. The study selection process is summarized in a PRISMA flow diagram in Figure 1.

Epidemiology and Prevalence of OSA in ICU Settings

Intensive Care Unit (ICU) usage is a complicated and financially expensive healthcare infrastructure. With approximately 5.7 million annual patient admissions in the US, ICU associated costs exceed $82 billion, greater than 4.1% of national healthcare expenses [23]. ICU hospitalizations are linked to a wide variety of complications, including but not limited to delirium, 1 in 6 patients developing sepsis-associated acute kidney injury, and the development of Post Intensive Care syndrome as characterized by a range of physical, mental, and cognitive impairments persisting after discharge [24,25,26]. In particular, sleep disturbances and disorders frequently occur in the ICU, with a prevalence of 66% [27]. Up to 80% of ICU patients reported experiencing significant sleep deprivation during their stay [28]. These disturbances are not only common but also clinically remarkable as impaired sleep quality in critically ill patients results in poor outcomes such as prolonged weaning, delayed extubation, and complete disappearance of REM sleep [29]. In general, untreated sleep apnea is associated with an elevated risk of morbidity and mortality [30].

OSA has been highly associated with hospitalizations. In patients hospitalized for CVD, OSA prevalence is estimated to be as high as 48% [31]. However, a low percentage –- 4.8-5.8% -- of hospitalized patients are formally diagnosed with OSA and provided CPAP therapy during the hospital stay [32]. Within the ICU, sleep disorders are known to be common yet frequently undiagnosed [33]. Approximately 68% out of 129 patients were discovered to have an apnea-hypopnea index (AHI) value ≥ 5, and 40% of those patients had an AHI >15 [34]. AHI, or apnea-hypoxia index, quantifies the episodes of apnea and hypopneas throughout sleep. An AHI <5 is normal, 5–14 indicates mild OSA, and ≥15 indicates moderate OSA [35]. Hence, an overwhelming majority of ICU patients exhibit varying degrees of OSA, indicating that the frequency and under-diagnosis of OSA in ICU is a critical issue. The proposed mechanisms linking obstructive sleep apnea to adverse outcomes in critically ill patients are illustrated in Figure 2.

ICU patients often have underlying medical conditions that exacerbate their likelihood of developing OSA, such as acute respiratory failure, sepsis, and COPD [36,37,38]. Naranjo et al., discovered that in patients hospitalized for COPD exacerbations with no prior formal diagnosis of OSA, screening revealed approximately 46.6% of patients had OSA, with greater OSA severity correlated with increased odds of hospital readmission and mortality [39]. Sepsis is known to activate a cascade of pro-inflammatory cytokines such as IL-6 and TNF-α [40]. Pro-inflammatory cytokines have been linked to the progression of lung disease via increased vascular permeability and decreased lung compliance, referring to the lung’s capacity to expand in response to pressure [41]. Sepsis-induced inflammation impairs gas exchange and inflicts respiratory distress, significantly impairing sleep and breathing in patients with underlying, undiagnosed OSA [42]. Furthermore, the most cited risk factors for OSA include male sex, obesity, and older age [43]. ICU admissions have been likewise found to be more likely for ages 50+, male sex, and obesity [44]. The equivalent risk factors for both OSA and ICU strongly underscore the likelihood that a significant portion of critically ill patients have underlying OSA, potentially exacerbating clinical outcomes and recovery times.

There is a lack of clinical data regarding the precise prevalence of OSA in the ICU setting due to diagnostic barriers. Historically, OSA data has been influenced by evolving definitions of hypopnea, AHI criteria, and changes in PSG (polysomnography methods) [45]. The lack of universal criteria for defining OSA presents a challenge for diagnoses, particularly in the ICU as PSG is highly time consuming, labor intensive, and expensive [46]. Furthermore, even diagnosing OSA was projected to cost nearly $2.4 billion, with CPAP therapy – the gold standard for treating OSA – costing around $3.4 billion in the US [47]. In the ICU where 80-90% of patients present with multiple morbidities, it’s plausible to assume that screening, diagnosis, and treatment for OSA will often not be of the utmost priority within the ICU [48]. It’s imperative to also understand that this dilemma extends to a global scale applied to other clinical settings as well.

2. Results

Clinical Implications of Unrecognized OSA in the ICU

The underdiagnosis of OSA has been linked with an increased duration of mechanical ventilation in critically ill patients. Key clinical outcomes and complications associated with obstructive sleep apnea in critically ill patients are summarized in Table 1. The pathophysiological hallmarks of OSA relating to intermittent hypoxia and increased negative intrathoracic pressure exacerbate respiratory efforts and increased difficulty of weaning patients off ventilation [49,50,51]. ICU patients with underdiagnosed OSA require significantly longer durations of invasive ventilation compared to non-OSA patients. For example, in pediatric patients, underdiagnosed OSA was associated with more than a 5-fold increase in the need for mechanical ventilation, as well as an extra day in the hospital [49]. Hospitalized adult patients with pneumonia or COPD with suspected OSA were also linked to higher rates of invasive and noninvasive ventilation, increased risk of clinical deterioration, and longer lengths of stay [50,51]. Reduced ventilatory drive, compromised upper airway tone, and blunted arousal tone in OSA patients all contribute to extubating delays and extend ventilator dependence.

Extubation failure is also a significant contributor to ICU patient mortality and morbidity, which is more common in patients with underdiagnosed OSA. The loss of pharyngeal muscle tone during sleep in OSA predisposed patients to upper airway collapse post-extubation, especially during the immediate period of post-sedation [52,53]. Rates as high as 30% have been reported in studies reporting re-intubation rates in ICU patients with high-risk STOP-BANG scores compared to 10-15% of the general population [54]. This phenomenon may further be compounded by unmonitored desaturations, unrecognized apneic events, and ineffective airway clearance during spontaneous breathing trials [55].

The perioperative period also presents numerous challenges to patients with underdiagnosed OSA. Most patients with moderate-to-severe OSA remain undiagnosed prior to surgery, lacking preventative risk management [56,57]. Notably, these patients have an increased risk for hypoventilation, desaturation, airway obstruction, and opioid sensitivity; all increased in risk following the administration of general anesthesia [58,59,60,61]. Post-operatively, patients with undiagnosed or high-risk OSA are nearly at double risk of complications compared to low-risk patients. The susceptibility of upper airway collapse and anesthesia-related respiratory depression may precipitate critical respiratory events during the recovery period [58,59]. The use of anesthetic and sedative agents can exacerbate upper airway collapse and depress central respiratory drive, which makes intraoperative and postoperative management more complex.

OSA induces substantial autonomic dysregulation with sympathetic overactivity and cyclical surges in blood pressure and heart rate during apneic episodes [11,60,61,62]. In the ICU, patients typically have limited physiological reserves, and the presence of OSA can further destabilize cardiovascular status. Episodes of hypoxia and intrathoracic pressure swings lead to an increase in myocardial oxygen demand while reducing coronary artery perfusion, leading to a heightened risk of life-threatening arrhythmias and myocardial injury. OSA has been shown to be independently associated with new-onset atrial fibrillation, ventricular ectopy, and sudden cardiac death in high-acuity patients. These life-threatening symptoms may even go unnoticed due to sedation masking typical OSA presentation and events [60,61]. Of note, women with OSA and acute coronary syndrome may face even greater long-term cardiovascular risks compared to men [63].

The implications of underdiagnosed OSA extend beyond the initial ICU stay. Recurrent hospitalizations increased 30-day ICU readmission rates, and higher long-term mortality have all been linked to untreated OSA [12,60,64,65]. Patients with unrecognized OSA have been shown to have a 1.8-fold increase of ICU readmission within 30 days and a 2.2-fold increase in one-year mortality compared to matched controls. Chronic intermittent hypoxia also contributes to higher systemic inflammation, endothelial dysfunction, and metabolic dysregulation, compounding risks for adverse patient outcomes, even independent from obesity as a risk factor. Neurocognitive impairments from undiagnosed OSA can also lead to a reduced quality of life, increased work-related injuries, and greater healthcare utilization [66]. Despite its prevalence, OSA remains underdiagnosed in high-risk populations, especially in the ICU, due to a lack of systemic screening [60,67]. Given these findings, there is a compelling need for early therapeutic intervention for ICU patients at risk for OSA.

3. Discussion

3.1. Diagnostic Challenges in the ICU

Atypical presentations in the critically ill, particularly in the ICU, can mask the symptoms of OSA, contributing to underdiagnosis and exacerbated outcomes. The typical clinical presentation of OSA include disruptive snoring, witnessed apneas, and excessive daytime sleepiness [68]. On average, 39.5% of ICU patients are receiving mechanical ventilation at any given hour [69]. Therefore, most ICU patients are not visually exhibiting classical symptoms that would arouse suspicion for clinical diagnosis of OSA. Furthermore, critically ill patients often present with atypical sleep, with a lack of sleep spindles and K complexes [70]. This poses a significant challenge in determining if the abnormal sleep pattern is a consequence of the underlying reason for ICU admission or a persistent condition attributable to OSA. Furthermore, undiagnosed and untreated OSA can significantly increase the likelihood of developing acute and post-operative delirium [71]. Delirium is particularly prevalent in the ICU, affecting 83% of ICU patients on mechanical ventilation with a 3.2-fold increase in 6-month mortality [72,73].

Diagnosis of true OSA in the ICU is clinically challenging in terms of differentiation from sedation effects, underlying respiratory failure (i.e. ARDS), and ventilator-induced breathing patterns. The usage of opioids and sedative medications within the ICU is significantly elevated, with a 56.1% reported prevalence [74]. Anesthetics and opioids are well-associated causes of respiratory depression, and opioids specifically impact the peripheral and central carbon dioxide chemoreflex loops to diminish respiratory capacity [75]. Specifically, opioids, in conjunction with benzodiazepines, are associated with increased episodes of apnea and hypoxemia in patients [76]. Opioids lead to shallow breathing, a slowed respiratory rate < 8 bpm, and a decreased SpO2, a close resemblance to the hypoventilation witnessed in OSA [77]. Neuromuscular blocking agents, or NMBAs, are frequently utilized for critically ill patients for mechanical ventilation and muscle relaxation. One common complication is residual neuromuscular block, in which upper airway muscles and pharyngeal muscle function are diminished [78]. With patients exhibiting hypoxic ventilatory response, the usage of anesthetics, neuromuscular blockers, and opioids can mask the symptoms of OSA, leading to misdiagnosis and/or lack of diagnosis. A proposed clinical workflow for screening and diagnosing obstructive sleep apnea in hospitalized patients is shown in Figure 3.

3.2. Current Approaches to Screening, Diagnosis, and Management

Currently, there are several validated questionnaires employed in outpatient and preoperative settings to estimate the probability of OSA. A summary of commonly used obstructive sleep apnea screening tools and their reported performance characteristics is provided in Table 2. The most widely used is the STOP-BANG Questionnaire that incorporates eight parameters: snoring, tiredness, observed apneas, high blood pressure, BMI > 35, age > 50, neck circumference > 40 cm, and male gender [79,80,81]. This method is favored for its ease of use, high sensitivity of up to 93% for moderate-to-severe OSA (though specificity is modest below 50%), and rapid administration. Alternatively, the Berlin Questionnaire stratifies patients based on their snoring behavior, daytime somnolence, and comorbid HTN/BMI, with a specificity around 70-80% [81]. The American Society of Anesthesiologists (ASA) has a similar OSA checklist intended for preoperative use to stratify patient risk prior to anesthesia, however this method lacks robust validation in ICU settings [82]. The application of these tools in critically ill patients presents significant limitations. Often, ICU patients are sedated, mechanically ventilated, or nonverbal, which makes subjective questionnaires appear to be inapplicable. Further, physiologic confounders such as fluid overload, altered state of consciousness, or acute respiratory distress can mask or mimic features of OSA [83,84,85]. Metrics such as neck circumference or BMI may be distorted by the presence of edema or critical illness-related catabolism, resulting in limited discriminatory power in the ICU. Despite these drawbacks, modified versions of the STOP-BANG and Berlin Questionnaire criteria have been implemented informally in some ICU studies, though no consensus has been reached regarding best practices [83,84].

Currently, PSG is the main standard for the formal diagnosis of sleep and sleep disordered breathing [85]. In the ICU, usage of PSG is simply not practical mainly due to the quantity of equipment needed and the deviating mental statuses of the patients [33,34]. In fact, the PADIS guidelines (Prevention and Management of Pain, Agitation/Sedation, Delirium, Immobility, and Sleep Disruption) established by the Society of Critical Care Medicine has deemed PSG utilization in ICU as unfeasible since 2018 [86]. Given the impracticality of utilizing full polysomnography (PSG) in ICU settings, simplified diagnostic alternatives have been explored, such as overnight pulse oximetry, which assesses nocturnal oxygen desaturation index (ODI) [87,88]. ODI presents a practical, portable, and cost-effective tool in resource-limited or high-acuity settings. However, oximetry may be confounded by supplemental oxygen, unstable hemodynamics and frequent desaturations unrelated to OSA. Alternatively, respiratory polygraphy can capture airflow, respiratory effort, and oxygen saturation without electroencephalography (EEG) [89]. This method retains acceptable diagnostic accuracy in non-intubated and stable patients, though data in ICU settings remain sparse. In one prospective study of 124 ICU patients, polygraphy detected moderate-to-severe OSA in 70% of subjects, suggesting significant underdiagnosis [33,34]. However, motion artifacts, supine positioning, and irregular respiratory patterns in critically ill patients may reduce diagnostic reliability. To overcome these diagnostic challenges, recent advances in machine learning (ML) algorithms and wearable biosensors have emerged trained on electronic heath record (EHR) data, including demographic variables, comorbidities, and physiologic trends for predicting underdiagnosed OSA [86,87,88,89]. ML models using easily accessible parameters such as age, BMI, heart rate, and select biomarkers have been successfully integrated into cloud-based markers to support clinical decision making [90]. Accelerometer-based actigraphy, thoracic effort bands, and peripheral arterial tone sensors are also increasingly being utilized in hospital settings [91,92]. FDA-approved devices are better equipped at sensing real-time respiratory event detection and autonomic metrics. However, recognizing the unique challenges of the ICU has led to the development of new frameworks and metrics with simplified clinical observations, such as respiratory effort during sleep or sedation weaning trials [89]. Alternatively, ventilator waveforms of apnea indices from overnight mechanical ventilation data can be used to infer OSA risk. While promising for the unique needs and challenges of ICU patients, these tools are in the early validation phases and have not yet been incorporated into standard practices [86].

With respect to treatment, effective management is highly dependent upon early identification and effective airway management that uniquely tailors to each patient’s needs. Currently, continuous positive airway pressure (CPAP) is utilized as the gold standard for moderate to severe OSA treatment. By applying a consistent positive pressure throughout the respiratory cycle, CPAP helps keep the airways open and decrease the AHI value back within the normal range [87]. CPAP also improves ventilation-perfusion matching (V/Q) and decreases the risk of atelectasis, effectively diminishing the likelihood of developing hypoxia and associated adverse outcomes [88]. However, most ICU patients present with contraindications to CPAP, including altered level of consciousness, inability to protect the airway, respiratory arrest, and unstable cardiorespiratory status [89].

For ICU patients with contraindications, there exists various non-CPAP treatment options, such as positional therapy, hypoglossal nerve stimulation, myofunctional therapy, and maxillofacial surgery [90]. An alternative avenue of treatment is low-dose fentanyl and/or dexmedetomidine to improve compliance until criteria for CPAP therapy is met [88]. Dexmedetomidine is an a2-agonist with analgesic, anxiolytic, and sympatholytic sedative properties that maximizes patient comfort while minimizing respiratory depression [91]. It has been shown to improve efficacy and comfort alongside BiPAP therapy, thereby optimizing outcomes [91]. BIPAP can be more effective than CPAP in ICU patients, providing higher pressure during inspiration and lower pressure during expiration [92] while lowering PaCO2 levels [92,93]. Given the association between hypercapnia, ICU readmissions, and elevated mortality rates [94], BiPAP can help optimize patient outcomes. Furthermore, BiPAP machines provide a backup rate for patients with central hypoventilation and unpredictable IPAP (inspiratory positive airway pressure) and EPAP (expiratory positive airway pressure) levels due to muscle weakness [95]. Critically ill patients in the ICU often present with multiple comorbidities as well as reduced respiratory drive because of sedation and opioid medication usage [96]. Hence, BiPAP can be of great benefit for ICU patients who are greatly vulnerable to rapid declines in respiratory function. Ultimately, a combination of proactive airway management, vigilant observation of respiratory status, and monitoring of medication usage must be undertaken to minimize the risk of worsening OSA in ICU patients.

3.3. Future Research Directions

Despite advancements and growing knowledge regarding the high prevalence of OSA in critically ill populations, there remain substantial knowledge gaps and areas for clinical improvement. The ICU setting poses several unique diagnostic challenges for patients due to their status of acuity, respiratory pathophysiology, and impracticality of conventional screening and diagnostic protocols [97,98,99]. For instance, current screening instruments, such as the STOP-BANG and BERLIN Questionnaires, were developed and validated in ambulatory and perioperative patient populations that rely heavily on patient reported symptoms, rendering them impractical for sedated, intubated, and nonverbal ICU patients [100,101,102,103]. There is an urgent need for the development of a tool that accounts for ICU-specific variables, such as ventilator waveforms, sedation levels, fluid shifts, and physiologic instability

Current literature remains sparse on prospective studies linking OSA screening or diagnosis within ICU settings to meaningful patient-centered outcomes, such as mortality, length of stay, ventilator duration, or readmission risk [104]. Most existing studies are retrospective, observational, or limited either by small sample size or heterogenous methodology. This necessitates high-quality, prospective cohort studies and randomized controlled trials to determine whether early management and targeted intervention of OSA in the ICU can alter clinical outcomes, especially in high-risk populations with heart failure, obesity, hypoventilation, or postoperative respiratory failure. [105,106,107,108]

The integration of OSA screening and diagnostic workflows in the ICU practice currently remains inconsistent and underdeveloped [109]. Developing institutional protocols, defining clear criteria for screening, setting diagnostic thresholds, and extubation management strategies is crucial [110]. The creation of ICU outpatient follow-up pathways for high-risk patients with OSA could facilitate easier definitive diagnosis and long-term treatment management. Engagement of multidisciplinary teams consisting of intensivists, sleep specialists, and respiratory specialists can also help ensure practical and sustainable protocol development. Timely diagnosis and CPAP treatment in high-risk hospitalized patients may reduce resource utilization, and as such should be investigated in ICU settings. Establishing evidence-based policies surrounding OSA management in critical care settings may improve both short and long-term outcomes, improving critical care standards.

4. Materials and Methods

A critical analysis of published literature was conducted using PubMed, EMBASE, and Scopus. Key search terms included “obstructive sleep apnea,” “ICU,” “critical illness,” and “sleep-disordered breathing.” The study selection process is summarized in a PRISMA flow diagram in Figure 1.

During the preparation of this work, the author(s) used Consensus to assist with initial literature surveying and to identify current research trends. All final article selection, critical appraisal, synthesis, interpretation, and manuscript writing were performed by the authors. After using this tool, the author(s) reviewed, verified, and edited the content as needed and take full responsibility for the accuracy and integrity of the published work.

5. Conclusions

OSA continues to be a prevalent disorder globally. OSA is associated with adverse outcomes, such as arrythmias, heart failure, insulin resistance, and inflammation, which is exacerbated the longer it’s undiagnosed. Notably, the ICU presents a challenge to the detection and diagnosis of OSA. Here we highlight various aspects of the ICU and its patient population that elevate the risk of undiagnosed OSA, including shared risk factors such as obesity. Although limited, previous research studies underscored the detrimental effects of delayed OSA recognition in critically ill patients, such as higher risk of hospital readmission, delirium, extubation delays, and elevated mortality risk. Current screening measures, including STOP-BANG and BERLIN questionnaire, fail to be applicable to ICU patients. Moreover, sedation and failure to protect the airway prevent patients from achieving maximal benefit via CPAP, the traditional gold standard treatment. Future research should focus on identifying the most effective diagnostic and treatment approaches for OSA in ICU patients to improve long-term clinical outcomes and reduce the burden on the healthcare system.