Sleep disordered breathing (SDB) frequently comes to medical attention for the first time when patients are hospitalized for diagnosis and treatment of an associated condition (e.g., poorly controlled hypertension, myocardial infarction, congestive heart failure, stroke, or problems related to management of diabetes mellitus). Diagnosis of SDB is generally performed in a specialized facility, which is often inconvenient and expensive for the hospitalized patient. Expectant peri-operative management of patients with sleep apnea is critical, particularly if they are previously undiagnosed. An ideal diagnostic strategy for these patients has not been defined. Continuous positive airway pressure (CPAP) is the mainstay of treatment of patients with sleep apnea. Unfortunately, it is often difficult for very ill patients to tolerate CPAP, unless it is administered with a high level of expertise.

Sleep disordered breathing (SDB), particularly obstructive sleep apnea (OSA), is a prevalent condition that is often undiagnosed. The development of obstructive apnea events is related to upper-airway anatomy and function. The relative contribution of these factors may vary widely. Obstruction most often occurs when upper-airway muscle tone is decreased relative to wakefulness. The upper airway collapses, with inspiration leading to obstructed breathing. The patient restores breathing at the expense of sleep continuity only to enter another phase of obstruction as sleep returns. This repetitive cycle of disordered breathing often produces episodic hypoxemia, increased stimulation of the sympathetic nervous system, and poor sleep quality.[1] Other sleep-related respiratory disturbances include Cheyne-Stokes breathing and central sleep apnea (CSA), characterized by oscillation or absence of respiratory effort.[2]

Population studies of middle-aged adults indicate that approximately 9% of women and 24% of men have SDB.[3] It has been estimated that 5% of all adults in Western countries have undiagnosed sleep apnea.[4] Even a relatively mild degree of OSA may be associated with adverse consequences including excess mortality,[5, 6] coronary artery disease manifestation,[7] stroke,[8] insulin resistance,[9] and increased risk of automobile accidents.[10] Studies have concluded that subjects with SDB are more likely to use health care resources and services before diagnosis.[11, 12] Therefore, incidental hospitalization of a patient with OSA may present a valuable opportunity for diagnosis. It also represents a potential challenge in management.

The impact of SDB on medical and surgical patients has not been rigorously investigated because it is largely speculative. Episodic hypoxemia and poor sleep quality may compromise recovery from medical illnesses. Furthermore, it is reasonable to assume that patients whose upper-airway patency is compromised during sleep are also at risk following administration of analgesia and anesthesia. This paper will briefly review data by which we may estimate the likelihood that OSA is present in association with commonly encountered medical conditions. Some reasonable precautions to avoid complications of SDB in medical and surgical patients will be presented. Finally, the inherent difficulties of diagnosis and treatment of a sick patient not previously known to have SDB will be discussed.

Obstructive Sleep Apnea in Hospitalized Patients
It has been suspected for many years that OSA is causally linked to cardiovascular complications. Until recently, investigations, purported to demonstrate this association, have been criticized for not meeting currently accepted standards of evidence.[13] The strengths and weaknesses of more recent studies have been discussed elsewhere.[14-17] It is clear, however, that groups of patients undergoing treatment for a variety of conditions demonstrate an increase in the occurrence of daytime sleepiness, snoring, and OSA. These studies are relevant because they highlight the likelihood that SDB will be encountered in hospitalized patients.

Hypertension
The relation between hypertension and OSA is now well established.[5, 18, 19] Even mild SDB is associated with an increased risk of cardiovascular disease.[20] The prevalence of OSA in hypertensive patients is not known. However, patients with hypertension that is difficult to control are particularly likely to have OSA.[21] Logan et al.[22] found that 83% of patients with drug-resistant hypertension had apnea/hypopnea indices (AHI) >/= ten events per hour. Effective treatment of patients with sleep apnea may result in decreases in systolic and diastolic blood pressure.[23-25]

Coronary Artery Disease and Myocardial Infarction
Numerous studies have suggested that a high proportion of patients with coronary artery disease (CAD) have OSA. Using a non-laboratory based overnight recording device, Schäfer et al.[26] found a 30.5% prevalence of OSA (defined as an AHI >/= 10) in patients with angiographically proven coronary artery disease versus 19.7% in controls. Other studies have reported prevalence of OSA in patients with CAD ranging from 14-65%.[27-30] Peker et al.[31] highlight the potential importance of OSA in patients with CAD and conclude that the presence of OSA was independently associated with an increased risk of cardiovascular mortality. Death occurred in 37.5% of patients with OSA over a 5-year follow-up period compared with 9.3% in patients without OSA.

Stroke
There is a high prevalence of SDB after stroke. Dyken et al.[33] found OSA on polysomnography in 77% of men and 64% of woman with recent strokes compared with 23% and 14%, respectively, in controls. Studying patients with stroke and transient ischemic attacks, Bassetti and Aldrich[34]] found an AHI >/= 10 in 62.5% of patients and only 12.5% of control subjects. Obstructive and central sleep apneas are common occurrences immediately following first time stroke.[35] Good et al.[36] reported 40% of stroke patients admitted to a rehabilitation unit demonstrated OSA identified by oximetry and confirmed with polysomnography. Parra et al.[37] reported that 71% of first time TIA and stroke patients had AHI >/= 10 in the acute phase. Wessendorf, Teschler, and Wang[38] reported prevalence of 61%, 44%, 32%, and 22% based on AHI cutoff points of 5, 10, 15, and 20 respectively. Good et al.[36] found functional outcome to be significantly worse in stroke patients with SDB. By contrast, Parra et al.[37] and Iranzo et al.[39] did not find a significant effect of OSA on functional outcome.

Recently, Wessendorf et al.[40] and Sandberg et al.[41] conducted treatment trials with CPAP in stroke patients. Improvements in depressive symptoms, sense of well-being, and nocturnal blood pressure were reported. Sandberg et al.[41] emphasized that compliance with CPAP is a problem for stroke patients especially in the presence of delirium and cognitive impairment.

Diabetes Mellitus
As a result of obesity or as an independent consequence of SDB, OSA has recently been linked to type 2 diabetes. Two recent studies have examined the relation between OSA and insulin resistance. Punjabi et al.[42] found that an AHI >/= 5 demonstrated an independent relation to glucose intolerance and insulin resistance despite controlling for body mass index (BMI) and AHI. Ip et al.[43] confirmed this observation.

Al-Delaimy et al.[44] looked at the relation between snoring and diabetes in the Nurses’ Health Study cohort. After adjusting for age and BMI, they found that, over a 10-year follow-up period, occasional and regular snorers demonstrated a significant increase in diagnosis with type 2 diabetes.

Congestive Heart Failure
Up to 50% of patients with stable, medically treated heart failure will show signs of SDB, most often in the form of central sleep apnea.[47] Mortality in congestive heart failure patients with untreated SDB is much higher than in those with heart failure alone.[48] Treatment with CPAP has been shown to significantly decrease mortality in patients with end-stage heart failure and SDB.[49] A study by Sin et al.[50] showed that patients with central sleep apnea waiting for heart transplant treated with CPAP had a significant increase in transplant-free survival.

Effects of Analgesia and Anesthesia on Patients With Sleep Apnea
Several recent reviews have highlighted the importance of OSA as a potential risk for complications related to anesthesia and analgesia after surgery.[51-53] However, the magnitude of this problem has not been well defined.[54] One uncontrolled study found that 17% of patients with OSA receiving general anesthesia, developed complications related to airway management.[55] Many of the factors responsible for upper-airway compromise in sleep are also present with anesthesia.[56] Anesthetic agents and narcotic medications increase the tendency for the upper airway to collapse. Furthermore, these agents impair the arousal response that terminates an episode of apnea, thereby increasing apnea severity.[57] Anesthetic agents alter several characteristics of sleep but these effects are difficult to separate from the effects of pain, pain management, and sleep deprivation in the hospital.[58, 59] As pain decreases after surgery and the effects of sleep deprivation accumulate, many patients experience a rebound of REM sleep during which sleep apnea is often most severe.[60, 61] The relation between severity of apnea and risk of sedation is ill defined. Gupta et al.[62] reported that even mild OSA represented a risk for orthopedic patients.

Prevention of Complications During Surgery for Patients With Obstructive Sleep Apnea
The greatest problem facing the anesthesiologist attempting to minimize complications related to sleep apnea is the fact that many, if not most, patients with OSA are undiagnosed.[4] Often, the anesthesiologist’s contact with the patient before surgery is brief. Patients with known OSA should be advised to notify the surgeon and anesthesia personnel of that diagnosis. It is reasonable to consider postponing elective surgery for patients with historical and physical examination findings suggestive of sleep apnea until diagnosis and treatment can be accomplished. A high level of expectant monitoring is necessary for patients who must proceed with urgent or non-elective surgery. Of particular concern are patients who have previously undergone surgical therapy for sleep apnea because there is a significant tendency for relapse a few years after treatment.[63]

Preparation Before Surgery
Once it is established or suspected that the patient has sleep apnea, a number of precautions are essential. The hallmark of safety administering anesthesia to patients with OSA is the concept that the patient’s airway must be controlled at all times. Unsupervised sedation before surgery should be avoided. The patient’s airway should be secured with either an oropharyngeal or nasopharyngeal airway during induction. A 3-5 minute period of pre-oxygenation may provide a brief cushion of safety.

Anesthesia Management
A system for intubation over a fiberoptic scope should be available and used if there is any doubt regarding ability to intubate the patient. Paralyzing agents, even short-acting agents, should be used with caution.[64] A surgeon should be available to perform an emergency tracheotomy if necessary.

Problems with intubation may be avoided by use of local or regional anesthesia. However, sedation may be more dangerous than intubation with airway control in these patients. Extubation is perhaps the greatest danger for patients with sleep apnea. The patient must be sufficiently awake to protect his airway before extubation can be safely accomplished. If the possibility of the presence of sleep apnea has been overlooked, this may be the time when unexpected difficulties develop.

Management After Surgery
Patient-administered pain control systems should be used with caution. Orders for narcotic analgesics or sedative medications should not be written on an “as needed” basis. The patient using CPAP therapy at home should be allowed to bring his own equipment to the hospital. Close observation remains important because his ability to use this equipment unaided may be impaired after surgery. Also, it is possible that the effects of surgery and analgesia may alter the appropriate CPAP settings. If the patient has not used CPAP previously, it may be necessary to initiate this therapy empirically.

Diagnosis of Sleep Disordered Breathing
Diagnosis of SDB is most often based on a polysomnography performed in a specialized facility.[65] For polysomnography to be effective, the patient must be able to sleep for extended periods of time. The sleep-disrupting environment of the hospital is well recognized. There is likely to be significant impact on sleep quality and continuity resulting from pain, anxiety, as well as patient-monitoring procedures, medication administration, routine nursing care, and hospital noise. Significant OSA can occur even in the absence of any suggestive physical findings or history.[66, 67] Limited sleep studies may have some utility but the role of this technology has been inconsistently validated. Oximetry alone is of little value because it is inadvisable to allow a patient under the stress of acute illness or surgery to become hypoxemic. Furthermore, there can be no certainty that the patient’s sleep in the hospital is a valid reflection of sleep at home. Consequently, it may be necessary to initiate therapy for presumed sleep apnea on an empiric basis. It must be emphasized that the diagnosis must be confirmed when the patient has recuperated significantly.

Empiric Treatment With Continuous Positive Airway Pressure
Key features of successful use of CPAP include effective mask fitting, correct institution of an optimal pressure level, and adequate patient education. All of these keys are potentially compromised in the setting of care of an acutely ill patient. Patients clearly should be protected during nocturnal sleep as well as naps, which may occur anytime when a patient is hospitalized. Most CPAP masks are not designed to be worn 24 hours per day and may produce serious skin breakdown as well as discomfort. Recently introduced masks that cover the entire face, such as the Respironics Total (Respironics, Pittsburgh, PA) mask, have become popular for non-invasive ventilation in the critical care unit. It is surprisingly well tolerated and may represent a useful option for some patients.

CPAP, or more specifically bi-level CPAP, is being used increasingly frequently for noninvasive ventilation. As a result, most respiratory therapy departments are familiar with mask ventilation. Unfortunately, long-term compliance with this therapy is disappointing.[68, 69] Most patients with sleep apnea do not require bi-level CPAP. However, it is very important that patients being started on CPAP therapy for sleep apnea begin treatment under the best possible conditions. Otherwise, they are likely to discard the treatment when they return home and even refuse subsequent evaluation and management. Careful monitoring and support of patients treated with CPAP for OSA has been shown to improve compliance.[70, 71] Some attempt at titration of pressure to identify optimal CPAP settings should be performed. There may be a role for automatically adjusting CPAP devices. The potential benefits and limitations of this technology have recently been reviewed in detail.[72] It should be emphasised that auto-CPAP is not an adequate substitute for polysomnography testing when the patient has stabilized. Controlled studies of different CPAP modalities during the transition from hospital to home care are needed.

OSA is commonly encountered in patients hospitalized for conditions that may or may not be directly associated to SDB. This fact represents an opportunity for diagnosis and a challenge for management. Unfortunately, medical education has historically devoted minimal time to this aspect of human health and diseases.[73] Careful attention to reports of excessive sleepiness or other features of sleep apnea in the patient’s history must be taken seriously. Physical findings that predispose to airway compromise should be noted. Major complications of surgery may well be prevented if health care providers exercise due caution regarding SDB. Acute medical or surgical illness, particularly if the level of central nervous system function is affected by the illness or its treatment, may amplify SDB manifestations.

The ideal approach to diagnosis and therapy under these circumstances is unknown. The relation between OSA in the hospital setting and persistent OSA at home “after the dust has settled” is often unpredictable and may vary greatly depending on specific circumstances. For example, episodic hypoxemia in a hospitalized patient receiving opiates may not be a valid predictor of the presence or severity of sleep apnea after recuperation. In most cases, a presumptive diagnosis of SDB made in a hospitalized patient should be corroborated by a standard polysomnography evaluation when the patient is stable. Empiric use of self-adjusting or bi-level CPAP may be helpful in some circumstances. An unfavorable initial experience with CPAP can complicate long-term compliance. Therefore, positive airway pressure therapy must be applied with a high level of expertise in sick patients who are naive regarding this treatment.