Sleep Testing for Obstructive Sleep Apnea (OSA)

CMS NCA document | source_status=Closed | review_type=New | public_comment_open=False | document_id=CAG-00405N

CMS finds that the evidence is sufficient to determine that the results of the sleep tests identified below can be used by a beneficiary’s treating physician to diagnose OSA, that the use of such sleep testing technologies demonstrates improved health outcomes in Medicare beneficiaries who have OSA and receive the appropriate treatment, and that these tests are thus reasonable and necessary under section 1862(a)(1)(A) of the Social Security Act. Therefore: Type I Polysomnography (PSG) is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have clinical signs and symptoms indicative of OSA if performed attended in a sleep lab facility. A Type II or a Type III sleep testing device is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have clinical signs and symptoms indicative of OSA if performed unattended in or out of a sleep lab facility or attended in a sleep lab facility. A Type IV sleep testing device measuring three or more channels, one of which is airflow, is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have signs and symptoms indicative of OSA if performed unattended in or out of a sleep lab facility or attended in a sleep lab facility. A sleep testing device measuring three or more channels that include actigraphy, oximetry, and peripheral arterial tone is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have signs and symptoms indicative of OSA if performed unattended in or out of a sleep lab facility or attended in a sleep lab facility.

Decision Memo: To: Administrative File: CAG # 00405N Sleep Testing for Obstructive Sleep Apnea (OSA) From: Steve Phurrough, MD, MPA Director, Coverage and Analysis Group Louis Jacques, MD Director, Division of Items and Devices Jean Stiller, MA Lead Analyst Ross Brechner, MD, MS (Stat.), MPH Lead Medical Officer Subject: Coverage Decision Memorandum for Sleep Testing for Obstructive Sleep Apnea (OSA) (CAG-00405N) Date: March 3, 2009 I. Decision CMS finds that the evidence is sufficient to determine that the results of the sleep tests identified below can be used by a beneficiary’s treating physician to diagnose OSA, that the use of such sleep testing technologies demonstrates improved health outcomes in Medicare beneficiaries who have OSA and receive the appropriate treatment, and that these tests are thus reasonable and necessary under section 1862(a)(1)(A) of the Social Security Act. Therefore: Type I Polysomnography (PSG) is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have clinical signs and symptoms indicative of OSA if performed attended in a sleep lab facility. A Type II or a Type III sleep testing device is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have clinical signs and symptoms indicative of OSA if performed unattended in or out of a sleep lab facility or attended in a sleep lab facility. A Type IV sleep testing device measuring three or more channels, one of which is airflow, is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have signs and symptoms indicative of OSA if performed unattended in or out of a sleep lab facility or attended in a sleep lab facility. A sleep testing device measuring three or more channels that include actigraphy, oximetry, and peripheral arterial tone is covered when used to aid the diagnosis of obstructive sleep apnea (OSA) in beneficiaries who have signs and symptoms indicative of OSA if performed unattended in or out of a sleep lab facility or attended in a sleep lab facility. II. Background We use the abbreviation PSG to refer to polysomnography or a polysomnogram furnished in a sleep laboratory facility. Unless we specifically describe an unattended use, we will always assume in this document that it has been attended. We note that some authors use the abbreviation NPSG to mean nocturnal PSG. We use the abbreviation HST (home sleep test) to refer to unattended multichannel sleep testing or multichannel sleep monitoring typically furnished in the beneficiary’s home. However, it does not exclude these tests being performed in other settings to include a sleep lab. OSA, sometimes referred to as Obstructive Sleep Apnea Hypopnea Syndrome-(OSAHS), is associated with significant morbidity and mortality. It is a commonly underdiagnosed condition that occurs in 4% of men and 2% of women (Young et al. 1993). The prevalence increases with age (up to 10% in persons 65 and older), as well as with increased weight. Complications associated with OSA include excessive daytime sleepiness, concentration difficulty, coronary artery disease, and stroke (Kokturk et al. 2005). It is estimated that 10% of patients with congestive heart failure (CHF) have OSA, which is independently associated with systemic arterial hypertension (Caples et al. 2005). Untreated OSA is associated with a ten-fold increased risk of motor vehicle accidents (Teran-Santos et al. 1999). The most common clinical presentation of patients with OSA is obesity accompanied by excessive daytime drowsiness (20% of adults with BMI > 30 have OSA), although other clinical findings associated with OSA include nocturnal choking or gasping, witnessed apneas during sleep, large neck circumference and daytime fatigue. Of the three different forms of sleep apnea (obstructive, central, or mixed), OSA is the most common. Patients suffering with sleep apnea may literally stop breathing (apnea) for a short period or have decreased breathing (hypopnea), repeatedly during sleep. The apnea episodes often last for a minute or longer, and can occur hundreds of times during a single night’s sleep. During the obstructive apnea episodes, either complete or partial obstruction of the airway occurs. The anatomic site of obstruction is thought to be the soft palate, extending to the base of the tongue. When patients with OSA fall asleep muscles of this region may relax to the point of permitting airway collapse and obstruction. When the airway closes, breathing stops and the sleeper awakens to open the airway. Arousals from sleep usually last only a few seconds, but these brief arousals disrupt continuous sleep and prevent persons from reaching deep stages of sleep (e.g., rapid eye movement sleep-REM), which is necessary in order for the body to rest and replenish strength. The patient repeats this cycle throughout the sleep period. OSA has often been defined by an apnea-hypopnea index (AHI) or respiratory disturbance index (RDI) of ≥ 5 events per hour during sleep (when using this less restrictive definition, the prevalence may be as high as 25% of the population) or by a higher threshold e.g. AHI of ≥ 15 per hour (the prevalence is approximately 3%). Medicare covers CPAP for the treatment of OSA if the beneficiary has an AHI or RDI ≥ 15 events/hour. Medicare also covers CPAP for the treatment of OSA if the beneficiary has a co-morbidity related to OSA and the AHI or RDI is ≥ 5. The key diagnostic finding in OSA is episodes of airflow cessation or reduction at the nose and mouth despite evidence of continuing respiratory effort. Other common clinical findings and measurements used by physicians in the diagnosis of OSA include oxygen desaturation, abnormal oxygen desaturation index, arterial pulsatile tone changes, measurement of airflow, measurement of breathing patterns, Multiple Sleep Latency Testing (MSLT), Maintenance of Wakefulness Testing, computerized EEG analysis, autonomic arousal detection, and body movement analysis (e.g., actigraphy). Diagnostic tests for OSA have historically been classified into four types. The most comprehensive is designated Type I attended facility based PSG, which is considered the reference standard for diagnosing OSA. Attended facility based polysomnogram is a comprehensive diagnostic sleep test including at least electroencephalography (EEG), electro-oculography (EOG), electromyography (EMG), heart rate or electrocardiography (ECG), airflow, breathing/respiratory effort, and arterial oxygen saturation (SaO 2 ) furnished in a sleep laboratory facility in which a technologist supervises the recording during sleep time and has the ability to intervene if needed. Overnight PSG is the conventional diagnostic test for OSA. The American Thoracic Society (ATS 1994) and the American Academy of Sleep Medicine (ASDA 1997) have recommended supervised PSG in the sleep laboratory over 2 nights for the diagnosis of OSA and the initiation of CPAP. Three categories of portable monitors (used both in attended and unattended settings) have been developed for the diagnosis of OSA. Type II monitors have a minimum of 7 channels (e.g., EEG, EOG, EMG, ECG-heart rate, airflow, breathing/respiratory effort, SaO 2 )-this type of device monitors sleep staging, so AHI can be calculated). Type III monitors have a minimum of 4 monitored channels including ventilation or airflow (at least two channels of respiratory movement or respiratory movement and airflow), heart rate or ECG, and oxygen saturation. Type IV devices may measure one, two, three or more parameters but do not meet all the criteria of a higher category device. Some monitors, e.g. WatchPAT use an actigraphy algorithm to identify periods of sleep and wakefulness. Young et al. (1993) note limited capacity to provide PSG testing to all persons with symptoms of OSA due to the high prevalence of OSA. Some studies have noted false-negative rates of 14 to 25% (Le Bon et al. 2000; Littner 2000). And as noted by Klingshott et al. (2000), the measures derived from PSG (e.g., AHI) correlate poorly with major consequences of OSA such as sleepiness and cognitive impairment. Loube et al. (1999) and others have also noted that these measures do not reliably predict the response to the standard therapy for OSA, i.e., nasal CPAP. PSG alternatives have been sought. Predictive algorithms (predictive formulae) to determine optimal CPAP (Flemons et al. 1994; Maislin et al. 1995; Rowley et al. 2000), screening oximetry (Whitlaw et al. 2005; Chiner et al. 1999), attended/unattended home diagnostic apnea monitoring devices (Sériés et al 1993; Golpe et al. 2002; Whitelaw et al. 2005), and questionnaires (e.g., Epworth Sleepiness Scale (ESS); Sleep Apnea Clinical Scores) have been developed to help diagnose OSA. Other strategies that have been suggested to reduce the delay, inconvenience and expense associated with sleep studies include split night studies (Yamashiro et al. 1995), partner titration, and home stepwise titration. A number of treatment approaches have been recommended for patients with OSA, depending on the severity of the disorder (e.g., the degree of clinical symptoms), as well as the objective level of nocturnal respiratory and sleep disturbance (e.g., daytime sleepiness or number of obstructive events per hour of sleep). For patients with severe OSA, nasal CPAP is the treatment of choice. Its regular use improves excessive sleepiness, cognitive performance, and quality of life (Jenkinson et al. 1999; Montserrat et al. 2001). In patients with severe OSA who can not tolerate nasal CPAP, surgical procedures (e.g., uvulopalatopharygnoplasty-UPPP, maxillofacial surgery) may be indicated. In patients with mild to moderate OSA, nasal CPAP may be indicated, though conservative measures such as weight reduction, avoidance of alcohol, avoidance of sleeping in a recumbent position, or intra-oral appliances may be better tolerated. III. History of Medicare Coverage We received an external request from Itamar Medical requesting a National Coverage Determination (NCD) to determine whether Home Sleep Testing (HST) devices measuring the peripheral arterial tone (PAT) signal (a measure of sympathetic activation), heart rate, blood oxygen saturation, and sleep time are reasonable and necessary for the diagnosis of obstructive sleep apnea (OSA). Itamar manufactures the Watch-PAT sleep test device that measures actigraphy, oximetry and peripheral arterial tone. Itamar also asked us to remove this technology from the Type IV classification in the current CPAP NCD and explicitly state that CPAP is covered in beneficiaries diagnosed with OSA using a clinical evaluation and a positive test using this technology. CMS has addressed the coverage of CPAP in three separate decisions in October, 2001, April 2005, and March 2008. In each of those decisions, we limited coverage of CPAP in patients with OSA to those patients whose diagnosis was based on specific testing modalities. Initially, we limited coverage to OSA diagnosed with PSG. In the latest decision, we expanded coverage to OSA diagnosed with several types of HST. However, we have not, at a national level, specifically addressed coverage of the tests themselves. In other words, we nationally cover CPAP for beneficiaries with OSA if diagnosed with these specific tests; yet, coverage of the specific tests has previously been left to local contractor discretion. Since Watch-PAT is only one of several diagnostic tests for OSA and we do not have an NCD on any of these tests, we have broadened the scope of this NCA to include other sleep test technologies. Based on our extensive review of the available evidence, and significant programmatic need to clarify coverage of sleep testing for OSA, we published our proposed decision concurrent with opening this NCA. Benefit Category Medicare is a defined benefit program. All services furnished and covered under the Medicare program must be medically reasonable and necessary, and appropriate for diagnosis and/or treatment of an illness or injury. Furthermore, physicians and nonphysician practitioners must be authorized by the State in which the services are furnished to render the services. An item or service must fall within a benefit category as a prerequisite to Medicare coverage: § 1812 (Scope pf Part A); § 1832 (Scope of Part B); § 1861(s) (Definition of Medical and Other Health Services). The above described services fall within the benefit category at §1861(s)(3) of the Social Security Act, “other diagnostic tests.” The Medicare regulations at 42 CFR 410.32(a) state in part, “…diagnostic tests must be ordered by the physician who is treating the beneficiary, that is, the physician who furnishes a consultation or treats a beneficiary for a specific medical problem and who uses the results in the management of the beneficiary’s specific medical problem.” IV. Timeline of Recent Activities Date Action December 23, 2008 CMS posts a tracking sheet, publishes a proposed decision memorandum, and opens the initial 30 day public comment period. January 22, 2009 Public comment period ended. V. Food and Drug Administration (FDA) Status Certain sleep test devices have been considered and cleared for marketing by the FDA under a 510(k) process. VI. General Methodological Principles When making NCDs, CMS evaluates relevant clinical evidence to determine whether or not the evidence is of sufficient quality to support a finding that an item or service falling within a benefit category is reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member. The critical appraisal of the evidence enables us to determine to what degree we are confident that: 1) the specific assessment questions can be answered conclusively; and 2) the intervention will improve health outcomes for Medicare beneficiaries. An improved health outcome is one of several considerations in determining whether an item or service is reasonable and necessary under § 1862(a)(1)(A) of the Act. A detailed account of the methodological principles of study design that are used to assess the relevant literature on a therapeutic or diagnostic item or service for specific conditions can be found in Appendix A. In general, features of clinical studies that improve quality and decrease bias include the selection of a clinically relevant cohort, the consistent use of a single good reference standard, and the blinding of readers of the index and reference test results. Public comment sometimes cites the published clinical evidence and may give CMS useful information. Public comments that give information on unpublished evidence such as the results of individual practitioners or patients are less rigorous and therefore less useful for making a coverage determination. CMS uses the initial public comments to inform its proposed decision. CMS responds in detail to the public comments on a proposed decision when issuing the final decision memorandum. VII. Evidence A. Introduction We recently conducted an exhaustive review of the evidence for a clinical benefit of the available diagnostic tests for OSA during the March 2008 reconsideration of the NCD on CPAP for OSA. Medicare National Coverage Determinations Manual, §240.4. A complete discussion of that review can be found at http://www.cms.hhs.gov/mcd/viewdecisionmemo.asp?id=204 . Evidence about the diagnosis of OSA was central to the CPAP NCD reconsideration. In that reconsideration the medical community did not dispute the use of CPAP to treat OSA. Rather, the crux of the issue was whether or not home sleep testing could adequately identify symptomatic beneficiaries with OSA for whom CPAP should be covered by Medicare. Thus the evidence review, technology assessments and the MEDCAC provide information that is useful to also address the current questions. We are providing a summary of the applicable evidence here, and are including relevant new evidence that has come to light since that review. The available evidence includes published peer reviewed medical literature, external technology assessments and recommendations from the Medicare Evidence Development and Coverage Advisory Committee (MEDCAC). B. Discussion of evidence reviewed 1. Questions & Outcomes of Interest Question 1: Is the evidence adequate to determine that attended facility based polysomnography accurately identifies patients with OSA who will benefit from treatment? Question 2: For which unattended out of facility sleep test technologies is the evidence adequate to determine that sleep testing accurately identifies patients with OSA who will benefit from treatment? As diagnostic tests, PSG and HST would not be expected to directly change health outcomes. Rather, a diagnostic test affects health outcomes through changes in disease management brought about by physician actions taken in response to test results. Such actions may include decisions to treat or withhold treatment, to choose one treatment modality over another, or to choose a different dose or duration of the same treatment. To some extent the usefulness of a test result is constrained by the available treatment options. As noted in the Background section, the number of practical treatment options for OSA is limited. Most patients are treated with CPAP; a few get oral appliances or surgery. A patient whose OSA is not readily controlled with CPAP may seek other treatment, continue CPAP with lesser benefit, or discontinue CPAP and not seek further medical treatment. In addressing the questions above, one of the factors we considered is whether there is sufficient evidence that the incremental information derived from PSG or HST leads to improved treatment of OSA by causing physicians to prescribe a different, more appropriate treatment than they would have prescribed without access to the test results. Outcomes of interest for a diagnostic test are not limited to determining its accuracy but also include beneficial or adverse clinical effects, such as changes in management due to test findings or preferably, improved health outcomes for Medicare beneficiaries. Ideally, we would see evidence that the systematic incorporation of PSG or HST results into a treatment algorithm leads treating physicians to prescribe different and better treatment than they would otherwise have prescribed, and that those patients whose treatment is changed by test results remain on the regimen and achieve better long term OSA control documented by repeated assessments over time. There is no anatomic or physiologic "gold standard" for the diagnosis of obstructive sleep apnea, in contrast to conditions such as cancer where a tissue biopsy result is the definitive standard reference. In studies that compare HST to facility-based PSG, the investigators have used the PSG result as the standard reference; i.e. the PSG result is used to define the true disease state for the individual patient. This is less than ideal since the true sensitivity and specificity of PSG in diagnosing OSA is not well documented and this deficiency poses a practical difficulty in diagnosing OSA. Given the absence of a true "gold standard" reference, the clinical application of terms such as sensitivity and specificity is not straightforward. Such evidence permits only the comparison of HST to facility-based PSG. It is problematic to make the inferential leap from there to a judgment on the ability of HST or PSG to accurately identify those patients who will, if untreated with CPAP, suffer the morbidity and mortality of obstructive sleep apnea. If an individual patient has conflicting results with these two tests, e.g. a negative HST in the face of a positive PSG, there is no available higher reference to determine whether the conflict arises from a false negative HST or a false positive PSG. 2. External technology assessments Systematic reviews are based on a comprehensive search of published studies to answer a clearly defined and specific set of clinical questions. A well-defined strategy or protocol (established before the results of the individual studies are known) guides this literature search. Thus, the process of identifying studies for potential inclusion and the sources for finding such articles is explicitly documented at the start of the review. Finally, systematic reviews provide a detailed assessment of the studies included. CMS commissioned two TAs from AHRQ for the March 2008 CPAP NCD reconsideration: Home diagnosis of Obstructive Sleep Apnea-Hypopnea Syndrome, and Obstructive Sleep Apnea-Hypopnea Syndrome: modeling different diagnostic strategies The full reports are available at the following CMS website: http://www.cms.hhs.gov/mcd/viewtechassess.asp?id=204 . Home diagnosis of Obstructive Sleep Apnea-Hypopnea Syndrome Ninety-three studies were included in a review of the literature. Eligible studies assessed the ability of sleep studies at baseline to predict response to CPAP treatment or CPAP use, the comparison of measurements with portable monitors and facility-based PSG, and the safety of sleep studies. The TA reported that the reference standard for the diagnosis of OSAHS is facility-based PSG, a comprehensive sleep study that records and evaluates a variety of cardiorespiratory and neurophysiologic signals during sleep time. It quantifies the severity of disturbances with the Apnea-Hypopnea Index (AHI). Higher AHI values imply more severe sleep disturbances. Typically, a value of 15 or more events/hour of sleep is considered to be suggestive of OSAHS. An AHI suggestive of OSAHS is neither sufficient nor necessary for the diagnosis of the condition, as the severity of symptoms has to be accounted for, and other conditions affecting sleep may need to be excluded. Baseline AHI is only modestly associated with response to CPAP use among people with high (pre-test) probability for OSAHS. The same is true for other indices obtained from sleep studies such as the mean or minimum O 2 saturation, apnea index, hypopnea index, frequency of arousals and other quantities. Based on limited data, the authors conclude that type II monitors may identify AHI suggestive of OSAHS with high positive likelihood ratios (> 10) and low negative likelihood ratios (< 0.1) both when the portable monitors were studied in the sleep laboratory and at home. Type III monitors may have the ability to predict AHI suggestive of OSAHS with high positive likelihood ratios and low negative likelihood ratios for various AHI cutoffs in laboratory-based PSG, especially when manual scoring is used. The ability of type III monitors to predict AHI suggestive of OSAHS appears to be better in studies conducted in the specialized sleep unit compared to studies in the home setting. Some studies of type IV monitors also showed high positive likelihood ratios and low negative likelihood ratios, at least for selected sensitivity and specificity pairs from ROC curve analyses. As with type III monitors, the ability of type IV monitors to predict AHI suggestive of OSAHS appears to be better in studies conducted in specialized sleep units. Medicare beneficiaries are older on average than the studied subjects (the median average age was approximately 50 years in the analyzed studies), and may more often have conditions other than OSAHS that affect sleep (e.g., Periodic Limb Movements in Sleep and Restless Leg Syndrome; cardiac insufficiency). These conditions may be misdiagnosed as OSAHS by sleep monitors that do not record channels necessary for the differential diagnosis of OSAHS. Therefore, some type III and type IV monitors may yield more false positives among Medicare beneficiaries, compared to what was observed in the assessed studies. For studies in the home setting, there are no direct data on whether and to what extent technologist support and patient education affect the comparison of portable monitors with facility-based PSG. For monitors that may be considered other than Type II, III, or IV, the authors found there is insufficient evidence to judge their value in diagnosing OSA. The TA differentiated Type IV monitors with three or more channels from those with one or two channels, finding greater diagnostic ability for the former. We note that the TA reviewed the Watch-PAT100 device as a Type IV device with three or more channels. Obstructive Sleep Apnea-Hypopnea Syndrome: modeling different diagnostic strategies The TA authors created a model to test the impact of different OSA strategies. When middle-aged people (50 years old) with symptoms and signs suggestive of OSAHS were tested in the home setting, approximately 10 percent of those with OSAHS are expected to remain undiagnosed; approximately 15 percent of those without OSAHS receive false-positive diagnoses. For older adults (70 years old), the expected number of misclassifications was larger, due to the expected increase in false positive diagnoses (30 percent). With the combination strategy that used home diagnosis and split-night PSG, almost 20 percent of middle-aged people with OSAHS received a (false) negative diagnosis, while the proportion of false positive results among 50 year-old people without OSAHS was very low (1 percent). The expected numbers were similar among older adults (70 years old). Both for middle-aged people and for older adults, the average time spent undiagnosed was practically negligible for the strategies that use home monitoring. In the combination strategy, people with positive diagnosis with the portable monitors received a final split-night PSG diagnosis within 15 weeks on average. When diagnosis of OSAHS and treatment initiation were managed outside the sleep laboratories in the home setting, middle-aged people with OSAHS spent on average ten weeks or nine percent of the total follow-up time in undiagnosed health states. Significantly, the corresponding mean time delay for middle-aged people is 27 weeks when they were managed with facility-based PSG. This number mainly reflects those with false negative diagnoses, who were never started on CPAP. The same delay is expected among older adults (70 years old). With the combination strategy, using home diagnosis and split-night PSG, correctly diagnosed people initiate CPAP after approximately 15 weeks. However, one fifth of the patients are not diagnosed and, overall, the average time spent while not on CPAP ("high-risk" states) becomes 33 weeks. Similar numbers are expected among older adults who have OSAHS. 3. Internal technology assessment Literature Search CMS performed an extensive literature search utilizing PubMed for randomized controlled trials (RCTs), systematic reviews, and series studies evaluating the technology used for the diagnosis of OSA. The literature search was limited to humans. There are currently several proposed mechanisms to diagnose OSA and determine the need for and benefit of OSA treatment, specifically CPAP. These include clinical diagnosis alone, PSG, home testing with various devices and a diagnosis made by using a trial of CPAP. Clinical Diagnosis Alone and Clinical Diagnosis with PSG Crocker et al. (1990) studied whether the number of PSGs required for diagnosis of OSA could be reduced in the population. They enrolled 100 consecutive patients (average age 50) screened by family and sleep physicians. The patients were then tested by PSG. A clinical model was created for predicting a diagnosis of OSA as compared to PSG and was applied to the next 114 consecutive patients, of whom IOS eventually had complete data. The model correctly classified 33 of 36 persons with OSA by correctly predicting an AHI > 15 and it correctly classified 35 of 69 patients by correctly predicting an AHI ≤ 15. In the model, BMI, reported apnea, age, and hypertension were statistically significant factors. The model had a sensitivity of 92% for predicting OSA when compared to PSG and a specificity of 51%. The authors concluded that clinical observation might reduce the need for PSG in the diagnosis of OSA by one-third. Deegan et al. (1996) compared the predictive value of certain clinical features to PSG for a diagnosis of OSA. Two hundred fifty consecutive patients (average age 45) were pre-screened by a physician and had a clinical assessment and administration of a sleep questionnaire, along with PSG. One hundred thirty six (54%) had an AHI ≥ 15 (considered positive for a diagnosis of OSA) and 114 (46%) had an AHI < 15 (not considered positive for OSA). Using clinical features and oximetry, 32.4% of patients could be confidently categorized, compared to PSG, as either having a true diagnosis of OSA or not having OSA. Significant factors in the model were BMI, alcohol intake, and age. The authors concluded that clinical observation may reduce the need for PSG in the diagnosis of OSA by approximately one-third. Haponik et al. (1984) asked whether or not PSG is necessary to assess the presence and severity of sleep-disordered breathing. They enrolled 37 patients (average age 50) with clinically suspected OSA, administered a questionnaire and did PSG testing. Compared to PSG (AHI ≥ 15 as cutoff for positive diagnosis of OSA) the clinical testing information had a sensitivity of 64% for a correct diagnosis of OSA and a specificity of 100%. The authors concluded that a single, brief clinical observation alone is an ineffective screening procedure for detecting OSA. Julià-Serdà et al. (1984) enrolled 225 consecutive referrals to a sleep clinic (average age 45 in the non-OSA group and 52 in the OSA group) with suspected OSA to determine whether or not cephalometry was useful in sparing PSG. All subjects had clinical assessment with an ESS questionnaire, physical examination and history. In addition they also had spirometry, cephalometry, and PSG testing. A statistical model was built to estimate a patient’s probability of a correct diagnosis of OSA as compared to PSG (using a cutoff value of AHI ≥ 10), based on clinical variables, physical examination, pulse oximetry, cephalometry, and soft palate and uvula measurements. The sensitivity of the model for a correct diagnosis of OSA as compared to PSG was 93% and the specificity was 83%. The authors concluded that cephalometry plus oximetry plus history and physical exam is capable of sparing the need for PSG in diagnosing OSA. Dixon et al. (1997) attempted to create a clinical model for predicting a correct diagnosis of OSA as compared to PSG in 99 pre-operative Laparoscopic Adjustable Gastric Banding patients with average age in their four groups ranging from 35 to 44. A thorough sleep history and physical examination were performed, checking for symptoms such as nocturnal choking, waking unrefreshed, morning headaches, excessive daytime sleepiness and poor sleep quality. An ESS was administered and all patients had a PSG test. The PSG was hand scored. For a PSG cutoff of AHI ≥ 15, independent predictors for a diagnosis of OSA were observed sleep apnea (the only positive symptom predictor of an AHI ≥ 15), male sex, higher BMI, age, fasting insulin and glycosylated hemoglobin A1c. From the model created, a scoring mechanism was established and a score of > 3 had a sensitivity of 89% for a correct diagnosis of OSA as compared to PSG and a specificity of 81% for moderate/severe OSA. The authors concluded that a simple method of predicting OSA in severely obese symptomatic subjects can assist in limiting the use of PSG to those with greater risk. Lim et al. (2006) performed a study to determine if a clinical model could be developed to predict OSA diagnosis from clinical diagnosis only. Seventy-one consecutive snorers (average age 44) referred for an evaluation for OSA were enrolled. OSA status was determined by clinical assessment based on symptoms suggestive of OSA as well as an ESS and BMI measurement. A PSG was administered and a clinical assessment model was created and used in identifying the ‘non-apneic snorers’ among patients referred with snoring. The model made use of the ESS score (using a cutoff of ≥ 15), the BMI (using a cutoff of ≥ 28), and the presence of symptoms such as nocturnal choking, witnessed apnea, daytime hypersomnolence or morning headaches. Compared to PSG using a cutoff of AHI > 10, the model had a sensitivity of 93.4% and a specificity of 60% for correctly diagnosing OSA. The authors concluded that identifying ‘non-apneic snorers’ in whom PSG could be avoided can be correctly accomplished via a clinical assessment if two out of three of the following are absent: 1) ESS score ≥ 15; 2) a BMI ≥ 28; and 3) the presence of specified symptoms such as nocturnal choking, witnessed apnea, daytime hypersomnolence or morning headaches. Hoffstein et al. (2006) utilized data from 594 patients with an average age of 47 who were referred to a sleep clinic for suspicion of sleep apnea and whom were all seen by the same physician to determine if it was possible to develop a clinical model to predict a correct diagnosis of OSA from a clinical exam. A PSG with a cutoff of AHI > 10 was used for the diagnosis of OSA. The independent predictors of a correct diagnosis of OSA as compared to PSG were age, sex, BMI, partner observation of apnea and pharyngeal exam findings (normal vs abnormal). Compared to PSG, the subjective (clinical) impression alone showed a sensitivity of 63% for a correct diagnosis of OSA and a specificity of 60%. The authors concluded that subjective impression alone is not enough to reliably identify patients with or without a correct diagnosis of OSA as compared to PSG. Garcia et al. (2003) studied whether or not they could predict a correct diagnosis of OSA with a clinical model. They enrolled 227 consecutive patients (average age 58) measuring clinical signs and symptoms and performing a PSG. They then took the next 102 patients and tested their model for clinical diagnosis of OSA (total 329). They utilized an AHI ≥ 30 as a cutoff for a correct diagnosis of OSA. In the model created, they utilized a cut point of 11 for the ESS and of 30 for BMI and included other significant and independent factors of age, sex, BMI, neck circumference history and the referring physician’s subjective feeling (dichotomized into ‘yes’ or ‘no’) as to each patient’s probability of having an AHI ≥ 30. Compared to PSG, the model had a sensitivity of 80% for a correct diagnosis of OSA and a specificity of 93%. The authors concluded that prior to diagnostic tests for OSA; clinical data can be useful for identifying patients suspected of having AHI ≥ 30. Kushida et al. (1997) attempted to predict OSA with a morphometric predictor model. Thirty patients (age range 15-75) were used to create the model and the model was then prospectively tested on the first consecutive 300 of a total of 423 patients referred for a diagnosis of OSA. All patients were also tested with PSG using a cutoff of AHI ≥ 10. The regression model included oral cavity measurements of the palatal height by two separate calipers measuring the distance between the mesial surfaces of the crowns of the second molars to obtain either the maxillary intermolar distance or the mandibular intermolar distance. BMI and neck circumference measurements were also made. The morphometric model had a sensitivity of 97.6% for a correct diagnosis of OSA as compared to PSG and a specificity of 100%. The authors concluded that the model may be clinically useful as a screening tool for OSA rather than as a replacement for PSG. Pillar et al. (1994) compared a clinical diagnosis of OSA to PSG (cutoff AHI ≥ 10). Eighty-six patients (average age 47) referred to a sleep clinic for suspicion of OSA were enrolled. The authors did not mention whether or not the subjects were consecutively enrolled. All patients answered a detailed sleep questionnaire, had a brief physical examination and had PSG testing. Compared to PSG (cutoff AHI ≥ 10), a clinical diagnosis of OSA had a sensitivity of 79% and a specificity of 50%. With regards to the model, the independent factors for a true diagnosis of OSA were neck circumference, age, self reporting of apnea and falling asleep unintentionally. Compared to PSG, the sensitivity was 92% and the specificity was only 18%. The authors concluded that clinical evaluation cannot replace PSG. Rauscher et al. (1993) enrolled 98 habitual snorers and 89 patients (average age 58 overall) with a positive diagnosis of OSA by PSG to see which snorers referred to a sleep laboratory need PSG for the diagnosis of OSA. A regression model was created that included weight, height, sex, witnessed episodes of apnea and falling asleep reading. This model was applied to 116 consecutive patients referred for investigation of heavy snoring. All patients with negative oximetry and a probability value < 0.31 for having OSA had an AHI < 10 by PSG. The authors concluded that snorers with negative oximetry classified as not having OSA by this model do not need PSG. Viner et al. (1994) examined whether or not history and physical examination can predict a correct diagnosis of OSA as compared to PSG. They enrolled 410 patients (average age 50) referred for clinically suspected OSA. They conducted a blinded comparison of history and physical examination versus results of nocturnal PSG utilizing a cutoff point of AHI ≥ 10. The regression model created included as significant independent factors age, BMI, sex, witnessed episodes of apnea and falling asleep reading. They noted that for p < 0.20 (a predicted probability of less than 20% of having OSA) the clinical model had 94% sensitivity and 28% specificity of correctly predicting a diagnosis of OSA as compared to PSG. Subjective impression alone had a sensitivity of 52% and a specificity of 70% for correctly predicting a diagnosis of OSA as compared to PSG. The authors concluded that in patients with a low predicted probability of having a correct diagnosis of OSA, approximately one-third do not need a PSG for diagnosis. Tsai et al. (2002) performed a study to create a decision rule for diagnostic testing in OSA. They enrolled 75 patients (average age 47) referred to a sleep clinic for suspicion of sleep apnea. No mention of consecutive selection was made. Each patient had portable RDI testing (using a cutoff of RDI ≥ 10) and nocturnal oxygen saturation measurements. During the feasibility phase, patients underwent routine clinical assessment plus the upper airway physical examination protocol (UAPP), performed by two investigators. Unreliable or time consuming measurements were eliminated from the UAPP based on clinical judgments and history of snoring and body position based on the consensus of the two investigators. A decision rule was developed using three predictors: a cricomental space (the perpendicular distance between the midpoint of the cricomental line, a straight line from the chin to the cricothyroid cartilage, and the skin of the neck) of 1.5 cm or less, a pharyngeal grade (I = palatopharyngeal arch intersects at the edge of the tongue; II = palatopharyngeal arch intersects at 25% or more of the tongue diameter; III = palatopharyngeal arch intersects at 50% or more of the tongue diameter; IV = palatopharyngeal arch intersects at 75% or more of the tongue diameter) of more than II and the presence of overbite. For patients with all 3 predictors (17%), the decision rule had a PPV of 95% and an NPV of 49% for a true diagnosis of OSA by PSG. Comparable performance was obtained in a validation sample of 50 patients referred for diagnostic testing. The authors concluded that their decision rule provides a simple, reliable and accurate method of identifying a subset patients with and, perhaps more importantly, without a true diagnosis of OSA. Home testing for OSA Types II, III & IV CMS reviewed the AHRQ TA assessment above and also found some additional evidence on HST. Ayappa et al. (2008) evaluated the ARES Unicorder, a self-applied, limited-channel portable monitoring device for the evaluation of sleep disordered breathing (SDB) using a prospective study with blinded analysis. Eighty patients with suspected OSA and 22 volunteers were enrolled of whom 97 were tested. Interventions used the ARES™ Unicorder at home for 2 nights using only written instructions. The number of men in the suspected OSA group was 60 and the number of women was 17, while in the volunteer group it was 9 and 11 respectively. The mean age in the suspected OSA group was 46 (range 26-74), while the mean age in the volunteer group was 36 (range 19-73). The mean BMI in the suspected group was 30 (range 21-70) the mean BMI was 24 in the volunteer group (range 19-32). Within 2 weeks, they returned to the laboratory for full nocturnal polysomnography (NPSG) with simultaneous monitoring with the Unicorder. NPSGs were scored manually to obtain an apnea-hypopnea index based on Medicare guidelines (AHI4%) and a respiratory disturbance index (RDI). ARES studies were autoscored and reviewed to obtain indices based on equivalent definitions i.e., AHI4%ARES, and apnea hypopnea (events with 1% desaturation) index (AHI1%ARES). Indices from the NPSG were compared to the in-lab ARES and in-home ARES indices using mean differences and the intraclass correlations (ICC). For the in-lab comparison, there was high concordance between AHI4%NPSG and AHI4%ARES (ICC = 0.96, mean difference = 0.5/hour) and RDINPSG and AHI1%ARES (ICC = 0.93, mean difference = 3.2/hour). For NPSG versus In-Home ARES comparison, there was good concordance between AHI4%NPSG and AHI4%ARES (ICC = 0.8, mean difference = 4.1/ hour) and RDINPSG and AHI1%ARES (ICC = 0.8 mean difference = 8.6/ hour). The diagnostic sensitivity of in-lab ARES™ for diagnosing OSA using an RDI cut-off of 15 per hour was 95% and specificity was 94%, with a positive likelihood ratio (LR+) = 17.04, and negative likelihood ratio (LR-) = 0.06. For in-home ARES data the sensitivity was 85% and specificity 91% (LR+ = 9.34, LR- = 0.17). There was good agreement between the manually scored NPSG OSA indices and the autoscoring ARES algorithm. The authors concluded that the ARES Unicorder provides acceptably accurate estimates of OSA indices compared to conventional laboratory NPSG for both the simultaneous and in-home ARES data and that the high sensitivity, specificity, and positive and negative likelihood ratios obtained in the group they studied supported the utility of an ambulatory limited-monitoring approach not only for diagnosing sleep disordered breathing but also to rule out OSA in suitably selected groups. Alvarez et al. (2008) aimed at evaluating the reliability of home respiratory polygraphy (using the Edentec Monitoring System polygraph) for the diagnosis of sleep apnea–hypopnea syndrome (OSA) and comparing the cost of this technique with that of nighttime polysomnography performed in a sleep laboratory. Using a prospective study design with a random sample of patients with clinically suspected OSA, the participants underwent both home respiratory polygraphy and nighttime PSG and were blinded as to the results of their first test. The study population was composed of 45 patients with a mean (SD) age of 52.3 (11) years of whom 21 (46.6%) were diagnosed with OSA, defined by an AHI> 10 by nighttime PSG. Comparison of the results between PSG and home polygraphy revealed statistically significant correlations for all comparisons. The optimal cutoff in this population was a RDI of 13.7 or more, for which the area under the receiver operating characteristic curve was 87.5% (95% confidence interval, 74.2%-95.4%). The authors concluded that home respiratory polygraphy is a reliable technique for the diagnosis of OSA and that uncertain results must be verified by nighttime PSG. Oximetry Both PSG and HST have an oximetry component, which monitors oxygen desaturation. A number of authors have claimed that just using the oximetry component alone can help in making a diagnosis of OSA (Nuber et al. 2000; Sériés et al 2005; Sériés et al.1993; Gyulay et al. 1993). Gyulay et al. (1993) studied 98 non-consecutive patients referred for suspicion of sleep apnea to a sleep clinic to compare clinical assessment with home oximetry in the diagnosis of OSA. All patients answered a questionnaire, had a history and physical exam, and had PSG testing using a cutoff value of AHI ≥ 15 for diagnosis of OSA. Physicians also independently estimated the likelihood of their patient having a true diagnosis of OSA on PSG testing. Compared to PSG, the independent clinical (physician) assessment had a sensitivity of 79% and a specificity of 50% for correctly diagnosing OSA at the cutoff value of AHI ≥ 15. Compared to PSG, oximetry with a desaturation of 2% had a sensitivity of 65% and a specificity of 74% for diagnosing OSA at the cutoff value of AHI ≥ 15. For desaturations of 3%, the corresponding sensitivity and specificity were 51% and 90%, respectively. If the percentage of sleep time spent at SaO 2 < 90 was ≥ 1%, the sensitivity for a true diagnosis of OSA as compared to PSG (AHI ≥ 15) was 93% and the specificity was 51%. The authors concluded that being at SaO 2 < 90 for < 1% of the time on home oximetry practically excludes OSA. As noted above, a number of studies have shown that oximetry measurement helps the diagnostic accuracy of OSA. Sériés et al. (1993) performed one of the earliest studies exploring this relationship. Using 240 consecutive patients with a confirmed (AHI > 10 on PSG) diagnosis of OSA (all were clinically suspected of having OSA because of loud snoring; nocturnal choking and awakenings or apneic events or all three reported by a bedmate; bad sleep quality; and daytime hypersomnolence), they found that oximetry had a 98% sensitivity for diagnosing OSA (AHI > 10), but a specificity of only 48%. Magalang et al. (2003) explored the relationship between oximetry and OSA. They noted that several quantitative indices derived from overnight pulse oximetry have been used to predict the presence of OSA: (1) number of episodes of oxyhemoglobin desaturations below a threshold-usually a 3% or 4% decline below baseline, (2) the cumulative time spent below an oxyhemoglobin saturation of 90%, and (3) the Δ [delta] index—a measure of the variability of the oxyhemoglobin saturation. The researchers wanted to compare these indices and determine if some combination of these indices predicted an individual’s AHI as measured by PSG. Using a derivation group which consisted of 224 consecutive patients, a prediction model was generated based on AHIs from the calculated quantitative indexes. The model was further validated using two groups of consecutive eligible patients (group 1 consisted of 101 patients and group 2 consisted of 191 patients). All patients underwent standard overnight PSG and measurement of arterial oxyhemoglobin (by pulse oximeter). The major findings of the study revealed that among the different oximetry indices, the Δ index was the best predictor of the presence of OSA, though the number of desaturation events provided similar levels of diagnostic accuracy (sensitivity of a Δ index of > 0.63 in the diagnosis of OSA was 91%, while the specificity was 59%). An aggregation of the model using combinations of all oximetry indices reduced the prediction error (r 2 = 0.70, p < 0.05) compared to using the Δ index alone (r 2 = 0.60), improving the precision of prediction of the AHI. The correlation between the predicted and actual AHI was 0.77 when using the Δ index alone, but improved to 0.83 when using a combination of all three oximetry indices. The authors note that one limitation of the study is that the prediction model was validated using overnight pulse oximetry obtained simultaneously with PSG data in the sleep laboratory. However, one advantage of this approach is it eliminated the potential confounder of night-to-night variability of AHI, as well ensuring that oximetry data were collected in exactly the same environment as the PSG data. Vazquez et al. (2000) studied the diagnostic performance of an automated digital oximetry analysis based on falls and recovery of oxygen saturation and compared the results to PSG. After excluding subjects not eligible for the study, 241 participants with suspected OSA were enrolled in the study and randomly assigned to either PSG or automated off-line analysis of the digitally recorded oximetry signal. Study outcomes included PSG-derived AHI, and oximeter-derived respiratory disturbance index (RDI). The study revealed that the PSG-derived AHI and the oximetry-derived RDI were strongly correlated (R = 0.97); the mean (± 2SD) of the differences between AHI and RDI was 2.18 (± 12.34)/h. Using a case definition of 15 episodes/hour for both AHI and RDI, the sensitivity and specificity were 98% and 88% respectively. The authors noted that one limitation of the applicability of this study was that the algorithm was evaluated by comparison with simultaneous PSGs. They also commented that a number of studies have shown a difference in RDI between home and hospital settings, despite using the same monitor and controlling for technical difficulty. But the authors were quick to note that by evaluating patients in the sleep laboratory, potential confounders (such as technical difficulties associated with remote monitoring, night-to-night variability, and the effects of the home environment on RDI) are eliminated. Devices measuring peripheral arterial tone, actigraphy and oximetry We were asked to perform a separate review of the Watch-PAT technology, as there has been some uncertainty expressed about how to classify this device in the current Type schema. Watch-PAT100 is an HST device which measures the peripheral arterial tone (PAT) and actigraphy (a measure of movement) which are recorded with an ambulatory wrist-worn device (Watch-PAT100). The PAT signal is a measure of the pulsatile volum