Screening for obstructive sleep apnea: an evidence-based analysis

American Journal of Otolaryngology–Head and Neck Medicine and Surgery 27 (2006) 112 – 118 www.elsevier.com/locate/amjoto

Screening for obstructive sleep apnea: an evidence-based analysis Kenny P. Pang, FRCSEd, FRCSI(OTO), FAMS(ORL)a,b,T, David J. Terris, MD, FACSa a

Department of Otolaryngology-Head and Neck Surgery, Medical College of Georgia, Augusta, GA, USA b Department of Otolaryngology, Tan Tock Seng Hospital, Singapore Received 28 June 2005

Abstract

Sleep disordered breathing is a spectrum of diseases that includes snoring, upper airway resistance syndrome, and obstructive sleep apnea (OSA). Obstructive sleep apnea is a common sleep disorder and is estimated to have an incidence of 24% in men and 9% in women. However, many authors believe that up to 93% of women and 82% of men with moderate to severe OSA remain undiagnosed. There is a strong link between sleep disordered breathing and hypertension, believed to be due to sleep fragmentation, intermittent hypoxemia, and increased sympathetic tone, which results in a higher mortality and morbidity rate among these patients. It is therefore desirable to attempt to diagnose all patients with OSA, to institute early treatment intervention, and to prevent development of cardiovascular complications. The gold standard for diagnosing OSA remains the attended overnight level I polysomnogram. However, in view of the limited resources, including limited number of recording beds, high cost, long waiting lists, and labor requirements, many authors have explored the use of clinical predictors or questionnaires that may help to identify higher-risk patients. Screening devices in the form of single or multiple channel monitoring systems have also been introduced and may represent an alternative method to diagnose OSA. The ideal screening device should be cheap, readily accessible, easily used with minimal instructions, have no risk or side effects to the patient, and be safe and accurate. We review a variety of clinical predictive formulae and several screening devices available for the diagnosis of OSA. D 2006 Elsevier Inc. All rights reserved.

1. Introduction Sleep is a basic human need in which there is a transient state of altered consciousness with perceptual disengagement from one’s environment. The average human spends 6 to 8 hours per day (about one third of his or her lifetime) sleeping. Contrary to common perception, sleep is an active process involving complex interactions between cortical, brainstem diencephalic, and forebrain structures. There is significant metabolism and oxygen consumption during this state of brest,Q and any disruption of oxygenation or interruption of this physiological process can lead to manifestations such as snoring, choking sensations, apneic episodes, poor concentration, memory loss, or daytime somnolence.

T Corresponding author. Department of Otolaryngology, Tan Tock Seng Hospital, Singapore, 11 Jalan Tan Tock Seng, Singapore, Republic of Singapore. Tel.: +65 63577742; fax: +65 63577749. E-mail address: [email protected] (K.P. Pang). 0196-0709/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.amjoto.2005.09.002

Snoring is caused by the vibration of the structures in the oral cavity and oropharynx—the soft palate, uvula, tonsils, base of tongue, epiglottis, and pharyngeal walls. It has historically been considered a nuisance and an objectionable social problem. However, snoring is important because it may represent an alarm to alert one to the possibility of a sleep disorder. Sleep disordered breathing (SDB) is a spectrum of diseases that includes snoring, upper airway resistance syndrome, and obstructive sleep apnea (OSA). Obstructive sleep apnea is a common sleep disorder. Young et al [1] studied 602 state employees with a formal overnight polysomnography and found that the incidence of SDB was 24% in men and 9% in women. Most of these patients are undiagnosed. It is estimated that up to 93% of women and 82% of men with moderate to severe OSA remain undiagnosed [2]. It is well documented that SDB has profound effects on the cardiovascular and respiratory systems and on neurocognitive function. The Sleep Heart Health Study and the Wisconsin Sleep Cohort [3,4] have demonstrated a strong

K.P. Pang, D.J. Terris / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 27 (2006) 112 –118

link between SDB and hypertension. This is believed to be due to sleep fragmentation, intermittent hypoxemia, and increased sympathetic tone [5]. This increased sympathetic tone is manifested not only during the nocturnal hypoxic events, but also during the day as systemic hypertension. There is convincing evidence of the association between SDB and cardiovascular disease [6]. The physiologic changes that are the result of recurrent apneas and hypoxemia can cause acute thrombotic events [7,8], atherosclerosis [9,10], and cerebrovascular accidents. There is a higher mortality rate among patients with cardiovascular disease who also have SDB [11,12]. Sleep-disordered breathing has also been associated with an increased risk of congestive heart failure [13], and treatment can improve heart failure symptoms. An essential public health priority is the diagnosis of all patients with OSA to prevent development of severe cardiovascular morbidity and mortality. The gold standard for diagnosing OSA remains the attended overnight level I polysomnogram. However, these studies suffer from limited resources, including recording beds, high cost, long waiting lists, and intense labor requirements. Moreover, elderly or sick patients often find the polysomnography (PSG) equipment too cumbersome, and may be reluctant to spend the night in the sleep laboratory. Several authors have therefore investigated the application of clinical predictors or questionnaires that may help identify higher-risk patients. Single and multiple channel monitoring systems have been evaluated for the ability to screen for OSA. The reliability and accuracy of these screening tools are largely unproven, however, and it is unclear if they are capable of predicting the severity of OSA. These issues are explored in this current analysis. 2. Clinical predictive models Symptoms of OSA include snoring, choking at night, witnessed apneic episodes, nocturia, and frequent arousals. Daytime manifestations are excessive daytime somnolence, poor concentration, poor memory, mood changes, and irritability. Puvanendran and Goh [14] studied 220 snorers who underwent an attended, hospital-based overnight polysomnogram and found that 87% of these habitual snorers have OSA. Tami et al [15] studied 94 habitual snorers and found an even lower incidence of only 72% of snorers having OSA. Vaidya et al [16] evaluated 309 snorers with a formal overnight polysomnogram and also found that 73% had OSA. When snoring is found in association with witnessed apneic events, the likelihood of finding OSA is higher [15]. The apneic events may be described as gasping for air during sleep or sensation of struggling or choking for breath during sleep resulting in frequent arousals [15]. Excessive daytime sleepiness is very sensitive but not specific for OSA because sleep deprivation can also give rise to daytime somnolence and a high Epworth (EPW) sleepiness scale score. Osman et al [17] studied 46 snorers with a full overnight polysomnogram and found that primary snorers

113

can also manifest high EPW scores because of other confounding factors. Only 41% of patients with OSA will have excessive daytime sleepiness, whereas 37% of primary snorers have excessive daytime sleepiness [18], as determined by Olson et al [18], in a study of 441 randomly selected community subjects who completed a detailed questionnaire and home PSG. However, many sleep specialists have demonstrated that the EPW scores are higher in individuals with OSA than in normals. Chung [19] illustrated this in 100 OSA patients and 61 controls, with mean EPW scores of 13.2 vs 7.5, respectively ( P b .001). There are several questionnaires and prediction models in the literature that acknowledge the higher incidence of a variety of these clinical symptoms in patients with OSA and attempt to use them to distinguish the OSA patients from the snorers. The most promising are outlined below: 1. Viner et al [20] developed a model that incorporated snoring, body mass index (BMI), age, and sex, and applied it in 410 patients; the model predicted only 52% of the patients who actually had OSA. The authors used stepwise linear logistic regression to develop 2 predictive models of sleep apnea: 1 based on the presence of characteristic clinical features, age, sex, and BMI, and 1 based on subjective clinical impression. They found that for patients with a predicted probability of apnea of less than 20%, the clinical model had 94% sensitivity and 28% specificity. 2. Maislin et al reviewed 427 patients with respect to snoring, gasping at night, witnessed apneas, age, sex and BMI. With their survey and formula, they were able to achieve a sensitivity of only 60% for the diagnosis of OSA [21]. The authors compared their survey data to respiratory disturbance index obtained from polysomnography. A multivariable apnea risk index including survey responses, age, sex and BMI was estimated using multiple logistic regressions. Predictive ability was assessed using receiver operating characteristic curves. The area under the receiver operating characteristic curve was 0.79 ( P b .0001). 3. The Berlin Questionnaire has 10 questions and includes the patient’s biomedical profile [22]. It uses a risk grouping system to predict the likelihood of the patient having OSA. Out of 744 patient responses, high-risk patients were found to have a higher BMI, loud snoring, hypertension, daytime sleepiness, and a propensity to fall asleep while driving [22]. It has been found to have a sensitivity of 86%, a specificity of 77%, and a positive predictive value of 89% for OSA. 4. Artificial neural nets have been used as prediction models. They assess 23 clinical variables including patient biomedical data, sleepiness, smoking and alcohol history, and physical examination [23]. The sensitivity from 405 patients evaluated was found to be 99%, with a specificity of 80%, a positive predictive

114

K.P. Pang, D.J. Terris / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 27 (2006) 112 – 118

value of 88.5% (Table 2). It has not found widespread application because it is labor-intensive and somewhat cumbersome. 7. Tsai et al [26] similarly investigated a large number of parameters and noted that there were 3 main reliable clinical symptoms of OSA: snoring, witnessed apneas, and hypertension; and 3 reliable signs of OSA: cricomental space less than 1.5 cm, pharyngeal grade higher than 2, and the presence of an overbite. When all of these are present, the positive predictive value was 95% for OSA. In their 75 patients, they also found that a cricomental space of more than 1.5 cm excluded OSA (negative predictive value of 100%).

Table 1 Predictive value of the various clinical predictive models Author

Viner et al [20] Maislin et al [21] Netzer et al [22] Kirby et al [23] Dixon et al [24] Kushida et al [25] Tsai et al [26]

Test

n

Sensitivity Specificity Negative Positive (%) (%) predictive predictive value (%) value (%)

410 94

28

427 60 Berlin Quest. ANN

744 86

77

405 99

80

99 96

71

Kushida 300 98 Index 75 40

89 98

88

100

88.5

100

96

100

95

Berlin Quest indicates Berlin Questionnaire; ANN, artificial neural net.

value of 88.1%, and a negative predictive value of 98% (Table 1) [23]. 5. Dixon et al [24] reviewed 99 patients suspected of having OSA and found that symptoms were poor predictors of OSA and that witnessed apneic spells were the only positive predictor. A higher BMI and a larger neck circumference were also predictive measures. Patients with these characteristics had very good correlation with an apnea hypopnea index (AHI) higher than 15, with a sensitivity of 96% and a specificity of 71% [24]. 6. Kushida et al [25], in 1997, described a complicated predictive formula using a morphometric model in 300 patients. This model included BMI, neck circumference, oral cavity measurements, palatal height, maxillary intermolar distance, mandibular intermolar distance, and overjet measurement, among others. When used properly, this model yielded a sensitivity of 97.6% and a specificity of 100%, with a positive predictive value of 100% and a negative predictive

Clinical predictors are notably unreliable in children. Wang et al [27] found that clinical symptoms such as snoring, witnessed apneic spells, daytime somnolence, mouth-breathing, and enuresis yielded a predictive value of only 30%. Similar results from Suen et al [28] failed to show any association between clinical symptoms and severity of OSA; with a predictive accuracy of only 51%. Goldstein et al [29] also found no statistical correlation between clinical predictive scoring and the existence of OSA in children. They found a positive predictive value of only 56%. Finally, Leach et al [30] found no difference in age, sex, symptoms or signs between 34 children with OSA and 54 normal children, with a predictive value of only 39%. Many authors favor the use of a combination of clinical symptoms, signs, and/or some form of questionnaire in the prediction of OSA. These prediction models have high sensitivity but low specificity for OSA [31]. There is, therefore, a need for some form of screening device to aid in the diagnosis of OSA. 3. Screening devices Several screening devices designed to assess the severity of OSA have emerged. They evaluate one or more

Table 2 Predictive value of the various devices Author

Test

n

Sensitivity (%)

Specificity (%)

Wiltshire et al [32]

Pulse oximetry NovaSom WatchPAT WatchPAT WatchPAT Somnocheck SleepStrip SleepStrip SNAP SNAP SNAP

100

36 – 100

23 – 99

51 30 42 102 51 288 20 59 31 60

91 89

83 78

83 84 86 97

95 71 80 87 57 63.3

Reichert et al [37] Ayas et al [38] Schnall et al [39] Bar et al [40] Ficker et al [44] Shochat et al [33] Lavie et al [46] Michaelson et al [49] Liesching [48] Su et al [50]

j Value measures degree of agreement.

90

Negative predictive value (%)

Positive predictive value (%)

Comments

15 studies

89

r r r r

= = = =

0.88 0.87, P b .0001 0.92, P b .0001 0.88, P b .0001

r = 0.73, P b .001

78

84

j = 0.23, P = .008 r = 0.92

K.P. Pang, D.J. Terris / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 27 (2006) 112 –118

parameters, and most are still undergoing validation. These devices serve to aid the physician in decision-making and should be used with prudence. The ideal screening device for OSA should be cheap, easily accessible, easily used with the minimal instructions, have no risk or side effects to the patient, and be safe and accurate. It should be capable of being issued by relatively unskilled staff and even sent through the mail to reduce patient travel and staff costs. 1. The pulse oximeter is a single monitoring device that has been commonly used as a screening tool for OSA. There are at least 15 studies of oximetry to date, many of which have varied results ranging from a sensitivity of 36% to 100% and a specificity of 23% to 99% [32,33]. Some of these were unattended, home-based studies, although the majority were done under full hospital supervision. Nixon et al [34] assessed 223 children with overnight pulse oximetry and compared it with PSG. They found no correlation with the pulse oximetry findings and the PSG. Kirk et al [35] likewise found poor correlation between the overnight pulse oximetry and the polysomnogram in a series of 58 patients. 2. QUISI is a single-channel self-applied ambulatory electroencephalogram (EEG) recording device that is designed to distinguish primary snoring from OSA. Fischer et al [36] conducted the first evaluation of its use in OSA in 2004 on 40 patients and found it to be inadequate. There was poor correlation between the QUISI readings and EEG, underscoring of stage 2, 3, and 4 sleep, and an overestimation of wakefulness. 3. NovaSom QSG is a 5-channel portable device used for home monitoring of OSA. It records nasal and oral airflow, oxygen saturation, respiratory effort and snoring intensity. The system uses patented audio digital signal processing technology to detect and analyze the patient’s respiratory sounds and convert them into airflow. The airflow sensor, which is worn on the patient’s upper lip, houses 2 microphones. One microphone captures the snoring sounds and ambient noise, whereas the other microphone captures the respiration sounds. The adaptive noise-canceling filters in the digital signal processing to subtract the snoring and ambient noise from the respiration process. The result is a pure respiration signal that has a fast response time and a linear correlation with measured airflow. The NovaSom QSG effort sensor is a thin Tygon tubing placed around the chest and is connected to a pressure transducer in the patient module. The finger sensor is a Nonin Adult Flexiform 7000A. It uses voice alerts to wake the patient up, in case of dislodgement of any of the sensors. When compared with the level I PSG, the home NovaSom QSG had a sensitivity of 91% and a specificity of 83%, whereas the in-laboratory NovaSom QSG

115

achieved a sensitivity of 95% and a specificity of 91% [37]. 4. A wrist-worn device, known as the WatchPAT 100, has been used to detect OSA. The PAT technology represents a unique and relatively new concept of noninvasive measurement of sympathetic activation levels that appears to be very accurate for detecting sleep-disordered breathing events (Fig. 1). This is a self-contained device worn around the wrist. The pressurizing mechanism of the probe allows it to be lightweight and silent, essential for a practical ambulatory device. Two finger probes extend from the main body of the device. One is the opticopneumatic sensor that detects the peripheral arterial tonometry (PAT) signal; the other measures arterial oxygen saturation. The body of the device also contains an actigraph (3-axis accelerometer for detection of limb activity), which is used to differentiate sleep from wakefulness. This technology uses a finger-mounted optic/pneumatic sensor that eliminates venous pulsations and continuously measures the pulse volume of the digit. Episodic vasoconstriction of digital vascular beds from sympathetic stimulation (mediated by a receptors) causes attenuation of the signal. Because discrete obstructive airway events (eg, apneas, hypopneas, and upper airway resistance) cause arousal from sleep, sympathetic activation, and peripheral vasoconstriction, these events are associated with attenuation of the PAT signal [38]. An automated computerized algorithm is used to calculate the frequency of respiratory events per hour of actigraphy-measured sleep. This algorithm also incorporates the PAT signal attenuation and the oxygen desaturation. Ayas et al [38] found good correlation between the WatchPAT 100 and the gold standard PSG. Schnall et al [39] found a high correlation between standard-polysomnography-scored apnea-hypopnea events and PAT-vasoconstriction events with concurrent tachycardia in an initial study with the bedside version of the system.

Fig. 1. Watch PAT-100 device, worn on the patient’s hand.

116

K.P. Pang, D.J. Terris / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 27 (2006) 112 – 118

Later, Bar et al [40] showed that detection of apnea and hypopnea events based on combined data from PAT and pulse oximetry was highly correlated with standard polysomnographic scored results, a finding that was confirmed by Pittman et al [41] usingboth manual and automatic analysis. O’Donnell et al [42] further explored the PAT response in patients with OSA. They experimentally induced upper airway obstruction and have shown that airflow obstruction in patients with OSA leads to a PAT signal attenuation in a bdose-responseQ manner, that is, greater airflow obstruction causes greater PAT attenuation. Further validation studies are necessary before this technology can be endorsed. 5. A new home apnomonitor has been used to establish the presence of OSA. It uses data from oronasal airflow, oxygen saturation, and fingertip plethysmography to distinguish central from peripheral apnea [43]. The respiratory flow curve becomes flat during an obstructive event, but the change in amplitude on plethysmography corresponds to thoracoabdominal motion. The change of amplitude in the plethysmogram does not correspond with thoracoabdominal motion under central sleep apnea conditions. These findings suggest that it is possible to distinguish between OSA and central sleep apnea, and to develop a new home-type apnomonitor [41]. 6. Somnocheck (SC) is a portable recording device that measures respiratory airflow (with thermistors), heart rate, oxygen saturation, body position, and tharocoabdominal movements. Ficker et al [44] found that it had good correlation with PSG when scored manually in 51 patients (correlation index, 0.98). The correlation with AHI obtained by automatic SC analysis was markedly lower (r = 0.83). When OSA was defined as an AHI greater than or equal to 10, the sensitivity of manual SC analysis was 97% and the specificity was 100%. The sensitivity of automatic analysis was 83% and its specificity, 95% [44]. 7. The SleepStrip is a small, lightweight device worn underneath the nose and above the upper lip (Fig. 2). The device is composed of flow sensors (oral and nasal thermistors), real-time analysis hardware and software, and a miniature display unit. Decreases or complete cessations of respiration are identified and recorded after a 20-minute calibration period. Patients are instructed to wear it just before bedtime and for at least 5 consecutive hours of sleep in their normal home environment. The SleepStrip score represents the number of respiratory events per hour of recording. From a multicenter trial, Shocat et al [33] studied 288 patients and obtained a sensitivity of 70% to 88% and a specificity of 57% to 94%, depending on the severity of the SDB. For AHI higher than 10, sensitivity was 86% and specificity was 57%; for AHI higher than 20, sensitivity was 80% and specificity

Fig. 2. The SleepStrip device is pasted on the patient’s face.

was 70%; for AHI higher than 40, the sensitivity was 80% and specificity was 86% [32]. Jaeho et al [45] described similar findings with the SleepStrip, which proved to be fairly reliable for screening of OSA. Lavie et al [46] showed that the SleepStrip was a reliable aid in screening for OSA in large populations. However, Hollingworth et al [47], mailed the Sleepstrips to 48 patients with instructions in the kit and found that the correlation of the SleepStrip data and the PSG data were poor. They qualified that this was likely due to poor compliance and that patients need more instructions than those given in the kit manual [47]. 8. Sleep sonography is the recording and analysis of snoring and has been used for home diagnosis of OSA. SNAP Laboratories uses technology that analyzes snoring recorded by a home microphone system. The microphone cannula is placed on the subject’s upper lip during sleep. The apparatus collects oronasal respiratory sound and airflow information, which are digitally recorded on a portable device. The recorded data are then returned to SNAP Laboratories, where they are analyzed using proprietary computer technology and software algorithms. The final test results are forwarded to the ordering physician. Liesching [48] validated the SNAP technology against the gold standard PSG in 31 patients and found modest correlation with a j value of 0.23 ( P = .008). SNAP study severity scores were overestimated in 13 of 31 patients (41.9%) compared with the polysomnography results. In most subjects (8 of the 13 boverestimatedQ patients or 8 of 31 total patients [25.8%]), the SNAP study diagnosed OSA when the patient had normal polysomnography findings. These results suggest that SNAP studies do not appear to accurately assess the severity of OSA. However, Michaelson et al [49] studied 59 patients with the SNAP technology, compared the data with an overnight PSG, and found a strong correlation coefficient between the SNAP and the PSG (0.916). Comparison of the SNAP and PSG

K.P. Pang, D.J. Terris / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 27 (2006) 112 –118

results when AHI was higher than 5 showed a sensitivity of 94% and a specificity of 86.6%. At an AHI N15, the sensitivity was 100% and the specificity was 88.5%. Su et al [50] revealed a correlation coefficient of 0.92 (AHI N 15), after comparing the SNAP with home polysomnography. The American Academy of Sleep Medicine, American College of Chest Physicians, and the American Thoracic Society recommended in 2003 that for patients strongly suspected of having OSA, a full comprehensive, attended overnight polysomnogram is required. A type 2 portable monitoring (PM) device (minimum of 7 channels) may be acceptable when incorporated into an attended PSG (including EEG, electro-oculography, electromyography, electrocardiogram, oxygen saturation, nasal airflow and respiratory effort) to diagnose OSA; however, data currently is lacking. A type 3 PM device (4-channel study) should be done in the hospital and attended by a technician to determine the presence of OSA. An unattended 4-channel study is not recommended. The type 4 PM device (single or double channel) is not recommended for the diagnosis of OSA [51].

4. Conclusion Obstructive sleep apnea is common, and most patients remain undiagnosed. These patients have a higher risk of hypertension and other cardiovascular morbidities if left untreated. Because the gold standard for the diagnosis of OSA (level 1 polysomnography), is cumbersome, expensive, labor-intensive, and requires expertise, it is desirable to explore clinical predictive models, questionnaires, and simple screening devices. Although there are several promising models and screening devices available, further validation will be necessary before widespread implementation can be done.

References [1] Young T, Palta M, Dempsey J, et al. The occurrence of SDB among middle-aged adults. N Engl J Med 1993;328:1230 - 5. [2] Young T, Evans L, Finn L, et al. Estimation of the clinically diagnosed proportion of sleep apnea syndrome in middle aged men and women. Sleep 1997;20:705 - 6. [3] Nieto JF, Young TB, Lind BK, et al. Association of SDB, sleep apnea & hypertension in a large community based study. JAMA 2000;283: 1829 - 36. [4] Peppard PE, Young T, Palta M, et al. Prospective study of the association between SDB & hypertension. N Engl J Med 2000;342: 1378 - 84. [5] Somers VK, Dyken ME, Clary MP, et al. Sympathetic neural in OSA. J Clin Invest 1995;96:1897 - 904. [6] Shakar E, Whitney CW, Redline S, et al. SDB & cardiovascular disease. Am J Respir Crit Care Med 2001;163:19 - 25. [7] Robinsky G, Miller M, Anlt k, et al. Spontaneous platelet activation and aggregation during OSA & its therapy with nasal CPAP. Chest 1995;108:625 - 30.

117

[8] Ranqemark C, Hedner JA, Carlson JT, et al. Platelet function and fibrolytic activity in hypertensive and normotensive sleep apnea patients. Sleep 1995;18:188 - 94. [9] Dean RT, Wilcox I. Possible atherogenic effects of hypoxia during OSA. Sleep 1996;16:S15 – S22. [10] Schulz R, Mahmoudi S, Hattar K, et al. Enhanced release of superoxide from polymorphonuclear neutrophils in OSA. Am J Respir Crit Care Med 2000;162:566 - 70. [11] Mooe T, Franklin KA, Holmstrom K, et al. Sleep Disordered Breathing & coronary artery disease: long term prognosis. Am J Resp Crit Care Med 2001;164:1910 - 3. [12] Peker Y, Hedner J, Kraiczi H, et al. Respiratory Disturbance Index: an independent predictor of mortality in coronary artery disease. Am J Respir Crit Care Med 2000;162:81 - 6. [13] Bradley TD. Right and left ventricular functional impairment and sleep apnea. Clin Chest Med 1992;13:459 - 79. [14] Puvanendran K, Goh KL. From snoring to sleep apnea in a Singapore population. Sleep Res Online 1999;2(1):11 - 4. [15] Tami TA, Duncan HJ, Pfleger M. Identification of OSA in patients who snore. Laryngoscope 1998;108(4 Pt 1):508 - 13. [16] Vaidya AM, Petruzzelli GJ, Walker RP, et al. Identifying OSA in patients presenting for LAUP. Laryngoscope 1996;106:431 - 7. [17] Osman EZ, Osborne J, Hill PD, et al. The Epworth’s Sleepiness Scale: can it be used for screening among snorers? Clin Otolaryngol 1999; 24(3):239 - 41. [18] Olson LG, King MT, Hensley MJ, et al. A community study of snoring and SDB. Symptoms. Am J Respir Crit Care Med 1995; 152(2):707 - 10. [19] Chung KF. Use of Epworth’s Sleepiness Scale in Chinese patients with OSA and normal hospital employees. J Psychosom Res 2000;49(5): 367 - 72. [20] Viner S, Szalai JP, Hoffstein V. Are the history and physical examination a good screening test of OSA? Ann Intern Med 1991; 115:356 - 9. [21] Maislin G, Pack AI, Kribbs NB. A survey screen for the prediction of apnea. Sleep 1995;18:158 - 66. [22] Netzer NC, Stoohs RA, Netzer CM, et al. Using the Berlin Questionnaire to identify patients at risk for the sleep apnea syndrome. Ann Intern Med 1999;131:488. [23] Kirby SD, Eng P, Danter W. Neural network prediction of OSA from clinical criteria. Chest 1999;116:409 - 15. [24] Dixon JB, Schachter LM, O’Brien PE. Predicting sleep apnea and excessive day sleepiness in the severely obese: indicators for polysomnogram. Chest 2003;123(4):1134 - 41. [25] Kushida CA, Efron B, Guilleminault C. A predictive morphometric model for OSAS. Ann Int Med 1997;127(8 Pt 1):581 - 7. [26] Tsai WH, Remmers JE, Brant R, et al. A decision rule for the diagnostic testing in OSA. Am J Respir Crit Care Med 2003;167: 1427 - 32. [27] Wang RC, Elkins TP, Keech D, et al. Accuracy of clinical evaluation in paediatric OSA. Otolaryngol Head Neck Surg 1998;118:69 - 73. [28] Suen JS, Arnold JE, Brooks LJ. Adenotonsillectomy for treatment of OSA in children. Arch Otolaryngol Head Neck Surg 1995;121: 525 - 30. [29] Goldstein NA, Sculerati N, Walsleben JA. Clinical diagnosis of pediatric OSA validated by polysomnogram. Otolaryngol Head Neck Surg 1994;111(5):611 - 7. [30] Leach J, Olson J, Hermann J, et al. Polysomnographic and clinical findings in children with OSA. Arch Otolaryngol Head Neck Surg 1992;118:741 - 4. [31] Harding SM. Prediction formulae for SDB. Curr Opin Pulm Med 2001;7(6):381 - 5. [32] Wiltshire N, Kendrick AH, Catterall JR. Home oximetry studies for the diagnosis of sleep apnea/hypopnea syndrome. Chest 2001;120: 384 - 9. [33] Shochat T, Hadas N, Kerkhofs M, et al. The SleepStrip: an apnea screener for the early detection of OSA. Eur Respir J 2002;19:121 - 6.

118

K.P. Pang, D.J. Terris / American Journal of Otolaryngology–Head and Neck Medicine and Surgery 27 (2006) 112 – 118

[34] Nixon GM, Kermack AS, Davis GM, et al. Planning adenotonsillectomy in children with OSA: the role of overnight oximetry. Paeds 2004;113(1 Pt 1):e19 – e25. [35] Kirk VG, Bohn SG, Flemons WW, et al. Comparison of home oximetry monitoring with laboratory polysomnography. Chest 2003; 124(5):1702 - 8. [36] Fischer Y, Junge-Hulsing B, Rettinger G, et al. The use of ambulatory, automatic sleep recording devise (QUISI version 1.0) in the evaluation of primary snoring and OSA. Clin Otolaryngol 2004;29(1):18 - 23. [37] Reichert JA, Bloch DA, Cundiff E, et al. Comparison of the NovaQ.S.G. Som, a new sleep apnea home-diagnostic system and polysomnography. Sleep Med 2003;4(3):213 - 8. [38] Ayas NT, Pittman S, MacDonald M, et al. Assessment of a wrist-worn device in the detection of OSA. Sleep Med 2003;4(5):435 - 42. [39] Schnall RP, Shlitner A, Sheffy J, et al. Periodic, profound peripheral vasoconstriction: a new marker of obstructive sleep apnea. Sleep 1999; 22:939 - 46. [40] Bar A, Pillar G, Dvir I, et al. Evaluation of a portable device based on peripheral arterial tone for unattended home sleep studies. Chest 2003;123:695 - 703. [41] Pittman S, Tal N, Pillar G, et al. Automatic detection of obstructive sleep-disordered breathing events using peripheral arterial tonometry and oximetry [abstract]. J Sleep Res 2000;9(Suppl):309. [42] O’Donnell CP, Allan L, Atkinson P, et al. The effect of upper airway obstruction and arousal on peripheral arterial tonometry in obstructive sleep apnea. Am J Respir Crit Care Med 2002;166:965 - 71.

[43] Mitani H, Ishiyama Y, Shirai Y. Novel apnomonitor for use at home. Rinsho Byori 2003;51(8):790 - 7. [44] Ficker JH, Wiest GH, Wilpert J, et al. Evaluation of a portable recording device (Somnocheck) for use in patients with suspected OSA. Respiration 2001;68(3):307 - 12. [45] Jaeho K, Yoo G, Choi S, et al. Comparison and analysis of SleepStrip and PSG in patients with OSA. Presented at the Spring Meeting of the Korean Society of Otolaryngology; 2001. [46] Lavie P, Herer P, Zomer J, et al. The SleepStrip. An aid for sleep apnea screening of large populations. Sleep Res 2000;9(Suppl 1):112. [47] Hollingworth L, Tooby M, Roberts D, et al. Practicality of the SleepStrip in postal screening for OSA. J Sleep Res 2003;12(2): 157 - 9. [48] Liesching TN. Evaluation of the accuracy of SNAP technology sleep sonography in detecting OSA in adults compared to standard PSG— clinical investigations. Chest 2004;125(3):886 - 91. [49] Michaelson PG, Allan PF, Chaney JC, et al. Validations of a portable home sleep study with 12 lead PSG: comparisons and insights into a variable gold standard. Presented at the American Academy of Otolarynology—Head and Neck Surgery Foundation Annual Meeting and OTO EXPO, Sept, 2004. [50] Su S, Baroody FM, Kohrman M, et al. A comparison of polysomnogram and a portable home sleep study in the diagnosis of OSA. Otolaryngol Head Neck Surg 2004;131:844 - 50. [51] Chesson AL, Berry RB, Pack A, et al. Practice parameters for the use of portable monitoring devices in the investigation of suspected OSA in adults. Sleep 2003;26(7):907 - 13.