ABSTRACT
Aim: to obtain the OSAS prevalence and risk factors of OSAS in obese early adolescents and to create a scoring system based on risk factors for diagnosing OSAS.
Methods: an observational study in Jakarta, November 2007 until December 2008 on obese adolescents aged 10- 12 years with snoring. Subjects underwent clinical examination, lung function test, paranasal sinus X-ray, and polisomnography. Measured outcomes were diagnosis of OSAS; sensitivity, specificity, predictive values, and likelihood ratios of a scoring system based on risk factors.
Results: the prevalence of OSAS in obese early adolescents is 38.2% using AHI cut-off point of >3 on PSG.
Tonsillar hypertrophy, adenoid hypertrophy, and neck circumference were the main risk factors. Scoring system was designed based on these results: OS= T + A + NC;
OS= OSAS score; T= tonsil hypertrophy (>T3 scored 1,
<T3 scored 0); A = adenoid hypertrophy (adenoid and nasopharynx ratio of >0.8 scored 1.5, <0.8 scored 0); NC
= neck circumference (> 34 cm scored 1; <34 cm scored 0). Children were most likely to have OSAS if they had a total score of 3.5. This scoring system has a sensitivity of 62% (95%CI 47 to 77%), specificity 100% (95%CI 100 to 100%), positive predictive value 100% (95%CI 100 to 100%), negative predictive value 81% (95%CI 73 to 89%), unlimited LR(+), LR(-) of 0.38 (CI 95% 0.6 to 0.56).
Conclusion: a scoring system based on tonsillar hypertrophy, adenoid hypertrophy, and neck circumfer- ence has sensitivity and specificity of 62% and 100% in diagnosing OSAS.
Key words: obesity, adolescent, obstructive sleep apnea syndrome, adenoid, tonsil, hypertrophy.
Risk Factors of Obstructive Sleep Apnea Syndrome in Obese Early Adolescents: A Prediction Model Using Scoring System
Bambang Supriyatno*, Mardjanis Said*, Bambang Hermani**, Damayanti R. Sjarif*, Sudigdo Sastroasmoro*
* Department of Child Health, Faculty of Medicine University of Indonesia - dr. Cipto Mangunkusumo Hospital.
Jl. Salemba 6, Jakarta Pusat 10430, Indonesia, **Department of ENT-HN, Faculty of Medicine University of Indonesia - Dr Cipto Mangunkusumo Hospital, Jakarta, Indonesia.
Correspondence mail to: [email protected].
INTRODUCTION
Obstructive sleep apnea syndrome (OSAS) is a complete or partial airway obstruction that causes significant physiologic disturbance with various clinical impacts. Apnea hypopnea index (AHI) is the frequency of apnea and/or hypopnea per hour which can be used as an indicator of OSAS severity.1,2 OSAS is suspected in children with snoring. Snoring is classified as occasional and habitual snoring. The prevalence of habitual snoring is around 3.2-12.1%, while that of occasional snoring is 28.1%, depending on the criteria used.2 Supriyatno et al. found that the incidence of snoring in children aged 5-13 was 31.6%.
In habitual snoring and occasional snoring, the incidence of OSAS was 26.4% and 5.2%, respectively.3 The prevalence of OSAS in children ranged from 0.3-2.4%,4 while in obese adolescences it ranged from 36-40%.5 Some studies reported adenoid and tonsillar hypertrophy, craniofacial disproportion, atopy, and obesity as risk factors for OSAS in children.6-9 Adenoid and tonsillar hypertrophy are the most common conditions that cause OSAS in children. In adolescent and adult population, obesity is the main risk factor of OSAS, but it has not been suggested in children under 5 years old.8,9 Neck circumference reflecting lipid accumulation in the neck which can cause upper airway narrowing is suggested to be more important than general obesity in the development of OSAS. Other hypothesis is that obese children have velopharynx which is easy to collapse and cause sleep breathing disorders.9
Definitive diagnosis of OSAS is established by polysomnography (PSG) during sleep as the gold standard examination.1,2,10 The presence of OSAS is
determined using various AHI cut off point, which are AHI >1, AHI >3, or AHI >5. The objectives of this study were to find out the prevalence and risk factors of OSAS in obese adolescents, and to create a scoring system for diagnosing OSAS based on risk factors.
METHODS
This was a cross sectional study carried out in Jakarta during November 2007 until December 2008.
Subjects were 10-12 year old children with obesity (BMI of >P95) and snoring symptoms. They were recruited consecutively from several elementary schools in Jakarta. We decided to include obese children above 10 years old since obesity at that age is likely to persist until adolescence or adulthood.
We performed weight and height measurement on all students and included those who fulfilled the inclusion criteria. We further classified subjects into obese (BMI of >P95) and superobese (BMI of >P95).
Children with craniofacial disproportion, adenotonsillectomy history and incomplete data were excluded. The sample size was 110 subjects.
Subjects fulfilling the inclusion criteria underwent comprehensive history taking and physical examina- tion to find out the general condition, degree of obesity, tonsil size, and neck circumference. Lung function test was done to assess lung function abnormality by measuring forced expiratory volume at 1 minute (FEV1) and peak expiratory flow rate (PEFR).
Obstructive disorder was defined if the FEV1 was less than 80% of Polgar standard value, restrictive disorder if the PEFR was less than 80% of standard value, and mixed disorder if both FEV1 and PEFR are less than <80%. Radiography examination was done to evaluate tonsillar-adenoid enlargement and sinusitis. Enlargement of adenoid was defined as the ratio of adenoid to nasopharynx of >0.8.
Tonsil size was assessed using the Brodsky criteria: T0 if tonsillectomy had been already done; T1 if the enlargement reached posterior pharyngeal arch but not over it; T2 if the enlargement was over posterior pharyngeal arch but did not reach uvula; T3 if the enlargement reached the uvula but not across mid- line; and T4 if both tonsils met each other (“kissing”).
Tonsil enlargement was defined if it was >T3.
The measurement of neck circumference was performed in the middle of the neck through cervical spine process and anterior middle neck (Adam’s apple in boys) using elastic Seca band and stated in centimeter. The measurement was done twice and averaged to get more precise value.
Thereafter, polysomnography with EmblettaÒ was performed to establish the diagnosis of OSAS. During polysomnography, the subjects’ sleeping activities, including snoring, effort of breathing, apnea episode and other activities were recorded. Polysomnography was performed while the subjects were sleeping. The polysomnography had several monitors: pulse oximetry, electroencephalography (EEG), breathing pattern (nose or mouth) and obstruction, and other subjects’
activities or movement. The results were evaluated the day after for the identification of breathing type, oxygen saturation, sleeping pattern (REM and NREM), apnea or hypopnea, and obstruction types. Apnea was defined as the stop of breathing for >10 seconds with oxygen saturation of <90%, while hypopnea was the stop of breathing for <5 seconds with oxygen satura- tion of <90%. Based on the data collected, the author determined whether the subjects had OSAS, defined if AHI >3. Radiograms were evaluated by a pediatric radiologist consultant using standard criteria, while the lung function test and polysomnography results were analyzed by the author.
The association between independent and dependent variables was analyzed with bivariate and multivariate analyses. Variables that had p values of
<0.25 in bivariate analysis were included to a model which was already matriculated for selective forward steps. We performed logistic regression analysis by Backward LR method and Hosmer Lemeshow goodness of fit, CI for exp (B), probability for stepwise entry 0.05-removal 0.05 which was then transform into a normogram. Multivariate analysis was done to find regression coefficient by controlling the independent variables resulting in a scoring system. Data were analyzed using SPSS 14.0.
This study has been approved by Ethics Committee Faculty of Medicine University of Indonesia and started after the permission had been obtained. Parental informed consents were obtained from every subject included.
RESULTS
Subject Characteristics
During the study period, 115 subjects fulfilled the inclusion criteria. Among them, only 110 could be analyzed because the other 5 subjects did not com- plete radiograph evaluation or polysomnography. The subjects consisted of 83 (75%) boys and 27 (25%) girls with the age range of 10-12 years old (mean age 10.5 years). The mean neck circumference was 34.6 cm.
The bivariate association of independent variables with OSAS is shown in Table 1. Habitual snoring, allergic rhinitis, tonsillar hypertrophy, adenoid hypertrophy, neck circumference and lung function test were significant risk factors for OSAS. Comorbid diseases such as asthma, allergic rhinitis and sinusitis were not found to be risk factors of OSAS in our subjects, however if all of those three conditions existed, the OR for developing OSAS was 6.776 (95%CI 1.742-26.324).
Risk Factors of OSAS Based on Multivariate Analysis On bivariate analysis, there were 7 variables (habitual snoring, allergic rhinitis, tonsillar hypertrophy, adenoid hypertrophy, neck circumference, asthma and lung function test) that had p values of <0.25. Those independent variables were included to a model which was already matriculated for selective forward steps.
After using logistic regression analysis by Backward LR method and Hosmer Lemeshow goodness of fit, CI for exp (B), probability for stepwise entry 0.05- removal 0.05, those variables were depicted in a normogram. After the normogram was made, a score
was addressed for each variable by dividing it with the lowest score to make the formula easier and simpler as shown in Table 2. Based on this model, area under curve calculated by ROC was 0.89 (95%CI 0.83-0.95).
The result of Hosmer-Lemeshow goodness of fit test was 0.81 with R2 of 0.81, supported that the model was quite good.
Based on Table 2 and cut-off point neck circum- ference of 34 cm, we stratified the risk of OSAS as follows: if total score was <2, then OSAS was unlikely,
Table 1. Bivariate association between several variables and OSAS (AHI >3) Predictors OSAS
Yes No P OR 95%CI
Age (years)
· Mean
· SD
10.40 (0.63)
10.54 (0.68)
0.283 1.398 0.759-2.574 Gender
· Male
· Female
30 12
53 15
0.441 0.708 0.293-1.708 Habitual snoring
· Yes
· No
34 8
10 58
0.000 24.650 8.875-68.465 Nutritional status
· Superobese
· Obese
34 8
58 10
0.550 0.733 0.264-2.035 History of asthma
· Yes
· No
15 27
17 51
0.229 1.667 0.722-3.847 Allergic rhinitis
· Yes
· No
23 19
23 45
0.037 2.368 1.077-5.210 Tonsillar hypertrophy
· Yes
· No
12 30
9 59
0.047 2.622 0.994-6.914 Adenoid hypertrophy
· Yes
· No
40 2
12 56
0.000 93.333 19.791-440.161 Sinusitis
· Yes
· No
26 16
37 31
0.440 1.361 0.621-2.984 Neck circumference
· Mean (cm)
· SD
36.89 (3.22)
33.23 (1.91)
0.000 2.280 1.659-3.135 Lung function test
· Abnormal
· Normal
31 11
33 35
0.009 2.989 1.295-6.858
Table 2. Predictor scores of OSAS
Predictor Criteria Score Simplified
score*
Hypertrophy 30 1.5
Adenoid
No hypertrophy 0 0
Hypertrophy (>T3) 20 1 Tonsil
No hypertrophy (<T2) 0 0
>34 20 1
Neck circumference (cm)
<34 0 0
Total score 70 3.5
Note: *Simplified score was original score divided by the lowest score (20), for example simplified adenoid score was 30/20 or 1.5.
if the score was between 2 and <3.5, then the subjects might or might not have OSAS, while if the score was 3.5, then OSAS was very likely.
Based on Table 3, the sensitivity was 26/42 (62%;
95%CI 47 to 77%), specificity 68/68 (100%; 95%CI 100 to 100%), positive predictive value (PPV) 26/26 (100%; 95%CI 100 to 100%), negative predictive value (NPV) 68/84 (81%; 95%CI 73 to 89%), unlimited LR(+), and LR (-) 0.38 (95%CI 0.26 – 0.56). Youden Index was 0.62 + 1-1 = 0.62, which meant good because more than 0 and near to 1.
was greater than girls, bivariate and multivariate analysis showed insignificant association between sex and OSAS. This may be because lipid accumulation in ado- lescents (10-12 years old) was not the same as adults.
Besides tonsillar enlargement, adenoid hypertrophy is also considered as the main risk factor OSAS in children. Adenoid hypertrophy, which is usually caused by inflammation, was suggested to be the main problem causing OSAS in children, while obesity played more roles in adolescent and adults. Children without obesity can have adenoid hypertrophy and suffer from OSAS which is improved after adenoidectomy.15-17
Neck circumference is an indicator of OSAS in adults. Neck circumference has a linear correlation with lipid accumulation surrounding the neck.14,15,18,19
It also had been used as an indicator for adult obesity rather than BMI. Unfortunately, few studies reported the cut off point of neck circumference for indicating obesity. Katz et al stated that male adults with neck circumference of more than 43.7 cm were more likely to have OSAS, while for female, 39.6 cm was proposed as the cut-off point.19 There had been no study reporting neck circumference as an indicator in childhood OSAS. This study found the mean of neck circumference or children with OSAS was 34.6 cm.
Many literatures combined tonsillar and adenoid hypertrophy as the main cause of OSAS.9,12,20-23 Based on that, the author used the term adenotonsillar hypertrophy. Katz et al.5 found that the inspiration effort was greater in children with OSAS compared to that of children without. Further investigations found that adenotonsillar hypertrophy played its role in OSAS by increasing the airway resistance detected by anterior rhinometry. Children with adenotonsillar hypertrophy showed airway narrowing during inspira- tion and the size of hypertrophy was correlated with the degree of narrowing.1,2,5
Arens et al.23 studied the correlation between neck lipid accumulation, tonsillar hypertrophy and OSAS in younger and older children. They reported that adenotonsillar hypertrophy had more important role in younger children because the adenoid had regressed in older ones.23 This study also done on teenagers still found that adenotonsillar hypertrophy was the main risk factor of OSAS, which might be influenced by other factors such as asthma and allergic rhinitis.
Neck circumference had been considered to have strong correlation with OSAS in adults, although not all experts agreed.19 There had been no study investi- gating the association between neck circumference and OSAS in children. This study found that neck circum- ference was the important indicator for OSAS in DISCUSSION
The prevalence of OSAS in this study was 38.2%, which was consistent with the previous study which found the prevalence of 36-60%.5 The difference of the result might due to the different definition of OSAS.
Using low AHI as the cut off point, the prevalence would be higher compared to that defined with higher AHI and this rule applied to general population of OSAS study (not only for obese children population).4 Establishing OSAS in adults is based on AHI of
>5, but considering that we should be more aggressive in children, some experts used AHI of >3 or >1 as the cut-off point for children.4 Another reason for early diagnosis of OSAS in children is to prevent the complication of delayed diagnosis and avoid unnecessary treatment such as tonsilloadenoidectomy.10 This study used AHI of >3 as the cut off point and found the OSAS prevalence of 38.2%. This supported recent studies which reported that 36-60% obese and snoring children will have OSAS.5 The more frequent the snoring (habitual snoring) and the higher the BMI, the greater possibility for having OSAS would be.5,9,11 In adults, gender is the main risk factor of OSAS with the ratio of male to female of 3-8 to 1.2,10,12,13 It is assumed that in obese male, the lipid is predominantly distributed in the chest and neck, while in female, it is accumulated in the waist. The lipid accumulation in the neck can result in increased pressure of upper airways. Based on this theory, many studies recently reported that OSAS is more likely found in male than female.1,2,14,15 In this study, although the number of boys
Table 3. Diagnostic value of OSAS scoring system compared to PSG as the gold standard
OSAS
Yes No Total
3.5 26 0 26
Total
score <3.5 16 68 84
Total 42 68 110
children with OR 2.28. This finding supported that lipid accumulation, by its mechanical mechanism, could serve as an indicator for OSAS. Neck circumference measurement is a simple screening test for obese and snoring children who are suspected of having OSAS.
More accurate lipid thickness measurements can be done by MRI.23
In adults, men with OSAS have greater neck circumference compared to that of women.24,25 Rowley et al25 suggested that upper airway resistance in men with OSAS was higher than that in women. Although the mechanism was not so clear, it was proposed that the longer circumference, the higher airway resistance would be. Increased circumference is assumed to cause lateral wall thickening, which is an important factor to increase airway resistance. Other possibility is that the wall thickening will cause increased pressure surround- ing the airway that leads to increased airway resistance.25 This study found that the mean neck circumference in OSAS children was 36.89 cm (SD 3.22), while in non-OSAS was 33.23 cm (SD 1.91).
Scoring System Usage Application
The scoring system has a good sensitivity, specificity, PPV, NPV, LR (+), and LR (-). The main risk factors for OSAS in children are tonsillar hypertrophy, adenoid hypertrophy, and neck circumference. Adenoid hypertrophy is usually accompanied by tonsillar hypertrophy so that it can be combined as a risk factor of OSAS.20,22
Compared to the Broullette formula,26 the scoring system proposed by the author was more applicable.
In Broullette formula, the important components were difficulty of breathing, snoring and apnea. Parents are usually difficult to detect apnea (a stop of breathing of
>10 seconds) when their children are sleeping, although they sleep together. Besides apnea, component of breathing difficulty was scored by 1 if it was infre- quent, 2 if frequent, and 3 if always, which was not easy to be done and very subjective.12 Unfortunately, the results of this study could not be applied to all children of all ages such as the Broullette formula.
CONCLUSION
The prevalence of OSAS in obese adolescents is 38.2% using AHI cut-off point of >3. Tonsillar hypertrophy, adenoid hypertrophy, and neck circum- ference are the main risk factors of OSAS in obese adolescents. We proposed a scoring system for obese early adolescent with snoring based on tonsillar hypertrophy, adenoid hypertrophy, and neck circum- ference. This scoring system has sensitivity of 62%
(95%CI 47 to 77%), specificity 100% (95%CI 100 to 100%), positive predictive value 100% (95%CI 100 to 100%), negative predictive value 81% (95%CI 73 to 89%), unlimited LR(+), LR(-) 0.38 (95%CI 0.6 to 0.56), and Youden Index is 0.62.
REFERENCES
1. Marcus CL, Carroll JL. Obstructive sleep apnea syndrome.
In: Loughlin GM, Eiger H, eds. Respiratory disease in children: diagnosis and management. Baltimore: William &
Wilkins; 1994. p. 475-91.
2. Schechter MS, Technical report: diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics.
2002;109:1-20.
3. Supriyatno B, Deviani R, Tumbelaka A, Kariani EBK, Rahajoe NN. Characteristics and risk factors of snoring and the prevalence of suspected obstructive sleep apnea in children. Pediatr Indones. 2005;45:40-5.
4. Lumeng JC, Chervin RD. Epidemiology of pediatric obstructive sleep apnea. Proc Am Thorac Soc. 2008;5:242- 52.
5. Katz ES, D’Ambrosio CM. pathophysiology of pediatric obstructive sleep apnea. Proc Am Thorac Soc. 2008;5:253- 62.
6. Chng SY, Thiam DY, Wang XS, Tan TN, Ong NBH. Snoring and atopic disease: a strong association. Pediatr Pulmonol.
2004;38:210-6.
7. Deegan MN. Clinical prediction rules in obstructive sleep apnea syndrome. Eur Respir J. 1997;10:1194-5.
8. Ramar K, Guilleminault C. Risk factors. In: Kushida CA, ed.
Obstructive sleep apnea. pathophysiology, comorbidities, and consequences. New York: Informa healthcare; 2007. p.197- 222.
9. Redline S, Tishler PV, Schluchter M, Aylor J, Clark K, Graham G. Risk factors for sleep-disordered breathing in children. Association with obesity, race, and respiratory problems. Am J Respir Crit Care Med. 1999;159:1527-32.
10. American Thoracic Society. Standards and indications for cardiopulmonary sleep studies in children. Am J Respir Crit Care Med. 1996;153:866-78.
11. Ng DK, Lam YY, Kwok KL, Chow PY. Obstructive sleep apnea syndrome and obesity in children. Hongkong Med J.
2004;10:44-8.
12. American Academy of Pediatrics, Section on Pediatric Pulmonology, Subcommittee on Obstructive Sleep Apnea Syndrome. Clinical practice guideline: diagnosis and management of childhood obstructive sleep apnea syndrome.
Pediatrics. 2002;109:701-12.
13. Books LJ. Obstructive sleep apnea syndrome in infants and children: clinical features and pathophysiology. In: Sheldon SH, Ferber R, Kryger MH, eds. Principles and practice of pediatric sleep medicine. USA: Elsevier Saunders; 2005. p.
223-9.
14. Noun BL, Sohar E, Laor A. Neck circumference as a simple screening measure for identifying overweight and obese patients. Obes Res. 2001;9:470-7.
15. Nixon GM, Kermack AS, McGegor CD, Davis GM, Manoukian JJ, Brown, et al. Sleep and breathing on the first night after adenotonsillectomy for obstructive sleep apnea.
Pediatr Pulmonol. 2005;39:332-8.
16. Suen JS, Arnold JE, Brooks LJ. Adenotonsillectomy for treatment of obstructive sleep apnea in children. Arch Otolaryngol Head Neck Surg. 1995;121:525-30.
17. Criscouli G, D’Amora S, Ripa G, Cinquegrana G, Mansi N, Impagliazzo N, et al. Frequency of surgery among children who have adenotonsillar hypertrophy and improve after treatment with nasal beclomethasone. Pediatrics. 2003;111:236- 8.
18. Noun BL, Laor A. Relationship of neck circumference to cardiovascular risk factors. Obes Res. 2003;11:226-31.
19. Katz I, Stardling J, Slutsky AS, Zamel N, Hoffstein V. Do patients with obstructive sleep apnea have thick necks? Am Rev Respir Dis. 1990;141:1228-31.
20. Ng KD, Chow PY, Chan CH, Kwok KL, Cheung JM, Kong FY. An update on childhood snoring. Acta Pediatr.
2006;95:1029-35.
21. Anuntaseree W, Kuasirikul S, Suntornlohanakul S. Natural history of snoring and obstructive sleep apnea in Thai school- age children. Pediatr Pulmonol. 2005;39:415-20.
22. Anuntasere W, Rookkapan K, Kuasirikul S, Thongsuksai P.
Snoring and obstructive sleep apnea in Thai school-age children: prevalence and predisposing factors. Pediatr Pulmonol. 2001;32:222-7.
23. Arens R, McDough JM, Costarino AT, Mahboubi S, Tayag-Kier CE, Maislin G, et al. Magnetic Resonance Imaging of the upper airway structure of children with obstructive sleep apnea syndrome. Am J Respir Crit Care Med. 2001;164:698-703.
24. Dancey DR, Hanly PJ, Soong C, Lee B, Shepard J, Hoffstein V. Gender differences in sleep apnea. The role of neck circumference. Chest. 2003;123:1544-50.
25. Rowley JA, Sanders CS, Zahn BR, Badr MS. Gender differences in upper airway compliance during NREM sleep:
role of neck circumference. J Appl Physiol. 2002;92:2535- 41.
26. Brouillette R, Hanson D, David R. A diagnostic approach to suspected obstructive sleep apnea in children. J Pediatr.
1984;105:10.
