ORIGINAL ARTICLE
Pre-operative parameters do not reliably identify post-operative respiratory risk in children undergoing adenotonsillectomy
Surendran Thavagnanam ,
1,2Saou Y Cheong,
3Karuthan Chinna,
4Anna M Nathan
1,2and Jessie A de Bruyne
1,21Department of Paediatrics,2University Malaya Paediatric and Child Health Research Group,3Faculty of Medicine and4Department of Social and Preventive Medicine, University of Malaya, Kuala Lumpur, Malaysia
Aim:Adenotonsillectomy is performed in children with recurrent tonsillitis or obstructive sleep apnoea. Children at risk of post-operative respira- tory complications are recommended to be monitored in paediatric intensive care unit (PICU). The aim of the study is to review the risk factors for post-operative complications and admissions to PICU.
Methods:A review of medical records of children who underwent adenotonsillectomy between January 2011 and December 2014 was per- formed. Association between demographic variables and post-operative complications were examined using chi-square and Mann–Whitney tests.
Results:A total of 214 children were identified, and of these, 19 (8.8%) experienced post-operative complications. Six children (2.8%) had respira- tory complications: hypoxaemia in four and laryngospasm requiring reintubation in a further two. Both of the latter patients were extubated upon arrival to PICU and required no escalation of therapy. A total of 13 (6.1%) children had non-respiratory complications: 8 (3.7%) had infection and 5 (2.3%) had haemorrhage. A total of 26 (12.1%) children were electively admitted to PICU and mean stay was 19.5 (SD13) h. No association between demographic characteristics, comorbid conditions or polysomnographic parameters and post-operative complications were noted. A total of 194 (90.7%) children stayed only one night in hospital (median 1 day, range 1–5 days).
Conclusion:The previously identified risk factors and criteria for PICU admission need revision, and new recommendations are necessary.
Key words: adenotonsillectomy; complication; obstructive sleep apnoea; post-operative.
What is already known on this topic
1 In children undergoing adenotonsillectomy, the rate of respira- tory complications ranges from 11% to 20%.
2 The most important predictors of post-surgical respiratory mor- bidity were young age, obesity and the initial severity of obstruc- tive sleep apnoea syndrome (OSAS).
3 As such, it is recommended that these children would need to be observed post-operatively in paediatric intensive care unit (PICU).
What this paper adds
1 No association between demographic characteristics, comorbid conditions or polysomnographic parameters and post-operative complications were noted.
2 The previously identified risk factors and criteria for PICU admis- sion need revision and new recommendations are necessary.
Adenoidectomy and/or adenotonsillectomy (AT) remain the most common surgical option for children with recurrent tonsillitis1and adenotonsillar hypertrophy causing obstructive sleep apnoea (OSA).2,3Previous studies have identified potential post-operative respiratory complications4–8in children following AT and a num- ber of risk factors have been identified such as age below 3 years, obesity, severe OSA (apnoea hypopnoea index (AHI) > 10 events/
h), genetic syndromes, craniofacial abnormalities and neuromus- cular diseases.4,9–12 These risk factors were identified from a
heterogeneous populations2and may not be supported by poly- somnography (PSG). Also varied definitions of post-operative com- plications were used, which may have overestimated the true number of complications. All of these broad classifications may not necessary be applicable these days since there have been improve- ment in medical and surgical techniques for AT.
In view of the possible post-operative complications, centres have adopted a blanket policy to electively admit the at-risked children to paediatric intensive care unit (PICU) following AT for close monitoring. Obtaining a bed in PICU is often difficult; there- fore, when a policy of routine admissions are in place, some of the AT cases are often cancelled when no PICU bed is available.
This lengthens the wait time for surgery and may indirectly worsen the child’s comorbidities. Recently, Theilhaber et al.
recommended routine post-operative ICU admission is not rou- tinely needed but only in the presence of early adverse events (AEs) in recovery period.13
Correspondence:Associate Professor Surendran Thavagnanam, Depart- ment of Paediatrics, University Malaya Medical Centre, Lembah Pantai, Kuala Lumpur 50603, Malaysia. Fax: +60 37949 4704; email:
[email protected] Conflict of interest: None declared.
Accepted for publication 16 September 2017.
formed to determine AHI as this determines the severity of OSA and the associated oxyhaemoglobin saturation nadir. Therefore, the AHI is thought to be an indicator of post-operative complications14–16; however, no consensus exists as to how sensi- tive these indicators really are. Recently, in the CHAT study, which evaluated the efficacy of early AT with respect to neurocognitive outcomes post-operatively, it was observed that the PSGs data or demographic characteristics, such as obesity, gender and race, were not associated as risk factors for post-AT complications.17
In this study, we looked at the casefiles of all children who had underwent AT at our tertiary centre to determine if we could identify certain risk factors for post-operative complications and looked at the outcome of children admitted to PICU post- operatively.
Methods
We performed a retrospective chart review of children between age 2 and 18 years who underwent adenoidectomy and/or AT for recurrent tonsillitis or clinically suspected OSA at University Malaya Medical Centre, between January 2010 and December 2014. The University Malaya Medical Centre Medical Ethics Committee approved the project (MEC Ref No: 996.7). Clinical parameters thought to be associated with post-operative compli- cations include age, gender, ethnicity, body mass index (BMI) and other comorbidities, which include craniofacial abnormali- ties, identifiable syndromes, cardiac abnormalities, including pul- monary hypertension and tonsillar size as per Brodsky, were analysed.18A total of 87 (41%) children had a PSG performed as part of their assessment for OSA. PSG parameters evaluated in this study included the AHI and SpO2nadir. A total of 40 (19%) children had AT based on clinical suspicion of OSA by our otolar- yngologists without a PSG. All cases were managed by surgeons and anaesthetist with a large paediatric experience at our centre.
The surgical methods for AT performed included cold dissection or diathermy. Anaesthesia used included a combination of an opi- oid, muscle relaxant and analgesia titrated to the patients’body weight.
PSGs were performed using a commercially available PSG sys- tem (Cadwell Kennewick, WA, USA/Compumedics, Victoria, Aus- tralia). Recordings were scored according to standard American Academy of Sleep Medicine.19OSA severity was classified on the basis of the obstructive AHI (OAHI) as normal (OAHI≤1 event/h);
mild OSA (OAHI between 1 and <5 events/h); moderate OSA (OAHI between >5 and <10 events/h) or severe OSA (OAHI≥10 events/h).
We also looked at the PICU registry during the study period to determine the number of patients that were admitted to PICU and noted if any further management were needed such as sup- plemental oxygen or non-invasive or invasive ventilation ther- apy. Respiratory AEs were defined as follows: a mild AE was defined as one of the following: oxygen saturation (SpO2) < 90%
treated by administration of supplemental O2or repositioning of the patient to improve the airway. A severe AE was defined as laryngospasm; bronchospasm treated by continuous positive air- way pressure or bi-level non-invasive ventilation, placement of oropharyngeal airway, re-intubation or unplanned admission to ICU. All ATs were performed using standard surgical techniques.
4-weeks post-adenoidectomy, the medical charts were reviewed, specifically to look for presence of haemorrhage; poor oral intake requiring intravenous fluid administration; presence of fever to indicate infection; and planned or unplanned admissions.
Statistical analysis
Statistical analyses were performed using IBM SPSS statistics for windows, version 22.0 (IBM Corp, Armonk, NY, USA). Continu- ous variables were presented as means and standard deviations, while categorical variables were presented as frequencies and percentages. Thet-tests and chi-square tests were used to test the associations between polysomnographic parameters and demo- graphic characteristics with complications. APvalue <0.05 was considered as statistically significant.
Results
A total of 214 children underwent adenoidectomy and/or AT during our study period. The mean (SD) age at surgery was 9.2 (3.9) years. The male to female ratio was 1.3:1 and 50% of the children were either overweight or obese. Of the 214 patients, nineteen (8.9%) had post-operative complications; six (2.8%) with respiratory complications and thirteen (6.1%) with non- respiratory complications (infections in eight (3.7%) and bleeding infive (2.3%) children) (Fig. 1).
There were no statistically significant associations between demographic variables (gender, race or obesity) and the presence of either respiratory or non-respiratory complications. As tonsillar hypertrophy was a predominant feature in most of these chil- dren, the size of tonsils was not associated with post-operative complications (P= 0.21) (Tables 1,2). In this study, there were 39 (18%) asthmatic patients who were stable on their preventa- tive therapies and asthmaper saywas not a risk for post-operative complications (P= 0.74). 24 children had other co-morbidities (Table 3) and of these, seven (29%) were electively admitted to PICU post-operatively however, none developed any complica- tions that required further interventions. They all were dis- charged home within 2 days of the surgery.
During the study period, 26 children were electively admitted to PICU and the mean stay in PICU was 19.5 (SD12.9) hours.
Of the six children with respiratory complications, three were admitted to PICU and the other three were observed in the ward.
Of the three children admitted to PICU, two required re- intubation in the post-operative period due to laryngospasm.
Both children were extubated within 2.5 hours upon arrival to PICU. The third child only required supplemental oxygen via nasal prongs due to hypoxemia. All the other 3 children in the ward required supplemental oxygen via nasal prongs only for less than 24 hours due to hypoxemia. No differences were noted in the demographic data of patients with respiratory complications who were admitted to PICU and the ward (Table 4).
Thirteen children had late post-operative complications (within 4 weeks post-AT) requiring further intervention. A total of 5 (2.3%) had tonsillar bleeding requiring re-exploration and eight (3.7%) had tonsillar infection requiring intravenous antibi- otics. Thirteen (6%) children had their planned surgery post- poned on one or more occasions, as they were to be admitted to
PICU post-operatively for observations but there were no back- up ICU beds available at the time of planned surgery.
Discussion
In our study, we found a low incidence of post-operative compli- cations in children following AT irrespective of obesity or PSG parameter. The complication rates in this study were low at 8.9%
and relatively similar to other studies.13 This could be due to newer surgical techniques used for AT, which helped to reduce the rates of primary haemorrhage and shorten recovery time and improved anaesthesia techniques used peri- and post-opera- tively.20,21 Of the complications, 2.8% were respiratory related and the others included bleeding and/or infection, which were 3.7% and 2.3%, respectively. These are comparable with other published studies.17,22
Obesity has been shown to be a significant risk factor for post- operative complications in children post-AT.23–26 Interestingly, we did notfind a similar association in our study.27One possible explanation for this is most of our study children had a BMI less than 30 kg/m2plus they did not have any other comorbidities. In morbidly obese children, their resting functional residual capacity is lower, and due to their increased workload, there is increased oxygen consumption and carbon-dioxide production.28,29These alterations in respiratory function and microatelectasis of lung bases during surgery may worsen the ventilation-perfusion n = 214 children
No complications (n = 195)
Respiratory complications
(n = 6)
Reintubated n = 2
Supplemental O2 n = 4
Non-respiratory complications
(n = 13)
Bleeding n = 5
Infection
n = 8 Fig. 1 Details of post-operative compli- cations.
Table 1 Demographic data of the study population based on the presence or absence of post-operative complications
No complications,
n= 195
Complications, n= 19
P value*
Age (years), meanSD 9.13.88 9.94.43 0.39 Gender,n(%)
Male 109 (92.4) 9 (7.6) 0.48
Female 86 (89.6) 10 (10.4)
Ethnicity,n(%)
Malay 146 (89.6) 17 (10.4) 0.30
Chinese 15 (100) 0 (0)
Indian 34 (94.4) 2 (5.6)
Body mass index (kg/m2), meanSD
20.76.52 229.33 0.43 Obesity,n(%)
Obese 76 (91.6) 7 (8.4) 0.67
Overweight 23 (85.2) 4 (14.8)
Underweight 29 (90.6) 3 (9.4)
Normal 67 (93.1) 5 (6.9)
Tonsillar size, meanSD 30.61 2.80.69 0.21
*P< 0.05 is considered significant. SD, standard deviation.
Table 2 Polysomnographic data of the study population based on the presence or absence of post-operative complications
No complications, n= 76
Complications, n= 11
P value*
AHI,n(%)
Normal 2 (32.9) 0 0.75
Mild 27 (31.0) 5 (45.5)
Moderate 14 (18.4) 3 (27.3)
Severe 33 (37.9) 3 (27.3)
AHI (events/h), meanSD
21.128.3 9.4210.3 0.20
SpO2nadir (%), meanSD
81.916.82 80.121.39 0.88
*P< 0.05 is considered significant. AHI, apnoea hypopnoea index; SD, standard deviation.
Table 3 Diagnosis of the study population and admissions to paediatric intensive care unit (PICU) and ward
Diagnosis Ward
admission (n= 52),n
PICU admission (n= 11),n
ADHD 1 ND
Renal 2 ND
Cardiac 1 1
Down 4 2
Neurology/Neurodevelopmental disorders 6 1
Skeletal dysplasia 1 1
Prader Willi Syndrome 2 1
Mucopolysaccharide syndrome ND 1
Asthma 35 4
ADHD, attention deficit hyperactivity disorder; ND, no data.
mismatch peri-operatively, resulting in respiratory complications post-operatively. With regards to the admission to PICU post- operatively, this decision can be dependent on the availability of services, such as high-dependency unit or acute care in the gen- eral wards.
Younger children below the age of 3 were previously shown to be at risk for post-operative respiratory complications,5,10thereby requiring routine admission to PICU post-operatively. Recently, Spencer et al.showed that this age group actually had a lower complication rate,30 especially in the children without other comorbidities and severe OSA. In our study, there werefive chil- dren below the age of 3 years and only one child developed lar- yngospasm post-operatively and required an unplanned admission to PICU. The other children were observed on the gen- eral surgical ward and had no further complications. We are unable to conclude with regards to this matter as our study only encompassed of a small number of patients in this age group.
In our centre, the site of post-operative care for high-risked children following AT is frequently discussed. Current consensus suggests these children should be admitted to PICU for close car- diorespiratory observation; however, guidance regarding the spe- cific levels of care is lacking. The British Working Party on Sleep Physiology recommends that high-risk children should have sur- geries performed in tertiary centres with PICU facilities avail- able.31In our study, 26 (12%) children were electively admitted to PICU, of which three required respiratory support. Two of these children who were reintubated immediately post- operatively due to laryngospasm and they were successfully extu- bated within 2.5 h of PICU admission. The third child was weaned off oxygen supplementation within 12 h of his PICU stay. A total of 6 (23%) children were transferred out of PICU within 8 h of admission post-operatively as PICU beds were required. A total of 13 (6%) children had their surgery postponed in this study, as there was no available PICU bed on the day of surgery. This led to increased anxiety and disappointment for patient and families and increased pressure on waiting lists. A total of 38% of these children had severe OSA with a mean AHI:
31.7 (SD14.7) per hour and mean SpO2 nadir: 76%
(SD7.7). There were no associations between ICU admission and post-operative complications identified in this study. There- fore, we recommend that criteria for routine admission to PICU
environment equipped with personnel who can manage paediat- ric emergencies and initiate respiratory support, such as oxygen and non-invasive ventilation, is an alternative in hospitals that do not have a PICU.
There were no statistically significant associations between PSG parameters and complications post-operatively in this study.
This observation is in contrast tofindings from other studies that showed that the AHI and the SpO2nadir were correlated with the presence of respiratory post-operative complications.14,15,32 PSGs are tedious, expensive, have a long wait list and may not be available in all centres. Therefore, one can dispute on the routine request for pre-operative PSG in the evaluation of OSA in view of our low post-operative complication rate and morbidity as seen in this study.33If there are concerns with regards to OSA, clini- cians can opt for the use of pulse overnight oximetry, which is now universally used as a screening tool for OSA severity.6This indirectly shortens diagnostic and treatment process for those with more severe disease, and aids clinicians to prioritise for ade- notonsillectomy and planning peri-operative care.34
Our findings should be interpreted in light of several limita- tions. This is a retrospective study with all the inherent limita- tions of design such as small sample size and 40% patients had a sleep study. Despite this, the study did look at the various risk factors as per American Academy of Pediatrics2and the American Academy of Otolaryngology–Head and Neck Surgery3and found no identifiable predictors for post-operative complications. When we analysed comorbidities such as asthma and other diseases with post-operative complications, no associations were found, which is in contrast to previous publications.11,35
Conclusion
We did not identify predictors of need for PICU in our cohort, though we advise caution regarding the young (under 3 years) and the obese, since other studies have demonstrated them to be predictive. Our study did not have sufficient numbers of either group for us to be able to support or challenge these statements.
Further research to identify patients at greatest risk of post- operative complications and new consensus and clinical pathways are warranted in the management of children undergoing AT.
Acknowledgement
This study met with our institutional ethical standards. No fund- ing was obtained for this study.
References
1 Mitchell RB, Pereira KD, Friedman NR. Sleep-disordered breathing in children: Survey of current practice.Laryngoscope2006;116: 956–8.
2 Marcus CL, Brooks LJ, Draper KAet al.Diagnosis and management of childhood obstructive sleep apnea syndrome.Pediatrics2012;130:
576–84.
3 Baugh RF, Archer SM, Mitchell RBet al.Clinical practice guideline:
Tonsillectomy in children.Otolaryngol. Head Neck Surg. 2011;144 (1 Suppl.): S1–30.
Table 4 Demographic data of the study population based on the presence of post-operative respiratory complications
Ward admission
(n= 3)
PICU admission
(n= 3) P value*
Age (years), meanSD 15.73.90 8.85.37 0.15 BMI (kg/m2), meanSD 37.75.26 25.28.02 0.09 AHI (SD), events/h 6.633.30 18.6714.51 0.23 SpO2nadir (%), meanSD 868.19 61.6731.02 0.26
Tonsils size 30 2.671.12 0.64
AsthmaAtopy (n) 1 2
*P< 0.05 is considered significant. AHI, apnoea hypopnoea index;
BMI, body mass index; PICU, paediatric intensive care unit; SD, stan- dard deviation.
4 Richmond KH, Wetmore RF, Baranak CC. Postoperative complications following tonsillectomy and adenoidectomy: Who is at risk? Int.
J. Pediatr. Otorhinolaryngol.1987;13: 117–24.
5 McColley SA, April MM, Carroll JL, Naclerio RM, Loughlin GM. Respira- tory compromise after adenotonsillectomy in children with obstructive sleep apnea.Arch. Otolaryngol. Head Neck Surg.1992;118: 940–3.
6 Nixon GM, Kermack AS, Davis GM, Manoukian JJ, Brown KA, Brouillette RT. Planning adenotonsillectomy in children with obstruc- tive sleep apnea: The role of overnight oximetry.Pediatrics2004;113 (1 Pt 1): e19–25.
7 Brown KA, Morin I, Hickey C, Manoukian JJ, Nixon GM, Brouillette RT.
Urgent adenotonsillectomy: An analysis of risk factors associated with postoperative respiratory morbidity. Anesthesiology 2003; 99:
586–95.
8 Rosen GM, Muckle RP, Mahowald MW, Goding GS, Ullevig C. Postoper- ative respiratory compromise in children with obstructive sleep apnea syndrome: Can it be anticipated?Pediatrics1994;93: 784–8.
9 Biavati MJ, Manning SC, Phillips DL. Predictive factors for respiratory complications after tonsillectomy and adenoidectomy in children.
Arch. Otolaryngol. Head Neck Surg.1997;123: 517–21.
10 Wiatrak BJ, Myer CM 3rd, Andrews TM. Complications of adenotonsil- lectomy in children under 3 years of age.Am. J. Otolaryngol.1991;
12: 170–2.
11 Goldman JL, Baugh RF, Davies Let al.Mortality and major morbidity after tonsillectomy: Etiologic factors and strategies for prevention.
Laryngoscope2013;123: 2544–53.
12 Gerber ME, O’Connor DM, Adler E, Myer CM 3rd. Selected risk factors in pediatric adenotonsillectomy.Arch. Otolaryngol. Head Neck Surg.
1996;122: 811–4.
13 Theilhaber M, Arachchi S, Armstrong DS, Davey MJ, Nixon GM. Rou- tine post-operative intensive care is not necessary for children with obstructive sleep apnea at high risk after adenotonsillectomy. Int.
J. Pediatr. Otorhinolaryngol.2014;78: 744–7.
14 Hill CA, Litvak A, Canapari Cet al.A pilot study to identify pre- and peri-operative risk factors for airway complications following adeno- tonsillectomy for treatment of severe pediatric OSA. Int. J. Pediatr.
Otorhinolaryngol.2011;75: 1385–90.
15 Jaryszak EM, Shah RK, Vanison CC, Lander L, Choi SS. Polysomno- graphic variables predictive of adverse respiratory events after pedi- atric adenotonsillectomy.Arch. Otolaryngol. Head Neck Surg.2011;
137: 15–8.
16 Shine NP, Coates HL, Lannigan FJ, Duncan AW. Adenotonsillar surgery in morbidly obese children: Routine elective admission of all patients to the intensive care unit is unnecessary. Anaesth. Intensive Care 2006;34: 724–30.
17 Konstantinopoulou S, Gallagher P, Elden Let al.Complications of ade- notonsillectomy for obstructive sleep apnea in school-aged children.
Int. J. Pediatr. Otorhinolaryngol.2015;79: 240–5.
18 Brodsky L, Moore L, Stanievich JF. A comparison of tonsillar size and oropharyngeal dimensions in children with obstructive adenotonsillar hypertrophy.Int. J. Pediatr. Otorhinolaryngol.1987;13: 149–56.
19 Standards and indications for cardiopulmonary sleep studies in chil- dren. American Thoracic Society.Am. J. Respir. Crit. Care Med.1996;
153: 866–78.
20 Hampson-Evans D, Morgan P, Farrar M. Pediatric laryngospasm.Pae- diatr. Anaesth.2008;18: 303–7.
21 Batra YK, Ivanova M, Ali SS, Shamsah M, Al Qattan AR, Belani KG.
The efficacy of a subhypnotic dose of propofol in preventing laryn- gospasm following tonsillectomy and adenoidectomy in children.Pae- diatr. Anaesth.2005;15: 1094–7.
22 Spektor Z, Saint-Victor S, Kay DJ, Mandell DL. Risk factors for pediatric post-tonsillectomy hemorrhage. Int. J. Pediatr. Otorhinolaryngol.
2016;84: 151–5.
23 Arachchi S, Armstrong DS, Roberts N et al.Clinical outcomes in a high nursing ratio ward setting for children with obstructive sleep apnea at high risk after adenotonsillectomy.Int. J. Pediatr. Otorhino- laryngol.2016;82: 54–7.
24 Kshirsagar R, Mahboubi H, Moriyama D, Ajose-Popoola O, Pham NS, Ahuja GS. Increased immediate postoperative hemorrhage in older and obese children after outpatient tonsillectomy.Int. J. Pediatr. Otor- hinolaryngol.2016;84: 119–23.
25 Austin AL, Kon A, Matteucci MJ. Respiratory failure in a child due to type 2 postobstructive pulmonary edema.Pediatr. Emerg. Care2016;
32: 23–4.
26 Gleich SJ, Olson MD, Sprung J et al. Perioperative outcomes of severely obese children undergoing tonsillectomy.Paediatr. Anaesth.
2012;22: 1171–8.
27 Kasle D, Virbalas J, Bent JP, Cheng J. Tonsillectomies and respiratory complications in children: A look at pre-op polysomnography risk fac- tors and post-op admissions.Int. J. Pediatr. Otorhinolaryngol.2016;
88: 224–7.
28 Chlif M, Keochkerian D, Choquet D, Vaidie A, Ahmaidi S. Effects of obesity on breathing pattern, ventilatory neural drive and mechanics.
Respir. Physiol. Neurobiol.2009;168: 198–202.
29 Eichenberger A, Proietti S, Wicky Set al.Morbid obesity and postop- erative pulmonary atelectasis: An underestimated problem.Anesth.
Analg.2002;95: 1788–92.
30 Spencer DJ, Jones JE. Complications of adenotonsillectomy in patients younger than 3 years.Arch. Otolaryngol. Head Neck Surg.2012;138:
335–9.
31 Royal College of Paediatric and Child Health, Working Party on Sleep Physiology and Respiratory Control Disorders in Childhood. Stan- dards for Services for Children with Disorders of Sleep Physiology.
London: Royal College of Paediatric and Child Health; 2009.
32 Koomson A, Morin I, Brouillette R, Brown KA. Children with severe OSAS who have adenotonsillectomy in the morning are less likely to have postoperative desaturation than those operated in the after- noon.Can. J. Anaesth.2004;51: 62–7.
33 Pang KP, Balakrishnan A. Paediatric obstructive sleep apnoea: Is a polysomnogram always necessary? J. Laryngol. Otol. 2004; 118:
275–8.
34 Kaditis A, Kheirandish-Gozal L, Gozal D. Pediatric OSAS: Oximetry can provide answers when polysomnography is not available.Sleep Med.
Rev.2016;27: 96–105.
35 Kalra M, Buncher R, Amin RS. Asthma as a risk factor for respiratory complications after adenotonsillectomy in children with obstructive breathing during sleep. Ann. Allergy Asthma Immunol. 2005; 94:
549–52.
