- Email: [email protected]
Polysomnography results in pediatric patients with mild obstructive sleep apnea: Adenotonsillectomy vs. watchful waiting
International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology journal homepage: www.elsevier.com/locate/ijporl
Polysomnography results in pediatric patients with mild obstructive sleep apnea: Adenotonsillectomy vs. watchful waiting Samuel J. Trosman a,*, David J. Eleff b, Jyoti Krishna c,1, Samantha Anne a a
Head and Neck Institute, Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195, United States Case Western Reserve University School of Medicine, 2109 Adelbert Rd, Cleveland, OH 44106, United States c Cleveland Clinic Sleep Disorders Center, Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195, United States b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 October 2015 Received in revised form 18 December 2015 Accepted 15 January 2016 Available online 27 January 2016
Objective: There is a lack of consensus and a paucity of data regarding how to best treat pediatric patients with mild obstructive sleep apnea. The objective of our study was to compare outcomes following adenotonsillectomy vs. observation in children with mild obstructive sleep apnea based on polysomnography results. Methods: A retrospective chart review was performed on children ages 9 months to 9 years with 2 or more polysomnograms completed at a tertiary care academic center. Children diagnosed with mild obstructive sleep apnea (obstructive apnea–hypopnea index 1–5) on polysomnography performed from 1999 to 2013 were included. Patients underwent adenotonsillectomy or watchful waiting for obstructive sleep apnea. The primary outcome was the change in apnea–hypopnea index. Results: There were 62 patients who met inclusion criteria for the study; 19 of the 62 patients were obese, while 15 had a craniofacial syndrome or hypotonia. Eighteen patients underwent adenotonsillectomy for mild obstructive sleep apnea while 44 were observed. The mean apnea–hypopnea index of patients after adenotonsillectomy improved from 3.50 (95% Confidence Interval [CI] 2.97–4.03) to 2.69 (95% CI 1.48–3.90), while the mean apnea–hypopnea index of the observation group worsened from 3.09 (95% CI 2.76–3.42) to 5.18 (95% CI 2.46–7.90). Between-group analysis showed significant improvement in the surgery group (p = 0.03), with a persistent improvement on multivariate analysis adjusting for baseline apnea–hypopnea index (p = 0.05). This difference was seen mostly in non-obese, nonsyndromic children (p = 0.04). There was no significant difference between groups amongst obese (p = 0.25) and syndromic (p = 0.36) patients. Conclusions: Adenotonsillectomy leads to a significant improvement in apnea–hypopnea index on follow-up polysomnography over an observational approach, especially in non-obese, non-syndromic children. A prospective, randomized trial is necessary to help determine appropriate treatment strategies for pediatric mild obstructive sleep apnea. ß 2016 Elsevier Ireland Ltd. All rights reserved.
Keywords: Obstructive sleep apnea Tonsillectomy Adenoidectomy Observation Apnea hypopnea index Polysomnography
1. Introduction Obstructive sleep apnea (OSA) is a common condition known to affect 1–4% of all children [1]. Adenotonsillectomy is considered to be the first line treatment for pediatric obstructive sleep apnea [2,3], with large prospective studies showing improvements in polysomnography findings, quality of life (QOL) metrics, and symptom severity scales compared to age-matched controls [4–6].
* Corresponding author. Tel.: +1 847 668 8435; fax: +1 216 445 9409. E-mail address: [email protected] (S.J. Trosman). 1 Present address: Department of Sleep Medicine, Akron Children’s Hospital, 215 W. Bowery St., Ste 6500, Akron, OH 44308, United States. http://dx.doi.org/10.1016/j.ijporl.2016.01.012 0165-5876/ß 2016 Elsevier Ireland Ltd. All rights reserved.
The efficacy of adenotonsillectomy in cases of mild OSA (AHI less than 5) is less clear. While surgery can have a positive impact on QOL even in cases of mild OSA [7], the degree of improvement in symptoms is not consistent in all children. Additionally, studies often exclude children with comorbid conditions such as obesity, asthma, and craniofacial abnormalities for which success rates may be lower [3]. Polysomnography (PSG) is the gold standard for diagnosing OSA yet, to our knowledge, there is a paucity of data regarding the effects of adenotonsillectomy on polysomnographic data in patients with mild OSA. The Childhood Adenotonsillectomy Trial (CHAT) compared PSG findings in children who were randomized to adenotonsillectomy vs. observation for OSA, however many of the children had moderate or severe OSA on their initial PSG [4]. In
26
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
addition, only patients ages 5–9 were included, thus the findings may not be generalizable to younger children who commonly present with concerns of sleep-disordered breathing. This helps explain the lack of consensus regarding how to best treat such patients. The objective of our study was to compare cure rates and the degrees of improvement following adenotonsillectomy vs. observation in children with mild OSA based on polysomnography results.
2.2. Intervention After the initial PSG, patients underwent observation or underwent tonsillectomy and adenoidectomy based on individualized clinician’s practice and parental input on impact on the child. Tonsillectomy was performed primarily in an extracapsular fashion using monopolar cautery. Adenoidectomy was performed using curettage, suction ablation and cautery, or microdebridement. Patients had repeat PSGs at various intervals based on clinical discretion.
2. Materials and methods
2.3. Statistical analysis
The study was approved by the Cleveland Clinic institutional review board. A retrospective chart review was performed on all pediatric patients under 10 years of age with mild OSA, defined as obstructive apnea–hypopnea index (OAHI, subsequently referred to as AHI) between 1 and 5, on PSG performed at a tertiary care academic center from 1999 to 2013 who also had a second PSG as ordered by their primary care physician, otolaryngologist and/or sleep medicine doctor. Inclusion criteria were that the patient (1) was seen clinically at our institution during the specified time period; (2) was under the age of 10 years at the time of the initial PSG; and (3) had 2 or more PSG’s performed within our health system available for review. Obstructive apnea/hypopnea index was chosen as the sole indicator of sleep apnea as per American Academy of Sleep Medicine (AASM) guidelines [8]. Patients who were 10 years or older at the time of initial PSG or who had undergone previous upper aerodigestive tract surgery, including adenoidectomy alone or combined adenotonsillectomy, were excluded. All patients who underwent use of positive pressure ventilation, including continuous positive airway pressure (CPAP), and dental appliances were also excluded. Information regarding baseline clinical characteristics and the timing and results of PSG findings were recorded from the electronic medical records.
Descriptive statistics, including means and standard deviations for continuous variables and percentages for categorical variables, were used to compare the treatment groups. Chi-square analysis was used to compare categorical data between groups, including whether or not patients showed resolution of OSA as defined above. Between-group comparisons of PSG changes were performed on log-transformed variables using the independent samples t test. In order to account for the baseline difference in AHI between treatment groups, a multivariate linear regression model was created using the log-transformed change in AHI as the dependent variable and the baseline AHI value as a covariate. Statistical analysis was performed using JMP (JMP1, Version h10i. SAS Institute Inc., Cary, NC, 1989–2007) software. Statistical significance was set at p 0.05.
2.1. Polysomnography and definitions All patients underwent standard overnight PSG at our institution in an AASM accredited laboratory that was interpreted by a pediatric physician board-certified/eligible in sleep medicine. Measured parameters included: bilateral electro-oculography using left and right leads; electroencephalography using bilateral frontal, central, and occipital leads; electromyography using mental, submental, and bilateral anterior tibialis leads; electrocardiography; continuous airflow monitoring with a thermistor and nasal pressure transducer; chest and abdominal effort; oxygen saturation using a pulse oximeter; end tidal carbon dioxide using a CO2 sensor; and body positioning via video monitoring. Obstructive apnea and hypopnea were defined based on AASM guidelines [8]. OSA was identified by the AHI and was used to screen patients with mild OSA. Resolution of mild OSA was defined as a subsequent PSG with an obstructive AHI value less than 1.0. Clinical characteristics such as medical comorbidities and the use of medications were retrieved from the medical records. Tonsil size was based on clinician’s transoral examination using the Brodsky grading scale [9] and was taken from the appointment at which the initial PSG was ordered, if available. Obesity was determined based on the body mass index (BMI) Z-score calculated using the Centers for Disease Control and Prevention 2000 growth charts [10]. A BMI Z-score in the 95th percentile (greater than 1.645) was considered obese. Patients were classified as syndromic if they had a previously diagnosed craniofacial syndrome or abnormality, Trisomy 21, cerebral palsy, or other neuromuscular disorder resulting in hypotonia.
3. Results 3.1. Patient characteristics There were 62 patients, ages 9 months to 9 years, identified with a diagnosis of mild OSA on polysomnography who had more than 1 PSG to review. Of the 62 patients, 18 patients underwent adenotonsillectomy for mild OSA while 44 patients were observed; 1 of the 18 surgical patients underwent partial intracapsular tonsillectomy. The clinical characteristics of the surgery and observation groups are shown in Table 1. There were no significant differences in age, gender, or baseline tonsil size between groups.
Table 1 Clinical characteristics of 62 patients with mild obstructive sleep apnea. Characteristics Gender Male Female Mean age, years (SD) Mean initial AHI, events/h (SD) Mean time between PSGs, months (SD)
Surgery (n = 18)
Observation (n = 44)
11 (61%) 7 (39%)
26 (59%) 18 (41%)
p-Value 0.88
3.1 (2.3) 3.5 (1.1)
4.5 (2.7) 3.1 (1.1)
0.12 0.16
16.0 (13.9)
16.7 (13.4)
0.54
Obesity Yes No
8 (44%) 10 (56%)
11 (25%) 33 (75%)
Syndromic Yes No
6 (33%) 12 (67%)
9 (20%) 35 (80%)
Tonsil size 1–2 3–4 Unknown
10 (56%) 7 (39%) 1 (6%)
27 (61%) 15 (34%) 2 (5%)
SD, Standard deviation.
0.13
0.28
0.69
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
27
Fig. 1. Polysomnography data for all 62 patients with mild OSA. Legend: Line diagram showing initial and follow-up AHI values in events per hour for (A) 18 surgery patients, and (B) 44 observation patients. The surgery group showed improvement compared to the observation group (p = 0.03).
Medical treatment of OSA was attempted in a minority of patients in the observation group, including 23% treated with nasal steroids and 11% treated with oral montelukast. 3.2. Polysomnography data—All 62 patients The average obstructive apnea/hypopnea index on the initial PSG was 3.50 (95% Confidence Interval [CI] 2.97–4.03) in the surgery group and 3.09 (95% CI 2.76–3.42) in the observation group. The mean interval between the initial and subsequent PSG was similar between the surgery and observation groups (16.0 months vs. 16.7 months, p = 0.54). The average time from surgery to follow-up PSG for patients undergoing adenotonsillectomy was 10.3 months. Five of the 18 total patients who underwent adenotonsillectomy for mild OSA showed resolution of OSA (defined as AHI < 1.0) on subsequent PSG, compared to 5 of 44 patients who were observed (28% vs. 11%, p = 0.14). The mean AHI for the surgery group improved from 3.50 to 2.69 (95% CI 1.48–3.90), while the mean AHI for the observation group worsened from 3.09 to 5.18 (95% CI 2.46–7.90) on subsequent PSG. Between-group analysis showed that improvement in AHI score was significantly greater in patients who underwent surgery than in those who were observed (p = 0.03, Fig. 1). 3.3. Outcomes in non-obese, non-syndromic patients with mild OSA
AHI for the 6 children who underwent surgery improved from 3.30 (95% CI 2.73–3.87) to 2.43 (95% CI 0.18–5.04), while the mean AHI for the 24 patients observed worsened from 2.95 (95% CI 2.50– 3.40) to 6.90 (95% CI 1.88–11.87) on subsequent PSG. Betweengroup analysis showed that improvement in AHI score was significantly greater in non-obese, non-syndromic patients who underwent surgery than in those who were observed (p = 0.04, Fig. 2). 3.4. Outcomes in obese children with mild OSA Out of 62 patients, there were 19 children (31%) who met criteria for pediatric obesity based on BMI Z-score. There was no significant difference between obese and non-obese patients in average initial AHI (3.47 vs. 3.09, p = 0.28), age (mean 4.6 years vs. 3.9 years, p = 0.34), or baseline tonsil size (p = 0.64, Table 2) Of the 19 obese patients, 8 children (42%) underwent adenotonsillectomy while 11 were observed. Only 1 of the 8 children who had surgery showed resolution of OSA on subsequent PSG, compared to none of the 11 children who were observed (13% vs. 0%, p = 0.42). The mean AHI for obese patients who underwent surgery did improve from 3.83 to 3.08, while the mean AHI for obese children who were observed worsened slightly from 3.20 to 3.40. Between-group analysis showed no significant difference in the change in AHI between obese children who underwent surgery and those who were observed (p = 0.25, Fig. 3).
Out of 62 patients, there were 30 (48%) that were not syndromic and did not meet criteria for pediatric obesity. Six of the 30 children underwent adenotonsillectomy for mild OSA while 24 were observed. There was no significant difference in baseline tonsil size between groups (p = 0.65, Table 2). Three of the 6 patients who underwent surgery showed resolution of OSA on follow-up PSG, while only 3 of 24 patients who were observed showed normalization of PSG findings (50% vs. 13%, p = 0.07). The mean
Table 2 Baseline tonsil size of healthy, obese, and syndromic patients. Group
Tonsil size
Surgery
Observation
p-Value
Non-obese, non-syndromic (n = 30)
1–2 3–4 Unknown
2 (33%) 3 (50%) 1 (16%)
13 (54%) 11 (46%) 0 (0%)
0.65
Obese (n = 19)
1–2 3–4
5 (63%) 3 (38%)
8 (73%) 3 (27%)
0.64
Syndromic (n = 15)
1–2 3–4 Unknown
3 (50%) 3 (50%) 0 (0%)
6 (67%) 1 (11%) 2 (22%)
0.27
Fig. 2. Polysomnography data for non-obese, non-syndromic patients with mild OSA. Legend: Bar graph showing mean AHI in events per hour with standard error measurements for 30 non-obese, non-syndromic patients. The surgery group showed improvement compared to the observation group (p = 0.04).
28
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
3.6. Multivariate analysis The results of a multivariate linear regression model with the change in AHI as the dependent variable and the baseline AHI values as a covariate are shown in Table 3. The difference in baseline AHI had a marginal impact on the change in AHI score from 1st to 2nd PSG (p = 0.09). After accounting for this difference, there remained an improvement in AHI in the surgery group compared to the observation group (p = 0.05). 4. Discussion
Fig. 3. Polysomnography data for obese patients with mild OSA. Legend: Bar graph showing mean AHI in events per hour with standard error measurements for 19 obese patients. There was no significant difference between the surgery and observation groups (p = 0.25).
3.5. Outcomes in syndromic children with mild OSA Fifteen of 62 patients (24%) were classified as syndromic by study criteria. The most common conditions seen were Down syndrome (n = 7), cerebral palsy (n = 4), DiGeorge syndrome (n = 1), Pierre– Robin sequence (n = 1), and other chromosomal abnormalities resulting in hypotonia (n = 2). There was no significant difference between syndromic and non-syndromic patients in average initial AHI (3.22 vs. 3.21, p = 0.88), age (mean 3.4 years vs. 4.4 years, p = 0.19), or baseline tonsil size (p = 0.27, Table 2). Of the 15 syndromic patients, 6 (40%) underwent adenotonsillectomy while 9 were observed. Two of the 6 patients showed resolution of OSA on subsequent PSG, compared to 2 of 9 patients who were observed (33% vs. 22%, p = 0.54). The mean AHI for syndromic patients who underwent surgery improved from 3.08 to 2.03, while the mean AHI for syndromic children who were observed improved from 3.31 to 2.84. Between-group analysis showed no significant difference in the change in AHI between syndromic children who underwent surgery and those who were observed (p = 0.36, Fig. 4).
The natural history of pediatric mild OSA remains unknown. Proponents of an observational approach cite a lack of neuropsychological impairment in children with mild OSA [11] as well as the low prevalence of tonsillar hypertrophy after children turn 8 years old [12]. However, in a prospective cohort study performed over 2 years, Li et al. found that mild OSA in children did not resolve spontaneously, and that nearly a third of children worsened by AHI score from baseline [13]. Also, in a recent prospective trial, Volsky et al. showed significant improvements in quality of life at 8 months follow-up in pediatric patients undergoing adenotonsillectomy for mild OSA compared to controls opting for observation [7]. However, QOL questionnaires have shown poor correlation with the severity of OSA in pediatric patients [14], thus it is uncertain whether surgery results in significant improvement in objective respiratory parameters when compared to watchful waiting in cases of mild OSA. Patients undergoing adenotonsillectomy showed, on average, a 0.8-point improvement in AHI on postoperative PSG. Twenty-eight percent of patients in the surgery group had normalization of their AHI postoperatively; this number increased to 50% when looking at non-obese, non-syndromic children, albeit with a very small sample size of 6 patients. This success rate is well below rates of 80% reported in meta-analyses [6] and 79% reported in the CHAT study [4]. It should be noted that the age group of our cohort was younger than the 5–9 year old children recruited by the CHAT trial. Nevertheless, the improvement in AHI score with surgery was significantly greater than that seen with observation; this was still present when accounting for differences in baseline AHI, though to a lesser degree which may be owing to the small sample size involved. Only 11% of patients in our observation group showed resolution of OSA, also a much lower percentage of patients than the 46% rate of resolution with observation seen in the CHAT study amongst slightly older children (average age 6.5 years vs. 4.5 years in our group) [4]. Not only did children who underwent watchful waiting not resolve their OSA, but the average AHI of the observation group worsened by 2.09 points over the course of the study in all patients and by nearly 4 points in non-obese, nonsyndromic patients. While the latter AHI values are somewhat skewed by a few children who worsened significantly, overall these results support the findings of Li et al. that mild OSA in the majority of pediatric patients does not resolve spontaneously, and
Table 3 Multivariate linear regression analysis with log-transformed AHI as the dependent variable, surgery vs. observation as the independent variable and baseline AHI as a covariate.
Fig. 4. Polysomnography data for syndromic patients with mild OSA. Legend: Bar graph showing mean AHI in events per hour with standard error measurements for 15 syndromic patients. There was no significant difference between the surgery and observation groups (p = 0.36).
Variable
Model parameter estimate (SE)
Medians ratio (95% CI)
p-Value
Baseline AHI value Surgery vs. observation
0.066 (0.038) 0.249 (0.126)
0.86 (0.72–1.02) 0.56 (0.32–0.99)
0.09 0.05
SE, Standard Error. CI, Confidence Interval.
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
that a significant proportion of these patients will get objectively worse over the course of 1–2 years [13]. Although adenotonsillectomy was not 100% effective in resolving mild OSA, surgery at a younger age normalized PSG findings in a significant proportion of patients, improved the average number of apneic events in nearly all patients, and seemed to protect patients from worsening of OSA over time. The lower rate of normalization of OSA in both our surgery and observation groups when compared to the CHAT trial may be attributed to a selection bias inherent to our retrospective review, in that it was likely more common for children to be referred for a second PSG if they were still symptomatic, while children who had improvement in symptoms did not undergo a repeat PSG and are not represented in this analysis. While the degree of bias is difficult to quantify, an observational comparison of the trends over time remains of interest. In our study, improvement with adenotonsillectomy in obese patients with mild OSA was not as pronounced as it was in healthy children. The mean AHI in obese children improved by an average of 0.75 despite having a higher initial AHI than non-obese patients, and only 13% showed resolution of OSA on subsequent PSG. Previous studies have suggested that pediatric obesity is a risk factor for persistence of OSA after adenotonsillectomy [15]. In a meta-analysis of sleep parameters after adenotonsillectomy in obese children, Costa et al. found that, while the mean AHI was significantly reduced and the oxygen saturation nadir increased after surgery, only 25% of patients had a postoperative AHI <2, although the starting AHI was much higher than in our study [16]. In the CHAT trial, while obese children also showed lower rates of resolution of OSA, 67% of obese patients normalized their AHI after adenotonsillectomy compared to 29% of observed patients, prompting the investigators to support a strategy of early adenotonsillectomy in obese children as well [4]. Interestingly, in our cohort, obese children with mild OSA who were observed did not show as severe of worsening of OSA over time as was seen in non-obese patients, and there was no significant difference in the change in AHI between obese patients who underwent surgery and those who were observed. Given the decreased rates of normalization of OSA in obese patients, this finding may suggest a role for watchful waiting in obese children with mild OSA. Non-invasive ventilation such as continuous positive airway pressure (CPAP) may be an alternative option in older children, although patient tolerance and compliance need to be monitored. Nasal steroids and oral montelukast have also shown improvement in children with mild OSA and can be considered [17–19]. Patients should be followed closely to ensure that symptoms don’t worsen, and weight loss programs should be implemented. Children with craniofacial anomalies or hypotonia can be much more difficult to treat for OSA due to several different anatomic sites of obstruction. Recent reviews have shown over 80% of patients with congenital craniofacial malformations referred for a PSG had evidence of OSA; adenotonsillectomy resulted in normalization of AHI in only 14–15% of these patients [20,21]. In our cohort, there was a 33% rate of resolution of mild OSA with a mean improvement of 1.05 in postoperative AHI, however there was no statistically significant change in AHI when compared to the observation group. The benefit of a modest improvement in OSA severity must be weighed against the risk of anesthesia and postoperative complications in syndromic patients, which may be substantial. It should be noted that, in our study, we use AHI as the sole objective measure of pediatric OSA while other large-scale trials have reported on several different PSG criteria [4]. While inclusion of O2 and CO2 data would be of interest, we are unsure of the utility of this in our patient population that a priori has mild OSA. Subsequent papers from the CHAT trial investigating overnight
29
end-tidal carbon dioxide (EtCO2) levels show weak to modest correlations with other polysomnographic indices, and in adjusted analyses, baseline total sleep time with EtCO2 >50 mmHg did not predict postoperative changes in cognitive and behavioral measurements [22]. In another paper from the trial, while oxygen criteria were not used to define OSA resolution, they were included in the analyses of predictors which concluded that lower snoring score, low AHI and normal waist circumference were the only independent predictors for spontaneous resolution of OSA [23]. Thus, it is our experience that, of the parameters recorded in the PSG, the AHI is the most commonly used measure of severity in mild cases of OSA. For this reason, we chose to restrict our analyses to AHI data alone. However, other factors need to be taken into account before a decision regarding surgery or observation for mild OSA can be made, including patient comorbidities, quality of life measures, and age, amongst others. AHI by itself is not the only indicator of the extent of disease burden and should not be used as the sole factor in decisionmaking. Our study has limitations that deserve mention. This review is purely retrospective and thus is limited by the quality of the medical records. The lack of randomization in the study likely lead to increased rates of follow-up and subsequent PSGs in the more symptomatic patients, including obese and syndromic patients who were at higher risk for persistent OSA. Clinicians may have been more likely to obtain a second PSG in the persistently symptomatic patient than in the child whose symptoms have resolved. In addition, the small sample size in the study meant that not only were many of the differences seen not highly statistically significant, but also there may have been differences between the groups that were not detected due to a lack of statistical power. A prospective, randomized trial comparing adenotonsillectomy and watchful waiting in children with mild OSA across a broad age range is needed in order to confirm or refute our results. Despite these limitations, our review demonstrates important observations regarding the natural course of mild OSA and the impact of adenotonsillectomy in healthy, obese, and syndromic children. 5. Conclusions This review shows that adenotonsillectomy for mild OSA in the pediatric population resulted in improvement in postoperative AHI most markedly seen among non-obese, non-syndromic children, while an observational approach not only didn’t lead to normalization of AHI in the majority of patients, but actually resulted in worsening of AHI values on subsequent PSG in several patients. Prospective, randomized trials are needed to further clarify the course of and the effect of surgical intervention for pediatric mild OSA. Previous presentation Preliminary data from this study were presented as a podium presentation at the American Society of Pediatric Otolaryngology (ASPO) Annual Spring Meeting on April 26, 2015 in Boston, MA. Disclosures The authors report no conflicts of interest to disclose. Acknowledgements Samuel J. Trosman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
30
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
References [1] J.C. Lumeng, R.D. Chervin, Epidemiology of pediatric obstructive sleep apnea, Proc. Am. Thorac. Soc. 5 (2) (2008) 242–252. [2] M.S. Schechter, Technical report: diagnosis and management of childhood obstructive sleep apnea syndrome, Pediatrics 109 (4) (2002) e69. [3] R.F. Baugh, S.M. Archer, R.B. Mitchell, R.M. Rosenfeld, R. Amin, J.J. Burns, et al., Clinical practice guideline: tonsillectomy in children, Otolaryngol. Head Neck Surg. 144 (Suppl. 1) (2011) S1–S30. [4] C.L. Marcus, R.H. Moore, C.L. Rosen, B. Giordani, S.L. Garrett, H.G. Taylor, et al., A randomized trial of adenotonsillectomy for childhood sleep apnea, N. Engl. J. Med. 368 (25) (2013) 2366–2376. [5] R.B. Mitchell, Adenotonsillectomy for obstructive sleep apnea in children: outcome evaluated by pre and postoperative polysomnography, Laryngoscope 117 (10) (2007) 1844–1854. [6] S.E. Brietzke, D. Gallagher, The effectiveness of tonsillectomy and adenoidectomy in the treatment of pediatric obstructive sleep apnea/hypopnea syndrome: a meta-analysis, Otolaryngol. Head Neck Surg. 134 (6) (2006) 979–984. [7] P.G. Volsky, M.A. Woughter, H.A. Beydoun, C.S. Derkay, C.M. Baldassari, Adenotonsillectomy vs. observation for management of mild obstructive sleep apnea in children, Otolaryngol. Head Neck Surg. 150 (1) (2014) 126–132. [8] R.B. Berry, R. Budhiraja, D.J. Gottlieb, D. Gozal, C. Iber, V.K. Kapur, et al., Rules for scoring respiratory events in sleep: update of the 2007 AASM manual for the scoring of sleep and associated events. Deliberations of the sleep apnea definitions task force of the American Academy of Sleep Medicine, J. Clin. Sleep Med. 8 (5) (2012) 597–619. [9] S.K. Ng, D.L. Lee, A.M. Li, Y.K. Wing, M.C. Tong, Reproducibility of clinical grading of tonsillar size, Arch. Otolaryngol. Head Neck Surg. 136 (2) (2010) 159–162. [10] C.L. Ogden, R.J. Kuczmarski, K.M. Flegal, Z. Mei, S. Guo, R. Wei, et al., Centers for Disease Control and Prevention 2000 growth charts for the United States: improvements to the 1977 National Center for Health Statistics version, Pediatrics 109 (1) (2002) 45–60. [11] S.L. Calhoun, S.D. Mayes, A.N. Vgontzas, M. Tsaoussoglou, L.J. Shifflett, E.O. Bixler, No relationship between neurocognitive functioning and mild sleep disordered breathing in a community sample of children, J. Clin. Sleep Med. 5 (3) (2009) 228–234.
[12] A. Akcay, C.O. Kara, E. Dagdeviren, M. Zencir, Variation in tonsil size in 4- to 17year-old schoolchildren, J. Otolaryngol. 35 (4) (2006) 270–274. [13] A.M. Li, C.T. Au, S.K. Ng, V.J. Abdullah, C. Ho, T.F. Fok, et al., Natural history and predictors for progression of mild childhood obstructive sleep apnoea, Thorax 65 (1) (2010) 27–31. [14] R.B. Mitchell, J. Kelly, Quality of life after adenotonsillectomy for SDB in children, Otolaryngol. Head Neck Surg. 133 (4) (2005) 569–572. [15] R. Bhattacharjee, L. Kheirandish-Gozal, K. Spruyt, R.B. Mitchell, J. Promchiarak, N. Simakajornboon, et al., Adenotonsillectomy outcomes in treatment of obstructive sleep apnea in children: a multicenter retrospective study, Am. J. Respir. Crit. Care Med. 182 (5) (2010) 676–683. [16] D.J. Costa, R. Mitchell, Adenotosillectomy for obstructive sleep apnea in obese children: a meta-analysis, Otolaryngol. Head Neck Surg. 140 (4) (2009) 455–460. [17] L. Kheirandish, A.D. Goldbart, D. Gozal, Intranasal steroids and oral leukotriene modifier therapy in residual sleep-disordered breathing after tonsillectomy and adenoidectomy in children, Pediatrics 117 (1) (2006) e61–e66. [18] L. Kheirandish-Gozal, D. Gozal, Intranasal budesonide treatment for children with mild obstructive sleep apnea syndrome, Pediatrics 122 (1) (2008) e149–e155. [19] A.D. Goldbart, S. Greenberg-Dotan, A. Tal, Montelukast for children with obstructive sleep apnea: a double-blind, placebo-controlled study, Pediatrics 130 (3) (2012) e575–e580. [20] M. Moraleda-Cibria´n, S.P. Edwards, S.J. Kasten, S.R. Buchman, M. Berger, L.M. O’Brien, Obstructive sleep apnea pretreatment and posttreatment in symptomatic children with congenital craniofacial malformations, J. Clin. Sleep Med. 11 (1) (2015) 37–43. [21] S.O. Zandieh, B.L. Padwa, E.S. Katz, Adenotonsillectomy for obstructive sleep apnea in children with syndromic craniosynostosis, Plast. Reconstr. Surg. 131 (4) (2013) 847–852. [22] S. Paruthi, C.L. Rosen, R. Wang, J. Weng, C.L. Marcus, R.D. Chervin, et al., End-tidal carbon dioxide measurement during pediatric polysomnography: signal quality, association with apnea severity, and prediction of neurobehavioral outcomes, Sleep 38 (11) (2015) 1719–1726. [23] R.D. Chervin, S.S. Ellenberg, X. Hou, C.L. Marcus, S.L. Garetz, E.S. Katz, et al., Prognosis for spontaneous resolution of OSA in children, Chest 148 (5) (2015) 1204–1213.
Contents lists available at ScienceDirect
International Journal of Pediatric Otorhinolaryngology journal homepage: www.elsevier.com/locate/ijporl
Polysomnography results in pediatric patients with mild obstructive sleep apnea: Adenotonsillectomy vs. watchful waiting Samuel J. Trosman a,*, David J. Eleff b, Jyoti Krishna c,1, Samantha Anne a a
Head and Neck Institute, Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195, United States Case Western Reserve University School of Medicine, 2109 Adelbert Rd, Cleveland, OH 44106, United States c Cleveland Clinic Sleep Disorders Center, Cleveland Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195, United States b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 October 2015 Received in revised form 18 December 2015 Accepted 15 January 2016 Available online 27 January 2016
Objective: There is a lack of consensus and a paucity of data regarding how to best treat pediatric patients with mild obstructive sleep apnea. The objective of our study was to compare outcomes following adenotonsillectomy vs. observation in children with mild obstructive sleep apnea based on polysomnography results. Methods: A retrospective chart review was performed on children ages 9 months to 9 years with 2 or more polysomnograms completed at a tertiary care academic center. Children diagnosed with mild obstructive sleep apnea (obstructive apnea–hypopnea index 1–5) on polysomnography performed from 1999 to 2013 were included. Patients underwent adenotonsillectomy or watchful waiting for obstructive sleep apnea. The primary outcome was the change in apnea–hypopnea index. Results: There were 62 patients who met inclusion criteria for the study; 19 of the 62 patients were obese, while 15 had a craniofacial syndrome or hypotonia. Eighteen patients underwent adenotonsillectomy for mild obstructive sleep apnea while 44 were observed. The mean apnea–hypopnea index of patients after adenotonsillectomy improved from 3.50 (95% Confidence Interval [CI] 2.97–4.03) to 2.69 (95% CI 1.48–3.90), while the mean apnea–hypopnea index of the observation group worsened from 3.09 (95% CI 2.76–3.42) to 5.18 (95% CI 2.46–7.90). Between-group analysis showed significant improvement in the surgery group (p = 0.03), with a persistent improvement on multivariate analysis adjusting for baseline apnea–hypopnea index (p = 0.05). This difference was seen mostly in non-obese, nonsyndromic children (p = 0.04). There was no significant difference between groups amongst obese (p = 0.25) and syndromic (p = 0.36) patients. Conclusions: Adenotonsillectomy leads to a significant improvement in apnea–hypopnea index on follow-up polysomnography over an observational approach, especially in non-obese, non-syndromic children. A prospective, randomized trial is necessary to help determine appropriate treatment strategies for pediatric mild obstructive sleep apnea. ß 2016 Elsevier Ireland Ltd. All rights reserved.
Keywords: Obstructive sleep apnea Tonsillectomy Adenoidectomy Observation Apnea hypopnea index Polysomnography
1. Introduction Obstructive sleep apnea (OSA) is a common condition known to affect 1–4% of all children [1]. Adenotonsillectomy is considered to be the first line treatment for pediatric obstructive sleep apnea [2,3], with large prospective studies showing improvements in polysomnography findings, quality of life (QOL) metrics, and symptom severity scales compared to age-matched controls [4–6].
* Corresponding author. Tel.: +1 847 668 8435; fax: +1 216 445 9409. E-mail address: [email protected] (S.J. Trosman). 1 Present address: Department of Sleep Medicine, Akron Children’s Hospital, 215 W. Bowery St., Ste 6500, Akron, OH 44308, United States. http://dx.doi.org/10.1016/j.ijporl.2016.01.012 0165-5876/ß 2016 Elsevier Ireland Ltd. All rights reserved.
The efficacy of adenotonsillectomy in cases of mild OSA (AHI less than 5) is less clear. While surgery can have a positive impact on QOL even in cases of mild OSA [7], the degree of improvement in symptoms is not consistent in all children. Additionally, studies often exclude children with comorbid conditions such as obesity, asthma, and craniofacial abnormalities for which success rates may be lower [3]. Polysomnography (PSG) is the gold standard for diagnosing OSA yet, to our knowledge, there is a paucity of data regarding the effects of adenotonsillectomy on polysomnographic data in patients with mild OSA. The Childhood Adenotonsillectomy Trial (CHAT) compared PSG findings in children who were randomized to adenotonsillectomy vs. observation for OSA, however many of the children had moderate or severe OSA on their initial PSG [4]. In
26
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
addition, only patients ages 5–9 were included, thus the findings may not be generalizable to younger children who commonly present with concerns of sleep-disordered breathing. This helps explain the lack of consensus regarding how to best treat such patients. The objective of our study was to compare cure rates and the degrees of improvement following adenotonsillectomy vs. observation in children with mild OSA based on polysomnography results.
2.2. Intervention After the initial PSG, patients underwent observation or underwent tonsillectomy and adenoidectomy based on individualized clinician’s practice and parental input on impact on the child. Tonsillectomy was performed primarily in an extracapsular fashion using monopolar cautery. Adenoidectomy was performed using curettage, suction ablation and cautery, or microdebridement. Patients had repeat PSGs at various intervals based on clinical discretion.
2. Materials and methods
2.3. Statistical analysis
The study was approved by the Cleveland Clinic institutional review board. A retrospective chart review was performed on all pediatric patients under 10 years of age with mild OSA, defined as obstructive apnea–hypopnea index (OAHI, subsequently referred to as AHI) between 1 and 5, on PSG performed at a tertiary care academic center from 1999 to 2013 who also had a second PSG as ordered by their primary care physician, otolaryngologist and/or sleep medicine doctor. Inclusion criteria were that the patient (1) was seen clinically at our institution during the specified time period; (2) was under the age of 10 years at the time of the initial PSG; and (3) had 2 or more PSG’s performed within our health system available for review. Obstructive apnea/hypopnea index was chosen as the sole indicator of sleep apnea as per American Academy of Sleep Medicine (AASM) guidelines [8]. Patients who were 10 years or older at the time of initial PSG or who had undergone previous upper aerodigestive tract surgery, including adenoidectomy alone or combined adenotonsillectomy, were excluded. All patients who underwent use of positive pressure ventilation, including continuous positive airway pressure (CPAP), and dental appliances were also excluded. Information regarding baseline clinical characteristics and the timing and results of PSG findings were recorded from the electronic medical records.
Descriptive statistics, including means and standard deviations for continuous variables and percentages for categorical variables, were used to compare the treatment groups. Chi-square analysis was used to compare categorical data between groups, including whether or not patients showed resolution of OSA as defined above. Between-group comparisons of PSG changes were performed on log-transformed variables using the independent samples t test. In order to account for the baseline difference in AHI between treatment groups, a multivariate linear regression model was created using the log-transformed change in AHI as the dependent variable and the baseline AHI value as a covariate. Statistical analysis was performed using JMP (JMP1, Version h10i. SAS Institute Inc., Cary, NC, 1989–2007) software. Statistical significance was set at p 0.05.
2.1. Polysomnography and definitions All patients underwent standard overnight PSG at our institution in an AASM accredited laboratory that was interpreted by a pediatric physician board-certified/eligible in sleep medicine. Measured parameters included: bilateral electro-oculography using left and right leads; electroencephalography using bilateral frontal, central, and occipital leads; electromyography using mental, submental, and bilateral anterior tibialis leads; electrocardiography; continuous airflow monitoring with a thermistor and nasal pressure transducer; chest and abdominal effort; oxygen saturation using a pulse oximeter; end tidal carbon dioxide using a CO2 sensor; and body positioning via video monitoring. Obstructive apnea and hypopnea were defined based on AASM guidelines [8]. OSA was identified by the AHI and was used to screen patients with mild OSA. Resolution of mild OSA was defined as a subsequent PSG with an obstructive AHI value less than 1.0. Clinical characteristics such as medical comorbidities and the use of medications were retrieved from the medical records. Tonsil size was based on clinician’s transoral examination using the Brodsky grading scale [9] and was taken from the appointment at which the initial PSG was ordered, if available. Obesity was determined based on the body mass index (BMI) Z-score calculated using the Centers for Disease Control and Prevention 2000 growth charts [10]. A BMI Z-score in the 95th percentile (greater than 1.645) was considered obese. Patients were classified as syndromic if they had a previously diagnosed craniofacial syndrome or abnormality, Trisomy 21, cerebral palsy, or other neuromuscular disorder resulting in hypotonia.
3. Results 3.1. Patient characteristics There were 62 patients, ages 9 months to 9 years, identified with a diagnosis of mild OSA on polysomnography who had more than 1 PSG to review. Of the 62 patients, 18 patients underwent adenotonsillectomy for mild OSA while 44 patients were observed; 1 of the 18 surgical patients underwent partial intracapsular tonsillectomy. The clinical characteristics of the surgery and observation groups are shown in Table 1. There were no significant differences in age, gender, or baseline tonsil size between groups.
Table 1 Clinical characteristics of 62 patients with mild obstructive sleep apnea. Characteristics Gender Male Female Mean age, years (SD) Mean initial AHI, events/h (SD) Mean time between PSGs, months (SD)
Surgery (n = 18)
Observation (n = 44)
11 (61%) 7 (39%)
26 (59%) 18 (41%)
p-Value 0.88
3.1 (2.3) 3.5 (1.1)
4.5 (2.7) 3.1 (1.1)
0.12 0.16
16.0 (13.9)
16.7 (13.4)
0.54
Obesity Yes No
8 (44%) 10 (56%)
11 (25%) 33 (75%)
Syndromic Yes No
6 (33%) 12 (67%)
9 (20%) 35 (80%)
Tonsil size 1–2 3–4 Unknown
10 (56%) 7 (39%) 1 (6%)
27 (61%) 15 (34%) 2 (5%)
SD, Standard deviation.
0.13
0.28
0.69
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
27
Fig. 1. Polysomnography data for all 62 patients with mild OSA. Legend: Line diagram showing initial and follow-up AHI values in events per hour for (A) 18 surgery patients, and (B) 44 observation patients. The surgery group showed improvement compared to the observation group (p = 0.03).
Medical treatment of OSA was attempted in a minority of patients in the observation group, including 23% treated with nasal steroids and 11% treated with oral montelukast. 3.2. Polysomnography data—All 62 patients The average obstructive apnea/hypopnea index on the initial PSG was 3.50 (95% Confidence Interval [CI] 2.97–4.03) in the surgery group and 3.09 (95% CI 2.76–3.42) in the observation group. The mean interval between the initial and subsequent PSG was similar between the surgery and observation groups (16.0 months vs. 16.7 months, p = 0.54). The average time from surgery to follow-up PSG for patients undergoing adenotonsillectomy was 10.3 months. Five of the 18 total patients who underwent adenotonsillectomy for mild OSA showed resolution of OSA (defined as AHI < 1.0) on subsequent PSG, compared to 5 of 44 patients who were observed (28% vs. 11%, p = 0.14). The mean AHI for the surgery group improved from 3.50 to 2.69 (95% CI 1.48–3.90), while the mean AHI for the observation group worsened from 3.09 to 5.18 (95% CI 2.46–7.90) on subsequent PSG. Between-group analysis showed that improvement in AHI score was significantly greater in patients who underwent surgery than in those who were observed (p = 0.03, Fig. 1). 3.3. Outcomes in non-obese, non-syndromic patients with mild OSA
AHI for the 6 children who underwent surgery improved from 3.30 (95% CI 2.73–3.87) to 2.43 (95% CI 0.18–5.04), while the mean AHI for the 24 patients observed worsened from 2.95 (95% CI 2.50– 3.40) to 6.90 (95% CI 1.88–11.87) on subsequent PSG. Betweengroup analysis showed that improvement in AHI score was significantly greater in non-obese, non-syndromic patients who underwent surgery than in those who were observed (p = 0.04, Fig. 2). 3.4. Outcomes in obese children with mild OSA Out of 62 patients, there were 19 children (31%) who met criteria for pediatric obesity based on BMI Z-score. There was no significant difference between obese and non-obese patients in average initial AHI (3.47 vs. 3.09, p = 0.28), age (mean 4.6 years vs. 3.9 years, p = 0.34), or baseline tonsil size (p = 0.64, Table 2) Of the 19 obese patients, 8 children (42%) underwent adenotonsillectomy while 11 were observed. Only 1 of the 8 children who had surgery showed resolution of OSA on subsequent PSG, compared to none of the 11 children who were observed (13% vs. 0%, p = 0.42). The mean AHI for obese patients who underwent surgery did improve from 3.83 to 3.08, while the mean AHI for obese children who were observed worsened slightly from 3.20 to 3.40. Between-group analysis showed no significant difference in the change in AHI between obese children who underwent surgery and those who were observed (p = 0.25, Fig. 3).
Out of 62 patients, there were 30 (48%) that were not syndromic and did not meet criteria for pediatric obesity. Six of the 30 children underwent adenotonsillectomy for mild OSA while 24 were observed. There was no significant difference in baseline tonsil size between groups (p = 0.65, Table 2). Three of the 6 patients who underwent surgery showed resolution of OSA on follow-up PSG, while only 3 of 24 patients who were observed showed normalization of PSG findings (50% vs. 13%, p = 0.07). The mean
Table 2 Baseline tonsil size of healthy, obese, and syndromic patients. Group
Tonsil size
Surgery
Observation
p-Value
Non-obese, non-syndromic (n = 30)
1–2 3–4 Unknown
2 (33%) 3 (50%) 1 (16%)
13 (54%) 11 (46%) 0 (0%)
0.65
Obese (n = 19)
1–2 3–4
5 (63%) 3 (38%)
8 (73%) 3 (27%)
0.64
Syndromic (n = 15)
1–2 3–4 Unknown
3 (50%) 3 (50%) 0 (0%)
6 (67%) 1 (11%) 2 (22%)
0.27
Fig. 2. Polysomnography data for non-obese, non-syndromic patients with mild OSA. Legend: Bar graph showing mean AHI in events per hour with standard error measurements for 30 non-obese, non-syndromic patients. The surgery group showed improvement compared to the observation group (p = 0.04).
28
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
3.6. Multivariate analysis The results of a multivariate linear regression model with the change in AHI as the dependent variable and the baseline AHI values as a covariate are shown in Table 3. The difference in baseline AHI had a marginal impact on the change in AHI score from 1st to 2nd PSG (p = 0.09). After accounting for this difference, there remained an improvement in AHI in the surgery group compared to the observation group (p = 0.05). 4. Discussion
Fig. 3. Polysomnography data for obese patients with mild OSA. Legend: Bar graph showing mean AHI in events per hour with standard error measurements for 19 obese patients. There was no significant difference between the surgery and observation groups (p = 0.25).
3.5. Outcomes in syndromic children with mild OSA Fifteen of 62 patients (24%) were classified as syndromic by study criteria. The most common conditions seen were Down syndrome (n = 7), cerebral palsy (n = 4), DiGeorge syndrome (n = 1), Pierre– Robin sequence (n = 1), and other chromosomal abnormalities resulting in hypotonia (n = 2). There was no significant difference between syndromic and non-syndromic patients in average initial AHI (3.22 vs. 3.21, p = 0.88), age (mean 3.4 years vs. 4.4 years, p = 0.19), or baseline tonsil size (p = 0.27, Table 2). Of the 15 syndromic patients, 6 (40%) underwent adenotonsillectomy while 9 were observed. Two of the 6 patients showed resolution of OSA on subsequent PSG, compared to 2 of 9 patients who were observed (33% vs. 22%, p = 0.54). The mean AHI for syndromic patients who underwent surgery improved from 3.08 to 2.03, while the mean AHI for syndromic children who were observed improved from 3.31 to 2.84. Between-group analysis showed no significant difference in the change in AHI between syndromic children who underwent surgery and those who were observed (p = 0.36, Fig. 4).
The natural history of pediatric mild OSA remains unknown. Proponents of an observational approach cite a lack of neuropsychological impairment in children with mild OSA [11] as well as the low prevalence of tonsillar hypertrophy after children turn 8 years old [12]. However, in a prospective cohort study performed over 2 years, Li et al. found that mild OSA in children did not resolve spontaneously, and that nearly a third of children worsened by AHI score from baseline [13]. Also, in a recent prospective trial, Volsky et al. showed significant improvements in quality of life at 8 months follow-up in pediatric patients undergoing adenotonsillectomy for mild OSA compared to controls opting for observation [7]. However, QOL questionnaires have shown poor correlation with the severity of OSA in pediatric patients [14], thus it is uncertain whether surgery results in significant improvement in objective respiratory parameters when compared to watchful waiting in cases of mild OSA. Patients undergoing adenotonsillectomy showed, on average, a 0.8-point improvement in AHI on postoperative PSG. Twenty-eight percent of patients in the surgery group had normalization of their AHI postoperatively; this number increased to 50% when looking at non-obese, non-syndromic children, albeit with a very small sample size of 6 patients. This success rate is well below rates of 80% reported in meta-analyses [6] and 79% reported in the CHAT study [4]. It should be noted that the age group of our cohort was younger than the 5–9 year old children recruited by the CHAT trial. Nevertheless, the improvement in AHI score with surgery was significantly greater than that seen with observation; this was still present when accounting for differences in baseline AHI, though to a lesser degree which may be owing to the small sample size involved. Only 11% of patients in our observation group showed resolution of OSA, also a much lower percentage of patients than the 46% rate of resolution with observation seen in the CHAT study amongst slightly older children (average age 6.5 years vs. 4.5 years in our group) [4]. Not only did children who underwent watchful waiting not resolve their OSA, but the average AHI of the observation group worsened by 2.09 points over the course of the study in all patients and by nearly 4 points in non-obese, nonsyndromic patients. While the latter AHI values are somewhat skewed by a few children who worsened significantly, overall these results support the findings of Li et al. that mild OSA in the majority of pediatric patients does not resolve spontaneously, and
Table 3 Multivariate linear regression analysis with log-transformed AHI as the dependent variable, surgery vs. observation as the independent variable and baseline AHI as a covariate.
Fig. 4. Polysomnography data for syndromic patients with mild OSA. Legend: Bar graph showing mean AHI in events per hour with standard error measurements for 15 syndromic patients. There was no significant difference between the surgery and observation groups (p = 0.36).
Variable
Model parameter estimate (SE)
Medians ratio (95% CI)
p-Value
Baseline AHI value Surgery vs. observation
0.066 (0.038) 0.249 (0.126)
0.86 (0.72–1.02) 0.56 (0.32–0.99)
0.09 0.05
SE, Standard Error. CI, Confidence Interval.
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
that a significant proportion of these patients will get objectively worse over the course of 1–2 years [13]. Although adenotonsillectomy was not 100% effective in resolving mild OSA, surgery at a younger age normalized PSG findings in a significant proportion of patients, improved the average number of apneic events in nearly all patients, and seemed to protect patients from worsening of OSA over time. The lower rate of normalization of OSA in both our surgery and observation groups when compared to the CHAT trial may be attributed to a selection bias inherent to our retrospective review, in that it was likely more common for children to be referred for a second PSG if they were still symptomatic, while children who had improvement in symptoms did not undergo a repeat PSG and are not represented in this analysis. While the degree of bias is difficult to quantify, an observational comparison of the trends over time remains of interest. In our study, improvement with adenotonsillectomy in obese patients with mild OSA was not as pronounced as it was in healthy children. The mean AHI in obese children improved by an average of 0.75 despite having a higher initial AHI than non-obese patients, and only 13% showed resolution of OSA on subsequent PSG. Previous studies have suggested that pediatric obesity is a risk factor for persistence of OSA after adenotonsillectomy [15]. In a meta-analysis of sleep parameters after adenotonsillectomy in obese children, Costa et al. found that, while the mean AHI was significantly reduced and the oxygen saturation nadir increased after surgery, only 25% of patients had a postoperative AHI <2, although the starting AHI was much higher than in our study [16]. In the CHAT trial, while obese children also showed lower rates of resolution of OSA, 67% of obese patients normalized their AHI after adenotonsillectomy compared to 29% of observed patients, prompting the investigators to support a strategy of early adenotonsillectomy in obese children as well [4]. Interestingly, in our cohort, obese children with mild OSA who were observed did not show as severe of worsening of OSA over time as was seen in non-obese patients, and there was no significant difference in the change in AHI between obese patients who underwent surgery and those who were observed. Given the decreased rates of normalization of OSA in obese patients, this finding may suggest a role for watchful waiting in obese children with mild OSA. Non-invasive ventilation such as continuous positive airway pressure (CPAP) may be an alternative option in older children, although patient tolerance and compliance need to be monitored. Nasal steroids and oral montelukast have also shown improvement in children with mild OSA and can be considered [17–19]. Patients should be followed closely to ensure that symptoms don’t worsen, and weight loss programs should be implemented. Children with craniofacial anomalies or hypotonia can be much more difficult to treat for OSA due to several different anatomic sites of obstruction. Recent reviews have shown over 80% of patients with congenital craniofacial malformations referred for a PSG had evidence of OSA; adenotonsillectomy resulted in normalization of AHI in only 14–15% of these patients [20,21]. In our cohort, there was a 33% rate of resolution of mild OSA with a mean improvement of 1.05 in postoperative AHI, however there was no statistically significant change in AHI when compared to the observation group. The benefit of a modest improvement in OSA severity must be weighed against the risk of anesthesia and postoperative complications in syndromic patients, which may be substantial. It should be noted that, in our study, we use AHI as the sole objective measure of pediatric OSA while other large-scale trials have reported on several different PSG criteria [4]. While inclusion of O2 and CO2 data would be of interest, we are unsure of the utility of this in our patient population that a priori has mild OSA. Subsequent papers from the CHAT trial investigating overnight
29
end-tidal carbon dioxide (EtCO2) levels show weak to modest correlations with other polysomnographic indices, and in adjusted analyses, baseline total sleep time with EtCO2 >50 mmHg did not predict postoperative changes in cognitive and behavioral measurements [22]. In another paper from the trial, while oxygen criteria were not used to define OSA resolution, they were included in the analyses of predictors which concluded that lower snoring score, low AHI and normal waist circumference were the only independent predictors for spontaneous resolution of OSA [23]. Thus, it is our experience that, of the parameters recorded in the PSG, the AHI is the most commonly used measure of severity in mild cases of OSA. For this reason, we chose to restrict our analyses to AHI data alone. However, other factors need to be taken into account before a decision regarding surgery or observation for mild OSA can be made, including patient comorbidities, quality of life measures, and age, amongst others. AHI by itself is not the only indicator of the extent of disease burden and should not be used as the sole factor in decisionmaking. Our study has limitations that deserve mention. This review is purely retrospective and thus is limited by the quality of the medical records. The lack of randomization in the study likely lead to increased rates of follow-up and subsequent PSGs in the more symptomatic patients, including obese and syndromic patients who were at higher risk for persistent OSA. Clinicians may have been more likely to obtain a second PSG in the persistently symptomatic patient than in the child whose symptoms have resolved. In addition, the small sample size in the study meant that not only were many of the differences seen not highly statistically significant, but also there may have been differences between the groups that were not detected due to a lack of statistical power. A prospective, randomized trial comparing adenotonsillectomy and watchful waiting in children with mild OSA across a broad age range is needed in order to confirm or refute our results. Despite these limitations, our review demonstrates important observations regarding the natural course of mild OSA and the impact of adenotonsillectomy in healthy, obese, and syndromic children. 5. Conclusions This review shows that adenotonsillectomy for mild OSA in the pediatric population resulted in improvement in postoperative AHI most markedly seen among non-obese, non-syndromic children, while an observational approach not only didn’t lead to normalization of AHI in the majority of patients, but actually resulted in worsening of AHI values on subsequent PSG in several patients. Prospective, randomized trials are needed to further clarify the course of and the effect of surgical intervention for pediatric mild OSA. Previous presentation Preliminary data from this study were presented as a podium presentation at the American Society of Pediatric Otolaryngology (ASPO) Annual Spring Meeting on April 26, 2015 in Boston, MA. Disclosures The authors report no conflicts of interest to disclose. Acknowledgements Samuel J. Trosman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
30
S.J. Trosman et al. / International Journal of Pediatric Otorhinolaryngology 83 (2016) 25–30
References [1] J.C. Lumeng, R.D. Chervin, Epidemiology of pediatric obstructive sleep apnea, Proc. Am. Thorac. Soc. 5 (2) (2008) 242–252. [2] M.S. Schechter, Technical report: diagnosis and management of childhood obstructive sleep apnea syndrome, Pediatrics 109 (4) (2002) e69. [3] R.F. Baugh, S.M. Archer, R.B. Mitchell, R.M. Rosenfeld, R. Amin, J.J. Burns, et al., Clinical practice guideline: tonsillectomy in children, Otolaryngol. Head Neck Surg. 144 (Suppl. 1) (2011) S1–S30. [4] C.L. Marcus, R.H. Moore, C.L. Rosen, B. Giordani, S.L. Garrett, H.G. Taylor, et al., A randomized trial of adenotonsillectomy for childhood sleep apnea, N. Engl. J. Med. 368 (25) (2013) 2366–2376. [5] R.B. Mitchell, Adenotonsillectomy for obstructive sleep apnea in children: outcome evaluated by pre and postoperative polysomnography, Laryngoscope 117 (10) (2007) 1844–1854. [6] S.E. Brietzke, D. Gallagher, The effectiveness of tonsillectomy and adenoidectomy in the treatment of pediatric obstructive sleep apnea/hypopnea syndrome: a meta-analysis, Otolaryngol. Head Neck Surg. 134 (6) (2006) 979–984. [7] P.G. Volsky, M.A. Woughter, H.A. Beydoun, C.S. Derkay, C.M. Baldassari, Adenotonsillectomy vs. observation for management of mild obstructive sleep apnea in children, Otolaryngol. Head Neck Surg. 150 (1) (2014) 126–132. [8] R.B. Berry, R. Budhiraja, D.J. Gottlieb, D. Gozal, C. Iber, V.K. Kapur, et al., Rules for scoring respiratory events in sleep: update of the 2007 AASM manual for the scoring of sleep and associated events. Deliberations of the sleep apnea definitions task force of the American Academy of Sleep Medicine, J. Clin. Sleep Med. 8 (5) (2012) 597–619. [9] S.K. Ng, D.L. Lee, A.M. Li, Y.K. Wing, M.C. Tong, Reproducibility of clinical grading of tonsillar size, Arch. Otolaryngol. Head Neck Surg. 136 (2) (2010) 159–162. [10] C.L. Ogden, R.J. Kuczmarski, K.M. Flegal, Z. Mei, S. Guo, R. Wei, et al., Centers for Disease Control and Prevention 2000 growth charts for the United States: improvements to the 1977 National Center for Health Statistics version, Pediatrics 109 (1) (2002) 45–60. [11] S.L. Calhoun, S.D. Mayes, A.N. Vgontzas, M. Tsaoussoglou, L.J. Shifflett, E.O. Bixler, No relationship between neurocognitive functioning and mild sleep disordered breathing in a community sample of children, J. Clin. Sleep Med. 5 (3) (2009) 228–234.
[12] A. Akcay, C.O. Kara, E. Dagdeviren, M. Zencir, Variation in tonsil size in 4- to 17year-old schoolchildren, J. Otolaryngol. 35 (4) (2006) 270–274. [13] A.M. Li, C.T. Au, S.K. Ng, V.J. Abdullah, C. Ho, T.F. Fok, et al., Natural history and predictors for progression of mild childhood obstructive sleep apnoea, Thorax 65 (1) (2010) 27–31. [14] R.B. Mitchell, J. Kelly, Quality of life after adenotonsillectomy for SDB in children, Otolaryngol. Head Neck Surg. 133 (4) (2005) 569–572. [15] R. Bhattacharjee, L. Kheirandish-Gozal, K. Spruyt, R.B. Mitchell, J. Promchiarak, N. Simakajornboon, et al., Adenotonsillectomy outcomes in treatment of obstructive sleep apnea in children: a multicenter retrospective study, Am. J. Respir. Crit. Care Med. 182 (5) (2010) 676–683. [16] D.J. Costa, R. Mitchell, Adenotosillectomy for obstructive sleep apnea in obese children: a meta-analysis, Otolaryngol. Head Neck Surg. 140 (4) (2009) 455–460. [17] L. Kheirandish, A.D. Goldbart, D. Gozal, Intranasal steroids and oral leukotriene modifier therapy in residual sleep-disordered breathing after tonsillectomy and adenoidectomy in children, Pediatrics 117 (1) (2006) e61–e66. [18] L. Kheirandish-Gozal, D. Gozal, Intranasal budesonide treatment for children with mild obstructive sleep apnea syndrome, Pediatrics 122 (1) (2008) e149–e155. [19] A.D. Goldbart, S. Greenberg-Dotan, A. Tal, Montelukast for children with obstructive sleep apnea: a double-blind, placebo-controlled study, Pediatrics 130 (3) (2012) e575–e580. [20] M. Moraleda-Cibria´n, S.P. Edwards, S.J. Kasten, S.R. Buchman, M. Berger, L.M. O’Brien, Obstructive sleep apnea pretreatment and posttreatment in symptomatic children with congenital craniofacial malformations, J. Clin. Sleep Med. 11 (1) (2015) 37–43. [21] S.O. Zandieh, B.L. Padwa, E.S. Katz, Adenotonsillectomy for obstructive sleep apnea in children with syndromic craniosynostosis, Plast. Reconstr. Surg. 131 (4) (2013) 847–852. [22] S. Paruthi, C.L. Rosen, R. Wang, J. Weng, C.L. Marcus, R.D. Chervin, et al., End-tidal carbon dioxide measurement during pediatric polysomnography: signal quality, association with apnea severity, and prediction of neurobehavioral outcomes, Sleep 38 (11) (2015) 1719–1726. [23] R.D. Chervin, S.S. Ellenberg, X. Hou, C.L. Marcus, S.L. Garetz, E.S. Katz, et al., Prognosis for spontaneous resolution of OSA in children, Chest 148 (5) (2015) 1204–1213.