Introduction

Pediatric sleep-disordered breathing (SDB) is common and represents a continuum of respiratory disturbances during sleep in children, ranging from primary snoring to obstructive sleep apnea (OSA)1,2. OSA is characterized by frequent episodes of airway obstruction and sleep fragmentation, disrupting the growth hormone/insulin-like growth factor-I axis (GH-IGF-1) and contributing to growth impairment3,4. Pediatric OSA confirmed by polysomnography has historically been estimated to affect 1%–6% of children2,5; however, questionnaire-based studies over the past decade have reported a higher prevalence of approximately 13%–20% among preschool children6. Adenotonsillectomy (AT) is the first-line treatment in pediatric OSA1,2. A significant increase in growth hormone secretion has been observed following AT, often accompanied by catch-up growth7,8.

Mild SDB is manifested as primary snoring, without frequent airway obstruction, sleep fragmentation, or oxygen desaturation, and affects approximately 5.75% to 9.61% of children2,5,9. Clinically, primary snoring, which is difficult to distinguish from mild OSA10, is characterized by habitual snoring occurring more than three nights per week with an obstructive apnea-hypopnea index (OAHI) of < 15, 9, 11. In contrast to OSA, primary snoring has received relatively little attention in research, as it is often regarded as a benign condition that does not significantly impact children’s growth12. However, increasing evidence has suggested that primary snoring might be associated with several adverse health outcomes13,14,15. Despite this, no standardized treatment protocol exists to date for primary snoring or mild SDB2. As a result, further research is needed to evaluate the potential benefits of AT on growth and sleep parameters in children with mild SDB.

The present study utilized data from the Pediatric Adenotonsillectomy Trial for Snoring (PATS), which is a randomized controlled trial (RCT) investigating children with mild SDB16,17,18. The primary outcomes of PATS are changes in executive behavior and vigilance; in the present study, we focused on growth outcomes not examined in the original study17. We hypothesized that AT provides growth benefit for children with mild SDB, in addition to improving sleep quality.

Methods

Data source

PATS is a prospective, multicenter, randomized, and single-blinded interventional study aimed to evaluate the short-term impact of AT in children with mild SDB17. The study included 459 children aged 3–12 years who presented symptoms of snoring and mild pediatric sleep apnea (OAHI ≤ 3), namely that they had at least three nights of snoring per week in the absence of oxygen saturation below 90%. Exclusion criteria included a history of tonsillectomy, recurrent tonsillitis, severe obesity (i.e., body mass index [BMI] z-score > 3), chronic conditions other than asthma, and significant developmental disorders. Children were randomized to either AT, performed within four weeks of randomization, or watchful waiting with supportive care, and were followed for 12 months after surgery or randomization. Blinded research staff conducted all assessments at randomization and at the 12-month follow-up. PATS was approved by the Institutional Review Board at the Children’s Hospital of Philadelphia (Approval No. 14-011214) and registered at ClinicalTrials.gov (ID: NCT02562040).

Study design

In the present study, we included data from 330 children who consented to share their data. Two analyses were performed, including one on growth outcomes and one on sleep parameters. Participants with missing data on the outcomes were excluded, resulting in a final sample of 272 children for the analysis of growth outcomes and 232 children for the analysis of sleep parameters, respectively.

Outcome measurements

The growth outcomes included height, weight, and BMI, expressed in sex- and age-specific percentiles, in addition to absolute values and age- and sex-adjusted z-scores19. Sleep parameters were measured by polysomnography, which was conducted at a pediatric research center and in absence of acute illness. Full-night polysomnography followed the AASM Manual (Version 2.2, pediatric standards) and was scored using Compumedics software. EDF/XML signals from the E-Series system were analyzed by blinded, certified technicians at Brigham and Women’s Hospital. OAHI was defined as the sum of events including obstructive apnea, mixed apnea, hypopnea with 3% drop in blood oxygen saturation, and arousal. In the present study, we included arousal index, respiratory disturbance index (RDI), peak End-tidal carbon dioxide (EtCO2), oxygen desaturation index (ODI), average oxygen saturation, and OAHI as sleep parameters. We also studied outcomes related to sleep quality, including sleep efficiency, sleep maintenance efficiency, sleep time, wake time after sleep onset, sleep latency, rapid eye movement (REM) sleep latency, and sleep architecture.

Statistical analysis

To compare changes in study outcomes between the AT and the watchful waiting groups at the 12-month follow-up, we applied linear regression after adjusting for the corresponding outcomes measured at randomization. We analyzed the study outcomes as percentiles in the main analysis and as absolute values or z-scores in the sensitivity analyses. To examine whether the results on BMI change would differ according to baseline BMI, we separately analyzed children with a relatively low (< 85th percentile) or high (≥85th percentile) BMI at randomization. To assess whether the results on growth outcomes would differ by age, sex, and other characteristics, we performed stratified analyses by age, sex, tonsil grading, history of asthma, and sleep parameters. As no overall association was noted between AT and BMI, the stratified analyses were performed for height and weight percentiles only. To assess whether the associations of AT with growth outcomes and sleep parameters would differ for children with primary snoring, we also conducted several sensitivity analyses restricted to children with an OAHI of < 1. Finally, while the original study of PATS adhered to the intention-to-treat (ITT) principle, our secondary analysis utilized an as-treated analytical approach, which offers a more nuanced representation of the actual clinical intervention effect20,21,22. To ensure methodological rigor given the presence of crossover (i.e., 5.2%), we further conducted an ITT analysis as another sensitivity analysis.

To make the results more clinically interpretable, regression coefficients (β) were converted into mean differences with 95% CIs, reflecting the changes in the outcome variables in relation to the exposure variable. All analyses were performed using IBM SPSS Statistics (version 27). A 2-sided p < 0.05 was considered statistically significant.

Results

Table 1 summarizes the baseline characteristics of the 330 children included in the present study, including 161 boys (48.8%) and 169 girls (51.2%), with a median age of 6.4 years. 233 (70.8%) of these children were white, and 134 (43.2%) were from households with an income of below $50,000. Asthma was reported in 83 (25.2%) children. Tonsillar hypertrophy was classified as Grade II in 123 children (37.3%), Grade III in 186 children (56.4%), and Grade IV in 21 children (6.4%). Additionally, 231 (70.0%) children had an OAHI of < 1 (median, 0.6; IQR, 0.3–1.1).

Table 1 Baseline characteristics of the study participants.
Full size table

After adjustment for outcomes measured at randomization, children in the AT group showed an increase of 2.74 percentiles in height (mean difference: 2.74, 95% CI: 0.33 to 5.15) and 2.79 percentiles in weight (mean difference: 2.79, 95% CI: 0.29 to 5.28) compared with the watchful waiting group at the 12-month follow-up (Table 2). Children in the AT group also showed an increase of 2.37 percentiles in BMI (mean difference: 2.37, 95% CI: -0.83 to 5.56) compared with the watchful waiting group at the 12-month follow-up, although the difference did not reach statistical significance (p = 0.146). The difference was mainly noted among children with a low baseline BMI (n = 207), whereas no clear difference was noted among children with a high baseline BMI (n = 123). Similar results were noted when studying absolute values or z-scores of height, weight, or BMI (Supplementary Tables 1 and 2). When the analysis was restricted to children with primary snoring, comparable results were observed (Supplementary Table 3).

Table 2 Associations of adenotonsillectomy with mean change in height percentile, weight percentile, and BMI percentile.
Full size table

Compared to the watchful waiting group, the AT group also showed greater improvement in sleep parameters, including OAHI (mean difference: -1.36, 95% CI: -2.07 to -0.65), RDI (mean difference: -0.47, 95% CI: -0.74 to -0.20), and ODI (mean difference: -0.99, 95% CI: -1.85 to -0.13) (Table 3). While better improvement was also observed in peak EtCO₂ and average oxygen saturation, these differences did not reach statistical significance. Regarding sleep architecture, we observed reduced Stage 1 sleep percentage (mean difference: -1.00, 95% CI: -1.90 to -0.11) and increased Stage 2 sleep percentage (mean difference: 2.06, 95% CI: 0.54 to 3.59) in the AT group. However, no statistically significant difference was noted for other outcomes related to sleep quality, likely due to limited power. Similar results were noted when restricting the analysis to children with OAHI < 1 (Supplementary Table 4).

Table 3 Associations of adenotonsillectomy with mean change in sleep parameters.
Full size table

In the stratified analysis, we found that the gain in height percentile in the AT group was mostly noted in children at 3–5 years (mean difference: 7.88, 95% CI: 2.71 to 13.04), but not in older children (> 5–8 or > 8–13 years) (Table 4). A higher point estimate was noted in boys than girls (mean difference: 3.84 vs. 1.53), as well as among children with grade III&IV tonsils (mean difference: 5.40; 95% CI: 2.55 to 8.26), non-asthmatic children (mean difference: 4.24; 95% CI: 1.60 to 6.88), children with ODI ≥ 1 (mean difference: 3.67, 95% CI: 0.66 to 6.68), and children with RDI ≥ 0.3 (mean difference: 3.30, 95% CI: 0.48 to 6.12), compared with other children. Similar results were noted in the stratified analyses for weight percentile, although gain in weight percentile appeared to be more pronounced among children with ODI < 1 or RDI < 0.3 (Supplementary Table 5). Finally, the sensitivity analyses using ITT approach rendered largely similar results (Supplementary Table 6).

Table 4 Associations of adenotonsillectomy with mean change in height percentile - stratified analysis by age, sex, tonsil grade, OAHI, asthma, ODI, or RDI.
Full size table

Discussion

To the best of our knowledge, the present study is the first to examine the influence of AT on the improvement of growth, in addition to sleep-related outcomes, in children with mild SDB. We showed that children with AT had higher height and weight, as compared to the watchful waiting group, at 12 months post-surgery, in addition to a general improvement in sleep parameters. Specifically, apart from reduction in OAHI, significant improvement was also noted in RDI and ODI. Interestingly, we found a significant difference in sleep architecture between the groups, including a decrease in Stage 1 and an increase in Stage 2 sleep among children with AT.

Previous studies often focused on children with OSA or lacked clear documentation of surgical indications, and the findings are inconsistent. A retrospective observational study of 815 children who underwent AT found increased weight and BMI but not height at 18 months post-surgery23. A RCT reported comparable findings although with a smaller sample size24. In contrast, more studies have demonstrated a beneficial effect of AT on both height and weight25,26,27. A comprehensive meta-analysis summarizing findings from 14 studies reported significant increase in both height and weight, as well as levels of IGF-1 and insulin-like growth factor-binding protein-3 (IGFBP-3) following tonsillectomy7. However, research on post-surgical growth in children with mild SDB is limited. Through a secondary analysis of data from a high-quality RCT, this study extends previous findings by demonstrating that AT promotes growth in both height and weight, even in children with mild SDB.

Despite limited research, AT hasbeen repeatedly shown to improve sleep-related outcomes and symptoms in children with mild SDB16. In one study, children with primary snoring who underwent AT were shown to demonstrate improved scores in Sleep-related Breathing Disorder scale at 5-year follow-up28. Another study showed that, even in children with mild SDB primarily characterized by tachypnea, postoperative improvement was observed in sleep-related outcomes, such as daytime fatigue, sleep terror, sleepwalking, and enuresis29. Our study adds on to these studies by showing that, in addition to the known sleep benefits, AT might also improve sleep-related respiratory parameters and sleep architecture among children with mild SDB. Previous studies have indeed shown that, while the sleep architecture is not significantly different between children with primary snoring and other children, Stage 1 sleep is positively associated with the severity of SDB30. Similar to our findings, the CHAT study also reported a reduction in Stage 1 sleep following AT among children with mild to moderate OSA, which indicates an improvement in SDB31.

Although mild SDB does not involve frequent airway obstruction, arousals, or ventilation abnormalities, it includes abnormal breathing patterns such as flow limitations and tachypnea, increased respiratory effort, and alterations in sleep microstructure11,29,32,33. These changes may lead to increased sleep instability and fragmentation which could subsequently disrupt the secretion pattern of growth hormones3. Mild SDB can also lead to a small number of events where blood oxygen levels drop by more than 3%. Children might be particularly sensitive to such intermittent hypoxic episodes as they can directly influence the release of neurotransmitters in the central nervous system and alter the hypothalamic-pituitary axis and peripheral endocrine glands, thereby reducing the release and biosynthesis of growth hormones4,34,35,36. As a result, adenotonsillectomy may promote growth by alleviating upper airway obstruction, which in turn ameliorates nocturnal hypoxia and enhances sleep quality, as evidenced by the significant improvement in key polysomnographic parameters (e.g., OAHI, RDI, and particularly ODI) observed in the present study, thereby creating favorable physiological conditions for growth hormone secretion such as elevated circulating levels of IGF-1 and IGFBP-37,8,37. Systemic inflammation associated with SDB can suppress the GH-IGF-1 axis25,38,39,40, whereas AT has been shown to alleviate chronic low-grade inflammation and relieve the inhibition of the GH-IGF-1 axis25,41,42,43. Furthermore, a reduced respiratory effort may lower energy expenditure, allowing more energy to be directed toward anabolic processes and growth44,45,46. Finally, overall improvements in health status may also contribute to better growth outcomes16,28.

The growth benefit following AT appeared to be more pronounced among younger children, boys, children with grade III&IV tonsils, and non-asthmatic children. Young age and adenotonsillar hypertrophy have been consistently identified as major risk factors for pediatric airway obstruction47,48. Further, the relatively narrow upper airway in young children makes tonsils and adenoids the primary anatomical basis of OSA in this age group49,50. Consequently, AT in young children with DSB may more effectively relieve airway obstruction, improve overall sleep quality, and create favorable physiological conditions for growth. Children with asthma, a condition characterized by chronic airway inflammation and corticosteroid use, often exhibit impaired height growth51,52. As a result, the presence of asthma might counteract the effect of AT in modulating growth. We observed a greater gain in height and weight following AT among boys than girls; however, whether there is indeed a sex-specific growth benight needs to be examined further. Finally, the potential impact of AT on weight gain in overweight or obese children continues to be a subject of debate23,53,54. In the present study, we observed little association between AT and BMI change in children with high BMI at baseline (≥85th percentile), in line with several recent studies55,56.

Strengths and limitations

Our study is based on data from a RCT with a relatively large sample size, alleviating most concerns on systematic and random errors like biases and confounding in observational studies. Our study is novel as it focused on children with mild SDB, a population that has been relatively underexplored in prior studies. However, the follow-up period of PATS was short, which limits our ability to assess the long-term impact of AT on growth. Additionally, the sample size was small in some of subgroup analyses, making false negative findings a concern due to limited statistical power. Moreover, we assessed growth using only anthropometric measures (height and weight) and did not measure growth biomarkers such as IGF-1 or IGFBP-3, limiting our ability of exploring the underlying mechanisms of post-AT growth. Finally, the study included children from one specific region, limiting the generalizability of the findings to other populations.

Conclusion

AT improved growth and sleep quality in children with mild SDB, suggesting that surgical intervention may offer additional health benefits beyond symptom relief, particularly in younger children with tonsillar hypertrophy.