The Effect of Adenotonsillectomy on Children’s Behavior and Cognitive Performance with Obstructive Sleep Apnea Syndrome: State of the Art

(1) Background: This systematic review was designed to analyze adenotonsillectomy’s role in treating behavioural disorders and sleep-related quality of life in pediatric OSAS. (2) Methods: Papers that report pre-operative and post-operative outcomes by using the Epworth sleepiness scale, OSA-18, NEPSY, Conners’ rating scale, BRIEF, PSQ-SRBD, PedsQL and CBCL. We performed a comprehensive review of English papers published during the last 20 years regarding behavioural disorders in OSAS patients and adenotonsillectomy. (3) Results: We included 11 studies reporting behavioral outcomes and sleep related quality of life after surgery. We investigated changes in behavior and cognitive outcomes after AT, and we found significant improvements of the scores post-AT in almost all studies. After comparing the AT group and control group, only one study had no difference that reached significance at one year post-AT. In another study, it did not show any significant improvement in terms of all behavioural and cognitive outcomes. The questionnaires on sleep-related quality of life after AT (PSQ-SRBD or ESS or OSA-18 or KOSA) may improve with positive changes in sleep parameters (AHI, ODI and SpO2). Furthermore, there is a significantly higher decrease in OSAS symptoms than the pre-AT baseline score. (4) Conclusion: Future studies should pay more attention to characterizing patient populations as well as rapid surgical treatments through existing criteria.


Introduction
Obstructive Sleep Apnea Syndrome (OSAS) is a condition characterized by repeated episodes of complete (apnea) or partial (hypopnea) cessation of airflow through the upper airways during sleep [1,2]. The prevalence of respiratory sleep disorders in preschool and school-age varies widely, with an estimated rate of primary snoring in children ranging from 8% to 27% and of OSAS from 1% to 5% [3][4][5][6][7][8].
On the other hand, in a retrospective study, Gozal et al. analyzed 797 subjects with low performance (LP) and 791 subjects with high performance (HP) among seventh and eighth graders attending public schools by using questionnaires. The authors reported snoring in early childhood in 103 LP children versus 40 in HP children (5.1%; OR: 2.79; confidence interval (CI): 1.88-4.15; p, 0.00001), with AT surgical intervention in 24 LP and 7 HP children (odds ratio: 3.40; confidence interval: 1. 47-7.84). These data suggested that neurocognitive morbidity may only be partially reversible after treatment and that residual deficits in the learning process could still remain many years after snoring has resolved. [33].
Moreover, Kohler and colleagues assessed by means of the Stanford Binet Intelligence Scale 5th edition, Neuropsychological Developmental Assessment (NEPSY) and polysomnography a total of 44 healthy snoring children (aged 3-12 years) at baseline and 6 months after adenotonsillectomy and reported the comparison with 48 age and gender matched non-snoring controls. In this case, neurocognitive deficits were reported at baseline in snoring children when compared to controls (10 point IQ difference, with p = 0.001, Cohen's d = 0.929 (large effect size) for full scale IQ); however, neurocognitive deficits did not improve 6 months after surgery relative to controls (Cohen's d = 0.137 (trivial effect size), although the range in frequency of desaturation was extremely reduced (from 0-53.1 to 0-5.6) [24].
In light of such heterogeneous results, it seemed reasonable to analyze literature data of the last 20 years on pediatric OSAS patients and provide a systematic review about the current correlation between AT and neurocognitive/behavioral disorders.
In particular, we examined the efficacy of AT on behavior through the use of validated questionnaires on sleep parameters compared from baseline to follow up after AT and, whenever possible, to subjects not treated surgically with AT (WWSC (Watchful Waiting with Supportive Care group) or control group).

Protocol Data Extraction
According to the PRISMA checklist for review and meta-analysis, we performed a systematic review of the current literature [55] (Figure 1), and this review protocol was registered on the International Prospective Register of Systematic Reviews (PROSPERO; registration number: 277325).
The authors P.DM and I. LM searched the Medline database via PubMed, EMBASE and Cochrane library from January 2001 to April 2021, solving any disagreements among the study members through a discussion.
We examined all the studies included, analyzing all available data and guaranteeing eligibility for all subjects. Main patient features, symptoms, diagnostic procedures, treatment modalities, outcomes scores and follow-up were collected. In order to analyze sleep quality, we analyzed data from AHI (

Electronic Database Search
PubMed/Medline, Embase, Web of Science, Scholar and the Cochrane Library electronic databases were searched for studies on adenotonsillectomy in OSA pediatric patients and neurocognitive and behavioral disorders over the last 20 years of literature (from 1 July 2001 to 1 July 2021) by two different authors. We used the following search keywords: "OSAS", "Obstructive Sleep Apnea Syndrome", "Sleep-Disordered Breathing", ''adenotonsillectomy", "cognitive disorders," "behavior", "neurocognitive function" and "quality of life".
All the papers' titles and abstracts available in the English language were analyzed; thus, we identified full-text articles screened for original data. The search process is summarized in Figure 1.

Inclusion Criteria
Studies that met the following criteria were included: (1) Cross-sectional studies, case controls, retrospective cohort studies, prospective cohort studies, primary science articles and epidemiological studies; (2) Studies regarding children with OSAS treated with adenotonsillectomy; (3) Studies using at least one validated questionnaire on the behavior of children with OSAS before and after adenotonsillectomy.
(4) All the studies reported detailed information on preoperative and postoperative OSA cognitive, behavioral and/or sleep outcomes, such as AHI, ODI, CRS-R, CBCL, NESPI and BRIEF.

Exclusion Criteria Selected
(1) Articles not published in English; (2) Case reports, letters to the editor and reviews; (3) Papers missing preoperative and postoperative continuous data.
The process undertaken is schematically presented in Figure 1.
For each study, we reported the following clinical characteristics: type of behavior and/or sleep questionnaires; correlation of the results of the questionnaires before and after tonsillectomy.

Retrieving Research
According to the PRISMA checklist for review and meta-analysis, we reviewed 210 articles. Before screening, 18 of them were removed because they were duplicate records, and 192 were assessed for eligibility. Of these, however, 117 were removed due to the full-text being unavailable, 40 were removed because they did not analyze our search target, 6 were removed because data were not available and 18 were removed because they were not written in English.
All studies analyzed the efficacy of AT on cognitive or behavioral measures outcomes assessed by validated questionnaires. The main results of Baseline Outcomes in WWSC or control and AT groups are summarized in Table 3. The main outcomes at follow-up and change from baseline to follow-up between groups are summarized in Table 4.

Neurocognitive Performance
Four articles, three RCTs [3,56,59] and one prospective cohort studio [58] evaluated the effect of AT on the results of the neurocognitive performance (Table 4).
In Marcus et al.'s study, the average on the NEPSY scores in comparison between early adenotonsillectomy group and WWSC group showed a difference but was not significant (p-value = 0.16) (Cohen's d = 0.15 (small effect size)) [3].
In Taylor
Marcus et al. reported a significant improvements among early adenotonsillectomy group than among WWSC group in behavioral disorders assessed via the caregiverreported Conners' Rating Scale, the teacher-reported data and the caregiver-reported BRIEF. However, they were not significantly different in terms of the teacher-reported version between the groups (p-value = 0.04) (Cohen's d = 0.29 (medium effect size)) [3].               In Hattiangadi Thomas et al.'s paper, at follow-up, Total Problems (change from baseline to 7 months: −1 (−6, 4), (p < 0.001)) and Internalizing (change from baseline to 7 mounths: −1 (−6, 6), (p = 0.04)) were evaluated. CBCL T-scores reduced more in eAT than WWSC. The eAT group also proved a significantly greater decrease in Thought Problems and Somatic Complaints [57].
In Paruthi et al.'s study, Hypercapnia and Behavioral Outcomes were correlated. The baseline percentage of TST with EtCO2 > 50 mmHg did not correlate with changes on behavioral assessments at follow-up (all p-value > 0.05), even after adjustments for age, sex, race and the treatment assignment (p > 0.3, Spearman: −0.059) [59]. Dillon et al. reported that the frequency of psychiatric disorders among controls changed minimally from baseline to follow-up (Cohen's d = 0.161 (trivial effect size)).
In contrast, the frequency of attention and disruptive behavior disorders in AT children dropped from 36.7% to 23.1% (p = 0.008) (Cohen's d = 0.359 (small effect size)). The overall prevalence of ADHD declined modestly from 27.8% to 20.5% (Cohen's d = 0.246 (small effect size)). ADHD remitted after AT (among 11 patients); rating changes on the DBDRS subscales corresponding to pre-operative diagnosis changed by an average of 51.1% [61].
The mean Pediatric Quality of Life (PedsQL), in which scores range from 0 to 100 and higher scores indicate better quality of life changing from baseline to 7 months, was −0.3 ± 0.2. In particular, 80% of the PedsQL total scores were < 0.33, suggesting high risk for obstructive sleep apnea. PedsQL scores were significantly related with AHI and the ODI at both baseline and follow-up but were not related with higher Etco2 values or other PSG measures. PedsQL (Child) and PedsQL (Parent) improved significantly after adenotonsillectomy (p-value: 0.006, (Cohen's d = −0.23 (medium effect size)) and <0.001 (Cohen's d = −0.37 (medium effect size)), respectively.
In Chervin et al., there were great differences in the behavioral hyperactivity index (p-value 0.003) and cognitive attention index (p-value = 0.020) at baseline compared to AT group and control group. In contrast, none of these differences reached significance at one year (p-value = 0.056 and p-value = 0.133, respectively). The frequency of Attention-Deficit/Hyperactivity Disorder was not significant, and the rates were not different from baselines after AT in both groups enrolled (Fisher's Exact Test, p-value = 0.23) [62].
In Chun T. Au et al., patients of both AT and WW groups had fewer behavioral problems at follow-up as reported by their parents on the CBCL compared to baseline (p < 0.05, Cohen's d = 0.15 (small effect size)).
However, there were no significant between-group differences in the changes of any CBCL score, and there were no significant differences in the changes in CPT parameters, CBCL scores, OSA-18 total score, daytime sleepiness scales and ADHD rating scale between the resolved and residual groups [64].
In Isaiah A. et al., there were no associations identified between any of the polysomnographic parameters, including the AHI, and the behavioral outcomes. Although the parentreported BRIEF MI score demonstrated the greatest improvement (Cohen d effect size = 0.5) in the early AT group, smaller improvements were identified for other parent reported behavioral outcomes (effect size for parent-reported Conners Global Index score was 0.3 in the early AT group). No statistically significant changes were identified for the teacherreported BRIEF MI or the Conners Global Index scores (effect sizes for teacher-reported Conners Global Index and BRIEF MI scores were 0.1 for each outcome) [65].

Sleep Related Quality of Life after Surgery
Six articles, five RCTs [3,56,60,63,65] and one prospective cohort study [64] evaluated the effect of AT on the results of the questionnaires of sleep related quality of life (Table 4).
In Marcus et al.'s study, the symptoms of obstructive sleep apnea syndrome measured with the use of the Epworth Sleepiness Scale, the generic and disease-specific measures of quality of life and the PSQ-SRBD were assessed by means of the PedsQL and OSA-18, respectively.
There was a more significant reduction in symptoms in the eAT group than in the WWSC group. At baseline, the PSQ-SRBD score and PedsQL score were as follows: 0.5 ± 0.2 and 76.5 ± 15.7 in WWSC group; and 0.5 ± 0.2 and 77.3 ± 15.3 in eAT group, with a change from baseline to 7 months after AT of −0.0 ± 0.2 (p < 0.001) and −0.3 ± 0.2 (p < 0.001) in two groups, respectively, for the PSQ-SRBD score and 0.9 ± 13.3 and 5.9 ± 13.6 in two groups, respectively, for the PedsQL score [3].
In Taylor et al.'s study, as in AHI outcomes, the regression analysis detected strong associations of improvement with positive changes in sleep parameters as measured by sleep questionnaires in the eAT group. The associations were weak (partial rs −0.15 to −0.30) and had small effect sizes (f 2 0.022-0.088) [56].
In Isaiah A et al.'s study, the regression models revealed a positive association between the baseline PSQ-SRBD score and the parent-reported severity of behavioral impairment. Another mean (SD) decrease was identified for PSQ-SRBD scores in the early AT group vs.

Discussion
Contrasting data are reported on the effect of adenotonsillectomy for OSAS in children in behavior, cognitive function and quality of life improvements. Garetz et al. suggested that adenotonsillectomy is associated with improvements in these fields above, but new, large, randomized and controlled studies are needed to provide definitive evidence of the benefits of this surgical procedure [66].
In a meta-analysis of Yu et al., after 6-12 months of observation, significant improvements in attention-executive function and verbal ability were found in children with OSAS treated with AT compared to their baseline level. The Hedges' g effect sizes of general intelligence, memory, attention-executive function and verbal ability compared to baseline level were medium (−0.37), medium (−0.36), trivial (−0.02) and medium (−0.45), respectively.
Moreover, the restoration of attention-executive function and memory was observed in children with OSAS after AT in comparison to healthy controls, and they say that rigorous randomized controlled trials should be conducted to obtain definitive conclusions [67].
Our review, which mainly includes RCT, reported better OSAS and sleep-related outcomes in tonsillectomies children, excluding the one that did not report PSG data. The six studies that included children with OSDB confirmed with PSG found that AHI (four studies) or ODI (two studies) or% TST EtCO2> 50 mmHg (2 studies) scores improved more in children receiving a tonsillectomy than in those who did not undergo surgery.
In six studies [3,56,59,[63][64][65], the sleep-related quality of life after AT (PSQ-SRBD or ESS or OSA-18 or KOSA) improved with positive changes in sleep parameters (one study), or showed a significantly higher decrease in OSAS symptoms (one study) or was significantly improved than the baseline score (three studies). It is possible that the reductions in OSA symptoms and improvements observed in quality of life were affected by parental expectations [68].
Investigated changes in behavior and cognitive outcomes after AT were found to improve scores significantly post-AT in all studies. In only one study, comparing the AT group and control group, no differences reached significance at one year after AT and, in another one, did not show a significant improvement in all behavior and cognitive outcomes. In Paruthi et al.'s study [59], the hypercapnia did not correlate with changes on the cognitive and behavioral assessments at follow-up. In Isaiah et al.'s study [61], the results of this study demonstrated that the treatment related changes in behavioral outcomes were causally attributable solely to the changes in parent-reported OSDB severity.
There were no significant correlations between polysomnographic parameters and behavioral outcomes, and neither baseline hypercapnia nor change in EtCO2 levels predicted baseline or change in cognitive and behavioral parameters. We have noted that surgery resulted in more significant improvements in the BRIEF and in the scores on the Conners' Rating Scale, measuring restlessness and impulsiveness and emotional lability, in the eAT group than in the WWSC group, and no significant difference between the two groups in the change from baseline to follow-up in our primary outcome was observed in terms of the attention and executive-function score of the NEPSY. The CBCL analysis confirmed an elevated prevalence of behavioral problems in children with OSAS at baseline. At follow-up, there was a highly significant improvement in Total Problems, Internalizing behaviors, Somatic Complaints and Thought Problems in children randomized to surgery than compared to WWSC. This review has shown a poor relation between neurocognitive/behavioral outcomes and polysomnographic parameters, even if improvements were found in the polysomnographic parameters and behaviors outcomes after AT. This may be because of the roles that affect many polysomnographic variables and other factors that affect behavior.
Of note, neither the PedsSQ nor the AHI predicted objectively measured attention or executive impairment at baseline or change after adenotonsillectomy.
The main practical implication of these results is that surgical candidacy for AT that is solely based on polysomnographic severity of SDB as measured by the AHI or other parameters is an unreliable predictor of behavioral outcomes. Conversely, standardized symptom-based questionnaires (e.g., PSQ-SRBD scale) should be considered as a useful adjunct for predicting behavioral outcomes in children undergoing AT.
Our systematic review, however, has several limitations, which are especially inherent in the nature of the evaluation tool performed for behavioral and cognitive performance assessment. Moreover, the samples enrolled in the indicated studies were not numerous, the study protocols were not adequate or there were no control groups. For this reason, further scientific evidence is needed to report data on the matter.

Conclusions
This review highlights the importance of future screening for behavioral symptoms and quality of life in children who present OSAS and compares behavioral symptoms and quality of life before and after AT as a predictive criterion. Currently, the PSG parameters provide clinicians with limited means to predict the improvement in neurobehavioral morbidity in OSAS.
Despite the need for further research, this review suggests that AT treatment of OSAS can improve the behavioral symptoms of children with OSAS, possibly avoiding the need for psychopharmacological treatment. The clear improvement of patients after AT provides new suggestive evidence for a cause-and-effect relationship between sleep breathing disorders and various negative results on behavioral, cognitive and mental health.
Future studies should pay more attention to characterizing patient populations and should pay attention to using existing criteria such as severity of respiratory disorders and risk factor assessment. Moreover, attention needs to be paid to the evaluation of other comorbidities such as neuro-cognitive disorders in order to distinguish patients who need surgical treatment immediately and patients who can wait.