Paediatric adenotonsillectomy, part 1: surgical perspectives relevant to the anaesthetist

Learning objectives.

By reading this article, you should be able to:

• •Distinguish between obstructive sleep-disordered breathing and obstructive sleep apnoea (OSA) syndrome.
• •Describe the anatomical and neuromotor endotypes of airway causing OSA in children.
• •Compare the indications, efficacy, recovery and morbidity of tonsillotomy with tonsillectomy.
• •Discuss the risk factors for persisting OSA after tonsillectomy in children.

Key points.

• •Obstructive sleep apnoea (OSA) syndrome is a multiorgan systemic inflammatory condition.
• •Obstructive sleep-disordered breathing is the most common indication for adenotonsillectomy (AT).
• •Overnight oximetry can diagnose OSA in children.
• •There is no consensus for the indications, approach or technique of AT to treat OSA.
• •Adenotonsillectomy normalises OSA sleep study variables in approximately 80% of children; the long-term treatment benefits of AT for OSA and sore throat are unclear.

Adenotonsillectomy (AT) is one of the most common surgeries performed in children worldwide. Obstructive sleep-disordered breathing (oSDB), a spectrum in which obstructive sleep apnoea (OSA) is an extreme form, is the main indication for AT and is typically performed as an outpatient. More than 530,000 children undergo AT each year in the USA, representing 16% of all paediatric surgeries. The number of ATs in the USA is 1.5–3 times that of other countries, including Canada and the UK. Within the USA, the rates of AT can vary up to four-fold.¹ As a common surgery associated with a high degree of variability in both practice and outcomes, evidence-based standardised care is desirable.

A lack of consensus on the criteria and severity of disease warranting surgery may explain the variation in practice. In the case of OSA, polysomnography (PSG) is the gold standard for diagnosis, but there is no consensus on the degree of abnormality on PSG that warrants surgery or criteria that increase the risk of perioperative respiratory adverse events (PRAEs). Further, because of limited access and cost, most children do not undergo PSG before AT. Variability in admission practice is widespread.² In the UK, only one in seven children who meet the Paradise guideline criteria for recurrent tonsillitis undergoes surgery.^1,3

Not surprisingly, inconsistent perioperative care translates into variable rates of morbidity and mortality, and high rates of revisiting and readmission to hospital after AT.⁴ Complications are mainly preventable, including pain, nausea/vomiting and dehydration.^4,5 Reported 30-day all-cause mortality rates also vary by country. Mortality rates (per 10,000) in the USA are from 0.6 (outpatients) to 4 (inpatients).^5,6 In Canada and the UK, the combined in-/outpatient mortality rates are 0.2 and 0.3, respectively.^4,7 Deaths have been attributed to both haemorrhage and respiratory events occurring both in, and more frequently, out of hospital.⁸

Reviews and related evidence-based guidelines exist for children undergoing AT.^3,9, 10, 11, 12, 13, 14 However, physicians generally report being overwhelmed by the large number of guidelines and their sometimes contradictory content. The purpose of this two-part review is to discuss controversies for anaesthetists caring for children undergoing AT. Discussion of the indications for surgery here is limited to oSDB and recurrent tonsillitis. The patient's and surgical characteristics are detailed here; considerations for anaesthesia are discussed in Part 2. The combined references in Parts 1 and 2 represent key publications, including a recently published UK guideline report. Further detailed references are available on request from the corresponding author.

Definitions

Obstructive sleep-disordered breathing

Obstructive SDB is defined as a syndrome of upper airway dysfunction during sleep characterised by snoring and/or increased respiratory effort that results from increased upper airway resistance and pharyngeal collapsibility.¹⁰ Adenotonsillar hyperplasia is the prototypical cause for oSDB in children. Obstructive sleep apnoea is the most severe form within the spectrum of oSDB, and is diagnosed based on attended overnight PSG, age, severity of symptoms and associated comorbidities. Obstructive sleep apnoea syndrome implies associated end-organ dysfunction, including cardiopulmonary disease, metabolic dysfunction and neurocognitive and behavioural disorders. The estimated prevalence of OSA is 1–5%, although the prevalence in syndromic children can be >50%.¹⁴ Severity of disease along the spectrum of oSDB is reflected by increasing frequency and degree of upper airway obstruction, non-restorative/fragmented sleep, and abnormal oxygenation and ventilation.

Tonsillitis

Recurrent acute tonsillitis refers to repeated episodes of a sore throat caused by viral or bacterial infection of the pharynx, palatine tonsils, or both. A distinction is made between positive and negative Group A, C, or G type of β-haemolytic streptococcus (BHS) infection because of concerns over associated morbidity. Associated Group A BHS infection accounts for approximately 5% of related medical consultations.³

Tonsillectomy

Extracapsular tonsillectomy consists of complete removal of the palatine tonsils and its surrounding capsule. Surgery typically includes excision of a portion of the anterior tonsillar pillar. It is increasingly popular practice to debulk the tonsil whilst leaving the thin surrounding capsule intact. The terminology used refers to the amount of tonsil removed: the exophytic portion (tonsillotomy), most of tonsil (intracapsular partial tonsillectomy) and all of the tonsil (intracapsular complete tonsillectomy).¹⁵

Disease burden

Obstructive sleep apnoea is associated with low socioeconomic status.^9,14 Children with OSA are frequent users of healthcare resources, and have excessive use of antibiotics for frequent throat and respiratory tract infections and asthma. If left untreated, OSA can change a patient's lifetime health course and can cause premature death.¹⁶ Associated behavioural disorders and systemic inflammation predispose these children to increased reports of head injury and early onset atheromatous cardiovascular disease, respectively. Even children with primary snoring (mild oSDB) show impaired behaviour (e.g. hyperactivity), physical symptoms (e.g. difficulty swallowing food), mild neurocognitive defects and reduced quality of life (QOL).^9,14

Airway anatomy

The upper airway is a complex structure. The pharyngeal airway is a collapsible muscular tube contained within a rigid bony chamber and surrounded by tissue; the proximal end extends to the inferior nasal turbinates. The muscular tube of the paediatric pharyngeal airway is smaller in diameter than in adults, and therefore maintains a higher muscle tone to guard against collapse. Obstructive apnoeas are therefore less common, and the apnoea–hypopnoea index (AHI), used to categorise OSA severity (none: 0; mild: 1–5; moderate: 6–10; and severe: >10) has lower diagnostic thresholds in children. Airway collapse occurs at lower (more negative) critical closing pressures.¹⁷ The retropalatal region is the site of maximal narrowing in the upper airway, particularly in children with OSA. The peak incidence of OSA (3–10 yrs of age) coincides with adenotonsillar growth, in which the growth rate exceeds the increase in airway diameter. The combined adenotonsillar volume correlates with the severity of OSA syndrome.¹⁷ Increased proximal or distal airway resistance can exacerbate underlying OSA.

In OSA in paediatrics, upper airway anatomical features can be classified into infants, child and pre-teen/teen phenotypes, and their associated anatomical endotypes (Table 1). Adenotonsillar hyperplasia is the prototypical endotype responsible for increased upper airway resistance in children. Adenotonsillar hyperplasia is the result of an interaction between environmental stressors (e.g. passive smoking) and an underlying genetic predisposition of the child. The risk of OSA increases in those with a positive history in first-degree relatives, and the prevalence is increased three- to four-fold in children of African descent compared with Caucasians. In addition, abnormal orofacial features at premature birth can also result in abnormal nasal airway resistance and associated oral facial hypotonia, mouth breathing and changes in maxillary–mandibular growth. Neuromuscular and genetic syndrome-related airway endotypes can span all ages. Multiple coexisting airway endotypes are possible (e.g. trisomy-21).¹²

Table 1.

Paediatric upper airway phenotypes in OSA by age and anatomical endotypes

Anatomical airway endotypes	Upper airway phenotypes
	Infant (0–1 yr)	Child (2–8 yrs)	Preteen/teen (9–21 yrs)
Lymphoid hyperplasia (adenoids with [+] or without [–] tonsils)	+/–	+++	++
Soft tissue
Obesity	+/–	++	+++
‘Genetic’ (e.g. mucopolysaccharidosis and Prader–Willi syndrome)	++	+++	++
Craniofacial syndromes
Vault and mandible (e.g. craniosynostosis and Pierre Robin syndrome)	+++	++	+/–
Foramen magnum (e.g. Arnold–Chiari malformation and achondroplasia)	++	++	+/–
Neuromuscular (e.g. cerebral palsy and trisomy-21)	++	+++	+++
Prematurity (<32–34 weeks)	+++	++	–
Inflammatory (e.g. asthma and sickle-cell disease)	+/–	+++	++

Airway pathophysiology

The pathophysiology of paediatric OSA is characterised by a narrowed upper airway that is susceptible to muscular collapse, secondary to negative inspiratory airway pressure whilst asleep. An imbalance between the surrounding bony structure and the tissue contained within has been described, with or without a predisposition to pharyngeal collapsibility. Upper airway neuromotor dysfunction is implied in children with OSA because of an absence of daytime symptoms and residual airway obstruction after AT.¹² Collapse of the pharyngeal airway is primarily caused by relaxation of the genioglossus muscle and other pharyngeal dilator and tensor muscles. Airway collapse occurs via two mechanisms as described in Figure 1.¹⁸ Obstructive sleep apnoea is associated with neuromotor (e.g. cerebral palsy) and muscular disorders (e.g. Duchenne muscular dystrophy) that may further decrease pharyngeal tone and increase sensitivity to anaesthetic agents.

Fig 1.

Pathophysiology of paediatric OSA. Airway collapse occurs via two mechanisms: (i) a decreased pharyngeal transmural pressure attributable to negative pressure on inspiration, and/or extra-lumen positive pressure from tissue and/or bony encroachment; and (ii) reduced longitudinal pharyngeal traction by the trachea caused by gravity or small lung volumes. The critical ‘closing’ pressure (P_close) is a measure of upper airway collapsibility, in which airflow ceases and is greater (less negative) in people with OSA (from Isono,¹⁸ modified with permission).

In addition to enhanced airway collapsibility, other neuromotor endotypes are associated with OSA (Table 2).^17,19,20 Most anaesthetic agents, in addition to reducing pharyngeal tone, impair ventilatory drive (particularly opioids) and arousal-related airway self-rescue. This makes the patient with OSA vulnerable to these drugs. Oxygen administration reduces ventilatory instability, and therefore can mask airway obstruction.²⁰

Table 2.

Paediatric OSA upper airway phenotypes by age and neuromotor endotypes. Protective mechanoreceptor and chemoreceptor reflexes ensure appropriate neuromuscular airway tone and ventilatory drive. The former decreases rapidly with the onset of sleep. The latter decreases in response to both hypoxaemia and hypercarbia. However, this results from high airway collapsibility and a blunted ‘airway self-rescue’ arousal threshold and increased instability of ventilatory control (high ‘loop gain’) increasing airway resistance load; sensitivity to oxygen and carbon dioxide is not directly affected. High loop gain refers to an exaggerated cyclical respiratory response to a respiratory disturbance. For example, airway tone and patency cycles between high, unobstructed to low, obstructed because of excessive ventilatory responses to hypo- and hypercarbia, respectively. Obesity may also contribute to blunted ventilatory and arousal responses because of leptin resistance

OSA neuromotor airway endotype	Receptor (location)	OSA phenotypes
		Infant (0–1 yr)	Child (2–8 yrs)	Preteen/teen (9–21 yrs)
Collapsibility (genioglossus)	CO2 (brainstem) Mechano (airway)	High	High	High
Ventilatory drive	O2 (peripheral) CO2 (brainstem)	?	Normal	?
Arousability ‘Airway self-rescue’	O2 (peripheral) CO2 (brainstem) Mechano (airway)	Blunted	Blunted	Blunted to high
Instability of ventilatory control ‘Loop gain’	O2 (peripheral) CO2 (peripheral and brainstem)	High	? High	? High

End-organ dysfunction

Obstructive airways cause pathophysiological sequelae and end-organ dysfunction, referred to as OSA syndrome (Fig. 2). Systemic inflammation caused by repeated nocturnal hypoxaemia with sympathetic nervous system stimulation is important in children with OSA.²¹ In patients with primary snoring (mild sleep-disordered breathing [SDB]), there is no PSG evidence of airway obstruction, but even these children demonstrate lower performance for behaviour, social interaction and neurocognition, and report symptoms of depression.

Fig 2.

Pathophysiology of paediatric OSA and end-organ dysfunction. GI, gastrointestinal; GORD, gastro-oesophageal reflux; HTN, hypertension; IQ, intelligence quotient.

In the genetically predisposed child, the oxidative stress from recurrent hypoxaemia results in increased concentrations of endorphins and reactive oxygen species. This results in upregulation of brainstem μ-opioid receptors and systemic-inflammation-prone epigenetic alterations.²¹ The former may explain increased sensitivity to the respiratory effects of opioids, and systemic inflammation may explain differences in pain perception and opioid requirements in groups with a higher prevalence of OSA (e.g. those of African descent). There are several inflammatory cascades that are central to the initiation and progression of disease morbidity related to OSA. These include local airway and systemic proliferation of proinflammatory cytokines. Dramatic swings in intrathoracic pressure and repetitive nocturnal release of catecholamines result in autonomic dysfunction and hypertension, and promote the formation of atheromatous lesions. Obstructive sleep apnoea syndrome is also associated with functional disruption of the vascular endothelium, further perpetuating an inflammatory response. The timing and selection of patients for AT to prevent long-term cardiovascular morbidity are unknown.

The interaction between OSA-related systemic inflammation and body habitus, behaviour disorders and impaired neurocognition or asthma is unclear.¹⁴ Most children with OSA have a normal body habitus. However, obese children have a two- to five-fold increased risk for OSA, and failure to grow is associated with SDB in preschool children. Behavioural problems are well established and include attention-deficit hyperactivity disorder, depression, social withdrawal, and aggressive tendencies.⁹ Obstructive sleep apnoea is an independent risk factor for metabolic syndrome and must be excluded in patients who are obese. Children with oSDB have a 3.6-fold increased risk of having severe asthma, and AT surgery may reduce exacerbations.¹⁴

Natural history of disease

Approximately one-third of those with primary snoring may progress to develop more severe disease.¹² Whilst OSA in childhood can resolve spontaneously, this is less likely in children with moderate-to-severe disease. Factors associated with persistent or worsening OSA include obesity, persistent tonsillar hypertrophy, male sex and being of African descent. Acute tonsillitis/tonsillitis-pharyngitis (with or without proven Group A β-haemolytic streptococcus) is usually self-limiting and only requires symptom support.

Diagnosis

Obstructive sleep-disordered breathing

The initial assessment of children with SDB is a medical history and physical examination. The diagnosis of oSDB is primarily a clinical one, in which parents describe habitual loud breathing/snoring (≥3 nights per week) with audible or witnessed pauses and bizarre positioning. Parental discomfort watching their child sleep is associated with moderate-to-severe OSA, as are loud snoring (heard through a door) and witnessed apnoeas.¹³ Those with both oSDB and OSA report daytime symptoms, including hyperactivity; inattention; poor concentration; and, in severe cases of OSA, somnolence. Anatomical conditions predisposing to OSA that increase upper airway resistance (Table 1) should also be assessed. In addition, OSA-related risk factors should also be considered, including poor school performance, history of prematurity (born <32–34 weeks), failure to thrive, gastro-oesophageal reflux disease, a family history of OSA and exposure to passive smoking.¹³ Unlike the validated adult ‘STOP-Bang’ OSA screening questionnaire, validation studies are required for abbreviated paediatric OSA questionnaires.22, 23, 24 It is the combined adenotonsillar (and not tonsillar) volume that correlates with the severity of OSA syndrome.²⁵ It is unclear whether anthropomorphic measures (e.g. neck circumference) improve the diagnostic performance of OSA questionnaires. Both flexible endoscopy and a lateral-view X-ray can be used to assess adenoid size, but they depend on the patient's cooperation and/or specialist expertise.

The gold-standard test to assess oSDB continues to be PSG. Polysomnography will determine conclusively whether or not the patient has OSA and its severity. During an attended PSG, a number of monitors record brain and eye muscle activity, HR, airflow, abdominal and chest wall movements, oxygen saturations and carbon dioxide. Current guidelines provide mixed messages as to the importance of PSG before AT in patients with oSDB. The reader is referred to the updated American Academy of Otolaryngology–Head and Neck Surgery guideline for PSG recommendations before AT.¹⁴ These include: symptomatic children <2 years of age; those with complex high-risk medical conditions (e.g. neuromuscular disorders) that require justification for surgery; the indication for tonsillectomy is uncertain or there is discordance between the physical examination and reported severity of oSDB.

Although controversial, overnight oximetry is another diagnostic option when access to PSG is limited. The McGill oximetry score (MOS) and oxyhaemoglobin desaturation indices (ODIs) represent the most common preoperative oximetry-based screening tools for OSA. The reader is directed to an excellent related review.²⁶ Importantly, the MOS has the advantage of being a single cost-effective test that provides a diagnosis of OSA, risk stratification for PRAEs, or both in healthy children aged >2 yrs.²⁷ Overnight oximetry was not endorsed in a recent UK report, which may reflect less experience because of a much lower reported prevalence of oSDB/OSA requiring AT compared with most other countries.^1,28 However, evidence for its utility in higher-risk (<3 yrs) healthy children has been reported there.²⁹ As an alternative, the oxyhaemoglobin desaturation indices of ODI 3% (>3.5) and ODI 4% (>1.5–2) have been used.¹² The ODI has not been used for perioperative risk stratification, and the utility of overnight oximetry in medically complex patients is unknown.

The role of drug-induced sleep endoscopy (DISE) and MRI in the assessment of SDB is evolving, and they are commonly used diagnostic tools for upper airway evaluation of persisting snoring and OSA after AT. Unfortunately, neither tool has been linked to improved outcomes after AT.

Finally, even though PSG is the preferred metric to diagnose OSA in children, alternative tests have been developed. Home-based sleep studies, including unattended PSG, nap studies and polygraphy, show promise, but they are not endorsed by the American Academy of Sleep Medicine.²⁶ Biological markers, although investigated, are not in mainstream clinical practice.

Recurrent sore throat

Recurrent tonsillitis is a clinical diagnosis. It relies on a patient's history, symptoms and laboratory tests to distinguish between a viral and bacterial infection. To guide laboratory testing for suspected β-haemolytic streptococcal tonsillitis, the McIsaac diagnostic scoring system is recommended.³

Indications for surgery

In general, the decision to remove the tonsils, thereby subjecting a child to potentially risky and painful airway surgery, must be taken with care because in many children their symptoms will resolve spontaneously within 6–12 months.³ That said, in selected children, tonsil removal can decrease upper airway resistance, improve or cure OSA and SDB, decrease the incidence of recurrent pharyngitis, and improve the child's health and QOL. Identification of these children is important because approximately a third of patients with primary snoring progress to more severe forms of SDB, including moderate-to-severe OSA.¹² The following is a summary of current literature concerning the decision to undergo AT.^3,10, 11, 12, 13, 14, 15 Other surgical approaches to treat OSA are beyond the scope of this article.

Primary snoring (mild SDB)

Adenotonsillectomy in the treatment of SDB is largely driven in North America by its potential to improve behaviour, cognition and QOL, independent of underlying OSA. Importantly, an ongoing trial (https://clinicaltrials.gov/: NCT02562040) investigating the benefits of AT compared with watchful waiting will provide further evidence whether surgery is justified.

Obstructive sleep apnoea

Treatment of children with OSA is indicated to arrest or reverse the associated multiple end-organ dysfunction. A consensus on the treatment thresholds to be reached, based on history and physical findings, associated comorbidities and sleep laboratory measurements, is lacking. Thresholds to define moderate (AHI >5 episodes h⁻¹) and severe (AHI >10 episodes h⁻¹) OSA are arbitrary in nature.^10,12 Moderate-to-severe OSA (AHI >5 episodes h⁻¹) is more likely to persist or worsen over time and should be prioritised.^12,30 As detailed previously, if PSG is unavailable, an MOS ≥1 or an abnormal ODI (ODI 3≥3.5 or ODI 4≥1.5–2) is considered an indication to undergo surgery. Beyond these diagnostic tests, comorbid conditions that may be improved with AT include low QOL, uncontrolled asthma, poor school performance, behavioural problems, metabolic syndrome, oral-motor dysfunction (e.g. articulation) and recurrent otitis media amongst others.^12,13,30 Genetic conditions that have a greater risk of pulmonary hypertension (e.g. Duchenne muscular dystrophy or trisomy-21) must also be considered. The surgical ‘window of opportunity’ to halt or reverse end-organ dysfunction is unclear.

Sore throat

The evidence for AT to treat patients with recurrent tonsillitis is modest at best. The current indication for surgery is per the Paradise criteria based on annual frequency of infections, and includes one or more of the following: temperature >38.3°C, cervical lymphadenopathy, tonsillar exudate and positive test for Group A BHS. Surgery is not indicated if there have been less than three episodes of tonsillitis in the previous 12 months.^3,14

Surgical procedure and technique

Traditionally, extracapsular tonsillectomy involves complete removal of the palatine tonsils and its surrounding capsule. To improve exposure, the surgery typically includes excision of a portion of the anterior tonsillar pillar. The procedure is performed by various techniques that are generally divided into ‘cold’ and ‘hot’ techniques. A cold technique uses sharp and blunt instruments to excise tissue followed by measures to control bleeding. A hot technique uses thermal instruments for incision, excision and simultaneous control of bleeding, and includes primarily mono- or bipolar (scissors or forceps) electrocautery and bipolar radio-frequency ionic dissociation (coblation), and various laser, harmonic scalpel, plasma knife and ultrasound technologies. In general, the hot compared with cold technique causes less intraoperative bleeding, but greater postoperative pain and an increased risk of delayed secondary bleeding from thermal injury of adjacent healthy tissue. Coblation, a non-heat-driven process, in which radio-frequency energy is applied to a conductive medium (usually saline), attempts to reduce thermal injury to surrounding tissues. However, no surgical technique is consistently superior in terms of postoperative pain, bleeding or wound healing.^15,31

Although controversial, partial tonsillectomy (debulking the tonsil whilst leaving the thin surrounding capsule intact using, e.g. a microdebrider) is increasingly popular to treat SDB and recurrent tonsillitis because of reported lower rates of pain and bleeding, and reduced utilisation of healthcare resources.^3,11,32 Its treatment efficacy in OSA is comparable with tonsillectomy. Functional recovery appears more favourable for partial tonsillectomy, although higher rate of tonsillar regrowth (2–6%) and recurrence of SDB symptoms have been reported.

Complications

Complications directly related to AT are categorised by when they occur (Table 3). Intraoperative bleeding is less common now because of the adoption of a hot surgical technique, but burns, airway fires and greater postoperative pain may occur. These patients have an increased risk for major PRAEs (laryngospasm and bronchospasm, apnoea and oxygen dependence). In children, SDB is an independent risk factor for critical PRAEs, and the incidence of respiratory complications after AT for OSA is nearly five-fold higher than in patients without OSA.¹⁴ The rate for a major PRAE after AT in children is 5.8% (95% confidence interval [CI]: 4.2–7.4%; p<0.001).¹⁴ Nausea and vomiting are common (21%; range: 13–32%) and are associated with aspiration, bleeding, prolonged length of stay and readmission.

Table 3.

Complications associated with AT in children

Onset	Description
Immediate	•Trauma or burn (secondary to surgical technique or airway device ignition) to eyes, lips, teeth, mandible, temporomandibular joint, cervical vertebra, tongue, uvula, soft palate, pharynx, carotid artery and larynx •Compression or dislodgement of airway device •Bleeding at surgical site
Intermediate (PACU ≤24 h)	•Emergence delirium •Pain, including referred otalgia •Major respiratory event, including post-obstructive pulmonary oedema and aspiration •Nausea and/or vomiting •Primary bleeding (caused by surgical technique), including need for blood transfusion •Death secondary to PRAE and/or bleeding
Delayed (>24 h)	•Pain, including referred otalgia •Nausea or vomiting •Dehydration, delayed feeding, or fever •Secondary bleeding (caused by wound healing), including need for blood transfusion up to 2 weeks •Speech disorders (e.g. velopharyngeal incompetence) •Injury to hypoglossal, glossopharyngeal, or vagal nerve resulting in transient or permanent dysphagia or altered taste •Persistent or recurrent OSA/oSDB •Tonsil regrowth or recurrent tonsillitis •Nasopharyngeal stenosis •Neck infection •Internal jugular vein thrombosis •Death secondary to PRAE and/or bleeding

PACU, post anesthesia care unit; PRAE, perioperative respiratory adverse event; OSA, obstructive sleep apnea; oSDB, obstructive sleep disordered breathing.

In Canada, crude hospital emergency department visits (12.4%) and readmission (2.7%) rates after AT are the highest amongst common paediatric ambulatory surgeries.⁴ Similar hospital revisit rates after AT have been reported in the USA. The highest hospital revisit rates are amongst adolescents (11–18 yrs), followed by very young children (0–3 yrs). Pain is amongst the most common reasons for hospital revisit/admission across all age groups.^4,5 Hospital visits because of pain, bleeding and febrile illness are more common in adolescents, whilst nausea, vomiting and dehydration are more common in younger children.¹¹

Provider surveys and malpractice data indicate that airway complications account for two-thirds of fatalities, which usually occur at home within the first 24 h after surgery. Obstructive sleep-disordered breathing diagnosis and recent administration of opioids feature prominently.^8,33 The remaining third of the fatalities have been attributed to surgical bleeding. The rate of haemorrhage after tonsillectomy is higher than after tonsillotomy (4.2% vs 1.5%), and rates of reoperation are also higher (2.2% vs 0.64%).³² Secondary bleeding peaks at postoperative Days 5–7, but can occur up to 14 days.¹⁵

Outcomes

Obstructive sleep-disordered breathing

Although oSDB cure rates are variable, surgery reliably improves various PSG-related measures and some clinical features. A meta-analysis reported the success rate of AT in normalising PSG was 82.9% (95% CI: 76.2–89.5%).¹⁴ Apnoea–hypopnoea index reduces to a greater extent in patients with mild-to-moderate OSA, whereas severe OSA is more resistant to treatment.¹¹ Treatment resistance is linked to obesity, in which persistent OSA after AT is 33–76%.³⁰ Risk factors for residual OSA include age >7 yrs, obesity, asthma, severe OSA, being of African descent and having Down syndrome.^11,30 Children with risk factors for residual OSA or who have persistent oSDB symptoms after AT should undergo follow-up PSG and/or DISE. Short-term improvements in QOL and negative behaviour, but not intelligence quotient, have been reported after AT.³⁰ There is no evidence for longer-term (2–3 yrs) differences in behaviour indices or neurocognitive improvement after surgery.³⁴

Recurrent tonsillitis

The effect of tonsillectomy on the incidence of throat infection is reported to be modest only in the short term (within 1 yr).³ There is limited evidence to suggest that tonsillectomy for recurrent infections improves QOL and reduces utilisation of healthcare.^3,11 The extent to which the tonsils contribute to a sore throat and consequently the potential impact of surgery is unclear.³

Conclusions

Adenotonsillectomy is one of the most common paediatric surgeries. Remarkably, despite the significant risk of morbidity and mortality, there is widespread practice variation in terms of surgical indication, approach and technique. Controversy remains regarding its long-term benefits to treat both oSDB and recurrent sore throat. Although both OSA and oSDB are associated with increased risk of PRAEs, being able to distinguish between the two is critical for the anaesthetist. Obstructive sleep apnoea is a systemic inflammatory disease with multiple end-organ effects, increased sensitivity to opioids and reported pain. The anaesthetic technique should be adjusted accordingly and can be informed by preoperative overnight oximetry, an increasingly accepted diagnostic tool for OSA. Residual symptoms of oSDB/OSA after AT imply severe baseline disease and/or an underlying medical complexity.

Declaration of interests

The authors declare that they have no conflicts of interest.

Acknowledgements

The authors would like to thank Dr Lisa Elden for reviewing the manuscript and providing feedback; Drs Debbie Schwengel and Stacey Ishman for their feedback regarding obstructive sleep apnoea endotypes in paediatrics and drugs affecting airway tone, respectively; and Johanna Spaans for editing the manuscript.

MCQs

The associated MCQs (to support CME/CPD activity) will be accessible at www.bjaed.org/cme/home by subscribers to BJA Education.

Biographies

Kimmo Murto MD FRCP is a paediatric anaesthetist working at the Children's Hospital of Eastern Ontario (CHEO) and an associate professor at the Department of Anesthesiology and Pain Medicine, University of Ottawa. He is past chair of the paediatric committee of the Society of Anesthesia and Sleep Medicine. His research interests include care of children undergoing tonsillectomy.

Julie Zalan MD FRCPC is an anaesthetist with specialty training in paediatrics working at Kingston Health Sciences Centre and an assistant professor at the Department of Anesthesiology and Perioperative Medicine, Queen's University, Kingston.

Jean-Philippe Vaccani MD FRCPC is a paediatric otolaryngologist working at CHEO, and an associate professor and vice-chair of education at the Department of Otolaryngology–Head and Neck Surgery, University of Ottawa. He has published previously on paediatric obstructive sleep apnoea.

Matrix codes: 1H02, 2D02, 3A02