Introduction
Obstructive sleep apnoea (OSA) is characterised by oxygen desaturation and reduced oronasal airflow despite preserved thoracic and abdominal respiratory effort.1 It is a common childhood problem affecting 1–3 per cent of all children. It can manifest itself with habitual snoring, mouth breathing and observed apnoeic episodes, restless sleep, enuresis and behavioural problems.Reference Rosen, Larkin, Kirchner, Emancipator, Bivins and Surovec2 Recent research has also shown that OSA may affect a child's ability to learn.Reference O'Brien, Mervis, Holbrook, Bruner, Klaus and Rugerford3 Adenotonsillectomy is generally accepted as an effective way of treating this problem in children with adenotonsillar enlargement.Reference Stradling, Thomas, Warley, Williams and Freeland4 It is a well-known fact that some children suffering with OSA do poorly after elective adenotonsillectomy, sometimes requiring paediatric intensive care unit level care. It is also well known that paediatric intensive care beds are a scarce and expensive commodity, making appropriate bed usage an important issue. In the literature there appears to be little consensus on how to identify the ‘high risk’ child pre-operatively. American guidelines advocate complex sleep studies; some advocate the use of overnight oximetry studies whilst others depend merely on clinical suspicion.5
We set out to clarify two aspects of elective paediatric adenotonsillectomy. Firstly, we aimed to identify regional surgical referral patterns for adenotonsillectomy in children with OSA to our tertiary referral centre. Secondly, we wanted to develop guidelines for referral to our regional unit to aid selection of those children likely to require intensive care following OSA adenotonsillectomy.
Methods
This study consisted of two parts, namely a questionnaire and a prospective study. The questionnaire was used to identify regional surgical referral patterns (Appendix I). It was sent to all consultant otolaryngologists in five surrounding hospitals, with no on-site paediatric intensive care facilities, in the region of West Yorkshire, UK. A prospective observational study was performed in our department over a three-year period between January 2002 and February 2005 to identify which children actually required paediatric intensive care admission following OSA adenotonsillectomy. All ‘high risk’ children undergoing adenotonsillectomy were included. Considered ‘high risk’ were those referred with severe OSA on the basis of (overnight) oximetry studies with documented nocturnal desaturation below 80 per cent, those admitted to hospital with concerns about frequent apnoeas, those with OSA plus a syndrome associated with airway problems, those with OSA and neurological conditions associated with altered muscle tone, and children with OSA and one or more co-morbidities as judged significant by a senior paediatric otolaryngologist and paediatric anaesthetist. All healthy children with otherwise ‘simple OSA’ were excluded. Data on most severe co-morbidity, pre-operative OSA investigation, planned post-operative destination, duration of paediatric intensive care unit admission and intra- and post-operative complications were collected.
Results
Questionnaire
The response rate for the questionnaire was 15 out of 20 (75 per cent). Three consultants indicated they would refer patients pre-operatively to the tertiary centre on the basis of a clinical history alone. Five consultants would refer on the basis of oximetry. Nine consultants stated they would refer syndromal children and children with previous peri-operative airway problems. Two consultants stated they would refer on the basis of polysomnography, which is not readily available in this region. Four consultants indicated they would never refer electively for OSA adenotonsillectomy (see Figure 1).
Fig. 1 Results of questionnaire.
Prospective study
A total number of 49 children met our inclusion criteria for the prospective study. The majority were children with one or more co-morbidities. Most were suffering from neurological and respiratory conditions or syndromes that are associated with disproportion of the upper airways such as Down's syndrome, Di George or CHARGE syndrome (coloboma of the eye, heart anomalies, atresia of the choanae, retardation of growth and/or development, genital and/or urinary abnormalities, ear abnormalities and deafness). Only one otherwise normal child with severe OSA was included (Figure 2). Oximetry studies were available in 59 per cent of cases. A total of 12 patients required elective paediatric intensive care unit admission following OSA adenotonsillectomy (Figure 3). No post-operative emergency admissions were recorded during this period. Figure 3 shows a co-morbidity profile for all patients that went to the paediatric intensive care unit. The majority consisted of neurological conditions and syndromes. Figure 4 shows the treatment required by the patients admitted to the paediatric intensive care unit. Nine patients were simply monitored and received supplementary oxygen. One required a nasopharyngeal airway. One patient required prolonged intubation. This was a child that was admitted with an acute airway obstruction secondary to upper and lower respiratory tract infection. One child required re-intubation. This was a child with Di George syndrome that initially did well but developed acute airway obstruction on the second post-operative day. Subsequently they underwent a tracheostomy followed by a mandibular advancement. The majority of the children admitted to the paediatric intensive care unit were discharged within 24 hours as shown in Figure 5.
Fig. 2 Profile of most severe co-morbidity of all included patients.
Fig. 3 Profile of most severe co-morbidity in patients requiring paediatric intensive care unit.
Fig. 4 Treatment in paediatric intensive care unit.
Fig. 5 Duration of paediatric intensive care unit admission.
Discussion
Our department performs around 600 paediatric adenotonsillectomies each year. The majority of children referred to our tertiary centre with a view to OSA adenotonsillectomy and potential paediatric intensive care unit level care will have their surgery done routinely without paediatric intensive care facilities having been arranged. This is because they are deemed to be a ‘significant risk’ on clinical assessment. This in turn means that the family have had to travel extra distances potentially for no difference in care.
In the literature there appears to be little consensus on how to identify the ‘high risk’ child pre-operatively. This is reflected by the current referral practice in our region, which shows little sign of uniformity. For example very few referred children have had overnight oximetry studies.
American guidelines advocate complex sleep studies involving overnight polysomnography and electroencephalogram monitoring as the gold standard.5 This is currently not widely available for children in the UK. Some, like Nixon et al., advocate the use of overnight oximetry studies which they consider a good tool to estimate the severity of nocturnal hypoxia.Reference Nixon, Kermack, Davis, Maoukian, Brown and Brouillette6 Others have reported, however, that overnight oximetry studies should not be used as a basis for deciding which children should proceed to adenotonsillectomy as normal overnight oximetry does not exclude significant obstructive sleep apnoea.Reference Rosen, Muckle, Mahowald, Goding and Ullevig7, Reference Brouillette, Morielli, Leimanis, Waters, Luciano and Ducharme8 Others merely depend on clinical suspicion and rely on the history. It has been shown that choking, observed apnoeic episodes, snoring, restless sleep and mouth breathing are the clinical signs correlating to the presence and severity of OSA.Reference Abisheganaden, Chan, Chee, Yap, Poh and Wang9, Reference Brouillette, Hanson, David, Klemka, Szatkowski and Fernbach10 This in combination with large tonsils and nasal obstruction on examination can function as a basis for deciding which child is likely to have (significant) OSA and therefore will benefit from adenotonsillectomy. It has been argued that clinical assessment is, at present, the most sensitive indicator of OSA, although it lacks specificity.Reference Kubba11
• It is a well-known fact that some children suffering with obstructive sleep apnoea have a higher incidence of respiratory complications after elective adenotonsillectomy, sometimes requiring paediatric intensive care
• There appears to be little consensus on how to identify the ‘high risk’ child pre-operatively
• This study aims to identify regional surgical referral patterns for adenotonsillectomy in children with obstructive sleep apnoea to a tertiary centre with paediatric intensive care facilities
• Clinical judgment without complex sleep studies by those experienced in this area was sufficient to detect complicated cases of OSA with co-morbidity requiring paediatric intensive care
The 49 ‘high risk’ patients included in the prospective audit consisted of a very heterogeneous group of children. Oximetry studies were available in only 58 per cent of cases. The co-morbidity profile shows a wide variety of problems, with only one normal child with severe OSA included in the study. The decision to arrange paediatric intensive care facilities was made pre-operatively by a senior consultant ENT surgeon and a senior consultant paediatric anaesthetist. These decisions were based on careful clinical assessment and in most cases without the benefits of sleep studies or oximetry studies. Clinical assessment was focused in particular on the coexistence of neurological conditions affecting the airway, syndromes associated with disproportionate upper airways, severe cyanotic congenital heart disease, morbid obesity, acute airway obstruction and respiratory failure. Only 12 children out of the 49 included required paediatric intensive care and no emergency admissions to the paediatric intensive care unit were recorded. A minority of the children admitted to the paediatric intensive care unit required further intervention in the form of invasive airway management. Those were also the patients that required a prolonged stay in intensive care. The majority required 1:1 nursing and oxygen administration and were discharged within 6–24 hours post-operatively, which indicates that they perhaps could have been managed in a paediatric high dependency setting rather than paediatric intensive care unit. Careful anaesthesia in combination with close post-operative monitoring of all children undergoing OSA adenotonsillectomy, regardless of their post-operative destination, however, cannot be over-emphasised.
This study is limited by a number of factors. Firstly, the numbers included are relatively small. Secondly, the size of the referred case load is undetermined; it is unknown which proportion of cases have been referred out of the total number of patients suffering with OSA and how many patients run into problems following surgery in peripheral centres. It is also conceivable that there may be a bias by setting this study in a tertiary centre, which may create a sense of security as all the ‘cavalry’ is at hand should complications arise. Other variables may consist of the anaesthetic technique used. Brown et al. postulated in 2004 that children with OSA might have reduced morphine requirements for analgesia related to their young age and an up-regulation of central opioid receptors due to recurrent hypoxaemia.Reference Brown, Laferriere and Moss12 In our centre only short acting opiates are used sparingly, if necessary at all, during general anaesthesia for OSA adenotonsillectomy.
Conclusion
The regional referral study showed that currently there is no regional consensus regarding the need for elective paediatric intensive care referral for OSA adenotonsillectomy.
Based on the results of the audit of ‘high risk’ patients we conclude that careful clinical assessment by an ENT surgeon and in particular by an experienced senior paediatric anaesthetist is sufficient to plan post-operative care. In our opinion complex sleep studies are not essential in predicting which child requires post-operative paediatric intensive care. On the basis of this study no strict criteria for referral to a centre with a paediatric intensive care unit for elective paediatric intensive care admission following OSA adenotonsillectomy have been determined.
It therefore stands that if the local unit to which an OSA child presents has little experience in clinically assessing OSA patients, a clinical assessment in a tertiary centre may be appropriate and expensive tests may not be.
We strongly recommend referral to a unit with paediatric intensive care facilities in elective paediatric intensive care admission for children with severe OSA plus:
(1) Neurological conditions affecting the airway.
(2) Syndromes associated with disproportionate upper airways.
(3) Severe cyanotic congenital heart disease.
(4) Morbid obesity.
(5) Acute airway obstruction.
(6) Respiratory failure.
APPENDIX I CONSULTANT QUESTIONNAIRE
PICU = paediatric intenstive care unit; HDU = high dependancy unit
We are trying to better define reasons that cause children undergoing adenotonsillectomy (for any indication) to be referred to a tertiary unit for their surgery in the Yorkshire region.
Please indicate your answers to the questions below by placing an X in the appropriate box.
1) Highest level of ward care available to your department
- PICU/HDU □
Day case unit □
- Short stay ward □
Comments □
2) Indication(s) for pre-operative referral to unit with a PICU
Obstructive sleep apnoea from clinical history □
Pulse oximetry demonstrated obstructive sleep apnoea □
Polysomnography demonstrated obstructive sleep apnoea □
Previous peri-operative airway problems □
Syndromic/neurological problems □
Other… □
3) Technique used for the majority of paediatric adenotonsillectomies □
4) Apart from reasons listed in part 2 do you find your anaesthetic colleagues refer some
patients to tertiary centres for surgery based on:
- Age □
Other factors □
- Concurrent morbidity □
