Children with craniofacial microsomia (CFM) are at risk of obstructive sleep apnoea (OSA). This systematic review provides an overview of the literature on the prevalence of OSA in children with CFM. A search was performed in PubMed, Embase, Cochrane Library, and Web of Science for articles on CFM and OSA. The following data were extracted from the articles: number of patients, patient characteristics, presence of OSA, polysomnography outcomes, and the treatments and outcomes of OSA. We included 16 articles on CFM and OSA, four of which reported the prevalence of OSA (range 7–67%). Surgical treatment was more often described in these patients than conservative treatment. According to the literature, OSA is related to CFM. However, as there have been no prospective studies and few studies have presented objective measurements, no definitive conclusions can be drawn. Prospective studies are needed to determine the prevalence of OSA in patients with CFM.
A common problem in children with a craniofacial anomaly is upper airway obstruction. This obstruction may be related to bilateral mandibular hypoplasia, as is the case in children with Pierre Robin sequence and Treacher Collins syndrome. The prevalence of upper airway obstruction, and more specifically of obstructive sleep apnoea (OSA), in children with Pierre Robin sequence and Treacher Collins syndrome is 12.5% and 46%, respectively. In the normal population this is 3–4%.
OSA is one of the clinical manifestations of upper airway obstruction and is characterized by snoring, laboured breathing during sleep, apnoea, and excessive daytime sleepiness. Complications of OSA include failure to thrive, pulmonary hypertension, cor pulmonale, and sudden death. Therefore, the accurate diagnosis and identification of risk groups is important.
As is the case for Pierre Robin sequence and Treacher Collins syndrome patients, craniofacial microsomia (CFM) patients have mandibular hypoplasia as a clinical characteristic. CFM is the result of a disturbance in the embryological development of the first and second branchial arches and is characterized by asymmetric underdevelopment of the facial structures, including the mandible, maxilla, ears, soft tissues, and facial nerves. CFM is most often regarded as a unilateral malformation, however the facial structures are involved bilaterally in 10% of cases and several recent publications have suggested that the contralateral side is abnormal in most cases as well, although not truly hypoplastic. The reported incidence rate ranges from 1 in 3500 to 1 in 20,000, which makes CFM the second most common facial birth defect after cleft lip and palate. CFM in combination with epibulbar dermoid and extra-craniofacial anomalies, such as heart, renal, and vertebral anomalies, is known as Goldenhar syndrome.
The most typical deformity of CFM is mandibular hypoplasia, which occurs in 89–100% of cases. The most commonly used classification of mandibular hypoplasia is the classification of Pruzansky modified by Kaban, in which mandibular hypoplasia is classified into four types. In type I, the mandibular ramus and temporomandibular joint (TMJ) are of normal shape but small. In type IIa, the mandibular ramus is abnormal in both size and shape, but the deformed TMJ is adequately positioned. In type IIb, the mandibular ramus and TMJ are abnormal in shape, size, and location. Type III deformity consists of an absent ramus, condyle, and TMJ.
As mandibular hypoplasia increases the risk of airway obstruction, patients with CFM are theoretically at risk of airway obstruction. Several authors have stated that patients with CFM should be screened routinely for OSA. Nevertheless, the exact prevalence of OSA in CFM and the severity of the pathology on which these statements are based are not mentioned in these expert opinions.
The aim of this review is to provide an overview of the literature regarding CFM and the prevalence and treatment of OSA based on the following key questions: (1) What is the prevalence of OSA in patients with CFM? (2) What are the treatment modalities for OSA reported in patients with CFM? (3) What is known about the follow-up after treatment for OSA in patients with CFM?
Within the group of craniofacial malformation patients, feeding difficulties are often closely related to upper airway obstruction. This topic is described separately in a second article entitled ‘Feeding difficulties in craniofacial microsomia: a systematic review’.
A search of public domain databases was performed to identify articles focusing on CFM and OSA. The search was conducted in the following databases: PubMed, Embase, Cochrane Library, and Web of Science (all searched up to 27 August 2014). In addition, we performed a manual search of secondary sources including references of the articles initially identified. The goal was to identify all studies addressing CFM in relation to OSA.
The following search terms were used: (((facial[tiab] OR face[tiab] OR hemifacial[tiab] OR orbitocranial[tiab] OR facies[tiab] OR cranial[tiab] OR mandibulofacial[tiab] OR otomandibular[tiab] OR craniofacial[tiab] OR faciocranial[tiab] OR hemimandibular[tiab]) AND (microsom*[tiab] OR asymmetr*[tiab] OR dysosto*[tiab] OR dysplasia[tiab] OR anomal*[tiab] OR deformit*[tiab] OR hypoplasia[tiab] OR syndrom*[tiab] OR malformation*[tiab]) OR (treacher collins[tiab] OR goldenhar[tiab]) OR (oculoauriculovertebral*[tiab] OR facioauriculovertebral*[tiab] OR (auriculo vertebral*[tiab]))) AND (airway obstruction*[tiab] OR obstructive airway *[tiab] OR nocturnal apnea[tiab] OR nocturnal apnoea[tiab] OR sleep apnea[tiab] OR sleep apnoea[tiab] OR osa[tiab] OR osas[tiab] OR osahs[tiab])) AND publisher[sb].
Data extraction and analysis
Two investigators (C.J.J.M.C. and B.I.P.) screened the studies independently. All articles on the prevalence and treatment of OSA in patients with CFM were included. Expert opinions were excluded. The full texts of all articles meeting the inclusion criteria and articles for which the abstract was lacking information were obtained.
Articles were graded on quality of evidence using the Oxford Centre for Evidence-Based medicine (CEBM) criteria. Data on the number of patients, patient characteristics such as gender, age, and severity of the CFM, the presence of OSA, polysomnography outcomes, and the treatment and outcome of OSA were tabulated when available.
Craniofacial microsomia and obstructive sleep apnoea
The search retrieved 1385 relevant articles. After removing duplicate articles and including further articles from the manual search of secondary sources, 835 articles were examined based on the title and abstract. A total of 749 articles were then excluded. Of the 86 articles retrieved for further examination, 16 were included in the present analysis ( Fig. 1 ).
What is the prevalence of OSA in patients with CFM?
Six studies and case series on the prevalence of OSA in CFM were found ( Table 1 ). OSA was defined as complete cessation of airflow for more than two breaths or 10 s and hypopnoea as a ≥50% reduction in respiratory airflow accompanied by a decrease of ≥3% in oxygen saturation (SaO 2 ), with a minimum of 30 episodes of obstructive apnoea in a 7-h sleep period. In some cases OSA was not defined at all. The prevalence of OSA in these studies varied from 7% to 67% ( Table 2 ).
|Reference||CEBM level of evidence||Methodology||Aim of the study|
|Cloonan et al.||III||Case–control study||To compare sleep outcomes in children with and without CFM|
|D’Antonio et al.||III||Cross-sectional study||To describe the occurrence and magnitude of pharyngeal and laryngeal anomalies in a population of patients with OAVS|
|Cohen et al.||III||Retrospective study||To determine the prevalence of OSA in CFM|
|Luna-Paredes et al.||III||Retrospective study||To determine the prevalence of airway obstruction symptoms in craniofacial anomalies|
|Sculerati et al.||III||Retrospective study||A delineation of clinical characteristics affecting the airway in craniofacially affected children|
|Sher et al.||III||Retrospective study||To describe the nasopharyngoscopic findings in patients with craniofacial anomalies|
|Reference||Method of diagnosis of OSA||No. of patients with CFM||Age, range or mean||Gender, M/F||Severity of CFM||No. of patients with OSA||Outcome of PSG|
|Cloonan et al.||PSQ, supplemental sleep questions||84/124||5.5–8.5 y||NR||Not possible to deduce||10||NR|
|D’Antonio et al.||Chart review, PSG||41||6 y 4 m||14/9||NR||5||NR|
|Cohen et al.||Chart review||38||8 y 3 m||21/17||Not possible to deduce||9||NR|
|Luna-Paredes et al.||Proactive screening programme, cardiorespiratory polygraphy||9||NR||NR||NR||5||AHI: 4.76 (mean)|
|Sculerati et al.||Chart review||41||NR||NR||NR||9||NR|
|Sher et al.||PSG||84||Birth to 24 y||NR||NR||6||NR|
Case–control study on the prevalence of OSA in CFM
Cloonan et al. described the prevalence of sleep disordered breathing (SDB) in children with CFM by studying 124 cases and 349 controls. Parents of children with CFM reported that their child snored more often than did the parents of otherwise healthy controls (29% vs. 17%). Eighty-four of the 124 cases received a supplementary questionnaire regarding history of airway interventions and the use of polysomnography (PSG) to classify the severity of SDB or OSA. A history of airway interventions was more often reported in children with CFM than controls (14% vs. 8%) and more children with CFM had undergone PSG than controls (20% vs. 2%). Of the 15 children with CFM who underwent PSG, 10 were diagnosed with SDB/OSA (67%); of the controls, four out of four were diagnosed with SDB/OSA.
Cross-sectional study on the prevalence of OSA in CFM
D’Antonio et al. evaluated the pharyngeal and laryngeal structure and function in patients with Goldenhar syndrome by physical examination, otolaryngological examination, and video-nasoendoscopy. Nine out of 41 patients (22%) reported symptoms of airway obstruction, of whom five (12%) had OSA documented by PSG ( Table 2 ).
Retrospective studies on the prevalence of OSA in CFM
Cohen et al. found a prevalence of OSA of 24%. Upper airway disorders fell within three categories. Patients in category 1 were asymptomatic for airway disorders, patients in category 2 were suspect for intermittent OSA or had experienced a perioperative apnoeic event, and patients in category 3 had a definite history of OSA. Like Cloonan et al., they found that patients with more severe mandibular and/or extra-craniofacial anomalies had a greater risk of OSA, i.e. 89% of the patients with a category 3 upper airway disorder exhibited a higher incidence of more severe mandibular involvement.
Luna-Paredes et al. found a prevalence of OSA of 56% in nine patients with Goldenhar syndrome, with OSA ranging from mild to severe.
Sculerati et al. studied 14 patients with Goldenhar syndrome, 12 with CFM, and 15 with microtia. Nine cases required a tracheostomy (22%). The exact diagnosis of the nine tracheostomized cases was not mentioned and the prevalence of OSA was also not mentioned.
Sher et al. evaluated 84 patients with facio-auriculo-vertebral sequence of whom six underwent endoscopy as they had evidence of upper airway obstruction. Five out of six had OSA and one out of six had obstructive awake apnoea, defined as a complete or partial cessation of air exchange during wakefulness caused by an observable obstruction in the upper airway.
What are the treatment modalities for OSA reported in patients with CFM?
Twelve studies were found concerning the treatment of OSA in craniofacial anomalies and/or CFM ( Tables 3 and 4 ).
|Reference||CEBM level of evidence||Methodology||Total no. of patients||No. of patients with CFM||Age of total group, mean (range)||Preoperative outcome of PSG||Treatment of OSA||Postoperative outcome of PSG|
|Burstein et al.||III||RS||28||4||5 y 3 m
(18 m to 17 y)
|Skeletal expansion||AI: 1
|Järund et al.||III||RS||21||1||3 y 7 m
(1 m to 2 y)
|Sher et al.||III||RS||266||84 (6 patients treated for OSA)||(Birth to 24 y)||NR||2 mandibular advancement, 3 tracheostomy, 1 refused treatment||NR|
|Cohen et al.||III||RS||16||7||4 y 8 m
(14 w to 16 y)
|Mandibular distraction||RDI: 1.7
|James and Ma||III||RS||25||7||(12 m to 13 y)||NR||Mandibular distraction||Decannulation|
|Sorin et al.||III||RS||18||5||2 y 8 m
(2 m to 8 y)
|Sculerati et al.||III||RS||251||41||4 y 10 m
(Birth to 26 y)
|Reference||CEBM level of evidence||Methodology||No. of patients||Age||Preoperative outcome PSG||Treatment of OSA||Postoperative outcome PSG|
|Morris et al.||V||Retrospective CR||1||PSG not possible with capped tracheostomy||Mandibular distraction||AHI: <5
|Hoch and Hochban||V||Retrospective CR||1||4 years||AI: >20
|Nasopharyngeal tube and mandibular distraction||NR|
|Kourelis et al.||V||Retrospective CR||1||17 years||AHI: 28
|Stellzig et al.||V||Retrospective CR||1||Birth||AHI: not given
|Orthopaedic myofunctional application||NR|
|McCarthy et al.||IV||Retrospective CS||2||Birth||AHI: 23
Retrospective studies on the treatment of OSA in CFM
Burstein et al. developed the airway zone concept in order to treat patients with severe OSA refractory to standard medical and surgical treatment. Zone 1 extends from the nares to the velum, zone 2 from the lips to the hypopharynx, zone 3 from the epiglottis to the trachea, and zone 4 from the subglottis to the bronchi. Their goal was to avoid permanent tracheostomy. The four patients with Goldenhar syndrome included in this study all suffered from an obstruction in zone 1 and zone 2 and underwent bony and soft tissue procedures. All patients responded well to skeletal expansion. Both the preoperative median apnoea index, defined as the total number of apnoeic events divided by the total sleep time and multiplied by 60, and the respiratory disturbance index improved postoperatively ( Table 3 ).
Järund et al. included one patient with Goldenhar syndrome and described the effects of continuous positive airway pressure (CPAP) in patients with craniofacial anomalies and OSA. Clinical characteristics were not given. According to Järund et al., the treatment of OSA with CPAP in children with craniofacial anomalies is superior to early surgery and tracheostomy ( Table 3 ).
All 16 patients in the study of Cohen et al. were considered tracheostomy candidates after conventional medical and surgical treatment of OSA had failed, e.g. tonsillectomy, adenoidectomy, uvulopalatoplasty, and relief of nasal obstruction, or had tracheostomies placed shortly after birth and could not be decannulated by standard protocols. In all cases, external mandibular distraction devices were used. Distraction was started on the third postoperative day and proceeded until the signs and symptoms of OSA resolved. In the patients without tracheostomies, a PSG was obtained preoperatively and postoperatively before removal of the distraction devices. The respiratory disturbance index was defined as the number of apnoeic and hypopnoeic events per hour of sleep. Unfortunately, specific clinical characteristics and outcomes could not be deduced for the patients with CFM ( Table 3 ).
Seven of the 25 patients studied by James and Ma were diagnosed with CFM, five bilaterally affected and two unilaterally. Objective data on the airway status of these patients before the tracheostomy were not available. The unilateral cases were treated with a combination of costochondral grafting and osteotomy. The exact treatment of the bilateral cases was not described. All CFM patients were treated successfully and decannulated ( Table 3 ).
Sorin et al. studied five patients with Goldenhar syndrome who underwent upper airway endoscopy before and after mandibular distraction. The results of the pre- and post-surgery airway endoscopy at the level of the oropharynx and supraglottis are shown in Table 5 . All of the Goldenhar patients were decannulated and remained tracheostomy-free during the time of the study ( Table 3 ).