Less invasive treatment of sleep-disordered breathing in children with syndromic craniosynostosis

The hospital’s electronic database was searched for all children with a diagnosis of SCS admitted between 01/01/2004 and 31/12/2016. Children with isolated craniosynostosis were excluded.

The following clinical data were collected: underlying diagnoses, age, gender, type of cleft (if any), presence of choanal stenosis or midface hypoplasia, different treatment modalities and any treatment prior to admission to our center. Children undergoing TPP treatment had sleep studies performed prior to treatment onset and after each (re-)fitting of the plate. All other children had sleep studies performed at least once/year.

Sleep studies were performed using a computerized system (Embla N 7000, MedCare, Reykjavik, Iceland). The study montage comprised the following channels and sensors: chest and abdominal wall movements (respiratory inductive plethysmography, MedCare), nasal pressure and linearized nasal airflow (nasal prongs and built-in pressure transducer, MedCare), pulse oximeter saturation (SpO₂) and pulse waveform (Radical, Masimo Inc., Irvine, USA), electrocardiogram (MedCare), and digital video via infrared camera (Panasonic; Tokyo, Japan). Recordings commenced in the evening and lasted for at least 8 h; all infants were studied in the supine position. Recordings were terminated if more than 3 desaturations to < 60% SpO₂ occurred, and not attempted if upper airway obstruction prior to fitting the palatal plate was so severe that it could only be managed by endotracheal intubation.

All recordings were manually analyzed for the presence of respiratory events using slightly modified standard criteria [26] as used in our previous work [22‐25]. In brief, total sleep time (TST) was determined from the first 10-min epoch without movement artifact or a distorted pulse waveform to the last such 10-min epoch; recordings comprising less than 3 h of TST were excluded. An apnea was scored if (i) the amplitude of the nasal airflow fell to < 20% of the average amplitude of the two preceding breaths, (ii) no airflow was detected at the mouth, and (iii) the event comprised at least two breath cycles (i.e. approximately 3–4 s). An obstructive apnea (OA) was scored if (i) the above criteria for apnea were fulfilled and (ii) out-of-phase movements of the chest and abdomen were present. A central apnea was scored if (i) criteria for apnea were fulfilled and (ii) no chest and abdominal wall movements were present. Mixed apneas were defined as those with both a central and an obstructive component, each lasting at least two breath cycles. In neonates and infants a mixed obstructive apnea index (MOAI) was calculated as the sum of mixed apneas plus OA per hour of TST. In children older than 12 months, hypopneas were also scored and the mixed-obstructive apnea-hypopnea index (MOAHI) determined: A hypopnea was scored if the nasal flow amplitude dropped by ≥30% of the pre-event amplitude lasting for at least 2 breaths and accompanied by a ≥ 3% desaturation. OSAS was defined as MOAI > 1 in infants and MOAHI > 1 in children older than 12 months (therefore denoted MOA(H)I in the following), and UARS [27] as more than 1 episode with nasal flow limitation/h, but without meeting OSAS criteria.

Desaturation events were visually confirmed to exclude spuriously low values. Events with a distorted pulse waveform signal within 7 s prior to their onset were considered artefactual and excluded. The number of desaturation events to < 80% SpO₂ was counted and expressed as desaturation index, defined as events per hour of TST (DI₈₀).

Children who had no sleep study performed because their upper airway obstruction had been so severe that they had arrived at our department already with an ET tube or tracheostomy in place were arbitrarily (and conservatively) assigned a MOA(H)I of 30 and a DI₈₀ of 3 for statistical analysis of their sleep study results.

After diagnosis, infants were admitted and monitored in the neonatal intensive care unit where they also underwent a baseline cardiorespiratory sleep study and fiberoptic nasopharyngoscopy without sedation to assess the type and localization of the UAO. This endoscopy usually took only 4–5 min. Children > 1 year were admitted to our pediatric sleep laboratory and fitting of the modified TPP was undertaken in the outpatient clinic.

Next, children had a maxillary imprint taken with a custom-made impression tray using alginate (Tetrachrom-Super-Alginat, ISO 1563, Klasse B, Typ I, Kaniedenta, Herford, Germany). This imprint covered the entire hard palate, the alveolar ridges and the vestibule. This procedure was carried out in the neonatal intermediate care unit under cardiorespiratory monitoring without sedation, but with a nasopharyngeal airway in place and in the presence of an experienced neonatologist. Older children had their imprints taken in the department of orthodontics without monitoring and nasopharyngeal airway in place. Then a plaster cast was produced using high precision dental plaster (Girodur Type IV, Synthetic Superhard Stone Plaster for Sectioned and Master Models DIN EN 26873, white, Girrbach Dental GmbH, Pforzheim, Germany). Using this cast, appliances were made from hard acrylic (autopolymerizing methylmethacrylate, Orthocryl, Dentaurum, Pforzheim, Germany). The TPP consisted of a palatal part that covered the hard palate and the cleft as well as the alveolar ridges and a velar extension of approximately 3 cm in length. The shape of the velar extension was modeled from dental wax and attached dorsally to the plaster cast. The length and the angle of the extension were chosen so that it was adjacent to the dorsum of the tongue. It was manufactured using a blue color to facilitate endoscopic evaluation in situ. After polymerization this prototype was polished using standard techniques.

Once a prototype of the plate was ready, infants had a repeat endoscopy to adjust the length and angle of the velar extension. The tip of the extension descended down to the vallecula epiglottica and the angulation was responsible for the anterior shifting of the base of the tongue and erection of the epiglottis, thereby widening the airway. Next, a tube made from hard acrylic was attached to the velar extension of the plate to form an artificial airway, followed by a repeat endoscopy with the endoscope introduced through this airway. If the airway appeared endoscopically and clinically open, the prototype was finished and a strengthening wire incorporated into the extension to safeguard the device against mechanical failure (Fig. 2). Two days later, its effectiveness was assessed by a second sleep study. If this sleep study still showed a MOA(H)I > 1, the plate was modified. Treatment in infants also comprised appropriate feeding techniques (finger feeding and Playtex Drop-Ins®, Playtex Products, Edgewell, North Bergen, NY, USA) and an orofacial stimulation therapy according to Castillo Morales®.

In infants and at the beginning of treatment, appliances were worn continuously and only removed for cleaning purposes. After 3–4 months of continuous treatment and in older children, the TPP was often only applied at nighttime. It was held in situ by adhesion and suction, by an adhesive cream (Blend-a-dent Super Haftcreme, Procter & Gamble, Schwalbach, Germany), and by extra-oral bows attached to the plate and fixed on the face with adhesive tape (Steri-Strip, 3 M Health Care, St. Paul, MN). Fitting of the plate was regularly controlled by the nursing staff. If the palatal part became too small, a new TPP was produced and fitted. After discharge, patients were seen at 6–8 weekly intervals at the orthodontic outpatient clinic; the next sleep study was performed 3 months after discharge. In general, new plates became necessary if a notch appeared on the alveolar ridges or sleep study results deteriorated again, which was often after approximately 3–6 months. In older children new plates became necessary with tooth eruption and skeletal growth.

Results are reported as median and range. Comparisons between sleep study results were done using software (Statistical Package for the Social Science, Version 18, IBM, New York, USA). For the analysis of sleep parameters, the Wilcoxon signed rank test was used; correlations were assessed by Spearman’s rank correlation coefficient.

In Germany, retrospective audits of own patient data, aimed at quality improvement, do not require ethical approval.

Twenty-five patients had an initial sleep study comprising ≥3 h of TST, of which 14 had OSAS (14 with a MOA(H)I > 1, 10 of these with a MOA(H)I > 3), and 5 had UARS. In 5 children, no sleep study was possible because of clinically severe OSAS: one had received a tracheostomy prior to admission, and in 4, OSAS had been so severe that they had already been intubated at their referring hospitals. In the remaining 4 children, no sleep study was performed; none had clinical signs suggestive of OSAS.

In 4 children we were able to study the aggravation of UAO with age: 2 children presented with Pfeiffer-, 1 with Crouzon- and 1 with Apert syndrome. The latter had an initially normal sleep study (MOAI = 0) as a neonate, but a repeat sleep study, performed at 4 months of age, showed a MOAI of 14.6. In the remaining patients split night sleep studies were performed (with and without ongoing treatment). In all 4 patients, UAO worsened with age and a positive correlation between MOAHI and age was found in repeat split-night sleep studies (Spearman’s rank correlation coefficient r_s = 0.66, p = 0.004).

Prevalence of OSAS varied by diagnosis (Table 1). Considering the total group, OSAS prevalence was 56%, that of UARS 15%, together yielding an SDB prevalence of 71%. In identified SCS this proportion varied between 80% in those with Pfeiffer-, Crouzon- or Apert syndrome, and 20% in those with Saethre Chotzen- and Cole Carpenter syndrome. Considering all patients with SCS, prevalence of SDB was 78% in those with and 22% in those without midface hypoplasia.

Out of the 19 children presenting with manifest OSAS, 11 were infants, three were toddlers, one was of preschool and 4 of school age.

Thirteen children (68%) underwent palatal plate treatment for their OSAS (Fig. 3), all but one via a modified TPP that had a tube attached to the pharyngeal spur (Fig. 2); in 9 patients this treatment was started during their first year of life (Table 2). Median duration of hospital stay in these patients was 60 days (range 10–104).

These are shown in Table 3. One patient initially had no obstructive apnea, but severe oxygen desaturations (DI₈₀ = 10), reaching down to 25% SpO₂. During TPP treatment, sleep study results improved in all patients (Table 3).

Weight and length were determined upon admission and discharge or at the next outpatient visit in all children receiving palatal plate treatment except for one toddler (n = 12) and expressed as standard deviation scores (SDS). Median SDS was < 0 for both parameters upon admission and decreased during treatment, although median SDS continued to be < 0 (Table 4). Differences were statistically significant for weight gain.

Out of the 9 patients with Apert-, Pfeiffer- and Crouzon syndrome, treatment could be discontinued in 1 child at age 11 years following surgical midface advancement. Of the remaining patients, 5 are continuing to wear their plates until midface surgery is planned at 12–14 years of age. One child (with Pfeiffer syndrome, see above) died, another also with Pfeiffer syndrome ceased to tolerate the TPP after 5.5 years of uneventful treatment and was continued on CPAP treatment. The remaining child underwent tracheostomy for severe respiratory distress due to recurrent mucus plugging of the tube attached to his palatal plate. (Fig. 3).

Ten patients received respiratory support via CPAP for some time mostly using a full-face mask: they had already undergone midface distraction at an early age, but suffered from residual OSAS (1), used CPAP on parental request (1), had started CPAP at another hospital (1) or used CPAP only temporarily for mild OSAS (1). Five patients were successfully switched from CPAP (2 ineffective nasal CPAP, 3 CPAP as bridging treatment) to TPP. The remaining patient was that mentioned above who suddenly ceased to tolerate TPP treatment (Fig. 4).

Three patients underwent surgical midface advancement at 6, 11 and 12 years, respectively: 2 had a tracheostomy in place, the third underwent surgery after 4.3 years of TPP treatment. Either treatment could be discontinued after surgery. The youngest patient re-developed OSAS 8 years following midface advancement and was then treated with CPAP.