Pierre Robin and breathing: What to do and when?

Robin sequence is characterized by mandibular retrognathia, airway obstruction, and glossoptosis; 80%–90% also have a cleft palate. Various treatment approaches exist, and although controlled studies are rare, objective assessment of treatment outcomes that address the leading clinical issues, namely obstructive sleep apnea and failure to thrive, are essential. Sleep‐disordered breathing may be detected using cardiorespiratory polygraphy or polysomnography. Pulse oximetry alone may miss infants with frequent obstructive apneas, yet no intermittent hypoxia. Among conservative treatment options, the Tubingen Palatal Plate with a velar extension shifting the tongue base forward is the only approach that corrects the underlying anatomy and that has undergone appropriate evaluation. Of the surgical treatment options, which are not necessarily the first line of therapy, mandibular distraction osteogenesis (MDO) is effective and has been most extensively adopted. Notwithstanding, it is puzzling that MDO is frequently used in some countries, yet hardly ever in others, despite similar tracheostomy rates. Thus, prospective multicenter studies with side‐by‐side comparisons aimed at identifying an optimal treatment paradigm for this potentially life‐threatening condition are urgently needed.

with particular emphasis on sleep studies and growth trajectories as guides to intervention and outcomes.

| Are sleep studies truly the only diagnostic tool to assess severity of in RS infants?
Clearly, in those with life-threatening UAO, a sleep study is not required to indicate the need for treatment. A large proportion of RS infants, however, exhibit more subtle clinical signs of UAO, often occurring only during sleep, and not necessarily leading to prominent oxygen desaturations ( Figure 1). However, is full polysomnography (PSG) strictly necessary to assess OSA severity? More specifically, is IH the sole contributor to the occurrence of long-term clinical problems? If this were the case, then IH can be readily detected using pulse oximetry. However, as frequent obstructive respiratory events not resulting in IH are also important determinants of RS-associated morbidities, PSG or other relevant diagnostic approaches may be more appropriate. In a systematic review comparing the diagnostic accuracy of various simpler alternatives to PSG in diagnosing OSA, cardiorespiratory polygraphy, but not oximetry, emerged as a valid alternative to PSG. 4 Of note, cardiorespiratory polygraphy is also the diagnostic tool that the first author employs in RS patients to guide treatment and monitor its effectiveness in the clinical setting at the University of Tuebingen in Germany, while the approach employed at the University of Missouri regularly utilizes PSG as the diagnostic tool. Notwithstanding, we should also allude to the fact that a recent American Academy of Sleep Medicine task force consensus failed to recommend the use of cardiorespiratory polygraphy as the gold standard, 5 such that individual center discretion and multidisciplinary team preference is probably advisable.
We should also remark that upon analysis of cardiorespiratory polygraphic studies, the focus is on the mixed-obstructive apnea index (MOAI), rather than the more commonly used obstructive apnea-hypopnea index (AHI), since the latter index requires the ability to detect electroencephalographic arousals, that is, requires a full PSG. Given that obstructive hypopneas are relatively uncommon in young infants, that is, are not substantial contributors to the obstructive AHI, 1,6 we consider the MOAI a more appropriate parameter for assessing OSA severity in RS infants.
Several issues that remain to be addressed and merit future evaluation include whether alternative methodologies used to quantify the severity of IH, such as the hypoxic burden 7 or autonomic probing using heart rate variability measures 8 may provide better estimates of OSA severity, and also aid in treatment decisions.
Of note, both of these measures would be available in either PSG or cardiorespiratory polygraphic recordings.
Would such an approach help in achieving a good long-term outcome? While not formally studied yet, we consider it encouraging that in a 7-year follow-up study of 34 consecutive children with isolated RS who had their postnatal airway obstruction identified and treated early and effectively (see below), 9 we found an intelligence quotient (IQ) within the normal range, which is in contrast to other reports in RS patients showing impaired neurodevelopment in a significant proportion of patients. 10,11 Similarly encouraging results are also reported from other programs employing a systematic approach to identifying and avoiding IH in RS infants. 12,13 Weight gain is the other measure of particular interest in RS infants. As stated above, poor weight gain is a frequent initial clinical manifestation in RS infants not requiring immediate intervention in the neonatal period, 14,15 and is also a well-known complication of OSA. 1 Failure to thrive is associated with impaired neurodevelopment, at least in preterm infants. 16 Weight gain may thus be a valuable parameter to monitor overall treatment effectiveness. 2 Based on the aforementioned considerations, we will review data on sleep study findings and weight gain outcomes as they are related with current therapeutic approaches in the management of infants with RS, while purposefully differentiating between conservative and surgical treatment.

| Prone positioning
This intervention was originally proposed by Robin himself, 17 who hypothesized that the narrow pharyngeal space in RS can be corrected by gravity moving the mandible forward in infants sleeping prone. In some case series, 50%-80% of patients with RS were reported as being satisfactorily treated by prone positioning only. 18,19 Furthermore, prone positioning was implemented by about 2/3 of the respondents in a recent survey focusing on interventions used in RS infants. 20 Very few studies however, objectively documented the effectiveness of prone positioning. A recent PSG-study in 18 infants with RS (mean age, 1.5 months) found a significantly higher sleep efficiency in the prone position, but no significant reductions in OSA severity. 21 Another study performed sleep studies in supine and nonsupine position in 27 cleft infants with a mean age of 6 months, in whom 56% had RS. Again, no significant differences emerged between the infants' positions regarding obstructive AHI or the nadir pulse oximeter saturation (SpO 2 ). 22 Given this apparent lack of effectiveness, combined with the fact that the prone sleep position is associated with a 14-times increased risk of sudden infant death, 23 prone positioning cannot currently be recommended as an intervention aimed at improving breathing in RS infants.

| Nasopharyngeal airway (NPA)
This simple device was first suggested by a British group, 24 and bridges the narrow pharyngeal space that characterizes RS infants by inserting an endotracheal tube into one of the nares so that its tip is placed immediately above the epiglottis (ascertained by endoscopy or x-rays). Of note, methods for estimating the required length of the tube without the need of endoscopy or x-rays, and securing it safely to the nose, have been described. 25 Sleep study data on the  with RS (three syndromic), with 20 being managed with a NPA and high-calorie nasogastric tube feeding. Infants were gradually weaned from both tubes under SpO 2 monitoring. During a mean duration of hospital stay of 60 days (range 25-162), 18 infants were reported to grow along their birth percentiles for weight, but only three infants were fully bottle-fed at discharge. Also, only 10 infants maintained their weight centile until the time of cleft repair. Unfortunately, no sleep study results were provided. 26 Another study reported weight gain in eight infants with a NPA (median age 50 days, range  days; four of these infants were also receiving supplemental oxygen).
Their weight gain increased from 86 g/week before to 255 g/week with the NPA, but no data on weight gain beyond hospital discharge or PSG results were reported. 27 The group that first described the use of a NPA also reported sleep study data on 63 infants managed with a NPA and indicated no residual UAO in 5, mild UAO in 39, and moderate UAO in 19 patients after use of the NPA for a median of 10 days; more detailed sleep study results, however, were not provided. 28 An algorithm to RS treatment that also involved the NPA was recently reported by a surgical group. They based their treatment decisions on sleep study results, and treated those with moderate or severe OSA with a NPA. If a repeat sleep study showed improvement (true for 10 of their 20 patients), the NPA was continued or combined with noninvasive respiratory support; the remainder were deemed to reflect failure of the NPA and underwent laryngoscopy and then progressed to tracheostomy or one of the other surgical interventions discussed below. 29 We should point out that a NPA exerts no stimulus on mandibular growth, nor does it favor the tongue to assume a more desirable horizontal position.
Thus, while certainly valuable as a temporary measure, more data on NPA's long-term effectiveness are needed before recommending this approach as a routine treatment option in RS infants.

| Pneumatic airway stenting using nasal continuous positive airway pressure (NCPAP) or intermittent positive pressure ventilation (NIPPV)
There are some case series on the use of NCPAP or NIPPV in RS. 30 In a single-center analysis of 81 RS patients, 7 (9%) were treated using NIPPV, starting at a mean age of 2 months and lasting for a mean of 17 months. Reported benefits included a decrease in the proportion of time spent with SpO 2 less than 90% from a mean of 14% to 1% (at a mean airway pressure of 8 cm H 2 O) and a decrease in mean transcutaneous CO 2 from 57 mmHg to 31 mmHg. All seven infants were discharged home with the device, which was used at home for an average of more than 8 h per day. No facial side effects were reported. 31 The same group also reported on the successful use of CPAP in neonates with RS. Of 44 such neonates, nine were successfully managed with CPAP, four underwent tracheotomy, while the remaining received a recommendation of prone positioning. 32 There is anecdotal evidence, however, that long-term NCPAP use in young children may result in mid-face hypoplasia. 33 This is particularly relevant to RS patients who may suffer from a hypoplastic maxilla anyway, 34 and yet this potential side effect has not yet been specifically studied systematically. Nonetheless, in our experience, CPAP is valuable as an intervention applied between admission and implementation of a more curative treatment.

| Tuebingen palatal plate (TPP)
The TPP consists of a palatal base plate that covers the hard palate and the cleft, as well as the alveolar ridges, and supports a velar extension (spur) of individual length (approximately 3 cm) that ends just above the epiglottis and shifts the base of the tongue forward, thereby opening the airway and correcting the underlying glossoptosis ( Figure 2). Fitting of the TPP is controlled using fiberoptic nasopharyngoscopy without sedation. During the endoscopic procedure, which usually takes only 1-2 min, the tip of the velar extension is checked and its angle, responsible for the forwardshifting of the tongue base, is adjusted so that it pushes the latter sufficiently forward to erect the epiglottis, thereby widening the pharyngeal space. Effectiveness of the TPP is regularly ascertained using sleep studies, the first being performed immediately before treatment onset, and subsequently before discharge, and at least 3 and 6 months after initiating TPP therapy ( Figure 1). The MOAI in these studies should be less than three events/h; if it is more than three events/h, the angle or length of the velar extension is modified.
Treatment is usually discontinued around 6-8 months of age, depending on sleep study results (which should show a MOAI ≤ 1 event/h) and the facial profile at the time.
The TPP is the only device used to treat RS infants that was evaluated in a controlled study design, that is, versus a conventional palatal plate used as a sham procedure. After 48 h of treatment, the median MOAI had declined from 13.8 events/h to 3.9 events/h with F I G U R E 1 Representative 2-min sections of polygraphic recordings done (a) at admission and (b) after 2 weeks of treatment with a Tübingen Palatal Plate (TPP) in a baby girl with isolated Robin sequence admitted at 4 weeks of age. Shown are (from top to bottom) nasal pressure, thoracic and abdominal breathing movements, pulse oximeter saturation (SpO 2 ; in beat-to-beat mode), pulse waveforms, transcutaneous CO 2 and heart rate. The top segment shows 5 obstructive apneas of 2-7 s duration (grey bars; overall, the infant had a MOAI of 93/h in this recording). The bottom segment shows no respiratory event (the MOAI had fallen to 3/h). CO 2 was at 38-42 mmHg throughout. In parallel, oral intake in this infant had changed from 100% gavage feedings to 80% intake via bottle feeding (Playtex)), weight from the 25th to the 40th centile. Note that no desaturations occurred in conjunction with the obstructive apneas; these events would have been missed by a simple recording of SpO 2 [Color figure can be viewed at wileyonlinelibrary.com] the TPP (p < .001), while it remained unchanged at 14.8 events/h with the sham procedure. 35 In an uncontrolled continuation of the study that involved 15 RS patients (median age at onset of treatment, 5 days), median MOAI fell from 17.2 events/h to 1.2 events/h after 3 months of treatment with the TPP. Notably, all infants had their feeding tubes removed before hospital discharge, and continued to gain weight at a mean rate of 24 g/day at discharge, and 19 g/day at the 3-month follow-up. All infants continued to be exclusively fed orally at follow-up. 36 This beneficial effect of the TPP on feeding ability may be related to the fact that it not only opens the airway, but also corrects glossoptosis, thus helping the tongue to assume its normal, horizontal position. A normal tongue function, with a downward-movement of the tongue to a position below the nipple at the time of peak vacuum, is crucial to successful oral feeding, 37 which may explain some of the feeding difficulties seen in RS infants.
To demonstrate that the TPP is also effective in more severe phenotypes of RS, the first author's group compared the change in the MOAI before and during TPP treatment in 122 infants with isolated RS, including 55 infants with a MOAI of more than 10 events/h in their initial sleep study (median 29, interquartile range [IQR], 15-51). By the time of the 3-month follow-up, the MOAI had decreased to 0.2 events/(0-1.3), which was very similar to the results found in infants with mild OSA, that is, an initial MOAI of 3-5 events/h. 38 The TPP is also effective in syndromic RS. In a case series of 68 consecutive patients with syndromic RS admitted to the Tuebingen center over a 7-year period, 56 patients completed TPP treatment, while in 4, no TPP was indicated because MOAI was less than 3 events/hr or because of the presence of significant laryngomalacia; three infants (4%) ultimately required a tracheostomy. 39 The remaining five patients did not tolerate TPP treatment, mostly related to problems with swallowing (e.g., in those with CHARGE association or Wiedemann-Beckwith syndrome). In contrast to those with isolated RS, the type of TPP used varied in these patients, with 23 using a plate with a perforated tube (n = 20) or a ring (n = 3) attached to the pharyngeal extension to prevent collapse of the lateral pharyngeal or laryngeal walls. Thus, syndromic RS patients showed an more than 80% reduction in the MOAI between admission and discharge. Also, in the 46 infants with complete data, Z-scores for weight at discharge were comparable to those documented at birth; that is, postnatal growth failure was avoided. Furthermore, the number of infants who were fed via nasogastric tube decreased from 23 to 7. 39 Of note, TPP treatment was also associated with mandibular catch-up growth, as evident from longitudinal measurements of the Jaw Index (JI), defined as the alveolar overjet (in mm) times maxillary arch/mandibular arch (also measured in mm). 40 In 20 RS patients, the Tuebingen group determined the JI upon admission, discharge and 3 months after discharge, and found a decrease from 8.8 (IQR, 6.3-11.3) on admission to 2.1 (2.0-4.0) at the 3-month follow-up (p < .001), which correspond to JI values similar to those found in healthy infants. 41 The objectively determined data confirm the clinical impression of a rapid improvement in the facial profile occurring during TPP treatment (Figure 3), although it should be kept in mind that the JI has not yet been validated.
Of note, the TPP approach has also been adopted by others. [42][43][44] For example, in a three-center study involving 49 consecutive RS infants, there was an overall decrease in median MOAI from 15.9 events/h (IQR, 6.3-31.5) on admission to 2.3 events/h (1.2-5.4) at discharge, thereby illustrating the similar outcomes to those initially reported in the authors' center. 45 Also, a group from Berlin reported F I G U R E 2 Tübingen Palatal Plate (TPP) production. Based on digital scanning of the oral cavity, a prototype is produced using CAD/CAM (left) and then assessed in the patient. If fitting is approrpriate, the final plate is manufactured, and includes a strengthening wire inside the spur and extraoral bows for better retention of the plate when the TPP is in place POETS ET AL.
| 1891 a case series of 132 RS infants who could be successfully managed with a variant of the TPP, and who sustained normal weight gain even though only four of these infants were discharged with a feeding tube. Unfortunately, no sleep study results were reported. 43 Since January 2019, the Tuebingen group has introduced a fully digital workflow, including scanning of the oral cavity (Trios 3, 3shape A/S) and a CAD/CAM workflow to build and print the TPP (Figure 3), so that necessary changes can be carried out more precisely and reproducibly compared to the conventional handcrafted method. 46 Taken together, these data show that the TPP is effective in both mild and severe RS, as well as in isolated and syndromic RS, results in resolution of OSA, promotes a normal weight gain trajectory, and helps to avoid a tracheostomy in all infants with isolated, and 85% of those with syndromic RS. 47 It has also been successfully applied by several RS teams and is associated with mandibular catch-up growth, that is, helps to correct the underlying anatomic defect in RS by allowing the tongue to assume a normal position. 41

| Tongue-lip adhesion (TLA)
This surgical approach was first proposed by Shukowsky, 48 and subsequently reported to result in better survival rates. 49 Success rates for this intervention, based on clinical criteria, vary, 19 and data on PSG results are scarce. Two studies compared sleep study results and clinical data in infants treated with TLA versus mandibular distraction osteogenesis (MDO; see below). Results showed that MDO was more effective in improving sleep study abnormalities, but MDO-treated patients were potentially also more prone to develop complications (e.g., wound infections; fracture of the bone anchoring the distractor). 50,51 In another follow-up study, 10 of 11 RS patients required additional interventions for airway or feeding problems following TLA. 52 Moreover, complications such as wound infections, adhesion dehiscence or scar formation of the lip and floor of the mouth were reported to occur in about 20%-25% of the patients. 15,19 Thus, it remains unproven whether TLA can indeed be recommended as a good surgical approach for most children with RS. 53

| Mandibular distraction osteogenesis (MDO)
This procedure aims at correcting the mandibular hypoplasia in RS by performing a bilateral vertical mandibular osteotomy, and placing pins for a multi-vector external (or internal) distractor. Beginning a few days after the operation, distraction is usually done at a rate of 1-2 mm/day 54 until the patient has a Class III occlusion (see below for more details regarding the MDO procedure). This overcorrection is considered necessary to sustain an adequate airway in case a (partial) relapse occurs following distraction. 55 Devices are usually removed 4-8 weeks after the end of the distraction period. Using an internal single-stage self-resorbable device is also possible. 56 Despite many studies reporting clinical success with the MDO technique, only a few have reported changes in weight gain. One study involving 10 patients even reported a decline in growth rate in seven patients in the first 12 months after MDO, despite continued tube feeding in three patients. 57 This may be related to the fact that dysphagia, often seen in RS, is not corrected by the MDO procedure. 58 A study in 17 infants reported full oral feeding being achieved in all infants by 3.5 months postoperatively, but provided no growth data. 59  Catch-up growth following MDO onset has also been reported in a retrospective analysis from Yale, in parallel to an increase in the proportion of infants being fully orally fed from 50% pre-operatively to 73% at 6 months after surgery. 62 Several case series have reported PSG results before and after MDO. 54,58,59,[63][64][65] All noted improvements or normalization of the AHI or the respiratory disturbance index (RDI) in the majority of the patients, but only three of the studies provided detailed PSG results.
One of the latter studies included seven patients, and reported a decrease in AHI from 60 ± 7.3 events/h before distraction to 1.6 ± 1.6 events/h at the end of the expander activation 6-8 weeks post-op. 54  while the third and largest study reported a decrease in AHI from a mean of 39.7 events/h (4.5-177) to 5.8 events/h (0-34) in a chart review of 28 infants in whom the MDO procedure was performed and changes in OSA documented by PSG. 65 The authors describing the self-resorbable device mentioned above also reported resolution of UAO (i.e., an RDI less than 2.0 events/) in all nine of their 14 nontracheostomized patients who were stable enough to tolerate a preoperative sleep study. 56 At the University of Missouri, mandibular distraction osteogenesis is performed using an internal device. 66,67 Although resorbable devices were initially used 56 to avoid secondary surgery to remove the hardware, the team has since moved to using metallic devices as the use of metallic, nonresorbable internal distractors was found to result in more effective and lasting mandibular advancement and resolution of OSA. An extensive conservative management protocol is used before MDO, including a craniofacial CT scan, laryngoscopy, PSG, and a swallow study. MDO is only considered if a trial of prone positioning and oxygen supplementation has failed. The proportion of referrals ultimately receiving MDO is less than 20% of those referred, and MDO is mainly pursued to avoid tracheostomy placement and facilitate feeding and weight gain. 64,68,69 Surgery is performed through bilateral submandibular external incisions ( Figure 4). Oblique corticotomies of the mandible are made; the CT scans are used to position the osteotomies to avoid injury to After a 48 h latency period, distraction proceeds at a rate of 2 mm/day with rhythm of 0.5 mm four times per day. 71 After 10 mm of distraction, rate decreases to 1 mm/day at 0.5 mm twice per day.
Distraction proceeds until overcorrection is achieved, which usually takes 2-3 weeks. After a further 8-12 weeks of consolidation, the hardware is removed.
Using this method, excellent success in avoiding tracheostomy, correcting OSA, and facilitating weight gain has been achieved. 68 Complications have been minimal, primarily involving temporary marginal mandibular palsy. 72 Patients are followed long term with clinical assessments and PSG.
In meta-analysis, however, complications such as pin site infections (in 18%), device failure (10%), persistent inferior alveolar nerve lesions (6%), and somewhat more rarely, damage to tooth buds resulting in long-term tooth loss, dentigerous cyst formation, or relapse of UAO symptoms and temporo-mandibular ankylosis were reported rather commonly. 73 As previously mentioned, it is puzzling that MDO seems to be the treatment of choice in some countries, while it is hardly ever used in others, despite equally low tracheostomy rates. 2,20

| Mandibular traction
This procedure, first described in 1937, 74 involves fixation of a percutaneous parasymphysial circumferential wire to the mandible, with traction being applied by suspension weights (50-200 g), and is left in place for 4-6 weeks. Its effect on UAO is unknown, and it requires long-term immobilization of the patient during a critical developmental period. Also, a long-term follow-up study showed persistence of retrognathia on cephalometric x-rays. 75 Therefore, based on the favorable outcomes using alternative interventions, this procedure is not currently recommended.

| CONCLUSION
A review of the current management approaches for RS infants suggests that the TPP may be the conservative treatment of choice, and has a wide applicability in both isolated or syndromic cases as well as in those with mild to severe clinical manifestations. Importantly, the TPP procedure is associated with normal long-term development, at least in patients with isolated RS. When surgical interventions are adopted, MDO is currently the most frequent surgical procedure being conducted and appears to yield favorable outcomes, albeit in the absence of high quality evidence. Nonetheless, what is needed is a head-to-head comparison between conservative approaches such as CPAP, the NPA or the TPP and surgical treatments such as MDO or TLA, and an identification of unique phenotypic features that justify the selection of one approach versus the other. As randomized trials are difficult to realize in this rare condition, this may be the only realistic way to reaching agreement on these differing clinical management pathways. Independent of the treatment approach being selected, repeated serial PSG or cardiorespiratory polygraphic recordings and careful monitoring of weight gain are critical for correct clinical decisions to be implemented at every stage. Further refinements of the information provided by the multichannel recordings may assist in improving therapeutic and outcome targets in the near future.

ACKNOWLEDGEMENT
We are grateful to the nursing and medical staff at our institutions for their dedication and support in caring for these babies, and to Wolfgang Buchenau, MD, for technical assistance in preparing this manuscript. Open Access funding enabled and organized by Projekt DEAL.