Towards Reaching Consensus in the Diagnosis and Management of Infants With Robin Sequence

Abstract

For good reasons, most recent advances in clinical care have been based on evidence from randomised controlled trials (RCT). However, there are situations where RCTs are impractical, particularly in rare diseases, where it is often difficult to obtain sufficient patient numbers to achieve statistical power for definite results from parallel group trials. This has led to treatments becoming popular without a proper scientific basis. A prime example for this is (Pierre-)Robin sequence (RS). Here, seeing a baby fight for breath, combined with these infants having an easily visible anatomic reason for their distress (i.e., a retropositioned or underdeveloped mandible) has resulted in interventions that may seem logical, but lack evidence for their effectiveness. One such intervention, still widely used today as a first-line treatment, is prone positioning, first suggested by Dr. Robin himself in his 1934 landmark paper on “glossoptosis due to atresia and hypothrophy of the mandible” [1]. He believed that in the prone position “the tongue naturally falls forward by its own weight”, thereby widening the narrow upper airway. This may seem logical even today, and compared to some surgical alternatives tried over the years is certainly a much less invasive option, yet several longitudinal studies have since shown that, although it improves symptoms of obstructive sleep apnea syndrome (OSAS) in some infants to some extent, a considerable proportion of infants with RS continue to have moderate to severe OSAS with this seemingly effective and harmless intervention [2, 3].

Moreover, placing infants prone may not only be of limited effectiveness, but may also not be at all harmless, as it is associated, at least in healthy infants, with a 13-times increased risk of sudden infant death syndrome (SIDS) [4]. Given the rarity of both SIDS and RS, we will never be in a position to prove that the prone position will increase the risk of sudden death to a similar extent as it has been done in infants without RS, but then there is also nothing to suggest that RS infants should be exempt from the (yet unknown) mechanisms mediating the association between the prone sleeping position and sudden infant death. This is why in our center no infant is discharged with a recommendation to use prone positioning during sleep, although we apply it on an interim basis after admission, before starting a more effective and definitive treatment.

The above is just one example of why it may be difficult to reach consensus on the clinical management of infants with RS. Such consensus, distilled into guidelines, however, is desperately needed given that RS can often be a life-threatening diagnosis requiring immediate intervention, yet only 15%–30% of infants are already diagnosed antenatally [5, 6], thus creating a situation where parents may have to make far-reaching decisions on their baby's treatment without being able to weigh the risks and benefits of the different approaches offered.

Against this background, it is commendable that Goel et al. undertook the effort of reaching consensus among their local colleagues dealing with these infants (and also parents of affected infants) on an incremental plan for diagnosing and treating upper airway obstruction in these infants. In the study published in Acta Paediatrica, they report the results of this endeavour. Noticeably, they took a three-tiered approach, with first reviewing the evidence, then seeking consensus among their local and national colleagues and finally also taking parents' views on board [7]. The novelty of their approach lies in its incremental nature, and that it focuses on both diagnostic modalities and various treatment options.

While this review was carefully crafted, it cannot cure the fact that evidence collected for several diagnostic and therapeutic techniques is rather scant and at times somewhat biased. For example, they quote a study from Gozal's group to support their claim that cardiorespiratory polygraphy (PG) will significantly underestimate the Apnea-Hypopnea Index (AHI) compared to full polysomnography (PSG), yet the study quoted to support this claim was done in children at a mean age of 7–8 years, not in infants [8]. Admittedly, in studies from our own and a French group, the obstructive apnea hypopnea index (OAHI) was also higher than the obstructive apnea index (OAI) that does not take hypopneas into account, yet in all 3 studies, this did not result in a change in treatment decisions [9, 10]. In our experience, PSG involves much longer wait times, is more costly to perform and more stressful to an already compromised infant, so that we only perform a PG in most infants referred with RS. Thus, we wish to question the validity of the recommendation to perform full PSG in every infant with RS.

Also, some statements would have benefitted from providing the underlying evidence. For example, while there is evidence to suggest that a face mask used for applying continuous positive airway pressure (CPAP) may lead to midface hypoplasia [11], Goel et al. claim that the same is true for High-Flow Nasal Cannula (HFNC) use, yet without providing any evidence.

Then there is the question of what to expect of the various treatments offered, i.e., merely that they bridge the narrow upper airway or that they correct the underlying anatomy? For example, the use of a nasopharyngeal airway or CPAP may bridge the narrow airway, yet may do little for acquiring a physiologic functioning of the neuromuscular structures involved. This is not just of theoretical concern, but given the form-follows-function paradigm [12], achieving a normal tongue position and swallowing function may be decisive for achieving mandibular growth. This, for example, may explain the mandibular catch-up growth recently reported for the Tübingen Palatal Plate (TPP) therapy, where the facial profile continued to approximate that of healthy infants in the second half of the 1st year of life, that is, well beyond the 4- to 6-month period where the TPP had been used [13]. One of the differences between the reports on the TPP and other centers' experience is that > 90% of infants treated with the former approach are fully orally fed while showing catch-up growth for weight by the time they reach their first birthday [14].

The problem then is, how shall we identify an optimal treatment paradigm if there are hardly any trials comparing different approaches? To our knowledge, the only randomised trial in this field compared the TPP against a sham procedure (demonstrating superiority of the former), yet this was a small crossover trial with an observation period of only 48 h [15]. Another approach is to compare outcomes from cohort studies. Here, agreement on a set of parameters considered relevant by all experts in the field, as now attempted by Goel et al., would be extremely valuable. For example, the Boston group recently compared their respiratory and feeding outcomes for RS infants treated with mandibular distraction osteogenesis against our data obtained in infants with the TPP. Here, similar respiratory outcomes were found, while growth and oral feeding success were better with the TPP [16]. A caveat, however, was that Resnick et al. used the OAHI as their primary outcome, while we used the OAI (excluding hypopneas).

Nonetheless, we need such data, ideally collected prospectively, to further facilitate consensus based on data rather than opinion. As rightly pointed out by Goel et al., agreeing on a common dataset for assessing treatment outcomes will be a first step in this direction.

The authors declare no conflicts of interest.