Discussion
Our study suggests that a high proportion of the children deceased from SUDI have a particular orofacial anatomy. Following our results, the height of the hard palate was significantly higher and the width significantly lower in the deceased children than in the living children, reinforced by the calculation of the height/width ratio, which seems to be an effective reflection of these anatomical specificities. This quantitative approach was also confirmed by a subjective evaluation of the ogival palate, which showed a significant difference in frequency between the two groups. The high significance of an ogival palate in our results supports the clinical evaluation first made by Rambaud et al. in (6) and is a first step to define it with quantitative measurements. Thus, it strongly reinforces the hypothesis that an ogival palate is associated with SUDI.
Craniofacial modifications are widely studied for their association with OSA (15-17). The optimal functioning of the upper airway (i.e., proper suction and swallowing, as well as nasal breathing) depend on many factors among which normal growth of the facial structures is one of the most important (18). A narrow nasomaxillary complex with an ogival palate can be associated with nasal obstruction and mouth breathing. The misuse of the nasal cavities and mouth breathing may consequently lead to dysfunction of the upper airway muscles, which may exacerbate the abnormalities of craniofacial structure (18-20). An orthodontic correction of this morphotype is one of the effective treatments for OSA, especially in children and young adults (20-22). Thus, if a relation between the orofacial structure and a breathing-related sleep disorder (including OSA and upper airway resistance syndrome) has been established over the years, the physiological process which leads to the morphological variations of the upper airway is debated. Indeed, it remains unclear whether the anatomical changes are the cause or the consequence of the obstructive sleep disorder (23, 24). The palatal morphology is resulting in different factors, which are particularly important during the first months of life (25). Feeding and sucking habits influence the growth of the palate through the forces involved in chewing and swallowing (26). Position and size of the tongue, length of the tongue frenula, strengthens of the masticatory muscles of the mastication are many factors susceptible to impact the palatal development (27, 28). A high vaulted palate has also been described in preterm infants (29). It could be explained by several factors, such as immature swallowing and sucking functions and/or prolonged orotracheal intubation (29, 30). The possibility of the ogival palate as an inherited phenotype must also be considered as a strong hypothesis, as OSA in families has been previously studied (7, 31). Antenatal and neonatal studies could help foster a better understanding of the origins of this anatomical disposition.
The link between OSA and SUDI was first suggested in the 1970s (9, 32-34). The descriptions of the narrowness, obstruction, and increased resistance of the upper airway presented in the literature have suggested an association between OSA and SUDI (31, 35). Dysfunction of the central nervous system has been suspected, too, through alteration of dysautonomic functions, especially in preterm infants; however, this dysfunction is probably less predominant as a lethal process (36, 37). Mechanical asphyxiation that is secondary to an acute upper airway obstruction is indeed recognized as one of the most common mechanisms of death in very young infants, especially suspected when intrathoracic petechiae are numerous (32, 34, 38). The role of airway obstruction is reflected by the effectiveness of prevention campaigns promoting proper sleep environments and sleep position, which mainly aim to reduce possible accidental asphyxiations (39). Thus, it becomes easy to consider that an unfavorable orofacial structure may be a major predisposition for acute airway obstruction when triggered by upper airway infection, by sleeping in a prone position, and/or by using inappropriate bedding (40).
In the current literature, the studies mostly concern school-age children or adults, and the clinical presentation of OSA in very young children is poorly described (41, 42). None of the parents of the deceased children in our study spontaneously described strong signs of breathing disorders during sleep, such as snoring, agitated sleep, or mouth breathing. However, these clinical data may seem banal or insignificant by parents and they are usually not actively asked about during interviews with parents. This information could be more easily gathered by using a standardized questionnaire that asks about OSA symptoms, as detailed by the International Classification of sleep disorders (snoring, obstructed breathing, movement arousals, neck hyperextension during sleep, inward rib-cage motion during inspiration) (43).
This study provides new perspectives to better understand and prevent SUDI. First, studying the orofacial structure in victims of SUDI may help researchers to better understand the protective role of pacifiers or dummies. Many hypotheses have been ventured: pacifiers may help to increase blood pressure during sleep but it is also supposed to enlarge the upper airways thanks to the genioglossus contraction and the mandibular movements (8, 44-46). In addition, the repeated suction and swallowing induced by the use of a pacifier could stimulate the growth and enlargement of nasomaxillary complex and the effectiveness of upper airway muscles (16, 47, 48). The correlation between palate structure and pacifier use in very young children needs to be evaluated in deceased and living children, with long-term cohort studies. Second, abnormalities in oral development could be one explanation for the increased risk of SUDI in preterm infants (49). Preterm children, especially boys, are subject to the alteration of the palatal morphology, which may be increased by lower gestational and longer orotracheal intubation (29). An attentive follow-up of palate structure, suction reflexes, and sleep breathing modality in these children could be an effective way to prevent the development of an ogival palate and may limit the risk of premature death.
Our study presents some limitations. First, the retrospective and monocentric design and the limited size of our sample limit the generalizability of the results, even if the magnitude of the effect allowed for a high statistical significance. A greater number of cases would have facilitated a subgroup analysis, especially regarding the age (less or more than one year). Second, the radiologists were not blinded to the status of the patients. A blind interpretation would have strengthened the results. Third, the choice of a case-control design might be criticized since it implies the comparison of deceased and alive children. However, we believe this design remains the best to investigate our research question with acceptable feasibility. Indeed, the incidence of the SUDI being, fortunately, low, a cohort study would need to recruit tens of thousands of children followed up at least 1 year to identify as much as the 32 presented cases, which would be difficult to implement, and ethically very arguable regarding the associated X-ray exposure. Furthermore, we are confident that the quality of the measurement remains the same, as the death of the child is not expected to modify the assessed radiological parameters. Fourth, the choice of the “controls” remains also an issue. We chose a 1:2 design as it increased the statistical power, and the representativeness of the control population, compared to a 1:1. But ideal controls would have been matched not only for the same age and sex but also with regard to their living environment and other identified risk factors for SUDI such as sleep environment.
The perspectives offered by this preliminary work are numerous. The radiological evaluations of other parameters, such as nasal piriform aperture, nasal septum deviation or mandibular position and measurement, could provide a complete description of the morphotype of children suddenly deceased. Subjective analysis of three-dimensional reconstructions of the hard palate could be very helpful to improve the evaluation of this parameter by different specialists (pediatricians, radiologists, general practitioners). Complementary studies on the correlations between CT images and clinical observations are essential in both living and deceased children to characterize possible subclinical OSA. The frequency of an ogival palate in the normal population of children under 2 years of age, which remains currently unknown, need also be determined, for a more precise interpretation of the frequency of the ogival palate in infants who died unexpectedly.
