Objectives. To test prospectively the hypothesis that an infant car seat modification to allow the infant’s head to rest in a neutral position on the trunk would prevent narrowing of the upper airway and thus reduce oxygen desaturation in preterm infants who are restrained in car seats.
Methods. Seventeen preterm infants who were approved for discharge were evaluated in a car seat for newborns, with and without a foam insert that provided a slot for the back of the infants’ head. Respiration timed inspiratory radiographs for assessment of upper airway dimensions were taken during quiet sleep in each position. Infants were monitored in each position for 30 minutes with continuous polygraphic recording of respiratory, cardiac, and nasal airflow activity and pulse oximetry.
Results. Placement of the insert in the car seat was associated with a larger upper airway space (mean ± standard deviation, 5.2 ± 1.3 vs 3.6 ± 1.4 mm). This radiologic improvement was associated with a significant reduction in the frequency of episodes of oxygen desaturation to <85% (1.5 ± 2.1 vs 3.5 ± 3.5 episodes/infant), of bradycardia <90 bpm (0.1 ± 0.3 vs 1 ± 1.7), and of arousal (median [25th, 75th], 2.5 [1.3, 4.0] vs 5.0 [4.0, 7.0]).
Conclusions. The cause of oxygen desaturation in preterm infants who are restrained in car seats is multifactorial. The present data strongly support the hypothesis that flexion of the head on the body is a significant contributor to these episodes and that the mechanism is posterocephalic displacement of the mandible, leading to narrowing of the upper airway. Critically, this pilot study demonstrates that the frequency of episodes of desaturation in a standard newborn car seat can be substantially reduced by placement of a simple foam insert that allows the infant to maintain the head in a neutral position on the trunk during sleep.
Premature infants are now widely recognized to be at high risk of oxygen desaturation and secondary central apnea while restrained in infant car seats.1,2 The American Academy of Pediatrics recommends that infants be formally tested before transport home in car seats.3 Despite use of rear facing, reclining car seats, up to 30% of premature infants may fail such testing,1 depending on the exact age of discharge. The only available clinical responses to this finding have been either to extend the hospital stay or to transport the infant entirely supine. The mechanism of this vulnerability has been unclear, and interventions such as sideways support of the infant’s head have proved ineffective.4
In 1976, Stark and Thach reported that small, preterm infants are vulnerable to hypoxia and apnea when their neck is flexed either by external pressure6 or spontaneously.5 They found that infants were particularly prone to head flexion when the infant had been placed in a more upright position.5 Reed et al7 concluded that this airway obstruction was attributable to the tongue’s impinging on the posterior pharyngeal wall.
It is a common parental and clinical observation that premature infants—and, indeed, some full-term infants—slouch the head forward when they fall asleep while restrained in car seats. The newborn infant has a prominent occiput, lying posterior to the line of the spine, which accentuates this tendency to flexion of the head.8 These observations suggest the hypothesis that that in some premature infants, the mechanism of oxygen desaturation in infant car seats may be excessive head flexion leading to restriction of the upper airway.8 To test this hypothesis, we examined the effect of a simple foam insert that when placed behind an infant in the car seat provides a slot for the head to lie in, allowing the head to maintain a neutral position on the spine, in the Frankfort plane,9 even during sleep. In this pilot study, we report the effect of placement of the insert on the infants’ airway dimensions, measured using respiration timed inspiratory radiographs,10 and on the frequency of episodes of oxygen desaturation, bradycardia, and arousal.
Preterm infants were recruited from the parent-infant nursery at National Women’s Hospital, Auckland. The parents of the infants were approached for consent for this study when the infants were eligible for discharge home. The study was approved by North Health Ethics Committee. The infants were studied in the radiology suite of National Women’s Hospital. The parent(s) and the pediatrician were in attendance throughout each study.
Car Seat Insert
A foam plastic insert was constructed such that when placed on the car seat behind the infant (Fig 1), space was left to accommodate the infant’s prominent occiput. This enabled the Frankfort plane (a line from the lower orbital margin to the auditory canal placed at 90 degrees to the spine) to be maintained. Infants were studied in an approved, back-facing young infant car seat. This car seat was modified by removing the “wing” on one side so that the horizontal X-ray beam would not be impeded (Fig 1). The X-ray plates were secured against the intact, left side of the seat.
Infants were tested 15 minutes after a feed. Each infant was studied for 30 minutes with the car seat insert in place and then for an additional 30 minutes after it had been removed. For the second monitoring period, the infant was lifted up gently and the insert was pulled out. In principle, it would have been preferable to randomize the order of study; however, the infants were consistently aroused by placement of the insert, whereas removal could be achieved without arousal, thus avoiding delays between the study periods. Because of time constraints, the sequential approach was adopted. The car seat was adjusted such that the infant’s back was lying at an angle of 40 degrees to the X-ray table. This angle was the mean that we observed when the car seat was placed in the front seat of a series of 10 cars (unpublished data).
Upper Airway Imaging
Inspiratory radiographs of the upper airway were taken, timed using a modified Graseby (Graseby Dynamics Limited, Herts, UK) MR 10-apnea monitor with the capsule of the monitor taped to the infant’s lateral abdomen at the level of the umbilicus.10 One radiograph was made in each of the 2 monitoring periods, when the infants were clinically quietly asleep. They were timed to inspiration as previous reports have shown that airway size varies with the phase of respiration, being smaller in inspiration than in expiration.10
Polygraphic recordings (Edentec, Edentec Mallinckrodt Inc, St Louis, MO) were made using terminals attached to record heart rate, respiratory rate, and excursion; oronasal thermistor for airflow; and pulse oximetry from the foot. In addition, a second oximeter was attached to the other foot for “real time” display. The Edentec polygraph was recorded for the entire time the infants were in the car seat. For this study, the number of the following events was recorded for each infant in each position: 1) oxygen saturation to <85%; 2) bradycardia to <90 bpm; and 3) arousal as defined by the combination of highly variable respiratory recordings, with increased heart rate and muscle activity. Episodes of obstructive apnea were identified on the basis of the combination of reduced airflow with increased respiratory effort. When persistent oxygen desaturation was observed on the real-time display, the infant was removed from the car seat and the study was terminated.
The effect of the insert on airway size was assessed by analysis of variance. Measurements of the upper airway at different levels and the effect of insert were treated as repeated measures. Comparison of the cephalometric measurements and the frequency of episodes of oxygen desaturation and bradycardia between the 2 recording periods was performed by Wilcoxon signed ranks test (SPSS, Chicago, IL). All data in the text are presented as mean ± standard deviation.
A total of 17 infants were studied, 9 female, 8 male, born at 32.0 ± 3.5 weeks, weighing 1792 ± 599 g. At the time of the study, the infants were 38 ± 40 days old and weighed 2472 ± 426 g. Ten of the 17 infants had had 1 or more episodes of apnea or bradycardia while in the hospital, but in all cases, no episodes had been recorded at least for the preceding week. With the insert in place, all infants were able to maintain their head in a neutral position (Fig 2A). When the insert was removed, in the majority of infants, the head tended to slump forward, with the chin pressed on the chest (Fig 2B).
The infants’ upper airspace space was wider when the infants were quietly asleep in the car seat with the insert in place compared with without the insert (overall mean ± standard deviation, 5.2 ± 1.3 vs 3.6 ± 1.4 mm; P < .001, analysis of variance; Fig 3). Fourteen infants had wider airways with the insert, whereas 3 infants showed no change.
Cephalometric analysis of the inspiratory radiographs showed that the distance between the space from nasion to the incisor in the lower jaw was reduced from 15.6 ± 3.5 mm to 12.5 ± 3.3 mm when the insert was removed (P < .001). Furthermore, the angle between the upper incisor to nasion to the lower incisor was increased from 9.1 ± 2.4° to 12.3 ± 3.5° (P < .001). These measurements indicate that on average, the jaw was being pushed backward and upward during quiet sleep in the car seat without the insert. Examination of individual radiographs demonstrated that the precise direction of displacement of the jaw was variable, with some infants showing a predominantly cephalic displacement, as illustrated in Fig 4, whereas in others, the direction of displacement was more posterior.
Sixteen of the 17 infants had satisfactory polygraphic recordings. In these infants, a total of 24 episodes of desaturation were identified with the insert in place, and 56 episodes were identified after the insert was removed. Overall, 8 of the 16 infants had at least 1 transient episode of desaturation with the insert in place, compared with 12 without the insert (P < .05, Wilcoxon signed ranks test). The insert was associated with a significant reduction in the frequency of episodes of oxygen desaturation to <85% (1.5 ± 2.1 vs 3.5 ± 3.5; P < .05, Wilcoxon signed ranks test) and of bradycardia <90 bpm (0.1 ± 0.3 vs 1 ± 1.7; P < .05). There were fewer episodes of arousal by the infants when the insert was in place, a total of 44 versus 84 without the insert (median [25th, 75th], 2.5 [1.3, 4.0] vs 5.0 [4.0, 7.0]; P < .005). Arousal occurred either before or during episodes of obstructive breathing in 8 cases with the insert and in 49 cases after removal of the insert (P < .001).
A range of breathing patterns were seen in the infants, including obstructive breathing (Fig 5B), hypopnea (Fig 5B), and periodic breathing. However, the difference between the groups was almost entirely attributable to a reduction in the frequency of obstructive apnea. With the insert in place, 3 episodes were identified as being associated with an obstructive pattern of breathing, and 21 were either nonobstructive or unable to be clearly classified. After removal of the insert, 26 episodes were identified as obstructive and 30 were identified as other types (P < .003, Fisher exact test).
The infant whose radiograph showed the most severe airway narrowing in the car seat without the insert momentarily had desaturation of oxygen to <60% and bradycardia to <70 bpm (Fig 5B). In contrast, 2 of the infants who showed repeated desaturation both with and without the insert in place had a pattern of periodic breathing. Discharge home of 1 of these infants was delayed. The parents of the other infants were counseled about the need for close supervision of their infant when restrained in a car seat.
The present data demonstrate for the first time that if premature infants who are restrained in a car seat are allowed to flex their head on the spine, then the upper airway is narrowed and that this narrowing is associated with a higher rate of episodes of oxygen desaturation and bradycardia and that the infants have evidence of more frequent arousals. This pilot study suggests that a simple foam insert is able to reduce significantly this airway narrowing.
In most developed countries, it is legally compulsory for children and infants to be transported in “approved” car restraints, to reduce the danger of rapid deceleration in a sudden stop or collision. Car seats for newborns are designed to be back facing so that the infant will be protected from rapid forward acceleration of the head. However, most of the infant car seats have been designed and tested using model infant dolls or mannequins, which do not reproduce the moveable anatomy of a living young infant. It is well established that when premature infants are restrained in such car seats, they demonstrate a high incidence of oxygen desaturation.1,2,11,12 The normal range of oxygen saturation in healthy term and preterm infants is reported to be 93% to 100%.13 In the present study, we used a lower threshold for desaturation (85%) to focus on clinically significant events.14 Consistent with the American Academy of Pediatrics recommendation that all premature infants be monitored in the car seat,3 all but 4 of the infants showed 1 or more significant episodes of desaturation in the car seat without the insert.
If the infant does develop episodes, then the only alternative is delayed discharge or the use of true horizontal car seats, which may not provide the same level of impact safety to the infant in some accident situations.15 It is a common observation that infants in car seats tend to slump forward or sideways. On the basis of this observation, in a recent study, infants’ heads were supported so that they could not move laterally—but forward flexion was not prevented.4 This did not improve oxygenation. These data are in contrast with the present study, in which forward flexion was specifically prevented.
We have previously shown that the infant upper airway anatomy is very different from that of the adult.16 The infant has a straight spine and a large head, the occiput of which protrudes back beyond the spinal line. The infant mandible is almost horizontal with an unstable temporomandibular joint. The mouth is filled by the tongue, with no teeth to keep the jaws apart, and the infant’s neck is very short. Consequently when the back of the shoulders and occiput are forced into line, the head must flex forward on the neck, and in most very small infants, the chin is pressed onto the chest. This in turn forces the unstable mandible upward and/or backward to carry the tongue onto the soft palate, narrowing the upper air space as in Fig 4B. We ensured that the nares were maintained clear of any obstruction throughout this study. These considerations suggest that 1 factor that might affect whether obstruction occurs may be the size of the infant’s tongue.17
In some of the infants in this study, the upper air space was almost obliterated in the car seat without the insert, and monitoring had to be interrupted when there was a severe fall in oxygen on the real-time oximeter, as was the case of the infant whose polygraph is shown in Fig 5B. When an insert was placed in the car seat allowing the infants’ head to maintain the normal upright posture on the spine, the radiographs showed normal upper airway space as in Fig 4A, and there was a marked reduction in the frequency of oxygen desaturation (Fig 5A). Consistent with this combination of findings, the effect of the insert on the infants’ breathing patterns was predominantly to reduce episodes of obstructive breathing with little effect on other patterns, such as periodic breathing.
This reduction in episodes of desaturation and of obstructive breathing was associated with a reduction in the frequency of arousal. Arousal can be attributable to many factors. In the present study, some arousals occurred in association with episodes of desaturation, as illustrated in the first episode in Fig 5B, whereas at other times, infants failed to arouse, as seen during the second episode in Fig 5B. This is consistent with a recent study that reported that behavioral arousal immediately preceded just 36% of the episodes of desaturation in infants who were sleeping prone.14 It is interesting that the second episode in Fig 5B in which arousal did not occur was associated with more severe desaturation, supporting the suggestion that more frequent arousals may provide some benefit for infants responding to potentially life-threatening events.18 In the present study, most of the change between the 2 infant observation periods was related to a reduction in the proportion of arousals that occurred before or during episodes of obstructive breathing, from 58% without the insert to 18% with the insert in place.
This study was conducted on premature infants who were considered to be ready for discharge home. It has been shown that even at term, some infants exhibit periods of oxygen desaturation when restrained in car seats.13,19 Furthermore, airway narrowing can occur during head flexion in older infants.20,21 These data, combined with the present results, suggest that airway compromise as a result of excessive head flexion in car seats probably applies to at least some more mature infants. This raises the possibility that this type of modification to avoid head flexion in car seats may be appropriate for many infants until they are able to sit unaided and are also able to mouth breathe easily, which is usually at approximately 6 months of age.
Four infants never developed any episodes of desaturation. It is striking and likely significant that these infants also showed no reduction in airway size on inspiratory radiography in the standard car seat. Conversely, however, in 2 cases in the present study, multiple episodes of desaturation occurred during monitoring both with and without the insert in place, in association with a pattern of periodic breathing. One of these infants was very premature at birth. These episodes occurred while a neutral airway position was demonstrated on radiography with the insert in place. Furthermore, their airway dimensions were comparable with the remainder of the group. This strongly supports the general concept that the cause of desaturation is often multifactorial; airway size and head position can be only 1 component of the problem.
Premature infants who were considered to be ready for discharge home were shown to have narrowing of their upper airways during quiet sleep when restrained in a conventional infant car seat. Consistent with many previous reports, under these conditions, many of the infants demonstrated episodes of oxygen desaturation and bradycardia and had disturbed sleep with frequent arousals. When an insert was placed in the car seat so that the head could be maintained in the neutral position on the trunk, their upper airway size and their vital signs were normalized, with a marked reduction in the frequency of desaturation, bradycardia, and arousal. Additional studies are required to evaluate crash safety, to establish how long it is necessary to continue to use the insert in infants who are born prematurely, and whether this approach may be beneficial for some term infants.
We gratefully acknowledge the financial support of the late H.B. Williams and the Turanga Trust. We thank Rosemary Stewart for skill in obtaining the radiographs and the parents whose infants participated in the study.
- ↵Bull MJ, Stroup KB. Premature infants in car seats. Pediatrics.1985;75 :336– 339
- ↵American Academy of Pediatrics, Committee on Injury and Poison Prevention and Committee on Fetus and Newborn. Safe transportation of premature infants. Pediatrics.1991;87 :120– 122
- ↵Roberts JL, Reed WR, Mathew OP, Menon AA, Thach BT. Assessment of pharyngeal airway stability in normal and micrognathic infants. J Appl Physiol.1985;58 :290– 299
- ↵Tonkin SL, Bennet L, Vogel S, Gunn AJ. Apparent life-threatening events associated with positional airways obstruction in infancy. Pediatr Res.2000;47 :120A
- ↵Gunn TR, Tonkin SL. Upper airway measurements during inspiration and expiration in infants. Pediatrics.1989;84 :73– 77
- ↵Bass JL, Mehta KA, Camara J. Monitoring premature infants in car seats: implementing the American Academy of Pediatrics policy in a community hospital. Pediatrics.1993;91 :1137– 1141
- ↵Merchant JR, Worwa C, Porter S, Coleman JM, deRegnier RA. Respiratory instability of term and near-term healthy newborn infants in car safety seats. Pediatrics.2001;108 :647– 652
- ↵Patel AL, Paluszynska D, Harris KA, Thach BT. Occurrence and mechanisms of sudden oxygen desaturation in infants who sleep face down. Pediatrics.2003;111(4) . Available at: http://pediatrics.org/cgi/content/full/111/4/e328–e332
- ↵Weber K. Crash protection for child passengers. A review of best practice. UMTRI Res Rev.2000;31 :1– 27
- ↵Goto K, Mirmiran M, Adams MM, et al. More awakenings and heart rate variability during supine sleep in preterm infants. Pediatrics.1999;103 :603– 609
- ↵Bass JL, Mehta KA. Oxygen desaturation of selected term infants in car seats. Pediatrics.1995;96 :288– 290
- ↵Carlo WA, Beoglos A, Siner BS, Martin RJ. Neck and body position effects on pulmonary mechanics in infants. Pediatrics.1989;84 :670– 674
- Copyright © 2003 by the American Academy of Pediatrics