OBJECTIVE: The purpose of this work was to compare the incidence of apnea, hypopnea, bradycardia, or oxygen desaturation in healthy term newborns placed in hospital cribs, infant car safety beds, or infant car safety seats.
METHODS: A consecutive series of 200 newborns was recruited on the second day of life. Each subject was studied while placed in the hospital crib (30 minutes), car bed (60 minutes), and car seat (60 minutes). Physiologic data, including oxygen saturation, frequency, and type of apnea, hypopnea, and bradycardia were obtained and analyzed in a blinded manner.
RESULTS: The mean oxygen saturation level was significantly different among all of the positions (97.9% for the hospital crib, 96.3% for the car bed, and 95.7% for the car seat; P < .001). The mean minimal oxygen saturation level was lower while in both safety devices (83.7% for the car bed and 83.6% for the car seat) compared with in the hospital crib (87.4%) (P < .001). The mean total time spent with an oxygen saturation level of <95% was significantly higher (P = .003) in both safety devices (car seat: 23.9%; car bed: 17.2%) when compared with the hospital crib (6.5%). A second study of 50 subjects in which each infant was placed in each position for 120 minutes yielded similar results.
CONCLUSIONS: In healthy term newborns, significant desaturations were observed in both car beds and car seats as compared with hospital cribs. This study was limited by lack of documentation of sleep stage. Therefore, these safety devices should only be used for protection during travel and not as replacements for cribs.
Provision of safe motor vehicle transportation for infants is a priority of parents and health professionals worldwide. The performance of car safety seats and beds in crash situations has dramatically improved.1 The car safety seat remains one of the most effective strategies available to protect children from injury and death in the early years of life.2 In a report by the National Highway Traffic Safety Administration for 2006, 452 children younger than 5 years died while riding in motor vehicles in the United States; almost one third were unrestrained.3
As optimized for crash performance, car seats require the placement of infants in the upright position. However, respiratory compromise in this position remains a concern. Early work by Willett et al4,5 on premature newborns emphasized the extent of this problem and has led to the performance of car safety seat testing on premature infants before discharge from the hospital.6–8
We and others have also found that term infants in car safety seats and beds develop respiratory compromise.9–12 Merchant et al10 showed that, in term and late preterm healthy newborn infants placed in car seats, oxygen saturation declined. Seven percent of infants were noted to have oxygen saturation values of <90% for periods >20 minutes.
The cause of this respiratory compromise is thought to be related to airway occlusion and chest wall compression, which is more likely to occur in the upright position than in the supine position.13–17 Tonkin et al16 demonstrated that preventing flexion of the head using a car seat insert increased airway size and reduced the frequency of oxygen desaturations in preterm infants. Because this correction of head position did not completely resolve the hypoxia, other mechanisms must be evoked. The infant chest wall is particularly vulnerable because of its increased compliance and its dependence on the intercostal muscles to act as a breaking mechanism, counteracting the natural tendency of the infant's lung to collapse. Any external compression of the chest wall by a car seat harness is likely to enhance this propensity of the lung to collapse. Placement of the infant in a car bed instead of a car seat should help resolve some of these position-related adverse effects. However, we found that significant desaturations occurred even when infants were placed in a car bed.12
Given the potential implications of such findings, we explored previously whether placement in a car bed, which presumably does not alter upper airway patency, is associated with less desaturation than placement in a traditional car seat.12 Sixty-seven term, healthy infants were randomly assigned to be monitored in either a car bed or a car seat. Infants in both groups spent a significant amount of the study time with oxygen saturation levels <95%. Because our initial study was limited by its small sample size and absence of a control group, we undertook the current study to test the hypothesis that greater respiratory compromise occurs when infants are placed in car beds or car seats compared with traditional cribs, using the percentage of time with an oxygen saturation of <95% as a primary outcome.
This prospective, randomized study was designed to test the hypothesis that respiratory compromise takes place when infants are positioned in car seats or car beds. We used the percentage of time with an oxygen saturation level of <95% as the primary outcome. Secondary outcomes included minimal oxygen saturation, total number of desaturation events, and mean oxygen saturation. The study was conducted between June and November 2006 in the Division for Neonatology, Maternity Hospital, University Medical Centre, Ljubljana, Slovenia. The institutional review board of the University Medical Centre, Ljubljana, and the Republic of Slovenia National Medical Ethics Committee approved the study. Informed consent was obtained from the parents of all of the participating infants. The inclusion criteria were a birth weight of 2.8 to 4.0 kg, a gestational age of 38 to 41 weeks, a normal newborn physical examination, normal vital signs, and a healthy mother. The exclusion criteria were an abnormal newborn examination, the need for pharmacologic treatment of the mother or child (other than vitamins and/or iron), and delivery by cesarean section. To overcome the limitations of our previous study,12 each subject was studied while placed sequentially in a hospital crib, a car bed, and a car seat, thus serving as his or her own control. After supine placement in a hospital crib for 30 minutes, each infant was randomly assigned by means of a computer-generated, random-number table to positioning either in a car seat first for 60 minutes (first position) followed immediately by an equal period in a car bed (second position) or to positioning in a car bed first for 60 minutes followed by an equal period in a car seat. We had initially planned to place each infant in each position for 60 minutes (total of 180 minutes), but decided to restrict the study to 150 minutes to limit time away from the infant's mother.
To address this limitation and to determine whether 60 minutes is too short a period of time to see respiratory compromise, a smaller study of 50 newborns (January to March 2007) was performed in which each infant was placed sequentially in a crib, a car bed, and a car seat in random order for 120 minutes in each position. Again, each infant served as his or her own control. Because of the length of the study, the infants were fed between each stage of the study. These families were fully informed about the extra length of the study.
Each study began ∼15 minutes after feeding and a diaper change. Infants were placed in a car seat (Römer Infant Safe Plus [Britax/Römer, Kindersichereit, GmbH, Ulm, Germany]) or a car bed (Aprica car bed [Aprica, Osaka, Japan]); these models were chosen because they were regarded as the best car bed or car seat available in Slovenia at the time of the study.
Each infant was placed in the car bed and car seat in a standardized manner. The car seat was placed at a 45° angle of recline. Secure positioning in the car seat was achieved by placing the infant with the buttocks against the back of the seat and tightening the harness to allow only 1 finger-width distance between the infant and the harness, as described previously.10,12 Additional blanket rolls were placed on either side of the infant and around the head to maintain midline positioning of the head and neck. This is not necessary18 but was performed to conform with previous studies.10,12 When necessary (ie, for infants with a short torso), an additional blanket roll was placed between the infant and the crotch belt to prevent the infant from sliding forward in the seat.
The car bed position was also standardized. The car bed was placed flat, and each infant was placed in it in an analogous manner as with the car seat, with the harness again tightened to allow 1 finger-width distance between the infant and the harness. Both the tightness of the harness and the fit of the infant within each safety device were confirmed with a check sheet.
All of the physiologic data were collected in a quiet room with ambient light and temperature by 2 experienced neonatal nurses.
Cardiorespiratory functions were monitored using the Tyco Sleep Apnea Recorder HypnoPTT (Puritan Bennett, Mallinckrodt Development France, Villers-les-Nancy Cedex, France). It also captured polygraphic recordings, including oxygen saturation level, respiratory wave form, and rate (through thoracic impedance); oronasal airflow (detected by neonatal thermistor); heart rate; and electrocardiogram. Oxygen saturation was measured from the left hand using Nellcor OxiMax neonatal oximetry technology.19
For every infant and for each position, the following parameters were recorded or calculated by using Hypnoscan software (Mallinckrodt Development France, Villers-les-Nancy Cedex, France): (1) mean oxygen saturation; (2) minimal oxygen saturation; (3) percentage of time with oxygen saturations of ≥95.0%, 90.0% to 94.9%, 85.0% to 89.9%, and <85.0%; (4) frequency and duration of bradycardia; (5) frequency, duration, and type of apnea (central or obstructive); and (6) frequency and duration of hypopnea. Apnea was defined as a ≥10-second period with airflow amplitude of <10% of baseline values on the oronasal thermistor signal, either with (obstructive) or without (central) evidence of respiratory effort on tracings from thoracic impedance.20 Hypopnea was defined as an airflow amplitude of 10.0% to 49.9% of surrounding baseline on the oronasal thermistor signal, lasting for ≥10 seconds and accompanied by a decrease in oxygen saturation of >4.0%.20 Bradycardia was defined as a heart rate of <70 beats per minute lasting for ≥5 seconds. We also determined the desaturation index (number of desaturations per hour), for which desaturation was defined as a decrease in oxygen saturation by >4% or <95% lasting for ≥10 seconds. Two experienced investigators independently reviewed all of the tracings to exclude periods with obvious artifacts. Each reviewer was not aware of the order of placement; accordingly, the analyses were performed in a blinded manner.
Statistical comparisons between groups were based on Mann-Whitney and Kruskal-Wallis analyses of ranks, as well as the χ2 test where appropriate. A P value of <.05 was needed for significance. Most P values were <.01, so a Bonferroni correction for multiple groups was not factored into the calculation. All of the data are presented as means ± SDs.
Clinical characteristics of infants and mothers are shown in Table 1. Each infant spent a similar amount of time in the car seat and the car bed (61.2 ± 2.4 minutes and 60.9 ± 3.3 minutes, respectively). According to the initial design of the study, the time spent in the control position (hospital crib) was shorter than that spent in the other positions by half (31.3 ± 2.4 minutes). Mean oxygen saturation level was highest in the hospital crib (97.9%), with lower means in both the car seat (95.7%) and the car bed (96.3%). The differences within each pair of groups were statistically significant. Minimal oxygen saturation level was lower in both safety devices (83.7% in the car bed and 83.6% in the car seat) compared with the hospital crib (87.4%). The time spent at a number of oxygen saturation levels was compared across groups (Table 2 and Fig 1). All of the groups had oxygen saturation levels <85% for ∼1% of the study duration. However, the percentage of total study time spent with an oxygen saturation level <95% was significantly higher in the car seat (23.9%) compared with the car bed (17.2%; P = .003) or hospital crib (6.5%; P < .001).
Notably, a subgroup of infants spent >50% of their time with oxygen saturation levels <95%. This occurred in only 1 infant (0.5%) while placed in the hospital crib. However, this prolonged desaturation occurred in 30 (15%) of the infants who were placed in the car seat (P < .001) and in 13 (6.5%) of the infants who were placed in the car bed (P < .001).
To understand the cause of these reductions in oxygen saturation, each infant's recording was scored for apneas. The obstructive apnea index (number of apneas per hour) was similar for infants who were placed in car beds (1.7 per hour) and hospital cribs (1.7 per hour) but slightly higher for infants who were placed in car seats (2.2 per hour; Table 2). The difference in obstructive apnea index between supine positioning in the car bed and upright positioning in the car seat did not reach statistical significance (P = .052). The hypopnea index (number of hypopneas per hour) was higher in the car seat position (2.9 per hour) compared with both the car bed (2.6 per hour) and the hospital crib positions (2.2 per hour); however, these differences did not reach statistical significance (P = .226; Table 2). No central apneas or bradycardias were recorded in any position.
Because of the crossover design of the study, each infant was placed sequentially in the car bed and the car seat. Thus, physiologic parameters for infants placed in the car bed as the first randomly assigned position could be compared with infants placed in the car bed as the second randomly assigned position. Likewise, a similar analysis could be done for infants placed in the car bed (Table 3). Infants generally did better during the first hour than the second.
To further exclude the possibility that the duration of time spent in each position had a significant effect on the occurrence of desaturations and apneas, we performed an additional study of 50 infants who were monitored during placement in a hospital crib, car bed, and car seat for 120 minutes in each position. These infants and mothers had similar characteristics as those in the larger study (Table 1). The observation times were similar in all 3 of the positions (hospital crib: 119.1 ± 5.2 minutes; car bed: 113.8 ± 16.1 minutes; car seat: 119.1 ± 6.7 minutes). The results obtained were nearly identical to those of the larger study with shorter observational times (Table 4).
Car seats and beds are essential safety devices. However, previous studies have found that placement of infants in these devices may alter respiratory performance. As in our previous study,12 we confirmed our hypothesis that the respiration of term infants is compromised in both car beds and car seats. Our primary and secondary outcomes, including mean oxygen saturation, minimal oxygen saturation, and percentage of time with oxygen saturation level <95%, were all significantly altered from the values obtained while infants were placed in cribs. Not all of the infants were equally affected, however. Indeed, a subgroup of infants spent >50% of their time with oxygen saturation levels <95%. The car bed performed better than the car seat on 1 parameter: mean oxygen saturation values recorded in the car bed were significantly higher (P < .001) than in the car seat. In the car seat, there was a trend toward more obstructive apneic events (P = .052), implicating obstructive apnea as a likely factor in the decreased oxygen saturations observed in car seats.
Another novel finding was demonstrated by our study. Because of the crossover design in which the infants were placed sequentially in each device, we were able to observe that infants generally fared better during the first hour than the second hour, regardless of the order of the device used (Table 3). This finding suggests that infants may be able to compensate for the respiratory limitation initially, but less so with time, either because of muscle fatigue or deeper sleep stage.
This clinical study has a number of limitations, which were addressed through additional analysis and should be considered in future studies. The time spent in a hospital crib was limited to just 30 minutes, and this might not be enough time for the infant to develop respiratory compromise, as documented by desaturations. However, comparison of our hospital crib group with the normal infants monitored in the Collaborative Home Infant Monitoring Study reveals similar findings in both groups.21,22 The most appropriate period of observation of infants in car seats or car beds is also unclear. Observation periods <90 minutes are often used in clinical practice and in predischarge testing for preterm infants (<37 weeks of gestation7,8,10–12,23), although infants are, in reality, frequently placed in these devices for prolonged periods of time. Thus, 60 minutes of placement in a car safety device may not be enough to document respiratory compromise.10 To mitigate these concerns, we performed an additional study of 50 infants who were monitored during placement in a hospital crib, car bed, and car seat for 120 minutes in each position. These infants had similar alterations in oxygen saturation when compared with the larger study (Table 4).
Sleep has an important influence on the respiratory physiology, and sleep states may be altered by body position. The intercostal muscles play an important role in limiting expiration and maintaining resting lung volumes; therefore, because these muscles are relaxed during active (rapid eye movement [REM]) sleep, oxygen saturation levels are lower in normal infants during REM sleep than during non-REM sleep.24,25 If body positioning associated with safety devices promotes REM sleep, oxygen saturation levels would be reduced. Although infants do sleep for prolonged periods between meals, we did not document sleep by observation or electroencephalography. This is a major limitation of this study.
In this study, the use of thermistors may have led to an underestimation of the number of hypopneas. This technical limitation might explain why there were significant changes in oxygen saturation between the groups, although the apnea-hypopnea index was similar for all of the groups. Thermistor is very likely to overestimate airflow at low levels; thus, the number of hypopneas events might be missed. Abdominal and chest wall bands to define abdominal and chest expansions may be more sensitive in defining obstructive apnea. In addition, nasal pressure monitoring is a more sensitive method to define reduced airflow. Thus, the use of these modifications may identify the true cause of desaturations in future studies.
Low oxygen saturation levels in infants placed in both car beds and car seats are concerning in light of recent data demonstrating the adverse effects of airway obstruction in older children.26,27 Even mild airway obstruction has been associated with behavioral problems and IQ deficits.28 Indeed, a number of deaths have occurred in car seats, some of which may be at least partly explained by airway obstruction.29 These devices are often used for many hours at a time for reasons other than travel,30 and infants placed in these devices for prolonged periods of time are, thus, at increased risk for recurrent hypoxic events.
Car safety seats and beds remain very important transport devices for newborn infants and older children. Their use is imperative for the safe transportation of infants. However, even in healthy term infants, apnea, hypopnea, and oxygen desaturations occur during placement in these devices. The use of these devices should, therefore, be restricted to protection from injury and death in traffic accidents, and they should never serve as a replacement for a crib. In addition, further modifications of car safety devices are clearly needed to minimize the respiratory compromise that has been consistently documented in current models.
This work was supported by Aprica (Osaka, Japan).
- Accepted April 10, 2009.
- Address correspondence to T. Bernard Kinane, MD, Mass General Hospital for Children, Pediatric Pulmonary, Fruit Street, Boston, MA 02114. E-mail:
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
Hypoxic events occur in preterm infants while they are in car seats, and some small studies have suggested that such events occur in term infants.
What This Study Adds:
This large, controlled study confirms that these hypoxic events occur in term infants while in both car seats and car beds. It supports the notion that car beds and car seats need to be redesigned to limit respiratory compromise.
- ↵American Academy of Pediatrics, Committee on Injury and Poison Prevention. Safe transportation of newborns at hospital discharge. Pediatrics.1999;104 (4):986– 987
- ↵Weber K. Crash protection for child passenger: a review of best practices. UMTRI Res Rev.2000;31 (3):1– 27
- ↵National Highway Traffic Safety Administration. Traffic safety facts 2006: a compilation of motor vehicle crash data from the Fatal Analysis Reporting System and the General Estimates System. 2006. Available at: www-nrd.nhtsa.dot.gov/Pubs/TSF2006FE.pdf. Accessed July 28, 2009
- ↵Pilley E, McGuire W. The car seat: a challenge too far for preterm infants? Arch Dis Child Fetal Neonatal Ed.2005;90 (6):F452– F455
- ↵American Academy of Pediatrics, Committee on Injury and Poison Prevention, Committee on Fetus and Newborn. Safe transportation of premature and low birth weight infants. Pediatrics.1996;97 (5):758– 760
- ↵American Academy of Pediatrics, Committee on Injury and Poison Prevention, Committee on Fetus and Newborn: Safe transportation of premature infants. Pediatrics.1991;87 (1):120– 122
- ↵Bass JL, Mehta KA. Oxygen desaturation of selected term infants in car seats. Pediatrics.1995;96 (2):288– 290
- ↵Merchant JR, Worwa C, Porter S, Coleman JM, deRegnier RAO. Respiratory instability of term and near-term healthy newborn infants in car safety seats. Pediatrics.2001;108 (3):647– 652
- ↵Kinane TB, Murphy J, Bass JL, Corwin MJ. Comparison of respiratory physiologic features when infants are placed in car safety seats or car beds. Pediatrics.2006;118 (2):522– 527
- 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 (1):290– 299
- ↵Tonkin SL, McIntosh CG, Hadden W, Dakin C, Rowley S, Gunn AJ. Simple car seat insert to prevent upper airway narrowing in preterm infants: a pilot study. Pediatrics.2003;112 (4):907– 913
- ↵Bass JL, Corwin M, Gozal D, et al. The effect of chronic or intermittent hypoxia on cognition in childhood: a review of the evidence. Pediatrics.2004;114 (3):805– 816
- ↵Tonkin SL, Vogel SA, Bennet L, Gunn AJ. Apparently life threatening events in infant car safety seats. BMJ.2006;333 (7580):1205– 1206
- Copyright © 2009 by the American Academy of Pediatrics