Car Seat or Car Bed for Very Low Birth Weight Infants at Discharge Home

Article Outline

• Abstract
• Methods
  • Population
  • Assessment Procedures
  • Data Collection
  • Sample Size and Statistical Analysis
• Results
• Discussion
• References
• Copyright

Objective

To compare the incidence of apnea, bradycardia, or desaturation in a car seat with that in a car bed for preterm very low birth weight (≤1500 g) infants.

Study design

Infants were studied for 120 minutes in a car seat and in a car bed. Apnea (>20 seconds), bradycardia (heart rate <80/min for >5 seconds), desaturation (Spo₂ <88% for >10 seconds), and absent nasal flow were monitored.

Results

We assessed 151 infants (median birth weight, 1120 g [range, 437 to 3105)]; median birth gestational age, 29 weeks [24 to 34]) in both devices. Twenty-three infants (15%) had ≥1 event in the car seat compared with 29 (19%) in the car bed (P = .4). Time to first event was similar in the car seat and car bed (mean, 54 to 55 minutes). In logistic regression analyses, bronchopulmonary dysplasia was a significant predictor for a car seat event and a lower gestational age at birth was a risk factor for a car bed event.

Conclusions

We found no evidence that an event is less likely in a car bed than in a car seat. Whichever device is used, very low birth weight infants require observation during travel.

Abbreviations: BPD, Bronchopulmonary dysplasia, GA, Gestational age, VLBW, Very low birth weight

In 1990, the American Academy of Pediatrics recommended that all newborn infants discharged from hospitals should be transported in infant car safety seats.¹ However, 12% to 30% of premature infants have been reported to have episodes of desaturation and bradycardia while in car seats,², ³, ⁴, ⁵ and studies by Bull et al⁶ suggest that a car bed can be adapted to accommodate very small infants. Based largely on these studies, the American Academy of Pediatrics in 1996 recommended that each preterm infant be monitored in a car safety seat before hospital discharge and that infants with documented desaturation, apnea, or bradycardia should travel in a supine or prone position in a car bed.⁷, ⁸ This recommendation is based on an assumption that these events are less likely in a car bed than in a car seat. Yet, there are no studies with preterm infants comparing the incidence of these events in a car seat with that in a car bed, particularly among very low birth weight (VLBW) infants, who are most likely to have these episodes.

We conducted a two-center, cross-over trial to determine whether preterm VLBW infants at discharge home had fewer events when placed supine in a car bed than in a car seat. We hypothesized that these infants have fewer events in a car bed than in a car seat. Our secondary objectives were to relate the duration of time in a car seat or bed to the likelihood of these episodes and to evaluate risk factors for their occurrence.

Back to Article Outline

Methods 

Population 

To increase the sample size and the generalizability of the findings, the study was conducted in two hospitals: Parkland Memorial Hospital (Dallas, Texas), a county hospital with a largely inborn population, and Memorial Hermann Children’s Hospital (Houston, Texas), a private hospital with a substantial proportion of maternal and neonatal referrals. All VLBW infants weighing ≤1500 g at birth and at <37 weeks’ gestational age (GA) became eligible when they were nearing discharge. At both hospitals, infants are considered ready for discharge when weighing >1800 g, are nipple-feeding well, have appropriate weight gain, and are free of apnea or bradycardia for >5 days. Informed parental consent was obtained for all participants, and the study was conducted as approved by the local institutional review board.

Assessment Procedures 

Because some infants (for example, those referred from other communities) may travel in a car seat for 2 hours, enrolled infants were studied for 120 minutes in a car seat and 120 minutes in a car bed. At least 1-hour recovery time was allowed between assessments. Depending on the availability of the research staff, the second assessment occurred as much as 24 hours after the first assessment. As is routine practice in our centers, both assessments occurred within 72 hours before the scheduled discharge for each infant. All study sessions occurred at a minimum of 30 minutes after feeding. Baseline data were recorded for 5 minutes before placement in the car seat or the car bed. Whether the infant was placed in the car seat first or the car bed first was randomly assigned by using sequentially numbered, sealed, opaque envelopes prepared by means of a random-number table by an investigator uninvolved in the care of the subjects. Infants were assessed in the Cosco infant car seat and in the Cosco Ultradreamride car bed. In the car seat, infants were reclined at a 45 degree angle and supported with cotton rolls, as recommend for positioning of the head and neck.¹ In the car bed, the infants were placed supine.

Each infant was monitored in both devices with a pulse oximeter, a cardiopulmonary monitor, and a nasal flow detector. Apnea, bradycardia, oxygen saturation, and nasal airflow were assessed by using the Eventlink 511 monitor by CAS Medical Systems, Inc (Branford, Conn). The 511 Monitor displays beat-to-beat heart rate, respiratory rate through thoracic impedance, oxygen saturation, and captures nasal airflow, electrocardiogram, pulse rate, and respiratory waveform. In addition, this monitor allows event detection and storage, based on adjustable parameters. The 20 seconds before the event is saved as the baseline, and the recording continues for 30 seconds after the event. Captured data can be reviewed and printed. In the absence of a clearly defined lower oxygen saturation limit in convalescent VLBW infants,⁹ we selected 88% as a lower cutoff. The monitor was set to identify apnea for >20 seconds, bradycardia <80 beats/min, and oxygen saturation <88%. The occurrence of events was reported to the attending physician to determine the appropriate evaluation and management of the infant.

Data Collection 

Before study, research nurses recorded demographic information, duration of mechanical ventilation, duration of oxygen therapy, presence or absence of bronchopulmonary dysplasia (BPD: oxygen administration at 36 weeks), intraventricular hemorrhage with ventricular dilation or periventricular echodensities (grade III/IV intracranial hemorrhage),¹⁰ and date of last apneic episode for each infant. While each study infant was in the car seat or car bed, pediatric nurse practitioners or neonatal nurses stayed in close proximity to the infant, recorded the oxygen saturation at baseline and at 30-minute intervals, responded to and documented monitor alarms, and intervened if the infant appeared cyanotic, had bradycardia lasting for 5 seconds or more, or had an oxygen saturation <88% for >20 seconds.

Two investigators independently reviewed all event tracings to exclude artifacts. Artifactual bradycardia was detected by inspection of the QRS complexes; artifactual apnea, by inspection of the respiratory impedance waveform and nasal airflow; and artifactual desaturation due to failure to detect the pulse, by a discrepancy between the internal pulse oximetry heart rate and the beat-to-beat heart rate. A consensus was reached on disputed tracings. The following definitions were used in characterizing the events: Apnea, cessation of breathing for >20 seconds or <20 seconds with associated bradycardia or cyanosis; central apnea, absence of respiratory effort with absence of nasal flow; obstructive apnea, absence of nasal flow in the presence of respiratory effort; mixed apnea, the combination of obstructive apnea followed by central apnea; bradycardia, heart rate less than 80 beats/min for 5 seconds or more; desaturation, oxygen saturation <88% lasting for 10 seconds or more.

Sample Size and Statistical Analysis 

It was difficult to predetermine an appropriate sample size because of the lack of data regarding the incidence of adverse events in safety devices in VLBW infants. Therefore, we determined the proportion of infants in the car seat arm of the study who had an event during the first 6 months into the study (18%).¹¹ To avoid bias in determining the sample size, we remained blinded to the findings in the car bed arm. To identify a 50% reduction in the proportion of infants with events in the car bed, we selected a sample size of 154 patients (two-tailed analysis; α error = 0.05; β error = 0.2).

Because each infant served as his or her own control, the proportion of infants with an event in the car seat was compared with that in the car bed by using the McNemar test for paired observations. The time to the first event was compared by using the Student t test. The mean oxygen saturation in the car seat and the car bed at baseline and at 30-minute intervals were compared by using the Friedman repeated-measures analysis of variance on ranks with the Dunn method for multiple comparisons. The oxygen saturations were compared at each time point by using the Wilcoxon signed rank test. χ² analysis and the Mann-Whitney rank sum test were used to compare infants with events versus those with no events in the car seat and the car bed. The Sigma Stat 3.1 software package (SPSS, Chicago, Ill) was used in the above analyses. All tests were two-tailed, with the type I error of 0.05.

Separate logistic regressions models were used for car seats and car beds to relate the likelihood of events to various risk factors: center, birth weight, GA, total ventilation days, BPD, grade III/IV intracranial hemorrhage, duration of apnea of prematurity, chronological age and weight at the time of test. Backward elimination procedures were used to determine the risk factors that best predict an event in the car seat or car bed. Adjusted odds ratio (OR) and 95% Wald confidence intervals (CI) were calculated for the significant risk factors. The SAS 9.1.3 software package (Cary, NC) was used in all logistic analyses.

Back to Article Outline

Results 

Between February and December 2002, 178 infants met inclusion criteria. Consent was obtained for 160, and 7 were not studied because they were discharged earlier than planned. The data could not be analyzed for two infants because of monitor malfunction. Thus, we assessed 151 infants; 77 were randomly assigned to the car seat first and 74 to the car bed first. The median (range) birth weight was 1120 g (437 to 3105); weight at study, 2545 g (1750 to 5670); GA at birth, 29 weeks (24 to 34); and GA at study, 38 weeks (31 to 56); 49% were boys; 54% were Hispanic; 29% were African American, 16% were Caucasian, and 1% were other racial groups.

Forty-three (28%) of the infants had at least one event in either or both transportation devices; 23 (15%) had an event in the car seat; and 29 (19%) had an event in the car bed (Table I). The absolute difference is 4% (95% CI = −4% to 12%, P = .4). In neither study hospital did events occur less commonly in a car bed than in a car seat (17.5% vs 16.5% at Parkland; 22% vs 13% at Hermann). The number of infants with events that prompted a nursing intervention (repositioning, suctioning, oxygen administration) was also comparable: 9 (6%) in the car seat versus 5 (3%) in the car bed (95% CI = −2% to 7%, P = .30). Two patients, one with each device, were removed from the device and retuned to a crib because of continued desaturation despite nursing interventions. No infant required intubation or bag and mask ventilation.

Table I. Distribution of infants according to device and whether an event occurred

The type of events was comparable for the car seat and car bed (Table II). No infant who received nursing interventions had electronic monitoring or clinically recognized episodes at enrollment in the study. The one infant receiving xanthine therapy at enrollment had no event in either device. The events requiring intervention in the car bed were attributed to emesis (n = 2), airway obstruction secondary to flexion of the neck (n = 1), previously undetected anemia (n = 1), or viral illness (n = 1). The events requiring intervention in the car seat were attributed to emesis (n = 2), flexion of the neck (n = 3), or anemia (n = 1) or had an unclear cause (n = 3).

Table II. Type of adverse events

Some infants are included in more than one category.

The mean time to first event was almost identical in the car seat and car bed (mean ± SD = 55 ± 42 vs 54 ± 34 minutes, respectively; range = 1 to 117 minutes for both devices). The car seat and car bed were also similar in the proportion of infants whose first event occurred after 60 minutes of observation (40% vs 52%) and after 90 minutes (30% vs 10%) (P > .05).

The median oxygen saturation (available for 146 infants) decreased from 100% at baseline to 98% at 120 minutes in the car seat (P < .001). The minimum oxygen saturation values at baseline, time zero, 30 minutes, 60 minutes, 90 minutes, and 120 minutes were 93%, 92%, 89%, 89%, 85%, and 90%, respectively. Oxygen saturation in the car bed decreased from 100% at baseline to 99% at 120 minutes (P < .001). The minimum oxygen saturations were 90%, 90%, 91%, 90%, 90%, and 91% at baseline, time zero, 30 minutes, 60 minutes, 90 minutes, and 120 minutes, respectively. Three infants in the car seat each had one oxygen saturation <90% recorded; none was recorded for infants in the car bed. Paired comparison of the median oxygen saturation of the two transportation devices revealed similar values at baseline, zero, 30 minutes, and 60 minutes; but lower values in the car seat than car bed at 90 minutes: 98% (96% to 100%) versus 99% (97% to 100%), P = .004; and 120 minutes: 98% (97% to 100%) versus 99% (97% to 100%), P < .001.

For the car bed, infants with at least one event were of lower GA, birth weight, and had longer mechanical ventilation than infants without events (Table III). For the car seat, BPD and GA approached statistical significance. In multivariable analysis, a lower GA at birth was the only predictor for a car bed event (OR, 1.28; 1.06 to 1.54, P < .01) per week decrease in GA; the presence of BPD was the best predictor for a car seat event (OR, 2.65; CI 1.03 to 6.81, P = 0.04).

Table III. Univariate analysis of infants with and without adverse events

N	Car bed event			Car seat event
	Yes	No	P	Yes	No	P
	29	122		23	128
BW (g)⁎	1040(858–1137)	1135(915–1280)	.02	1102(866–1269)	1121(899–1270)	NS
GA (wk)⁎	28(26–29)	29(28–31)	.006	28(26–29)	29(27–30)	.06
Cesarean (%)	76	57	NS	65	60	NS
Male (%)	48	52	NS	56	48	NS
Hispanic (%)	48	56	NS	65	52
Test weight⁎	2675(2250–2839)	2480(2080–3070)	NS	2750(2567–3386)	2432(2090–3025)	NS
Test age (wk)⁎	39(37–40)	38(36–41)	NS	39(37–42)	38(36–40)
Duration of MV⁎	6(1–20)	1(0–6)	.004	6(1–20)	1(0–7)	NS
BPD (%)	28	21	NS	39	20	.07
IVH (%)	0	4	NS	0	4	NS
Age of apnea resolution (wk)⁎	35(34–37)	35(33–37)	NS	35(33–37)	35(33–39)	NS

BW, Birth weight; GA, gestational age; MV, mechanical ventilation; BPD, bronchopulmonary dysplasia; IVH, grade III/IV intracranial hemorrhage.

Back to Article Outline

Discussion 

To our knowledge, this is the first randomized study comparing responses of infants to car beds and car seats. Our study has three major findings: (1) As reported by other investigators,², ³, ⁴, ⁵ apnea, bradycardia, and desaturation episodes may still occur at discharge when VLBW infants are placed in a transportation device, particularly among infants who were born most prematurely or who are recovering from BPD; (2) we found no evidence that these episodes are less likely in a car bed than a car seat; (3) a brief observation period in a transportation device is not sufficient to identify infants at risk, but the longer these infants remain in such devices, the more likely oxygen saturation is to fall.

The proportion of our VLBW infants with one or more episodes—15% in the car seat and 19% in the car bed—is in the range previously reported for premature infants observed in car seats.², ³, ⁴, ¹² This proportion may be higher in populations whose mean birth weight and GA is lower or whose incidence of BPD is higher. Although most events resolved spontaneously, 6% of infants in the car seat and 3% of those in the car bed received nursing intervention to terminate events. Although other investigators have noted that VLBW infants have subclinical events well past term,¹³, ¹⁴ the testing in this study may have unmasked a new or underlying problem that had been missed by the caregivers before the test.

The appropriate period of observation in a car seat or car bed is unclear.¹⁵ Most reports have involved a 90-minute interval,², ³, ⁴, ¹² although shorter periods may be used in clinical practice. A short observation period may not detect infants at risk, whereas a long period adds burden on the nursery staff and may prompt continued hospitalization of infants who experience minor episodes but could be safely discharged. The time to first event approached 1 hour, with a wide range (1 to 117 minutes) for both transportation devices. The optimal observation period in a transportation device and the proper interpretation of the results cannot be well defined without further study of the reliability of findings and their relation to the risk of important events while riding in a car.

Our results for oxygen saturation monitoring are similar to those of Merchant et al,¹² who reported a significant decline in the mean to 94% during a 90-minute car seat test. These declines may have little clinical significance. However, 7% of their infants and 2% of ours had oxygen saturation values <90%. The difference between studies may reflect the use of continuous recording in the study of Merchant et al and a difference in nursing attention to optimize positioning. Oxygen saturation <85% has been associated with reduced cerebral oxygenation in VLBW infants.¹⁶ The potential significance of these events on the short- and long-term outcomes of VLBW infants is unclear.

Bronchopulmonary dysplasia was the only significant predictor for failing a car seat test. Functional residual capacity is reduced in infants with BPD,¹⁷ and their tidal volumes decrease after placement in a car seat.⁴ It is plausible that car seat placement worsened an already compromised pulmonary status in infants with BPD and contributed to the events.

Infants with events in the car bed were of lower GA and birth weight and required longer mechanical ventilation than those with no events, and each 1-week decrease in the GA increased the chance of events by ∼30%. These findings outline the precariousness of the cardiopulmonary status of the preterm VLBW infants and are consistent with the lability of the oxygen saturation previously described in this high-risk population.¹³, ¹⁴

In conclusion, VLBW infants who are ready for discharge are as likely to have apnea, bradycardia, or oxygen desaturation when placed in a car seat, compared with a car bed. The role of car seats in protecting young infants from injury and death in motor vehicle accidents is well documented.¹⁸ However, preterm VLBW infants should be closely observed and travel time limited, irrespective of whether car beds or car seats are used.

Back to Article Outline

References 

1. American Academy of Pediatrics Committee on Accident and Poison Prevention. Safe transportation of newborns discharged from the hospital. Pediatrics. 1990;86:486–487
   • View In Article
   • MEDLINE
2. 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
   • View In Article
   • MEDLINE
3. Willett LD, Leuschen MP, Nelson LS, Nelson RM. Risk of hypoventilation in premature infants in car seats. J Pediatr. 1986;109:245–248
   • View In Article
   • Abstract
   • Full-Text PDF (281 KB)
   • CrossRef
4. Willett LD, Leuschen MP, Nelson LS, Nelson RM. Ventilatory changes in convalescent infants positioned in car seats. J Pediatr. 1989;115:451–455
   • View In Article
   • Abstract
   • Full-Text PDF (414 KB)
   • CrossRef
5. Ojadi VC, Petrova A, Mehta R, Hegyi T. Risk of cardio-respiratory abnormalities in preterm infants placed in car seats: a cross-sectional study. BMC Pediatr. 2005;5:28
   • View In Article
   • MEDLINE
   • CrossRef
6. Bull MJ, Weber K, Stroup KB. Automotive restraint systems for premature infants. J Pediatr. 1988;112:385–388
   • View In Article
   • Abstract
   • Full-Text PDF (2183 KB)
   • CrossRef
7. American Academy of Pediatrics Committee on Injury and Poison Prevention and Committee on Fetus and Newborn. Safe transportation of premature and low birth weight infants. [see comment] Pediatrics. 1996;97:758–760
   • View In Article
   • MEDLINE
8. Bull M, Agran P, Laraque D, Pollack SH, Smith GA, Spivak HR, et al. American Academy of Pediatrics Committee on Injury and Poison Prevention Transporting children with special health care needs. Pediatrics. 1999;104:988–992
   • View In Article
   • MEDLINE
   • CrossRef
9. Askie LM, Henderson-Smart DJ, Irwig L, Simpson JM. Oxygen-saturation targets and outcomes in extremely preterm infants. N Engl J Med. 2003;349:959–967
   • View In Article
   • CrossRef
10. Papile LA, Burstein J, Burstein R, Koffler H, Koops B. Relationship of intravenous sodium bicarbonate infusions and cerebral intraventricular hemorrhage. J Pediatr. 1978;93:834–836
    • View In Article
    • Abstract
    • Full-Text PDF (256 KB)
    • CrossRef
11. Gould AL, Shih WJ. Modifying the design of ongoing trials without unblinding. Stat Med. 1998;17:89–100
    • View In Article
    • MEDLINE
12. 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
    • View In Article
    • CrossRef
13. Poets CF, Stebbens VA, Alexander JR, Arrowsmith WA, Salfield SA, Southall DP. Oxygen saturation and breathing patterns in infancy, 2: preterm infants at discharge from special care. Arch Dis Child. 1991;66:574–578
    • View In Article
    • CrossRef
14. Poets CF, Stebbens VA, Alexander JR, Arrowsmith WA, Salfield SA, Southall DP. Arterial oxygen saturation in preterm infants at discharge from the hospital and six weeks later. J Pediatr. 1992;120:447–454
    • View In Article
    • Abstract
    • Full-Text PDF (1440 KB)
    • CrossRef
15. Williams LE, Martin JE. Car seat challenges: where are we in implementation of these programs?. J Perinat Neonatal Nurs. 2003;17:158–163
    • View In Article
    • MEDLINE
16. Yamamoto A, Yokoyama N, Yonetani M, Uetani Y, Nakamura H, Nakao H. Evaluation of change of cerebral circulation by SpO2 in preterm infants with apneic episodes using near infrared spectroscopy. Pediatr Int. 2003;45:661–664
    • View In Article
    • MEDLINE
    • CrossRef
17. Hjalmarson O, Sandberg KL. Lung function at term reflects severity of bronchopulmonary dysplasia. J Pediatr. 2005;146:86–90
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (115 KB)
    • CrossRef
18. Bull MJ, Sheese J. Update for the pediatrician on child passenger safety: five principles for safer travel. Pediatrics. 2000;106:1113–1116
    • View In Article
    • CrossRef