Car Seat Tolerance Screening for Late-Preterm Infants

This was a retrospective medical record review of late-preterm infants born between January 1, 2013, and December 31, 2017, who were admitted at the University of Maryland Children’s Hospital (Baltimore, MD) and who qualified for a CSTS as a result of being born prematurely. The study was approved by our institutional review board. Inclusion criteria included (1) birth GA of 34 and 0/7 to 36 and 6/7 weeks, (2) survival to discharge, and (3) discharge from our institution. Exclusion criteria included (1) discharged on a home ventilator, (2) deemed inappropriate for a CSTS by medical team, including discharge to hospice care, and (3) parents declined a CSTS.

CSTSs are performed on all infants born <37 weeks’ GA, regardless of associated comorbidities, for 90 to 120 minutes, and in the location of discharge. Subjects are tested in their personal car seat, and certified child passenger safety technicians are available to assist in positioning. Our guidelines specify that infants should not undergo a CSTS until deemed within 24 to 48 hours of anticipated discharge. Car seats are placed in a buggy that simulates the appropriate angle of the car seat base. Failure criteria include (1) apnea >20 seconds, (2) bradycardia <80 beats per minute for >10 seconds, and (3) desaturation <90% for >10 seconds.

Baseline characteristics were evaluated between those who failed versus passed an initial CSTS. For continuous, binary, and categorical variables, t-tests, Wilcoxon rank-sum, χ², and Fisher-exact tests were used, as appropriate. These variables included birth GA, BW, sex, race, delivery mode, singleton versus multiple gestation, location (NICU versus NBN), maternal medications (anesthesia, antenatal steroids), resuscitation requirements with positive-pressure ventilation, maternal group B Streptococcus status, respiratory requirements (intubation, surfactant, mechanical ventilation, continuous positive airway pressure [CPAP], or low-flow nasal cannula [LFNC]), infant medications (caffeine, postnatal steroids, antireflux, diuretics), intraventricular hemorrhage (IVH), gastrostomy tube, tracheostomy, and characteristics at the time of the CSTS, including weight, postmenstrual age, LOS, and CSTS result. Infants born <35 and 0/7 weeks and/or <2.1 kg are automatically admitted to the NICU at our institution.

Multivariable modeling was performed to identify predictors of increasing LOS. We a priori included CSTS result, mode of delivery (infants born via cesarean delivery stay longer than those born via vaginal delivery), opiate-positive urinary toxicology result for mother and/or infant (neonates are observed inpatient for a minimum of 5 days to monitor for withdrawal symptoms), sex, and race. Assessment was performed for confounders and collinearity. We included variables that achieved statistical significance (P < .05), confounders, and a priori variables in the final model. Statistical analyses were performed by using SAS 9.3 (SAS Institute, Inc, Cary, NC).

We identified 918 subjects, of whom 42 failed (4.6%). Before the CSTS, 54.1% (n = 497) were admitted in the NICU, 37% (n = 339) were exclusively in the NBN, and 8.9% (n = 82) spent time in both locations. Over time, there was an increase in percentage admitted to both locations (1.8% in 2013 to 12.6% in 2017, P = .0021). In the overall cohort, 29% were white (n = 266), 60% were African American (n = 554), and 51% were male (n = 470). Incidence of failure was 3.5% for infants born at 34 weeks, 4.2% for infants born at 35 weeks, and 5.2% for infants born at 36 weeks. There was no significant change in failure rate by year.

There was a range of comorbidities in those admitted to the NICU at any time before CSTS failure (n = 28), including pneumothorax (n = 2), severe pulmonary hypertension, gastroschisis, imperforate anus, fetal hemolysis requiring transfusions, cardiac defects (tetralogy of fallot and double outlet right ventricle), in utero opiate exposure resulting in neonatal abstinence syndrome (n = 2), and genetic disorders (n = 6), including trisomy 21, trisomy 13, and Cornelia de Lange with micrognathia. Seventeen (61%) of the NICU cohort had respiratory distress requiring support.

CSTS failure was due to a desaturation event in 64% (n = 27), bradycardia in 7% (n = 3), a combination of bradycardia and desaturation in 12% (n = 5), and apnea in 5% (n = 2). Specifics of failure were not documented in 12% (n = 5). A repeat CSTS was passed by 74% (n = 31), whereas 24% (n = 10) failed a repeat CSTS. Follow-up was not documented in the remaining 2% (n = 1).

Of the 10 who failed a second CSTS, 30% (n = 3) passed a third CSTS, and 10% (n = 1) passed a fourth CSTS after varying periods of observation. Twenty percent (n = 2, both from the NBN) failed >2 CSTSs and underwent polysomnography and required oxygen at discharge. Another 20% (n = 2) passed a third CSTS on supplemental oxygen without undergoing polysomnography. The remaining 20% (n = 2) had no further testing documented.

Infants admitted exclusively to the NICU had lower mean birth GA (35 and 2/7 ± 0.9 weeks) than those admitted to the NBN (36 and 2/7 ± 0.5 weeks) or both locations (35 and 6/7 ± 0.7 weeks) (P < .0001). Incidence of failure for those in the NICU was 4.2% (n = 21). Infants who failed were more likely to have been diagnosed with IVH, required a gastrostomy tube, required a tracheostomy, received postnatal steroids, and required respiratory support (Table 1). There were no differences at the time of CSTSs when it came to medications, weight, or postmenstrual age.

Infants who spent time in both locations before the CSTS had the highest failure rate of 8.5% (n = 7). Twenty-two infants (27%) were initially admitted to the NICU and subsequently transferred to the NBN, of whom 2 failed (9.1%). Sixty infants (73%) were initially admitted to the NBN and subsequently transferred to the NICU, of whom, 5 failed (8.3%). There were no significant differences in GA, age at test, weight, or clinical and demographic characteristics between those who passed versus failed (Table 2). One infant who failed a CSTS also failed a critical congenital heart disease (CCHD) screen and underwent echocardiography without evidence of congenital heart disease. The remainder passed a CCHD screen.

Incidence of failure for those exclusively in the NBN before a CSTS was 4.1% (n = 14). Those who passed were born more prematurely than those who failed (36 and 2/7 vs 36 and 4/7 weeks, P = .046), although there was no difference in BW, hours of life at the CSTS, or weight at the CSTS (Table 3). Infants who failed a CSTS were more likely to be born via cesarean delivery (71% vs 39%, P = .0166).

In those infants admitted exclusively to the NBN at the time of CSTS failure, 1 had trisomy 21 and 1 was an infant of a diabetic mother. . Three infants were subsequently transferred to the NICU because of repeated or severe CSTS failure. However, the majority (n = 9) of late-preterm infants from the NBN were not noted to have specific comorbidities besides prematurity.

Of the 3 infants (21.4%) who required subsequent NICU admission because of a CSTS result, 2 were admitted because of repeated CSTS failures. Both underwent polysomnography, were found to have obstructive apnea with significant time spent desaturated, and were discharged from the hospital on supplemental oxygen. The third had severe oxygen desaturation during the initial CSTS, an echocardiogram demonstrated CCHD, and the infant underwent cardiac repair. A CCHD screen had not yet been performed at the time of the CSTS. Besides this infant who underwent cardiac surgery, all of the infants passed a predischarge CCHD screen despite failing the CSTS.

Median (interquartile range [IQR]: 25th percentile to 75th percentile) crude LOS was longer for those who failed versus passed the CSTS in the entire cohort (median: 12 days [IQR: 4–21] vs 7 days [IQR: 3–14], P = .003). Median crude LOS from time of initial CSTS to discharge was longer for those who failed versus passed (median: 2.5 days [IQR: 1–7] vs 1 day [IQR: 0–2]), P < .0001). However, this did not adjust for other potential confounding variables that could affect LOS. To identify the most significant predictors of LOS in our cohort, we performed multivariable linear regression modeling (Table 4). Significant predictors of increasing LOS included requiring NICU admission, requiring mechanical ventilation, and requiring LFNC therapy. LOS was inversely associated with birth GA and BW, meaning infants born with higher birth GA and with higher BW had shorter LOS. Despite longer crude LOS in infants who failed a CSTS, when adjusting for these predictors and a priori variables (opiate toxicology, race, sex, mode of delivery), the CSTS result was not significantly associated with LOS.

We evaluated a large, modern cohort of late-preterm infants to determine incidence and predictors of CSTS failure in this population. We found a difference in failure rates on the basis of admission locations before CSTS, with those admitted to both the NBN and NICU having higher failure rates than those who were only in 1 location. Although crude LOS was longer for those infants who failed an initial CSTS, when adjusting for clinical factors such as location of admission and respiratory support requirements, CSTS outcome was not associated with increased LOS.

The overall preterm birth rate in the United States has risen annually for the past few years, driven almost entirely by an increase in late-preterm births.²  They have fewer comorbidities and are more mature in general than their early-preterm counterparts but are still at risk for adverse outcomes. The shift in designation of this group from “near-term” to “late-preterm” was meant to emphasize that these infants are in fact still premature, with physiologic needs unique to their developmental level.^4,10  Late-preterm infants have higher incidence of respiratory distress syndrome, transient tachypnea, apnea, and respiratory failure than term infants.^10–12  They have increased risk of developmental delays, often out of proportion for their level of illness in the hospital.¹³  Significant cortical and cerebellar growth and maturation occur during the late-preterm period, indicating that even mild insults during this time frame could lead to significant consequences on later brain development.¹⁰  It is therefore of the utmost importance to optimize cardiorespiratory status in this group before discharge, which is the goal of the CSTS: to identify those infants who have unstable cardiorespiratory status when placed in the semiupright position as well as to ensure infants deemed appropriate for discharge have the same stability in the car seat as they do while supine in their crib.

Data are limited on CSTS results in the late-preterm population. Davis et al⁸  found they had more than double the CSTS failure rates of those born early preterm (5.6% vs 2.4%). Although late-preterm infants made up only 60% of the population, they accounted for ∼80% of CSTS failure. They hypothesized this may be due to the fact that early-preterm infants undergo continuous monitoring in the NICU, allowing a better estimation of cardiorespiratory maturity before initiating a CSTS. Late-preterm infants are often admitted to the NBN and are only on continuous monitors during the CSTS, which may be the only time that underlying cardiorespiratory immaturity is uncovered. In addition, they were tested at younger postnatal ages and therefore were less physiologically mature at the time of testing. Subsequently, Smith et al⁹  focused on late-preterm infants undergoing CSTSs and found a much higher incidence of failure (15% failure rate in the NICU patients and 28% in the NBN patients). Those who failed were more likely to be born via cesarean delivery and similarly had younger postnatal ages at the time of testing.

To better identify predictors for failure, we chose to evaluate late-preterm infants in 3 groups. We found a failure incidence that ranged from 4.1% to 8.5% depending on locations of admission. Those exclusively admitted to the NBN had the lowest incidence of failure (4.1%). This is lower than previously reported, but researchers in previous studies reported only where the infants were admitted at the time of the CSTS, not if they had required NICU observation before transfer to the NBN. It is not uncommon for late-preterm infants to spend some time in the NICU to monitor physiologic maturity but be transferred to the NBN once deemed appropriate to room in with their family. Presumably, only those deemed the most mature and least likely to have cardiorespiratory events remained in the NBN throughout. Smith et al⁹  noted 85% were admitted to the NBN, whereas only ∼46% of our subjects spent any time in the nursery. Our institution may therefore be more conservative with criteria for NBN admissions for late-preterm infants, which could explain the lower failure rate.

Those who failed from the NICU had a higher failure rate (4.2%) than the NBN and were more likely to have required respiratory support. Over time, semiupright placement in the car seat leads to lower mean saturations and increased risk of ABD events,¹⁴  which could exacerbate respiratory status in those who required more respiratory support to begin with. They were the most immature and had the most comorbidities, although they underwent continuous monitoring throughout their stay, allowing providers, in theory, to more precisely assess timing of cardiorespiratory maturity before CSTS.

Those admitted to both locations were, at some point, thought to be mature enough for the NBN but also demonstrated evidence of immaturity or distress that required NICU observation. We found the highest rate of failure in those who spent time in both locations (8.5%); however, there were no other significant clinical predictors of failure. The significant increase in late-preterm infants admitted to both locations over time (1.8%–12.6% from 2013 to 2017) may reflect a culture shift from conservative management in the NICU to prioritizing rooming in with their mothers for bonding and early breastfeeding initiation. This may mean some are admitted to the NBN inappropriately while still displaying immature feeding, thermoregulation, and cardiorespiratory status. Clearly this group is at high risk of immaturity and requires closer observation when admitted or transferred to the NBN than term infants.

Although the utility of CSTSs to identify infants at longer-term risk of adverse cardiopulmonary events has not been proven, previous studies noted a substantial number of infants who required escalation of care after failed CSTSs. Davis et al⁸  found 35% of infants who failed in the NBN required subsequent NICU admission for an average of 7 days. Smith et al⁹  had similar results, with 55% of those failing CSTSs from the NBN requiring NICU admission, and 28% of those requiring supplemental oxygen for a range of 1 to 13 days. These infants were not admitted per any protocol after a failed CSTS; instead, their providers deemed them ill enough to escalate level of care on the basis of their vital sign changes during CSTSs.

We also found a concerning number who required escalation of care after a failed CSTS. Whether the car seat position exacerbated their tenuous respiratory status or they were already having underlying ABD events is unclear and warrants further study, but it has been shown that routine predischarge CCHD screening does not predict CSTS failure.¹⁵  One infant who failed was diagnosed with cyanotic heart disease and would have been picked up on their CCHD screen. One infant failed both the CCHD screen and CSTS. All remaining infants passed a CCHD screen around the time they failed their CSTS. What seems clear is that late-preterm infants who fail repeated (≥2) CSTSs are at high risk of underlying pathology and require further workup. Although a majority passed a third test, 40% of those with repeated failures required supplemental oxygen for safe discharge from the hospital. Those who failed did have increased overall crude LOS, but when accounting for GA, BW, and evidence of cardiorespiratory immaturity (LFNC and ventilator requirements), the CSTS result itself was not a significant predictor. Instead, it was the cardiorespiratory immaturity noted or discovered by performing a CSTS that affected LOS.

Limitations of our study include the retrospective design and risk of incomplete data. However, our child passenger safety team introduced a CSTS checklist that improved documentation in the medical record. Although our failure cutoffs are some of the most commonly chosen,⁶  many institutions use different vital sign parameters, which could affect generalizability. One reason we may have found lower overall incidence of failure is our access to a robust child passenger safety team trained in car seat positioning for all infants. All NBNs and NICUs should ensure availability of certified child passenger safety technicians to assist with staff education and proper positioning for neonates.

Finally, the most important factor related to car seat safety at discharge for all infants regardless of GA is thorough, proper education for all parents, guardians, and child care providers. The vast majority of infant deaths in sitting devices occur in car seats, with the car seat being used as directed in fewer than 10% of cases.¹⁶  Caregivers should be taught proper placement of their infant in the car seat. They should be educated that a passed CSTS does not ensure absolute safety in the car seat, and direct monitoring during travel is important. Caregivers should be advised to minimize time in the car seat, using the seat for supervised travel only.

Although late-preterm infants are generally healthier than their early-preterm counterparts, they are still at risk for adverse cardiorespiratory events during a critical period in brain development. In this study, a concerning number demonstrated unstable cardiorespiratory status in their semiupright car seat, even in those presumed to be the most mature with the fewest comorbidities. Those infants who failed repeated CSTSs frequently had underlying respiratory morbidities that required escalation of care and close follow-up. Although further study is warranted, LOS was not associated with the CSTS result but rather with the cardiorespiratory immaturity noted or discovered by performing a CSTS.

ABD

apnea, bradycardia, and desaturation

BW

birth weight

CCHD

critical congenital heart disease

CPAP

continuous positive airway pressure

CSTS

car seat tolerance screen

GA

gestational age

IQR

interquartile range

IVH

intraventricular hemorrhage

LFNC

low-flow nasal cannula

LOS

length of stay

NBN

newborn nursery