Variation in Car Seat Tolerance Screen Performance in Newborn Nurseries

The study was performed through the Better Outcomes through Research for Newborns (BORN) network, established by the Academic Pediatric Association in 2010. At the time of the survey, the network included 106 NBN sites at academic and community medical centers across 35 states, with an annual birth rate of >400 000 infants. At each member NBN, the main contact individual was encouraged to forward the survey to the staff member most knowledgeable about CSTS protocols in their nursery. The study was determined to be exempt from human subjects research review by the Institutional Review Board at the University of Maryland, Baltimore.

Survey data were collected and managed by using Research Electronic Data Capture system, a secure Web-based platform for survey design and database management maintained at the University of Maryland, Baltimore.^15,16  The survey was adapted from a previous CSTS survey published in 2013.¹¹  Between February 2019 and June 2019, an e-mail link to the survey was sent to the designated representative at each BORN center, followed by up to 5 additional e-mail reminders. The survey included questions about location (state), academic affiliation, average annual deliveries, and whether the site routinely performed a predischarge CSTS. For those sites that did perform a CSTS, the survey addressed inclusion and failure criteria, duration, timing, location of testing, follow-up of a failed CSTS, and use of car beds. For those sites that did not perform CSTS, reasons for not performing the test were explored in the survey. All site respondents were also questioned regarding their individual attitudes related to CSTS, including its utility in assessing discharge readiness, optimal duration, and ideal inclusion and failure criteria.

We received responses from 79.2% (n = 84) of the BORN sites, representing 35 states. The median (interquartile range [IQR]) annual delivery rate was 3100 (2200–4500) infants, and 84.5% (n = 71) of sites were academically affiliated. Respondent training backgrounds included 79.8% in pediatrics (n = 67), 17.8% in neonatology (n = 15), and 2.4% unknown (n = 2). When asked about CSTS performance, 90.5% (n = 76) responded that their site did perform a predischarge CSTS, whereas 9.5% (n = 8) did not. There were no significant differences in training background, academic affiliation, or annual deliveries between the sites that did and did not perform CSTS.

Of the 76 nurseries that performed predischarge CSTS, 98.7% (n = 75) considered birth GA as an inclusion criterion. Of these, 98.7% (n = 74) tested all infants born at <37 weeks, and 1.3% (n = 1) only tested those born at <35 weeks. Additionally, 81.6% (n = 62) included a minimum BW in addition to birth GA. The most common BW used was <2.5 kg by 82.3% (n = 51) of those sites.

When asked about additional inclusion criteria used to determine eligibility for CSTS at their site, 7.6% (n = 21) noted no additional factors, although most nurseries did include other diagnoses and comorbidities, as shown in Table 1. Hypotonia, oxygen requirement during admission, and hemodynamically significant congenital heart disease (CHD) were the most common additional factors considered in decisions to perform CSTS.

Nurses were responsible for performing the CSTS at all sites, but several also noted the involvement of additional staff (Table 1). Although only 6% specifically noted involvement of trained child passenger safety technicians (CPSTs) in performance of CSTS, 31.6% did have CPSTs available to assist if needed. Most had a staff member position the infant in the car seat for the CSTS, whereas ∼8% had the parent or guardian do the positioning while supervised by staff.

When it came to timing of the CSTS in relation to the last feeding, 64.5% (n = 49) did not specify timing, 25% (n = 19) recommended waiting at least 30 or 60 minutes after the last feeding, 9.2% (n = 7) recommended testing within 30 minutes of the last feeding, and 1.3% (n = 1) noted they test “between feeds.” Timing in relation to discharge was not specified by 78.9% (n = 60), whereas the remaining 21.1% (n = 16) recommended CSTS performance within 12 hours of anticipated discharge.

The majority of sites followed the AAP guidelines to test between 90 and 120 minutes (Table 1). Forty-eight sites specified a maximum duration, with the most common being 90 minutes, although the times ranged from 30 minutes to 3 hours. Maximum duration of testing was not specified by 36.8% (n = 28), implying that a CSTS is performed for the entire duration of the car ride home, even if >90 to 120 minutes.

Of those NBNs that perform CSTS, 74 (97.4%) reported on failure criteria and follow-up of a failed CSTS. Spo₂ constituting failure ranged widely from <85% to <93%, with the most common criterion being <90% (Table 2). The most common minimum duration of desaturation constituting failure of the CSTS was >20 seconds but ranged from any drop below the threshold to >60 seconds.

The minimum heart rate that constituted failure due to bradycardia varied from <60 to <100 bpm, with <80 bpm used most commonly (Table 2). Minimum duration also varied from any drop below the specified threshold to requiring >60 seconds of bradycardia. The most common definition of apnea was a breathing pause >20 seconds.

Sites were presented with the case of an infant who failed an initial CSTS but after removal from the car seat was without distress and had complete resolution of abnormal vital signs. In this scenario, most sites would plan to retest in a car seat at a future point (93.2%; n = 69). Four (5.4%) automatically admitted to a higher level of care, and 1.4% (n = 1) automatically discharged in a car bed. Whereas a majority of sites (70.3%) considered alternative plans for discharge after 2 failed CSTSs, 10.8% continued to repeat CSTS until the infant passed (Table 3).

When queried on the use of car beds after a failed CSTS, 20.3% (n = 15) never discharged in a car bed, whereas 1.4% (n = 1) automatically discharged in a car bed after 1 failed CSTS (Table 3). Half of the nurseries would consider car beds in the setting of repeated failures but did not routinely use them. For those considering discharge in a car bed (n = 59), more than half performed a CSTS in the car bed, 30.5% would discharge an infant in the car bed without testing, and the remaining 15.3% noted that the decision was provider dependent.

When it came to transitioning from a car bed to a car seat, fewer than one-third had the ability to schedule follow-up appointments either in the NBN, NICU, or specialty clinic for repeat CSTS (Table 3). Primary care physician (PCP) appointments were the only follow-up in 67% of sites. Most sites (67.8%) made no recommendations for repeat testing in a car seat to ensure safe transition from a car bed, leaving the decision up to the PCP. The remaining sites used various criteria for retesting, including weight, postmenstrual age ranging from 40 to 44 weeks, and timing after discharge (range of 2–6 weeks). One site did not recommend repeating outpatient CSTS and instead recommended transitioning to a car seat once achieving 40 weeks’ postmenstrual age.

For the 8 sites not performing CSTS, 7 noted that this was due to poor data regarding utility and outcomes. Three also noted a lack of staff and/or resources to perform the test. Single sites noted the following reasons: the Canadian Paediatric Society statement,¹³  which rescinded a recommendation for routine predischarge CSTS; desire to not give the family a false sense of safety from a passed CSTS; and limited availability of car beds, making discharge plans inconsistent.

Eighty-two site respondents answered survey questions on their attitudes regarding CSTS (Table 4). When asked if the CSTS is a “good way to assess cardiorespiratory readiness for discharge in at-risk newborns,” 39.2% of those who performed CSTS felt it was, 45.1% were unsure, and 25.7% felt it was not, as did 100% of the sites that did not perform routine CSTS (“no CSTS group”). Survey respondents in the no CSTS group did, however, recommend predischarge CSTS on certain infants, with half recommending testing infants with hypotonia and/or those with hemodynamically significant CHD and more than half recommending testing infants with airway malformations. Overall, the most common recommendations by survey respondents (n = 82) for performing CSTS included screening infants with hypotonia (82.9%; n = 68), airway malformations (78%; n = 64), hemodynamically significant CHD (63.4%; n = 52), and prematurity (61%; n = 50).

Providers were asked, “At what Spo₂ and for what duration would you become clinically concerned enough to continue monitoring and potentially delay discharge?” in the following scenario: “You are observing an otherwise healthy 36-week infant in the nursery who has their pulse ox on after undergoing a predischarge critical cyanotic heart disease screen. The infant is supine in a crib when you notice an oxygen desaturation.” There was no difference in the median Spo₂ level between the respondents from programs that did versus those that did not perform routine CSTS (89% vs 88%, respectively; P = .4). Overall, the median duration of concern was 20 seconds (IQR 20–30) for the entire cohort.

We performed a national survey of NBNs via the BORN network to determine management strategies and potential variation in practice of the CSTS. We found that >90% of NBNs did perform a predischarge CSTS and that the vast majority follow AAP recommendations for screening all infants born prematurely and testing duration. As with previous studies in NICUs, we found wide variation in failure criteria cutoffs for Spo₂ and heart rate. Most NBNs did repeat a CSTS after an initial failure, but timing of repeat testing varied. Although discharge in a car bed after CSTS failure was not common, we found that almost half of NBNs did not perform a similar period of observation in a car bed before discharge and that two-thirds of NBNs discharge infants in car beds with only PCP follow-up.

CSTS performance has been widely implemented in US NICUs, with 96.5% performing a CSTS as of 2020.¹²  In our study, we found a similar rate in NBNs (90.5%) across 35 states, which is a large increase from the 22% reported in 2003.¹⁴  This is important information because late-preterm infants are often cared for in NBNs to avoid NICU admissions, improve bonding, and facilitate breastfeeding. The sites that did not perform routine CSTS identified the lack of data on utility as the main driver. However, half believed that infants with hypotonia and those with hemodynamically significant CHD should undergo CSTS, whereas almost two-thirds responded that infants with airway malformations should undergo CSTS. Clearly, the usefulness of the CSTS remains unclear to clinicians, which emphasizes the need for rigorous study to determine which, if any, patients would benefit from routine screening.

Although there are no studies evaluating CSTS results and sudden unexplained infant death (SUID), there are data linking car seat position and SUID. In one recent study of SUID with no other known extrinsic risk factors (such as sleeping prone, bed-sharing, soft bedding, or having their head covered) researchers found that 27% of infants had been put to sleep in a semiupright car seat or bouncy seat.¹⁷  In another study evaluating infant deaths in sitting devices, researchers found that 63% occurred in car seats.¹⁸  The association between semiupright car seat positioning and SUID is concerning and requires further study. In a recent study of CSTS in NICU patients, researchers found that although patients who failed had longer lengths of stay, they also had significantly decreased odds of readmission within 30 days.¹⁹  In a study focused exclusively on NBN infants who were automatically admitted to the NICU for continuous monitoring after one failed CSTS, researchers found that 39% had ongoing apneic events while supine.⁷  And in a study evaluating late-preterm infants, researchers found that 21% of those who failed >1 CSTS from the NBN ultimately required NICU admission for significant hypoxia and that two-thirds subsequently required supplemental oxygen for safe discharge.²⁰  It remains unclear whether CSTS failure unnecessarily prolongs inpatient admission or properly identifies infants at risk for unstable cardiorespiratory status before discharge, thereby preventing adverse postdischarge outcomes, which is why more research is needed.

The AAP and the Committee on Fetus and Newborn both state that infants are not safe for hospital discharge until “physiologically mature and stable cardiorespiratory function has been documented for a sufficient duration.”^21,22  This must apply not only while supine in a crib but also while semiupright in a car seat. However, in the absence of standardized screening and failure criteria, each institution has been left to determine which vital signs are clinically concerning for unstable cardiorespiratory function, leading to variation in CSTS practice. Since the 1980s, Spo₂ constituting failure in the literature has ranged from <85% to <95%.²³  In 2 recent surveys of NICUs, the Spo₂ and duration constituting failure ranged widely from <80% to <93% and from any drop below the threshold to requiring >60 seconds, respectively.^11,12  Similarly, we found that NBNs also have a wide range of Spo₂ failure cutoffs, with values ranging from <85% to <93% and a similar duration range.

One of the main barriers to further research is the heterogeneity of CSTS protocols in NICUs and NBNs. Consensus on failure criteria across the nation is important not only to encourage standardization of care but also to allow the performance of large multicenter trials to determine the utility of the CSTS. This remains challenging because data are limited on what constitutes a safe lower Spo₂ in high-risk infants near the time of discharge. Interestingly, a recent review of CSTS literature,²⁴  as well as both the NICU surveys^11,12  and our NBN survey, revealed that the most commonly used Spo₂ failure cutoff was <90%. When providers in our study were asked what Spo₂ would cause enough clinical concern to potentially delay a discharge from the NBN, the median value we found was 89%. Because Spo₂ <90% is the most commonly used criteria in the NICU literature as well as in our NBN study, this value could be used to standardize CSTS failure criteria via expert consensus. Early-preterm infants with Spo₂ target values <90% had increased inpatient mortality compared with those with Spo₂ target values >90% during their hospitalization,²⁵  although it is unclear how applicable these data are to otherwise healthy late-preterm infants in NBNs. However, there is a precedent for using Spo₂ <90% as a screening cutoff, namely in the CCHD screen. When performing routine critical CHD screens, the AAP and American Heart Association recommend that any infant found to have Spo₂ <90% requires a “comprehensive evaluation for causes of hypoxemia.”²⁶  Therefore, applying this value to CSTS as identifying unstable cardiorespiratory status in the car seat seems reasonable.

Less variation exists for the definitions of clinically concerning apnea and bradycardia. More than 83% of NBNs defined apnea as a breathing pause lasting >20 seconds. The most common criterion to define bradycardia was a heart rate <80 bpm, used by 65% of NBNs in our study and, similarly, by 70% of NICUs.¹¹ 

The AAP clinical report recommends “interventions to reduce the frequency of desaturation and episodes of apnea and bradycardia” when significant cardiorespiratory events occur in the car seat.⁹  Examples include employing certified CPSTs to optimize positioning, use of car beds, supplemental oxygen, continued hospitalization, and further medical assessment. The AAP clinical report states that infants with cardiorespiratory events in the semiupright position “should travel in a supine or prone position in an…approved car bed after an observation period” similar to the CSTS.⁹  We found that ∼5% of sites routinely admit to a higher level of care and that ∼1% routinely discharge in a car bed after an initial failed CSTS. Despite no specific mention in the clinical report, we found that the most common practice in NBNs is to retest in a car seat at a future point with the goal of sending the infant home in a rear-facing car seat in a semiupright position. The avoidance of car beds for discharge despite the AAP report is likely due to multiple factors, including greater complexity in their use and the cost to families. In addition, rear-facing car safety seats have been more extensively crash tested and therefore have better documentation of protection than car beds.²⁷  Studies of preterm and term infants have revealed similar rates of desaturation and bradycardic events in both devices,^28–31  indicating that car beds are not necessarily a safer choice for all infants who fail their initial CSTS, especially when a car bed observation screen is not routinely performed, as was the case in ∼45% of NBNs surveyed. As recommended by the AAP,⁹  if discharge is planned in a car bed, a similar period of observation (ie, car bed screen) should be performed given the risk of unstable cardiorespiratory status in NBN patients that has also been demonstrated in car beds.³²  Infants for whom a car bed might be beneficial will be better identified by performing such a car bed screen. If an infant is discharged in a car bed, a plan should be established, by the time of discharge, to evaluate the infant’s safety for transitioning to a rear-facing car seat. The PCP should be contacted before discharge and should be involved in all follow-up planning. PCPs may not have the resources to perform a period of observation to assess cardiorespiratory safety in the outpatient setting, so leaving follow-up decisions only to the PCP is not always appropriate. Additional consultations may be required to safely transition to a rear-facing car seat, such as performing polysomnography or repeating a CSTS in a specialty clinic.

Strengths of our study include the use of the BORN network, which provided a large, diverse, nationally representative sample of NBNs. We had a strong survey response rate (79%), representing NBNs from 35 states, making these data widely generalizable. We were able to obtain data on CSTS practice as well as data on newborn providers’ attitudes toward CSTS performance. These data will help identify focus areas for future research on this controversial topic. One limitation is that data were based on survey response by an assigned site representative rather than direct observation of the written CSTS protocol, so any differences between the practice and survey response could not be assessed. In this survey, we did not address whether the car seat used during the CSTS was the same one used at discharge or whether the sites had an established policy in place for education of caregivers regarding car safety seat use,³³  which would be important for future studies. In addition, data on providers’ attitudes were based solely on the specific respondent and were not necessarily representative of all newborn providers at their institution.

Although we found a large degree of variability in implementation of CSTS in a cross section of NBNs across the United States, we observed trends regarding screening practices and failure criteria that can inform future practice and policy. This variability suggests the need for more specific guidance by the AAP if the CSTS is to be reproducible across sites, and to help in understanding if this test is clinically important in the detection of cardiorespiratory instability in the semiupright position in at-risk infants.

We thank the members of the BORN network for their participation in this study and specifically Beth King for her invaluable assistance throughout the study.

AAP

American Academy of Pediatrics

BORN

Better Outcomes through Research for Newborns

BW

birth weight

CHD

congenital heart disease

CPST

child passenger safety technician

CSTS

car seat tolerance screen or screening

GA

gestational age

IQR

interquartile range

NBN

newborn nursery

PCP

primary care physician

Spo₂

hemoglobin-oxygen saturation

SUID

sudden unexplained infant death