#431 - [Journal Club] - 📌 Should We Stop Routine Car Seat Testing for Preterm Infants?

In this Journal Club episode, Ben and Daphna review a systematic review and meta-analysis from JAMA Network Open questioning the clinical value of predischarge car seat tolerance screening (CSTS). Driven by data suggesting that testing does not reduce 30-day mortality or hospital readmissions, they discuss the high failure rates, varying definitions of bradycardia and desaturation, and the unintended consequence of prolonged NICU stays. They also highlight the practical reality of CSTS in ensuring parents actually have an appropriate car seat at discharge. Tune in for a critical look at whether this 1991 AAP recommendation still holds up today!

----

Predischarge Car Seat Tolerance Screening in Preterm and At-Risk Full-Term Infants: A Systematic Review and Meta-Analysis. King BC, Dalvie N, Hay S, Jensen EA, Zupancic JAF.JAMA Netw Open. 2026 Feb 2;9(2):e2558197. doi:0.1001/jamanetworkopen.2025.58197.

Transcript

Ben Courchia MD (00:00.6) Hello everybody, welcome back to the Incubator podcast. We're here for a new episode of Journal Club. Daphna, good morning. How are you today?

Daphna Yasova Barbeau MD (00:08.549) Good morning. I'm doing well. I feel like we didn't get to talk all I wanted to talk about on the steroid paper. But we're moving on today. If only we had a more active community somewhere on social media where we could continue the discussion. No, I don't want to do that.

Ben Courchia MD (00:18.306) I'm sorry about that. It's kind of my fault. Daphna wants to build her own social media platform. She's the next tech mogul.

Daphna Yasova Barbeau MD (00:38.981) That's comical in this exact situation. Anyways, I want to talk about car seats today. This was an article in JAMA Network Open, Predischarge Car Seat Tolerance Screening in Preterm and At-Risk Full-Term Infants: A Systematic Review and Meta-Analysis. Lead author Brian King, and this whole group—Dr. Dalvie, Dr. Hay, Dr. Jensen, Dr. Zupancic—so quite a group.

Ben Courchia MD (01:17.174) All are friends of Beth Israel.

Daphna Yasova Barbeau MD (01:16.604) That's right. Where does the pre-discharge car seat tolerance screening come from? It has been recommended by the American Academy of Pediatrics, first endorsed in 1991. That was based on observations from some small studies showing that preterm infants and full-term born infants with risk factors like low birth weight can have vital sign instability when they're semi-upright in a car seat. So we really moved to universal screening for the premature baby and some of these at-risk full-term infants. Now, the pendulum is in some ways swinging a little bit. There was a Cochrane review published in 2006 that found no randomized clinical trial that addressed this question. The Canadian Paediatric Society removed recommendations for car seat testing in 2016. Kaiser Permanente in California moved away from routine car seat testing in 2016, and they have subsequently published data on more than 40,000 infants after that change, showing no change in the rates of measured adverse outcomes after not doing routine car seat testing. That's really what they wanted to evaluate. This was a systematic review and meta-analysis to estimate the association of car seat testing with length of stay and post-discharge outcomes, and to estimate the overall proportion of infants who fail an initial and subsequent test. They looked at English language articles published before June 2025. They had the following inclusion criteria: one, a newborn population before hospital discharge; two, car seat testing with established failure criteria. They included all study types, but categorized them into three subgroups: one, randomized control trials; two, non-randomized intervention studies that included a comparison group of infants that did not undergo or did not have a documented passed car seat test prior to discharge; and three, single-group observational studies that reported car seat tolerance failure rates. The exclusion criteria were: one, observational studies that described clinical instability without reporting car seat test failure; two, observational studies of adherence to testing recommendations; three, surveys of clinical practice; and four, studies that were limited to certain subgroups like, for example, just congenital heart disease. They looked at the literature and found 857 unique articles.

Daphna Yasova Barbeau MD (04:08.155) Further review identified 76 articles for full-text review. 55 of these were excluded based on the exclusion criteria. That left them with 21 studies. Most studies were single-group and lacked a comparison group of infants who did not undergo car seat testing. Three studies, which was 14% of the studies, included a comparison group discharged without a passed car seat test and were categorized as non-randomized interventional studies. This included 54,000 participants, 27,000 of which were the control group without a car seat test. Most of the studies were of preterm infants, so gestational age less than 37 weeks. This was about 67% of the studies. Studies could have included babies admitted to the NICU or the well-baby nursery, and that will become important later on. Regarding car seat test failure criteria, all the included studies were consistent with their definition of apnea. They used 20 seconds, but interestingly, they used variable thresholds to define bradycardia and desaturation events. And we will talk about that.

Ben Courchia MD (05:33.102) Do you think we'll revise this definition one day, like this 20-second definition of apnea?

Daphna Yasova Barbeau MD (05:37.816) Yeah, maybe. I think it was really interesting to see what the criteria were. Again, pretty consistent with the apnea duration, but the bradycardia cutoff wasn't even that consistent. 76% of studies used less than 80 beats per minute, but there was a wide variation of bradycardia duration: any duration in 14%, 10-second duration in 28%, five-second duration in 10%, persistent bradycardia in 10%. There was also a really wide range of desaturation cutoffs. What do you even decide is a failure? Apnea? Sure. Everybody agreed on 20 seconds of apnea. But was it any desaturation? That was nearly 10% of studies. The desaturation cutoff was 93% in 9% of studies, 90% in 24% of studies, and 85% in 10% of studies. People were using cutoffs anywhere from under 93% to 85%. To me, those are two very different types of desaturations. That's one thing I think to be considered. Among the 21 included studies, there was an overall estimate of 8.62 failures per 100 patients. Now, among the subgroup of patients undergoing testing in the nursery, the estimate was 9.86 per 100 patients compared with 7.11 failures per 100 patients among infants who were being tested in the NICU. So there were actually more failures in the newborn nursery than there were in the NICU. Interestingly, eight studies included patients undergoing the initial testing in the nursery or the NICU. In six of those studies, the first test failure rate was higher in the nursery as compared to the babies in the NICU.

Daphna Yasova Barbeau MD (07:59.804) Among preterm infants, the overall estimate was 8.56 failures per 100 patients compared with an estimate of 10.72 failures per 100 patients among full-term born infants. That gave me pause that they were seeing more failures in full-term born infants. But again, full-term infants that get selected for car seat testing have other comorbidities or risks for failing the test. That's why we're including them. Six studies included both preterm and full-term born infants undergoing their initial test, and in five out of the six studies, there were higher first test failure rates among term infants compared to preterm infants. What about repeat car seat tests? 11 studies with a total of just over 900 babies reported results for repeat testing among infants who failed the initial screening. The timing of repeat testing was variable. We always have this discussion. In six studies, repeat testing was performed between 12 and 48 hours. In two studies, the repeat testing was delayed for more than 48 hours. In three studies, they didn't report the repeat test interval. The overall estimate for the 11 included studies was 24% of babies—so 24 out of 100 patients who failed the first car seat test failed a second test, so one in four. Two studies reported results of repeat testing among infants regardless of whether they failed the initial test, which I thought was interesting. In Davis et al., 11% of moderately preterm infants (six out of 53 infants) who passed an initial car seat test failed one of two repeats 24 to 48 hours later. I thought this was a very interesting fact. Additionally, in DeGrazia et al., the second study, among the 41 preterm infants who passed their initial car seat test, 10% failed repeat testing just 12 to 36 hours later. What about length of stay? Only two studies compared pre-discharge length of stay between patients who underwent testing before discharge compared to patients discharged without it. They weren't able to pool it just because the samples were so different.

Daphna Yasova Barbeau MD (10:18.651) But in one study, they reported no difference in unadjusted overall length of stay when comparing 21,000 infants before and 20,000 infants after discontinuation of routine car seat testing. In the other study, they reported no significant difference in length of stay, maybe six hours longer for the babies who failed the car seat test. However, when they stratified the analysis in this study, they found that routine car seat testing was associated with a significantly longer adjusted length of stay for those babies who were being tested in the NICU. Those babies on average stayed an additional nearly 36 hours longer, which makes sense based on the average retesting interval. They mentioned based on these two studies, there's a low certainty of evidence that car seat testing is associated with overall average pre-discharge length of stay, which is actually not what it feels like, but that's why we do studies. And post-discharge outcomes. Three studies described post-discharge outcomes with a comparison group. They weren't able to pool these either based on differences in study design and outcome reports. But two studies reported death within 30 days of discharge. The Jensen et al. paper observed no deaths in the cohort of infants discharged without a passed car seat test. Braun et al. observed fewer deaths after stopping routine pre-discharge car seat tests, an odds ratio of 0.26. After adjustment for other clinical factors, the difference was not significant. All three studies reported no significant difference in rates of readmission within 30 days after discharge, and a meta-analysis of data from the three studies similarly showed no difference in an unadjusted readmission rate. Two studies included death or readmission within 30 days as a composite outcome, and there were no significant differences between the groups in either study or in the pooled rate of death or readmission within 30 days. So they conclude...

Daphna Yasova Barbeau MD (12:35.961) ...that testing was not associated with a reduction in 30-day mortality or hospital readmission. Approximately 10% of infants fail initial testing, of whom one in four will fail their repeat testing. I wanted to underscore that even babies who passed their test often failed a subsequent test. I thought that might be potentially the most interesting piece of data. They note that failure and repeat testing potentially contributes to prolonged hospitalization in this subset of infants without clear evidence of benefit. As such, current recommendations for routine car seat testing in all preterm infants may merit reevaluation. I will say one thing that our population's car seat testing has been very effective at, and that is identifying how many families do not actually have car seats at the time of discharge. If we didn't test them for car seats, I think they would go home without one. Actually, we've had the Broward County Sheriff's Department come in and do discharge car seat teaching with families. They say the proportion of babies that are not in car seats on any random stop is frightening. So, I mean, for that reason alone...

Ben Courchia MD (13:44.703) Yeah, teaching. Yeah. Frightening.

Daphna Yasova Barbeau MD (13:57.55) ...in our neck of the woods, identifying a car seat. That's not the purpose of the test. It causes everybody a lot of anxiety. The nurse that has to do it, the family, it may cause the baby anxiety. I don't know.

Ben Courchia MD (14:12.11) But that's not the reason for the test. I invite you all to listen to a recording that we did at Hot Topics in Neonatology with Michael Narvey about car seat testing. All of my thoughts are encompassed in that episode, because like we were discussing, putting a baby on a leveled floor that doesn't move in any direction...

Daphna Yasova Barbeau MD (14:18.512) Hmm. Yes. Mm-hmm.

Ben Courchia MD (14:39.672) ...for two hours, how is that simulating a car ride? Either you put the baby in a car seat on a flat surface and say, "I'm doing a fitting," and you fit the baby for the car seat. But you're just simulating a baby being in a car seat for two hours. You're not simulating a drive. Does every baby need a car seat test? I don't know. This is something that we do that is not causing a lot of angst amongst the community, so we keep doing it, but we don't have a good reason for it. We all feel very embarrassed when the baby fails the car seat and the parents are like, "What now?" And we just do it again. "If you pass, go home tomorrow." It feels so silly, but it is what it is.

Daphna Yasova Barbeau MD (15:18.199) We just wait. And they're like, "How long do you wait?" And you're like, "I don't know." Sure. Now, I will say, car seat testing and identification of the car seat is very important. There are lots of babies who...

Ben Courchia MD (15:39.678) Agreed. "Oh, I don't have the base." Or sometimes you see parents bringing a car seat that's way too big for their former preemie. And you're like, "Oh my God, this kid is way too big."

Daphna Yasova Barbeau MD (15:44.738) Yeah. Or you can even see that within a set of multiples. Like one kid fits fine in the car seat and the other kid just does not. It matters. Not every car seat is built for every baby. So that's a different discussion. All right. Sounds good.

Ben Courchia MD (15:54.367) Mm-hmm. Anyway. All right. I will see you tomorrow.