Screening for Cardiopulmonary Events in Neonates
Screening for Cardiopulmonary Events in Neonates
The infant car seat challenge (ICSC), or period of observation in a car safety seat before discharge to monitor for episodes of apnea, bradycardia and desaturation, is one of the most common tests performed on preterm neonates in the United States. However, the utility of the ICSC to identify infants at risk for adverse cardiopulmonary events in the car seat remains unclear. Minimal evidence exists to guide clinicians in performance of this test including appropriate inclusion criteria and failure criteria. In this article, the origins of the ICSC are discussed as well as potential etiologies of desaturations and bradycardia in the car seat position. Current literature on implementation, inclusion and failure criteria, incidence of failure and data on the meaning of a 'passed' vs 'failed' ICSC are discussed. Emphasis is made on minimizing time in car seats and seated devices given concern over the risk of desaturations.
Research by the National Highway Traffic Safety Administration found that proper use of child safety seats reduces fatal motor vehicle injuries by 71% in children younger than 1 year of age. For that reason, the American Academy of Pediatrics (AAP) began recommending that all neonates and young children be transported in approved rear-facing semiupright car safety seats in 1974.
At the same time, advances in obstetrical management and neonatal care led to improvements in survival of preterm and critically ill neonates. Studies in the 1980s alerted providers that preterm neonates were at increased risk of apnea and drops in Hb-O2 saturation in arterial blood (desaturations) when placed in their car safety seats. Because of this, the AAP began recommending a period of observation in a car safety seat before discharge to monitor for episodes of apnea, desaturations and bradycardia for all infants born <37 weeks gestational age (GA), a test now referred to as the car seat test or infant car seat challenge (ICSC). In the United States alone, ~500 000 infants are born prematurely (<37 weeks GA) each year, making the ICSC potentially one of the most commonly performed tests on neonates. However, despite the widespread implementation of the test, very little information is available for clinicians to evaluate its usefulness, which is a major concern. The aim of the ICSC is to identify infants at risk of cardiopulmonary compromise in the semiupright position, but limited information exists on failure rates, appropriate vital sign parameters for a 'failure' and whether this test will identify infants who truly are at risk of adverse events. What a 'failed' ICSC means for the infant is still unclear, leading to much confusion and frustration. A small number of reviews have been published attempting to elucidate this controversial topic, although often uncovered more questions than answers. What we do know is that infants spend significant amounts of time in their car seats and other semiupright positioning devices after discharge. Parents surveyed about infant seating device practices reported that 94% of infants <5 months old spend >30 min in a seating device daily, averaging 5.7 h daily. Parents reported that 38% of infants spent >1 h specifically in a car seat daily, with 33% for 1 to 2 h and 5% for 3 to 4 h. Therefore, identifying cardiopulmonary safety of neonates in the seated position, such as in a car seat, is of the utmost importance.
The purpose of this review is to provide a comprehensive summary of existing literature on: (1) the origin of the ICSC recommendations, (2) current ICSC practices and (3) the evaluation of the ICSC as a tool to identify infants at risk of cardiopulmonary compromise in the semiupright position. Available data related to respiratory safety in the car seat and ICSCs are discussed.
In 1985, Bull and Stroup were among the first to report that low birth weight (LBW) and prematurity pose unique dilemmas when it came to safe transportation of infants. At that time, federal standards did not specify minimum size or weight for which a car safety seat was deemed appropriate. They noted that a number of car seat models allowed neonates weighing <2.2 kg (5 lbs) to slouch, had shoulder straps that fit too loosely and had lap pads that covered the neck and face. They recommended various techniques including supportive blanket rolls to minimize this risk, but concern remained about the fit and safety of preterm infants.
Based on these findings, Willett et al. hypothesized that premature infants may be at increased risk of cardiopulmonary compromise in the car seat. In an article published in 1986, they compared 20 preterm infants (12 with a history of apnea of prematurity and concurrently on methylxanthine treatment and 8 with no history of apnea or treatment) to 10 healthy term infant controls while in the car seat position. They evaluated oxygen levels: (1) at baseline, (2) during a 30 min period in the car seat and then (3) for 30 min afterwards while placed prone. They noted the preterm infants had significant decreases in Hb-O2 saturation during the car seat interval, spent longer times saturating <85% and had more frequent Hb-O2 saturation drops to <80% than the term group. Time spent <90% and <85% was longest in preterm infants with a history of apnea, followed by preterm infants without a history of apnea, both statistically significant when compared with term controls. They did not note any episodes of apnea in either group.
Subsequently, Willett et al. hypothesized that preterm infants in the upright car seat position would be at increased risk of obstructive apnea and/or decreased minute ventilation due to decreasing tidal volume, and that this may result in hypoxia and bradycardia. Sixty-two infants admitted to the neonatal intensive care unit (NICU), both term and preterm, underwent multichannel recordings and 50 had complete pulmonary function tests. They noted that infants spent significantly more time with Hb-O2 saturation <85% and had more short apneas (lack of airflow and respiratory effort for 5 to 20 s) during the car seat interval than while supine. They found that none of the term infants (n=9) had an episode of bradycardia, desaturation and/or apnea that required intervention during the car seat interval but that 30% of the preterm infants did (n=16 of 53). However, they found that pulmonary function seemed to improve while in the car seat, with stable tidal volumes, increased compliance and decreased pulmonary resistance overall in the cohort. When performing a subgroup analysis comparing <32 weeks GA, 32 to 36 weeks and >36 weeks, the two most preterm groups were noted to have improved resistance, compliance and work of breathing while in the car seat despite abnormal changes on the multichannel recordings. They were unable to uncover pulmonary function abnormalities to account for the vital sign changes while in the semiupright position. They concluded that the cause of respiratory compromise in the car seat is multifactorial and may be different for each infant.
Similarly in 1989, Carlo et al. reported that former preterm infants corrected to 8 to 10 months of age had decreased pulmonary resistance, increased lung compliance and no change in minute ventilation in the semiupright position compared with supine. They did note that neck flexion significantly increased pulmonary resistance and decreased minute ventilation. Questions remained regarding the cause of cardiorespiratory compromise while in the semiupright position as it was not easily explained by pulmonary mechanics.
An initial pilot study of a car seat insert that allowed the head and neck to lie in neutral position demonstrated maintenance of a larger upper airway space radiographically and significant reduction in the frequency of desaturation events, suggesting that neck flexion is a major contributor to abnormal respiratory status in the car seat. A larger randomized controlled trial of this car seat insert showed a reduced rate of obstructive respiratory events and severity of Hb-O2 desaturation events (<85%) in healthy full-term infants, but did not affect the overall rate of moderate desaturations. Similar studies of inserts that prevent lateral flexion have shown no improvement in oxygenation. Neck flexion alone therefore does not explain desaturation events in all neonates, although it may contribute to events.
Hertz et al. performed polysomnography (PSG) on 28 preterm infants in the car seat and found that during car seat time, the presence of at least one abnormal breathing event (apnea, desaturation, bradycardia or periodic breathing) was significantly higher than while prone, and that periodic breathing was more frequent while in the car seat. However, they noted that periodic breathing had a significant sleep-state effect and was more common during active sleep, regardless of position of the infant. Active sleep was found to be significantly higher while in the car seat position. The authors concluded that it may be the increase in active sleep while in the car seat, rather than the actual position, that may account for periodic breathing and perhaps other respiratory abnormalities.
The exact cause of desaturations while in the semiupright position is still unclear. Clearly, desaturation events in the car seat are multifactorial and preventing and treating these events continues to be complicated. Immature respiratory control and prolonged apnea of prematurity, low tone and risk of improper neck flexion and upper airway position, and sleep state-associated effects all likely have a role to varying degrees.
Because of these concerns, in 1991 the AAP began recommending a period of observation in a car safety seat before discharge to monitor for episodes of apnea, desaturations and bradycardia for all preterm infants born <37 weeks GA, now known as the ICSC. For those infants who had evidence of Hb-O2 desaturations or bradycardia during an ICSC, the initial recommendation was that they 'travel prone in an alternative seating device'. They stated that a supine position could be considered but was noted to be concerning due to 'potential for increased risk of aspiration'; however, with information regarding SIDS (sudden infant death syndrome) and back-to-sleep, prone positioning is now discouraged even for travel. It was also noted that parents should be counseled to avoid infant swings, infant carriers or other equipment that places the infant in the semiupright position. In 1996, the recommendation for a period of car seat observation was reiterated, and hospitals were encouraged to 'develop policies to include this evaluation in their discharge planning process'. Infants who 'failed' this observation period were recommended at that time to travel in a supine or prone position in an alternative safety device, and again minimizing travel and other upright equipment was emphasized. Again, prone positioning has since been discouraged.
An attempt was made by the AAP to standardize car seat testing with the 2009 recommendation, suggesting performing ICSCs for a minimum of 90 to 120 min or the duration of the car ride home, whichever is longer. Additional recommendations were to consider inclusion of infants who were full term but possessed increased risk of cardiorespiratory compromise while in the semiupright position, such as those with neuromuscular disorders, congenital heart disease and micrognathia. However, the criteria for a 'failed' test were still left up to individual institutions and clinician groups as was the management of this group of infants. It was noted that 'if events documented on cardiorespiratory monitoring in a car safety seat are deemed significant by the treating physician or hospital policy, interventions to reduce the frequency of desaturation and episodes of apnea and bradycardia are recommended'. Suggestions for this management included supplemental oxygen, travel supine in a car bed after a similar period of observation or further medical management. It was advised that a repeat ICSC should be considered in those infants who failed initially before transitioning from a car bed to a semiupright car seat.
After the early recommendations of the AAP to perform ICSCs, clinicians were left to decide the duration of the test. A 2003 survey of ICSC practices noted that 33% of NICUs were testing for 30 min, 24% for 60 min and 33% for 90 min. In 2007, Salhab et al. demonstrated that the proportion of infants whose first event occurred after 60 min of observation was 40%, and after 90 min was 30%, indicating that a shorter testing duration may miss clinically significant events. Current AAP guidelines recommend testing for 90 to 120 min or the duration of the car ride home, whichever is longer. This evidence also demonstrates the increased risk of Hb-O2 desaturation events with increased duration of time in the semiupright position, making it important to emphasize limiting time in this position.
There are no current AAP guidelines for what level of desaturation should constitute a 'failed' ICSC. A 2013 survey of NICUs performing ICSCs demonstrated that the definition of Hb-O2 saturations constituting a failure ranged from <80% to <93%, and ranged from any drop up to >60 s. The most commonly used desaturation level was <90%, with 37% of NICUs using this level; the most common definition of a bradycardia was <80 b.p.m., used by 70% of NICUs. In 2010, Bass acknowledged that though the definition of bradycardia <80 b.p.m. is widely accepted, a safe lower Hb-O2 saturation level is not universally agreed upon and recommended <93% as a cutoff for failure based on a review of normal Hb-O2 saturation ranges in newborns. More recent prospective studies have had a variety of Hb-O2 saturation failure cutoffs, generally between 85 and 90%, with 90% remaining the most commonly used cutoff in published literature. The wide range in the literature reflects the wide ranges used in clinical practice, making many studies difficult to generalize.
Standardization of failure criteria has remained challenging despite data on oxygenation ranges in healthy former preterm infants. Multiple studies have evaluated overnight recordings of arterial oxygen to obtain normative baseline data in otherwise healthy preterm infants at the time of discharge as well as in full-term infants. Median baseline Hb-O2 saturations were 98 to 99% in both term and preterm infants at the time of discharge (corrected to a mean GA of 36 to 37 weeks depending on the study), although the preterm infants had a wider range of baseline saturations than did full-term infants (88% to 100% vs 97% to 100%). Median Hb-O2 saturations increased in a preterm group to 100% at 6 weeks after discharge (range 95 to 100%). Despite improvement with time, the lower range of Hb-O2 saturations remains below those born at term. Without longer-term studies on those preterm infants with clinically unrecognized low baseline Hb-O2 saturations, it is difficult to determine clinical significance. Although preterm infants with a history of bronchopulmonary dysplasia whose baseline Hb-O2 saturations with sleep are persistently <92% have poorer growth velocities than do those with saturations >92%.
Episodes of Hb-O2 desaturations (intermittent hypoxia, IH) were frequently found on recorded oximetry in otherwise healthy preterm infants once deemed ready for discharge, but were often not detected clinically. When analyzing the CHIME data, a 2011 study demonstrated that 79% of former preterm infants (once corrected to a minimum of 36 weeks) and 65% of term cohorts had ≥1 IH episode (defined as a drop in Hb-O2 saturation <90%). Interestingly, there were no differences in the frequency or severity of IH between symptomatic neonates (noted to have at least one episode of apnea or bradycardia documented in nursing note during the 5 days before discharge) and those who were symptom-free. The median number of IH episodes in the preterm cohort was 4 with a wide range of 1 to 98, indicating that IH is commonly seen in this population. It is possible that the ICSC is capturing these IH events, which are commonly occurring anyway, as it is a time of intensive monitoring.
However, even if low baseline arterial saturations and IH events occur commonly in otherwise healthy preterm infants once corrected to term and ready for discharge, the potential influence of such hypoxemia on clinical and developmental outcomes remains uncertain. Samuels et al. studied 91 former preterm infants who had an apparent life-threatening event (ALTE) and then underwent 8 to 12 h recordings of breathing movements, electrocardiograms and arterial oxygen saturation and compared these with 110 healthy preterm infants made at 6 weeks after discharge. Twenty-one percent (n=19) of those infants experiencing an ALTE had hypoxic episodes, almost 7% (n=6) had low baseline Hb-O2 saturation (<95%), and 19% (n=17) had both. Thirty-one of 33 infants subsequently treated with supplemental oxygen had significant reduction or cessation of ALTE events, leading the authors to conclude that early recognition and treatment of abnormalities in episodic or baseline hypoxemia may reduce risk of further ALTE in former preterm infants. Longer-term follow-up of infants after discharge in the CHIME study demonstrated that those who had 5 or more cardiorespiratory events (apnea and/or bradycardia) had significantly lower Bayley-II Mental Developmental Index scores than those who did not have events. And, a large systematic review of literature related to chronic or IH in children demonstrated adverse impact on development, cognitive performance and behavior. This seems to indicate that early identification and treatment of hypoxemia is necessary to prevent adverse cardiopulmonary outcomes, which is one of the goals of the ICSC. However, it remains unclear whether the ICSC achieves this goal.
One concern when attempting to standardize ICSCs is the variation in monitors and averaging times. The number and percentage of desaturations identified depends on the duration and the averaging time set on the monitor. Longer averaging times are often used clinically because they have lower likelihood of false alarms related to motion artifact common in neonates. However, these longer averaging times may underestimate the frequency of brief desaturations. Conversely, shorter averaging times increase the rate of false alarms, but are more sensitive to detect brief intermittent desaturations. Recent prospective studies use 2-s averaging times to evaluate brief desaturation events, but national standards for averaging times for clinical use do not exist. Vagedes et al. found that 96% of Hb-O2 desaturations <80% in healthy preterm infants were brief, lasting <20 s, demonstrating the importance of averaging time on evaluating desaturation events.
It remains up to each NICU clinician group to determine failure criteria for their unit. The Canadian Paediatric Society Statement acknowledges this lack of a clear definition of a significant cardiopulmonary abnormality, but empirically defined it as 'two or more episodes of oxygen desaturation <88% for 10 s or more during a 90 min period of observation.'
ICSC implementation in the United States was evaluated in a 2013 survey of Level II and III NICUs in New England and New York, noting that 11% did not perform any form of an ICSC. Of those who did perform ICSCs, 17% did not test all born <37 weeks GA and used more premature birth GAs as inclusion criteria. Although there are no specific recommendations about including LBW infants in testing, 45% of NICUs surveyed did test infants regardless of birth GA if they were LBW (<2.5 kg). Only 55% had tests that lasted a minimum duration of 90 min, recommended by the most updated AAP policy.
A number of studies have attempted to evaluate incidence and predictors of failure to limit the scope of the testing to the most at-risk infants. Failure rates in the literature range widely from 4 to 30%, but studies have been limited by small sample size and lack of consensus on appropriate failure definitions. Predischarge weight <2000 g has been linked to increased risk of failure. And infants with a history of apnea of prematurity also appear to be at increased risk.
The largest study evaluating incidence of failure retrospectively reviewed data on >1100 infants born preterm and found a 4.3% overall failure rate in this population, with failure criteria ranging from Hb-O2 saturations <88% to <90% at the participating hospitals. The only significant risk factor was, contrary to the authors' hypothesis, advancing birth GA (odds ratio for failure was 1.07 for each additional day of GA). On subgroup analysis, the failure rate for early preterms (<34 weeks birth GA) was 2.4%, and for late preterms (34 to 36+6/7 weeks) was 5.6% despite being tested at approximately the same corrected GA. The authors hypothesized that the early preterm infants were tested at chronologically older ages and were therefore more mature than the late preterm infants, but the etiology and ramifications of these findings are still unclear. Although this indicates that late preterm infants are still at risk of abnormal respiratory patterns in the car seat.
The care of late preterm infants remains controversial, as this group rarely requires as much support as early preterms, but are still premature and at risk of immature respiratory control. As noted earlier, a number of NICU protocols limit testing to early preterm infants. However, there is evidence that late preterms may have higher rates of desaturation in the car seat position than early preterm infants. Merchant et al. noted poor fit in the car seat in 24% of late preterm infants and 4% of term infants (who of note were LBW). Mean Hb-O2 saturations declined for both late preterm and term infants during the 90 min study, but 12% of the late preterm infants vs none of the term infants had significant apnea or bradycardia while in the car seat.
The most recent AAP guidelines note that many hospital protocols also include ICSCs for infants other than those born preterm, who may also be at risk for cardiopulmonary compromise in the semiupright car seat position, including infants with congenital heart disease, hypotonia and micrognathia. In fact, up to 45% of NICUs surveyed include LBW as an inclusion criterion for testing, even in term neonates. However, data are limited to suggest which, if any, term infants would benefit from ICSCs.
In 1995, Bass et al. evaluated term infants who 'in the judgment of their pediatrician were felt to be at risk for O2 desaturations or apnea' for reasons such as: LBW, respiratory symptoms, genetic disorders and history of apnea. They found 28.6% had desaturations <90% during the test. Simsic et al. evaluated ICSCs for infants after cardiac surgery and found that 4 of 66 infants (6%) failed (defined as drop >15% from baseline Hb-O2 saturation), all of whom were full term. However, the authors acknowledged that these were highly selected populations, all with medical issues that prompted the testing.
A cross-over study of healthy full-term infants (38 to 41 weeks at birth) in both the car seat and car bed tested 150 infants for 60 min in each device (to limit time away from families) and a subgroup of 50 infants for 120 min to evaluate the hypothesis that respiratory compromise does occur in these devices. They found that the percent of total time with Hb-O2 saturation <95% during the 60 min observation was significantly higher (23.9%) while in the car seat than in the car bed or while supine in a crib (17.2% and 6.5%, respectively). Fifteen percent spent >50% of their time with Hb-O2 saturations <95% in the car seat (majority 90 to 94%). They noted the results were almost identical for the 120 min group. They also noted that regardless of order of device tested, infants saturated better during the first hour, indicating that limiting time in either device should be recommended. A similar study evaluating at Hb-O2 saturations in term infants while in a car seat for 30 min demonstrated mean Hb-O2 saturations were significantly lower when in a 45° car seat position (95.8%) when compared with supine in bed (98.7%) or a car bed (98.8%). They did note moderate desaturations (<90%) in 27% while in the car seat and none while in the crib, although this was a small study of only 15 subjects. However, the clinical significance of these desaturations noted in healthy full-term neonates remains unclear.
A Cochrane Review from 2006 attempted to assess whether predischarge ICSCs prevent morbidity and mortality in preterm infants, but found no trials for inclusion and concluded that it is unclear whether the ICSC is beneficial or harmful to patients. There is very little data on what a failed ICSC means for our patients.
A few groups have performed PSG during ICSCs to evaluate the ability of the test to identify infants at risk of abnormal breathing patterns. Elder et al. studied 18 former preterm infants within 1 to 2 days of discharge during sleep in the car seat and while supine. A failed ICSC was defined by Hb-O2 saturation <90% with distress, apnea >20 s or any respiratory distress, whereas a failure of the PSG occurred with recurrent desaturations <90% over 5 min associated with obstruction, or apnea. They found that a clinically observed ICSC was more sensitive in predicting infants with increased obstruction in the car seat (sensitivity 80%) than failure of the PSG (sensitivity 60%), but had poor reliability and a low positive predictive value for both (57% and 43% respectively). ICSCs may detect some infants at increased risk specifically of airway obstruction, but does not catch all infants at risk. However, the small sample size makes further generalizations difficult.
Schutzman et al. performed a retrospective analysis of preterm infants comparing continuous PSG for 12 to 24 h within 48 h of discharge to a 60 min ICSC. Of the 313 infants who had PSG and an ICSC, 56.5% failed the PSG, defined as apnea >20 s, bradycardia <80 for 3 to 5 s and desaturation <85% for 3 to 5 s, or evidence of reflux on pH probe. Of those who failed PSG, a large majority (89.3%) passed an ICSC, indicating that ICSCs are not as effective as PSG in identifying abnormal respiratory events before discharge. However, the PSG included all time spent supine, prone and feeding, unrelated to car seat position. The actual ability of the ICSC to identify safety while specifically in the semiupright car seat position was not addressed, so it continues to be difficult to identify what a 'failed' ICSC means.
For infants who fail an ICSC, some units proceed to a car bed. However, the evidence available on car beds indicates that positioning in the car bed does not reduce the number of significant desaturations when compared with car seats in either term or preterm infants. And with limited evidence on safety and increased burden on families, car beds do not seem to be an easy fix for infants who desaturate in the car seat position.
We have very little data on what failed ICSCs mean for our patients. However, when an infant fails, they receive further evaluation—whether that is repositioning and retesting, testing in a car bed or further clinical observation. Infants who pass are generally discharged without further investigation. How certain can we be that infants who pass an ICSC are safe for travel in this position? Degrazia attempted to answer this by performing repeated ICSCs on the same neonates and found that 90% who passed an initial ICSC went on to pass a second test. Davis et al. performed three repeated ICSCs on neonates and found that an initial passed test had a positive predictive value of 89% to predict two subsequent passes. However, both studies identified infants who passed an initial test and went on to fail a subsequent test indicating that even those infants who pass an ICSC are at risk for desaturations and therefore all infants need to be closely monitored. Infants with an unstable result (fail after initial pass) had significantly lower weights at the time of testing than those with repeated passes.
As described previously, Elder et al. compared PSG with ICSCs, and despite finding a low positive predictive value, they did find high negative predictive values for clinically observed ICSCs (91%), suggesting that a pass of the clinical test is reassuring.
Abstract and Introduction
Abstract
The infant car seat challenge (ICSC), or period of observation in a car safety seat before discharge to monitor for episodes of apnea, bradycardia and desaturation, is one of the most common tests performed on preterm neonates in the United States. However, the utility of the ICSC to identify infants at risk for adverse cardiopulmonary events in the car seat remains unclear. Minimal evidence exists to guide clinicians in performance of this test including appropriate inclusion criteria and failure criteria. In this article, the origins of the ICSC are discussed as well as potential etiologies of desaturations and bradycardia in the car seat position. Current literature on implementation, inclusion and failure criteria, incidence of failure and data on the meaning of a 'passed' vs 'failed' ICSC are discussed. Emphasis is made on minimizing time in car seats and seated devices given concern over the risk of desaturations.
Introduction
Research by the National Highway Traffic Safety Administration found that proper use of child safety seats reduces fatal motor vehicle injuries by 71% in children younger than 1 year of age. For that reason, the American Academy of Pediatrics (AAP) began recommending that all neonates and young children be transported in approved rear-facing semiupright car safety seats in 1974.
At the same time, advances in obstetrical management and neonatal care led to improvements in survival of preterm and critically ill neonates. Studies in the 1980s alerted providers that preterm neonates were at increased risk of apnea and drops in Hb-O2 saturation in arterial blood (desaturations) when placed in their car safety seats. Because of this, the AAP began recommending a period of observation in a car safety seat before discharge to monitor for episodes of apnea, desaturations and bradycardia for all infants born <37 weeks gestational age (GA), a test now referred to as the car seat test or infant car seat challenge (ICSC). In the United States alone, ~500 000 infants are born prematurely (<37 weeks GA) each year, making the ICSC potentially one of the most commonly performed tests on neonates. However, despite the widespread implementation of the test, very little information is available for clinicians to evaluate its usefulness, which is a major concern. The aim of the ICSC is to identify infants at risk of cardiopulmonary compromise in the semiupright position, but limited information exists on failure rates, appropriate vital sign parameters for a 'failure' and whether this test will identify infants who truly are at risk of adverse events. What a 'failed' ICSC means for the infant is still unclear, leading to much confusion and frustration. A small number of reviews have been published attempting to elucidate this controversial topic, although often uncovered more questions than answers. What we do know is that infants spend significant amounts of time in their car seats and other semiupright positioning devices after discharge. Parents surveyed about infant seating device practices reported that 94% of infants <5 months old spend >30 min in a seating device daily, averaging 5.7 h daily. Parents reported that 38% of infants spent >1 h specifically in a car seat daily, with 33% for 1 to 2 h and 5% for 3 to 4 h. Therefore, identifying cardiopulmonary safety of neonates in the seated position, such as in a car seat, is of the utmost importance.
The purpose of this review is to provide a comprehensive summary of existing literature on: (1) the origin of the ICSC recommendations, (2) current ICSC practices and (3) the evaluation of the ICSC as a tool to identify infants at risk of cardiopulmonary compromise in the semiupright position. Available data related to respiratory safety in the car seat and ICSCs are discussed.
Etiology of Desaturations in the Car Seat and the Origins of the ICSC
In 1985, Bull and Stroup were among the first to report that low birth weight (LBW) and prematurity pose unique dilemmas when it came to safe transportation of infants. At that time, federal standards did not specify minimum size or weight for which a car safety seat was deemed appropriate. They noted that a number of car seat models allowed neonates weighing <2.2 kg (5 lbs) to slouch, had shoulder straps that fit too loosely and had lap pads that covered the neck and face. They recommended various techniques including supportive blanket rolls to minimize this risk, but concern remained about the fit and safety of preterm infants.
Based on these findings, Willett et al. hypothesized that premature infants may be at increased risk of cardiopulmonary compromise in the car seat. In an article published in 1986, they compared 20 preterm infants (12 with a history of apnea of prematurity and concurrently on methylxanthine treatment and 8 with no history of apnea or treatment) to 10 healthy term infant controls while in the car seat position. They evaluated oxygen levels: (1) at baseline, (2) during a 30 min period in the car seat and then (3) for 30 min afterwards while placed prone. They noted the preterm infants had significant decreases in Hb-O2 saturation during the car seat interval, spent longer times saturating <85% and had more frequent Hb-O2 saturation drops to <80% than the term group. Time spent <90% and <85% was longest in preterm infants with a history of apnea, followed by preterm infants without a history of apnea, both statistically significant when compared with term controls. They did not note any episodes of apnea in either group.
Subsequently, Willett et al. hypothesized that preterm infants in the upright car seat position would be at increased risk of obstructive apnea and/or decreased minute ventilation due to decreasing tidal volume, and that this may result in hypoxia and bradycardia. Sixty-two infants admitted to the neonatal intensive care unit (NICU), both term and preterm, underwent multichannel recordings and 50 had complete pulmonary function tests. They noted that infants spent significantly more time with Hb-O2 saturation <85% and had more short apneas (lack of airflow and respiratory effort for 5 to 20 s) during the car seat interval than while supine. They found that none of the term infants (n=9) had an episode of bradycardia, desaturation and/or apnea that required intervention during the car seat interval but that 30% of the preterm infants did (n=16 of 53). However, they found that pulmonary function seemed to improve while in the car seat, with stable tidal volumes, increased compliance and decreased pulmonary resistance overall in the cohort. When performing a subgroup analysis comparing <32 weeks GA, 32 to 36 weeks and >36 weeks, the two most preterm groups were noted to have improved resistance, compliance and work of breathing while in the car seat despite abnormal changes on the multichannel recordings. They were unable to uncover pulmonary function abnormalities to account for the vital sign changes while in the semiupright position. They concluded that the cause of respiratory compromise in the car seat is multifactorial and may be different for each infant.
Similarly in 1989, Carlo et al. reported that former preterm infants corrected to 8 to 10 months of age had decreased pulmonary resistance, increased lung compliance and no change in minute ventilation in the semiupright position compared with supine. They did note that neck flexion significantly increased pulmonary resistance and decreased minute ventilation. Questions remained regarding the cause of cardiorespiratory compromise while in the semiupright position as it was not easily explained by pulmonary mechanics.
An initial pilot study of a car seat insert that allowed the head and neck to lie in neutral position demonstrated maintenance of a larger upper airway space radiographically and significant reduction in the frequency of desaturation events, suggesting that neck flexion is a major contributor to abnormal respiratory status in the car seat. A larger randomized controlled trial of this car seat insert showed a reduced rate of obstructive respiratory events and severity of Hb-O2 desaturation events (<85%) in healthy full-term infants, but did not affect the overall rate of moderate desaturations. Similar studies of inserts that prevent lateral flexion have shown no improvement in oxygenation. Neck flexion alone therefore does not explain desaturation events in all neonates, although it may contribute to events.
Hertz et al. performed polysomnography (PSG) on 28 preterm infants in the car seat and found that during car seat time, the presence of at least one abnormal breathing event (apnea, desaturation, bradycardia or periodic breathing) was significantly higher than while prone, and that periodic breathing was more frequent while in the car seat. However, they noted that periodic breathing had a significant sleep-state effect and was more common during active sleep, regardless of position of the infant. Active sleep was found to be significantly higher while in the car seat position. The authors concluded that it may be the increase in active sleep while in the car seat, rather than the actual position, that may account for periodic breathing and perhaps other respiratory abnormalities.
The exact cause of desaturations while in the semiupright position is still unclear. Clearly, desaturation events in the car seat are multifactorial and preventing and treating these events continues to be complicated. Immature respiratory control and prolonged apnea of prematurity, low tone and risk of improper neck flexion and upper airway position, and sleep state-associated effects all likely have a role to varying degrees.
AAP Recommendations for Testing: Evolution of the ICSC
Because of these concerns, in 1991 the AAP began recommending a period of observation in a car safety seat before discharge to monitor for episodes of apnea, desaturations and bradycardia for all preterm infants born <37 weeks GA, now known as the ICSC. For those infants who had evidence of Hb-O2 desaturations or bradycardia during an ICSC, the initial recommendation was that they 'travel prone in an alternative seating device'. They stated that a supine position could be considered but was noted to be concerning due to 'potential for increased risk of aspiration'; however, with information regarding SIDS (sudden infant death syndrome) and back-to-sleep, prone positioning is now discouraged even for travel. It was also noted that parents should be counseled to avoid infant swings, infant carriers or other equipment that places the infant in the semiupright position. In 1996, the recommendation for a period of car seat observation was reiterated, and hospitals were encouraged to 'develop policies to include this evaluation in their discharge planning process'. Infants who 'failed' this observation period were recommended at that time to travel in a supine or prone position in an alternative safety device, and again minimizing travel and other upright equipment was emphasized. Again, prone positioning has since been discouraged.
An attempt was made by the AAP to standardize car seat testing with the 2009 recommendation, suggesting performing ICSCs for a minimum of 90 to 120 min or the duration of the car ride home, whichever is longer. Additional recommendations were to consider inclusion of infants who were full term but possessed increased risk of cardiorespiratory compromise while in the semiupright position, such as those with neuromuscular disorders, congenital heart disease and micrognathia. However, the criteria for a 'failed' test were still left up to individual institutions and clinician groups as was the management of this group of infants. It was noted that 'if events documented on cardiorespiratory monitoring in a car safety seat are deemed significant by the treating physician or hospital policy, interventions to reduce the frequency of desaturation and episodes of apnea and bradycardia are recommended'. Suggestions for this management included supplemental oxygen, travel supine in a car bed after a similar period of observation or further medical management. It was advised that a repeat ICSC should be considered in those infants who failed initially before transitioning from a car bed to a semiupright car seat.
Duration of ICSC
After the early recommendations of the AAP to perform ICSCs, clinicians were left to decide the duration of the test. A 2003 survey of ICSC practices noted that 33% of NICUs were testing for 30 min, 24% for 60 min and 33% for 90 min. In 2007, Salhab et al. demonstrated that the proportion of infants whose first event occurred after 60 min of observation was 40%, and after 90 min was 30%, indicating that a shorter testing duration may miss clinically significant events. Current AAP guidelines recommend testing for 90 to 120 min or the duration of the car ride home, whichever is longer. This evidence also demonstrates the increased risk of Hb-O2 desaturation events with increased duration of time in the semiupright position, making it important to emphasize limiting time in this position.
Failure Criteria
There are no current AAP guidelines for what level of desaturation should constitute a 'failed' ICSC. A 2013 survey of NICUs performing ICSCs demonstrated that the definition of Hb-O2 saturations constituting a failure ranged from <80% to <93%, and ranged from any drop up to >60 s. The most commonly used desaturation level was <90%, with 37% of NICUs using this level; the most common definition of a bradycardia was <80 b.p.m., used by 70% of NICUs. In 2010, Bass acknowledged that though the definition of bradycardia <80 b.p.m. is widely accepted, a safe lower Hb-O2 saturation level is not universally agreed upon and recommended <93% as a cutoff for failure based on a review of normal Hb-O2 saturation ranges in newborns. More recent prospective studies have had a variety of Hb-O2 saturation failure cutoffs, generally between 85 and 90%, with 90% remaining the most commonly used cutoff in published literature. The wide range in the literature reflects the wide ranges used in clinical practice, making many studies difficult to generalize.
Standardization of failure criteria has remained challenging despite data on oxygenation ranges in healthy former preterm infants. Multiple studies have evaluated overnight recordings of arterial oxygen to obtain normative baseline data in otherwise healthy preterm infants at the time of discharge as well as in full-term infants. Median baseline Hb-O2 saturations were 98 to 99% in both term and preterm infants at the time of discharge (corrected to a mean GA of 36 to 37 weeks depending on the study), although the preterm infants had a wider range of baseline saturations than did full-term infants (88% to 100% vs 97% to 100%). Median Hb-O2 saturations increased in a preterm group to 100% at 6 weeks after discharge (range 95 to 100%). Despite improvement with time, the lower range of Hb-O2 saturations remains below those born at term. Without longer-term studies on those preterm infants with clinically unrecognized low baseline Hb-O2 saturations, it is difficult to determine clinical significance. Although preterm infants with a history of bronchopulmonary dysplasia whose baseline Hb-O2 saturations with sleep are persistently <92% have poorer growth velocities than do those with saturations >92%.
Episodes of Hb-O2 desaturations (intermittent hypoxia, IH) were frequently found on recorded oximetry in otherwise healthy preterm infants once deemed ready for discharge, but were often not detected clinically. When analyzing the CHIME data, a 2011 study demonstrated that 79% of former preterm infants (once corrected to a minimum of 36 weeks) and 65% of term cohorts had ≥1 IH episode (defined as a drop in Hb-O2 saturation <90%). Interestingly, there were no differences in the frequency or severity of IH between symptomatic neonates (noted to have at least one episode of apnea or bradycardia documented in nursing note during the 5 days before discharge) and those who were symptom-free. The median number of IH episodes in the preterm cohort was 4 with a wide range of 1 to 98, indicating that IH is commonly seen in this population. It is possible that the ICSC is capturing these IH events, which are commonly occurring anyway, as it is a time of intensive monitoring.
However, even if low baseline arterial saturations and IH events occur commonly in otherwise healthy preterm infants once corrected to term and ready for discharge, the potential influence of such hypoxemia on clinical and developmental outcomes remains uncertain. Samuels et al. studied 91 former preterm infants who had an apparent life-threatening event (ALTE) and then underwent 8 to 12 h recordings of breathing movements, electrocardiograms and arterial oxygen saturation and compared these with 110 healthy preterm infants made at 6 weeks after discharge. Twenty-one percent (n=19) of those infants experiencing an ALTE had hypoxic episodes, almost 7% (n=6) had low baseline Hb-O2 saturation (<95%), and 19% (n=17) had both. Thirty-one of 33 infants subsequently treated with supplemental oxygen had significant reduction or cessation of ALTE events, leading the authors to conclude that early recognition and treatment of abnormalities in episodic or baseline hypoxemia may reduce risk of further ALTE in former preterm infants. Longer-term follow-up of infants after discharge in the CHIME study demonstrated that those who had 5 or more cardiorespiratory events (apnea and/or bradycardia) had significantly lower Bayley-II Mental Developmental Index scores than those who did not have events. And, a large systematic review of literature related to chronic or IH in children demonstrated adverse impact on development, cognitive performance and behavior. This seems to indicate that early identification and treatment of hypoxemia is necessary to prevent adverse cardiopulmonary outcomes, which is one of the goals of the ICSC. However, it remains unclear whether the ICSC achieves this goal.
One concern when attempting to standardize ICSCs is the variation in monitors and averaging times. The number and percentage of desaturations identified depends on the duration and the averaging time set on the monitor. Longer averaging times are often used clinically because they have lower likelihood of false alarms related to motion artifact common in neonates. However, these longer averaging times may underestimate the frequency of brief desaturations. Conversely, shorter averaging times increase the rate of false alarms, but are more sensitive to detect brief intermittent desaturations. Recent prospective studies use 2-s averaging times to evaluate brief desaturation events, but national standards for averaging times for clinical use do not exist. Vagedes et al. found that 96% of Hb-O2 desaturations <80% in healthy preterm infants were brief, lasting <20 s, demonstrating the importance of averaging time on evaluating desaturation events.
It remains up to each NICU clinician group to determine failure criteria for their unit. The Canadian Paediatric Society Statement acknowledges this lack of a clear definition of a significant cardiopulmonary abnormality, but empirically defined it as 'two or more episodes of oxygen desaturation <88% for 10 s or more during a 90 min period of observation.'
Implementation
ICSC implementation in the United States was evaluated in a 2013 survey of Level II and III NICUs in New England and New York, noting that 11% did not perform any form of an ICSC. Of those who did perform ICSCs, 17% did not test all born <37 weeks GA and used more premature birth GAs as inclusion criteria. Although there are no specific recommendations about including LBW infants in testing, 45% of NICUs surveyed did test infants regardless of birth GA if they were LBW (<2.5 kg). Only 55% had tests that lasted a minimum duration of 90 min, recommended by the most updated AAP policy.
Incidence and Predictors of Failure
A number of studies have attempted to evaluate incidence and predictors of failure to limit the scope of the testing to the most at-risk infants. Failure rates in the literature range widely from 4 to 30%, but studies have been limited by small sample size and lack of consensus on appropriate failure definitions. Predischarge weight <2000 g has been linked to increased risk of failure. And infants with a history of apnea of prematurity also appear to be at increased risk.
The largest study evaluating incidence of failure retrospectively reviewed data on >1100 infants born preterm and found a 4.3% overall failure rate in this population, with failure criteria ranging from Hb-O2 saturations <88% to <90% at the participating hospitals. The only significant risk factor was, contrary to the authors' hypothesis, advancing birth GA (odds ratio for failure was 1.07 for each additional day of GA). On subgroup analysis, the failure rate for early preterms (<34 weeks birth GA) was 2.4%, and for late preterms (34 to 36+6/7 weeks) was 5.6% despite being tested at approximately the same corrected GA. The authors hypothesized that the early preterm infants were tested at chronologically older ages and were therefore more mature than the late preterm infants, but the etiology and ramifications of these findings are still unclear. Although this indicates that late preterm infants are still at risk of abnormal respiratory patterns in the car seat.
ICSC and Late Preterms
The care of late preterm infants remains controversial, as this group rarely requires as much support as early preterms, but are still premature and at risk of immature respiratory control. As noted earlier, a number of NICU protocols limit testing to early preterm infants. However, there is evidence that late preterms may have higher rates of desaturation in the car seat position than early preterm infants. Merchant et al. noted poor fit in the car seat in 24% of late preterm infants and 4% of term infants (who of note were LBW). Mean Hb-O2 saturations declined for both late preterm and term infants during the 90 min study, but 12% of the late preterm infants vs none of the term infants had significant apnea or bradycardia while in the car seat.
ICSC in Term Infants
The most recent AAP guidelines note that many hospital protocols also include ICSCs for infants other than those born preterm, who may also be at risk for cardiopulmonary compromise in the semiupright car seat position, including infants with congenital heart disease, hypotonia and micrognathia. In fact, up to 45% of NICUs surveyed include LBW as an inclusion criterion for testing, even in term neonates. However, data are limited to suggest which, if any, term infants would benefit from ICSCs.
In 1995, Bass et al. evaluated term infants who 'in the judgment of their pediatrician were felt to be at risk for O2 desaturations or apnea' for reasons such as: LBW, respiratory symptoms, genetic disorders and history of apnea. They found 28.6% had desaturations <90% during the test. Simsic et al. evaluated ICSCs for infants after cardiac surgery and found that 4 of 66 infants (6%) failed (defined as drop >15% from baseline Hb-O2 saturation), all of whom were full term. However, the authors acknowledged that these were highly selected populations, all with medical issues that prompted the testing.
A cross-over study of healthy full-term infants (38 to 41 weeks at birth) in both the car seat and car bed tested 150 infants for 60 min in each device (to limit time away from families) and a subgroup of 50 infants for 120 min to evaluate the hypothesis that respiratory compromise does occur in these devices. They found that the percent of total time with Hb-O2 saturation <95% during the 60 min observation was significantly higher (23.9%) while in the car seat than in the car bed or while supine in a crib (17.2% and 6.5%, respectively). Fifteen percent spent >50% of their time with Hb-O2 saturations <95% in the car seat (majority 90 to 94%). They noted the results were almost identical for the 120 min group. They also noted that regardless of order of device tested, infants saturated better during the first hour, indicating that limiting time in either device should be recommended. A similar study evaluating at Hb-O2 saturations in term infants while in a car seat for 30 min demonstrated mean Hb-O2 saturations were significantly lower when in a 45° car seat position (95.8%) when compared with supine in bed (98.7%) or a car bed (98.8%). They did note moderate desaturations (<90%) in 27% while in the car seat and none while in the crib, although this was a small study of only 15 subjects. However, the clinical significance of these desaturations noted in healthy full-term neonates remains unclear.
What Does a 'Fail' Mean?
A Cochrane Review from 2006 attempted to assess whether predischarge ICSCs prevent morbidity and mortality in preterm infants, but found no trials for inclusion and concluded that it is unclear whether the ICSC is beneficial or harmful to patients. There is very little data on what a failed ICSC means for our patients.
A few groups have performed PSG during ICSCs to evaluate the ability of the test to identify infants at risk of abnormal breathing patterns. Elder et al. studied 18 former preterm infants within 1 to 2 days of discharge during sleep in the car seat and while supine. A failed ICSC was defined by Hb-O2 saturation <90% with distress, apnea >20 s or any respiratory distress, whereas a failure of the PSG occurred with recurrent desaturations <90% over 5 min associated with obstruction, or apnea. They found that a clinically observed ICSC was more sensitive in predicting infants with increased obstruction in the car seat (sensitivity 80%) than failure of the PSG (sensitivity 60%), but had poor reliability and a low positive predictive value for both (57% and 43% respectively). ICSCs may detect some infants at increased risk specifically of airway obstruction, but does not catch all infants at risk. However, the small sample size makes further generalizations difficult.
Schutzman et al. performed a retrospective analysis of preterm infants comparing continuous PSG for 12 to 24 h within 48 h of discharge to a 60 min ICSC. Of the 313 infants who had PSG and an ICSC, 56.5% failed the PSG, defined as apnea >20 s, bradycardia <80 for 3 to 5 s and desaturation <85% for 3 to 5 s, or evidence of reflux on pH probe. Of those who failed PSG, a large majority (89.3%) passed an ICSC, indicating that ICSCs are not as effective as PSG in identifying abnormal respiratory events before discharge. However, the PSG included all time spent supine, prone and feeding, unrelated to car seat position. The actual ability of the ICSC to identify safety while specifically in the semiupright car seat position was not addressed, so it continues to be difficult to identify what a 'failed' ICSC means.
For infants who fail an ICSC, some units proceed to a car bed. However, the evidence available on car beds indicates that positioning in the car bed does not reduce the number of significant desaturations when compared with car seats in either term or preterm infants. And with limited evidence on safety and increased burden on families, car beds do not seem to be an easy fix for infants who desaturate in the car seat position.
What Does a 'Pass' Mean?
We have very little data on what failed ICSCs mean for our patients. However, when an infant fails, they receive further evaluation—whether that is repositioning and retesting, testing in a car bed or further clinical observation. Infants who pass are generally discharged without further investigation. How certain can we be that infants who pass an ICSC are safe for travel in this position? Degrazia attempted to answer this by performing repeated ICSCs on the same neonates and found that 90% who passed an initial ICSC went on to pass a second test. Davis et al. performed three repeated ICSCs on neonates and found that an initial passed test had a positive predictive value of 89% to predict two subsequent passes. However, both studies identified infants who passed an initial test and went on to fail a subsequent test indicating that even those infants who pass an ICSC are at risk for desaturations and therefore all infants need to be closely monitored. Infants with an unstable result (fail after initial pass) had significantly lower weights at the time of testing than those with repeated passes.
As described previously, Elder et al. compared PSG with ICSCs, and despite finding a low positive predictive value, they did find high negative predictive values for clinically observed ICSCs (91%), suggesting that a pass of the clinical test is reassuring.
Source...