Variation and Temporal Trends in Delivery Room Management of Moderate and Late Preterm Infants

Vermont Oxford Network (VON) is a nonprofit, voluntary worldwide community of practice dedicated to improving the quality, safety, and value of care for newborns. This is a retrospective observational cohort study of prospectively collected data from the VON database of all NICU admissions.¹¹  The database includes: (1) VLBW (≤1500 g) infants who are admitted anywhere in a hospital or die anywhere in a hospital within 28 days of birth; and (2) infants >1500 g admitted to a NICU (defined as a location within a hospital where continuous positive airway pressure [CPAP] or intermittent mechanical ventilation can be provided, not including the DR or other infant stabilization location) within 28 days of birth, and (3) infants >1500 g who die anywhere in the hospital within 28 days of birth.¹²  Participating VON members submit data on eligible infants until death, discharge from the hospital, or transfer to another center. All data undergo automated checks for quality and completeness upon submission and are deidentified. The Institutional Review Board at the University of Vermont determined that use of the VON database for this study was not human subjects’ research.

Infants included in the study cohort were born between 30^0/6 and 36^6/7 weeks’ gestation from January 1, 2011 to December 31, 2020 at VON centers in the United States contributing to the all NICU admissions database. Infants with any congenital anomalies were excluded.

Interventions of interest included “DR interventions,” interventions performed during initial resuscitation in the DR or in a resuscitation area immediately following birth before NICU admission as defined in the VON Manual of Operations.¹²  Oxygen is the receipt of any supplemental oxygen (>21%). Positive pressure ventilation (PPV) is the receipt of any positive pressure ventilations via face mask. The provision of CPAP includes CPAP applied through the nose, inclusive of CPAP administered via a face mask that is covering the nose, without the administration of intermittent ventilations. Endotracheal tube (ETT) ventilation refers to any ventilation received through an ETT. Chest compressions include any external cardiac massage. Epinephrine includes the administration of any epinephrine via intravenous, intracardiac, or intratracheal routes. Any DR intervention includes the receipt of any of the above listed therapies during the initial resuscitation. We also assessed surfactant administered in the DR. We examined pneumothorax, given that exposure to ventilation, noninvasive or invasive,^13,14  may result in pneumothorax, which was defined as infants with extrapleural air on chest radiograph or who required needle thoracentesis during NICU admission. The data analyzed include all interventions provided to the infant in the DR, not just the highest level of intervention.

In the VON data, NICUs are categorized into 4 types based on surveys completed by members. Type A NICUs with restrictions on ventilation transfer infants to another center for assisted ventilation based on infant characteristics (ie, gestational age) or the duration of ventilation. Type A NICUs without ventilation restrictions do not perform 1 of 8 surgeries (omphalocele repair, ventriculoperitoneal shunt, tracheoesophageal fistula or esophageal atresia repair, bowel resection or reanastomosis, meningomyelocele repair, cardiac catheterization, or cardiac surgery requiring bypass). Type B NICUs perform at least 1 of the aforementioned 8 surgeries, except cardiac surgery requiring bypass. Type C NICUs perform cardiac surgery requiring bypass. These NICU types correspond to the current AAP neonatal levels of care, with type A NICUs with restrictions on ventilation corresponding to level II units, type A and B NICUs corresponding to level III units, and type C NICUs corresponding to level IV units.

Maternal and neonatal characteristics were examined for the entire cohort. Proportions of DR interventions were calculated at the patient- and hospital-level. These calculations were completed for the entire cohort, stratified by each week of completed gestation (30 to 36), and by moderate (30 to 33) and late (34 to 36) preterm subgroups. We used logistic regression models with linear time trends to test for changes in DR interventions over time at the patient level. Rates of DR interventions were compared by NICU type. Risk-adjusted hospital rates of any DR intervention were calculated using indirect standardization with a logistic regression model, accounting for infant gestational age, birth weight z-score, sex, multiple gestation, mode of delivery, 1-minute Apgar, and hospital location above 4000 feet elevation. The 1-minute Apgar score was included in the model as it reflects the infant’s baseline status before DR interventions. Analyses were completed using R version 4.0.2 (Vienna, Austria).

There were 616 110 eligible infants, with a median gestational age of 34 weeks and birth weight 2165 g. Maternal and infant characteristics are shown in Table 1. In this cohort, 308 (0.05%) infants died; of those 115 (37%) did not receive any DR interventions.

Tables 2 and 3 shows the frequency of DR interventions. In the overall cohort, 55.3% of infants received any DR intervention, which ranged from 89.7% to 44.2% among infants born at 30 and 36 weeks’ gestation, respectively. The most commonly provided intervention was supplemental oxygen (50.3%), followed by CPAP (36.2%), and PPV (24.1%). The frequency of each DR intervention decreased with increasing gestational age.

The frequency of DR interventions changed markedly over the study period (Fig 1). The largest change was in CPAP use, which increased significantly from 17.9% in 2011 to 47.8% in 2020 (P ≤.001). The provision of PPV also increased from 22.9% to 24.9% (P ≤.001). The rate of ETT ventilation decreased from 6.9% to 4.0% (P ≤.001), as did DR surfactant administration (3.5% to 1.3%, P ≤.001). Supplemental Fig 5 shows the frequency of DR interventions in the moderate and late preterm subgroups over time. The frequency of PPV, CPAP, and ETT ventilation were examined by gestational age (Fig 2). The rate of CPAP use increased for all gestational age groups, with a 2.2-fold increase (from 34.9% to 77.5%) among infants born at 30 weeks’ gestation and a 3.5-fold increase (10.4% to 36.8%) among those born at 36 weeks’ gestation. The provision of DR PPV increased in all gestational age groups, whereas DR ETT ventilation decreased across all gestational ages.

Pneumothorax was reported in 1.6% of infants and when stratified by gestational age had a U-shaped distribution (2.2% of infants born at 30 weeks, 1.2% of infants born at 33 weeks or 34 weeks, and 2.5% of infants born at 36 weeks). In the whole cohort, rates of pneumothorax decreased over the study period from 1.9% to 1.6% (P ≤.001).

During the study period, 483 VON centers contributed data to the all NICU admissions database. Of participating centers in the cohort, 17.2% (n = 83) were type A with ventilation restrictions, 38.3% (n = 185) were type A without ventilation restrictions, 32.5% (n = 157) were type B, and 12.0% (n = 58) were type C. Maternal and infant characteristics by NICU type are shown in Supplemental Table 3.

The frequencies of hospital-level DR interventions are shown in Fig 3. Median rates of any DR intervention were 54% (interquartile range [IQR] 48%–62%). The largest variation in hospital-level practices, illustrated by the widest IQR, was seen in DR CPAP (median rate: 36% [IQR 23%–46%]). The frequency of DR interventions at the hospital-level among moderate and late preterm infants is shown in Supplemental Fig 6. Patterns in usage by gestational age at the hospital-level mirrored those at the infant-level (Supplemental Table 5). Risk-adjusted rates of any DR intervention did not differ by NICU type for the cohort overall (P = .07) (Fig 4), nor when stratified by moderate (P = .45) and late (P = .05) preterm subgroups (Supplemental Fig 7).

In this large, contemporary observational study, we found that the majority of MLP infants receive interventions in the DR, which, as expected, decreases with increasing gestational age. We report a significant increase in DR CPAP utilization for all gestational age groups between 2011 and 2020. Although rates of DR CPAP were highest in the most immature infants in the study, the increase over time was largest in the most mature infants. Hospital-level analyses demonstrate variation in DR respiratory management, particularly DR CPAP, of MLP infants. The varied DR management of MLP infants illustrates differences in the interpretation and application of resuscitation guidelines as well as opportunities to improve care by reducing variation.

Prior studies of respiratory management of MLP infants often focused on the support provided after NICU admission and report decreases in the use of invasive ventilation and increases in noninvasive ventilation among infants born before 34 weeks’ gestation.^15,16  Although the use of invasive ventilation has decreased over time, marked variation remains across hospitals, which parallels our findings.¹⁵  However, the DR is a distinct environment from the NICU, where care is focused on supporting postnatal transition. The use of DR practices may be influenced by providers’ experience, the number of team members present, and local protocols. Although resuscitation guidelines are meant to be universal, limited guidelines address the nuances of MLP infant DR care, likely contributing to the variation observed in this study and highlighting an opportunity for future improvement.

The significant increase in DR CPAP utilization, with no associated increase in pneumothorax during the NICU admission, is a notable finding. During the study period, there were 2 publications that could have influenced DR management. First, is the release of the seventh edition of the NRP Textbook of Neonatal Resuscitation. The major changes in the seventh edition were: recommending delayed cord clamping for stable infants, intubating infants before the initiation of chest compressions, and increased emphasis on the use of electrocardiography during PPV and cardiopulmonary resuscitation.¹⁷  It is unlikely that revisions to the seventh edition of NRP contributed to the observed temporal trends, as they did not introduce changes to initiating CPAP or PPV in the resuscitation algorithm.

The second noteworthy publication was the 2014 AAP Committee of Fetus and Newborn policy statement regarding the respiratory support of preterm infants at birth.¹⁸  Unlike prior statements which supported the early administration of surfactant, historically via an ETT, the new statement recommended early use of CPAP followed by selective surfactant administration. The studies and associated meta-analysis informing these recommendations were conducted in infants <30 weeks’ gestation with a primary outcome of death or bronchopulmonary dysplasia.^19–22  Practice changes following this statement largely targeted patients at highest risk of death or bronchopulmonary dysplasia, extremely preterm and VLBW infants. Although the Committee of Fetus and Newborn statement used the word preterm throughout, it did not differentiate between extremely, very, moderate, and late preterm populations. It is plausible that promotion of DR CPAP among the most preterm patients influenced DR management of MLP infants. The perception that 17.3% of MLP infants warranted DR CPAP in 2011 compared with 47.3% of MLP infants in 2020 suggests practice changes in the aggressiveness of DR interventions. We speculate this may result from providers applying evidence-based interventions for 1 population (eg, extremely preterm infants) to a different group of patients (eg, MLP infants). This suggests an ongoing need for education surrounding the evidence and generalizability of DR studies and associated recommendations to different infant populations.

In this MLP population, we report a median frequency of DR intervention of 54% at the hospital-level. Unfortunately, contextualizing this finding within published estimates of DR interventions is challenging given variation in the DR interventions reported and limited generalizability of study populations. To date, many estimates surrounding the frequency of DR interventions in term populations reflect data from lower resource settings^23,24  or focus exclusively on a specific intervention, such as PPV, chest compressions, or epinephrine administration.^25,26  Alternatively, there are a subset of studies focused solely on extremely preterm infants, DR interventions, and outcomes.^27,28  Some studies have stratified findings by gestational age, including a study across 4 British hospitals of infants 32 to 33 and 34 to 36 weeks’ gestation reporting 37% and 15% received DR resuscitation (PPV or intubation), respectively,²⁹  and a population-based study of all infants born in a single Norwegian hospital that reported 77% of infants 28 to 33 weeks’ gestation and 10% of infants 34 to 36 weeks’ gestation received DR interventions (inclusive of CPAP only, PPV, intubation, chest compressions, administration of epinephrine, or a fluid bolus).³⁰  Similarly, a US study from the Neonatal Research Network reported that 76% of infants 29 to 33 weeks’ gestation received some DR intervention beyond drying and stimulation.⁵  These gestational age ranges do not parallel our study, making comparisons challenging. Notably, in the Norwegian study, the provision of CPAP only among infants 34 to 36 weeks’ gestation was 4.6% compared with the 25.8% and 31% reported in infants 34 and 36 weeks’ gestation in the current US-based cohort of infants admitted to the NICU. Although differences exist between neonatal resuscitation algorithms used in Europe and the United States, both are based on the International Liaison Committee on Resuscitation treatment recommendations. The magnitude of the difference in CPAP use likely reflects hospital-level variation and differences in resuscitation practices and interpretation of guidelines; this variation is likely not limited to international comparisons.

Given hospital variation in DR interventions, we also examined differences by NICU type (ie, level of care). Although studies of extremely preterm and VLBW infants support the delivery of the highest risk infants in hospitals with high-level neonatal care,^31,32  data regarding the optimal location for the delivery and management of MLP infants are limited. Thus, MLP infants may be born in hospitals with a wide range of neonatal capabilities. In our adjusted analysis we found no difference in the median rates of DR interventions provided by NICU type. This finding suggests that the increase in DR interventions, like CPAP, and associated temporal trends, are not driven by 1 type of hospital and reflect a widespread culture change.

This study has limitations. First, the study cohort does not include all VON members as only a portion of VON units contribute to the all NICU admissions database. Second, this sample does not include infants who may have received DR interventions and were not admitted to the NICU. Infants ≤34 weeks are regularly admitted to a NICU and thus estimates of DR interventions in this subpopulation are likely accurate. However, there is significant between-unit variation in NICU admission criteria by gestational age and birth weight. This variation will affect estimates of DR intervention frequency for infants born at 35 and 36 weeks’ gestation who may not be routinely admitted to the NICU, which may lead to overestimation (eg, if DR interventions increase the likelihood of NICU admission) or underestimation (eg, if infants receive DR interventions and are not admitted to the NICU) of DR interventions. Whereas admission thresholds vary by hospital, changes within hospitals are unlikely and as a longitudinal study, each hospital can serve as its own control. Third, these data do not include the use of a supraglottic airway as that item was added to the VON database in 2018. However, this is not a commonly used rescue airway in the MLP population and is unlikely to have a large effect on DR intervention estimates. Fourth, the route of surfactant administration continues to shift from invasive approaches via an ETT to minimally invasive approaches. Although we reported a decrease in surfactant administration in this study, the route of surfactant is not available in the dataset. Fifth, the timing of pneumothorax is not collected in the VON data. Though prior studies using similar definitions of pneumothorax during NICU admission have demonstrated associations between DR CPAP and CPAP among infants ≥35 and ≥32 weeks’ gestation, respectively.^13,14  Finally, these data do not differentiate between “necessary” and “unnecessary” DR interventions. Thus, the clinical indicators driving the increased provision of DR interventions, such as CPAP, are unknown.

This study also has a number of strengths. First, these data reflect the care provided to more than 615 000 infants across 483 hospitals, making these findings more generalizable than prior studies. Second, centers in this study were included independent of their approach to DR interventions. Third, the granularity of DR data available in VON (interventions provided and patient characteristics) provide more specific estimates of DR intervention frequency and allow for assessment of patient- and hospital-level variation. Fourth, this is a dedicated study of MLP infants and demonstrates the high proportion of MLP infants who receive DR intervention, which may inform discussions and decisions surrounding provider attendance at MLP infant deliveries.

In this large, observational cohort study, we found marked variation in the provision of DR interventions to the MLP population at the individual- and hospital-level. We report a greater than 2.5-fold increase in DR CPAP use during the study period, without an associated change in pneumothorax rates. These findings increase our understanding of DR care for MLP infants and highlight marked variation in DR interventions and associated interpretation of resuscitation guidelines, highlighting opportunities to study, optimize, and improve DR care for this at-risk population.

We thank our colleagues who submit data to the Vermont Oxford Network on behalf of infants and their families. The list of centers contributing data to this study are listed in Supplemental Table 5.

DR

delivery room

NICU

neonatal intensive care unit

PPV

positive pressure ventilation

CPAP

continuous airway pressure

ETT

endotracheal tube