Objective. To study the effect of two different delivery room (DR) policies on the rate of endotracheal intubation and mechanical ventilation (EI/MV) and short term morbidity in extremely low birth weight infants (ELBWI; <1000 g, ≥24 weeks).
Methods. Retrospective cohort study of 123 inborn ELBWIs born in 1994 and in 1996. DR policies have changed. Until 1994, ELBWIs were intubated immediately after delivery when presenting the slightest signs of respiratory distress or asphyxia after initial resuscitation using a face mask and a handbag. During 1995, the guidelines for respiratory support were changed. In 1996, continuous (15 to 20 seconds), pressure controlled (20 to 25 cm H2O) inflation of the lungs using a nasal pharyngeal tube, followed by continuous positive airway pressure (CPAP; 4 to 6 cm H2O) was applied to all ELBWIs immediately after delivery to establish a functional residual capacity and perhaps to avoid EI/MV.
In addition to the changes in respiratory support, the prevention of conductive and evaporative heat loss was improved in 1996. For analysis of morbidity and mortality, infants were matched for gestational age and birth weight.
Results. The rate of EI/MV in the DR decreased from 84% in 1994 to 40% in 1996. In 1996, 25% of the ELBWIs were never intubated (7% in 1994), but 35% of the ELBWIs needed secondary EI/MV, primarily because of respiratory distress syndrome (RDS).
Initial ventilator settings, ventilator days, mortality, and morbidity were not different between ELBWIs with EI/MV in the DR and infants with secondary EI/MV attributable to RDS in the intensive care unit. ELBWIs with no EI/MV that was caused by RDS had a lower morbidity (ie, bronchopulmonary dysplasia, intraventricular hemorrhage >grade 2 and/or periventricular leukomalacia), mortality, and fewer hospital days (mean: 79 vs 105 days). The incidence of gastrointestinal adverse effects like feeding intolerance or necrotizing enterocolitis was not increased in 1996.
Paco 2 was significantly higher at admission to the neonatal unit in ELBWIs with CPAP in 1996 (54 ± 15 mm Hg, 7.2 ± 2.0 kPa) compared with infants with EI/MV in 1994 (38 ± 11 mm Hg, 5.1 ± 1.5 kPa. A total of 26% of spontaneously breathing infants had hypercapnia (Paco 2 ≥60 mm Hg [8.0 kPa]), compared with 7% of infants with EI/MV in 1994. Within the first few hours of life, Paco 2decreased to 46 (32 to 57) mm Hg (6.1 [4.3 to 7.6] kPa) in never intubated ELBWIs (n = 17), but increased to 70 (57 to 81) mm Hg (9.3 [7.6 to 10.8] kPa) in ELBWIs (n= 14) with RDS and secondary EI/MV (age 5.5 [1 to 44] hours).
Conclusions. In our setting, the individualized intubation strategy in the DR restricted EI/MV to those ELBWIs who ultimately needed it, without increasing morbidity or mortality in infants with secondary EI/MV attributable to RDS. We speculate that an individualized intubation strategy of the ELBWI is superior to immediate intubation of all ELBWIs with slight signs of respiratory distress after birth.
- EI/MV =
- endotracheal intubation and mechanical ventilation •
- ELBWI =
- extremely low birth weight infant •
- DR =
- delivery room •
- CPAP =
- continuous positive airway pressure •
- RDS =
- respiratory distress syndrome •
- NICU =
- neonatal intensive care unit •
- Spo2 =
- pulse oximetry percent saturation •
- NEC =
- necrotizing enterocolitis •
- PDA =
- patent ductus arteriosus •
- IVH =
- intraventricular hemorrhage •
- BPD =
- bronchopulmonary dysplasia •
- ROP =
- retinopathy of prematurity
Criteria for endotracheal intubation and mechanical ventilation (EI/MV) of extremely low birth weight (ELBWIs) infants are not based on valid controlled trials1 and EI/MV is associated with an increased incidence of sepsis2 and pulmonary morbidity such as air leaks and chronic lung disease.3 Immediate respiratory support in the delivery room (DR) using continuous positive airway pressure (CPAP) has been reported to decrease the rate of EI/MV in LBWIs.4 ,5Morbidity or mortality of these infants was not reduced, but this may be different in ELBWIs.
Until 1994, the DR management of ELBWIs at the University of Ulm, Germany, included immediate EI/MV if slight signs of respiratory distress syndrome (RDS) were evident.6 In 1995, this management was replaced by an individualized intubation strategy; EI/MV was performed only when clear respiratory failure had been established to avoid overtreatment.
Sustained positive pressure inflation by face mask immediately after delivery has been shown to be an effective method to establish a functional residual capacity in term-asphyxiated neonates.7 ,8 This method of initial respiratory support was applied to ELBWIs. The established functional residual capacity was maintained by immediate subsequent application of nasal pharyngeal CPAP. This strategy was applied to all inborn ELBWIs in 1996.
The aim of this study was to investigate the effect of this new DR policy on the rate of EI/MV, mortality, and morbidity in ELBWIs. Therefore, we compared data of the ELBWI born in 1994, the last year of the immediate intubation policy, with data of 1996, the first year after implementation of the sustained positive pressure inflation and individualized intubation.
The neonatal intensive care unit (NICU) at the University of Ulm, Germany, is the only tertiary care center in an area managing ∼17 000 deliveries per year. Greater than 80% of all ELBWIs of this area are born in the hospital. All inborn ELBWIs with a weight of <1000 g and a gestational age of ≥24 weeks who were born in 1994 (n = 56) and 1996 (n = 67) received neonatal care and were included in this retrospective cohort study. The limit of 24 weeks of gestation (based on anamnestic and early sonographic data) was used, because generally, active obstetric treatment was not recommended on fetal indication below this age in our hospital. Thus, 5 ELBWIs (1994, n = 3; 1996,n = 2) with a gestational age of <24 weeks (all survived) were excluded.
The infants were grouped according to the type of respiratory support administered because of RDS. Diagnosis of RDS was established according to the following radiologic9 signs: group A, EI/MV in the DR; group B, CPAP in the DR but secondary EI/MV in the NICU attributable RDS; and group C, CPAP in the DR and no EI/MV caused by RDS.
Nine infants in group C developed pulmonary failure with EI/MV later on attributable to reasons other than RDS (ie, sepsis or central apnea) that probably were not influenced by the DR management. Because these ELBWIs had sufficient spontaneous breathing after DR treatment, they were analyzed together with the ELBWI with no EI/MV until discharge.
The clinical risk index for infants was evaluated for all ELBWIs to compare the initial neonatal risk of the different cohorts.10
Informed consent on DR management and all neonatal intensive care procedures was obtained routinely from the parents before birth.
The intervention was a change of the DR management. In 1994, initial respiratory support was as follows. After oropharyngeal and nasal suctioning, the lungs were inflated using a neonatal resuscitation bag via face mask, followed by EI/MV. In 1996, initial respiratory support was as follows. After suctioning, a pressure controlled (20 cm H2O) inflation of the lungs was sustained (15 seconds), using a nasal pharyngeal tube and a mechanical ventilator (Reanimat, Fa. Weyer, Hürten-Herwig, Germany). Mouth and other nostril were closed manually during inflation. This procedure was repeated with an increased pressure (25 cm H2O) if the heart rate remained <100 minute−1 and/or if the infant remained cyanotic. After initial inflation, a CPAP of 4 to 6 cm H2O was maintained. If necessary, intermittent mandatory ventilation (peak inspiratory pressure of 20 to 25 cm H2O and a rate of 60 minute−1) using the nasal pharyngeal tube was applied for several minutes until there was sufficient spontaneous ventilatory effort. EI/MV was performed if the heart rate did not increase to >100 minute−1, if pulse oxymetry percent saturation (Spo 2) remained <80% (N-200 Nellcor Puritan Bennett, Pleasanton, CA), or if the infant had apnea or showed marked and increasing dyspnea (Table 1.)
Additional Treatment and Monitoring in the DR, Transport to the NICU
To prevent postnatal hypothermia, a radiant warmer (Ohio Airco Inc, Madison, WI) and warmed blankets were used in 1994. In 1996, the ELBWIs were placed in a resuscitation unit, equipped with an additional spotlight, a radiant warmer (Variotherm REA-KC, Ceramotherm, Fa. Weyer, Hürten-Herwig, Germany), and a heated (38.5°C) silicone gel mattress. After drying with warmed blankets, infants were wrapped immediately including the head (with only the face uncovered), arms, and legs in a sterile, transparent plastic film to avoid conductive and evaporative heat loss. During all additional procedures this cover was kept closed as tightly as possible without restricting respiratory efforts.
The Spo 2 sensor (Oxisensor II N-25, Nellcor Puritan Bennett, Pleasanton, CA) was placed at the right wrist and covered against light with a foam rubber bandage.
In 1994, the ELBWIs were transported to the NICU without an intravenous line and placement of umbilical arterial lines occurred after admission to the NICU. In 1996, an infusion of 10% glucose was initiated in the DR via a peripheral venous line, and umbilical arterial lines were inserted under sterile conditions immediately after respiratory stabilization in ELBWIs with EI/MV (optional if without EI/MV). Once the arterial line was placed, bleeding was prevented by a tight purse string suture including all umbilical vessels. The catheter was fixed with five double knots, one opposite to the other. The saline-filled arterial line was closed with a screw top during transport. The position was corrected according to the radiograph after admission at the NICU.
A preheated (38°C), nonhumidified transport incubator (Transport-Inkubator 5400, Fa. Dräger, Lübeck, Germany) was used for the transfer to the NICU. CPAP was maintained during transport with a mechanical ventilator (Babylog 1HF or 2000, Fa. Dräger, Lübeck, Germany). Heart rate and Spo 2were monitored (Propaq Encore, Protocol Systems, Inc, Beaverton OR) during transport.
In 1994 and in 1996, all deliveries were attended by one neonatologist at the level of an attending and one fellow or resident. Additional assistance was provided by a midwife if necessary. An open nasogastric tube was inserted at the NICU to prevent significant gastric distension caused by pharyngeal CPAP.
The primary outcome criterion was the rate of EI/MV until discharge. Secondary outcome criteria were survival and morbidity, such as necrotizing enterocolitis (NEC; >Bell stage IIB),11patent ductus arteriosus, treated with indomethacin or surgical ligation, air leak, ventilator days, intraventricular hemorrhage (IVH) grades 3 and 412/periventricular leukomalacia,13 bronchopulmonary dysplasia (BPD), defined as Fio 2 >.21 at 36 weeks' postmenstrual age, retinopathy of prematurity (ROP) >stage 214 and hospital days. Blood gas data and initial ventilator settings were compared in relation to the different DR strategies as well.
A total of 34 items were collected from the patient records. Data are given as mean ± 1 SD or as median and range. Data were analyzed using the Fisher's exact two-tailed test and the Mann-WhitneyU test (Student Systat Version 1.0; Systat, Inc, Evanston, IL). For comparison of the groups A 1994 and A, B, and C 1996, the potential bias of gestational age and birth weight was eliminated by the analysis of matched pairs: A 1994 and A 1996, 22 pairs; A 1994 and B 1996, 10 pairs; and A 1994 and C 1996, 19 pairs. The level of significance was set at P < .05. Bonferroni correction for multiple testing was not used for the explorative statistics.
Prenatal betamethasone dosage was increased from 8 to 12 mg during 1996.15 This was the only apparent systematic change in obstetric management.
Conventional and high frequency oscillatory ventilation were performed according to the protocol of a multicenter, randomized trial, comparing these two ventilatory strategies.16 This protocol was applied to all ELBWIs born in 1994 and in 1996, the enrollment for this trial was started in 1995. Criteria for surfactant therapy16 did not include prophylaxis in the DR. Extubation criteria were not changed during the study period.
The principle of hyperalimentation (500 kJ/kg body weight) was used in both cohorts. In 1994, oral feeding was started with breast milk or formula. In 1996, breast milk or formula was diluted with equal parts of sterile water up to a daily intake of 100 mL/kg body weight. Parenteral nutrition was not changed during the study period.
There were no other systematic changes in neonatal management during this period.
Rate of EI/MV
In 1994, 47 of the infants (84%) underwent EI/MV in the DR, compared with 27 infants (40%) in 1996 (P < .001). A total of 4 ELBWIs (7%) were never intubated in 1994 (gestational age, 29.7 [27.4 to 32] weeks; birth weight, 915 [870 to 980] g) compared with 17 infants (25%) in 1996 (P < .01).
RDS ≥grade 2 was diagnosed in 36% of all ELBWIs in 1994 and 40% of all ELBWIs in 1996. In 1994, RDS (n = 4) and sepsis (n = 1) indicated secondary EI/MV in 5 of the 9 ELBWIs with no EI/MV in the DR. Data of the groups B 1994 and C 1994 were not analyzed further because of the small number. In 1996, RDS indicated secondary EI/MV (age 5.5 [1 to 44] hours) in 14 infants from group B, but other reasons for EI/MV were found in 9 infants with an initially uncompromised lung function (subgroup C 1996), including apnea (n = 6), sepsis (n = 2), and pulmonary hemorrhage (n = 1). The age at the start of EI/MV in these 9 infants was 96 (21 to 340) hours. The respiratory support in relation to gestational age is presented in Fig 1.
Mortality and Morbidity
Mortality was not different in ELBWIs with EI/MV immediately after delivery, compared with infants who underwent secondary EI/MV in the NICU. In 1996, significantly fewer infants had an IVH >grade 2 and/or BPD, and the duration of hospitalization was shorter.
None of the infants without any EI/MV until discharge died, had IVH >grade 2, periventricular leukomalacia, BPD, or ROP >grade 2. The duration of hospitalization for these infants was significantly shorter than that in all other groups. Detailed data on mortality and morbidity are presented in Tables 2A and B.
The first arterial blood gas examination was obtained within 1 hour after admission to the NICU. In 1996, the Paco 2 was higher and the pH was lower in ELBWI with CPAP, compared with infants with EI/MV in 1994 and in 1996. Hypocapnia (Paco 2 <30 mm Hg, [4.0 kPa]) at time of first blood gas examination was present in 13 ELBWIs in 1994 (26 [21 to 29] mm Hg; 3.5 [2.8 to 3.9]) and in 8 ELBWIs in 1996 (24.5 [17 to 27]; 3.3 [2.3 to 3.6]). All of them had EI/MV in the DR. The incidence of hypoxemia (Spo 2 <80%) was not different at the time of admission (1994, n = 8; 1996, n = 10). In group C 1996, the initially high Paco 2 decreased within a few hours. During the first day of life, the Paco 2 was 48 (32 to 57) mm Hg (6.4 [4.3 to 7.6] kPa) in group D 1996 and did not increase (47 [39 to 49] mm Hg, 6.3 [5.2 to 6.5] kPa) during the first week. In contrast to group C 1996, Paco 2 increased to 70 (57 to 81) mm Hg (9.3 [7.6 to 10.8] kPa) in the ELBWIs of group B 1996 before secondary EI/MV (mean age 5 hours). Fio 2 increased to 0.60 (0.5 to 1.0) and respiratory acidosis occurred with a pH of 7.16 (7.08 to 7.21).
The initial ventilator settings for the infants of the groups A 1994, A 1996, and B 1996 were not different. Detailed data are presented inTable 3.
Additional Treatment and Monitoring in the DR, Transport to the NICU
Umbilical arterial lines were placed successfully in 77% in 1994 and in 61% in 1996 of all ELBWIs (71% in group A 1996). No respiratory deterioration requiring EI/MV and no dislocations of the arterial line or umbilical bleedings occurred with the transfer of the ELBWIs into the transport incubator or during transport (∼5 minutes) to the NICU. Reliable data of Spo 2 within the first 10 minutes of life were obtained in all ELBWIs who were stabilized in the DR with CPAP (B 1996 and C 1996), in 80% of intubated ELBWIs (A 1996), and in all ELBWIs during transport.
Adverse Effects Possibly Associated With Sustained Pharyngeal Inflation and CPAP
None of the ELBWIs had any pulmonary air leak at time of admission to the NICU (chest radiography), and no infant developed pulmonary air leak while on CPAP.
We observed no gastric distension resulting in perforation. The incidence of NEC was similar in 1994 and in 1996. Of the 5 infants with NEC in 1996, 4 had EI/MV in the DR. The volume of enteral nutrition (breastmilk or formula) per kilogram of body weight at day 14 was not different in 1994 (78 ± 48 mL) and 1996 (92 ± 53 mL). Furthermore, there was no significant difference in enteral intake at day 14 among groups A, B, and C in 1996.
In 1996, there was a trend toward more infants having received 2 doses of betamethasone, and in 55% of these ELBWIs, the dosage was increased from 8 to 12 mg (A 1996, 54%; B 1996, 43%; C 1996, 62%). With the same indications for surgical delivery (multiple gestation, chorioamnionitis, premature rupture of membranes, oligoamnion or anhydramnion, and abnormal position), more infants were delivered vaginally in 1996 (13 vs 2 in 1994; P < .05). Incidences of oligoamnion/anhydramnion or fetal distress (abnormal cardiotocogram or umbilical Doppler pattern) did not differ significantly among the groups.
In 1994, more ELBWIs had high frequency oscillation (85% vs 50% in 1996; P < .001).
Surfactant was given in the DR to 26% of the ELBWIs in 1994 and 8% of the ELBWIs in 1996 with EI/MV. The total number of infants treated with surfactant (57%) and with dexamethasone (25%)17 was similar in 1994 and in 1996. Additional data on surfactant is presented in Table 3.
The demographic and admission data are shown in Tables 4A and B. Data of all ELBWIs were available for 12 of the 34 items, data were missing (<5% of ELBWIs) for 21 items, and umbilical arterial pH was not measured in 13 infants (11%). Three infants in 1994 (Goldenhar syndrome [matched pairs: secondary EI/MV attributable to RDS], cytochrome C oxidase [complex 4] deficiency [matched pairs: EI/MV in the DR], total anomalous pulmonary vein connection) and 1 infant in 1996 (male pseudohermaphroditism [matched pairs: group A 1996]) had major anomalies. The age at admission to the NICU was similar in 1994 and in 1996: median 25 (11 to 45) minutes. In 1996, fewer infants presented with hypothermia (body temperature <35°C) (Tables 4A and B). The incidence of multiple gestation was similar in 1994 (n = 11) and in 1996 (n = 15). In each year, there were 3 sets of twins with fetofetal transfusion syndrome. There was a trend toward more females in 1996, this was significant in the group of matched pairs, A 1994 versus C 1996. Infants with EI/MV in the DR had a lower arterial blood pressure at time of admission, compared with the spontaneously breathing ELBWIs. There was no significant difference in the incidence of hypoglycemia (<1.7 mmol/L) at the time of admission to the NICU in 1994 and in 1996.
The primary result of our study is that there were 25% of all ELBWIs who were never intubated until discharge. A total of 13 of these ELBWIs had a gestational age of <28 weeks and the majority (8/13) had no intrauterine growth retardation, accelerating lung maturation.
We observed a decrease in the rate of EI/MV in the DR from 84% to 40%, but none of the theoretic adverse effects of the sustained inflation and pharyngeal CPAP such as air leaks or impaired feeding tolerance.
The only randomized prospective study on this issue was reported by Drew in 1982.6 The infants in that study were more mature (>2 weeks) and had a mean birth weight of 1022 g. A total of 26% of the infants survived without EI/MV, but infants with secondary EI/MV had a significantly increased mortality (91%), compared with infants with EI/MV intubated immediately in the DR. Moreover, 26% of the infants who were not ventilated died, and Drew reported an increased incidence of air leaks, metabolic acidosis, and a higher ventilatory demand in infants with secondary EI/MV.
None of these adverse effects were observed in our infants. A total of 35% had secondary EI/MV for different reasons. Pulmonary function was not compromised in the DR in 9 of these infants. The secondary EI/MV at a median age of 5.5 hours in our ELBWIs with RDS (B 1996) was not associated with a greater demand for ventilatory support (increased pressure or Fio 2) or a higher cumulative dose of exogenous surfactant, compared with the infants who underwent EI/MV in the DR. There also was no difference regarding ventilator days, air leaks, BPD, and IVH (>grade 2). Mortality was 13% in ELBWIs with secondary EI/MV versus 33% in the infants with EI/MV in the DR.
How can our findings be explained? Intubation criteria in both studies were similar: Paco 2 >70 mm Hg, hypoxia with Fio 2 >.60, and RDS with severe recurrent apneas. Respiratory problems were the primary cause of death in Drew's study, but only 50% of the deaths were related directly to pulmonary morbidity in our ELBWIs. Morbidity may have changed because of improved prenatal care (ie, increased use and dosage of betamethasone). Improved neonatal monitoring (transcutaneous devices, pulse oximetry) probably allowed earlier detection of clinical deterioration in spontaneously breathing ELBWIs, and the general advance in neonatal intensive care during the last 14 years may have improved treatment, once secondary respiratory failure has been established (eg, surfactant and ventilation techniques).
Surfactant was given as treatment for RDS but not for prophylaxis, because this was not scheduled in the protocol of a multicenter trial on high frequency oscillatory ventilation at this time.16Whether prophylactic surfactant therapy would have changed our results in favor of immediate EI/MV in the DR is speculative and should be the subject of a prospective trial.
The skill of neonatal staff and the technical facilities should not be underestimated. With ∼150 inborn VLBWIs per year, the Ulm University NICU is covered 24 hours per day by experienced staff, and the neonatologist on duty is present within 10 minutes during the night or weekend. This is possibly an important factor, positively influencing our results. They are consistent with those of other studies reporting on low rates of EI/MV (30%, <1500 g,512%, ≤1000 g, and ≥24 weeks18) without increased mortality or morbidity in infants supported with CPAP.
We permitted higher levels of Paco 2 in the spontaneously breathing infants, compared with the ELBWIs with EI/MV in the DR. Paco 2 values at the age of 1 hour were higher in the infants with nasal pharyngeal CPAP, compared with the infants with EI/MV; however, the first Paco 2after admission was not useful to identify those infants who needed secondary EI/MV. In our infants (C 1996) who were not ventilated, the initially high Paco 2 decreased within a few hours to a lower range (<50 mm Hg, 6.7 kPa), but in infants with RDS and secondary EI/MV (B 1996), there was an additional increase of Paco 2 to 70 (57 to 81) mm Hg (9.3 [7.6 to 10.8] kPa) before intubation. We infer from our data that initial hypercapnia should be tolerated to differentiate ELBWIs not requiring EI/MV. Whether Paco 2 levels, as seen in our infants, are associated with an impaired outcome is unclear. The target range for Paco 2 in preterm infants is controversial and as yet, no prospective study has examined at this issue. The morbidity data of our ELBWIs who were not ventilated favor the hypothesis that a Paco 2 of even >60 mm Hg (8.0 kPa) in the first hours of life is not associated with an increased short term morbidity or mortality necessarily, at least in spontaneously breathing ELBWIs with nasal pharyngeal CPAP.
The individualized intubation strategy in the DR is a requirement to detect ELBWIs with no need for EI/MV. However, the increased number of ELBWIs who were not ventilated in 1996 may not result only from the change in the DR policy but also from the different cointerventions. More infants were treated with high frequency oscillation in 1994. However, the results of the multicenter high frequency oscillation trial16 as well as a subanalysis of the infants of our unit showed no significant influence of the ventilation mode on the outcome variables of this study. The higher dose of prenatal betamethasone, the trends toward more females and more ELBWIs with intrauterine growth retardation in 1996, and the improved thermal management may have contributed to the significant decrease of BPD and IVH without increased mortality.
In conclusion, with the current prenatal and postnatal therapy, there was an increasing number of ELBWIs in our population who presented only slight signs of RDS and did not need EI/MV immediately after birth or even until discharge. Neonatologists should be encouraged to differentiate these ELBWIs to avoid overtreatment if their facilities are adequate.
A DR intubation strategy that targets treatment according to the symptomatology of each infant seems to be effective. Such efforts avoid futile intensive care procedures and save funds. However, when using this strategy, one must be aware that a secondary EI/MV, if not performed in time when respiratory failure becomes evident, is a potential risk for increased mortality and morbidity. Thus, the skill of the staff in the NICU should match the skill in the DR.
The retrospective design of this study does not allow a claim that secondary EI/MV is not associated with any disadvantages. However, our results favor the hypothesis that an individualized intubation strategy is superior to immediate intubation of all ELBWIs after delivery. This hypothesis should be tested in a prospective randomized trial before this DR strategy is recommended generally.
We maintain the individualized intubation strategy in the DR and observe a continued low incidence of morbidity.
We thank B. Kapotic for helpful advice on the manuscript.
- Received June 12, 1998.
- Accepted October 29, 1998.
Reprint requests to (W.L.) Department of Pediatrics, University of Ulm, Prittwitzstr, 4389070 Ulm, Germany.
This work was presented in part by the European Society for Paediatric Research meeting, Szeged, September 1997; and inPediatric Research 1997; 42: 405, 123A.
- ↵Tyson J. Immediate care of the newborn infants. In: Sinclair J, Bracken M, eds. Effective Care of the Newborn Infant. Oxford, UK: Oxford University Press; 1992:21–29
- Avery ME,
- Tooley W,
- Keller J,
- et al.
- Poets CF,
- Sens B
- Committee for the Classification of Retinopathy of Prematurity
- Copyright © 1999 American Academy of Pediatrics