A Kribs, Department of Neonatology, Children’s Hospital, University of Cologne, Kerpener Str. 62, D-50931 Cologne, Germany. Tel: +49-221-4783578 | Fax: +49-221-4786451 | Email: firstname.lastname@example.org
Aim: To evaluate the outcome of a cohort of extremely low gestational age newborn infants (ELGAN) below 26-week gestation who were treated following a revised, gentle delivery room protocol to assist them in the transition and adaptation to extrauterine life.
Methods: A cohort of infants with a gestational age (GA) below 26 weeks (study group; n = 164) was treated according to a revised delivery room protocol. The protocol included an optimized prenatal management, strict use of continuous positive airway pressure (CPAP), avoiding mechanical ventilation and early administration of surfactant without intubation. The parameters management of respiratory distress syndrome, survival, neonatal morbidity and neurodevelopmental outcome were compared with a historical control group (n = 44).
Results: Seventy-four per cent of the study group infants were initially treated with CPAP and surfactant administration without intubation. In comparison with the control group, significantly less children were intubated in the delivery room (24% vs. 41%) and needed mechanical ventilation (51% vs. 72%; both p < 0.05). Furthermore, compared with the historical control overall mortality (20% vs. 39%), rate of bronchopulmonary dysplasia (18% vs. 37%) and IVH > II° (10% vs. 33%) in survivors were significantly lower during the observational period (all p < 0.05). Neurodevelopmental outcome was normal in 70% of examined study group infants.
Conclusions: A revised delivery room management protocol was applied safely to infants with a GA below 26 completed weeks with improved rates of survival and morbidity.
• A revised gentle delivery room management policy, which aimed at assisting preterm infants in their adaptation to extrauterine life without mechanical ventilation, was applied safely to 164 ELGAN infants below 26 weeks of gestation.
• With this approach, mortality, rates of bronchopulmonary dysplasia (BPD) and severe IVH were significantly lower than in a historical control cohort of ELGAN.
•70% of infants treated following the revised protocol had no cognitive impairment.
There is growing evidence that many interventions performed to make survival of extremely low gestational neonates (ELGAN) possibly carry a risk of inducing adverse long-term consequences (1,2). As shown in a recent trial, delivery room cardiopulmonary resuscitation in ELGAN is a prognostic marker for higher mortality and neurodevelopmental impairment (3). Therefore, the authors stressed the need for improved delivery room strategies for these fragile patients. Referring to this, Jobe (4) and other authors (5) emphasized that only few VLBW preterm infants need resuscitation, while the majority only requires assistance to allow transition and adaptation to extrauterine life. In this context, Vento recently (6) referred to the ‘first golden minutes of the extremely low gestational age neonate’ and placed emphasis on the importance of a gentle delivery room approach.
In line with these reflections, we hypothesized that a gentle delivery room management protocol aiming at avoiding mechanical ventilation might improve both neonatal mortality and morbidity rates in ELGAN infants.
With the intention to validate this assumption, we revised our delivery room management protocol in 2001. Our revised protocol focussed on avoiding mechanical ventilation during the first hours of life and on supporting the preterm infants’ own vitality. This approach was in accordance with existing German and European guidelines (7,8) and was also applied to the most premature infants.
The objective of the present study is to give a detailed description of our approach, to report the outcomes, to compare them to a historical control cohort and to discuss these results with respect to current knowledge on physiological processes of adaptation to extrauterine life and available data from other centres. It is important to stress that the present work is a retrospective study and that therefore both further prospective trials are necessary to fully validate our experience and that our results may not be directly translatable into other neonatal units without taking into consideration that this approach requires specific training and a special delivery room ‘milieu’ that we developed over many years.
Patients and methods
This study comprised infants with a GA below 26 0/7 weeks who were born in our centre between 1 January 2000 and 31 December 2007. While the study cohort covered all infants born between 15 November 2001 and 31 December 2007, infants born between 1 January 2000 and 14 November 2001 formed the control cohort of our study.
In our level 3 neonatal intensive care unit, we provide care for preterm infants at the border of viability from 22-week gestation on. In Germany, maternal transferal into a level 3 unit is mandatory in case of imminent preterm delivery after completion of 24-week gestation (7). Between 22- and 24-week gestation, it is optional and transferal to a level 3 unit is organized if parents ask for it.
In our unit, all parents are prenatally informed about the chances and risks for their infant(s) and the therapeutic options at the actual GA are discussed on the basis of local and national data on short- and long-term outcome. Then, informed consent for all therapeutic interventions is obtained from the parents.
In general, every delivery of an infant with a GA of more than 22 full weeks of gestation is supervised by a senior neonatologist. The delivery room management approach is chosen based on the parents’ wishes and the infants’ condition.
In November 2001, we revised our delivery room management protocol to a more gentle approach of assisting infants in their transition and adaptation to extrauterine life. All attending neonatologists were trained in the protocol and agreed to perform the new protocol as the ‘centre-specific standard management’. We now performed a retrospective analysis of neonatal mortality and morbidity after introduction of this revised protocol and compared these data to a historical control.
Revised management protocol for the observational period (15 November 2001–31 December 2007)
Prenatal management included antenatal steroids and antibiotics in accordance with obstetrical guidelines. All foetuses from 22-week GA at risk of preterm delivery were considered candidates for antenatal treatment with corticosteroids (two doses of 8 mg betamethasone within 24 h). If the foetus was not delivered, antenatal steroids were not repeated regularly but another single dose of 8 mg of betamethasone was given before delivery if more than 14 days had passed since the initial steroid treatment.
In cases of imminent foetal distress, Caesarean section was performed from 22-week GA on. Regional anaesthesia was preferred to general anaesthesia whenever possible. Caesarean section (modified Misgav Ladach method) was performed by an experienced obstetrician who tried to extract the complete amniotic cavity containing foetus and placenta gently from the uterus. If successful, the amniotic sac was opened after the child was placed on the delivery room unit. The placenta was held above the infant for approximately 2 min by the midwife to allow transfusion of placental blood to the child. During placental transfusion, resuscitation manoeuvres were started as usual. If this approach was not possible, delayed cord clamping was performed in most patients by holding the foetus below the level of the mother′s introitus for 30–45 sec. During delayed cord clamping, the infant′s heart rate was not controlled and no respiratory aid applied. However, the child’s status was assessed visually by either a midwife or an obstetrician.
Delivery room management
Immediately after birth, the infants were placed on the delivery room unit with a prewarmed mattress and a radiant heater and wrapped in a polyethylene cover (Neowrap®, Fisher & Paykel Healthcare Limited, Auckland, New Zealand). Only in case of blood- or meconium-stained amniotic fluid, the oral cavity was suctioned. The baby received sustained continuous positive airway pressure (CPAP) via a face mask with a variable flow CPAP device (Benveniste valve, Dameca, Copenhagen, Denmark) (9) to recruit lung volume. A pulse oximeter was positioned on the right-hand side and used to monitor heart rate and oxygen saturation (SpO2). The gas flow was humidified and warmed to a temperature of 38°C. The temperature of the gas admixture was checked with a temperature sensing device (Fisher & Pykel) in the inspiration tube 20 cm distant from the gas outlet. FiO2 was initially set to 0.6 and a gas flow of 15 L/min was used resulting in a positive end-expiratory pressure (PEEP) of about 8 cm H2O. A gastric tube was inserted after 10 min to prevent abdominal gas accumulation.
Depending on the infants’ breathing efforts, heart rate and SpO2, both gas flow and supplemental oxygen were adapted. The target range for heart rate was >100/min after 1.5 min. If heart rate was not increasing >100/min after initiation of CPAP (Flow 14l, FiO2 0.6), flow was increased by 2 L/min (repeat three times for 30 sec). If heart rate was <100/min after 3 min, sustained inflation (30 sec) or bag and mask ventilation were performed (repeat three times, pressure limit 25 mbar, if unsuccessful 30 mbar) at the discretion of the neonatologist in charge. If heart rate remained <100/min, intubation was performed. Target of oxygen saturation was >85% after 10 min. Intubation was also performed if the infant did not commence to breathe after gas flow was increased to a maximum of 20 L/min (resulting in a PEEP of 14 cm H2O), and sustained inflation and/or bag and mask ventilation had been tried.
Criteria for surfactant application were evaluated after 10 min and included clinical signs of severe dyspnoea as defined by a Silverman Score >5 (10) and/or the necessity of FiO2 > 0.3 and/or >15 L/min of flow to keep SpO2 > 85%. All infants requiring surfactant received 100 mg/kg of a bovine surfactant preparation (Survanta®; Abbott, Wiesbaden, Germany) via a thin endotracheal catheter during spontaneous breathing with CPAP (11) at about 30 min of age as previously described.
The infants were then placed in an incubator and connected to an infant flow nCPAP generator (eme, Brighton, UK) or a Babylog 8000 ventilator (Draeger, Luebeck, Germany), respectively. All intubated infants received high-frequency oscillation ventilation (HFOV) following a high volume strategy as described elsewhere (12) (mean airway pressure 8–10 cm H2O, frequency 6–8 Hz).
Further revisions of the protocol that were introduced after the end of the observational period (from 2008)
After the studies of Wang and Escrig (13,14) were published in 2008, we changed initial FiO2 from 0.6 to 0.3. After Dawson′s nomogram for heart rate was published (15), we slightly adapted our protocol and changed the target range for heart rate from 100 to 120/min after 3 min. Our revised, currently applied protocol is shown in detail in a flow chart (Fig. 1).
Conventional management during the control period (1 January 2000 until 14 November 2001)
Prenatal management included counselling of parents and use of antenatal steroids in consent with the parents. Gentle extraction with intact amniotic sac and late cord clamping were not performed routinely. Resuscitation in the control period was performed according to ILCOR guidelines (16). If the infant was not breathing, PPV with FiO2 0.6 was applied and heart rate re-evaluated after 30 sec. If heart rate was <60/min, chest compressions were applied; if heart rate was 60–100/min, PPV was continued and intubation considered. If the infant was breathing, it was also the declared aim to stabilize the babies with CPAP. Infants who required surfactant were intubated for surfactant administration and put on mechanical ventilation. Criteria for surfactant application did not differ in both groups. Analogous to study group infants, HFOV following a high volume strategy was used.
Other therapeutic interventions during the control and the observational period
After admission to the NICU, all infants were treated following the same standards independent of their GA. Secondary intubation or re-intubation was performed in case of recurrent apnoea and bradycardia not responding to stimulation, or in case of respiratory failure defined as a pH <7.20 or FiO2 > 0.5 to maintain the paO2 in the intended range of 45–60 mmHg for more than 2 h.
Besides an intended paO2 of 45–60 mmHg and a pH >7.20, no strict limits were set for paCO2 provided that pH values maintained ≥7.20. Arterial hypotension (mean arterial pressure lower than GA in weeks) was treated if there was evidence of poor tissue perfusion.
All patients were screened for a patent ductus arteriosus with continuous left to right shunt at least once during the first 48 h of life. In case of ductal patency with L-R shunting, pharmacological treatment for closure was induced according to the German guideline with indomethacin (7). Intravenous fluids are started with 70–80 mL/kg/day. Enteral feeding was started on the first day of life with colostrums, if available. Minor changes were performed during the study period: gentamicin was replaced by tobramycin, daily protein intake was raised on day one from 1 to 3 g/kg/d daily and we stopped supplementing folic acid and zinc on a regular basis in the course of the observational period.
Data collection and statistics
Infants with a GA below 26 0/7 weeks who were born in our centre between 15 November 2001 and 31 December 2007 were included in this study. Control group infants were born from January 2000 to 14 November 2001. The following data were extracted:
Obstetrical data: Use of any antenatal steroids, use of tocolytic agents, perinatal risks, mode and place of delivery.
Neonatal data: GA at birth estimated on the basis of the last menstrual period or first trimester ultrasound investigation, birthweight, Apgar Score, duration of mechanical ventilation, CPAP.
Outcome data: Death before discharge, BPD defined as need for supplemental oxygen at 36 weeks, intraventricular haemorrhage (IVH) as defined by Papile, cystic periventricular leucomalacia (PVL), necrotising enterocolitis (NEC) with need for surgical intervention, retinopathy of prematurity (ROP) with need for surgical intervention, small bowel perforation.
Neurocognitive outcome in the study group was assessed by routine follow-up using the following standardized tests: Bayley Scales of Infant Development II and III, Kaufman Assessment Battery for Children and the German ‘Münchner funktionelle Entwicklungsdiagnostik’.
Outcome measures were compared with Fisher’s exact test and Mann–Whitney U-test, respectively, using the SPSS Package version 13.0 for Windows (IBM SPSS Inc, Chicago, IL, USA). All p-values are two-sided.
Before publishing our results, we presented the manuscript to our local ethics committee, which consented that an evaluation for the change of protocol had not been required in 2001.
Before we changed our approach in 2001, we performed noninvasive surfactant application in individual cases of very immature children the parents of whom did not consent to intubation and mechanical ventilation. These infants presented good spontaneous breathing but clearly needed surfactant. After explaining to the parents that surfactant application might be possible following an experimental not evidence-based approach without intubation and mechanical ventilation to apply surfactant, they consented. After we observed how well the infants did following this approach and considering the growing evidence of the negative effects of PPV to very immature preterm lungs, we decided to change our protocol and defined the here presented approach as our new centre-specific standard management for very immature babies. As the changes were not performed within a clinical study, the ethics committee was not informed in 2001, a time where ethical regulations in Germany were less strict. Furthermore, two multi-centred, randomized controlled trials for surfactant application without intubation had been initiated in 2007 and 2008, for which the ethical boards of all participating institutions had confirmed that this approach was ethical. Therefore, retrospectively, our protocol was approved by our ethics committee.
In total, clinical data of 208 infants were compared. One hundred sixty-four infants were treated following our revised delivery room management policy (study group). Data of these children were compared with 44 ELGAN infants defined as a historical control group. As highlighted in Table S1, demographic data and perinatal risks were similar in both groups. Significant differences between the two groups were observed for the following criteria: 5-min Apgar scores were higher in the control subgroup of infants born after 22-week GA, clinically suspected chorioamnionitis was present significantly more often in study patients born after 25 weeks of gestation and the control subcohort of children born after 25 weeks of gestation comprised significantly more boys (Table S1).
Table 1. Characteristics of study group (s) and control group (c)
In both cohorts, infants with a GA of 22 weeks were less frequently born by Caesarean section (23% vs. 86% study group, p < 0.001, 50% vs. 95% control group, p < 0.05) and received prenatal steroids significantly less often than more mature infants (62% vs. 95% study group, p = 0.001, 25% vs. 84% control group, p < 0.05).
The comparison of respiratory distress syndrome (RDS) management both in the delivery room and during the first 3 days of life is shown in Table S2. In the study group, 74% of all infants received only CPAP or CPAP and surfactant administration without intubation in the delivery room. In the control group, surfactant application without intubation was not available. Thus, only 11 control group infants (eight survivors; three nonsurvivors) were managed using CPAP alone. In line with this, significantly more control group infants were intubated in the delivery room (41% vs. 24%, p < 0.05) and were ventilated for RDS within the first 72 h of life (72% vs. 51%, p < 0.05).
Outcome parameters of both groups are presented in Table S3. Intriguingly, mortality was significantly higher in the control group compared with the study group (39% vs. 20%, p < 0.05). Regarding the study group alone, mortality was significantly higher in infants with a GA of 22 completed weeks than in the more mature infants (62% vs. 17%, p < 0.05). In contrast, there was no significant difference in mortality between control group infants born after 22-week GA and more mature infants (50% vs. 38%, ns).
Moreover, the incidence of BPD in the control group was twice as high as in the study group (37% vs. 18%, p < 0.05), and BPD occurred significantly more frequently in the more immature patients of 22- and 23-week GA (study cohort: 32% vs. 14%, p < 0.05). Similarly, control group infants significantly more often experienced higher-grade IVH (grade III or IV: 33% vs. 10%, p < 0.05). In contrast, there were no differences regarding low-grade IVH (grade I or II), PVL, ROP needing laser therapy, NEC requiring surgery and small bowel perforation.
Data on neurocognitive outcome were available for 96 survivors (73%). Sixty-seven (70%) were classified as normal, 12 (13%) showed minor (IQ 70–84) and 17 (18%) major cognitive impairment (IQ < 70). Data for control group infants were only available for 12 infants (44%). One child had minor (8%), three children major (25%) and eight children (67%) no cognitive impairment.
Delivery room management of infants at the border of viability is an issue of intense discussion. In 2001, we revised our delivery room management protocol to gently support extreme premature infants in their transition and adaptation to extrauterine life. Intriguingly, we observed that it is not only safely possible to treat even the most immature infants with such an approach but that our strategy resulted in both improved rates of survival and morbidity in ELGAN infants.
With respect to survival, the overall mortality in our study group infants with a GA below 26 completed weeks was remarkably low (20%) in our cohort as compared with 24–74% in other studies (17,18). Yet, our results are in line with similar observations by other groups, who reported that the mortality of infants <1000 g who received mechanical ventilation on the first day of life was almost twice as high (20%) as compared with infants who were supported with CPAP (19). In addition, the study by Thomas et al. also supports the assumption that avoiding mechanical ventilation in the first hours of life has contributed in large parts to the improved survival in our analysis. While both safety and efficacy of early CPAP have been established recently in several well-conducted trials (20,21), it is our impression that the combination of early surfactant administration during spontaneous breathing with CPAP support and without any positive pressure ventilation is the most important point of our revised protocol. This combination is also the most striking difference to similar protocols from other groups (20,22,23) and might therefore explain in large parts the good overall outcome in our study cohort.
Although surfactant administration in combination with CPAP support has been shown to effectively reduce the need for mechanical ventilation (24), still 77% of the most immature of the infants of our study cohort needed mechanical ventilation for RDS within 3 days after birth. Those very immature babies have more immature lungs producing less surfactant and have more difficulties in gas exchange because their gas exchange surface is smaller and the distance between alveolar sacks and capillary is greater. Thus, the high rate of infants needing mechanical ventilation is most likely attributed to lung immaturity and not indicative of an insufficient response to our treatment approach.
Of further importance, the flow rate and CPAP pressure that we applied during neonatal transition were higher than traditionally used in most units. As the COIN study (20) showed an increased incidence of pneumothorax using a CPAP level of 8 cm of water, we anticipate that our approach could raise concerns regarding negative pulmonary effects. However, it has to be stressed that we only used high flow levels during initial neonatal transition (e.g. for 10–30 min after birth). If the infants continued to require high flow and/or FiO2 levels (>0.3), surfactant application was mandatory. In contrast, in the COIN study, rescue surfactant was only given after intubation and CPAP failure was defined as needing treatment with an FiO2 level >0.6. Furthermore, Mulrooney et al. (25) investigated physiologic responses of preterm lambs to CPAP and found that a CPAP of 8 cm H2O improved lung function during initial neonatal transition to air breathing relative to a CPAP of 5 cm H2O without increasing pneumothorax.
Apart from avoiding mechanical ventilation whenever possible, we consider several other steps to be important for patients’ outcome. One additional factor in our opinion is an aggressive obstetric management with early Caesarean section in cases of foetal distress even in the most immature of infants. The rate of Caesarean sections was not different in the study and control group with higher rates of Caesarean sections performed under general anaesthesia in the control group. Yet, less of the most immature infants of 22-week GA were born by Caesarean section as compared with more mature infants. We think that very immature infants may profit as much as more mature infants from an active obstetrical approach with Caesarean section performed in imminent foetal distress, although we could only partially realize this approach in our study. In line with this assumption, a recent trial reported improved neurodevelopmental outcome values and reduced mortality and IVH rates in ELBW infants who were delivered by Caesarean section than in ELBW infants who were delivered vaginally (26). Delayed cord clamping has been shown to both reduce IVH and to protect from motor dysfunction in VLBW infants (27) and therefore should in our opinion be integrated into current management protocols for ELGAN infants.
As survival of extremely premature infants is often burdened with severe morbidity, it is feared that lower mortality rates might imply higher morbidity. In the present study, we observed an incidence of BPD of 18% in the study group compared with 30–70% in the literature (17,20,21) and 37% in our control group. This observation is in line with another recent trial in which infants between 26 0/7 and 28 6/7 weeks of GA who were treated with early CPAP and surfactant without intubation had lower need for oxygen on day 28 compared with children treated with early CPAP and rescue intubation (24).
The rate of severe cerebral lesions (IVH III and IV) in the study group was only 1/3 compared with control group infants. Other studies showed similar results with IVH III and IV rates being two to four times as high if ELBW infants were mechanically ventilated one the first day of life (19,28).
We think that our described mode of delivery room management is an adequate way to care for infants at the border of viability. It may have the impact of improving survival without an increase in neonatal morbidity. Long-term outcome in our study group showed severe disability in 18% of examined survivors (29). Studies analysing neurocognitive outcome of infants of similar GAs reported rates of severe disabilities in 20–40% of patients (30). There is a well-grounded hope that a better short-term outcome could also be related to a better long-term outcome. Studies (19,28) confirm this observation showing that children <1000 g treated with CPAP compared with mechanical ventilation on the first day of life had significantly better neurocognitive outcome and significantly higher developmental quotient.
Despite the encouraging results of our approach, our study has several limitations: first, it is a single-centre retrospective observational investigation without stratification, not blinded and not randomized. Therefore, it is probably difficult to extract general conclusions from our results. In particular, one could hypothesize that perhaps control cohort babies born in 2000 and 2001 being 22 or 23 weeks of gestation were not resuscitated with the same intensity and proactive approach as later on when we changed the protocol simply because obstetricians and neonatologists were less confident in the possibilities of survival of these babies. In this context, it is of importance to note that the very immature babies included in the control cohort of our study were only a fraction of all infants actually born at 22 or 23 weeks of gestation in 2000 and 2001 while most of these children were not actively treated at that time. Infants who were treated were resuscitated very proactively because of an explicit parental wish. Additionally, they were vigorous after birth, which is indicated by the significantly higher 1- and 5-min Apgar scores in the control group children of 22-week GA compared with the study group infants. After we observed the high survival rates of these children, we started to treat more extremely immature babies more proactively even if the infants were less vigorous after birth. A second limitation is the huge difference in the number of infants in the observational and control group that could have also influenced our results. Similarly, although boys and girls were evenly distributed comparing the total study and control cohorts, the subgroup of infants born after 25 weeks of gestation consisted of significantly more boys in the control than in the study group, so that gender differences might have confounded the results of this subcohort. Third, multiple changes both in obstetrical and neonatal procedures were performed almost simultaneously. Therefore, it is not possible to separate the effects of these individual changes in the obstetric and neonatal field, and we cannot analyse how obstetrical or neonatal procedures like delayed cord clamping or HFOV may have influenced our results. Due to these difficulties, we cannot recommend adopting our approach without restrictions and we think that future trials are necessary to accurately assess the impact of any of the individual steps of our approach on infants′ outcomes. Lastly, the neurodevelopmental follow-up data for children of the control group are rather scarce.
Finally, we consider it of central importance to stress that the approach we outline in the present manuscript might not be readily applicable in other neonatal units. It is important to note that our present protocol was developed and implemented after extensive training and constant supervision and that a special delivery room ‘milieu’ had been developed over many years.
Nonetheless, we think that our data in summary substantiate recent proposals to rather assist very immature preterm infants in their transition to extrauterine life than to resuscitate them (4). Furthermore, our data indicate that the greatest potential benefit of a gentle delivery room management policy can be envisioned for the most fragile infants at the border of viability. Yet, evidence-based data on optimal delivery room management approaches for the most immature of infants are still limited, and well-designed randomized controlled trials are needed to compare the impact of different strategies. We therefore hope that our observations invite neonatal specialists to expand discussion and research on an optimal delivery room management in extremely premature babies.
We thank André Oberthuer for critical reading of the manuscript. We also thank Peter Mallmann and Markus Valter, responsible obstetricians for optimized prenatal management.