Neonatal resuscitation after severe asphyxia--a critical evaluation of 177 Swedish cases.

AIM: To evaluate neonatal resuscitation of infants born with severe asphyxia.
METHOD: All case records of the 472 claims for financial compensation due to suspected medical malpractice in conjunction with childbirth in Sweden between 1990 and 2005 were scrutinized. Inclusion criteria were: gestational age > or =33 completed weeks, planned vaginal onset of delivery, a reactive CTG at onset of labour, neonatal asphyxia (defined as metabolic acidosis [pH of < 7.05 and/or a base excess of < -12]), or an Apgar score <7 at 5 min. It was assessed that 177 infants suffered from cerebral palsy or early death due to severe asphyxia presumably caused by malpractice around labour.
RESULTS: Median Apgar score at 5 min was 3, indicating that all infants needed immediate and extensive resuscitation. There was insufficient adherence to guidelines concerning neonatal resuscitation, including delayed initiation of excessive resuscitation in 19 infants, lack of satisfactory ventilation in 79 infants, and untimely interruption of resuscitation in 38 infants.
CONCLUSIONS: Compliance with guidelines for resuscitation of severely asphyctic newborn may be improved, especially concerning ventilation and prompt paging for skilled personnel in cases of imminent asphyxia. Documentation of neonatal resuscitation must be improved to enable reliable evaluation.

INTRODUCTION

Experience over the last century has demonstrated that perinatal mortality can be reduced by improved obstetrical and neonatal care (1). With the aim to avoid errors in care by implementing system-based changes, a systematic review of the pitfalls and mistakes in the clinical practice of perinatal medicine can be useful.

In a previous Swedish nationwide report, we identified 177 newborn infants over a 16-year period who had suffered from severe birth asphyxia presumably due to delivery-related malpractice. The most common causes of obstetrical errors were that established guidelines for foetal surveillance were not followed (i.e. negligence to supervise foetal well-being), signs of foetal asphyxia were overlooked, misuse of oxytocin, and a non-optimal choice of mode of delivery. We concluded that foetal surveillance and attention to signs of asphyxia must be improved, that there is a need to improve cooperation between professionals in the labour unit and, to create security barriers (2).

Less is known about potential flaws in the immediate management of the asphyxiated newborn infant. Although national guidelines for neonatal resuscitation have been in place for many years (3), compliance with and the effectiveness of these guidelines has not been studied. Video uptakes from emergency rooms indicate that deviations from protocols and errors in implementing guidelines are common (4). The aim of this study was to scrutinize the initial resuscitation procedure of the previously identified 177 newborn infants with severe asphyxia (2).

METHODS

We retrieved information on all 472 claims from 1990 to 2005 concerning suspected malpractice during pregnancy, delivery and the neonatal period, sent to the County Council's Mutual Insurance Company under the Patients' Advisory Committee, Sweden (PAC, Person Skade Reglering AB in Swedish). We included infants with a gestational age of ≥33 completed gestational weeks, planned vaginal delivery, a normal CTG tracing at admission to the delivery unit, and severe asphyxia-related neurological outcomes. We defined labour-related asphyxia as an Apgar score of < 7 at 5 min and, if the acid base status was measured in the umbilical cord at delivery or shortly thereafter, metabolic acidosis as a pH of < 7.05 and/or a base excess (BE) of < –12 (5,6). In all, we included 177 infants with labour-related asphyxia, presumably caused by malpractice in connection with labour. The selections of the cases and the investigation procedure have previously been described in detail (2).

A structured protocol, including clinical data retrieved from the obstetric and neonatal records, was used. We collected information on pregnancy and maternity care, onset of labour, presentation and mode of delivery, time of paging of a physician, the date and time of birth, gestational age, gender, birth weight, Apgar scores at 1, 5 and 10 min, umbilical cord acid base status, acts of resuscitation, the length of stay (LOS) in the neonatal intensive care unit (NICU), degree of hypoxic ischemic encephalopathy (HIE) and investigations during the first weeks of age. We also retrieved information on the long-term follow-up, including information on morbidity (cerebral palsy [CP] and other neurological disorders), degree of disability or death from the paediatric medical records.

All medical documents were scrutinized and computerized by one of us (SB), a specialist in obstetrics for a period of 15 years, a graduate of the American Neonatal Resuscitation Provider Program, and also an instructor of team training in neonatal resuscitation at the Centre for Education in Paediatric Simulator (CEPS) at Södersjukhuset in Stockholm (1). The reviewer SB needed access to all information from each case record and was therefore not blinded to the final outcome.

National guidelines on neonatal resuscitation in place during the period of the investigation are presented in Figure 1. (3) According to these guidelines, (revised in 2006), clearing the airways from meconium and initiating bag and mask ventilation with initially 40% oxygen should be performed immediately in all newborn infants with a heart rate below 100 beats per min (bpm). In cases of asystolia, endo-tracheal intubation and administration of thoracic compressions should immediately be performed. Early intubation (within the first 2–3 min after birth) was recommended in cases with persistent bradycardia to ensure free airways and satisfactory ventilation. Adrenalin, either through the endo-tracheal tube or intravenously, should be administered if the heart rate did not rise despite satisfactory ventilation. Correction of metabolic acidosis should be performed in cases of refractory bradycardia and terminal apnoea (2 mmol buffer/kg i.v. for 5 min). Interruption of resuscitation should be considered if there is no evidence of heart activity after 15 min or if no spontaneous breathing or body movements have occurred during the first 30 min after birth.

Figure 1

National Swedish guidelines for neonatal resuscitation 1996–2006.

According to national clinical practice in obstetrics, the paediatrician should be paged before the birth in cases of imminent asphyxia, in breech and instrumental deliveries, and immediately when deliveries are complicated by shoulder dystocia. In cases of unexpected asphyxia, the desirable time for the arrival of skilled assistance was set at being within 4 min after birth for the paediatrician and within 15 min for the neonatologist (back up).

Based on these guidelines, we established the following cases of failure events in conjunction with neonatal resuscitation:

Artificial ventilation not started within 1 min after birth in cases of apnoea and foetal heart rate <100 bpm.

Endotracheal intubation not performed within 3 min after birth in cases of asystolia or in cases of persistent apnoea and bradycardia.

Thoracic compressions not performed despite asystolia.

Adrenaline not provided despite asystole or persistent bradycardia.

Correction of metabolic acidosis not performed in cases of asystole or persistent bradycardia.

Paediatrician or neonatologist not present before birth despite knowledge of delivery complication.

Paediatrician or neonatologist not arrived within 4 min after birth in cases of unexpected asphyxia.

Resuscitation not interrupted in spite of asystole for >15 min or lack of spontaneous breathing or movements > 30 min after birth.

All statistical analyses were performed in SPSS for Windows, Rel. 14.0. 2005. Chicago, IL, USA: SPSS Inc. Data are presented as median (range) values or as numbers and proportions (%). The study was approved by the research ethics committee at Karolinska Institutet (number 1589–2006).

RESULTS

All 177 mothers attended antenatal care on a regular basis. All cases had a reactive CTG recording at admission for labour, indicating a non-asphyxiated foetus. The onset of labour was spontaneous in 130 cases and induced in 47 cases. In six of the 48 spontaneously delivered infants, labour was complicated by shoulder dystocia and nine infants were born in breech presentation. Sixty-five infants were delivered by emergency caesarean section (CS), 62 by vacuum extraction and 2 by forceps. Twelve of the 62 deliveries by vacuum extraction were complicated by shoulder dystocia. In twenty-five deliveries, trial of labour with vacuum extraction or forceps was performed before converting to emergency CS.

The median Apgar score was 5 (range 0–7) 10 min after birth, and at an older post-natal age there was rarely any notation of an Apgar score. A total number of 32 (18%) infants died, 16 in the neonatal period and 16 during later follow-up.

Information about the time of the initiation of artificial ventilation was noted in 167 of all 177 infants. Despite clear indications, artificial ventilation was not initiated within 1 min after birth in seven infants (4.2%) while waiting for a physician. Two of these had an Apgar score of ≤3 at 10 min and one infant died (Table 1). The median time for mask ventilation was 8 min (range 0–30 min), and mask ventilation proceeded for more than 15 min in 34 infants. Eight of these infants had an Apgar score of ≤3 at 10 min and seven infants died.

Table 1

Failure events during neonatal resuscitation of 177 asphyxiated newborn infants and outcome

	n	(%)	Apgar score 0–3 at 10 min (n)	Deaths1 n	(%)
1. Non satisfactory/unsatisfactory resuscitation	84	(47)
(a) Artificial ventilation not started within 1 min after birth	7		2	1	(14)
(b) Endotracheal intubation not performed within 3 min after birth	72		26	17	(24)
(c) Thoracic compressions not performed despite asystole2	5		1	0
2. Not satisfactory drug administration	23
(d) Adrenaline not provided despite asystole1	14		6	5	(24)
or persistent bradycardia3	7
(e) No correction of metabolic acidosis3 despite asystole or persistent bradycardia	2		2	1	(50)
3. Late initiation of resuscitation	19
(f) Paediatrician, neonatologist or anaesthesiologist not present before the birth in spite of known delivery complication	11		2	3	(27)
(g) Paediatrician, neonatologist or anaesthesiologist not arrived within 4 min after birth in cases of unexpected asphyxia (n = 48)	8		4	4	(50)
4. Resuscitation not interrupted	38
(h) Resuscitation not interrupted:
○ In spite of asystole for more than 15 min	8		8	7	(88)
○ In spite of lack of spontaneous breathing during the first 30 min after birth	30		15	8	(27)

There were 16 neonatal deaths, and another 16 during follow-up (total deaths 32/177 = 18%).

33 neonates had asystole at 1 min of age.

24 neonates had persistent bradycardia at 5 min.

Endotracheal intubation was performed in 112 cases, and the time of intubation was noted in 93 infants. Median time for intubation was 6 min (range 0–180 min). In 72 infants, intubation was not, despite clear indications, performed within 3 min after birth. Twenty-six of these infants had an Apgar score of ≤3 at 10 min and 17 died (Table 1). Twenty-eight infants were intubated after more than 10 min of age. The time of the initiation of spontaneous breathing was noted for 119 infants of all 177 infants (67%), and the median time was 18 min (range 1–70 min). Forty-one infants were subjected to chest compressions. Despite asystole at one min of age (n = 33), thoracic compressions were not performed on five infants, one of which had an Apgar score of ≤ 3 at 10 min (Table 1).

Adrenaline was not provided in 14 of 33 infants with asystolia at one min and nor was it provided in seven of 24 infants with persistent bradycardia. In all 21 infants where adrenaline was not provided, six infants had an Apgar score of ≤ 3 at 10 min and five of these infants died (Table 1). An acid base status at birth or shortly after birth was analyzed in 107 of all the infants (60%). The median pH at or shortly after birth was 6,83 (Range 6,57–7,20) and BE –22,5 (–10 to –40) mmol/L. One hundred thirty-five infants were treated with buffer. The time for the correction of acidosis was noted in 47 (35%) infants, and the median time for correction was 15 min (range 1–120 min). In two infants with asystolia or persistent bradycardia there was no correction of metabolic acidosis. Both had an Apgar score of ≤3 at 10 min and one of these infants died (Table 1). Four infants had a post-natal asphyxia after a traumatic delivery with a pH >7,05, a BE >–12 mmol/L, but an Apgar score of < 7 at 5 min.

Information relating to the attending physician and his/her time of arrival at the resuscitation was not routinely noted. Of 92 vaginal-risk deliveries (65 instrumental, 18 with shoulder dystocia and nine breech deliveries), the presence of a physician (paediatrician, neonatologist or anaesthesiologist) before birth was noted in 58 (63%) deliveries. Despite knowledge of a delivery complication none of the physicians were paged in advance in 11 cases, of which two infants had an Apgar score of ≤ 3 at 10 min and three died (Table 1). In cases of unexpected asphyxia (n = 48), the time of arrival of the neonatologist, paediatrician or the anaesthesiologist was noted in 42 cases. The time of arrival exceeded 4 min in the case of eight infants, four of whom had an Apgar score of ≤ 3 at 10 min and died later (Table 1).

Resuscitation was not interrupted in spite of asystolia for more than 15 min in 8 infants, all of whom had an Apgar score of ≤ 3 at 10 min and 7 eventually died. Similarly, resuscitation was not interrupted in spite of lack of spontaneous breathing or spontaneous body movements during the first 30 min after birth in 30 infants. Fifteen of these infants had an Apgar score of ≤ 3 at 10 min, of which eight infants died (27%, Table 1).

All infants were admitted to the NICU. Neonatal outcomes following resuscitation are summarized in Table 2. The median length of stay in the NICU was 24 days (range 1–120). Of 128 infants born at hospitals without facilities for ventilator treatment, 48 (38%) were transferred to a university hospital shortly after birth. Seventy-eight (44%) infants were treated on ventilators for more than 24 h. Almost all (94%) infants suffered from convulsions during the neonatal period. EEG was performed on all infants and was pathological in 80%.

Table 2

Neonatal outcomes and characteristics following severe asphyxia and resuscitation

	n	(%)	Length of stay or duration, days	Range
NICU	177	100	24	(1–120)
Ventilator therapy >24 h	78	(44)	4	(1–19)
Convulsions	167	(94)
Anticonvulsive medication at discharge	113	(64)
Transportation to university hospital	48	(27)

Information on HIE missing in 40 cases (23%).

Examination by cerebral CT scan was performed on 137 (77%) of all infants, and the scan was assessed as pathological in 96 infants (70% of those scanned). In addition, 58 (33%) were examined with magnetic resonance imaging (MRI) of the brain, and the MRI was abnormal in 49 infants. The degree of HIE was diagnosed in 137 infants. Among those, HIE I was diagnosed in 10, HIE II in 69 and HIE III in 58 infants (Table 2).

MORTALITY AND LONG-TERM MORBIDITY

Information on mortality, neurological disorders, co-morbidity and degree of impairment has been published previously (2). Briefly, there were 16 neonatal deaths, and another 16 infants died during the follow-up. Information on morbidity was for 45 infants based on case records with a limited follow-up period. These infants were considered to be suffering from an unspecified CP syndrome with diagnosed encephalopathy during the neonatal period. Nine of these infants died before a final diagnosis could be made. Sixty-nine of the 116 children with information about the type of CP had dyskinetic CP, 18 spastic tetraplegia, 21 spastic diplegia and 8 had hemiplegia.

DISCUSSION

We are very much aware of the difficulty of evaluating the efforts of resuscitation in these most severely asphyxiated infants since the final outcome was poor for all, but nevertheless we could evaluate compliance with the guidelines relating to neonatal resuscitation. In accordance with our previous findings on insufficient adherence to current guidelines on foetal surveillance (2), we found indications of similar insufficiencies with respect to neonatal resuscitation and that resuscitation was not performed according to guidelines in the case of 84 infants (47%) (Table 1). Since ventilation is the most important issue in cases of neonatal resuscitation, we consider the most perturbing findings to be a delayed initiation of extensive resuscitation, lack of satisfactory ventilation by insurance of free airways by an early intubation, and untimely interruption of resuscitation (Fig. 1 and Table 1).

Foetal surveillance at admission to the delivery ward registered no irregularities in any of the cases, presumably indicating a healthy foetus. After delivery, umbilical cord blood gas and acid-base assessment are the most objective determinations of the foetal metabolic condition, but acid-base status at birth or shortly thereafter was available only in 60% of all cases. On the other hand, Apgar score at 5 min, which is considered to be a reliable indicator of asphyxia in the absence of malformations, was noted in all cases (7). Median pH at birth was 6, 83, median BE was –22,5 and the median Apgar score at 5 min was 3, indicating that the neonates in the study group suffered from profound asphyxia at birth and were in great need of an immediate initiation of neonatal resuscitation. The documentation of the resuscitation was generally poor and incomplete, which is a warning sign itself that ought to be elucidated. Despite the difficulty of prioritizing documentation in these stressful situations, documentation is necessary, not least in severe cases where infants might have been injured by malpractice in conjunction with labour. A simple and straight forward pre-printed protocol simplifies documentation in cases of neonatal resuscitation (8). The poor documentation also emphasizes that the results from our evaluation must be interpreted with caution.

Although there were several shortcomings during the neonatal resuscitation, we consider it most likely that the infants were primarily damaged by asphyxia in conjunction with labour. Immediately after birth, the most important shortcomings noted were, according to the national guidelines for neonatal resuscitation, unsatisfactory ventilation or drug administration and a late initiation of or untimely interruption of resuscitation.

Even if the majority of failure events defined in this study is uncontroversial, failure to interrupt resuscitation after 15 or 30 min is more contestable. There could be clinical situations when the attending physician would have considered continuing resuscitation despite a longstanding asystolia or persistent apnoea. For example, the parents may have opposed discontinuation of resuscitation. Drugs administered to the mother immediately before delivery may have depressed spontaneous breathing in the newborn infant, demanding continued artificial ventilation after birth. An unclear diagnosis of intrapartum asphyxia—for other reasons than depressant drugs—may also have added to a decision of continued resuscitation beyond the stipulated points of no return.

Although the value of routine correction of acidosis in newborn infants has been rightfully questioned, we note that in our study of severely asphyxiated infants, there were two patients with asystolia or persistent bradycardia who did not receive any correction of metabolic acidosis (9,10). Both had an Apgar score of ≤3 at 10 min and one of these infants died. An acid-base status was available in 107 out of 177 infants and 135 infants were treated with buffer. The time for correction of acidosis was noted in only 47 (35%) infants. Among these infants, correction of acidosis was in many cases delayed (median time for correction 15 min (range 1–120 min) after birth). The point of time for correction of acidosis was missing in too many cases to provide meaningful indication whether an earlier correction could have improved the outcome.

Post-asphyxial management at the neonatal intensive care unit is also of vital importance for the chance of intact survival (11,12). Aggressive prevention and treatment of complications such as recurrent acidosis and hypoxia, severe hypoglycaemia, systemic hypotension, convulsions and brain oedema are necessary (13,14). Recently, moderate hypothermia has been reported as an evolving therapy with promising results in cases of neonatal asphyxia (13,15). Quality assurance of the post-asphyxial management remains to be done.

Neuroimaging gives the possibility to time-relate a brain injury. Clinical consequences are dependent on the topography of the lesion. In cases of asphyxia during labour, basal ganglia and thalamus lesions occur, predominantly because of acute and severe perfusion failure in the term or near term born infant. Newborn infants with pure basal ganglia and thalamus lesions tend to have dyskinetic CP at follow-up, often with severe motor problems but less pronounced cognitive deficits. Infants with additional central and especially hippocampal involvement are usually severely retarded in both their motor and cognitive development and develop severe bilateral spastic CP (16,17). The dominating types that occurred among our cases were dyskinetic CP and spastic tetraplegia, which are thought to be the most common subtypes of CP that are associated with intrapartum hypoxic events (14). Although some infants in the study-group may have had encephalopathy prior to labour, we find it probable that suboptimal care around delivery aggravated the encephalopathy in these infants. The median LOS at NICU was 24 days, 44% were treated for more than 24 h in a ventilator, and almost all had convulsions with pathological EEG during the neonatal period, which indicates how severely injured these infants were at birth. Neonatal resuscitation is important in cases of asphyxia. Even if the percentage of newborns requiring assistance may be small, the implications of not receiving correctly performed neonatal resuscitation can cause lifelong impairment or even death. This report does not intent to cover all neonatal resuscitation in connection with malpractice during delivery. However, given our outcome definition of asphyxia at birth and the long-term outcomes of these infants, the study probably mirrors the initial care of the most severely asphyctic neonates.

Ten percent of all newborns require some assistance to adjust to the extra uterine environment, while 1% needs extensive resuscitative measures to survive (1). The crucial time for improvement in cases of severe asphyxia at birth is around 10 min, and it is necessary that labour staff available initiates resuscitation and that more skilled staff participates within a few minutes (1). Asphyxia was unexpected in 48 of the infants, and in eight of these cases neither a paediatrician, nor a neonatologist or an anaesthesiologist had arrived to assist within 4 min of the birth, which may have aggravated the asphyxia. Four of these infants had an Apgar score of ≤ 3 at 10 min and died later. In addition, resuscitation was noted delayed by the absence of skilled persons in 11 of the 92 cases of known delivery complications, but notation about presence before birth was missing in another 34 complicated deliveries. Due to the possibility of unexpected asphyxia, it is necessary that all staff attending childbirths continuously train how to anticipate and handle complications in conjunction with labour (18) while waiting for skilled personnel to arrive.

CONCLUSIONS

There are possibilities for improvement in the immediate neonatal resuscitation within our labour units. The most important contributions may be made by improving compliance with the guidelines concerning ventilation, and the paging for the early assistance of skilled personnel in cases of imminent asphyxia. It is crucial that all of the staff on the labour ward is familiar with how to initiate extensive neonatal resuscitation. Every case of unexpected asphyxia, also those that recover without sequelae, should be scrutinized to enable the creation of security barriers and improvements in each labour unit, concerning both obstetrical care and neonatal resuscitation (19). In addition, the documentation of neonatal resuscitation must be improved to enable accurate and reliable evaluation.