The continuing fall in incidence of hypoxic-ischaemic encephalopathy in term infants


Correspondence: Dr K. Dodd, Derbyshire Children's Hospital, Uttoxeter Road, Derby DE22 3NE, UK.


Objective To examine trends in incidence of hypoxic-ischaemic encephalopathy in term infants over a twenty-one year period.

Design A retrospective analysis of medical records of all term infants admitted to a neonatal unit with hypoxic-ischaemic encephalopathy during the years 1992–1996 (period C) and a comparison with data from the years 1976–1980 (period A) and 1984–1988 (period B) from the same unit (previously published1).

Setting A District Health Authority in Central England serving a population of about 450,000.

Sample All term infants admitted with clinical features of hypoxic-ischaemic encephalopathy.

Main outcome measures Incidence of three grades of hypoxic-ischaemic encephalopathy, disability and mortality.

Results In each five year period there were similar numbers of births. Over the time-span of this study the stillbirth rate and neonatal mortality rate has consistently fallen. The overall incidence of hypoxicischaemic encephalopathy in term infants was significantly lower (P < 0.001; OR 0.42 CI 0.29–0.59) in the present study period (C) compared with the earlier study period B (1.9 vs 4.6 per 1000 total live births). The fall in moderately and severely affected infants between the present and the first study period was significant (1.2 vs 2.6 per 1000 total live births, P < 0.001: OR 0.46 CI 0.29–0.72). The number of deaths and the incidence of cerebral palsy in survivors fell progressively over the 21 years spanned by this study.

Conclusion This study shows that the incidence of hypoxic-ischaemic encephalopathy and its sequelae in term infants has fallen significantly. The use of cardiotocography and caesarean section rates have risen but the relative contributions of changes in clinical practice are uncertain.


In low risk pregnancies the incidence of stillbirths or neonatal deaths due to intrapartum asphyxia is regarded as a sensitive measure of the quality of intrapartum care2–4. The 4th annual report of the Confidential Enquiry into Stillbirths and Deaths in Infancy5 recently identified sub-optimal care in many intrapartum deaths.

In an earlier study we reported a fall in incidence of hypoxic-ischaemic encephalopathy in term infants during the period 1984–1988 (period B) compared with 1976–1980 (period A) in babies admitted to Derby City Hospital Neonatal Unit1.

The aim of this study was to determine whether there had been a significant fall in incidence of moderate and severe hypoxic-ischaemic encephalopathy in the same neonatal unit, by looking at the period 1992–1996 (period C). Information was also obtained on the subsequent morbidity and mortality of the three grades of hypoxic-ischaemic encephalopathy in the latest study period. The findings were looked at in the context of other changes occurring over the time period of this study such as demographic changes and clinical practice.


The infants

In the first study1, for practical reasons, constrained by time and resources, data was studied from a total of 10 years. As complete data were available from 1976–1988, a three year gap between the study periods was chosen to achieve the longest possible time span for the study. In this present study, a further three year gap was left between study periods in order to compare trends over a total of 21 years. The medical records of all infants admitted to Derby City General Hospital Neonatal Unit over 37 weeks of gestation from 1992–1996 were therefore reviewed.

The study was restricted to term infants as the neurological syndrome of hypoxic-ischaemic encephalopathy, seen following a period of perinatal asphyxia, characteristically occurs in babies of this gestation.

The records where the infant had features of hypoxic-ischaemic encephalopathy, as defined by Levene et al.6, were identified by L. W. and J.S. The illness was graded according to the following criteria6: Grade 1 (mild): irritability, hyperalert, mild hypotonia, poor sucking; Grade 2 (moderate): lethargic, seizures, marked abnormality of tone, required tube feeding; Grade 3 (severe): comatose, prolonged seizures, severe hypotonia, failure to maintain spontaneous respiration.

The records were cross-checked using information from the neonatal unit audit data. In the previous study the medical records of all infants greater than 37 weeks of gestation admitted to the same neonatal unit in two five year periods A and B were reviewed to identify those infants with hypoxic-ischaemic encephalopathy. By combining the two sets of data we had information on babies with hypoxic-ischaemic encephalopathy in three 5-year periods spanning 21 years (1976–1996). The overall admission rate (5%–7% of live births) and the neonatal unit's admission policy for hypoxic-ischaemic encephalopathy (i.e. all infants with features suggestive of hypoxic-ischaemic encephalopathy are admitted) were unchanged throughout the study. It is, however, possible that a small number with Grade 1 hypoxic-ischaemic encephalopathy were not admitted to the Neonatal Unit but remained on the postnatal ward.

Infants with encephalopathy clearly attributable to a cause other than asphyxia were excluded. One baby initially identified with encephalopathy and included in the study was subsequently excluded as post mortem findings were consistent with a mitochondrial cytopathy.


The medical notes of all the infants with hypoxic-ischaemic encephalopathy also contained information on follow up and outcome. Some babies with Grade 3 disease died. All those with Grade 2 or 3 disease were followed up in clinic by K.D., though the majority were discharged by 18 months if progress was satisfactory. The follow up neurological assessments were performed with prior knowledge of the perinatal history and grade of hypoxic-ischaemic encephalopathy. The diagnosis of cerebral palsy, in all those with suspect development, was confirmed by K.D. who has a specialist interest in child development.

In the previous study periods, all those recorded to have severe motor disability such as spastic quadriplegia, hemiplegia or diplegia were included in the analysis. Others noted to have less precise problems such as delayed development or speech problems were excluded. In the present study period information on milder disability was noted in addition to those with severe motor disability.


A proforma was completed for each patient identified. The following information was collected: maternal age and parity, maternal smoking and alcohol consumption, ethnic group, gestational age, length of first and second stage of labour, mode of delivery, indications for any assisted delivery, presence of meconium staining of amniotic fluid, abnormalities of cardiotocography, birthweight, head circumference, sex, Apgar scores at 1 and 5 minutes and details of any resuscitation at birth. The records were complete for all these items. Data on fetal and cord blood gas analysis was incomplete and therefore unhelpful for the purposes of this enquiry.


The City Hospital Neonatal Unit serves a large district Health Authority with a population of around 450,000. In each study period the total number of births was similar. However, in the five year period 1992–1996 the majority of babies were born in the Derby City Hospital maternity unit with the remaining 1.3% of babies born at home. This situation differs from the two previous periods studied when a proportion of babies were born in other local maternity hospitals or maternity homes which have now closed (Table 1). All babies identified as having hypoxic-ischaemic encephalopathy in period C were born in Derby City Hospital.

Table 1.  Number of births in each five year period according to place of birth. Values are given as n (%).
PeriodDistrict general hospitalMaternity homesMaternity hospital AMaternity hospital BHomeTOTAL
1976–198016,462 (66)5394 (22)821 (3.3)1908 (7.7)240 (0.9)24,824
1984–198821,872 (90)868 (3.6)373 (1.5)879 (3.6)273 (1.3)24,265
1992–199624,471 (98.7)0 (0)0 (0)0 (0)343 (1.3)24,804

A wider population enquiry was not made and it is possible that a very ill infant born at home could have been admitted to a neonatal unit in a neighbouring health authority. As the proportion of home births is very small the calculations of incidence were based on the assumption that this was negligible.


Statistical analysis used both χ2 tests and odds ratios. Odds ratios were calculated to identify the relative risk of hypoxic-ischaemic encephalopathy between time periods.


Incidence of hypoxic-ischaemic encephalopathy

The total number of affected infants was 48 in period C compared with 189 and 112 in periods A and B, respectively (Fig. 1). The overall incidence of hypoxic-ischaemic encephalopathy in term infants for the latest study period was 1.9 per 1000 live births of all gestations compared with 7.7 and 4.6 for periods A and B respectively. This fall in total incidence from period B to period C was statistically significant (P < 0.001; OR 0.42 CI 0.29–0.59).

Figure 1.

Total incidence of hypoxic-ischaemic encephalopathy according to Grade, and incidence of stillbirths.

Twenty-nine infants were moderately to severely affected in the latest study period compared with 63 and 44 in periods A and B, respectively. The fall in incidence of moderately to severely affected infants (Grades 2 and 3) showed a downward trend (P < 0.1 OR 0.64 CI 0.39–1.05) when comparing periods C (1.2 per 1000 live births) and B (1.8 per 1000 live births) and a significant fall when comparing the periods C and A (2.6 per 1000 live births, P < 0.001: OR 0.46 CI 0.29–0.72) (Fig. 1).

The number of stillbirths has fallen steadily during the time course of this study (280 in period A, 138 in period B and 104 in period C). The incidence of stillbirths has dropped from 11.3 per 1000 total births in the first period studied to 4.2 in the latest.

Other factors

No differences were found compared with the previous study in relation to maternal age. It was again noted that hypoxic-ischaemic encephalopathy was more common in males (64%). It was also noted that only 18% of affected infants were > 41 weeks of gestation compared with 38% and 48% in the previous two study periods (A and B, respectively).


During the 21 years spanning this study the overall rates of assisted vaginal delivery have not changed significantly but the caesarean section rate has nearly doubled from 7.6% to 14.6%. Those with moderate to severe (Grades 2 and 3) encephalopathy were less likely to have been born by emergency caesarean section or instrumental delivery. In period C 16/19 (84%) of infants with Grade 1 hypoxic-ischaemic encephalopathy were born by assisted vaginal delivery or caesarean section. This compares with 22/29 (75%) for the infants with moderate to severe encephalopathy. In the previous study periods the assisted vaginal delivery/caesarean section rate was also higher in the Grade 1 group compared with those with moderate to severe encephalopathy1. For those with Grades 2 and 3 encephalopathy there has been no change in the assisted vaginal delivery rate but the section rate has risen. All caesarean sections performed on affected infants in the most recent study period were carried out as an emergency.

Cardiotocography was used in all the infants who subsequently developed hypoxic-ischaemic encephalopathy in the 1992–1996 study period. Fetal distress was given as the indication for all assisted vaginal deliveries and all caesarean sections in those who subsequently developed Grade 1 hypoxic-ischaemic encephalopathy. Of those with Grades 2 and 3 encephalopathy delivered by caesarean section, fetal distress was also recognised in all of them in comparison with fewer than 50% (3/7) who had an assisted vaginal delivery.


Eighty-five percent of infants with hypoxic-ischaemic encephalopathy in the present study period had a depressed response at birth (Apgar score 0.3 at 1 min) compared with 67% and 52% in the previous two periods (A and B, respectively). It had been noted in the earlier study that infants may suffer severe cerebral insult before birth and yet show satisfactory response at birth but this was not confirmed in this study.


Overall the number of deaths fell from 22 (92%) to 12 (80%) to 7 (58%) over periods A, B and C respectively (Fig. 2). The total number of infants who subsequently developed cerebral palsy has fallen from 11 to six to two in the same periods.

Figure 2.

Outcome of infants with Grades 2 & 3 encephalopathy in three five year periods. ▪ dead; ▪ abnormal; □ normal.

A comparison of outcomes for Grade 2 and Grade 3 hypoxic-ischaemic encephalopathy for each study period is also shown in Fig. 2. Those that were classified as abnormal had a severe motor disability with definite signs of cerebral palsy at 18 months. Of those with Grade 3 hypoxic-ischaemic encephalopathy in period A, there were only two survivors and both were severely disabled. In period B there were three survivors of whom one had spastic quadriplegia with epilepsy. In period C there were five survivors; one child survived intact, two were severely disabled with spastic quadriplegia, one had an Erbs palsy secondary to shoulder dystocia at birth but was otherwise normal and one had sensorineural hearing loss but no other neurological problems.

Of those infants who had Grade 2 encephalopathy in period C, five had problems identified at follow up but none had a severe motor disability (one Erbs palsy, one epilepsy and a squint, one high tone deafness, one global developmental delay but no evidence of cerebral palsy and one mild speech delay). In period A, nine (23%) had cerebral palsy, of whom six had a spastic quadriplegia, one a diplegia and one a mild hemiplegia. In period B, five (17%) had cerebral palsy, of whom four had a spastic quadriplegia and one had a mild hemiplegia.

Of those with Grade 1 encephalopathy, one child in the recent study period had mild nonpersistent upper motor neurone signs in her lower limbs. In the previous two study periods all were normal at follow up.



This study shows a further significant fall in the incidence of hypoxic-ischaemic encephalopathy in term infants in the period 1992–1996 compared with the earlier study periods. In addition there was a statistically significant difference in the incidence of those moderately to severely affected (Grades 2 and 3) when comparing periods A and C, and a downward trend when comparing B and C. A decline in hypoxic-ischaemic encephalopathy has been reported throughout the whole Trent region over the period 1992–1997 by Field et al.7. From an international perspective, an incidence of hypoxic-ischaemic encephalopathy of 1.8 per 1000 live births has been reported in Sweden8 and an incidence of 3.8 per 1000 term live births for moderate or severe encephalopathy in Australia9. The decline in incidence in hypoxic-ischaemic encephalopathy should be seen against a background of falling stillbirth and neonatal mortality rates during the study period.

The aetiology of hypoxic-ischaemic encephalopathy may be heterogeneous with both antepartum and intra-partum events identified as risk factors9,10. Many factors could account for the decline in hypoxic-ischaemic encephalopathy including improved socio-economic status, better antenatal care and other demographic factors together with changes in the organisation of services and in obstetric, paediatric and anaesthetic practice. The methodology however does not allow specific conclusions to be drawn.

Changing practice

Organisation of delivery services

Since 1976 the organisation of local maternity services has altered with over 98% of births now taking place in the City Hospital Maternity Unit (Table 1). The remainder (1.3%) are born at home and the proportion of these has not changed. There have been major improvements in obstetric staffing levels.

Obstetric practice

During the 21-year time span studied considerable change has taken place in obstetric practice. The caesarean section rate has risen and the use of cardiotocography is now widespread. The limitations of cardiotocography in identifying infants at risk of asphyxia are well recognised11,12. The interpretation of cardiotocographs by midwifery and obstetric staff is also an area for concern and was highlighted by the recent CESDI report5.

In the latest study period, all of those who subsequently developed Grade 3 hypoxic-ischaemic encephalopathy had an abnormal cardiotocograph. This was a higher detection rate than the earlier study periods. 11 of 12 (92%) had an operative delivery. These deliveries were often associated with obstetric catastrophes such as cord prolapse or placental abruption. The remaining infant was born by unassisted delivery which was complicated by shoulder dystocia. The implication of this may be that in spite of immediate intervention adverse outcomes were not prevented.

A negative association of newborn encephalopathy with elective section has been reported10. None of the sections performed on babies with hypoxic-ischaemic encephalopathy in this latest study period were carried out electively.


In 1976 when the study began there was no neonatal consultant and only one paediatric senior house officer. The first neonatal consultant was appointed in 1978 and by 1987 there was full 24 hour cover by senior house officers and at middle Grade level. In the last four months of the study the Derbyshire Children's Hospital moved to the City Hospital site resulting in paediatric middle grade staff being on site 24 hours a day.


The first specialist obstetric anaesthetist was appointed in 1978. Since then there has been a steady increase in the anaesthetic commitment to obstetrics, so that there is now continuous consultant cover for the unit. A dedicated resident on call anaesthetist for the labour ward has been provided since 1983, and since that time there has been a 24 hour service for epidurals. Anaesthetic practice has changed progressively over the period, with an increase in the epidural rate, and over the last few years a move from general anaesthesia to spinal anaesthesia for obstetric operations has occurred.


The death rate associated with hypoxic-ischaemic encephalopathy has continued to fall throughout the time course of this study. Disability was confined to those with Grade 2 or 3 hypoxic-ischaemic encephalopathy in all the study periods and the numbers of those with cerebral palsy has fallen. The improving prognosis of those infants with Grade 3 hypoxic-ischaemic encephalopathy could reflect a tendency to ventilate infants more readily and hence lead to them being reclassified from Grade 2 to 3 but would not affect overall figures for Grades 2 and 3 together.

Two children in the latest study period (one Grade 2, one Grade 3) were noted to have sensorineural hearing loss but no other disability. Others have reported sensorineural hearing loss in survivors of severe birth asphyxia13. It appears to be a plausible but uncommon association. Other risk factors would need to be excluded.


We are grateful to Professor S. Arulkumaran, Dr J. Collier, Dr L. Dann, Mr J. Hodgkins and Dr J. McIntyre for their helpful comments.