Article Text



Statistics from


P. O. D. Pharoah. FSID Unit of Perinatal and Paediatric Epidemiology, Department of Public Health, University of Liverpool, Liverpool L69 3GB

Background: Twins are at greater risk of perinatal death and serious morbidity than singletons and, within twins, monozygous (MZ) are at greater risk than dizygous (DZ) twins.

Aim: To quantify stillbirth and infant mortality and serious morbidity in MZ and DZ twins and to propose a hypothesis that unifies the increased risk in MZ twins and explains the pathogenesis of several congenital anomalies and cerebral palsy.

Methods: Analysis of all twin birth and death registrations in England and Wales 1993–2000. Survey of surviving twins born in England and Wales 1993–1995. Weinberg’s rule applied to mortality data to estimate number of MZ and DZ twins. When it is inappropriate to apply Weinberg’s rule, like are compared with unlike sex twins as a partial proxy for zygosity.

Results: The relative risk (RR) for MZ compared with DZ was 15.1 for both twins to be stillbirths, and 1.8 for one twin to be a stillbirth. If only one twin was a livebirth, infant mortality was 148.3 in MZ and 66.5 in DZ, a RR of 2.2. When both twins are livebirths, MZ have a significantly increased risk of being born of very low birthweight (<1500 g) (RR 2.5). Among livebirths, like compared with unlike sex twins (as a partial proxy for MZ) are at significantly increased risk of cerebral palsy (CP), congenital cardiac valve atresias, and stenoses, bowel atresias, and isolated hydrocephalus.

Conclusion: Fetal damage in a MZ conception may range from early, mid, or late fetal death, infant death through to severe morbidity. Any combination of this range may occur in both of a twin pair, for example one early fetal loss + one congenital anomaly will present as the congenital anomaly in a singleton; CP in one twin and a congenital anomaly in the other, and so on.

Hypothesis: The damage is ischaemic and is attributable to haemodynamic instability in a monochorionic MZ conception. The location (organ or part of an organ), severity, and timing of the ischaemic episode may affect either or both twins and accounts for the increased mortality risk and the pathogenesis of cerebral palsy and several congenital anomalies.


E. S. Draper1, J. Zeitlin2, P. Truffert2,3, B. N. Manktelow1, D. J. Field1. 1University of Leicester, UK; 2INSERM U149, Paris, France; 3NICU, Maternité Jeanne de Flandres, Lille, France

Background: Although births <33 weeks gestation only account for up to 2.0% of all births, they constitute more than 30% of perinatal deaths. The organisation of obstetric and neonatal care for these births has an affect upon their survival. This study compares the characteristics and mortality patterns of preterm births in European regions with similar demographic characteristics, but distinct approaches to the organisation of perinatal care.

Methods: All preterm births 22–32 weeks of gestation born in Trent and NPC were studied in 1997. Total births for these regions were 55 633 and 54 810, respectively. In Trent there are 19 maternity units;16 have an associated neonatal unit, whereas in NPC there are 50 maternity units with 15 neonatal units. Data from the Trent Neonatal Survey (an on going data collection for babies admitted for NIC) and CESDI were merged for the Trent region. In NPC, data were from the EPIPAGE study, a population based study of all births <33 weeks gestation in nine regions in France.

Results: Trent had a 40% higher very preterm birth rate (1180 v 824 births). Mortality was higher in NPC. Mortality differences were most marked for antepartum, intrapartum fetal, and labour ward deaths: 27.4% in NPC v 18% in Trent. The organisation of perinatal care partially explained these differences in mortality. Stillbirth and labour ward death rates were higher in maternity hospitals without on site neonatal units (level I units). In both regions, birth in level II units was associated with higher intrapartum and labour ward deaths, but with higher in-hospital mortality in NPC only.

Conclusions: Despite similar standards of living and access to health care, these two regions had strikingly distinct very preterm populations. The greatest disparities between regions and levels of care were observed for antepartum, intrapartum, and labour ward deaths. Obstetric practices or aetiological differences in preterm births may account for this.


S. V. Glinianaia, C. Wright, J. Rankin, S. Sturgiss, M. Ward-Platt, U. Wariyar, M. Renwick. School of Population and Health Sciences, Faculty of Medical Sciences, University of Newcastle, and Multiple Pregnancy Register Steering Group, Regional Maternity Survey Office, Newcastle upon Tyne, UK

Introduction: The continued increase in multiple birth rates over the past decade has raised concern about their rising contribution to fetal and infant mortality and to childhood disability. This retrospective cohort study examines rates of early and late fetal and neonatal losses using population based data from a Northern Multiple Pregnancies Register.

Methods: Multiple pregnancies are notified to the register as soon as they are detected, either antenatally or perinatally, irrespective of whether they resulted in a spontaneous abortion, termination of pregnancy, or registered birth. The register is unique within the UK and in Europe, in that the surveillance of twin pregnancies begins in the first trimester.

Results: A total of 1818 twin pregnancies were recorded during 1998–2001; 1686 of these resulted in at least one live birth or stillbirth, giving a twinning rate of 14.0 per 1000 maternities. Nearly 65% of twin pregnancies were detected before 13 weeks of gestation. Chorionicity was determined in 84.4% of 1686 twin maternities with at least one stillbirth or live birth. The fetal loss rate before 24 weeks of gestation was 95.2 per 1000 twins (346/3636). The stillbirth (fetal loss at or after 24 gestational weeks) and infant mortality rates were 16.9 per 1000 twin births and 28.8 per 1000 twin live births, respectively. Monochorionic twins are at higher risk of stillbirth that dichorionic: 45.3 v 10.1 per 1000 births (RR = 4.5, 95% CI 2.6 to 7.8).

Conclusion: Data from the Northern Multiple Pregnancy Register show that nearly 10% of registered twins have been lost before 24 weeks of gestation. Twins, in particular monochorionic, are also at high risk of stillbirth and infant mortality. The Multiple Pregnancy Register not only allows monitoring of trends in multiple birth rates and mortality, but also aetiological research and long term follow up studies.


L. Parker, S. V. Glinianaia, J. Rankin, R. Bell, M. S. Pearce, C. Wright. Schools of Clinical Medical Sciences and Population and Health Sciences, Faculty of Medical Sciences, University of Newcastle, Newcastle upon Tyne, UK

Background: Perinatal mortality is affected by a number of risk factors including birthweight, maternal age, plurality, and sex. The distribution of these factors has changed over the past 20 years, thereby potentially influencing observed rates of perinatal mortality.

Aim: To examine the contribution of changes in risk factors in the birth population to observed trends in perinatal mortality.

Methods: Data on all perinatal deaths (stillbirths occurring after 28 week’s gestation and neonatal deaths) during 1982–2000 were obtained from the Northern Perinatal Mortality Survey, a population based survey of all perinatal and infant deaths in the Northern Region. Perinatal mortality rates were directly standardised for 2000, based on the birth population structure for 1982.

Results: Perinatal mortality declined by 20% between 1982 and 1990 and 1991 and 2000. There was an increase in LBW babies, in births to mothers aged >30, and in twin births, but no change in sex distribution. Standardising individually for these risk factors, in particular for birthweight, lowered perinatal morality rates. The adjusted perinatal mortality rate was further reduced when standardisation considered all factors combined (unadjusted 7.9 per 1000 births; adjusted 6.8 per 1000). The most frequent causes of perinatal death were antepartum anoxia, congenital malformation, and prematurity. Standardising for population risk factors reduced the cause specific rates by 13% for antepartum anoxia, by 21% for congenital malformation, and by 43% for prematurity.

Conclusions: The demographic characteristics of our population have changed over the past 20 years in important respects, which influence the population risk of stillbirth and neonatal death. If birth population characteristics had remained unchanged, the observed mortality rates would be considerably lower. Standardising for birthweight had the greatest effect, although changes in maternal age and multiple births have also influenced trends in observed mortality. Furthermore, adjusting for risk factor profile differentially affects rate of perinatal death by cause.


S. Sinha, A. V. I. Soler, T. Lacaze-Masmonteil, J. Gadzinowski, R. Segal, R. D’Agostino, T. Wiswell, J. Massaro. The James Cook University Hospital, Middlesbrough and 22 other Participating Centres in Europe and North America

Objective: To compare the safety and efficacy of the novel surfactant Surfaxin®, which contains KL4-peptide, against the predominately used animal derived (porcine) surfactant, Curosurf®, in the prevention of RDS.

Methods: A masked, international, multi-center, randomised, controlled, non-inferiority trial with 22 participating centres. Inclusion criteria: 1) GA 24 to <29 weeks; 2) birthweight 600–1250 g; 3) successful intubation at birth; and 4) informed consent. Primary outcome: incidence of being alive without bronchopulmonary dysplasia (BPD) at day 28. Secondary outcomes: including mortality, air leak, BPD (36 weeks), PVL, IVH, PDA, NEC, ROP, pulmonary haemorrhage, sepsis, and neurodevelopmental and respiratory outcome at 6 and 12 months. The study was approved by individual IRBs and monitored by an international Data Safety Monitoring Board.

Results: The baseline demographics of the babies enrolled in two groups (Surfaxin = 119, Curosurf = 124) were similar (BW 929 v 937 g, GA 27.0 v 27.1 weeks), while a similar proportion received antenatal steroids (~85%). There were no significant differences in the frequency of oxygen desaturation, bradycardia, or apnea during surfactant administration. There were also no significant differences in any of the outcome parameters as listed in the table, or, in the incidence of PDA, NEC, or ROP. Neurodevelopmental and respiratory outcomes (at 6 and 12 months, PMA) are currently being assessed.

Abstract G5

Conclusions: Surfaxin is as effective as Curosurf and appears to be just as safe.


P. Clarke1, S. J. Mitchell1, R. Wynn2, S. Sundaram3, V. Speed4, E. Gardener1, S. Farrimond2, D. Roeves5, M. J. Shearer5. 1Hope Hospital, Salford; 2Royal Manchester Children’s Hospital, Manchester; 3Royal Bolton Hospital, Bolton; 4Billinge Hospital, Wigan; 5Haemophilia Reference Centre, St Thomas’s Hospital, London, UK

Background: Neonatal vitamin K stores are precarious. Without vitamin K prophylaxis preterm infants may be at particular risk of vitamin K deficiency bleeding (VKDB), but the optimal dose and route is unclear.

Aim: To compare the vitamin K status of preterm infants during the first week of life and when established on full enteral feeds, following three regimes of vitamin K prophylaxis after delivery.

Methods: Infants born at <32 week’s gestation were randomised to receive vitamin K1 0.5 mg intramuscularly (IM) (group 1), 0.2 mg IM (group 2) or 0.2 mg intravenously (IV) (group 3) on day 1. Blood samples taken at 5 days postnatal and 2 weeks after establishment of full enteral feeds were analysed for plasma vitamin K, prothrombin time (PT), and prothrombin concentration.

Results: Of 73 infants enrolled, 69 had a day 5 blood sample. Full enteral feeds were established at a median (IQR) of 10 (8–18) days, and 60 infants completed the study with a second blood sample taken at 25 (22–32) days. Baseline characteristics (mean (SD) for groups 1, 2, and 3, were, respectively: gestational age 28.0 (2.5), 28.4 (2.4), and 27.9 (2.6) weeks; birthweight 991 (387), 1091 (375), and 1054 (352) g. Median (range) vitamin K1 concentrations (ng/mL) are shown in the table. Concentrations did not differ between groups on day 5 (p = 0.14), but were significantly higher in group 1 when established on feeds (p = 0.04). Two infants in group 3 had undetectable levels when on full breast milk feeds.

Abstract G6

Conclusions: Preterm infants given 0.2 mg or 0.5 mg vitamin K at birth generally have very high plasma concentrations during the first week of life. With a reduced dose (0.2 mg) given IV, vitamin K can fall to undetectable levels in breast fed infants and this may increase their risk of late VKDB.


C. Snowdon, D. Elbourne, J. Garcia. Medical Statistics Unit, LSHTM and 5 UK neonatal units

Background: Postmortem (PM) rates are declining worldwide. In the UK, perinatal pathology has been particularly undermined by recent controversies. Previous research has not considered the potentially serious consequences for randomised controlled trials that include pathology studies.

Methods: 5 perinatal pathologists responded to brief questions to inform the qualitative study which was based on semi-structured interviews with 26 neonatologists linked to the CANDA and/or INNOVO neonatal trials, and with 18 bereaved parents whose babies were enrolled in one of these trials.

Results: Pathologists expressed concern over the difficulties experienced in UK perinatal pathology and the impact on research of inadequate levels of samples. Interviews with neonatologists revealed some discomfort over approaching bereaved parents for PMs. Neonatologists were concerned to protect vulnerable parents and varied in whether they saw this as compatible with inclusion in trial related pathology studies, especially where the cause of death seemed apparent. Interviews with parents demonstrated a range of reactions from being clear that they did not want a PM to feeling that they needed the information from the PM, as well as parents who were initially discomforted but who then made the decision to go ahead. Parents who elected to have a PM did so for their own needs, or to contribute altruistically to a trial, or for both reasons. Raising the subject was generally not seen as inappropriate and no parents stated that they felt pressured into their decision.

Conclusions: The study provides support for a careful approach, but also suggests that for some parents the degree of caution and selectivity exercised by the neonatologists may not be entirely appropriate and that it may be possible to approach more parents without undermining their wellbeing. Research is needed to document and gain an understanding of the consent process and to examine the usefulness and impact of different consent forms. It should assess whether professionals might benefit from training in the skills which go along with helping parents to come to their decisions.


G. M. Taylor1, F. E. Alexander2, S. W. D’Souza3. 1University of Manchester, Cancer Immunogenetics, St Mary’s Hospital, Manchester; 2Public Health Sciences, University of Edinburgh; 3Academic Unit of Child Health, St Mary’s Hospital, Manchester

Background: Smoking during pregnancy is an important cause of low birthweight.1 It is thought to cause childhood asthma and contributes to adult diseases, including cardiovascular disease and type 2 diabetes. Low birthweight due to smoking may be influenced by polymorphisms in certain maternal genes,2,3 but there are no data on interactions between fetal genetic polymorphisms, maternal smoking, and birthweight. However, certain HLA-DQ alleles are associated with birthweight in children with juvenile onset diabetes.4

Aims: To determine whether fetal HLA-DQ alleles affect birthweight due to smoking during pregnancy.

Methods: We studied an unselected series (n = 552) of full term singleton newborns of white UK Caucasian parents. Birthweight and smoking data were abstracted from the case notes. HLA-DQ typing was carried out on DNA extracted from cord blood samples using standard molecular typing methods. Mean birthweight in relation to DQ type was calculated for newborns of smoking (n = 211) and non-smoking (n = 341) mothers. Statistical comparisons were adjusted for birthweight confounding factors.

Results: Maternal smoking reduced mean birthweight by 243 g. Multivariate analysis (correcting for confounding variables) showed that newborns with DQA1*0101 and DQB1*0501 had further reductions (in addition to smoking alone) of 229 and 239 g, respectively, due to interaction with maternal smoking. In contrast, newborns with DQA1*0201 and DQB1*0201 were “protected” from the effects of maternal smoking by 269 g and 293 g, respectively. There was no evidence that fetal DQ alleles were associated with a change in birthweight in the absence of maternal smoking.

Conclusions: Certain fetal HLA-DQ alleles interact with smoking to increase or decrease smoking induced effects on birthweight. Fetal DQ type might be useful in predicting morbidity in the offspring of smoking mothers.






J. H. Simpson1, F. L. R. Williams1, C. Delahunty1, H. Van Toor2, S. -Y. Wu3, T. J. Visser2, R. Hume1 with collaboration from the Scottish Preterm Thyroid Group1. 1Tayside Institute of Child Health, University of Dundee; 2Erasmus University, Rotterdam; 3University of California, California

Aims: To clarify developmental trends and inter-relationships between thyroxine (T4), free thyroxine (FT4), thyroxine binding globulin (TBG), thyroid stimulating hormone (TSH), triiodothyronine (T3), reverse triiodothyronine (rT3), and thyroxine sulphate (T4S) levels in umbilical cord and fetal blood sera (n = 710, 15–42 weeks gestation), and to correlate infant levels (23–42 weeks gestation) to maternal values (n = 454, 16–45 years) and to those of non-pregnant women (n = 252, 16–46 years).

Methods: The study encompassed four distinct groups: 1) a cohort of mothers and infants delivered at 23–34 weeks gestation who were part of a multi-centre study; 2) a consecutive sample of mothers and infants delivered at 35–42 weeks gestation in a single centre at Ninewells Hospital, Dundee; 3) a cohort of fetuses 15–20 weeks gestation following therapeutic termination of pregnancy in a single centre at Ninewells Hospital, Dundee; and 4) a cohort of non-pregnant women, recruited from Dundee.

Results: Full term umbilical cord sera TSH, TBG, and all iodothyronine levels, except T3, are higher than non-pregnant women. Maternal third trimester T4, T3, and TBG levels are higher, and FT4, TSH, rT3, and T4S levels and T4:TBG ratios are lower than corresponding infant umbilical cord values. In umbilical cord and fetal sera T4, T3, and TBG levels increase with gestation until term; T4S, rT3 levels increase and peak in the late second early third trimester and then decline to term; and T4:TBG ratios increase with gestation until late second trimester and plateau to term.

Conclusions: The late second early third trimester is a critical transition period in fetal thyroid hormone metabolism, which is interrupted by premature birth and may contribute to postnatal thyroid dysfunction.


C. Tomlinson1, H. McDevitt2, S. F. Ahmed2, M. White2. 1Princess Royal Maternity Hospital, Glasgow, UK; 2Bone and Endocrine Research Group, Yorkhill NHS Trust, Glasgow, UK

Introduction: Quantitative ultrasound (QUS) assessment of speed of sound (SoS) is increasingly being used in adults, and to a lesser extent in paediatrics, as a marker of bone health. In this study we assessed the change in speed of sound in VLBW infants over time using the Omnisense 7000P. The software with the scanner provides an absolute measure of bone SoS and a z score showing variation from the mean value for patient age.

Methods: Infants were eligible for the study if they were born ⩽32 weeks gestation, <1500 g birthweight and parents consented to the study. 24 infants with a median gestational age of 28 weeks (range 24–32) and median birthweight 1080 g (625–1500) were studied. Both tibia were scanned shortly after birth and weekly thereafter until the infant reached term corrected age or were discharged.

Results: The initial SoS measurement for each infant was within the range expected for preterm infants (mean z score = 0.01, range −1.5 to 1.3). There was a correlation between gestational age and SoS (r = 0.77) but no correlation with birthweight (r = 0.38) or sex (p = 0.89).

Conclusion: The SoS from 24 to 40 weeks gestation is expected to increase from ~2750 to ~3100 m/s. However, in our group of patients the SoS did not increase after delivery and in fact fell. The mean fall in SoS from admission to discharge of 70 m/s (SD 81 m/s) was statistically significant (p<0.01). For those infants born at 24–26 weeks gestation the maximal fall in SoS occurred at ~10 weeks of age compared with ~4 weeks for more mature infants (p<0.01).

Conclusion: In this group of VLBW infants the initial SoS is within the expected range compared with the manufacturers database. However, rather than increasing after birth, the SoS was found to be lower at discharge than the initial measurement. Further studies are needed to establish QUS as a tool for assessing bone mineralisation in preterm infants and the abnormal trajectory of SoS values in these infants.


M. G. Gnanalingham, A. Mostyn, M. E. Symonds, T. Stephenson. Centre for Reproduction and Early Life, Institute of Clinical Research, University Hospital, Nottingham NG7 2UH

Introduction: Brown adipose tissue specific uncoupling protein (UCP)-1 mRNA peaks at the time of birth, before rapidly declining, and is then normally no longer detectable at any other stage of the life cycle in large mammalian species. UCP2 is a member of the inner mitochondrial protein superfamily, whose exact function is not known. Postulated roles include thermogenesis like UCP1 and fluid exchange within the lung. The extent to which UCP2 mRNA in the lung may exhibit a similar ontogenic pattern of expression to UCP1 is not, however, known.

Methods: Lung samples (n = 4–7) were taken from fetuses at 140 days (term ≅ 147 days), from lambs at 4–6 hours, 7 and 30 days, and from adolescent lambs at 6 months, after humane euthanasia. Total lung RNA was extracted, reverse transcribed, and UCP2 mRNA abundance measured by RT-PCR using oligonucleotide primers designed specifically to ovine UCP2. Results are given as means SE in arbitrary units, as a ratio of 18S rRNA and are expressed as a percentage of a reference sample. Differences between groups were analysed by one way ANOVA with post hoc Bonferroni.

Results: See table. UCP2 mRNA was highly abundant in the 140 day fetal lung, peaking at 4–6 hours after birth and then gradually declining by one month of postnatal age, and was barely detectable at 6 months of age.

Abstract G11

Conclusions: UCP2 mRNA abundance in the lung shows a similar ontogeny to brown adipose tissue UCP1, peaking soon after birth, and is very low in older animals. These findings suggest a critical role for UCP2 at birth. This may include thermogenesis, lung fluid clearance, and/or establishment of independent breathing in the newborn.


J. Tooley, M. Thoresen. Department of Clinical Science, South Bristol (Child Health), University of Bristol

Background: Monitoring of cardiovascular status in the extremely unwell baby is limited to the continuous measurement of blood pressure (BP). There is concern that BP values do not accurately reflect end organ perfusion. The measurement of cardiac output (CO) (for mainly technical reasons) has not been possible, except intermittently via echocardiography. Pulse contour analysis of the arterial BP waveform1 has been shown in adults to be accurate in determining stroke volume (SV), systemic vascular resistance (SVR), and CO.

Aim: To determine whether continuous measurement of CO using pulse contour analysis is possible in the 1.5 kg piglet, and to determine whether CO is a more effective marker of tissue perfusion than BP.

Methods: 10 newborn piglets underwent anaesthesia, intubation, and UAC insertion for the measurement of invasive BP followed by a 45 min hypoxic insult by reducing the inspired O2 to ~5%. Continuous measurement of CO, SV, and SVR was derived by real time analysis of the arterial wave form trace (PulseCo, Lidco Ltd, UK). Perfusion was measured in the thigh muscle using laser Doppler perfusion (Oxford Optronix Ltd, UK). Actual CO was also determined at baseline and the end of the study by lithium dilution (Lidco Ltd).

Results: Continuous recording of CO was possible in all animals. During hypoxia BP fell below baseline levels in all animals, However, cardiac output only fell in 5/10 animals. Tissue perfusion during hypoxia was more closely correlated with CO (R2 0.52 (IQR 0.46 to 0.56) than with BP (R2 0.24 (IQR 0.09 to 0.4). There was poor correlation between BP and CO (R2 0.22 (IQR 0.13 to 0.45).

Conclusions: Continuous non-invasive analysis of CO is possible in a 1.5 kg neonatal model using pulse contour analysis. CO has a closer relationship to tissue perfusion than BP. This methodology is applicable to the term and preterm infant in assessing cardiovascular function.



D. Shah1, M. Pavlovic2, D. Thompson2, H. Wang2, M. Kean2, T. Inder1,2. 1Royal Children’s Hospital, Melbourne, Australia; 2Howard Florey Institute, University of Melbourne, Australia

Aim: To gain macroscopic and structural information on the effect of premature birth and presence of white matter injury (WMI) on cerebellar development.

Method: 118 scans, 103 of ex-preterm at term corrected, and 15 of controls were reviewed initially. A subset of the full term (n = 7) and premature infants (n = 13) underwent model based segmentation of the cerebellum into gray matter, un-myelinated white matter, myelin, basal ganglia, and cerebral spinal fluid (CSF) was performed. An average was calculated for the cerebellum tissue volumes (CV). Diffusion tensor images were acquired using line scan. Quantitative measures of the water diffusion, apparent diffusion coefficient (ADC), and relative anisotropy (RA) were calculated for axial images at the vermis level for which two regions of interest (ROI) were placed in each hemisphere; one in the vermis and one in the white matter.

Results: On qualitative review, 36 of the 103 ex-preterm infants had evidence of cerebellar hypoplasia, 12 had decreased myelination, two had focal signal change, and one infant had focal haemorrhage. Of the 15 controls, one had evidence of cerebellar hypoplasia. On volumetric analysis, the preterm infants had lower mean cerebellar volumes than the term controls (20.1 mm3v 22.9 mm3) and the preterm infants with significant WMI had lower volumes than the preterm infants without. Preterm infants with WMI had lower RA and higher ADC values as compared with controls, corresponding with increasing RA values with brain maturation.

Conclusion: There is evidence of significant alteration in cerebellar structure at term in preterm infants, which may be caused by developmental abnormalities of fibres or destruction of fibre tracts as a result of WMI. Impaired cerebellar development may be associated with increased neurodevelopmental disability.


S. Victor, R. E. Appleton, M. Beirne, A. M. Weindling. Department of Child Health, University of Liverpool and Department of Neurology, Royal Liverpool Children’s Hospital NHS Trust

Introduction: Clinicians treat hypotension to maintain organ perfusion in sick preterm infants including that to the brain. However, the critical level of mean blood pressure (BP) below which cerebral perfusion is compromised has not been clearly defined.

Aim: To determine if EEG could be used to determine the critical level of BP for cerebral perfusion.

Methods: A prospective, observational study was performed on 34 very low birth weight infants. Infants with serious intra-ventricular haemorrhage and high cord base excess were excluded. BP (range: 16–46 mm Hg) was measured using arterial catheters. Four channel digital EEG recordings were performed for 1 hour on the first two days after birth and the recording with lowest BP was chosen for analysis. The EEG was analysed by (a) spectral analysis to give relative power (RP) of delta (0–3.5 Hz), theta (4–7.5 Hz), alpha (8–12.5 Hz), and beta (13–30 Hz) bands, and (b) manual calculation of median, 90th centile and percentage interburst interval. EEG was also reported qualitatively by two of the authors (MB and REA), who were unaware of the BP. Cerebral and peripheral fractional oxygen extraction were measured during each EEG recording.

Results: Infants had a median gestational age of 27 weeks (range: 23–30). Four infants had abnormal EEG (low amplitude and prolonged interburst intervals). The BP of all four infants was below 23 mm Hg. A cubic regression curve was fitted and values of RP of delta band (Rsq = 0.61; p<0.001) and interburst intervals (Rsq = 0.44, p = 0.001) were within normal limits at BP levels above 23 mm Hg. Cerebral and peripheral fractional oxygen extraction did not show any relationship to BP.

Conclusion: EEG may be a useful determinant of the adequacy of cerebral perfusion in very low birth weight infants. BP below 23 mm Hg was associated with an abnormal EEG. Cerebral perfusion may be maintained at BP levels much lower than currently accepted by clinicians.


K. J. Riley1, J. S. Wyatt1, D. Peebles2, M. De Haan1. 1Department of Paediatrics and Child Health; 2Department of Obstetrics and Gynaecology, Royal Free and University College Medical School, University College London, Gower Street Campus, 5 University Street, London WC1E 6JJ

Introduction: The preterm brain is vulnerable to injury in the perinatal period. The aetiology of the injury is multifactorial and predicting those infants at increased risk can be difficult. The aim of this study was to investigate the relationship between placental histology and neurodevelopmental outcome in a cohort of extremely low birthweight infants.

Methods: All infants with a birthweight of <1000 g admitted to the neonatal intensive care unit at University College London Hospital between January 1995 and December 2001 were recruited for long term neurodevelopmental follow up. An initial dataset which included placental histology results was collected while the infant was an inpatient. At 1 year of age (corrected for prematurity) all infants were recalled for neurodevelopmental assessment using the Griffiths Neurodevelopmental Assessment Scale.

Results: 290 infants fulfilled the recruitment criteria. Maternal placental histology results were available in 178. 35% had normal histology, 30% had histology suggestive of chorioamnionitis and/or funisitis, and 27% had evidence of placental hypoxia/ischaemia. There was a significant relationship between birthweight for gestational age and placental histology (p<0.001). Using fetal growth charts 55% of those with normal histology, 12% with histological evidence of infection, and 96% of infants with placental hypoxia/ischaemia had birth weights less than the tenth centile. There was no significant difference in the total developmental quotient on the Griffiths Developmental Scales between the histological groups (94.3 (SD 13.1), 95.4 (SD 15.1), and 96.5 (SD 15.1), respectively). Analysis of the sub-scales revealed similar findings and were not significantly different.

Conclusions: There is a significant relationship between placental histology and birthweight centile calculated from fetal growth charts. There is no significant difference in neurodevelopmental outcome between the histopathological groups.


T. Austin1, M. Noone1, M. Sellwood1, J. Henty2, J. H. Meek1, J. S. Wyatt1. 1Department of Paediatrics and Child Health, University College London; 2Department of Medical Physics and Bioengineering, University College London

Background: Doppler ultrasound is frequently used to assess cerebral haemodynamics in the newborn infant undergoing intensive care. The inability to measure the diameter of the cerebral arteries means that only non-quantitative information on cerebral blood flow velocity (CBFV) can be obtained using this technique. Various indices have been described, one of the most commonly used being Pourcelot’s resistance index (RI). In vitro studies have shown the RI to correlate inversely with flow.1 Quantitative measurements of cerebral blood flow (CBF) can be obtained using near infrared spectroscopy (NIRS).2

Aim: We aimed to determine the relationship between the RI in the anterior cerebral artery and global CBF in preterm infants by comparing paired measurements of RI and CBF.

Method: Sixteen preterm infants born at a median (range) gestational age of 24 (23–31) weeks were studied. The median age of study was 1.5 (1–43) days post-delivery. Measurements of Hb and HbO2 were made using a near infrared spectrophotometer and CBF was measured using the oxygen bolus technique.3 CBFV measurements were obtained from the anterior cerebral artery using an Acuson XP instrument and the RI calculated from analysing and averaging 5 waveforms. The data were analysed using Pearson Product Moment Correlation (SigmaStat, SPSS Inc).

Results: The median (range) RI was 0.8 (0.56–1.1) and CBF 11 (4–20) ml 100g−1 min−1. Three of the infants had evidence of absent or reversed end diastolic flow velocity associated with a patent ductus arteriosus (RI>1.0). There was no correlation between RI and CBF (correlation coefficient 0.04, p = 0.89).

Conclusion: Measurement of Doppler flow velocity in the anterior cerebral artery does not appear to provide clinically useful information about cerebral perfusion in preterm infants.



View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.