- Email: [email protected]
Cerebral and renal artery blood flow velocity before and after birth
Early Human Development 46 ( 1996) 165- 174
Cerebral and renal artery blood flow velocity before and after birth S.T. Kempley”,
S. Vyas, S. Bower, K.H. Nicolaides, H. Gamsu
Children Nationwide Neonatal Centre and Harris Birthright Reseurch Centrr fbr Fetrrl Mrdkinr. King’s College Hospital Medical School. London, UK
Received 16 January 1996; revised 29 March 1996; accepted 2 April 1996
Abstract Objective: To document perinatal changes in cerebral and renal artery haemodynamics 111 premature growth-retarded and normal term infants. Design: Longitudinal study of individual infants. Doppler ultrasound measurements of blood flow velocity (BFV) in the middIe cerebral and renal arteries were obtained before delivery, soon after delivery and during the first week of postnatal life. Setting: Teaching hospital obstetric and neonatal units. Subjects: 13 severely growth retarded infants born at 28-36 weeks gestation, and eight normally grown infants born at term. Results: In both groups, BFV in the cerebral artery was significantly lower in the first few hours after birth than in fetal life, but subsequently increased to reach pre-delivery values by the end of the first week. In contrast, BFV in the renal artery during the first postnatal day was not significantly different from fetal values, but it also increased during the subsequent week. In six of the preterm growth-retarded infants, fetal blood gases were measured in samples obtained by cordocentesis, and in these cases an increase in blood oxygen content at birth was documented. Conclusions: Cerebral artery BFV falls at birth and is relatively ION during the time that premature infants are at the greatest risk of developing periventricular haemorrhage. Keyword.?: Blood flow velocity (BFV); Cerebral artery; Renal artery: Preterm
*Corresponding author. Royal London Hospital, Whitechapel. London El IBB. UK 0378-3782/96/$15.00 0 1996 Elsevier Science Ireland Ltd. All rights reserved PII SO37X-3782(96)01754-9
166
S.T. Kempley et al. I Early Hurnun Development 46 (1996) 165-174
1. Introduction Intrauterine growth retardation due to placental insufficiency is associated with fetal hypoxaemia [l]. Doppler ultrasound studies have demonstrated that in such fetuses, blood flow velocity (BFV) in the cerebral circulation is increased [2], presumably to compensatefor the hypoxaemia and thus maintain cerebral oxygenation. Although in the managementof affected pregnancies,preterm delivery is often necessaryto prevent intrauterine death, such prematurity places the infant at increased risk of conditions such as respiratory distress syndrome and periventricular haemorrhage (PVH). Hambleton and Wigglesworth [3] suggestedthat PVH may be due to rupture of capillaries in the germinal matrix, the consequenceof acute increase in blood flow. However, there is evidence suggesting that haemorrhageoccurs primarily following episodes of hypotension [4,5] and therefore reduced cerebral blood flow. Doppler ultrasound studies of full term infants delivered vaginally have demonstrated that BFV in the middle cerebral artery falls following birth and subsequently increases during the first few days of life [6]. This suggeststhat during the first few days of life, when premature infants are at greatestrisk of PVH [7,8], cerebral artery BFV may be relatively low. In preterm infants BFV in all cerebral arteries increasesduring the first few days of life [9] but there are as yet no studies comparing fetal and neonatal cerebral BFV values in this group. The aim of this study was to determine whether changesin the cerebral circulation similar to those of term infants occur in growth-retarded premature infants, by measuring BFV before birth and during the first week of life. To determine whether any observed changesare specific to the cerebral circulation, we also measuredBFV in the renal artery. For some of these infants intrauterine cord blood gas measurements allow us to relate the changesin BFV to changesin oxygenation. The findings are compared with BFV measurements obtained from normally grown infants delivered at term by caesareansection. 2. Methods In 13 growth-retarded infants that were born at 28-36 (median 32) weeks gestation, serial Doppler ultrasound measurementsof BFV and pulsatility index (PI) were obtained from the middle cerebral and renal arteries. The Doppler studies were carried out within 72 h before delivery (median 22 h) and on four occasions after birth (within 12 h and 36 h after birth and on days 2-4 and 5-9). Not all babies were studied on all occasionseither becausethey were clinically unstable or becausethey died or were discharged from the unit. Patients were selectedsequentially from small-for-gestational age premature babies admitted to the neonatal unit following delivery by caesareansection undertaken becauseof deteriorating fetal or maternal condition arising in a pregnancy known to be complicated by fetal growth retardation (fetal abdominal circumference measured by ultrasonography < 25th centile for gestation). All had birthweights < 3rd
S.T. Kempley et al. I Early Human Development J6 (1496) I6*5%174
centile for gestation and 12 weighed < 1.5 kg. To be included, Doppler studies must have been performed both antenatally and postnatally, so that case selection was limited by the availability of obstetric and neonatal researchers. At the time of the postnatal Doppler measurements,mean intra-arterial blood pressure(measuredvia an umbilical catheter attached to a dome transducer, Medex, Rossendale,UK), and blood gases(ATL 300 blood gas analyzer,WA) were recorded. In six cases,prenatal blood gaseswere measuredin umbilical venous blood samples obtained by cordocentesis,which was performed immediately after the fetal Doppler measurements.Cordocentesis was performed to assist in clinical management and was not a planned part of this study. In these cases, blood oxygen content was calculated from the haemoglobin concentration and blood gas measurementsusing the dissociation characteristics of fetal haemoglobin. Values before and after delivery were compared. The values in the preterm group were compared to those from eight infants that were delivered at 37-41 weeks (median 38 weeks) by elective caesareansection for previous section, cephalopelvic disproportion, or breech presentation. All infants were normal and healthy and their birthweight was > 10th centile for gestation. Doppler studies were performed 3-7 h before delivery and again 3-7 h, 20-24 h and day 5 after birth, just before discharge from hospital. At the time of the Doppler studies, mean arterial blood pressurewas recorded using an oscillometric method (Dinamapp, Critikon, Ascot, UK) and oxygen saturation was measuredusing a pulse oximeter (Ohmeda Medical, Louisville, CO). Invasive measurementsof theseparameterscould not be performed for ethical reasonsas the subjectswere all well babies; the data are presented but have not been used for any calculated values. Approval for the studies in both the preterm and term groups was obtained from the hospital ethical committee and the parents gave informed consent. All Doppler studies were performed while the infants were asleep or lying quietly.
2.1. Dopplrr studies
Flow velocity waveforms were obtained from one of the middle cerebral arteries and from the left renal artery using range-gatedpulse-wave Doppler ultrasound. The placement of sample volumes was guided by colour-flow imaging for the fetal measurements,and by simple real-time imaging for the postnatal studies. Intensityweighted mean BFV and the PI (Systolic-Diastolic/Mean velocity: Gosling and King [lo]) were measuredand the average of at least four cardiac cycles was used. The intensity weighted mean velocity averagesthe varying velocities found between the centre and the periphery of a vessel during laminar flow, by integrating Doppler signal intensity and Doppler shift following fast Fourier transform. High PI indicates increasedvelocity waveform pulsatility which can result from either high downstream vascular resistance,or left-to-right shunting through the ductus arteriosus. An inverse relationship therefore normally exists between BFV and PI. Velocity was measured directly on-screen using inbuilt software and for PI a spectrum analyzer (Doptek 9000, Doptek, Chichester, UK) was used. Whenever the angle of insonation of the
168
S.T. Kempley et al. I Early Human Development 46 (1996) 165-174
vessel differed from 0”, it was measuredand the velocity was corrected for angle of insonation. Antenatal measurementswere performed by an obstetrician using a ~-MHZ linear array transducer (Acuson 128, Acuson, CA), and the postnatal measurementswere performed by a neonatologist using a ~-MHZ mechanical sector transducer (Sonos 100, Hewlett-Packard,Uxbridge, UK). Coefficients of variation between the operators and equipment were calculated from paired measurementsof middle cerebral artery BFV in each of 10 healthy newborns. The mean difference between the obstetric and the neonatal equipment was only 0.45 cm/s (2% of the measurement;t = 0.20, P = O.SS),and the standarddeviation of the error was 2.8 cm/s (14% of the measurement; this compares with a 12% intra-observer variation which we found previously for measurementsof postnatal cerebral artery BFV averaged over four cardiac cycles). 2.2. Statistical analysis
To compare antenatal BFV with the first postnatal measurement,data from each infant were only included for analysis if a measurementhad been obtained from the same artery both before and after delivery. Student’s paired t-test was use to determine the significance of any differences. The mean and 95% confidence intervals of BFV within each time period were then calculated, and unpaired t-tests used to determine the significance of any differences between the various periods. PI measurementswere significantly skewed, and paired Wilcoxon rank-sum test was used to determine the significance of differences in median values. Comparison of BFV values between preterm and term infants was by unpaired Student’s t-test, and comparison of PI values was by the Mann-Whitney u-test.
3. Results
In both the preterm and term infants, the mean middle cerebral artery BFV at the first postnatal measurementwas significantly lower and the mean PI was significantly higher than in fetal life (Table 1). Before delivery, the mean cerebral artery BFV in the preterm, growth-retarded fetuses (15.1 cm/s) was not significantly different from that of the term, appropriately grown fetuses (17.6 cm/s, t = 1.1, P = 0.30). There was no significant change at birth in renal artery BFV or PI in either group. Before delivery, the mean renal artery BFV in the preterm, growth-retarded fetuses (4.7 cm/s) was significantly lower than in the term, appropriately grown fetuses (9.7 cm/s, t = 3.1, P < 0.01). PI was significantly higher in the preterm, growthretarded fetuses (3.26 vs. 2.14, Mann-Whitney P < 0.001). For the infant who developed a parenchymal periventricular haemorrhage,cerebral artery BFV fell from 13.0 cm/s before delivery to 4.6 cm/s after delivery. This postnatal value was relatively low, but three other preterm growth-retardedbabies had lower values. None of the other infants displayed ultrasound evidence of periventricular haemorrhage or periventricular leucomalacia (PVL), although some were discharged back to their referring units before signs of PVL would have been apparent.
S.T. Kempley et al. / Early Human Development 46 (1996) 16.5-174
lb9
Table 1 Blood flow velocity and pulsatility index before delivery and at the time of the first postnatal measurement for the preterm, growth-retarded infants and the term infants Middle cerebral artery Antenatal
Renal artery Postnatal
Antenatal
Postnatal
4.7 (1.2) 9.7 (3.X)
6.3 (2.2) 9.5 (2.0)
3.26 (0.49)
3.63 13 1) 2.14 (0.98)
Mean (SD.) velocity (cm/s) Preterm IUGR Term
15.1 (5.1) 17.6 (4.8)
7.2 (3.5)***
10.7 (2.4)**
Median (interquartile range) Pukatility Index Preterm IUGR Term
1.48 (0.61) 1.47 (0.71)
2.18 (1.5)** 1.96 (0.85)*
Postnatal measurements which were significantly (*P < 0.05; **P < 0.01; ***p < 0.001).
2.14 (0.49)
different from the antenatal measurement are indicated
There were no statistically significant differences in antenatal or postnatal BFV between the growth-retarded infants who died and those who survived (Fostnatal MCA: 8.1 cm/s for survivors, 5.1 cm/s for deaths, P = 0.22; Postnatal Renal: 6.2 cm/s for survivors, 6.4 cm/s for deaths).
3. I. Postnatal trends
The mean and 95% CI of BFV and PI in the middle cerebral and renal arteries of the preterm and term groups during each time-period are shown in Fig. 1 and Fig. 2
Renal-Preterm
01 -1
I
I,,
0
1
2
3
,
/
I
I
4
5
6
7
Age (days)
0' -1
0
1
2
3
4
5
6
7
Age WW
Fig. 1. Blood flow velocity in the Middle Cerebral Artery (MCA, upper graphs) and in the Renal Artery (lower graphs) for the premature growth-retarded infants and the term infants. (Means and 95% CIs, with significant differences between antenatal and postnatal values shown below, and significant differences between the first and later postnatal values shown above the error bars; *P < 0.05; **I’ < 0.01 J.
170
S.T. Kempley et al. I Early Human Development 46 (1996) 165-I 74
MCA-Term
MCA-Preterm
I07 i T
I
;
0;
-1
Renal-Term
I
I,
0
1
2
Renal-Preterm
I
I,,
3
4
,
5
6
01
7
-1
0
1
Age (days)
2
3
4
5
6
7
Age (days)
Fig. 2. Pulsatility index in the Middle Cerebral Artery (MCA, upper graphs)and in the Renal Artery (lower graphs) for the premature growth-retarded infants and the term infants. (Medians and 95% CIs, with significant differences between antenatal and postnatal values shown below, and significant differences between me first and later postnatal values shown above the error bars; *P < 0.05; **P < 0.01). As would be expected, changes in PI were the inverse of changesin BFV. and the numbers of infants studied, their blood pressure and blood gasesare shown in Table 2. Middle cerebral artery BFV was significantly lower in growth-retarded infants at 12
Table 2 Numbers of infants studied and characteristicsat the time that the postnatal measurementswere performed <12 h
12-24 h
3 days
1 week
10 32
8 32
9 41
(29-47)
(19-50)
6 34 (34-38)
No. receiving mechanical ventilation
7.38 (7.13-7.53) 4.7 (3.6-11.1) 8.1 (5.2-11.6) 6
7.36 (7.17-7.44) 5.3 (4.1-8.3) 8.0 (5.3-l 1.6) 2
7.27 (7.24-7.34) 6.9 (4.9-7.3) 8.5 (8.0-12) 3
7.33 (7.27-7.35) 5.1 (5.1-5.6) 7.6 (5.1-8.1) 2
Term infants II Median (range)arterial blood pressure(mmHg)
8
8
39
41
Premature growth-retarded infants n Median (range)arterial blood pressure(mmHg) Median (range) arterial blood gases PH
Median (range) oxygen saturation (%)
(34-45)
(38-52)
94 (88-96)
94 (91-96)
-
(39-41)
6 47 (40-52)
93 (92-97)
S.T. Kempley et al. I Early Human LIeveIopment46 (1996) 16.5-174
1‘II
h of postnatal age, compared with the term infants (Fig. 1: Growth-ret 7.1 cm/s, term 10.7 cm/s, t = 2.23, P < 0.05). There was no significant ence between later measurements.At all ages renal artery BFV was significantly lower in the premature growth-retarded infants than in the term infants. By 1 week of age, middle cerebral and renal artery BFV had increaseds in both groups (Figs. 1 and 2). Changes in pulaatihty index were not sta&tic&ly significant, but were generally inversely related to the changesin BFV Exclusion of data from babies who died made no difference to these trends. 3.2. Comparison with changes in oxygenation For the infants who had cordocentesisperformed, antenatal and postnatal data are displayed in Table 3. Oxygen tension and blood oxygen content increased significantly after delivery (P < 0.01, P < 0.05, respectively). Mid&e cerebral @fery BFV fell significantly (P < 0.01). The fall in middle cerebral artery BFV was &set by a comparableincreasein oxygen content in cases3-6. In cases1 and 2 there were substantial falls in BFV (71% and 8256, respectively) but very little increase in oxygen content (8% and 20%). Both of these infants weighed less than 750 g; one was anaemic and the other was hypocarbic. They both survived with no ultrasound evidence of cerebral pathology. Table 3 Oxygen tension (PO,) and haemoglobin (Hb) for the infants who bad cordocentesisperfmmed Case no. (Birthweight, g) 1 (636) 2
(f5f36) 3 (788) 4 (804) 5
(836) (838) Antenatal mean (S.D.) Postnatal mean (SD.) Significance (F-value)
PO, @Pa)
Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal
Hb Wdl)
0, content
MCA BFV (cm/s)
4.4 7.5 4.7 8.9 1.7 8.8 3.3 11.1 2.8 8.0 3.9 6.9
10.3 9.8 15.2 15.4 14.5 15.1 15.8 16.1 14.1 14.0 14.8 17.8
5.0 5.4 7.3 8.8 1.8 8.7 5.9 9.5 4.8 8.2 6.6 10.0
22.6 6.5 15.‘7 2.9 14.3 4.2 7.6 4.9 20.7 11.1 9.1 5.1
3.5 (1.1) 8.5 (1.5) < 0.01
14.1
5.2 (1.9) 8.4
15.0
Gw 14.7 (2.7) NS
(1.6) < 0.05
(6.0) 5.8 (2.9) < 0.01
Mean Art. BP @m&T)
Pp, &Pa)
30 32 47 32 29
5.6 4.0 4.7 3.6 6.6 4.0 5.1 5.9 5.3 8.3 5.9 5.7 5.5 (0.7) 5.3 (l.Sj NS
AntenataI values are based on analysis of umbilical venous blood, and postnatal vahres on analysis of arterial blood. Oxygen content is calculated from the dissociation characteristics of fetal haernoglobin. using PO,, Pcoz, pH and haemoglobin values.
172
XT. Kempley et al. I Early Human Development 46 (1996) 165-l 74
4. Discussion This study has demonstratedthat in both growth-retarded preterm infants as well as normally grown term infants, velocity of blood flow in the middle cerebral artery is much lower on the first day of postnatal life than in the fetus. The fall in cerebral artery BFV at birth was accompaniedby an increase in pulsatility index. By the end of the first week, cerebral artery BFV had increased to a level close to that of the fetus. On the first postnatal day renal artery BFV was close to that of the fetus, but by the end of the first week it had risen considerably. All of the antenatal measurementswere performed before the onset of labour, to ensure that they representedthe ‘resting’ fetal state.This meant that all of the infants studied were delivered by caesareansection, and we cannot be sure of the changes which would follow a normal vaginal delivery. However, our findings were similar to those of Meerman et al. [6] who studied infants born vaginally at term. There are three possible explanations for the fall in cerebral artery BFV seen at birth, with the subsequentincrease over the next week. Firstly, it is possible that the reduced cerebral artery BFV on the first day does not represent reduced cerebral blood flow. Instead, it could be due to dilatation of the major cerebral arteries, with volume flow through the artery being close to fetal levels. We cannot exclude this possibility as it is impossible to measurethe diameter of these small vessels using ultrasound imaging. However, a generalised vasodilatation of proximal and distal cerebral arteries should be associatedwith a decreasein pulsatility index, and the PI actually increasedat birth (Fig. 2). Blood flow velocity in the cerebral arteries has been shown to be closely related to cerebral blood flow as measuredby Xenon clearance [ 111. Clearance techniques have also demonstrateda postnatal increase in cerebral blood flow which is similar to the increase in BFV found in this study [12] suggesting that the changesobserved probably do represent changes in volume flow rather than changes in vessel diameter. Secondly, the fall in cerebral artery BFV at birth could representa passive response of the cerebral circulation to changes in the amount of cardiac output available for cerebral perfusion, as a result of the complex changesin cardiovascular function and anatomy which occur at birth. Left ventricular output increasesat birth and falls over the next few days [13] the opposite of the changes in cerebral artery BFV we observed. However, the blood flow available to the fetus for organ perfusion is equal to the combined ventricular output, minus placental blood flow. Doppler studies of the human fetus suggest that at term, combined cardiac output is close to 450 ml/kg/mm, with a placental blood flow of 120 ml/kg/n& [14]. In the early neonatal period the blood flow available to the infant for organ perfusion will be equal to left ventricular output, minus the volume of left-to-right ductal shunting. In the term infant left ventricular output is in the region of 240 ml/kg/min immediately after birth, falling to 200 ml/kg /min over the next 3 days [ 151, with ductal shunting accounting for 60 ml/kg /min shortly after birth and decreasing rapidly thereafter [16]. It is therefore possible that less cardiac output is available for cerebral perfusion shortly after birth, but that the amount increasesrapidly during the first day as ductal closure occurs.
XT. Kempley et al. t Early Human Devetopmmt 46 (19%) 165-174
173
Finally, the fact that BFV fell in the cerebral, but not the renal arteries, suggests that these changescould be due to an increase in cerebral vascular resistance,rather than to a fall in cardiac output. The increase in cerebral Pi after delivery would support this view. Cerebral vascular resistance could have increased in response to the rise in arterial oxygen tension which occurred at birth. However, the later increase in middle cerebral artery BFV cannot be explained on this basis, as oxygen saturation did not change over the first week in either group (Table 2). Only if cerebral oxygen demand increases dramatically over the first few days of postnatal life could metabolic autoregulation account for this increase. For those growth-retarded infants in whom blood gases had been measured antenatally, we were able to calculate blood oxygen content before and after delivery. As none of these infants had been transfusedat this stage,it was safe to assumethat they had predominantly fetal haemoglobm, and the oxygen dissociation characteristics of fetal haemoglobin were used. We documented an increase in arterial oxygen content at birth, but the reduction in mean cerebral artery BFV was greater than the increase in mean arterial oxygen content. Oxygen delivery through an artery will be equal to the product of volume blood flow and arterial oxygen content. If the diameter of the vessel remains constant between fetal and postnatal life, then oxygen delivery would have decreasedin the two infants weighing less than 7.50g who had substantial falls in cerebral artery BFV with little or no increasein blood oxygen content (cases1 and 2). It is noteworthy that following delivery, some of these infants were hypocarbic, hypotensive or anaemic. In the other four infants, oxygen delivery would have stayed constant or increased. For the term infants it is unlikely that postnatal oxygen delivery would have been lower than that occurring in fetal life, as the fall in cerebral artery BFV was not as great. Periventricular haemorrhageis uncommon in the fetus, or after 1 week of postnatal age [7,8]. Our data suggestthat during the period of greatestrisk (i.e. the first 24-48 postnatal hours) cerebral artery BFV is relatively low. The fall in BFV at birth may result from an increase in cerebral vascular resistance,or from changesin available cardiac output, with consequent reduced cerebral blood flow. An increase in blood oxygen content may compensatefor the lower velocity, but this may not be complete in small or sick preterm growth-retarded infants. Biochemical markers of cerebral injury suggestthat perinatal events may be important in the etiology of PVH 1173and periventricular leucomalacia [ 181, even when these are not manifest until later. A major change in cerebral blood flow at birth could be the initiating event which culminates in later PVH. If elective delivery of the premature growth-retardedfetus is to lead to improved cerebral oxygenation, care should be taken to avoid hypocarbia, hypotension and anaemiain the first few hours after birth, as all of thesecould lead to a reduction in cerebral oxygen delivery to below fetal levels.
References [I] Nicolaides, K.H., Soothill, FYW.,Rodeck, C.H. and Campbell, S. (1986): Ultrasound-guidedsampling of umbilical cord and placental blood to assessfetal wellbeing. Lancet, i, 1065-1067.
114
S.T. Kempley et al. I Early Human Development 46 (1996) 165-174
[2] Vyas, S., Nicolaides, K.H., Bower, S. and Campbell, S. (1990): Middle cerebral artery flow velocity waveforms in fetal hypoxemia. Br. J. Obstet. Gynaecol., 97, 797-803. [3] Hambleton, G. and Wigglesworth, J.S. (1976) Origin of periventricuhu haemorrhagein the preterm infant. Arch. Dis. Child., 51, 651-659. [4] Miall-Allen, V.M., de Vries, L.S., Dubowitz, L.M.S. and Whitelaw, A.G. (1989): Blood pressure fluctuation and intraventricular haemorrhagein the preterm infant of less than 31 weeks gestation. Pediatrics, 83, 657-661. [5] Mehrabani, D., Gowen, C.W. and Kopelman, A.E. (1991): Association of pneumothorax and hypotension with intraventricular haemorrhage.Arch. Dis. Child., 66, 48-51. [6] Meerman, R.J., van Bel, F., van Zwieten, P.H.T., Oepkes, D. and den Ouden, L. (1990): Fetal and neonatal cerebral blood flow velocity in the normal fetus and neonate:a longitudinal ultrasound study. Early Hum. Dev., 24, 209-217. [7] De Crespigny, L., Mackay, R., Murton, L., Roy, R. and Robinson, P. (1982): Timing of neonatal cerebrovascularhaemorrhagewith ultrasound. Arch. Dis. Child., 57, 231-233. [8] Szymonowicz, W. and Yu, VY.H. (1984): Timing and evolution of periventricular haemorrhage in infants weighing 1250 g or less at birth. Arch. Dis. Child., 59, 7-12. [9] Evans, D.H., Levene, MI., Shortland, D.B. and Archer, L.N.J. (1988): Resistanceindex, blood flow velocity, and resistance-areaproduct in the cerebral arteries of very low birth weight infants during the first week of life. Ultrasound Med. Biol., 14, 103-110. [ 101 Gosling, R.G. and King, D.H. (1974): Arterial assessmentby Doppler-shift ultrasound.Proc. R. Sot. Med., 67(b), 447. [ll] Greisen, G., Johansen, K., Ellison, PH., Ftedriksen, P.S., Mali, J. and Friis-Hansen, B. (1984): Cerebral blood flow in the newborn infant: comparisonof Doppler ultrasound and xenon clearance.J. Pediatr., 104, 411-418. [12] Greisen, G. (1986): Cerebral blood flow in preterm infants during the first week of life. Acta Paediatr. Stand., 75, 42-51. 1131AgamY., Hiraishi, S., Oguchi, K., Misawa, H., Horiguchi, Y., Fujino, N. et al. (1991): Changesin left ventricular output from fetal to early neonatal life. J. Pediatr., 119, 441-445. [14] Sutton, M.G., Plappert, T. and Donbilet, P. (1991): Relationship between placental blood flow and combined ventricular output with gestational age in the normal human fetus. Cardiovasc. Res., 25, 603-608. [IS] Winberg, P., Jamtson, M., Marions, L. and Lundell, B.PW. (1989): Left ventricular output during postnatal circulatory adaptation in healthy infants born at full term. Arch. Dis. Child., 64, 1374-1378. [16] Drayton, M.R. and Skidmote, R. (1987): Ducms arteriosus blood flow during first 48 hours of life. Arch. Dis. Child., 62, 1030-1034. [17] Amato, M., Huppi, l?, Gambon,R. and Scheider,H. (1989): Biochemical timing of peri-intraventricular haemorrhage assessedby perinatal CPK-BB isoenzyme measurements.J. Perinat. Med., 17, 447-452. [18] Russell, G.A.B., Jeffers, G. and Cooke, R.W.I. (1992): Plasma hypoxanthine: a marker for hypoxicischaemic induced periventricular leucomalacia? Arch. Dis. Child., 67, 388-392.
Cerebral and renal artery blood flow velocity before and after birth S.T. Kempley”,
S. Vyas, S. Bower, K.H. Nicolaides, H. Gamsu
Children Nationwide Neonatal Centre and Harris Birthright Reseurch Centrr fbr Fetrrl Mrdkinr. King’s College Hospital Medical School. London, UK
Received 16 January 1996; revised 29 March 1996; accepted 2 April 1996
Abstract Objective: To document perinatal changes in cerebral and renal artery haemodynamics 111 premature growth-retarded and normal term infants. Design: Longitudinal study of individual infants. Doppler ultrasound measurements of blood flow velocity (BFV) in the middIe cerebral and renal arteries were obtained before delivery, soon after delivery and during the first week of postnatal life. Setting: Teaching hospital obstetric and neonatal units. Subjects: 13 severely growth retarded infants born at 28-36 weeks gestation, and eight normally grown infants born at term. Results: In both groups, BFV in the cerebral artery was significantly lower in the first few hours after birth than in fetal life, but subsequently increased to reach pre-delivery values by the end of the first week. In contrast, BFV in the renal artery during the first postnatal day was not significantly different from fetal values, but it also increased during the subsequent week. In six of the preterm growth-retarded infants, fetal blood gases were measured in samples obtained by cordocentesis, and in these cases an increase in blood oxygen content at birth was documented. Conclusions: Cerebral artery BFV falls at birth and is relatively ION during the time that premature infants are at the greatest risk of developing periventricular haemorrhage. Keyword.?: Blood flow velocity (BFV); Cerebral artery; Renal artery: Preterm
*Corresponding author. Royal London Hospital, Whitechapel. London El IBB. UK 0378-3782/96/$15.00 0 1996 Elsevier Science Ireland Ltd. All rights reserved PII SO37X-3782(96)01754-9
166
S.T. Kempley et al. I Early Hurnun Development 46 (1996) 165-174
1. Introduction Intrauterine growth retardation due to placental insufficiency is associated with fetal hypoxaemia [l]. Doppler ultrasound studies have demonstrated that in such fetuses, blood flow velocity (BFV) in the cerebral circulation is increased [2], presumably to compensatefor the hypoxaemia and thus maintain cerebral oxygenation. Although in the managementof affected pregnancies,preterm delivery is often necessaryto prevent intrauterine death, such prematurity places the infant at increased risk of conditions such as respiratory distress syndrome and periventricular haemorrhage (PVH). Hambleton and Wigglesworth [3] suggestedthat PVH may be due to rupture of capillaries in the germinal matrix, the consequenceof acute increase in blood flow. However, there is evidence suggesting that haemorrhageoccurs primarily following episodes of hypotension [4,5] and therefore reduced cerebral blood flow. Doppler ultrasound studies of full term infants delivered vaginally have demonstrated that BFV in the middle cerebral artery falls following birth and subsequently increases during the first few days of life [6]. This suggeststhat during the first few days of life, when premature infants are at greatestrisk of PVH [7,8], cerebral artery BFV may be relatively low. In preterm infants BFV in all cerebral arteries increasesduring the first few days of life [9] but there are as yet no studies comparing fetal and neonatal cerebral BFV values in this group. The aim of this study was to determine whether changesin the cerebral circulation similar to those of term infants occur in growth-retarded premature infants, by measuring BFV before birth and during the first week of life. To determine whether any observed changesare specific to the cerebral circulation, we also measuredBFV in the renal artery. For some of these infants intrauterine cord blood gas measurements allow us to relate the changesin BFV to changesin oxygenation. The findings are compared with BFV measurements obtained from normally grown infants delivered at term by caesareansection. 2. Methods In 13 growth-retarded infants that were born at 28-36 (median 32) weeks gestation, serial Doppler ultrasound measurementsof BFV and pulsatility index (PI) were obtained from the middle cerebral and renal arteries. The Doppler studies were carried out within 72 h before delivery (median 22 h) and on four occasions after birth (within 12 h and 36 h after birth and on days 2-4 and 5-9). Not all babies were studied on all occasionseither becausethey were clinically unstable or becausethey died or were discharged from the unit. Patients were selectedsequentially from small-for-gestational age premature babies admitted to the neonatal unit following delivery by caesareansection undertaken becauseof deteriorating fetal or maternal condition arising in a pregnancy known to be complicated by fetal growth retardation (fetal abdominal circumference measured by ultrasonography < 25th centile for gestation). All had birthweights < 3rd
S.T. Kempley et al. I Early Human Development J6 (1496) I6*5%174
centile for gestation and 12 weighed < 1.5 kg. To be included, Doppler studies must have been performed both antenatally and postnatally, so that case selection was limited by the availability of obstetric and neonatal researchers. At the time of the postnatal Doppler measurements,mean intra-arterial blood pressure(measuredvia an umbilical catheter attached to a dome transducer, Medex, Rossendale,UK), and blood gases(ATL 300 blood gas analyzer,WA) were recorded. In six cases,prenatal blood gaseswere measuredin umbilical venous blood samples obtained by cordocentesis,which was performed immediately after the fetal Doppler measurements.Cordocentesis was performed to assist in clinical management and was not a planned part of this study. In these cases, blood oxygen content was calculated from the haemoglobin concentration and blood gas measurementsusing the dissociation characteristics of fetal haemoglobin. Values before and after delivery were compared. The values in the preterm group were compared to those from eight infants that were delivered at 37-41 weeks (median 38 weeks) by elective caesareansection for previous section, cephalopelvic disproportion, or breech presentation. All infants were normal and healthy and their birthweight was > 10th centile for gestation. Doppler studies were performed 3-7 h before delivery and again 3-7 h, 20-24 h and day 5 after birth, just before discharge from hospital. At the time of the Doppler studies, mean arterial blood pressurewas recorded using an oscillometric method (Dinamapp, Critikon, Ascot, UK) and oxygen saturation was measuredusing a pulse oximeter (Ohmeda Medical, Louisville, CO). Invasive measurementsof theseparameterscould not be performed for ethical reasonsas the subjectswere all well babies; the data are presented but have not been used for any calculated values. Approval for the studies in both the preterm and term groups was obtained from the hospital ethical committee and the parents gave informed consent. All Doppler studies were performed while the infants were asleep or lying quietly.
2.1. Dopplrr studies
Flow velocity waveforms were obtained from one of the middle cerebral arteries and from the left renal artery using range-gatedpulse-wave Doppler ultrasound. The placement of sample volumes was guided by colour-flow imaging for the fetal measurements,and by simple real-time imaging for the postnatal studies. Intensityweighted mean BFV and the PI (Systolic-Diastolic/Mean velocity: Gosling and King [lo]) were measuredand the average of at least four cardiac cycles was used. The intensity weighted mean velocity averagesthe varying velocities found between the centre and the periphery of a vessel during laminar flow, by integrating Doppler signal intensity and Doppler shift following fast Fourier transform. High PI indicates increasedvelocity waveform pulsatility which can result from either high downstream vascular resistance,or left-to-right shunting through the ductus arteriosus. An inverse relationship therefore normally exists between BFV and PI. Velocity was measured directly on-screen using inbuilt software and for PI a spectrum analyzer (Doptek 9000, Doptek, Chichester, UK) was used. Whenever the angle of insonation of the
168
S.T. Kempley et al. I Early Human Development 46 (1996) 165-174
vessel differed from 0”, it was measuredand the velocity was corrected for angle of insonation. Antenatal measurementswere performed by an obstetrician using a ~-MHZ linear array transducer (Acuson 128, Acuson, CA), and the postnatal measurementswere performed by a neonatologist using a ~-MHZ mechanical sector transducer (Sonos 100, Hewlett-Packard,Uxbridge, UK). Coefficients of variation between the operators and equipment were calculated from paired measurementsof middle cerebral artery BFV in each of 10 healthy newborns. The mean difference between the obstetric and the neonatal equipment was only 0.45 cm/s (2% of the measurement;t = 0.20, P = O.SS),and the standarddeviation of the error was 2.8 cm/s (14% of the measurement; this compares with a 12% intra-observer variation which we found previously for measurementsof postnatal cerebral artery BFV averaged over four cardiac cycles). 2.2. Statistical analysis
To compare antenatal BFV with the first postnatal measurement,data from each infant were only included for analysis if a measurementhad been obtained from the same artery both before and after delivery. Student’s paired t-test was use to determine the significance of any differences. The mean and 95% confidence intervals of BFV within each time period were then calculated, and unpaired t-tests used to determine the significance of any differences between the various periods. PI measurementswere significantly skewed, and paired Wilcoxon rank-sum test was used to determine the significance of differences in median values. Comparison of BFV values between preterm and term infants was by unpaired Student’s t-test, and comparison of PI values was by the Mann-Whitney u-test.
3. Results
In both the preterm and term infants, the mean middle cerebral artery BFV at the first postnatal measurementwas significantly lower and the mean PI was significantly higher than in fetal life (Table 1). Before delivery, the mean cerebral artery BFV in the preterm, growth-retarded fetuses (15.1 cm/s) was not significantly different from that of the term, appropriately grown fetuses (17.6 cm/s, t = 1.1, P = 0.30). There was no significant change at birth in renal artery BFV or PI in either group. Before delivery, the mean renal artery BFV in the preterm, growth-retarded fetuses (4.7 cm/s) was significantly lower than in the term, appropriately grown fetuses (9.7 cm/s, t = 3.1, P < 0.01). PI was significantly higher in the preterm, growthretarded fetuses (3.26 vs. 2.14, Mann-Whitney P < 0.001). For the infant who developed a parenchymal periventricular haemorrhage,cerebral artery BFV fell from 13.0 cm/s before delivery to 4.6 cm/s after delivery. This postnatal value was relatively low, but three other preterm growth-retardedbabies had lower values. None of the other infants displayed ultrasound evidence of periventricular haemorrhage or periventricular leucomalacia (PVL), although some were discharged back to their referring units before signs of PVL would have been apparent.
S.T. Kempley et al. / Early Human Development 46 (1996) 16.5-174
lb9
Table 1 Blood flow velocity and pulsatility index before delivery and at the time of the first postnatal measurement for the preterm, growth-retarded infants and the term infants Middle cerebral artery Antenatal
Renal artery Postnatal
Antenatal
Postnatal
4.7 (1.2) 9.7 (3.X)
6.3 (2.2) 9.5 (2.0)
3.26 (0.49)
3.63 13 1) 2.14 (0.98)
Mean (SD.) velocity (cm/s) Preterm IUGR Term
15.1 (5.1) 17.6 (4.8)
7.2 (3.5)***
10.7 (2.4)**
Median (interquartile range) Pukatility Index Preterm IUGR Term
1.48 (0.61) 1.47 (0.71)
2.18 (1.5)** 1.96 (0.85)*
Postnatal measurements which were significantly (*P < 0.05; **P < 0.01; ***p < 0.001).
2.14 (0.49)
different from the antenatal measurement are indicated
There were no statistically significant differences in antenatal or postnatal BFV between the growth-retarded infants who died and those who survived (Fostnatal MCA: 8.1 cm/s for survivors, 5.1 cm/s for deaths, P = 0.22; Postnatal Renal: 6.2 cm/s for survivors, 6.4 cm/s for deaths).
3. I. Postnatal trends
The mean and 95% CI of BFV and PI in the middle cerebral and renal arteries of the preterm and term groups during each time-period are shown in Fig. 1 and Fig. 2
Renal-Preterm
01 -1
I
I,,
0
1
2
3
,
/
I
I
4
5
6
7
Age (days)
0' -1
0
1
2
3
4
5
6
7
Age WW
Fig. 1. Blood flow velocity in the Middle Cerebral Artery (MCA, upper graphs) and in the Renal Artery (lower graphs) for the premature growth-retarded infants and the term infants. (Means and 95% CIs, with significant differences between antenatal and postnatal values shown below, and significant differences between the first and later postnatal values shown above the error bars; *P < 0.05; **I’ < 0.01 J.
170
S.T. Kempley et al. I Early Human Development 46 (1996) 165-I 74
MCA-Term
MCA-Preterm
I07 i T
I
;
0;
-1
Renal-Term
I
I,
0
1
2
Renal-Preterm
I
I,,
3
4
,
5
6
01
7
-1
0
1
Age (days)
2
3
4
5
6
7
Age (days)
Fig. 2. Pulsatility index in the Middle Cerebral Artery (MCA, upper graphs)and in the Renal Artery (lower graphs) for the premature growth-retarded infants and the term infants. (Medians and 95% CIs, with significant differences between antenatal and postnatal values shown below, and significant differences between me first and later postnatal values shown above the error bars; *P < 0.05; **P < 0.01). As would be expected, changes in PI were the inverse of changesin BFV. and the numbers of infants studied, their blood pressure and blood gasesare shown in Table 2. Middle cerebral artery BFV was significantly lower in growth-retarded infants at 12
Table 2 Numbers of infants studied and characteristicsat the time that the postnatal measurementswere performed <12 h
12-24 h
3 days
1 week
10 32
8 32
9 41
(29-47)
(19-50)
6 34 (34-38)
No. receiving mechanical ventilation
7.38 (7.13-7.53) 4.7 (3.6-11.1) 8.1 (5.2-11.6) 6
7.36 (7.17-7.44) 5.3 (4.1-8.3) 8.0 (5.3-l 1.6) 2
7.27 (7.24-7.34) 6.9 (4.9-7.3) 8.5 (8.0-12) 3
7.33 (7.27-7.35) 5.1 (5.1-5.6) 7.6 (5.1-8.1) 2
Term infants II Median (range)arterial blood pressure(mmHg)
8
8
39
41
Premature growth-retarded infants n Median (range)arterial blood pressure(mmHg) Median (range) arterial blood gases PH
Median (range) oxygen saturation (%)
(34-45)
(38-52)
94 (88-96)
94 (91-96)
-
(39-41)
6 47 (40-52)
93 (92-97)
S.T. Kempley et al. I Early Human LIeveIopment46 (1996) 16.5-174
1‘II
h of postnatal age, compared with the term infants (Fig. 1: Growth-ret 7.1 cm/s, term 10.7 cm/s, t = 2.23, P < 0.05). There was no significant ence between later measurements.At all ages renal artery BFV was significantly lower in the premature growth-retarded infants than in the term infants. By 1 week of age, middle cerebral and renal artery BFV had increaseds in both groups (Figs. 1 and 2). Changes in pulaatihty index were not sta&tic&ly significant, but were generally inversely related to the changesin BFV Exclusion of data from babies who died made no difference to these trends. 3.2. Comparison with changes in oxygenation For the infants who had cordocentesisperformed, antenatal and postnatal data are displayed in Table 3. Oxygen tension and blood oxygen content increased significantly after delivery (P < 0.01, P < 0.05, respectively). Mid&e cerebral @fery BFV fell significantly (P < 0.01). The fall in middle cerebral artery BFV was &set by a comparableincreasein oxygen content in cases3-6. In cases1 and 2 there were substantial falls in BFV (71% and 8256, respectively) but very little increase in oxygen content (8% and 20%). Both of these infants weighed less than 750 g; one was anaemic and the other was hypocarbic. They both survived with no ultrasound evidence of cerebral pathology. Table 3 Oxygen tension (PO,) and haemoglobin (Hb) for the infants who bad cordocentesisperfmmed Case no. (Birthweight, g) 1 (636) 2
(f5f36) 3 (788) 4 (804) 5
(836) (838) Antenatal mean (S.D.) Postnatal mean (SD.) Significance (F-value)
PO, @Pa)
Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal Antenatal Postnatal
Hb Wdl)
0, content
MCA BFV (cm/s)
4.4 7.5 4.7 8.9 1.7 8.8 3.3 11.1 2.8 8.0 3.9 6.9
10.3 9.8 15.2 15.4 14.5 15.1 15.8 16.1 14.1 14.0 14.8 17.8
5.0 5.4 7.3 8.8 1.8 8.7 5.9 9.5 4.8 8.2 6.6 10.0
22.6 6.5 15.‘7 2.9 14.3 4.2 7.6 4.9 20.7 11.1 9.1 5.1
3.5 (1.1) 8.5 (1.5) < 0.01
14.1
5.2 (1.9) 8.4
15.0
Gw 14.7 (2.7) NS
(1.6) < 0.05
(6.0) 5.8 (2.9) < 0.01
Mean Art. BP @m&T)
Pp, &Pa)
30 32 47 32 29
5.6 4.0 4.7 3.6 6.6 4.0 5.1 5.9 5.3 8.3 5.9 5.7 5.5 (0.7) 5.3 (l.Sj NS
AntenataI values are based on analysis of umbilical venous blood, and postnatal vahres on analysis of arterial blood. Oxygen content is calculated from the dissociation characteristics of fetal haernoglobin. using PO,, Pcoz, pH and haemoglobin values.
172
XT. Kempley et al. I Early Human Development 46 (1996) 165-l 74
4. Discussion This study has demonstratedthat in both growth-retarded preterm infants as well as normally grown term infants, velocity of blood flow in the middle cerebral artery is much lower on the first day of postnatal life than in the fetus. The fall in cerebral artery BFV at birth was accompaniedby an increase in pulsatility index. By the end of the first week, cerebral artery BFV had increased to a level close to that of the fetus. On the first postnatal day renal artery BFV was close to that of the fetus, but by the end of the first week it had risen considerably. All of the antenatal measurementswere performed before the onset of labour, to ensure that they representedthe ‘resting’ fetal state.This meant that all of the infants studied were delivered by caesareansection, and we cannot be sure of the changes which would follow a normal vaginal delivery. However, our findings were similar to those of Meerman et al. [6] who studied infants born vaginally at term. There are three possible explanations for the fall in cerebral artery BFV seen at birth, with the subsequentincrease over the next week. Firstly, it is possible that the reduced cerebral artery BFV on the first day does not represent reduced cerebral blood flow. Instead, it could be due to dilatation of the major cerebral arteries, with volume flow through the artery being close to fetal levels. We cannot exclude this possibility as it is impossible to measurethe diameter of these small vessels using ultrasound imaging. However, a generalised vasodilatation of proximal and distal cerebral arteries should be associatedwith a decreasein pulsatility index, and the PI actually increasedat birth (Fig. 2). Blood flow velocity in the cerebral arteries has been shown to be closely related to cerebral blood flow as measuredby Xenon clearance [ 111. Clearance techniques have also demonstrateda postnatal increase in cerebral blood flow which is similar to the increase in BFV found in this study [12] suggesting that the changesobserved probably do represent changes in volume flow rather than changes in vessel diameter. Secondly, the fall in cerebral artery BFV at birth could representa passive response of the cerebral circulation to changes in the amount of cardiac output available for cerebral perfusion, as a result of the complex changesin cardiovascular function and anatomy which occur at birth. Left ventricular output increasesat birth and falls over the next few days [13] the opposite of the changes in cerebral artery BFV we observed. However, the blood flow available to the fetus for organ perfusion is equal to the combined ventricular output, minus placental blood flow. Doppler studies of the human fetus suggest that at term, combined cardiac output is close to 450 ml/kg/mm, with a placental blood flow of 120 ml/kg/n& [14]. In the early neonatal period the blood flow available to the infant for organ perfusion will be equal to left ventricular output, minus the volume of left-to-right ductal shunting. In the term infant left ventricular output is in the region of 240 ml/kg/min immediately after birth, falling to 200 ml/kg /min over the next 3 days [ 151, with ductal shunting accounting for 60 ml/kg /min shortly after birth and decreasing rapidly thereafter [16]. It is therefore possible that less cardiac output is available for cerebral perfusion shortly after birth, but that the amount increasesrapidly during the first day as ductal closure occurs.
XT. Kempley et al. t Early Human Devetopmmt 46 (19%) 165-174
173
Finally, the fact that BFV fell in the cerebral, but not the renal arteries, suggests that these changescould be due to an increase in cerebral vascular resistance,rather than to a fall in cardiac output. The increase in cerebral Pi after delivery would support this view. Cerebral vascular resistance could have increased in response to the rise in arterial oxygen tension which occurred at birth. However, the later increase in middle cerebral artery BFV cannot be explained on this basis, as oxygen saturation did not change over the first week in either group (Table 2). Only if cerebral oxygen demand increases dramatically over the first few days of postnatal life could metabolic autoregulation account for this increase. For those growth-retarded infants in whom blood gases had been measured antenatally, we were able to calculate blood oxygen content before and after delivery. As none of these infants had been transfusedat this stage,it was safe to assumethat they had predominantly fetal haemoglobm, and the oxygen dissociation characteristics of fetal haemoglobin were used. We documented an increase in arterial oxygen content at birth, but the reduction in mean cerebral artery BFV was greater than the increase in mean arterial oxygen content. Oxygen delivery through an artery will be equal to the product of volume blood flow and arterial oxygen content. If the diameter of the vessel remains constant between fetal and postnatal life, then oxygen delivery would have decreasedin the two infants weighing less than 7.50g who had substantial falls in cerebral artery BFV with little or no increasein blood oxygen content (cases1 and 2). It is noteworthy that following delivery, some of these infants were hypocarbic, hypotensive or anaemic. In the other four infants, oxygen delivery would have stayed constant or increased. For the term infants it is unlikely that postnatal oxygen delivery would have been lower than that occurring in fetal life, as the fall in cerebral artery BFV was not as great. Periventricular haemorrhageis uncommon in the fetus, or after 1 week of postnatal age [7,8]. Our data suggestthat during the period of greatestrisk (i.e. the first 24-48 postnatal hours) cerebral artery BFV is relatively low. The fall in BFV at birth may result from an increase in cerebral vascular resistance,or from changesin available cardiac output, with consequent reduced cerebral blood flow. An increase in blood oxygen content may compensatefor the lower velocity, but this may not be complete in small or sick preterm growth-retarded infants. Biochemical markers of cerebral injury suggestthat perinatal events may be important in the etiology of PVH 1173and periventricular leucomalacia [ 181, even when these are not manifest until later. A major change in cerebral blood flow at birth could be the initiating event which culminates in later PVH. If elective delivery of the premature growth-retardedfetus is to lead to improved cerebral oxygenation, care should be taken to avoid hypocarbia, hypotension and anaemiain the first few hours after birth, as all of thesecould lead to a reduction in cerebral oxygen delivery to below fetal levels.
References [I] Nicolaides, K.H., Soothill, FYW.,Rodeck, C.H. and Campbell, S. (1986): Ultrasound-guidedsampling of umbilical cord and placental blood to assessfetal wellbeing. Lancet, i, 1065-1067.
114
S.T. Kempley et al. I Early Human Development 46 (1996) 165-174
[2] Vyas, S., Nicolaides, K.H., Bower, S. and Campbell, S. (1990): Middle cerebral artery flow velocity waveforms in fetal hypoxemia. Br. J. Obstet. Gynaecol., 97, 797-803. [3] Hambleton, G. and Wigglesworth, J.S. (1976) Origin of periventricuhu haemorrhagein the preterm infant. Arch. Dis. Child., 51, 651-659. [4] Miall-Allen, V.M., de Vries, L.S., Dubowitz, L.M.S. and Whitelaw, A.G. (1989): Blood pressure fluctuation and intraventricular haemorrhagein the preterm infant of less than 31 weeks gestation. Pediatrics, 83, 657-661. [5] Mehrabani, D., Gowen, C.W. and Kopelman, A.E. (1991): Association of pneumothorax and hypotension with intraventricular haemorrhage.Arch. Dis. Child., 66, 48-51. [6] Meerman, R.J., van Bel, F., van Zwieten, P.H.T., Oepkes, D. and den Ouden, L. (1990): Fetal and neonatal cerebral blood flow velocity in the normal fetus and neonate:a longitudinal ultrasound study. Early Hum. Dev., 24, 209-217. [7] De Crespigny, L., Mackay, R., Murton, L., Roy, R. and Robinson, P. (1982): Timing of neonatal cerebrovascularhaemorrhagewith ultrasound. Arch. Dis. Child., 57, 231-233. [8] Szymonowicz, W. and Yu, VY.H. (1984): Timing and evolution of periventricular haemorrhage in infants weighing 1250 g or less at birth. Arch. Dis. Child., 59, 7-12. [9] Evans, D.H., Levene, MI., Shortland, D.B. and Archer, L.N.J. (1988): Resistanceindex, blood flow velocity, and resistance-areaproduct in the cerebral arteries of very low birth weight infants during the first week of life. Ultrasound Med. Biol., 14, 103-110. [ 101 Gosling, R.G. and King, D.H. (1974): Arterial assessmentby Doppler-shift ultrasound.Proc. R. Sot. Med., 67(b), 447. [ll] Greisen, G., Johansen, K., Ellison, PH., Ftedriksen, P.S., Mali, J. and Friis-Hansen, B. (1984): Cerebral blood flow in the newborn infant: comparisonof Doppler ultrasound and xenon clearance.J. Pediatr., 104, 411-418. [12] Greisen, G. (1986): Cerebral blood flow in preterm infants during the first week of life. Acta Paediatr. Stand., 75, 42-51. 1131AgamY., Hiraishi, S., Oguchi, K., Misawa, H., Horiguchi, Y., Fujino, N. et al. (1991): Changesin left ventricular output from fetal to early neonatal life. J. Pediatr., 119, 441-445. [14] Sutton, M.G., Plappert, T. and Donbilet, P. (1991): Relationship between placental blood flow and combined ventricular output with gestational age in the normal human fetus. Cardiovasc. Res., 25, 603-608. [IS] Winberg, P., Jamtson, M., Marions, L. and Lundell, B.PW. (1989): Left ventricular output during postnatal circulatory adaptation in healthy infants born at full term. Arch. Dis. Child., 64, 1374-1378. [16] Drayton, M.R. and Skidmote, R. (1987): Ducms arteriosus blood flow during first 48 hours of life. Arch. Dis. Child., 62, 1030-1034. [17] Amato, M., Huppi, l?, Gambon,R. and Scheider,H. (1989): Biochemical timing of peri-intraventricular haemorrhage assessedby perinatal CPK-BB isoenzyme measurements.J. Perinat. Med., 17, 447-452. [18] Russell, G.A.B., Jeffers, G. and Cooke, R.W.I. (1992): Plasma hypoxanthine: a marker for hypoxicischaemic induced periventricular leucomalacia? Arch. Dis. Child., 67, 388-392.