F162 Archives of Disease in Childhood 1995; 72: F162-F167

Blood pressure, heart rate, and skin temperaturein preterm infants: associations withperiventricular haemorrhage

S W D'Souza, H Janakova, D Minors, R Suri, J Waterhouse, G Appleton, C Ramesh,D G Sims, M L Chiswick

Department of ChildHealth, St Mary'sHospital, HathersageRoad, ManchesterM13 OJHS W D'SouzaR SuriC Ramesh

North WesternRegional NeonatalUnitG AppletonD G SimsM L Chiswick

School of BiologicalSciences, MedicalSchool, University ofManchesterH JanakovaD MinorsJ Waterhouse

Correspondence to:Dr S W D'Souza.Accepted 7 February 1995

AbstractThe mean arterial blood pressure(MABP), heart rate, and skin tempera-ture were monitored every 15 minutes inthe first 10 days after birth in 34 preterminfants, gestational age 24 to 33 weeks.Ultrasound brain scans carried out dailyshowed that a periventricular haemor-rhage (PVH) occurred in a subgroup ofinfants (n= 15) of lower birthweight andgestational age. In infants without PVHthe daily median of MABP increasedwith birthweight and postnatal age; thatof heart rate was not affected by post-natal age, body weight, or gestationalage; and that ofskin temperature showeda slight fall with postnatal age. In infantswith PVH, on or before the day of PVH,daily medians of MABP and skin tem-perature were not significantly differentfrom those of infants without PVH, butthe daily median of heart rate tended tobe slightly higher. The percentage ofpositive correlations between the 96 15minute values per day for heart rate andMABP increased with postnatal age andwith birthweight, but did not differ ininfants who developed a PVH. The coef-ficient of variation (CV) of the 96 15minute values for MABP tended to behigher in infants on the day ofPVH, anda similar trend was apparent on the daybefore.The processes of development of blood

pressure, heart rate, and skin tempera-ture are similar in infants with or withoutPVH but at lower gestational ages alteredblood pressure control may cause brainhaemorrhage.(Arch Dis Child 1995; 72: F162-FI67)

Keywords: periventricular haemorrhage, bloodpressure, heart rate, temperature.

Studies in very low birthweight infants haveshown that in the first 48 to 96 hours of lifeblood pressure is influenced by birthweightand postnatal age. 1-3 A rise in blood pressurewas associated with an increase in birthweightand an increase in postnatal age. The risk of aperiventricular haemorrhage (PVH) wasgreater in infants who had respiratory distresssyndrome and required ventilation.46 PVHhas been associated with changes in cerebralblood flow velocity, which reflected changes insystemic blood pressure.7 The possibility that

changes in systemic blood pressure may affectthe cerebral circulation in a 'pressure passivestate' therefore has implications for the aetiol-ogy of PVH.8 9Most haemorrhages seem to start in the

germinal layer matrix, as thin-walled bloodvessels are vulnerable to damage due to dis-turbances in perfusion, caused by an increasein blood pressure,'0 a fall in blood pressure,"1or a fluctuating blood pressure.7 In suchinfants breathing asynchronously with theventilator, Rennie et al reported a significantlygreater variability in cerebral blood flowvelocity but found no characteristic wavepattern in those at risk from PVH.12 In venti-lated infants changes in blood pressure mayoccur as a result of handling during nursingcare procedures'3 and due to a pneumotho-rax14 which may lead to a PVH. Comparedwith infants who do not develop a PVH, thoseat risk were reported to spend a significantlygreater proportion of time with wide swings inmean blood pressure - that is, with a highercoefficient of variation (CV) in blood pres-sure.3 7 This finding is not consistent with astudy by Miall-Allen et al, 14 who showed that areduction in the CV for systolic blood pressurepreceded the development of a PVH.

In this study we monitored at 15 minuteintervals, for three to 10 days in the first 10days after birth, the mean arterial bloodpressure (MABP), heart rate, and skin temper-ature in an unselected group of very low birth-weight infants, to investigate a link with thedevelopment of a PVH. As baroreceptorreflexes might be poorly developed and thismight influence the development of a PVH wealso investigated the association between heartrate and MABP. Skin temperature was mea-sured as we expected changes in tissue bloodflow after birth to support transitional changesin metabolism in order to maintain body tem-perature.'5 The infants were followed up afterleaving hospital to investigate the incidence ofsubsequent neurological abnormality.

MethodsBetween February 1992 and March 1993 atotal of 34 preterm infants admitted to theNeonatal Medical Unit at St Mary's Hospital,Manchester, were included in the study. Theirgestational ages ranged from 24 to 33 weeks(table). All had the idiopathic respiratorydistress syndrome and received intermittentpositive pressure ventilation. None had a PVHon an ultrasound brain scan immediately after

on June 27, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.72.3.F

162 on 1 May 1995. D

ownloaded from

http://fn.bmj.com/

Blood pressure, heart rate, and skin temperature in preterm infants F163

Clinical conditions in first 10 days of life in relation to development ofPVH*

With PVH Without PVH pConditions (n= 15) (n= 19) Value

Site of deliverySt Mary's Hospital 5 9 NSAnother hospital 10 10

Mode of deliveryVaginal 11 9Caesarean section 4 10 NS

Apgar score1 minute 3-20 (0-52) 4 95 (0 52) 0-0245 minutes 5-67 (0 69) 8-16 (0 32) 0 004

Birthweight (g) 1070 (70) 1382 (125) 0 039Gestational age (week) 27-4 (0 64) 29-5 (0 59) 0-021Sex ratio (male:female) 10:5 8:11 NSDuration of ventilation (hours)24 15 16 NS

Maximum peak airway pressure(cm H2O) 25-80 (1-40) 24-00 (1 10) NSLowest arterial PaO2 28-3 (1-68) 36-1 (2 56) 0-017Highest arterial Paco2 67-1 (3 43) 60-1 (2-41) 0 10Lowest arterial pH 7-11 (0-02) 7-17 (0-02) 0-12FI2O (%) 82-0 (4 57) 75-1 (5 38) NSPneumothorax 2 0 NSPancuronium/morphine 3 0 0-08Inotropes 6 0 0 004Artificial surfactant (Exosurf) 5 11 NS

*Results are expressed number or mean (SE). NS=non-significant.Mean values were compared using Student's t test. Numbers were compared by x2 distributionor Fisher's exact test.

birth. The infants were an unselected groupwho were eligible for entry into the study ifthey had an umbilical artery catheter in placein the first few hours after birth, if they had noobvious congenital malformation, and if aneonatal monitor designated for research wasavailable.

MEAN ARTERIAL BLOOD PRESSURE, HEARTRATE, AND SKIN TEMPERATURE MONITORINGFor each infant the MABP, heart rate, andskin temperature were continuously recordedon a neonatal monitor (Hewlett Packardmonitor, Model 78834A). Blood pressure wasmonitored via an arterial catheter that incor-porated an oxygen electrode and a side lumenwhich was used for measuring blood pressure.Coupled to the catheter was a pressuretransducer from which blood pressure signalswere relayed to the neonatal monitor.Continuous monitoring of skin temperatureand heart rate was carried out using skinelectrodes attached to the neonatal monitor.For measurement of the skin temperature askin electrode incorporating a metal plate wasplaced on the anterior abdominal wall. Themetal plate was attached to a thermistor con-nected to the neonatal monitor.

Electrocardiogram (ECG) electrodes wereattached to the right and left chest wall and theright lower limb. The heart rate was calculatedfrom the ECG which was continuously dis-played on the neonatal monitor alongside theskin temperature and blood pressure. The neo-natal monitors were interfaced to a micro-computer (Hewlett Packard, Model VectraES 12). Every 15 minutes the MABP, heartrate, and skin temperature were recorded on afloppy disk. These recordings were carried outdaily in the first 10 days after birth, unless theinfant had died at an earlier age. Occasionallyrecordings were missed because of: occlusionof the intra-arterial catheter or air in thecatheter; removal of electrodes whilst the babywas being x rayed; or because the thermistor

had not been properly placed on the baby'sskin.The infants received clinical care indepen-

dently of this study (table). Pancuronium (0 1mg/kg) or morphine (0 1 to 0-2 mg/kg) weregiven when deemed necessary to assist ventila-tion, if the infant was observed to be strugglingagainst the ventilator and the arterial partialoxygen pressure was poorly maintainedbecause of active expiration against theinspiratory phase of the ventilator cycle.Plasma or 5% albumin (10-15 mg/kg) wasgiven slowly if the infant was considered to bepoorly perfused, together with inotropic sup-port (dopamine or dobutamine, 5-10 pug/kg/minute). The inotrope infusion was started at2-5-5-0 ,ug/kg/minute and increased at inter-vals. Exogenous surfactant (Exosurf 67-5mg/kg) was administered into the endotra-cheal tube in the first 72 hours of age whenoxygenation was considered to be inadequatedespite increasing ventilatory support.Vitamin E (20 mg/kg) was administered intra-muscularly routinely after admission to theunit.

CRANIAL ULTRASOUND SCANSThese were carried out daily during the studyperiod using an Acuson 128 sector scannerwith a multihead probe; 3 5-7 5 MHz.During each scan the ultrasound probe wasplaced over the infant's fontanelle. It wasinitially positioned in the coronal plane andangled towards the face and then towards theocciput. The probe was later rotated through90 degrees and aligned along the sagittalsuture. The lateral ventricles were seen in thesagittal section by angling the probe to theright and left. The scans were analysed on thesame day. PVH was confirmed by the pres-ence of an area of echogenicity in the germinalmatrix of the subependymal region. Somehaemorrhages extended into the ventricles orinto the substance of the brain. Areas of intra-parenchymal echodensities were followed up;in some there was complete resolution whilein the remaining areas the development ofcysts was noted.

STATISTICSFor each day's recording of MAIBP, heart rate,and skin temperature, the medians and 10-90centile ranges of the 96 15 minute values werecalculated. For MABP, a statistic analogous tothe coefficient of variation was calculated as:10-90 centile range/median. This was analysedafter transformation to logarithms to allow forits skew distribution. The correlation coeffi-cient between MABP and heart rate was calcu-lated for each day's results.

It was considered that the variables ofinterest might be affected by postnatal age,gestational age, and birthweight. Investi-gations of this are complicated by the fact thatpostnatal age varies within infants while theother two measures vary among infants. Theanalysis was therefore done in two stages.First, the dependence on postnatal age was

on June 27, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.72.3.F

162 on 1 May 1995. D

ownloaded from

http://fn.bmj.com/

F164 D'Souza, Janakova, Minors, Suri, Waterhouse, Appleton, et al

600 C0 TE 50-

. +1

co~ &.&*cn40 ..

30B+1 B+2 B+3 B+4 B+5 B+6 B+7 B+8 B+9 B+1O

Postnatal age

Figure 1 Mean (SE) arterial blood pressure 0- - inlight (1 5 kg) infants without PVH in the first 10 daysafter birth. B+ 1, day 1 after birth; B+2, second day afterbirth, etc. Blood pressure is not shown if the number ofbabies in the subgroup is less than three.

investigated by fitting parallel regression lines tothe infants without PVH. If there was evidenceof a dependence on postnatal age, the results forall the infants were adjusted to a postnatal age of5 days using the common slope, b; otherwisethe mean of the available values was calculated.The difference between the infants with or with-out PVH was then studied by an analysis ofcovariance, using gestational age and birth-weight as potential covariates.

ResultsOf the 34 infants, 15 (44%) developed a PVH.In 11 this was diagnosed in the first four daysafter birth, including four infants with asubependymal haemorrhage, three with anintraventricular haemorrhage and dilated ven-tricles, and four with an intraventricular haem-orrhage and periventricular infarction(leucomalacia). The remaining four of the 15infants had haemorrhages diagnosed between6 and 9 days of age, including one with asubependymal haemorrhage, two with anintraventricular haemorrhage and dilatedventricles, and one with an intraventricularhaemorrhage and periventricular leuco-malacia. The five infants with periventricularleucomalacia later developed cystic lesions.None of the infants without PVH had similarischaemic brain damage.The table shows the clinical conditions of

the infants in the first 10 days of life. Infantswho developed a PVH had significantly lowerApgar scores at one and five minutes of ageand they had a significantly lower birthweightand gestational age (Student's t test). Infantswho did not develop a PVH were, therefore, ina better condition at birth than the others inwhom PVH was subsequently diagnosed byultrasound scanning. The duration of ventila-tion, the maximum peak airway pressure, andthe maximum FIO2 were not significantly dif-ferent, but the severity of the respiratorydistress syndrome seemed to be greater ininfants who developed brain haemorrhages, asjudged by the lowest arterial P02 (p=0-017),highest arterial Pco2 (p=0~10), and lowestarterial pH (p=Ol-2, Student's t test). Six ofthese infants who developed a PVH hadreceived plasma and inotrope (p=0 004).As there were differences in birthweight and

gestational age between the two groups of

infants, a straightforward comparison betweenthem in terms of MABP, heart rate, and skintemperature would be difficult to interpret.Therefore, the possible effect of birthweight,gestational age, and postnatal age would needto be assessed.

MEAN ARTERLAL BLOOD PRESSUREThe daily median of MABP (mm Hg) ininfants without PVH was investigated by divid-ing the infants into light ( 1 5 kg) subgroups (fig1). A plot of mean (SE) values against post-natal age in the three subgroups showed a risein blood pressure with postnatal age and withthe mean birthweight of the subgroup.Regression analysis indicated that the MABPincreased very significantly with postnatal age,the regression coefficient being + 1 255mm/day ((SE 0-159); p

Blood pressure, heart rate, and skin temperature in preterm infants F165

Heavier Lighter LighterWithout PVH With PVH

100

~~~.o~~60r

0)--

c0 )

l -40

(L 20

0 - _ - _LfO LOO LI-

Postnatal age (days)

Figure 3 Percentage of total correlations between the 9615 minute recordings of heart rate andMABP that arepositive in infants with and without PVH showing relationto birthweight and postnatal age.

increase or decrease in heart rate with post-natal age. When the means of the heart ratevalues were studied, there was no significantcorrelation with gestational age or birthweight,but there was a suggestion that heart rate wasincreased in the infants with PVH. Thus themean difference between the groups on the daybefore PVH was +9 1 beats/minute, PVHgroup higher (95% confidence interval= + 0-5,+17*7; p

F166 D'Souza, J7anakova, Minors, Suri, Waterhouse, Appleton, et al

Recently, blood pressure and heart rate werefound to be affected by gestational age.22 Ininfants of under 34 weeks' gestation there wasa negative correlation - that is, the two vari-ables were in antiphase - but after 34 weeks'gestation infants shared a positive correlation,indicating that the two variables were inphase.22 We have shown that this relation isalso affected by postnatal age. In our groups ofinfants the correlation between the dailymedians of heart rate and blood pressure wasmore likely to be positive over a less preciseperiod than that previously reported.22 Wefound a rise in the proportion of positivecorrelations between heart rate and MABPwith postnatal age and birthweight.Cunningham et al 22 suggest that asynchrony

in the development of the autonomic nervoussystem may account for this transition in phaserelation. However, this would ignore a plethoraof other possible influences, including theadaptation of chemo- and baroreceptors toextra-uterine life, changes in the central nervoussystem control mechanisms in the face ofincreased input from these receptors, or thesecretion of some endocrine factor which maybias the system to a positive phase relation.23 Inour study the synchronous relation of heart rateto MABP in infants with PVH was similar tothat in a comparable group of infants withoutPVH. Brain haemorrhage had taken placeirrespective ofthe percentage ofpositive correla-tions between heart rate and blood pressure.We found that the CV for MABP was not

influenced by postnatal age, gestational age, orbirthweight: but it was higher for the PVHgroup on the day before and on the day thehaemorrhage was diagnosed. Previous studieshave not considered the influence ofbirthweightand postnatal age on the CV for blood pressure,although the mechanisms regulating these vari-ations are not completely understood.7 14 Infetal animals such variability in blood pressuremay reflect fluctuations in sympathetic tone,and pharmacological blockade of sympatheticactivity greatly attenuates the naturally occur-ring variability in heart rate and MABP.24 Asinfants who developed a PVH had lower Apgarscores and a greater severity of respiratory dis-tress syndrome, hypoxia and acidosis may havemade them more prone to changes in bloodpressure and caused an increase in heart rate.We suggest that this may occur partly throughincreasing sympathetic tone.

During periods of hypoxia, Jansen et alfound that blood flow to the cerebrum, mid-brain, medulla, pons and cerebellum wasincreased.25 The cerebral vasculature of thefetus does not appear to autoregulate, in thatduring hypoxia there is a linear relationbetween systemic blood pressure and cerebralblood flow.26 In similar circumstances it isreasonable to speculate that impaired cere-brovascular autoregulation had resulted inPVH in our infants. The infant's apparentpropensity for increases in cerebral blood flowto dangerous levels may have occurred with the'normal' rise in blood pressure which takesplace after birth due to a pressure passive stateof the cerebral circulation.8 Thus the postnatal

increase in MABP (fig 1) may be delivered tothe germinal layer matrix when there is super-imposed hypoxia and cause disruption ofthin-walled blood vessels.9 10

It seems to us that infants who developed aPVH had been born too soon and instability inblood pressure may have been as a result oftheir immaturity, adding to their difficulty inadapting to an extra-uterine environment.Otherwise, their MABP and skin temperaturewere similar to those in infants of greater gesta-tional age and without a PVH, taking intoaccount their bodyweight, gestational age, andpostnatal age. By contrast, there was a higherheart rate and CV for MABP than in infantswithout a PVH. The extent to which thiswould influence cerebral blood flow velocityrequires further study. Although it remains ourview that hypotension should be treated, wedid not find that therapeutic intervention withplasma and inotropic drugs prevented thedevelopment of PVH.We are grateful for help received from the medical and nursingstaff in the Neonatal Medical Unit, and from Dr S Russell andDr S Rimmer in the Department of Radiology, St Mary'sHospital. The help we also received from consultant paediatri-cians at other hospitals in studying outcome in our infants isgratefully acknowledged. We thank Professor R D H Boyd forhis helpful comments, Professor M J R Healy for help with sta-tistics, and Mrs Elaine Evans for secretarial assistance.

1 Weindling AM. Blood pressure monitoring in the newborn.Arch Dis Child 1989; 64: 444-7.

2 Watkins AMC, West CR, Cooke RWI. Blood pressure andcerebral haemorrhage and ischaemia in very low birth-weight infants. Early Hum Dev 1989; 19: 103-10.

3 Bada HA, Korones SB, Perry EG, Arheart KL, Ray LD,Ponayrous M, et al. Mean arterial blood pressure changesin premature infants and those at risk for intraventricularhaemorrhage.J Pediatr 1990; 117: 607-10.

4 Dykes FD, Lazzara A, Aliman P, Blumstein B, Schwartz J,Braun AW. Intraventricular haemorrhage: a prospectivestudy of ethiopathologies. Pediatrics 1980; 66: 42-9.

5 Cooke RWI. Factors associated with periventricular haem-orrhage in very low birth weight infants. Arch Dis Child1981; 56: 425-31.

6 Levene MI, Fawer CL, Lamont RF. Risk factors in thedevelopment of intraventricular haemorrhage in thepreterm neonate. Arch Dis Child 1982; 57: 410-7.

7 Perlman JM, McMenamin JB, Volpe JJ. Fluctuatingcerebral blood flow velocity in the respiratory distresssyndrome. N Engl J Med 1983; 309: 204-9.

8 Lou CL, Lassen NA, Friis-Hansen B. Impaired autoregula-tion of cerebral blood flow in the distressed newborninfant. J Pediatr 1979; 94: 118-21.

9 Pape K. Etiology and pathogenesis in intraventricular haem-orrhage in newborns. Pediatrics 1989; 84: 382-5.

10 Hambleton G, Wigglesworth JS. Origin of intraventricularhaemorrhage in the preterm infant. Arch Dis Child 1976;51: 651-9.

11 Miall-Allen VM, De Vries LS, Whitelaw AGL. Meanarterial blood pressure and neonatal cerebral lesions. ArchDis Child 1987; 62: 1068-9.

12 Rennie JM, Smith M, Morley CJ. Cerebral blood flowvelocity variability in infants receiving ventilation. ArchDis Child 1987; 62: 1247-51.

13 Omar SY, Greisen G, Ibrahim MM, Youssef AM, Friis-Hansen B. Blood pressure responses to care procedures inventilated preterm infants. Acta Paediatr Scand 1985; 74:920-4.

14 Miall-Allen VM, deVries LS, Dubowitz LMS, WhitelawAGL. Blood pressure fluctuation and intraventricularhaemorrhage in the preterm infant of less than 31 weeks'gestation. Pediatrics 1989; 83: 657-61.

15 Walker AM. Circulatory transitions at birth and the controlof the neonatal circulation. In: Hanson MA, SpencerJAD, Rodeck CH, eds. Fetus and neonate: physiology andclinical applications. Volume 1. The circulation. Cambridge:Cambridge University Press, 1993: 160-6.

16 Dawes GS, Johnston BM, Walker DW. Relationship ofarterial blood pressure and heart rate in fetal, newbornand adult sheep. JPhysiol (Lond) 1980; 309: 405-7.

17 Hanson MA. The importance of baro- and chemoreflexes inthe control of the fetal cardiovascular system. Jf DevPhysiol 1988; 10: 491-511.

18 Padbury JF, Diakomanolis ES, Hobel CJ, Perlman A, FisherDA. Neonatal adaptation: sympathoadrenal response toumbilical cord cutting. Pediatr Res 1981; 15: 1483-7.

19 Padbury JF, Pole DH, Newnham JP, Lam RW. Neonataladaptation: greater sympathoadrenal response in pretermthan full term fetal sheep at birth. Am Jf Physiol 1985; 248:E443-9.

on June 27, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.72.3.F

162 on 1 May 1995. D

ownloaded from

http://fn.bmj.com/

Blood pressure, heart rate, and skin temperature in preterm infants F167

20 Heymann MA, Iwamoto HS, Rudolph AM. Factors affect-ing changes in the neonatal systemic circulation. Ann RevPhysiol 1981; 43: 371-83.

21 Broughton-Pipkin F, Symond EM. Renin-angiotensinsystem in early life. In: Beard RW, Nathanielsz PW, eds.Fetal physiology and medicine. 2nd edn. London:Butterworths, 1984: 459-80.

22 Cunningham S, Deere S, McIntosh N. Cyclical variation ofblood pressure and heart rate in neonates. Arch Dis Child1993; 69: 64-7.

23 Hanson MA. The control of the heart rate and blood pres-sure in the fetus: theoretical considerations. In: HansonMA, Spencer JAD, Rodeck CH, eds. Fetus and neonate:

physiology and clinical applications. Volume 1. Thecirculation. Cambridge: Cambridge University Press,1993: 1-22.

24 Segar JL, Merrill DC, Smith BA, Robillard JE. Role of sym-pathetic activity in the generation of heart rate and arterialpressure variability in fetal sheep. Pediatr Res 1994; 35:250-4.

25 Jensen A, Kemzel W, Kastendieck E. Fetal sympathetic activ-ity, transcutaneous P02 and skin blood flow duringrepeated asphyxia in sheep. JDev Physiol 1987; 15: 309-23.

26 Tweed WA, Cote J, Pash M, Lou H. Arterial oxygenationdetermines auto-regulation of cerebral blood flow in thefetal lamb. Pediatr Res 1983; 17: 246-9.

on June 27, 2021 by guest. Protected by copyright.

http://fn.bmj.com

/A

rch Dis C

hild Fetal N

eonatal Ed: first published as 10.1136/fn.72.3.F

162 on 1 May 1995. D

ownloaded from

http://fn.bmj.com/