Kai KonigDan M. CasalazEmily J. BurkeAndrew Watkins

Accuracy of non-invasive blood pressuremonitoring in very preterm infants

Received: 14 April 2011Accepted: 11 December 2011Published online: 6 March 2012� Copyright jointly held by Springer andESICM 2012

Electronic supplementary materialThe online version of this article(doi:10.1007/s00134-012-2499-y) containssupplementary material, which is availableto authorized users.

K. Konig ()) � D. M. Casalaz �E. J. Burke � A. WatkinsDepartment of Paediatrics, Mercy Hospitalfor Women, 163 Studley Road,Melbourne, VIC 3084, Australiae-mail: kaikonig@gmail.comTel.: ?61-3-84584444Fax: ?61-3-94595389

Abstract Purpose: Blood pres-sure (BP) monitoring is an essentialprocedure in intensive care. There iscontroversy about the reliability ofnon-invasive BP measurements invery preterm infants. This prospectivetrial compared non-invasive BP mon-itoring with BP monitoring via anumbilical arterial catheter (UAC) inthis population. Methods: Preterminfants born at less than 32 weeksgestation requiring a UAC for clinicalmanagement were eligible. Enrolledinfants had up to three BP measure-ments on the right arm (RA) and rightleg (RL) when in a resting state. UAC-BP measurements were noted imme-diately after the non-invasive BP wasdisplayed on the monitor. Measure-ments were analysed in subgroupsaccording to birth weight: no greaterthan 750 g, 751–1,000 g, above1,000 g. Statistical analysis reportsmedian, range, and Bland–Altmananalysis. Results: Sixty infants wereincluded. Median (range) gestationalage was 26.4 weeks (23.6, 31.2); birthweight 924 g (581, 1,518). A total of1,865 measurements were performed(RA: 935, RL: 930). Mean difference

(95% limits of agreement) for infantsno greater than 750 g: RA 2.53 mmHg(-11.18, 16.24), RL -0.804 mmHg(-12.65, 11.04); for infants751–1,000 g: RA 3.535 mmHg (-9.6,16.7), RL -1.239 mmHg (–13.14,10.66); for infants above 1,000 g: RA-1.65 mmHg (-13.47, 10.17), RL-4.101 mmHg (-14.17, 5.96).Conclusions: Although the averagedifferences between invasive and non-invasive BP measurements areacceptable, the range of under- andoverestimation of non-invasive BPmeasurements is large and not consis-tent, making reliance on non-invasivemodalities to guide circulatory man-agement problematic. If arterial BPmonitoring is not available, our resultssuggest measuring non-invasive BP onthe leg in preterm infants with a birthweight no greater than 1,000 g.

Keywords Blood pressure �Monitoring �Umbilical arterial catheter �Preterm infant

Introduction

Haemodynamic monitoring of very preterm infantsremains challenging due to the infants’ unique postnatalcirculatory transition and the size and weight of thesepatients. Standard adult intensive care modalities such as

pulmonary artery catheters are not available for the pre-term population. The neonatologist needs to rely onclinical signs, non-invasive modalities, and—with somelimitations—biochemical markers. Over the last decade,functional echocardiography has gained increased popu-larity. Kluckow and Evans [1] initially described and

Intensive Care Med (2012) 38:670–676DOI 10.1007/s00134-012-2499-y PEDIATRIC ORIGINAL

established measurement of blood flow in the superiorvena cava (SVC) as a surrogate for systemic blood flow.They demonstrated in a cohort of 130 preterm infants thatlow SVC flow within the first 48 h of life is associatedwith early brain injury and poorer neurodevelopmentaloutcome [2, 3]. Further studies looking at SVC flow andother haemodynamic markers such as capillary refill time,blood pressure (BP), central–peripheral temperature dif-ference, urine output, and hyperkalemia did not showconvincing correlation [4, 5]. While echocardiographyadds valuable information for the clinician, it requires ahigh level of skill and expertise, and it is not a modality ofcontinuous monitoring. On the other hand, BP can bemonitored continuously via indwelling arterial catheters.While this provides real-time information, there isincreasing controversy about the exact definition of lowblood pressure, and when and how to treat it. Severalstudies reported an association between hypotension andmajor adverse outcomes [6–15]. However Dempsey andBarrington [16] recently reviewed BP thresholds foradverse outcome and treatment options for hypotensionand concluded that there was little evidence to supportcommon approaches to the management of low bloodpressure in preterm infants. Dempsey et al. [17] alsoreported similar outcomes between extremely low birthweight infants that had normal BP or untreated BPlower than their gestational age but otherwise signs ofgood perfusion. Logan et al. [18] recently studiedindicators of hypotension during the first day of life inextremely low gestational age neonates and the risk ofdevelopmental delay at 2 years of age. After adjustingfor confounders, none of the indicators for hypotensionwere associated with poor neurodevelopmentaloutcome.

Despite the debate, monitoring of BP remains standardcare in very preterm infants in the neonatal intensive careunit (NICU). While invasive, intra-arterial BP monitoringvia an umbilical arterial catheter (UAC) or a peripheralarterial line is the widely accepted gold standard, it carriessignificant risks, in particular the risk of bleeding, infec-tion, thrombosis or vasospasm [19]. Non-invasive devicesare available for BP measurements even for the tiniestinfants although their accuracy in newborns is contro-versial. Only one study has been published exclusivelyreporting results of paired arterial and non-invasivemeasurements in 12 very preterm infants. This studyshowed that the oscillometric device tended to overesti-mate BP in infants with low BP [20]. Other studies carriedout in mixed preterm and term infant population dem-onstrated both over- and underestimation of the arterialmean BP, with overestimation the more common phe-nomenon [21–23].

The aim of this study was to assess the accuracy ofnon-invasive BP measurements in a large cohort of verypreterm infants.

Methods

Patients

The study was conducted at the Mercy Hospital forWomen, a tertiary perinatal centre in Melbourne, Aus-tralia (approximately 6,000 births per year), betweenOctober 2006 and June 2009. Very preterm infants born atless than 32 weeks gestational age who required anindwelling UAC for clinical reasons were eligible. Babieswith congenital heart disease, limb abnormalities, ormoribund patients were excluded. Only babies with aUAC in an appropriate position were recruited. The studyinvestigators were not involved in the decision-makingconcerning UAC insertion and UAC removal. Informedparental consent was obtained prior to study commence-ment. The institutional human research ethics committeeapproved the study.

Study protocol

Non-invasive cuff BP measurements were performed byNICU nursing staff who received an education sessionabout the study protocol prior to study commencement.Prior to the first cuff measurement, the study investigatorsmeasured the circumference of the right arm and the rightleg. The appropriate cuff sizes were chosen according tothe manufacturer’s recommendations with measurementsperformed on the right arm and right leg to ensure con-sistency. This approach facilitated an upper and lowerlimb assessment. Measurements were only performedduring routine care when the infant was in a quiet restingstate, to minimise the impact on the infants’ resting andsleeping periods. Initially, the cuff was placed at the rightupper arm. Arterial BP was monitored for at least 3 min.If the arterial BP was stable for at least 3 min, the firstmeasurement was performed. If the baby remained set-tled, the procedure was repeated up to two times. Thesame procedure was then performed with the cuff on theright leg. The arterial BP was read when the monitorshowed each new cuff measurement.

The study of each baby was completed when themanaging neonatologist decided that the UAC was nolonger clinically necessary in the care of the baby.

All BP measurements were performed with the Intel-liVue Patient Monitor System MP70 (Multi-Measurement-Server M3001A, Philips Australia, NorthRyde, New South Wales, Australia), the standard moni-toring system of the NICU. The arterial wave form wascontinuously displayed on the patient monitor. InvasiveBP monitoring was performed with a UAC in high posi-tion (5 or 3.5 Fr, Vygon SA, Ecouen, France). The lengthof both catheters was 38 cm. The UACs were connectedto a transducer kit (Transpac�, ICU Medical Aust Pty

671

Limited, Baulkham Hills, NSW, Australia). According tothe manufacturer, the damping coefficient of the trans-ducer kit is between 0.18 and 0.22 (personalcommunication). The transducer was initially zeroed atthe level of the right atrium as per NICU protocol andmaintained by a continuous infusion (0.5 ml/h) of 0.45%NaCl with heparin 1 U/ml for infants below 1,000 g birthweight and 0.9% NaCl with heparin 1 U/ml for infantsabove 1,000 g birth weight. The system was calibratedand checked for bubbles before measurements wereperformed.

Statistical analysis

Paired mean BP measurements were analysed in sub-groups according to birth weight: no greater than 750,751–1,000, above 1,000 g; for upper and lower limbmeasurements. Results are presented in Bland–Altmanplots using GraphPad Prism version 5.00 for Windows(GraphPad Software, Inc., La Jolla, California, USA) todisplay mean difference (bias), standard deviation (SD),and 95% limits of agreement (mean difference betweenmeasures ± 2SD) between non-invasive and invasivemeasurements. The analysis was adjusted for multiplemeasurements.

Results

Sixty preterm infants born at less than 32 weeks gestationwere recruited during the study period. Median gesta-tional age was 26.4 weeks (range 23.6, 31.2), and medianbirth weight was 924 g (range 581, 1,518). A total of1,865 paired measurements were performed, 935 on theright arm and 930 on the right leg. A median of 30 pairedmeasurements was recorded for each baby (range 5, 58).Twenty-six (43.3%) infants were female, 11 (18.3%) wereoutborn and transferred postnatally to our NICU. Fifty-six(93.3%) infants had at least one dose of antenatal steroidadministration. Thirty-four (56.6%) infants were deliv-ered via caesarean section. Eighteen (30%) infants had a5 min Apgar score of no greater than 5, and 51 (85%)were intubated at birth. Fifty-six (93.3%) received sur-factant and mechanical ventilation. Forty infants (66.6%)were managed with inotropes. Three (5%) study infantssuffered from a severe intraventricular haemorrhage gradeIII or IV, three (5%) had necrotising enterocolitis, four(6.6%) required laser treatment for severe retinopathy ofprematurity, and 28 (46.6%) developed bronchopulmo-nary dysplasia, defined as need for respiratory support orsupplemental oxygen at the corrected age of 36 weeksgestation. Three (5%) study infants died.

Infants with birth weights no greater than 750 g

Sixteen study infants were born with a birth weight of nogreater than 750 g. A total of 552 paired measurementswere performed: 302 paired measurements on the rightarm and 250 paired measurements on the right leg.Bland–Altman analysis for right arm measurementsrevealed a bias of 2.53 mmHg (SD 6.994), the cufftending to overestimate BP. The 95% limits of agreementwere from -11.8 to 16.24 mmHg (Fig. 1).

Right leg measurements slightly underestimated arte-rial BP with a bias of -0.804 mmHg (SD 6.043). The95% limits of agreement were from -12.65 to11.04 mmHg (Fig. 2).

20 40 60

-20

0

20

40

Bias 2.530

Mean Blood pressure (mmHg)Mea

sure

men

t d

iffe

ren

ce (

mm

Hg

)

Fig. 1 Bland–Altman plot showing right arm mean BP versusUAC mean BP in infants with birth weights no greater than 750 g

20 40 60

-20

0

20

40

Bias -0.8040

Mean Blood pressure (mmHg)

Mea

sure

men

t d

iffe

ren

ce (

mm

Hg

)

Fig. 2 Bland–Altman plot showing right leg mean BP versus UACmean BP in infants with birth weights no greater than 750 g

672

Infants with birth weights 751–1,000 g

Twenty-five study infants were born with a birth weightof 751–1,000 g. A total of 783 paired measurements wereperformed: 398 paired measurements on the right arm and385 paired measurements on the right leg. Bland–Altmananalysis for right arm measurements revealed a bias of3.535 mmHg (SD 6.718), overestimating the arterial BP.The 95% limits of agreement were from -9.633 to16.7 mmHg (Fig. 3).

The bias for right leg measurements was -1.239mmHg (SD 6.072) reflecting slight underestimation. The95% limits of agreement were from -13.14 to 10.66mmHg (Fig. 4).

Infants with birth weights above 1,000 g

Nineteen study infants were born with a birth weight ofabove 1,000 g. A total of 530 paired measurements wereperformed: 263 paired measurements on the right arm and267 paired measurements on the right leg. Bland–Altmananalysis for right arm measurements revealed a bias of-1.65 mmHg (SD 6.03). The 95% limits of agreementwere from -13.47 to 10.17 mmHg (Fig. 5).

The bias for right leg measurements was -4.101mmHg (SD 5.135). The 95% limits of agreement werefrom -14.17 to 5.964 mmHg (Fig. 6). Both upper andlower limb measurements underestimated the arterial BP.

Discussion

BP monitoring is frequently performed in very preterminfants as hypotension has been associated with increasedmortality and morbidities such as peri/intraventricularhaemorrhage, as well as increased risk of adverse long-term neurodevelopment [6–15]. Measurement of BPcommonly guides circulatory management with volumereplacement or inotropes and vasopressors [24, 25].Continuous BP monitoring via an intra-arterial catheter isfeasible even in the tiniest infants. Because indwellingarterial lines have well-described risks, there are thereforeadvantages in using non-invasive techniques if suchtechniques are reliable [19].

Three studies have been conducted recently to assessthe accuracy of non-invasive BP monitoring with oscil-lometric devices in neonates. Dannevig et al. [22]compared three oscillometric BP devices in 20 infantsbetween 25 and 42 weeks gestation in 2005. They per-formed 32 paired measurements with each device andshowed that one device overestimated arterial BP whilethe other two underestimated it. O’Shea and Dempsey

20 40 60

-20

0

20

40

Bias 3.535

Mean Blood pressure (mmHg)Mea

sure

men

t d

iffe

ren

ce (

mm

Hg

)

Fig. 3 Bland–Altman plot showing right arm mean BP versusUAC mean BP in infants with birth weights between 751 and1,000 g

20 40 60

-20

0

20

40

Bias -1.239

Mean Blood pressure (mmHg)

Mea

sure

men

t d

iffe

ren

ce (

mm

Hg

)

Fig. 4 Bland–Altman plot showing right leg mean BP versus UACmean BP in infants with birth weights between 751 and 1,000 g

20 40 60

-20

0

20

40

Bias -1.650

Mean Blood pressure (mmHg)

Mea

sure

men

t d

iffe

ren

ce (

mm

Hg

)

Fig. 5 Bland–Altman plot showing right arm mean BP versusUAC mean BP in infants with birth weights above 1,000 g

673

[23] studied 25 infants with birth weights between 560and 4,050 g; 21 of those infants were born preterm. Atotal of 332 measurements were performed with threeoscillometric devices. All devices overestimated thearterial measurements with an average difference between3.3 and 8.4 mmHg. Meyer et al. [26] found betteragreement between non-invasive and invasive BP mea-surements. In 50 preterm infants between 24 and35 weeks gestation, 200 measurements were performed at5, 12, 18, and 24 h after birth. The average difference wasbetween -0.36 and -0.81 mmHg, depending on the birthweight.

Papadopoulos and colleagues [27] chose a differentapproach to study three oscillometric BP monitors forneonates by using a simulator. Their group demonstratedacceptable agreement of two devices but inaccuratereadings of the third device tested. This experimentaldesign is obviously superior in assessing the repeatabilityof BP devices given that the same measurement can berepeated numerous times under the same conditions. Thedrawback of such experimental studies is that in vitroresults cannot be fully translated into routine clinicalpractice where BP monitoring is performed on individualpatients with constant changes in their clinical condition.

The results of our study demonstrate both over- andunderestimation by non-invasive measurements. In thegroup of babies born with a birth weight no greater than750 g the oscillometric method overestimated arterial BPvalues while measured on the leg it slightly underesti-mated. A similar picture was found in the group with abirth weight between 751 and 1,000 g. This is particularlyconcerning, as these babies are the most vulnerablepatients with the highest risk of adverse outcome andoften require circulatory support in their first days of life.In infants with birth weights above 1,000 g, both arm andleg measurements tended to underestimate the arterial BP,

leg measurements being more prone to this than armmeasurements. Although the average differences betweenmethods appear to be small for all groups (betweenminimum -0.8 and maximum -4.1 mmHg), the 95%limits of agreement need a critical appraisal. For allgroups the range comprises at least 20 mmHg. Given thata common rule of thumb is to choose the gestational ageof a very preterm infant as the lowest acceptable mean BPand guide initiation of circulatory support based on thesecut-off values, it is obvious that reliance on non-invasivemodalities is problematic. Subgroup analysis of mea-surements of infants with certain risk factors, such aschorioamnionitis, intrauterine growth restriction, low5 min Apgar score, or outborns requiring immediatepostnatal transfer to our NICU, revealed a similar picture(see detailed analysis and plots in the online resource). Inour own clinical practice we would consider assessmentincluding functional echocardiography and invasive BPmonitoring necessary in such a clinical scenario. In lightof the current controversy about the clinical significanceof current definitions of neonatal hypotension and whe-ther treatment of hypotension actually improvesimportant outcomes, we believe that meticulousassessment of infants at risk of circulatory compromiseis paramount.

To our best knowledge, this study is so far the largeststudy with more than 1,800 paired measurements com-paring non-invasive versus invasive BP measurements inpremature neonates. In our study design we elected tolimit eligibility to preterm infants less than 32 weeksgestation as this subgroup represents the majority ofNICU patients requiring continuous BP monitoring. Wealso decided to strictly use UA measurements with UACin an appropriate position only, and to perform oscillo-metric measurements only on the right arm and leg toavoid as much methodological bias as possible.

All nursing staff involved received an education ses-sion prior to the start of the study to ensure a uniformapproach when measuring BP non-invasively. Measure-ments were only done when a baby was in a quiet restingstate but the definition of quiet and resting may slightlydiffer between staff. In addition, arterial BP monitoring isnot free of errors although it is considered as the goldstandard. Umbilical catheters are small in diameter andmade of soft polyvinyl chloride which may kink, resultingin incorrect measurements. We addressed this potentialerror by checking the UA on a daily basis until thecatheter was removed. It is, however, possible to missminor kinking or kinking that is hidden under tapes orsuture material. Another potential error can occur due tosmall air bubbles in the catheter which can lead todamping with high diastolic and low systolic readings[28]. Despite all efforts we cannot exclude that theseerrors may have affected our measurement results. Sucherrors are likely to be uncommon and to occur randomly,so are unlikely to have introduced a systematic bias. Our

20 40 60

-20

0

20

40

Bias -4.101

Mean Blood pressure (mmHg)

Mea

sure

men

t d

iffe

ren

ce (

mm

Hg

)

Fig. 6 Bland–Altman plot showing right leg mean BP versus UACmean BP in infants with birth weights above 1,000 g

674

findings can be taken as a realistic assessment of whathappens in normal careful practice in modern NICUs.

Our study showed a remarkable difference betweenupper and lower non-invasive measurements in infantswith a birth weight no greater than 1,000 g. This is oncontrast to O’Shea and Dempsey [23] who only found aminimal average difference of 0.13 mmHg. Kunk andMcCain [29] also studied preterm infants and did not finda significant difference between upper arm and calfmeasurements. Both studies’ infants had a median birthweight of 1,629 and 1,790 g, respectively, so it is possiblethat the demonstrated overestimation on arm measure-ments and under-reading of leg measurements in ourstudy are limited to extremely preterm infants. Wespeculate that this could be related to the minimal musclemass in these babies, in particular in preterm infants’upper limbs.

Conclusion

Our study results strongly question the value of non-invasive oscillometric BP measurements in very preterminfants as the limits of agreement are large, indicating thepossibility of significant error. Reliance on non-invasiveBP monitoring to guide circulatory management isproblematic in this population. If arterial BP monitoring isnot available or indicated, our results suggest that mea-surement of non-invasive BP on the leg in preterm infantswith a birth weight no greater than 1,000 g is the bestalternative, although the potential for error remainssignificant.

Acknowledgments The study investigators would like to thank allthe families for their cooperation and all NICU nurses for theexcellent support in conducting this study.

References

1. Kluckow M, Evans N (2000) Superiorvena cava flow in newborn infants: anovel marker of systemic blood flow.Arch Dis Child Fetal Neonatal Ed82:F182–F187

2. Kluckow M, Evans N (2000) Lowsuperior vena cava flow andintraventricular haemorrhage in preterminfants. Arch Dis Child Fetal NeonatalEd 82:F188–F194

3. Hunt RW, Evans N, Rieger I, KluckowM (2004) Low superior vena cava flowand neurodevelopment at 3 years invery preterm infants. J Pediatr145:588–592

4. Osborn DA, Evans N, Kluckow M(2004) Clinical detection of low upperbody blood flow in very prematureinfants using blood pressure, capillaryrefill time, and central-peripheraltemperature difference. Arch Dis ChildFetal Neonatal Ed 89:F168–F173

5. Kluckow M, Evans N (2001) Lowsystemic blood flow and hyperkalemiain preterm infants. J Pediatr139:227–232

6. Miall-Allen VM, de Vries LS,Whitelaw AG (1987) Mean arterialblood pressure and neonatal cerebrallesions. Arch Dis Child 62:1068–1069

7. Watkins AM, West CR, Cooke RW(1989) Blood pressure and cerebralhaemorrhage and ischaemia in very lowbirthweight infants. Early Hum Dev19:103–110

8. Low JA, Froese AB, Galbraith RS,Smith JT, Sauerbrei EE, Derrick EJ(1993) The association between pretermnewborn hypotension and hypoxemiaand outcome during the first year. ActaPaediatr 82:433–437

9. Goldstein RF, Thompson RJ Jr, OehlerJM, Brazy JE (1995) Influence ofacidosis, hypoxemia, and hypotensionon neurodevelopmental outcome invery low birth weight infants. Pediatrics95:238–243

10. Cunningham S, Symon AG, Elton RA,Zhu C, McIntosh N (1999) Intra-arterialblood pressure reference ranges, deathand morbidity in very low birthweightinfants during the first seven days oflife. Early Hum Dev 56:151–165

11. Al-Aweel I, Pursley DM, Rubin LP,Shah B, Weisberger S, Richardson DK(2001) Variations in prevalence ofhypotension, hypertension, andvasopressor use in NICUs. J Perinatol21:272–278

12. Synnes AR, Chien LY, Peliowski A,Baboolal R, Lee SK (2001) Variationsin intraventricular hemorrhageincidence rates among Canadianneonatal intensive care units. J Pediatr138:525–531

13. Martens SE, Rijken M, Stoelhorst GM,van Zwieten PH, Zwinderman AH, WitJM, Hadders-Algra M, Veen S (2003) Ishypotension a major risk factor forneurological morbidity at term age invery preterm infants? Early Hum Dev75:79–89

14. Fanaroff JM, Wilson-Costello DE,Newman NS, Montpetite MM, FanaroffAA (2006) Treated hypotension isassociated with neonatal morbidity andhearing loss in extremely low birthweight infants. Pediatrics117:1131–1135

15. Batton B, Zhu X, Fanaroff J, KirchnerHL, Berlin S, Wilson-Costello D,Walsh M (2009) Blood pressure, anti-hypotensive therapy, andneurodevelopment in extremely preterminfants. J Pediatr 154:351–357 357 e351

16. Dempsey EM, Barrington KJ (2007)Treating hypotension in the preterminfant: when and with what: a criticaland systematic review. J Perinatol27:469–478

17. Dempsey EM, Al Hazzani F,Barrington KJ (2009) Permissivehypotension in the extremely lowbirthweight infant with signs of goodperfusion. Arch Dis Child FetalNeonatal Ed 94:F241–F244

18. Logan JW, O’Shea TM, Allred EN,Laughon MM, Bose CL, Dammann O,Batton DG, Engelke SC, Leviton A(2011) Early postnatal hypotension anddevelopmental delay at 24 months ofage among extremely low gestationalage newborns. Arch Dis Child FetalNeonatal Ed 96:F321–F328

19. Hermansen MC, Hermansen MG(2005) Intravascular cathetercomplications in the neonatal intensivecare unit. Clin Perinatol 32:141–156 vii

20. Diprose GK, Evans DH, Archer LN,Levene MI (1986) Dinamap fails todetect hypotension in very lowbirthweight infants. Arch Dis Child61:771–773

21. Chia F, Ang AT, Wong TW, Tan KW,Fung KP, Lee J, Khin K (1990)Reliability of the Dinamap non-invasivemonitor in the measurement of bloodpressure of ill Asian newborns. ClinPediatr (Phila) 29:262–267

675

22. Dannevig I, Dale HC, Liestol K,Lindemann R (2005) Blood pressure inthe neonate: three non-invasiveoscillometric pressure monitorscompared with invasively measuredblood pressure. Acta Paediatr94:191–196

23. O’Shea J, Dempsey EM (2009) Acomparison of blood pressuremeasurements in newborns. Am JPerinatol 26:113–116

24. Raab E, Seri I (2006) Diagnosis andtreatment of the hypotensive very lowbirth weight infant during postnataltransition. J Perinatol 26:657–659

25. Dempsey EM, Barrington KJ (2009)Evaluation and treatment ofhypotension in the preterm infant. ClinPerinatol 36:75–85

26. Meyer S, Sander J, Graber S,Gottschling S, Gortner L (2010)Agreement of invasive versus non-invasive blood pressure in pretermneonates is not dependent on birthweight or gestational age. J PaediatrChild Health 46:249–254

27. Papadopoulos G, Mieke S, Elisaf M(1999) Assessment of the performancesof three oscillometric blood pressuremonitors for neonates using a simulator.Blood Press Monit 4:27–33

28. Weindling AM (1989) Blood pressuremonitoring in the newborn. Arch DisChild 64:444–447

29. Kunk R, McCain GC (1996)Comparison of upper arm and calfoscillometric blood pressuremeasurement in preterm infants.J Perinatol 16:89–92

676