lactate study using umbilical cord blood

Original Article

Lactate study using umbilical cord blood: Agreement between LactatePro hand-held devices with blood gas analyser and evaluation of lactatestability over time

Tina Y. SU, Mifanwy REECE and Seng C. CHUAObstetrics and Gynaecology, Westmead Hospital, Sydney, New South Wales, Australia

Background: Lactate measurements have become increasingly preferred over pH analysis in the evaluation of fetalacidaemia in labour. In a busy labour ward, often the umbilical cord may be sampled late and as a result yield unreliablelactate values.Aim: To investigate the agreement of hand-held device Lactate Pro with a reference method blood gas analyser andevaluate the stability of umbilical cord lactate values over time.Methods: Prospective study carried out at elective caesarean section. Sixteen umbilical cords were double clampedimmediately after delivery with paired arterial and venous blood samples collected by an independent researcher, atvarying time intervals, and processed by two Lactate Pro devices and a reference method blood gas analyser.Results: A significant difference of �0.41 to 0.10 mmol/L was found when different groups of Lactate Pro devices werecompared with blood gas analyser at lactate values up to 5.70 mmol/L, with average lactate value of 2.45 mmol/L. Overtime, there is progressive rise in lactate samples obtained from the umbilical cord.Conclusion: Lactate Pro devices have a significant difference, but when used in clinical practice on cord blood afterdelivery, this is unlikely to be meaningful. In intrapartum fetal surveillance, a systematic overestimation might lead tounnecessary intervention. It is possible to retrospectively predict the likely level of lactate at birth in delayed cord samples.

Key words: agreement, blood gas analyser, comparison, lactate, umbilical cord blood.

Introduction

Fetal lactate analysis is increasingly being used as a tool todetect metabolic acidaemia and the decision for urgentdelivery in the presence of a nonreassuring cardiotocogram(CTG) in birth units worldwide. Lactate values have beenfound to correlate with, and are comparable to, fetal pH,base deficit and perinatal outcome.1 It has also been shownthat fetal lactate has better predictive value than pH for Apgarscore <4 at five minutes as well as in relation to moderate-to-severe hypoxic–ischaemic encephalopathy in scalpsamples.2 In addition, pH analysis has been demonstratedto be associated with more intrapartum sampling failure(11–20%) compared with lactate analysis.3,4 As a result,lactate measurements have become an attractive alternative

to the traditional pH assessment via fetal blood samplingused since 1962.5–7

Hand-held lactate analysers have the advantages of easeof use, shorter time to result and lower cost comparedwith blood gas analysers, which commonly require a largersample of blood (up to 75 lL) as well as a processingtime up to 30 minutes.1,4,8–11 This includes the Accusportmeter which has been determined by Pennell et al.12 tocorrelate well with a blood gas analyser, in addition tobeing cheaper and more portable. One of the mostcommonly used devices is the Lactate Pro (Arkray, KDK,Kyoto, Japan), which has better correlation in tests dealingwith fetal blood concentrations in comparison with theAccusport lactate analyser.13

An earlier study sponsored by the Lactate Promanufacturer had concluded that there was a strongcorrelation between the Lactate Pro and existing blood gasanalysers up to 18 mmol/L.14 So far, a small number ofstudies have investigated the correlation between lactatemeasurements from this hand-held device with blood gasanalysers using fetal blood.8,13,15

In addition, current literature available shows conflictingresults on the change of pH and lactate over time whencord blood is not sampled immediately at birth. Thisincludes varying outcomes on the reliability of the pH or

Correspondence: Dr Tina Su, Obstetrics and Gynaecology,Westmead Hospital, Sydney, NSW, Australia. Email:[email protected]

Conflict of interest: None of the authors have a conflict ofinterest.

Received 10 July 2012; accepted 16 February 2013.

© 2013 The Authors 375ANZJOG © 2013 The Royal Australian and New Zealand College of Obstetricians and Gynaecologists

Australian and New Zealand Journal of Obstetrics and Gynaecology 2013; 53: 375–380 DOI: 10.1111/ajo.12081

Th e Australian and New Zealand Journal of Obstetrics and Gynaecology

lactate values, when evaluated at different time intervalsafter birth at different temperatures as well as in variousstorage media.16–22

In the first part of the study, lactate values wereobtained using Lactate Pro devices and compared withthose obtained simultaneously with a reference methodblood gas analyser. This allowed comparison of LactatePro performance with results obtained from otherpreceding studies. In the second part of the study, lactatevalues were measured up to 60 minutes after delivery.The rate of change in lactate levels over time in umbilicalcord samples was then compared with syringe samples,which has previously been shown to remain stable for upto 30 minutes after delivery.19

Materials and Methods

A prospective study was carried out between June andNovember 2008 in Westmead Hospital, Sydney, Australia.Westmead Hospital has a tertiary-level obstetric unitwhere over 5000 deliveries take place annually. Sixteenumbilical cords were collected at elective caesareansections in low-risk women after 37 weeks of gestation,following informed consent.

Protocol

Each umbilical cord was double clamped for the longestlength available immediately after delivery, placed in akidney dish and passed to the independent researcher forsampling of all cords. Outside the operating room,arterial samples were obtained first followed by venoussamples using 23 gauge needles into prelabelledpreheparinised syringes kept on ice to be used as‘control’ and then analysed at 5 to 20-minute intervalsup to 60 minutes. At the same time, umbilical cordarterial and venous samples were collected via 23 gaugeneedles, and also analysed at 5 to 20-minute intervals to60 minutes. The maximum number of samples able tobe obtained per umbilical cord was processed at timeintervals as quickly as allowable by the same blood gasanalyser and Lactate Pro device. The first samples werecollected from within 1 to 10 minutes of birth.

Equipment

Two Lactate Pro hand-held devices (marked LPro1 andLPro2) and two Radiometer ABL735 Blood GasAnalysers were used, one located in the Neonatal IntensiveCare Unit (NICU; ABG1) while the other is within theoperating theatre (ABG2). The arterial and venoussamples were all analysed at one or the other of the bloodgas analysers to compare with a Lactate Pro device at thesame time running in parallel. Some samples wereprocessed by both Lactate Pro devices to allow checkingfor interdevice differences.Both blood gas analysers are calibrated daily and comply

with Quality Control measures as per Manufacturer’s

Manual. The Lactate Pro devices were calibrated with thecalibration strip that accompanies each packet of 25 strips,before starting the new pack.

Statistical analysis

Results were analysed by paired t-tests. Values arepresented as mean differences of the groups comparedand standard deviation (SD), with P-value of <0.05considered significant. Bland–Altman graphs were plottedwith mean difference of the group on Y-axis against themean lactate value on X-axis. Linear mixed-effects modelwith SPSS software was used in the second part of thestudy with calculation of change from initial values overtime.As this study conforms to the standards established by

the NHMRC23 for ethical quality review, ethics approvalwas not sought.

Results

Part one: agreement study

A total number of 479 lactate values were obtained fromthe Lactate Pro devices and blood gas analysers. However,17 results were excluded as these samples were notprocessed by a Lactate Pro device and a blood gasanalyser at the same time to allow valid comparison. Thisleft 462 lactate values or 231 paired values for analysis ingroups for comparison as listed in Table 1.A summary of the comparison groups and their mean

differences with limits of agreement is presented inTable 2.From Table 2, it can be seen there is a mean difference

of 0.10 mmol/L between the two hand-held Lactate Prodevices, with a mean lactate value of 2.53 mmol/L. Thereis also a consistent difference in LPro1 and ABG1 of�0.41, which is also significant, whereby the Lactate Prodevice has consistently lower values of lactate comparedwith blood gas analyser 1. On the other hand, LPro1

Table 1 Groups of lactate analysers for comparison

Groups Devices compared No. of samples

1 Lactate Pro device 1 with Lactate Prodevice 2 (LPro1 – LPro2)

48

2 Lactate Pro device 1 with blood gasanalyser 1 (LPro1 – ABG1)

62


123


7


65

6 Lactate Pro device 1 with combinedblood gas analyser 1 + 2(LPro1 – ABG 1 + 2)

185

376 © 2013 The Authors

ANZJOG © 2013 The Royal Australian and New Zealand College of Obstetricians and Gynaecologists

T. Y. Su et al.

when compared with combined samples of both blood gasanalysers shows only a mean difference of �0.13, which issignificant, with mean lactate value of 2.51. The combinedanalysis of ABG1 and ABG2 was carried out to addressthe potential error related to ABG1 measuring higherlactate values compared with LPro1. In turn, both hand-held Lactate Pro devices were compared with blood gasanalyser 2 which showed minimal differences, but theseresults were not significant. LPro2 compared with ABG1only had a small number of seven samples and was notsignificant and as such a Bland–Altman plot for this groupwas not generated.Figure 1 demonstrates the level of agreement on Bland–

Altman plots which show linear regression analysis (meandifference with 2 SD) between the 2 analysers compared.The Y-axis represents difference between lactate values ofthe two devices compared for each sample, and the X-axisrepresents the mean of the lactate values from the twodevices from the corresponding sample. The closer thecluster is around the value of 0.00 on the Y-axis(difference) the higher the level of agreement there isbetween the two devices compared.

Part two: analysis of lactate over time

A total of 176 cord venous lactate values were comparedwith 88 syringe venous lactate values, and 124 cord arteriallactate values compared with 91 syringe arterial lactatevalues. These are plotted on two graphs against time toshow change over time when samples are kept in syringe onice compared with room temperature in umbilical cord.In Figure 2 increasing change of lactate over time is

clearly demonstrated in samples extracted from theumbilical cord compared with samples kept in syringe,with a greater rate of change seen in the arterial samplesover time. The significant change in lactate value overtime is calculated from an estimate of the within samplechange in value per unit time based on the initial value.

Discussion

The Lactate Pro meter is a convenient and versatile toolcurrently used by many institutions to assess fetal acidosisin the presence of a nonreassuring intrapartum CTG.24 Inour study, we have compared the Lactate Pro hand-helddevices with a blood gas analyser (ABL735) to assess the

agreement of these point-of-care devices as well as theiraccuracy using the blood gas analyser as referencemethod. Up to lactate values of 5.70 mmol/L (arterial)and 4.90 mmol/L (venous) in umbilical cord bloodobtained from low-risk elective caesarean deliveries, wecan conclude that there is acceptable agreement of lactatevalues with a minimal difference of �0.41 mmol/L.However, it has to be noted that this minimal differencemay be more prominent given the more narrow cut-off forreassuring fetal status of 4.8 mmol/L used with LactatePro, compared with the suggested level of 5.4 mmol/Lused when measured with an acid-base analyser.2,8,25

There are significant differences at mean lactate reading of2.31–2.53 mmol/L, but the differences are unlikely to beclinically meaningful. This finding is similar to an earlierstudy conducted with 120 samples which comparedLactate Pro and Accusport to a reference methodcolorimetric plasma lactate analyser, which found LactatePro had better correlation with a similar mean differenceof �0.40 mmol/L to our study of 231 samples.8 Anotherstudy involved 118 samples comparing Nova Lactate Plusand Lactate Pro to a reference method, which foundLactate Pro values to correlate well up to 6.0 mmol/L witha mean bias of �0.3 � 0.4 mmol/L.8 At their full range oflactate values up to 14.7 mmol/L, the mean bias was�0.5 � 0.7 mmol/L.8 It can be noted on the Bland–Altman plots, there appear to be greater differences atprogressively higher values of lactate. We were only ableto obtain few higher values from our series of low-riskelective caesarean deliveries. A study by Kruger et al.26

demonstrated significant correlation between fetal scalplactate within 60 minutes of delivery to umbilical cordarterial and venous lactate. It is then possible that theresults of an umbilical cord study may apply to fetal scalpanalysis. As a result of this umbilical cord study, a furtherstudy involving fetal scalp sampling from high-riskdeliveries is under way to investigate higher levels oflactate at Westmead Hospital. This is evaluated withclinical context and will include neonatal outcome toreassess the threshold for immediate delivery if higherreadings of lactate prove unreliable.The second part of the study has demonstrated a

progressive rise in lactate over time for samples that remainin the umbilical cord compared with those extractedimmediately into a syringe and kept on ice. There was asignificant increase of 0.03 mmol/L per minute over time

Table 2 Mean differences with limits of agreement of the lactate groups compared

Groups forcomparison

Mean difference(mmol/L) SD N P-value

Limits ofagreement

Mean of thegroups

Mean lactate(mmol/L) SD

LPro1 - LPro2 0.10 0.26 48 0.010 �0.42 0.62 (LPro1,LPro2) 2.53 0.79LPro1 - ABG1 �0.41 0.26 62 <0.001 �0.93 0.11 (LPro1,ABG1) 2.47 0.64LPro1 - ABG2 0.01 0.34 123 0.747 �0.67 0.69 (LPro1,ABG2) 2.53 0.88LPro2 - ABG1 �0.14 0.20 7 0.106 �0.54 0.26 (LPro2,ABG1) 2.36 0.52LPro2 - ABG2 0.01 0.25 65 0.884 �0.50 0.50 (LPro2,ABG2) 2.31 0.80LPro1 - ABG (1 + 2) �0.13 0.37 185 <0.001 �0.87 0.61 (LPro1, ABG1 + 2) 2.51 0.81


Lactate study agreement and time effect

COMPARISON BETWEEN LACTATE PRO DEVICE 1 AND BLOODGAS ANALYSER 1

COMPARISON BETWEEN LACTATE PRO DEVICE 1AND BLOOD GAS ANALYSER 2

COMPARISON BETWEEN LACTATE PRO DEVICES 1 AND 2 (a)

(b)

(c)

COMPARISON BETWEEN LACTATE PRO DEVICE 2 ANDBLOOD GAS ANALYSER 2

(d)

COMPARISON BETWEEN LACTATE PRO DEVICE 1 ANDBOTH BLOOD GAS ANALYSERS COMBINED

(e)

Figure 1 (a) The two Lactate Pro devices are compared with each other, with the differences of LPro1 to LPro2 plotted against themean lactate values of the two devices. Broad dashed line is the mean difference of 0.10 mmol/L, with fine dashed lines being limits ofagreement from �0.42 to 0.62. (b) Lactate Pro (LPro1) is compared with blood gas analyser (ABG1), with differences plotted againstmean lactate values. Broad dashed line is the mean difference of �0.41 mmol/L, with fine dashed lines being limits of agreement of �0.93to 0.11. (c) Lactate Pro (LPro1) is compared with blood gas analyser (ABG2), with differences plotted again mean lactate values. Broaddashed line is the mean difference of 0.01 mmol/L, with fine dashed lines being limits of agreement of �0.67 to 0.69. (d) Lactate Pro(LPro2) is compared with blood gas analyser (ABG2), with differences plotted again mean lactate values. Broad dashed line is the meandifference of 0.01 mmol/L, with fine dashed lines being limits of agreement of �0.50 to 0.50. (e) Lactate Pro (LPro1) is compared withboth blood gas analyser results combined (ABG1 + 2), with differences plotted against mean lactate values. Broad dashed line is the meandifference of �0.13 mmol/L, with fine dashed lines being limits of agreement of �0.87 to 0.61.



T. Y. Su et al.

in venous samples and an even higher increase of0.05 mmol/L per minute in arterial samples. The rise inlactate values over time may be due to the ongoingmetabolism of cord blood or vessel endothelium and issupported by two other studies.18,20 The higher differencein arterial samples is likely an effect from the morelactacidemic nature of arterial cord blood, but could alsoreflect the greater metabolic activity in the arterial vesselwall.20 These findings are similar to the rate of changederived by White et al.20 being 0.037 mmol/minute for30 minutes and Dessolle et al.18 with 0.062 mmol/minutein umbilical cord samples at room temperature. This hasclinical implications as often cord gases are not collectedimmediately into a syringe at time of birth due to the needto assist in neonatal resuscitation of a compromisednewborn. Hence, delayed analysis of umbilical cord bloodcan lead to falsely high lactate readings being obtained,

which can have significant medico-legal implications. Thefalsely high lactate level may portray the neonate to be inpoorer condition than its actual state or lead tomisdiagnosis of hypoxic–ischaemic encephalopathy.18 Inany litigation case, the arterial lactate level can beinterpreted as vital evidence. From our study, we havecalculated the absolute change per minute of arterial andvenous blood lactate levels separately over time in roomair. This can be considered to predict the more likelybaseline lactate level at birth, when there has been delayedcord blood collection. For example, an estimate of lactatevalue at birth can be made by subtracting 0.05 multiplied byx minutes since delivery from the delayed arterial lactatevalue obtained. Similarly, estimation of venous lactate valuecan be made by subtracting 0.03 multiplied by x minutessince delivery from the delayed lactate value. However, itmust be used with caution as our algorithm reflects ourspecific group of low-risk deliveries only, where furtherstudies involving large sample and diverse obstetric rangeare required for more meaningful extrapolation.One of the strengths of our study includes having one

independent researcher for collection of all samples toensure consistency and reliable technique. Collections ofrepetitive samples from the same umbilical cord also allowmore reliable comparison with change over time. Thelimitations of our study include a small sample size of 16umbilical cords obtained from low-risk clinical situations.This allowed us to analyse only the normal range oflactate values up to 5.7 mmol/L. Further studies areneeded to evaluate the change in lactate at higher readingswith a greater sample size. Also, the reference methodblood gas analyser (ABG1) is located in NICU whichcontributed to the longer time delay (up to 10 minutesfrom delivery) taken to process the first samples of eachcord with this blood gas analyser. However, the sampleswere able to be processed simultaneously by the LactatePro device and a reference method blood gas analyser toallow for agreement part of the study. The consistentoverestimation by ABG1 when compared with LPro1 mayreflect a likely systematic difference attributable to theblood gas analyser. This is because the two Lactate Prodevices were compared with each other and showed asimilar result with a minimal difference of 0.10. Due totime constraint of sample processing, syringe samples kepton ice were not all able to be warmed prior to processing,and this may have uncertain effect on the analysis. Insummary, when measuring umbilical cord blood, theLactate Pro hand-held device does not have clinicallysignificant difference when compared with a referencemethod blood gas analyser. Over time, there is aprogressive increase in lactate value from samples obtainedfrom blood remaining in the umbilical cord, with a greaterdifference seen in arterial samples.

Acknowledgement

We wish to thank Karen Byth statistician at WestmeadHospital for her invaluable contribution to our data analysis.

VENOUS LACTATE VALUES (mmol/L) OVER TIME OBTAINED FROM SYRINGE VERSUS UMBILICAL CORD

0 10 20 30 40 50 60 70 80Minutes

2

4

2

4

mm

ol/L

Type: Cord

Type: Syringe

ARTERIAL LACTATE VALUES (mmol/L) OVER TIME OBTAINED FROM SYRINGE VERSUS UMBILICAL CORD

0 10 20 30 40 50 60 70 80Minutes

3

6

3

6

mm

ol/L

Type: Cord

Type: Syringe

(a)

(b)

Figure 2 (a) showing stable venous lactate values obtained fromsyringe samples compared with definite increasing values overtime from samples kept in umbilical cord. (b) showing stablearterial lactate values obtained from syringe samples comparedwith even more marked increasing values over time from sampleskept in umbilical cord.


Lactate study agreement and time effect

References1 Westgren M, Divon M, Horal M et al. Routine measurements

of umbilical artery lactate levels in the prediction of perinataloutcome. Am J Obstet Gynecol 1995; 173: 1416–1422.

2 Kruger K, Hallberg B, Blennow M et al. Predictive value offetal scalp blood lactate concentration and pH as markers ofneurologic disability. Am J Obstet Gynecol 1999; 181: 1072–1078.

3 Tuffnell D, Haw WL, Wilkinson K. How long does a fetal scalpblood sample take? BJOG 2006; 113: 332–334.

4 Westgren M, Kruger K, Ek S et al. Lactate compared withpH analysis at fetal scalp blood sampling: a prospectiverandomised study. Br J Obstet Gynaecol 1998; 105: 29–33.

5 Saling E, Schneider D. Biochemical supervision of the foetusduring labour. J Obstet Gynaecol Br Commonw 1967; 74: 799–811.

6 Wiberg-Itzel E, Lipponer C, Norman M et al. Determinationof pH or lactate in fetal scalp blood in management ofintrapartum fetal distress: randomised controlled multicentretrial. BMJ 2008; 336: 1284–1287.

7 Saugstad O. Is lactate a reliable indicator of tissue hypoxia inthe neonatal period? Acta Paediatr 2002; 91: 17–19.

8 Ridenour R, Gada R, Brost B, Karon B. Comparison andvalidation of point of care lactate meters as a replacement forfetal pH measurement. Clin Biochem 2008; 41: 1461–1465.

9 Ramanah R, Martin A, Riethmuller D et al. Value of fetalscalp lactate sampling during labour: a comparative study withscalp pH. Gynecol Obstet Fertil 2005; 33: 107–112.

10 Nordstrom L. Fetal scalp and cord blood lactate. Best PractRes Clin Obstet Gynaecol 2004; 18: 467–476.

11 Westgren M, Kublickas M, Kruger K. Role of lactatemeasurements during labor.Obstet Gynecol Surv 1999; 54: 43–48.

12 Pennell C, Tracy M. A new method for rapid measurement oflactate in fetal and neonatal blood. Aust N Z J Obstet Gynaecol1999; 39: 227–233.

13 Nordstrom L, Chua S, Roy A, Arulkumaran S. Qualityassessment of two lactate test strip methods suitable forobstetric use. J Perinat Med 1998; 26 (2): 83–88.

14 Pyne D, Boston T, Martin D, Logan A. Evaluation of thelactate pro blood lactate analyser. Eur J Appl Physiol 2000; 82:112–116.

15 Luttkus A, Fotopoulou C, Sehouli J et al. Technical performanceof lactate biosensors and a test-strip device during labour. ZGeburtshilfe Neonatol 2010; 214 (2): 62–67.

16 Sykes GS, Molloy PM. Effect of delays in collection oranalysis on the results of umbilical cord blood measurements.Br J Obstet Gynaecol 1984; 91: 982–992.

17 Lynn A, Beeby P. Cord and placenta arterial gas analysis: theaccuracy of delayed sampling. Obstet Gynecol Surv 2007; 62:761–762.

18 Dessolle L, Lebrec J, Darai E. Impact of delayed arterial cordblood sampling for lactate assay. Neonatology 2009; 95: 224–229.

19 Strickland DM, Gilstrap LC, Hauth JC, Widmer K. Umbilicalcord pH and pCO2: effect of interval from delivery todetermination. Am J Obstet Gynecol 1984; 148: 191–194.

20 White C, Mok T, Doherty D et al. The effect of time,temperature and storage device on umbilical cord blood gasand lactate measurement: a randomised controlled trial. JMatern Fetal Neonatal Med 2012; 25: 587–594.

21 Sinn J, Lloyd J, Todd D et al. Umbilical cord blood lactatein normal infants: comparison between two methods ofmeasurement. J Paediatr Child Health 2001; 37: 24–27.

22 Nordstrom L, Ingemarsson I, Persson B et al. Lactate in fetalscalp blood and umbilical artery blood measured duringnormal labor with a test strip method. Acta Obstet GynecolScand 1994; 73: 250–254.

23 National Health and Medical Research Council of Australia(NHMRC). When Does Quality Assurance in Health CareRequire Independent Ethical Review? [Accessed 28 March2008.] Available from URL: http://www.nhmrc.gov.au/publications/synopses/e46syn.htm.

24 Borruto F, Comparetto C, Treisser A. Prevention of cerebralpalsy during labour: role of foetal lactate. Arch Gynecol Obstet2008; 278: 17–22.

25 Heinis A, Spaanderman A, Klein Gunnewiek J, Lotgering F.Scalp blood lactate for intra-partum assessment offetal metabolic acidosis. Acta Obstet Gynecol Scand 2011; 90:1107–1114.

26 Kruger K, Kublickas M, Westgren M. Lactate in scalp andcord blood from fetuses with ominous fetal heart rate patterns.Obstet Gynecol 1998; 92: 918–922.



T. Y. Su et al.

lactate study using umbilical cord blood

Documents