Wiberg-Itzel, Eva MD, PhD; Akerud, Helena MD, PhD; Andolf, Ellika MD, PhD; Hellström-Westas, Lena MD, PhD; Winbladh, Birger MD, PhD; Wennerholm, Ulla-Britt MD, PhD
Despite every intention of providing good fetal monitoring during labor, adverse fetal outcomes still occur.1,2 The failure of monitoring is probably the result of the inadequacy of the methods used, but can also be the result of staff failure to react to warning signals from the fetus. The cardiotocography gives a high number of false-positive alarm signals, ie, of those with a pathologic cardiotocography trace, only a small proportion are actually hypoxic at birth.3,4 A pathologic cardiotocography trace during ongoing labor is therefore recommended in combination with a fetal blood sample for pH or lactate analysis.5–7 However, fetal blood sampling is invasive and noncontinuous, and sampling failure is reported to be as high as 20%.8 Furthermore, the measurement of pH and lactate in fetal blood has an unsatisfactory low predictive value for adverse fetal outcome at birth.7
Since the 1970s, it has been known that amniotic fluid contains a high concentration of lactate, but its origin has so far been unknown.9,10 Human uterine smooth muscle cell metabolism and the mechanism by which these cells provide the energy needed during labor are still unclear. We have previously shown that human myometrial cells can produce and deliver lactate to the extracellular medium.11 The production of lactate increases during hypoxia. If uterine contractions are irregular and long-standing, this may thus lead to an increased production of lactate that seems to diffuse into the amniotic fluid. Irregular and long-standing myometrial contractions will lead to myometrial tissue hypoxia with a decreased placental blood flow over time and with increased risk of hypoxia for the fetus.
The aim of the present study was to estimate the association among a high concentration of lactate in amniotic fluid (as a possible marker of uterine tissue hypoxia during delivery), pathologic cardiotocography trace, and adverse neonatal outcome at delivery.
MATERIALS AND METHODS
This is a secondary analysis of data from a study performed in two labor wards (Soder Hospital in Stockholm and Central Hospital in Karlstad) in Sweden between November 2006 and May 2008.12
During the study period, approximately 2,000 women were eligible to be included in this trial. The inclusion criteria were pregnant women with a full-term (37 weeks or greater), low-risk, singleton pregnancy in cephalic presentation. The women had to be in spontaneous and active labor. The definition of active labor was a cervical dilation 3 cm or greater and regular and painful contractions at a frequency of at least four contractions every 10 minutes. When the lactate measurement device was available for use, parturients who fulfilled the inclusion criteria were asked for their consent to have them participate in the study. In total, 25 of 850 of the women initially included did not meet the criteria for eligibility for inclusion and they were therefore excluded from analysis (Fig. 1).
Samples of amniotic fluid for lactate analysis were collected from flowing amniotic fluid by the midwife at every vaginal examination (median five times per delivery; range, 1–16 times) during the labor process.
If no spontaneously flowing amniotic fluid was available, a small disposable intrauterine catheter was used to gather approximately 150 μL of amniotic fluid vaginally. The amniotic fluid sample was then collected in a test tube, and lactate concentration in amniotic fluid was analyzed immediately at the bedside in the delivery room. Results were available within 15 seconds but they were not communicated to the woman in labor, the attending midwife, or the obstetrician in charge, in line with the study design. The last sample of amniotic fluid collected within 30 minutes before delivery was used in our calculation to examine associations between high lactate value in amniotic fluid as a marker of uterine hypoxia and adverse neonatal outcome at birth.
All 825 deliveries included in the study underwent cardiotocography monitoring during labor, of which 734 (89%) had a continuous recording during the last 30 minutes before delivery. All cardiotocography recordings were then blindly reviewed by two of the authors (E.W.I. and H.A.). The review was made in accordance with the Federation of Gynaecology and Obstetrics guidelines for the use of fetal monitoring.13 According to the Federation of Gynaecology and Obstetrics guidelines, a normal cardiotocography was described as having a baseline of 110–150 beats per minute, a baseline variability of 5–25 beats per minute, and no decelerations. A pathologic cardiotocography was defined as baseline heart rate less than 110 beats per minute or greater than 150 beats per minute, heart rate variability less than 5 beats per minute, pathologic decelerations (severe variable, severe repeated early, prolonged late, or sinusoidal), or all of these. Bradycardia was defined as a decrease in the fetal heart rate below 110 beats per minute for more than 3 minutes. Cardiotocography recordings during the last 30 minutes before delivery were used in this calculation.
The umbilical cord was clamped immediately after delivery, before the newborn's first cry, and arterial and venous blood samples were drawn from a double-clamped segment of the umbilical cord and analyzed within a few minutes (pH, base deficit, and lactate). All cord blood analyses were performed using the Radiometer automated ABL 800 Flex analyzer available in the labor wards. The base deficit was calculated from the blood compartment applying the algorithm used by Radiometer blood gas analyzers, recently reported to show a higher association with neonatal depression than base deficit calculated from the extracellular fluid compartment.
Because hemoglobin concentration in cord blood was not recorded, we used the general approximation of a hemoglobin concentration of 150 g/L.
In 130 deliveries with a pathologic cardiotocography trace, the attending obstetrician decided to sample fetal scalp blood for lactate analysis.14–17 At the same time, a sample of amniotic fluid was obtained for lactate analysis. Exact times for all sampling occasions (fetal blood per amniotic fluid) as well as routine parameters from the partogram and medical files were recorded in the study protocol.
The midwife or the pediatrician on call determined the Apgar score at 1, 5, and 10 minutes after delivery.
Clinical markers and symptoms that have been closely associated with intrauterine asphyxia were defined as adverse neonatal outcome: 1) pH less than 7.10 in umbilical artery; 2) metabolic acidosis, ie, pH less than 7.05 and base deficit greater than 12 in umbilical artery; 3) 5-minute Apgar score less than 7; 4) meconium aspiration syndrome and the need for neonatal intensive care treatment; 5) the need for resuscitation resulting from fetal distress; 6) development of hypoxic–ischemic encephalopathy grade 1–3; 7) admission to the neonatal intensive care unit for more than 24 hours; or all of these. Some of the neonates who were classified as having adverse neonatal outcome had more than one of the factors identified.
Lactate concentration in amniotic fluid was determined using a bedside device. The device measures lactate concentration in amniotic fluid between 0.5 and 25.0 mmol/L and has a coefficient of variation of approximately 3% at a lactate concentration of 11 mmol/L.12 The device was modified for this study to allow for the blind testing and was placed at the bedside. The results of the amniotic fluid lactate analysis were concealed from the clinician, midwives, and the woman undergoing labor during the whole labor process but were shown after delivery.
A cutoff value denoting a high concentration of lactate in amniotic fluid was set at greater than 10.1 mmol/L in line with the results of our previous publications.12,18
To investigate the possible influence of meconium and blood in the measurement of lactate in amniotic fluid, 10 mL of nonmeconium-stained amniotic fluid was mixed with 2, 4, or 6 mL of meconium.
The concentration of lactate in amniotic fluid was then analyzed repeatedly. Meconium concentration did not influence the measurement of the lactate concentrations in this in vitro test. Similar tests were performed with a mixture of blood and amniotic fluid. With a high concentration of blood, lactate concentration could be affected because red blood cells contain a large amount of lactate. Amniotic fluid samples with a large amount of blood have therefore been removed from our analysis.
Lactate in fetal scalp blood samples was analyzed in this study according to the guidelines effective at the participating hospitals. A commercially available microvolume test strip device was used. This method has previously been evaluated in intrapartum fetal monitoring.7,18,19 Guidelines for interpretation of the fetal scalp blood analyses were as follows: lactate less than 4.2 mmol/L normal, 4.2–4.8 mmol/L preacidemia, and greater than 4.8 mmol/L acidemia. In cases of preacidemia, repeated sampling of fetal scalp blood was recommended within 20–30 minutes if no other indication for intervention was present. The attending clinician decided on the time and mode of delivery in fetuses with acidemia.
This is a secondary analysis of data collected in a study in which the association between high levels of lactate in amniotic fluid and an operative intervention resulting from dysfunctional labor was examined.12 The power calculation for the main study showed that 850 women had to be included to ensure clinical significance (can be shown on request). The power calculation was performed in Sample Power 2.0.
Background data of the population studied and fetal outcome was presented as frequencies (%), years, days or median (range). Wald chi-square statistics and the Mann–Whitney test were used to compare deliveries with and without adverse neonatal outcome (Tables 1 and 2).
To evaluate the predictive capabilities of lactate concentrations in AF as a possible marker of uterine tissue hypoxia during delivery, we calculated sensitivity, specificity, and positive (PPV), negative (NPV) predictive values (Table 3).
Logistic regression20 was used to estimate the association between the primary outcomes (ie, adverse neonatal outcome) and amniotic fluid lactate level greater than 10.1 mmol/L at last sample before delivery (yes or no), pathologic cardiotocography trace when attending the delivery ward (yes or no), cardiotocography trace with a tachycardia within 30 minutes before delivery (yes or no), cardiotocography trace with bradycardia within 30 minutes before delivery (yes or no), lactate in fetal scalp blood greater than 4.2 mmol/L (yes or no), and lactate in fetal scalp blood greater than 4.8 mmol/L (yes or no).
Our model strategy was as follows: first, unadjusted associations with each risk factor were studied in a univariable model. Second, the adjusted associations with respect to the risk factors measured were studied in a multivariable model. All variables were tested using Wald chi-square statistics and were considered significant if P<.05, two-tailed.
Statistical analyses were performed in SPSS 17.0, and the statistical Package Statistica for Windows 8. The study was approved by the regional ethics committee at Karolinska Institute, Stockholm (2006/718-31/3).
Figure 1 shows a flow diagram of the 850 participating women. Table 1 describes baseline demographic data for the study population. A total of 6.3% (52 of 825) of all neonates were classified as having an adverse neonatal outcome at delivery according to the definition.
No significant differences in demographic data were seen between the groups except that more operative interventions were made in the group with adverse neonatal outcome and also that more boys had adverse neonatal outcome.
In the group with high amniotic fluid lactate concentrations greater than 10.1 mmol/L at the last sampling occasion before delivery, significantly more neonates had an adverse neonatal outcome as compared with deliveries with amniotic fluid lactate less than or equal to 10.1 mmol/L (11.4% compared with 2.9%), resuscitation was performed more frequently (1.5% compared with 0%), and a higher number of newborns were admitted to the neonatal intensive care unit (3.9% compared with 1.6%). The two neonates with hypoxic–ischemic encephalopathy grade 2 both belonged both to the group with high amniotic fluid lactate, whereas there were no newborns in the group with lower amniotic fluid lactate that developed hypoxic–ischemic encephalopathy (0.06% compared with 0%).
The fetal heart rate was monitored by cardiotocography in all deliveries included in the study. Among the 52 deliveries resulting in adverse neonatal outcome, 42 of 52 (81%) were considered to have a pathologic cardiotocography recording within 30 minutes before delivery. The remaining 10 deliveries were considered to have a normal cardiotocography.
Bradycardia, according to the definition by the Federation of Gynaecology and Obstetrics, occurred in eight of 48 deliveries (16.7%) with severe fetal outcome and showed the most significant association with adverse neonatal outcome at birth (odds ratio [OR] 7.4, 95% confidence interval [CI] 3.0–18.1). If bradycardia was present together with lactate in amniotic fluid greater than 10.1 mmol/L, the odds of delivering a neonate with an adverse neonatal outcome increased almost fourfold (OR 10.7, 95% CI 3.7–31.7) compared with deliveries with bradycardia and a low amniotic fluid lactate value (OR 3.3, 95% CI 0.7–15.6). Adjustment for all potential risk factors did not change the magnitude of the association between a high level of lactate in amniotic fluid at the last sampling occasion before delivery and the adjusted odds of delivering a neonate with an adverse fetal outcome. In our multivariate model (Table 4), the odds of delivering a neonate with adverse fetal outcome were increased approximately 10 times (OR 9.8, 95% CI 1.11–87.46) if the lactate level in amniotic fluid was high (greater than 10.1 mmol/L) at the last sampling occasion before delivery.
In the group with adverse neonatal outcome (n=52), a sample of fetal blood for analysis of lactate was obtained in 13 cases within 60 minutes before delivery as a result of an abnormal cardiotocography. None of these 13 samples had a high value of lactate in fetal blood (ie, lactate greater than 4.8 mmol/L). However, 12 of 13 (92%) had a high value of lactate in amniotic fluid (greater than 10.1 mmol/L) when lactate in amniotic fluid was analyzed retrospectively. All 13 had a severe fetal outcome at delivery.
In the remaining 39 deliveries with adverse neonatal outcome, no fetal blood samples were obtained during labor. In these cases, cardiotocography was either considered to be normal by the midwife or obstetrician in charge and no further action was taken or so abnormal that delivery was carried out immediately. When reviewing the analysis of lactate in amniotic fluid in these 39 deliveries, 28 of 39 (72%) had a high level of lactate (greater than 10.1 mmol/L) in amniotic fluid present at the last sampling occasion before delivery.
If lactate concentration in AF greater than 10.1 mmol/L was used as a possible marker of uterine tissue hypoxia during delivery, a sensitivity of 62% (58–65%), a specificity of 73% (59–84%), a PPV of 97% (95–98%), and a NPV of 11.4% (8.3–15.4%) was presented. Same figures for CTG used the last 30 minutes before delivery was: sensitivity 90% (77–96), Specificity 50% (46–54), PPV 98% (96–99) and NPV 12% (8–14) (Table 3).
The result of this large prospective observational cohort study of 825 healthy delivering women demonstrates an overall increased incidence of neonatal complications if a high level of lactate in amniotic fluid was present shortly before delivery.
During normal labor contractions, the uteroplacental blood flow decreases but rapidly restores when the contraction is over. When labor contractions are irregular and ineffective, the uteroplacental blood flow is reported to be reduced for a longer period or constantly. As a consequence, gas exchange in the fetal-maternal unit is impaired, resulting in fetal hypoxia–ischemia and lactic acidosis.
Several experimental studies have shown that the first compensatory mechanism of the hypoxic fetus is tachycardia followed by increasing bradycardia if the hypoxia persists.21 In this study, 61% of the cardiotocography recordings 30 minutes before delivery were considered abnormal according to the guidelines. Nonetheless, only 9.4% of the abnormal traces were associated with adverse neonatal outcome at birth. This result demonstrates the limitation of using a cardiotocography recording as the only means of fetal surveillance during labor and reflects the findings of other studies showing the low predictive value of an abnormal cardiotocography on fetal outcome (Table 3)22 Among the deliveries that had a pathologic cardiotocography during the last 30 minutes and actually resulted in adverse neonatal outcome, almost 80% also had also a high level of amniotic fluid lactate greater than 10.1 mmol/L at the last sampling occasion before delivery. Consequently, high amniotic fluid lactate seems to be a good marker of uterine hypoxia and, when used together with a cardiotocography recording, the predictive value for adverse neonatal outcome is increased. Fetal bradycardia together with a high concentration of lactate in amniotic fluid before delivery indicated the highest risks for adverse neonatal outcome.
For the best obstetric care, close fetal surveillance is crucial, and physicians must carefully consider when to intervene.2 In 1962, Saling introduced a new method of sampling fetal scalp blood during labor for analyzing pH as an indicator of hypoxia.23–25 This technique has been regarded since then as one of the “gold standards” for identifying intrapartum fetal hypoxia. However, the well-known limitations of the method (invasiveness and a high frequency of sampling failure) have stimulated alternative developments, including during the last decade the evaluation of microvolume techniques for measuring lactate in fetal scalp blood.15–18 In the present study, a large number of fetal blood samples for analysis of lactate in fetal blood were performed as a result of an abnormal cardiotocography trace pattern, but none of those in which a high value of lactate in fetal blood (greater than 4.8 mmol/L) was present had an adverse neonatal outcome. It is of course of great importance to have as few false-negative and false-positive tests as possible when a clinical test is used, especially given that these results always lead to a decision. This knowledge has to be taken in consideration when fetal scalp blood is used as the “gold standard” in fetal monitoring. To use lactate in amniotic fluid as a marker of uterine tissue hypoxia together with an already existing opportunity for monitoring might be a new and important clinical way to identify a risk group that requires intensified fetal monitoring during labor.
The interpretation of this work is that fetal monitoring during labor is difficult. Signs of severely threatened fetal well-being may be hard to detect even with careful monitoring. When the cardiotocography recordings were analyzed retrospectively in this work, subtle signs of fetal distress could be detected in 13 deliveries, even if the signs were not obvious to the midwife or obstetrician in charge during ongoing labor. In 12 of these 13 deliveries, a lactate value in amniotic fluid greater than 10.1 mmol/L at the last sampling occasion before delivery was shown.
Some limitations of this study should be noted. The number of deliveries with adverse outcome was limited probably as a result of the inclusion criteria, ie, all women included had to be healthy with a normal pregnancy and spontaneous onset of labor. This means that all the deliveries belonged to a selected low-risk pregnancy group and the incidence of fetuses delivered with an adverse neonatal outcome was low. A further limitation is that not all the women eligible to be included in this trial were actually included because the lactate measurement device could only be used for one delivery at a time. However, this can reasonably be regarded as a random dropout. Future studies in different contexts should be completed to evaluate appropriate cutoff points for these populations as well as the broader clinical usefulness of this information.
The hypothesis that hypoxia of the myometrium with increased lactate production reflected in an increased concentration of lactate in amniotic fluid is very appealing. It is a reasonable explanation of why adverse neonatal outcome is overrepresented in the group with a lactate value in amniotic fluid greater than 10.1 mmol/L. It may also explain why the duration of labor itself not is reported as influencing the outcome, according to previous studies.2 If the myometrial tissue is well oxygenated, the neonatal outcome may not be affected. Like with many obstetric concerns, considering when to intervene should entail an evaluation of the risks of further expectant management compared with the risks of intervention. The lactate concentration in amniotic fluid together with cardiotocography appears to be a promising noninvasive method of improving risk assessments in bedside obstetrics.
In summary, we have found that the use of cardiotocography recordings together with an analysis of lactate concentration in amniotic fluid could be a useful predictor of fetal outcome in labor. The method is easy, noninvasive, and safe for the mother and her unborn child.
Our findings have important clinical implications in view of the fact that children are stillborn with an unexpected adverse neonatal outcome, even with what is considered to be careful fetal surveillance.
1. Altman MR, Lydon-Rochelle MT. Prolonged second stage of labor and risk of adverse maternal and perinatal outcomes: a systematic review. Birth 2006;33:315–22.
2. Hellman LM, Prystowsky H. The duration of the second stage of labor. Am J Obstet Gynecol 1952;63:1223–33.
3. Hon EH, Hess OW. Instrumentation of fetal electrocardiography. Science 1957;125:553–4.
4. Hon EH, Petrie RH. Clinical value of fetal heart rate monitoring. Clin Obstet Gynecol 1975;18:1–23.
5. Bretscher J, Saling E. pH values in the human fetus during labor. Am J Obstet Gynecol 1967;97:906–11.
6. Westgren M, Kuger K, Ek S, Grunewald C, Kublickas M, Naka K, et al. Lactate compared with pH analysis at fetal scalp blood sampling: a prospective randomised study. Br J Obstet Gynaecol 1998;105:29–33.
7. Wiberg-Itzel E, Lipponer C, Norman M, Herbst A, Prebensen D, Hansson A, et al. Determination of pH or lactate in fetal scalp blood in management of intrapartum fetal distress: randomised controlled multicentre trial. BMJ 2008;336:1284–7.
8. Tuffnell D, Haw WL, Wilkinson K. How long does a fetal scalp blood sample take? BJOG 2006;113:332–4.
9. Fadel HE, Northrop G, Misenhimer HR, Harp RJ. Acid-base determinations in amniotic fluid and blood of normal late pregnancy. Obstet Gynecol 1979;53:99–104.
10. Brace RA. Physiology of amniotic fluid volume regulation. Clin Obstet Gynecol 1997;40:280–9.
11. Åkerud H, Ronquist G, Wiberg-Itzel E. Lactate distribution in culture medium of human myometrial biopsies incubated under different conditions. Am J Physiol Endocrinol Metab 2009;297:E1414–9.
12. Wiberg-Itzel E, Pettersson H, Andolf E, Hansson A, Winbladh B, Akerud H. Lactate concentration in amniotic fluid: a good predictor of labor outcome. Eur J Obstet Gynecol Reprod Biol 2010;152:34–8.
13. Schiermeier S, Pildner von Steinburg S, Thieme A, Reinhard J, Daumer M, Scholz M, et al. Sensitivity and specificity of intrapartum computerised FIGO criteria for cardiotocography and fetal scalp pH during labour: multicentre, observational study. BJOG. 2008;115:1557–63.
14. Nordstrom L, Ingemarsson I, Kublickas M, Persson B, Shimojo N, Westgren M. Scalp blood lactate: a new test strip method for monitoring fetal wellbeing in labour. Br J Obstet Gynaecol 1995;102:894–9.
15. Westgren M, Kublickas M, Kruger K. Role of lactate measurements during labor. Obstet Gynecol Surv 1999;54:43–8.
16. Kruger K, Kublickas M, Westgren M. Lactate in scalp and cord blood from fetuses with ominous fetal heart rate patterns. Obstet Gynecol 1998;92:918–22.
17. Kruger K, Hallberg B, Blennow M, Kublickas M, Westgren M. Predictive value of fetal scalp blood lactate concentration and pH as markers of neurologic disability. Am J Obstet Gynecol 1999;181:1072–8.
18. Wiberg-Itzel E, Pettersson H, Cnattingius S, Nordstrom L. Association between lactate concentration in amniotic fluid and dysfunctional labor. Acta Obstet Gynecol Scand 2008;87:924–8.
19. Shimojo N, Naka K, Uenoyama H, Hamamoto K, Yoshioka K, Okuda K. Electrochemical assay system with single-use electrode strip for measuring lactate in whole blood. Clin Chem 1993;39:2312–4.
20. Hosmer DW, Lemenshow S, Applied logistic regression. 2nd ed. Wiley Series in Probability and Mathematical Statistics. New York (NY): Wiley; 2000.
21. Karlsson M, Tooley JR, Satas S, Hobbs CE, Chakkarapani E, Stone J, et al. Delayed hypothermia as selective head cooling or whole body cooling does not protect brain or body in newborn pig subjected to hypoxia–ischemia. Pediatr Res 2008;64:74–8.
22. Thacker SB, Stroup DF. Continuous electronic heart rate monitoring for fetal assessment during labor. The Cochrane Database of Systematic Reviews 2000, Issue 2. Art. No.: CD000063. DOI: 10.1002/14651858.CD000063.
23. Saling E. A new method for examination of the child during labor. Introduction, technic and principles [in German]. Arch Gynakol 1962;197:108–22.
24. Saling E. Microblood studies on the fetus. Clinical application and 1st results [in German]. Z Geburtshilfe Gynakol 1964;162:56–75.
25. Saling E, Schneider D. Biochemical supervision of the foetus during labour. J Obstet Gynaecol Br Commonw 1967;74: 799–811.