Are we ready for widespread implementation of lactate POCT in fetal blood sampling? An analytical and clinical verification of the novel statstrip POCT analyzer

Abstract

Objectives

Currently, pH values are used in fetal scalp blood sampling as a parameter to rule out fetal distress during the delivery. Due to a high number of pre-analytical errors in pH measurements, lactate measurement is already extensively examined as an alternative. Our objectives were to confirm the analytical performance of the StatStrip lactate POCT analyzer and to compare pH and lactate as a marker of fetal distress.

Design

and methods: Fetal blood scalp, umbilical and arterial blood test results (n = 100) were analyzed to compare the current POC blood gas analyzer including lactate (iSTAT-1) with the new POCT analyzer (StatStrip) to test the analytical performance. Furthermore, in all fetal scalp blood tests with a lactate en pH measurement from 2021 to 2023, clinical delivery data was collected to perform a clinical verification.

Results

The lactate concentration on the StatStrip analyzer correlated well with the iSTAT-1 (Pearson's r ≥ 0.95). 73 Fetal scalp blood tests showed 18% discrepant results when comparing pH and lactate with regard to fetal distress and consequent delivery intervention. Lactate showed more false positive results than pH (4 versus 1), but no false negatives as opposed to pH (0 versus 1).

Conclusions

The lactate Statstrip and iSTAT-1 POCT analyzers were analytically equivalent. The clinical verification study showed that lactate is a good predictor of fetal distress, although more false positive results were found in our limited dataset. However, unnecessary interventions due to failed pH measurements might be prevented when a lactate measurement is introduced.

Highlights

• •Evidence for lactate as reliable marker to rule out fetal distress during labor piles up.
• •Implementation of lactate instead of or in combination with pH seems not yet widespread.
• •Lactate Statstrip POCT analyzer is valid for measurement in fetal scalp blood.
• •Lactate is a reliable marker to rule out for metabolic acidosis in our study.
• •Further testing is required to confirm decision value and pre-acidosis interval.

Abbreviations

CTG

cardiotocogram

ED

emergency department

FBS

fetal blood sampling

IC

intensive care

LCI

lower confidence interval

NPV

negative predictive value

pCO2

partial pressure of carbon dioxide

PFD

pulmonary function department

PPV

positive predictive value

UCB

umbilical cord blood

UCI

upper confidence interval

1. Introduction

Fetal monitoring during labor is important to prevent adverse outcome arising from metabolic acidosis. Fetal heart rate monitoring is used as an indicator of fetal distress and is measured by the cardiotocogram (CTG). The CTG also measures uterine contraction. Although a normal CTG can rule out hypoxia, an abnormal CTG has a false positive outcome in at least 50% of cases [1]. Therefore, Saling [2] introduced fetal blood scalp sampling (FBS) so that additional information about the well-being of the fetus can be provided in case of an abnormal CTG. The pH value serves as the gold standard, with acidosis being defined as pH ≤ 7.20 requiring immediate intervention (e.g. delivery of the baby) [2]. A pH value between 7.20 and 7.25 is considered pre-acidosis. Repeating FBS after 30 min is advised. With pH values ≥ 7.25, it is assumed that there is no fetal distress. However, pH samples might provide false results as a result of admixture of air, amniotic fluid or maternal blood [3] or may be unsuccessful because of pre-analytical errors such as clotting or too little sample volume. Failure of pH testing may result in unnecessary intervention if fetal distress cannot be ruled out.

Lactate is an alternative parameter to use in FBS, since this is the end product of the anaerobic dissimilation and therefore a measure of the imbalance between oxygen supply and demand. Current lactate POC analyzers have practical advantages: less blood volume is needed, which means that the measurement will fail considerably less often (1–3%) compared to pH-testing (10–23%) and the results come in faster (13 s (StatStrip) compared to 2 min (iSTAT-1)) [[4], [5], [6]]. Furthermore, the prediction of neonatal outcome with lactate values is equivalent to pH-values with no significant difference in the number of delivery interventions [7,8]. However, there is no consensus in lactate cut-off values for indicating direct intervention; acidotic cut-off values range between 4.8 and 6.6 mmol/L depending on the used test device and the preferred test characteristics: i.e. high sensitivity or high specificity [7,9,10].

In this study, an analytical verification of the StatStrip lactate POCT analyzer was performed on FBS samples. In addition, a single-center peripheral hospital clinical verification was performed to compare pH and lactate FBS test results as a diagnostic marker of fetal distress in the Obstetrics department (1650 deliveries/year). Finally, an optimal and safe decision limit for the lactate value on the StatStrip LAC/Hb/Hct was investigated.

2. Materials & methods

2.1. Protocol for FBS

A blood sample is taken from the fetal scalp in case of fetal heart rate at risk of acidosis, according to the FIGO classification for CTG monitoring [11] and national guideline fetal monitoring of the Dutch College of Obstetricians and Gynecologists (NVOG) [12]. One hundred mL of blood is sampled by the midwife or gynecologist by using a capillary tube after drying of the fetal scalp with swab and applying paraffin for proper drop formation. The sample is analyzed immediately in the delivery room using both the iSTAT and the StatStrip lactate POCT analyzer. 95 μL Is used for iSTAT measurement of pH and lactate and 0.6 μL of fetal blood is used for StatStrip lactate POCT analyzer measurement. According to the department protocol, the gynecologist considers only the scalp pH value in deciding whether labor should be prolonged or immediate delivery is needed. Sampling may be repeated several times during labor, depending on pH results and the following CTG classification.

2.2. Analytical verification

The intended use of the lactate POCT meter is for FBS measurements, but there was a limited amount of FBS samples available with a limited range of lactate concentrations. Therefore, for the analytical verification, umbilical cord samples were added, as well as samples from the Emergency Department (ED), the Intensive Care (IC) and the Pulmonary Function Department (PFD) with a lactate value > 4 mmol/L.

For the analytical performance of the StatStrip LAC/Hb/Hct (serial no. 262248522097), the repeatability and reproducibility were determined. The repeatability was determined by testing three samples with different lactate concentration ten times consecutively within 45 min. For this purpose, 0.6 μL whole blood sample was used per StatStrip Lactate strip (article no. 61440). The reproducibility was determined by testing two Quality Controls (StatStrip LAC Control Solution Level 1 (0.3-0.6–0.9 mmol/L) (article no. 47553) and Control Solution Level 2 (5.0-6.3–7.5 mmol/L) (article no. 47553)). These were measured on ten different days.

A method comparison followed between iSTAT-1 lactate and StatStrip LAC/Hb/Hct lactate. FBS samples were tested on both the iSTAT-1 (serial no. 300471-B) using the CG4+ cassette (article no. 03P8525) and on the StatStrip LAC/Hb/Hct. UCB and samples from the ED, IC and PFD were tested on the iSTAT-1, StatStrip and additionally on the ABL-90 Flex (serial no. 1393 090R 0511N006).

The statistical analyses were performed using the R statistical package (version February 1, 5033).

2.3. Clinical verification

pH and lactate test results from January 2020 to April 2023 were reviewed from FBS samples tested on iSTAT-1 (serial no. 300471-B) (before January 2020, only FBS pH was measured). In November 2022, Statstrip lactate was introduced for this study, but not yet for clinical decision making. Samples with pCO₂ concentrations lower than 37 mmHg were excluded for further analysis, because of suspected pre-analytical air admixture [13].

Test characteristics were calculated for multiple decision values, including specificity, sensitivity, positive predictive value (PPV) and negative predictive value (NPV). Cross-tables were made with lactate and pH and their performance in predicting fetal distress. From these cross-tables, concordant and discordant test results were calculated.

In case of a discrepancy, the following additional clinical data were collected from the electronic patient record: CTG, delivery intervention, time of FBS and time of birth, APGAR-score, pH concentrations of umbilical cord blood (UCB) and notes from the gynecologists and pediatrician. The overall judgement of clinical performance data from the electronic patient record of the neonate at birth was done by an experienced gynecologist.

Ethical approval

All women giving birth at the St Jansdal hospital are informed that their data are computerized and can be used to evaluate medical practices. Under Dutch regulations, this study is exempt from medical ethical committee review, being an observational study and conducted for the purpose of quality improvement of standard care. The extra measurement on the Statstrip lactate was performed on blood that was already collected in regular clinical care.

3. Results

3.1. Analytical verification

3.1.1. Repeatability

In the intra-run variation, the following Coefficients of variation (CV%) were obtained: 4.2%, 4.0% and 7.4% at a concentration of 5.2 mmol/L (4.6–5.7 mmol/L), 7.0 mmol/L (6.6–7.5 mmol/L) and 13.3 mmol/L (12.2–14.9 mmol/L), respectively (n = 10). No outliers were observed according to the average ± 2*SD calculation. The claimed CVs from the package insert were 3.4% and 5.8% at a concentration of 10.52 mmol/L and 2.16 mmol/L respectively (n = 20) [12]. These CVs were higher than the CVs from the manufacturer and therefore, after extra pre-analytical instructions from the manufacturer, a repeated intra-run experiment showed a CV of 2.9% at a concentration of 6.3 (6.0–6.5 mmol/L).

3.1.2. Reproducibility

In the inter-run variation, a CV% of 16.7% and 8.0% were found at a concentration of 0.7 mmol/L (0.5–0.8 mmol/L) and 8.4 mmol/L (7.5–9.2 mmol/L), respectively (n = 10). No outliers were observed (< average ± 2*SD). The claimed CV% in the package insert were 5.1% and 4.7% at a concentration of 1.71 mmol/L and 10.4 mmol/L, respectively (n = 80) [12].

3.1.3. Method comparison

For the method comparison, a total of 100 samples were collected (12 FBS samples, 42 UCB samples and 42 samples from the ED/IC/PFD). Fig. 1 shows Passing-Bablok regressions and Bland altman plots of iSTAT-1 and StatStrip lactate concentrations for the different sample types. For FBS samples, a Pearson's r of 0.95 was calculated (Fig. 1A), for UCB samples 0.97 (Fig. 1B) and for ED/IC/PFD samples 0.98 (Fig. 1C).

• A.iSTAT versus Statstrip lactate – sample type: fetal scalp blood (FBS)

Fig. 1.

Passing-Bablok regression (left) and Blant Altman difference plots (right) of iSTAT lactate and StatStrip lactate with data of (A) FBS samples (n = 12, r = 0.95), (B) UCB samples (n = 42, r = 0.97) & (C) ED/IC/PFD (n = 38, r = 0.98). The current method (iSTAT lactate) is displayed on the x-axis and the new method (StatStrip lactate) on the y-axis. For all PB regression fit lines, the 0.95-confidence bounds were calculated with the bootstrap (quantile) method.

The mean bias for all sample types ranged from −0.10 (−2.92%) to 0.13 (−0.72%).

The same Passing-Bablok regression plots were made for the ABL-90 Flex lactate and StatStrip LAC/Hb/Hct lactate (supplemental fig 1). A Pearson's r of 0.95 was calculated for UCB samples and 0.98 for ED/IC/PFD samples. The bias was −0.49 (6.3%) for UCB samples and 0.46 (4.1%) for ED/IC/PFD samples.

In FBS and UCB samples, a correlation with R² = 0.5 was calculated in the comparison between iSTAT-1 pH and StatStrip lactate.

3.2. Clinical validation

pH and lactate measurements of 100 FBS samples were collected determined on iSTAT-1 and/or StatStrip. A summary of the sample characteristics in this clinical validation are shown in Fig. 2. The mean scalp pH value was 7.28 and the mean scalp lactate concentration 4.3 mmol/L. The pH scalp value was normal for 73% of samples (pH > 7.25), pre-acidotic for 17% (pH 7.21–7.25) and acidotic for 10% (pH < 7.20).

Fig. 2.

Flow chart for included FBS samples. Discrepant results for pH and lactate showed a different conclusion for acidosis, pre-acidosis or a normal metabolic status. Cut-off values were lactate ≥5.2 and pH < 7.20 NB the 100 samples included in the clinical validation were not the same samples as used for the analytical validation.

In total, 84 of 100 FBS samples were reliable based on pCO₂ values. In Table 1, five different decision values mentioned in other guidelines or studies [[7], [8], [9], [10]] were used to calculate the test characteristics for our StatStrip LAC/Hb/Hct measurement for FBS testing.

Table 1.

Test characteristics for different decision values determined on 84 reliable FBS samples for the prediction of acidosis (pH is the golden standard).

Decision value (≥mmol/L)	4.8	5.0	5.2	5.7	6.6
Specificity (%)	79	82	83	85	96
Sensitivity (%)	77	69	69	69	54
PPV (%)	40	41	43	45	70
NPV (%)	95	94	94	94	92

The sensitivity (69%) and NPV (94%) remain the same for the decision values 5.0–5.7 mmol/L, which can be explained by the limited number of data points in this range. At the decision value of 6.6 mmol/L, the sensitivity decreases to 54% and the NPV to 92%. The specificity varies from 79% at 4.8 mmol/L to 96% at 6.6 mmol/L, with the PPV varying from 40% to 70%.

Table 2 shows a cross-table for pH and lactate as a marker of fetal distress with the following decision values: acidosis: pH ≤ 7.20; lactate ≥6.6 mmol/L, pre-acidosis: pH 7.21–7.24; lactate 5.3–6.6 mmol/L) normal values: pH ≥ 7.25; lactate ≤5.2 mmol/L, conform the national guideline and a current large retrospective research [10].

Table 2.

Cross-table of reliable FBS samples (n = 84) measured on iSTAT-1 in period January 2020 to April 2023, Grey: concordant results, orange: detailed patient history in Table 3.

In 61 cases (grey marked), pH and lactate were concordant (71%). In 9 cases (orange marked); one parameter showed a normal metabolic state or pre-acidosis, while the other parameter showed acidosis (11%).

In Table 3, additional clinical data are shown for these 9 cases. None of these newborns needed to be admitted to a Neonatal Intensive Care Unit. The CTG image, UCB pH value, APGAR-score and overall clinical picture were determined to reflect on the condition of the child after birth. For all cases, the time interval between FBS and time of birth was <75 min, except one (127 min). False positive or false negative results for pH and/or lactate were attempted to be identified by using umbilical cord pH and BE, Apgar scores and clinical signs of asphyxia as markers for fetal distress. However, these parameters were discordant as well in most cases. Eventually, we classified false positive and false negative Statstrip lactate results based on clinical signs of asphyxia as endpoint.

Table 3.

ix clinical cases with discrepant test results between pH and lactate. CS: Cesarean Section OVD: Operative Vaginal Delivery, SVD: Spontaneous Vaginal Delivery ^aCut-off values were lactate ≥5.2 and pH < 7.20, ^cClinical signs were judged to identify asphyxia after birth: green cases were judged as non-asphyxia and yellow as possible/mild asphyxia FP: False positive, FN: False negative based on clinical signs of asphyxia.

4. Discussion

The analytical verification of the StatStrip LAC/Hb/Hct lactate determination showed sufficient results with respect to reproducibility and repeatability. Our repeatability results underline the need for careful pre-analytic standardization of Statstrip lactate measurement to obtain low CVs. In the method comparison with iSTAT and ABL-90 Flex, a Pearson's r of 0.95 or higher was calculated with a negligible bias.

Our data showed a weak correlation between pH and lactate concentrations (R² = 0.50 with FBS samples and R² = 0.50–0.55 with UCB samples). The same correlation was seen in earlier studies with larger study populations with R² = −0.5 [8] and an agreement rate estimated by Cohen's Kappa of −0.36 [12]. Apparently, pH and lactate are two different parameters of hypoxemia that show limited correlation to each other, but both are good predictors of fetal distress [8].

Because of the limited number of samples, it was not possible to designate an optimal safe decision limit on the basis of our data. The calculated test characteristics differed minimally with a change in the decision limit. In our study, we therefore used a decision value of 5.2 mmol/L for our clinical validation based on a recent large prospective study (n = 3334) with Statstrip lactate for predicting an APGAR score <7 at 5 min [10].

Of 85 reliable FBS samples, pH and lactate values resulted in a different clinical decision (i.e. termination of the delivery or not) for fetal distress in 11% of cases, which is comparable to other studies [14,15].

When the clinical data and the course of labor were considered, we believe no definitive conclusions can be made since some parameters indicate an abnormal metabolic state, whereas others are normal. Thus, although we added a classification of false positive and false negative results for lactate or pH, we actually found such a classification does not cover the complexity of fetal distress and its determination. For example, in case 1 and 2: we classified the lactate as false positive based on the absence of signs of asphyxia. But the low umbilical cord pH and BE and Apgar score do indicate metabolic distress during labor. We therefore conclude that pH and lactate are different parameters measuring fetal distress; they are not interchangeable, but complementary. It appeared that both pH and lactate were false positive in some cases. There seemed no false negatives in our small dataset: i.e. asphyxia was never missed with pH or lactate. Thus, the negative predictive value (NPV) of pH and lactate is excellent in our study. Therefore, pH and lactate are both suitable to rule out fetal distress and are thereby a good extra test in preventing unnecessary obstetric interventions based on an abnormal CTG.

Previously published studies also confirm a good usability of both parameters with no significantly different number of deliveries with FBS sampling based on pH versus lactate [7,8], with the exception of one smaller study [15]. In our study, we found one case with a deep acidotic pH concentration with a normal lactate concentration, showing no signs of fetal distress after birth with a cesarean section, indicating a false positive pH.

Unfortunately, data on failed samples in our hospital were not available, but upon inquiry the staff experienced a high failure rate, discouraging FBS. We speculate that in a smaller peripheral hospital with less (complex) deliveries it might be even more important to have an easy-to-use POC test, because the staff is less experienced.

A strength of our study is the inclusion of a clinical verification. Some studies about lactate and pH FBS only include an analytical verification [16], whereas only a few comparable clinical studies were published [8,14]. We believe clinical outcome measurements are the preferred endpoint to judge the value of a new diagnostic marker. However, we experienced it to be complex to affirm fetal distress, because diagnostic and clinical parameters were not concordant (f.e. an acidotic umbilical cord pH with normal Apgar scores). Furthermore, the clinical condition of a neonate at birth does not always reflect the situation at the time of FBS. Therefore, cases with a prolonged time (>75 min) between FBS and birth are less useful in the interpretation of fetal distress. Moreover, an obstetric intervention such as a caesarian section is done to improve the clinical condition of the neonate at birth. Apgar-scores are therefore influenced by the intervention and might not match the situation at FBS.

Some hospitals choose to measure pH and lactate simultaneously in FBS samples assuming the combination improves sensitivity and specificity. Combined analyses are not unanimously recommended as they are thought the be increasing the number of interventions without decreasing metabolic acidaemia at birth [7]. In a study with 241 FBS tests by Liljeström et al. [14] combined analyses with pH and lactate showed the amount of interventions was slightly higher in the combined analysis group (27%) than in a single test group (22%). 41% Of all combined tests were discrepant for pH and lactate. We found it remarkable that even in the case of a positive pH ánd lactate concentration indicating fetal distress, not always an operative delivery was performed. In this study, no infants were born with metabolic acidosis. The possibility of using the combined test or a single test should be evaluated in future studies, as well as the optimal cut-off values.

Recently, the Dutch obstetrical guideline was revised and lactate was added as an equal alternative to pH for FBS [17]. Gynecologists are therefore experienced with the pH measurement and might be hesitant to implement a new parameter in the acute care setting of fetal distress during labor. Moreover, gynecologists might fear a higher false positive rate with lactate measurements, because the lactate concentration may be elevated during prolonged labor, sepsis or maternal fever without fetal distress [6]. However, in our study no higher false positive rates were found for lactate. Furthermore, although the pH was the golden standard for years, our limited dataset suggests that pH is also not always reflecting the fetal condition correctly either. Moreover, in several large Scandinavian studies, no differences in operative delivery rates were found using lactate [7,10].

Our study and all previous studies show convincing evidence for safe implementation of lactate as a marker for fetal distress during labor.

5. Conclusion

Our study shows a good analytic performance of the StatStrip lactate POCT analyzer and confirms its clinical applicability to rule out fetal distress during labor. Therefore, it is a good candidate to complement or replace the current pH POCT measurement, because of a significant reduction in pre-analytical errors.

Formatting of funding sources

This work was supported by Nova Biomedical by providing the StatStrip LAC/Hb/Hct POCT analyzer on loan during this study.

Units

Lactate is measured in mmol/L.

CRediT authorship contribution statement

Nienke Spronk: Formal analysis, Investigation, Validation, Writing – original draft. Madeleine SQ. Kortenhorst: Conceptualization, Formal analysis, Supervision, Validation, Writing – review & editing. Jasmijn A. van Balveren: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Supervision, Validation, Writing – original draft, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix ASupplementary data

The following is the Supplementary data to this article:

Data availability

Data will be made available on request.