Admission ultrasonography as a means of identifying unrecognized fetal compromise in term singleton pregnancies at the onset of labor – a feasibility study

Abstract Objectives Apparently uncomplicated low-risk pregnancies, especially first time births, account for a significant proportion of adverse birth outcomes. Improved risk stratification with a simple bedside scan on admission in early labor could potentially reduce adverse intrapartum outcomes. The aim of this feasibility study was to assess a cohort of low-risk subjects with admission ultrasonography at the onset of labor with a view to conducting a future randomized controlled trial (RCT). The objectives were three-fold; i) to determine the logistics of performing a labor admission ultrasound scan ii) to establish whether abnormal ultrasound features can be identified and iii) whether they are associated with emergency delivery and/or poor condition of the neonate at birth. Methods We performed a prospective cohort study of 295 participants with term singleton cephalic pregnancies admitted in early labor or for labor induction with non-fetal indications. The setting was a university teaching hospital in Ireland with almost 8000 births annually. A bedside ultrasound scan was performed to assess fetal biometry, amniotic fluid volume and placental maturity. Patients and their babies were followed up until hospital discharge. The outcomes of interest included image quality, time to perform a scan, oligohydramnios (Single Deepest Pool ≤ 2 cm), small for gestational age (SGA; abdominal circumference <10th centile), mature placenta (Grannum 2 or 3), pathological CTG, emergency cesarean section (CS), fetal acidosis (cord arterial pH <7.10 or base excess <-12.0), low Apgar score <7 at 5 min and neonatal unit admission. Results Image quality was optimal in 274 of the 295 scans (93%) and 271 (92%) were completed in less than 10 min. Of this low-risk population, 67 of 294 (23%) had oligohydramnios, 11 (4%) were small for gestational age and 87 (30%) had a mature placenta (Grannum grade 2). The incidence of pathological CTG and emergency CS was higher among patients with oligohydramnios than those with a normal scan but did not reach statistical significance; Odds Ratio 3.40 (95% Confidence Intervals 0.55 to 20.92) and OR 1.43 (95% CI 0.66 to 3.08) respectively. The mean birthweight was significantly lower in those with oligohydramnios −139 g (95% CI −248 to −30) and admission scan detected SGA −357 g (95% CI −557 to −137). Adverse perinatal outcomes were uncommon with a higher incidence of fetal acidosis (pH < 7.10) in the oligohydramnios group. The incidence of neonatal unit admission >24 h was higher in the oligohydramnios group but not statistically significant; OR 3.75 (95% CI 0.61 to 22.97). Results for SGA alone were non-significant and results for oligohydramnios and SGA combined were similar to those for oligohydramnios alone. Conclusions Admission ultrasonography is feasible in a routine clinical setting, but evidence of benefit is weak and does not currently justify a randomized controlled trial.


Introduction
A third of births in Ireland are by cesarean section (CS) with similar rates in Europe and the United States of America [1][2][3].Despite a doubling of CS rates in high income countries over the past two decades, rates of cerebral palsy among term normally formed infants have remained constant, affecting between 1.5 and 3 per 1000 births [4].Five hundred babies start labor apparently healthy each year in the United Kingdom and die a delivery-related death [5].The challenge lies in determining which fetuses have the reserve to tolerate the demands of labor and which have limited reserve that will be exhausted prior to safe delivery [6][7].
Antenatal assessment of fetal wellbeing has advanced over the years and fetuses with evidence of compromise are monitored closely and have a planned delivery.However, a major challenge for lowrisk pregnancies is that the interval between a normal fetal assessment and the onset of labor is variable and largely unpredictable.As such, there may be deterioration in the fetal status due to progressive placental insufficiency, and this may be reflected in how the fetus copes with the demands of labor.The planned place of birth and approach to fetal monitoring may no longer be appropriate for the clinical circumstances.What is required is a reliable inexpensive admission test of fetal well-being that can be performed and interpreted at the onset of labor by staff of average expertise.The admission CTG has been widely used as a screening tool but has performed poorly in terms of discriminating between normal and abnormal fetal reserve [8][9].A systematic review concluded that routine admission cardiotocography in low-risk patients increases the risk of CS without any benefit and should not be recommended [10].
Admission ultrasonography is worth consideration given that antenatal ultrasonography has been shown to reduce adverse perinatal outcomes [11].As ultrasound machines have become more compact, mobile and cheaper, they have become more readily available on labor wards and are being used increasingly in labor [12][13][14][15].Admission ultrasonography has the additional advantage of confirming the presentation and avoiding inadvertent breech presentation late in labor.
There are several possible approaches including assessment of fetal biometry, amniotic fluid volume and placental maturity.The aim of this feasibility study was to establish the logistics of performing admission ultrasonography at the onset of labor, whether it identifies evidence of fetal compromise, and if there is an association with emergency delivery and adverse perinatal outcomes that would warrant a future randomized controlled trial (RCT).

Study setting and participants
The Coombe Hospital, Dublin, is a consultant-led university teaching hospital with 7500-8000 births annually.The induction rate in 2020 was 35%, the assisted vaginal birth rate was 14% and the CS rate was 33%.
Maternal and neonatal care is provided by an interdisciplinary team of midwives, obstetricians, anesthesiologists, and neonatologists.Labor ward protocols for induction of labor, fetal monitoring, CTG interpretation, and operative delivery are in accordance with Royal College of Obstetricians and Gynecologists (RCOG) approved guidelines [16].
The study was limited to patients with a singleton pregnancy, cephalic presentation, gestational age greater than 36 weeks, and an apparently normal pregnancy.Participants were eligible if they presented in spontaneous labor or for labor induction with nonfetal indications (e.g.post-dates, pelvic girdle pain, or maternal request).Induced patients were included if they had a favorable cervix and were awaiting transfer to the labor ward for artificial rupture of membranes and/or oxytocin.Patients with a pre-diagnosed obstetric complication including abnormal fetal growth, oligohydramnios, polyhydramnios, congenital fetal abnormality, or pre-eclampsia were excluded, as were patients with multiple pregnancy, limited English or under 18 years of age.Patients were also excluded if they had confirmed spontaneous rupture of membranes as an assessment of amniotic fluid volume would not be informative.Women who had an abnormal CTG on presentation were transferred directly to the labor ward and were excluded from recruitment to the study.

Recruitment
Recruitment of participants to the study followed a three-stage process: 1.All potentially eligible patients received information about the study prior to labor.Posters, parent education classes and written leaflets were used to explain the project purpose and design.The leaflet contained contact details to allow patients to discuss the study further if they wished.2. Once a patient presented in early labor or for induction of labor, a research fellow sought written informed consent if the following criteria were satisfied: i. the midwife looking after the patient assessed them to be capable of providing informed consent.ii. the patient had adequate pain control.iii. the patient had not used systemic opiates in the last four hours.3. Once consent had been given the study participants received an admission ultrasound scan and were not consulted again unless they requested further information.

Ultrasound assessment
An ultrasound scan was performed to assess fetal biometry, amniotic fluid volume, and placental maturity.The aim was to achieve high-quality accurate measurements as quickly as possible given that study participants were in early labor or awaiting artificial rupture of membranes and oxytocin.The scan images were classified as optimal (if all relevant anatomical landmarks could be identified) or sub-optimal, and the time taken to complete the scan was recorded in minutes [17].The biometry included the biparietal diameter, head circumference, abdominal circumference (AC), and femur length.The estimated fetal weight (EFW) was calculated from the fetal biometry using the Hadlock formula for validation with the actual birth weight.For the abdominal circumference, three measurements were taken and averaged.The measurements were plotted on the UK system Fetal Medicine centile charts and an AC below the 10th centile was classified as small for gestational age (SGA) and above the 90th centile as large for gestational age (LGA).Amniotic fluid was assessed in two ways: single deepest vertical pool (SDP) and amniotic fluid index (AFI).An AFI <5th centile for gestational age or an SDP 2cm was classified as oligohydramnios and an SDP >8cm was classified as polyhydramnios.The scan was classified as 'normal' if there was no evidence of SGA or LGA (10-90th centile), normal amniotic fluid (no oligohydramnios or polyhydramnios), and if optimal images had been achieved.Placental maturity was assessed using fetal ultrasound defined granum grades which formed an exploratory aspect of the study.A Grannum grade 2 placenta was classified as mature and Grannum grade 3 as hyper mature [18].
All scans were performed by a single research fellow who had a one-month period of training and validation by a Fetal Medicine specialist.As this was an observational study and there is no evidence currently that admission ultrasonography is either beneficial or harmful, the scan findings were documented by the researcher on a study datasheet and were only interpreted after the study had been completed.The protocol required the researcher to highlight unexpected findings of clinical importance to the labor ward staff, for example, undiagnosed breech presentation, although this was not required during the study.

Data collection
Information on the following patient characteristics was extracted from the electronic records: maternal age, marital status, socioeconomic group, nationality, body mass index (BMI), primary carer, smoking, alcohol use, and history of illicit drug use.Data were recorded for past obstetric history, gynecological history, complications during pregnancy and any history of medical disorders.Gestational age was estimated from the calculation based on first day of the last menstrual period, however, the routine booking ultrasound scan estimate was preferred if the dates were uncertain or there was a discrepancy of more than seven days.Labor factors included admission CTG, use of either prostaglandin and/or oxytocin, presence of meconium, maternal pyrexia, fetal blood sampling, use of regional analgesia, duration of first and second stages of labor, and duration of active pushing.

Outcome measures
Abnormal intrapartum events indicative of fetal compromise included pathological CTG, low pH on fetal blood sampling, and delivery by emergency CS.The CTG was classified separately for the first and second stages of labor as normal, suspicious, and pathological according to RCOG criteria [16].The mode of delivery was classified as spontaneous vaginal delivery (SVD), assisted vaginal birth (AVB), or cesarean section (CS).Cesarean sections were for suspected fetal compromise or labor dystocia and in most cases both indications were listed.Neonatal outcomes included birth weight, low Apgar scores at 1 and 5 min (classified as Apgar score of 3 at 1 min or <7 at 5 min), fetal acidosis on paired cord blood results (classified as arterial pH <7.10 or base excess <À12.0),neonatal unit (NNU) admission, prolonged NNU admission (>24 h) and whether resuscitation was required.Detailed data were extracted for those admitted to the neonatal unit including details on ventilation, encephalopathy, therapeutic hypothermia, neonatal abnormalities, and perinatal death.Perinatal death includes stillbirths and early neonatal deaths (defined as the death of a baby within the first 7 days of life).

Statistical analysis
Data analysis and reporting were completed according to STROBE guidelines for cohort studies.Descriptive statistics were used to describe the ultrasound image quality and the time taken to complete the scan.Categorical and continuous variables were used to describe the patient characteristics, labor and delivery details, and perinatal outcomes of the cohort in relation to each of the ultrasound categories; normal, oligohydramnios, SGA, mature placenta (Grannum grade 2) and oligohydramnios/SGA.Logistic regression analyses were performed to measure the association between abnormal ultrasonography and adverse labor and perinatal outcomes.Results are reported as proportions, odds ratios (OR), and 95% confidence intervals (CI) for categorical variables, and difference between means and 95% CI for continuous variables.Data analysis was performed with the statistical package SPSS (Version 20.0).

Results
We recruited 295 patients between March 2019 and March 2020.The admission CTG in all cases was normal.Image quality was optimal in 274 of the 295 scans (93%) and 271 (92%) were completed less than 10 min.(Table 1) The amniotic fluid measurement detected oligohydramnios in 67 of 294 cases (23%) and there was one case of previously undetected polyhydramnios.The placental appearance was mature (Grannum grade 2) in 87 (30%) cases and there were no cases of hyper mature placenta (Grannum grade 3).A higher proportion than expected had a head circumference below the 10th centile (63 of 295; 21%) which may reflect molding in labor or a deeply engaged head in the late third trimester.Small for gestational age (SGA; abdominal circumference <10th centile) was detected in 11 cases (4%) with an estimated fetal weight (EFW) below 2500 g in two cases.
The risk profile of the study population was largely representative of the hospital population with half the participants nulliparous, approximately half receiving midwifery-led care, and approximately a third delivering after 41 weeks' gestation.(Supplemental Table 1) The proportion being induced (49%) was higher than the background population (35%), reflecting the consent procedures for study participation.All study participants delivered within 24 h of scan completion.Factors associated with a higher incidence of mature placenta included older maternal age (!40 years) OR 1.89 (95% CI 0.37-9.58)and maternal smoking OR 1.95 (95% CI 0.71-5.39).The duration of the first and second stages of labor was similar in each group.The incidence of pathological CTG and emergency CS was higher among participants with oligohydramnios than those with a normal scan but did not reach statistical significance; Odds Ratio 3.40 (95% Confidence Intervals 0.55 to 20.92) and OR 1.43 (95% CI 0.66 to 3.08) respectively.(Table 2) Results for SGA alone were non-significant, reflecting the small sample size.Results for oligohydramnios and SGA combined were similar to oligohydramnios alone.(Supplemental Table 2).Mature placenta was not associated with significant differences in adverse labor events.
The mean birthweight was significantly lower in those with oligohydramnios À139 g (95% CI À248 to À30) and admission scan detected SGA À357 g (95% CI À557 to À137).(Table 2) Adverse perinatal outcomes were uncommon with a higher incidence of fetal acidosis (pH < 7.10) in the oligohydramnios group.There was one case of neonatal encephalopathy and one case of neonatal trauma both in the normal scan group.The incidence of neonatal unit admission >24 h was higher in the oligohydramnios group although not statistically significant; OR 3.75 (95% CI 0.61 to 22.97).SGA alone and mature placenta were not associated with statistically significant differences in adverse perinatal outcomes.(Supplemental Table 2) The results for SGA combined with oligohydramnios were similar to oligohydramnios only.A sample size calculation indicates that a cohort of 3900 participants would be required to demonstrate a statistically significant doubling of the incidence of fetal acidosis in association with oligohydramnios detected on admission ultrasonography, and the sample size for SGA would be almost 12,000 participants.

Principal findings
This prospective cohort study has demonstrated the feasibility of performing labor admission ultrasonography in a timely manner, achieving optimal images with a portable scan machine and basic operator training.We detected oligohydramnios in over 20% of our low-risk population and a small number of previously undetected SGA babies.The evidence for an association between an abnormal scan (oligohydramnios alone or with SGA) and pathological CTG, emergency CS and prolonged NNU admission is weak and does not justify a randomized controlled trial.Further evaluation in large cohorts would be required in the first instance with a sample size of several thousand participants.

Results in context
A number of approaches have been suggested using intrapartum ultrasonography as a means to identify fetal compromise.These include oligohydramnios, fetal biometry, and the biophysical profile score (BPS).
Although BPS has showed promise in two studies, the limiting factor is the length of time taken to complete the procedure and it is not therefore feasible for lowrisk women accessing routine care [19][20].In contrast, our assessment of amniotic fluid, biometry, and placental maturity could be achieved within ten minutes, which is faster than an admission CTG.An alternative approach would be to use Doppler ultrasonography which can be achieved quickly but requires a different skillset [21].
There are conflicting data on the detection of oligohydramnios in early labor as a predictor of adverse outcomes.A meta-analysis of ten studies in the antenatal or intrapartum period evaluated whether oligohydramnios (defined as AFI <5cm) was associated with increased risks of CS for "fetal distress," low Apgar score (<7 at 5 min) and low umbilical artery pH (<7.00) [22].The incidence of oligohydramnios at 22% was similar to our study, as was the association with emergency CS, but the data on fetal acidosis were inconclusive.A subsequent randomized controlled trial (RCT) allocated one thousand women in labor to either measurement of AFI or a single deepest pocket (SDP) and neither technique in isolation predicted pregnancies at higher risk of CS for "fetal distress" [23].A Cochrane systematic review of antenatal assessment of amniotic fluid reported on four trials including 3125 women and concluded that the SDP measurement seems a better choice since the use of the AFI increases the rate of diagnosis of oligohydramnios without improving perinatal outcomes [24].We assessed both AFI and SDP in our study and reported oligohydramnios in relation to SDP.
Data on fetal biometry performed in labor are limited and the few studies that have been completed are conflicting.A prospective study of 192 women conducted during the active phase of labor found that the accuracy of clinical and ultrasonographic estimation of EFW using the Hadlock formula (biparietal diameter (BPD), abdominal circumference (AC) and femur length (FL)) correlated well with actual birth weight, except in cases of fetal macrosomia where the sensitivity was poor [25].A further study found that EFW calculated within forty-eight hours of delivery was accurate with an error rate of 9% [26].In contrast, a prospective study of 1034 women in labor reported limited accuracy of EFW when done at term especially in cases of fetal macrosomia [27].These studies focus on birthweight correlation and fetal macrosomia rather than perinatal outcome.Our findings suggest that SGA babies can be identified with admission ultrasonography, albeit in small numbers, and this information could inform safe intrapartum care.Measurements based on AC are probably more reliable than EFW given the changes to fetal head dimensions in response to labor.

Clinical Implications
Admission ultrasonography to assess amniotic fluid and basic biometry is worth consideration given that one in five low-risk women had oligohydramnios and one in twenty were SGA.None of these cases would have been identified by admission CTG.We previously demonstrated that abdominal ultrasound is acceptable to women in labor and to clinicians looking after them, furthermore, obstetric trainees can acquire the skills to perform an accurate ultrasound assessment in labor within a short timeframe [15].A standardized approach would appear to be feasible for a health professional with basic training in obstetric ultrasound.As with antenatal studies, we found little evidence that assessment of placental maturity at the onset of labor contributed to identification of fetal compromise.This probably reflects the hospital policy of routine dating scans and induction of labor before 42 weeks' gestation.Assessment of placental maturity, particularly Grannum grade 3, is likely to be of more benefit in settings where induction before 42 weeks is either declined or not routinely offered [18].

Research Implications
One important lesson learned from obstetric research is that the two processes of identification and intervention need to be evaluated separately [13].The first stage is to be able to identify the subset of fetuses particularly vulnerable to intrapartum hypoxia that may be detectable by admission ultrasonography [28].It may be the case that a combined approach will be more successful including admission ultrasonography together with Doppler studies and CTG.The second stage would be to establish a management protocol for these patients with robust interventions designed to reduce risk without introducing unnecessary intervention.Pregnancies complicated by suspected placental insufficiency are managed as high-risk with heightened surveillance and a low threshold for operative delivery [11,16,29].In the case of apparently uncomplicated pregnancies, there would be greater confidence in managing patients expectantly where adequate fetal reserve had been documented.The observation in the Netherlands that the delivery related perinatal death rate is higher in "low-risk" pregnancies than in "high-risk" pregnancies emphasizes the importance of identifying "at-risk" fetuses among "low-risk" patients and managing them appropriately [29].

Strengths and limitations
This study has a number of strengths.We included a representative sample of low-risk patients who were assessed consistently with every scan completed by the same operator.The patients were managed according to routine hospital protocols without knowledge or influence of the scan findings.The outcome data were recorded prospectively and were accurate and complete.While the study provides useful data on admission ultrasonography and hypothesis-generating associations between abnormal scan findings and adverse intrapartum and neonatal outcomes, it was not powered to provide definitive conclusions for implementation into clinical practice.It may also be the case that the accuracy and speed of assessment would be less good if we had excluded women undergoing induction.The frequent use of third-trimester scanning in our clinical setting with access to portable scan machines in clinics may have limited our ability to identify new findings at labor admission.Admission ultrasonography may perform differently in settings with less access to antenatal scanning, particularly in low-resource settings.

Conclusions
This feasibility study suggests that while admission ultrasonography can be incorporated into an admission assessment in a timely manner, the evidence of benefit is weak, and does not currently justify a randomized controlled trial.If admission ultrasonography is to be tested for clinical efficacy, it will require further evaluation within very large cohorts and in combination with other modalities such as Doppler.

Table 2 .
Labor and perinatal outcomes in relation to abnormal findings on the admission scan.