PROPENSITY-ID: Sonographic and Clinical Assessment of Head Position Prior to Instrumental Delivery and Labour Outcomes

Introduction Instrumental delivery is commonly performed during the second stage of labor with the aim to expedite delivery in conditions of second stage dystocia or non-reassuring fetal condition and is estimated to account for approximately 3-15% of all births. The appropriate use of vacuum extractor and forceps is crucial to avoid unnecessary and potentially challenging second-stage caesarean sections, however a 4-6% failure rate has been reported following attempted instrumental delivery. When this occurs, caesarean section is required to deliver the fetus. However, caesarean section following failed instrumental delivery is acknowledged to be associated with an increased incidence of maternal and fetal complications including neonatal trauma, skull fracture, intracranial haemorrhage, neonatal acidaemia and transfer to NICU as well as postpartum haemorrhage, surgical injury and third- or fourth-degree perineal tears compared to a successful vacuum delivery.

The decision to expedite delivery by means of caesarean section during the second stage of labor has also been associated with maternal complications including uterine incision extension, incidental cystotomy and with an increased frequency of fetal injury. Therefore, the decision as to which intervention to perform is crucial to optimize the maternal and neonatal outcomes when an obstetric intervention is indicated during the second stage of labor.

Fetal head position and station are among the major determinants of successful instrumental delivery. Instrumental delivery is recommended following the ascertainment of the fetal head position when the station is below the level of the maternal ischial spines (level 0). A secondary analysis of a randomized controlled trial conducted on a group of women submitted to instrumental vaginal delivery has shown that the risk of failed vacuum extraction is increased in the event of inaccurate placement of the instrument, which is favoured by a high fetal station and a non-occiput anterior position - i.e. a malposition - of the fetal head. Fetal head position and station have been traditionally assessed by digital examination. However, such evaluation has been shown to yield limited accuracy and poor reproducibility. Akmal et al. compared the accuracy of vaginal examination versus transabdominal sonography in the ascertainment of the head position prior to instrumental delivery showing an overall 27% rate of misdiagnosis at vaginal examination, which was mainly accounted by cases in occiput posterior position and station at the level of the ischial spines. Wong et al conduced a randomized controlled trial (RCT) on 50 patients undergoing instrumental delivery by means of vacuum extractor due to second stage dystocia by measuring distance between the centre of the cignon and the flexion point between the cases submitted (n=25) and those not submitted (n=25) to intrapartum US in addition to the vaginal examination. The mean distance was lower in the patients submitted to intrapartum sonography in addition to clinical examination compared to those who had only vaginal examination (2.1+1.3 vs 2.8+1.0 cm, p<0.05).

Conversely, intrapartum sonography has been demonstrated to be a reliable tool for the assessment of the fetal head station and position, particularly in the second stage of labor. The transabdominal approach is considered the gold standard method for the diagnosis of the occiput position, while the transperineal approach is commonly used for the objective measurement of the fetal station. The US parameters have also been shown to be more accurate than clinical examination in predicting the outcome of instrumental vaginal delivery. On this basis, the International Guidelines endorse the use of intrapartum sonography in conditions of abnormal progression of the first or of the second stage of labor and in the event of uncertainty of the occiput position following vaginal examination prior to perform an instrumental delivery.

Hypothesis and Justification for the study

While grade "A" evidence supports the role of intrapartum sonography in the ascertainment of the occiput position, to date three randomized studies as well as two systematic reviews have failed to demonstrate a role of intrapartum sonography in improving labor outcomes. This can be explained by the fact that, given the low frequency of such adverse outcomes, a large number of randomized cases is warranted to demonstrate a potential benefit of ultrasound in adjunct to clinical examination when compared to clinical examination only. Two of these studies were indeed underpowered to demonstrate clinical benefits in terms of failed instrumental delivery and perinatal outcomes. The R.I.S.POS.T.A. trial was adequately powered to investigate whether the use of intrapartum US is associated with a reduction of the frequency of failed vacuum delivery compared to standard clinical management of labor, however the study was preliminary concluded upon advice of a data safety monitoring committee based on exceedingly slow recruitment rate and an unexpectedly low frequency of failed vacuum delivery. On this basis, the Authors hypothesized that most practitioners commonly perform sonography before attempted vacuum extraction even in the absence of any evidence supporting its clinical benefit, and that randomization was considered only when the fetal extraction was considered easy.

Propensity score matching analysis allows to design and analyse an observational - i.e. non-randomized - dataset to make it mimic some of the particular characteristics of a RCT. Based on the assumption that a properly conducted RCT is highly unlikely to be feasible, the investigators hypothesize that propensity score matching analysis may allow to perform an adequately powered evaluation of the potential clinical benefits - in terms of successful instrumental delivery and maternal and fetal outcomes - of intrapartum sonography in adjunct to clinical examination compared to conventional labor management.

Study objectives

To evaluate whether the assessment of the fetal occiput position by intrapartum sonography before instrumental delivery improves labor outcome.

The two primary outcomes will be vaginal delivery and composite adverse perinatal outcome. Composite adverse perinatal outcome will be defined based on the occurrence of either shoulder dystocia, acidaemia (as defined by arterial cord pH <7.0 and/or a base excess >12.0), Apgar score <7 at 5 minutes, neonatal injury (e.g. intracranial haemorrhage, skull fracture), NICU admission, ischemic-hypoxic encephalopathy and death.

Secondary outcomes will include maternal complications including third and four-degree tear, postpartum haemorrhage and surgical injury, as well as sequential instrumental delivery by forceps.

Study design

International multicentre prospective data collection which will involve tertiary maternity units with reported instrumental delivery rates around 5% or above. A minimum number of 130 patients is expected to be collected by each participating Unit.

The study will consist in a multicentre prospective data collection of two parallel groups:

• Group 1: patients submitted to instrumental delivery preceded by the use of ultrasound for the assessment of the occiput position by clinicians who routinely perform intrapartum sonography as an adjunct to clinical examination prior to vacuum delivery
• Group 2: patients having submitted to instrumental delivery without the adjunct of by clinicians not performing intrapartum sonography prior to vacuum delivery

Patients will be counselled regarding the aim of the study on admission to the labor ward. The acquisition of verbal consent for the pseudoanonymized data collection in the event of instrumental delivery will be a prerequisite for study enrolment. Data will be prospectively collected from case notes of women submitted to instrumental delivery and sent to the coordinating centre (University of Parma) for analysis at the end of the study period.

Statistical methods

In order to reduce the impact of the important differences in baseline characteristics between the intrapartum sonography + clinical examination and the clinical examination only groups, the investigators will perform a propensity score (PS) matching analysis to assess the effect of US as an adjunct to clinical examination prior to instrumental delivery on the occurrence of failed instrumental delivery, adjusting for the confounding bias caused by this imbalance. Compared with classic multivariate adjustments, the PS permits finer adjustments for wider sets of covariates. PS will be defined as the conditional probability of having US given the measured covariates in order to balance covariates in the two groups.

To obtain the PS, a logistic regression model will be fitted with intrapartum sonography + clinical examination as the dependent variable, then the investigators will model the conditional probability of having a failed instrumental delivery as a function of baseline and those clinical characteristics associated with having a vacuum delivery. The investigators will use the PS to match, without replacement, each case of intrapartum sonography + clinical examination prior to vacuum delivery with the clinical examination only that had the closest PS in 1:1 ratio, in order to optimize the precision of the estimate of association and limit bias. The investigators will also accept cases only if the difference in PS between matched cases is small (preset caliper), to obtain a better balance between the intrapartum sonography + clinical examination group and clinical examination only group as matched samples. The investigators will compute standardized differences for all variables included in the PS before and after matching, to assess the effect of matching on the imbalance, and will aim a 10% standardized difference to be the limit for a correct balance. After matching, the investigators will compare the failed vacuum delivery rate between the intrapartum sonography + clinical examination group and the clinical examination only as matched group. Finally, the investigators will calculate the odds ratio to quantify the association between intrapartum sonography and failed vacuum delivery using univariate logistic regression analysis fitted by generalized estimating equations to account for matched data.

Sample size and power

Based on available data, the estimated failure rate of instrumental delivery ranges between 4-6%. The baseline risk of failed instrumental delivery in the control group (clinical examination only) will be estimated at around 4%. Based on the assumption that the ultrasound assessment of the fetal head position could halve the frequency of failed instrumental delivery up to 2%, the total sample size of "matched" patients required in order to obtain statistical power of 80% (P <0.05) would be 2282, i.e. 1141 cases per arm. The prospective multicentre design of the study is expected to allow the enrolment of an over two-fold higher number of cases compared to the required sample size.

Based on published data, the investigators estimate that this sample size is enough also to investigate the second primary outcome of the study, i.e. composite adverse perinatal outcome as above defined.