Pelvic Floor Ultrasound Before Induction of Labour and Labour Outcome

Overview

Pregnancy and vaginal birth are the main causes of pelvic floor disorders such as urinary incontinence and pelvic organ prolapse. The levator hiatus - the opening in the pelvic floor muscles through which the baby passes - can be measured on three-dimensional transperineal ultrasound at rest, during pelvic floor contraction and during a Valsalva manoeuvre. Its size and its ability to stretch have been related to how long labour lasts and to how the baby is eventually born, but the evidence comes mostly from women in spontaneous labour, in whom the exact time labour began cannot be established. Women having their labour induced offer a better model, because the time each stage of labour begins is documented. Recently, automated software has made the measurement of the levator hiatus fast and reproducible, which makes it realistic to use in everyday practice. This prospective observational study will enrol 140 women with a singleton pregnancy at 36 weeks or more who are scheduled for induction of labour. Before induction begins, each woman will have a transperineal ultrasound scan of the pelvic floor at rest, on maximum contraction and on maximum Valsalva. The scan does not involve any vaginal probe, radiation or injection, and takes about 15 minutes. The times of each phase of labour, the way the baby is born and the condition of the perineum after birth will be recorded prospectively from the clinical record. The scan will be repeated before hospital discharge and about one month after birth, together with validated pelvic floor symptom questionnaires. The main question is whether the area of the levator hiatus measured on Valsalva before induction is related to the duration of the active, pushing phase of the second stage of labour. The study will also examine the duration of the first stage, the mode of delivery, and how the pelvic floor changes in the first month after an induced birth. Study procedures do not influence clinical care in any way: induction and labour are managed according to standard departmental protocols, and the research ultrasound results are not disclosed to the clinical team.

1. Background and rationale Pregnancy and vaginal delivery are the main determinants of pelvic floor disorders (PFDs), which affect up to 25-30 % of women during their lifetime and generate substantial healthcare expenditure. Levator ani muscle (LAM) injury - in the form of overdistension (hiatal ballooning) or frank avulsion - occurs in approximately 15-30 % of vaginal deliveries, and in 5-20 % of cases does not recover within the first year after birth, leading to urinary incontinence, pelvic organ prolapse and anal incontinence later in life. Three- and four-dimensional transperineal ultrasound (TPUS) has become the reference imaging modality for the assessment of LAM integrity and function. Levator hiatus area (LHA) measured at rest, on maximum pelvic floor contraction and on maximum Valsalva manoeuvre reflects both the dimensions and the distensibility of the hiatus. An inverse relationship has been described between hiatal dimensions and labour outcome: women with a larger and more distensible hiatus, and in particular those with the greatest dynamic change of the anteroposterior (AP) diameter from rest to Valsalva, have a shorter active second stage, while levator ani co-activation during Valsalva is associated with a longer active second stage and a higher rate of operative delivery. Two practical obstacles have limited the clinical translation of these observations. First, manual offline analysis of 3D volumes is time-consuming and requires a substantial learning curve, with limited interobserver reproducibility. This limitation has been largely overcome by automated segmentation software: Smart Pelvic™ (Mindray Medical, Shenzhen, China) has recently been validated against manual tracing in 66 pregnant women studied at rest, on maximum contraction and on Valsalva by Resta et al. (J Perinat Med 2024;52:165-70), with excellent intraobserver, interobserver and intermethod agreement (ICC \> 0.9) and a measurement time reduced from 162.5 ± 20.0 s to 9.3 ± 3.9 s per volume; that validation study provides the methodological basis for the acquisition and measurement protocol adopted here. Second, most published series have studied heterogeneous populations of women in spontaneous labour, in whom the onset of the first stage cannot be dated precisely. Induction of labour (IOL) offers a uniquely controlled clinical model: the moment of onset of the cervical ripening process and of the active phase can be documented prospectively, allowing an unbiased measurement of the duration of the first stage, and of the passive and active phases of the second stage. IOL now accounts for approximately one in three deliveries in high-income settings and is associated with a longer labour and a higher risk of intrapartum caesarean section, making the identification of predictors of labour progress in this population clinically relevant. Finally, the natural history of the levator hiatus after induced labour, and its relationship with antepartum hiatal biometry, has not been described. Serial assessment at discharge and at one month postpartum allows the immediate impact of delivery on hiatal dimensions and the early recovery process to be quantified in the same women. This study will therefore combine an automated, reproducible antepartum measurement of the levator hiatus with precisely dated labour outcomes and short-term postpartum pelvic floor imaging. 2. Objectives 2.1 Primary objective To evaluate the association between antepartum levator hiatus area on maximum Valsalva manoeuvre, measured by automated software from 3D TPUS volumes at ≥ 36 weeks' gestation, and the duration of the active (pushing) second stage of labour in women undergoing induction of labour and delivering vaginally. 2.2 Secondary objectives 1. To evaluate the association between hiatal biometry (LHA and AP diameter at rest, on contraction and on Valsalva; hiatal distensibility indices; levator co-activation) and: \- duration of the active phase of the first stage of labour; * duration of the passive (non-pushing) second stage; * total duration of labour from the start of the active phase to delivery; * mode of delivery (spontaneous vaginal, operative vaginal, intrapartum caesarean section) and indication for caesarean section; * failed induction and need for oxytocin augmentation. 2. To assess whether hiatal biometry provides information on labour duration and mode of delivery independent of established maternal and fetal predictors (maternal age, height, body mass index, ethnicity, parity, gestational age, cervical status at induction, method of induction, epidural analgesia, fetal head circumference and estimated fetal weight, birthweight). 3. To describe the change in hiatal biometry from the antepartum assessment to hospital discharge and to one month postpartum, overall and according to mode of delivery. 4. To estimate the incidence of hiatal ballooning and of levator ani avulsion at discharge and at one month postpartum, and to identify antepartum and intrapartum predictors of these findings. 5. To describe the relationship between postpartum hiatal findings and pelvic floor symptoms assessed by validated questionnaires at one month. 6. To confirm, in this cohort, the reproducibility of automated LHA measurement and its agreement with manual tracing (intraobserver, interobserver and intermethod). 2.3 Exploratory objective To develop and internally validate a multivariable model combining hiatal biometry with maternal and fetal characteristics for the prediction of a prolonged active second stage and of intrapartum caesarean section. 3\. Study design and setting Single-centre (or multicentre, with identical protocol at each site) prospective observational cohort study conducted at \[Institution\], a tertiary referral maternity unit with approximately \[n\] deliveries per year. Planned recruitment period: \[start date\] to \[end date\] (approximately 18 months), followed by 1 month of follow-up for the last participant recruited. The study is purely observational: ultrasound assessment is performed for research purposes only, results are not disclosed to the clinicians managing labour, and the management of induction and labour follows standard departmental protocols throughout. 4\. Study population 4.1 Inclusion criteria * Age ≥ 18 years; * Singleton pregnancy with cephalic presentation; * Gestational age ≥ 36+0 weeks, dated by first-trimester crown-rump length; * Scheduled for induction of labour with a planned attempt at vaginal delivery, for any clinical indication; * Assessment feasible before the start of cervical ripening or, at the latest, before the administration of the first pharmacological or mechanical ripening agent; * Written informed consent. 4.2 Exclusion criteria * Multiple pregnancy or non-cephalic presentation; * Known fetal structural or chromosomal anomaly; * Previous caesarean section or previous uterine surgery; * Planned caesarean section, or contraindication to vaginal delivery arising before the ultrasound assessment; * Connective tissue disease, neuromuscular disease or neuropathy affecting the pelvic floor; * Previous pelvic floor or anti-incontinence surgery; * Active labour or ruptured membranes with regular contractions at the time of the scheduled assessment; * Inability to perform an adequate Valsalva manoeuvre or pelvic floor contraction after standardised instruction; * Inability to provide informed consent or to understand the study questionnaires. 4.3 Recruitment Consecutive eligible women admitted to the induction of labour service will be approached by a member of the research team on the day of admission. Written information will be provided and consent obtained before the antepartum scan. A screening log of all eligible women, including reasons for non-participation, will be kept to allow a STROBE flow diagram. 5\. Study procedures and timing of assessments Assessment Timing Content T0 - Antepartum ≥ 36+0 weeks, on admission for IOL, before the first ripening agent Maternal history and anthropometry; fetal biometry; cervical assessment; 3D/4D TPUS at rest, contraction and Valsalva; baseline pelvic floor symptom questionnaires Labour Continuous Prospective recording of labour milestones, interventions and delivery data from the partogram and electronic records T1 - Discharge 24-96 h postpartum, before hospital discharge 3D/4D TPUS at rest, contraction and Valsalva; tomographic ultrasound imaging (TUI) for levator avulsion; perineal trauma review T2 - One month 28-42 days postpartum 3D/4D TPUS at rest, contraction and Valsalva; TUI; pelvic floor symptom questionnaires; breastfeeding and pelvic floor rehabilitation status Women delivering by intrapartum caesarean section remain in the cohort and undergo the full postpartum assessment schedule; they contribute to the analyses of hiatal change and of mode of delivery, and are censored in the time-to-event analyses of labour duration. 6\. Ultrasound assessment 6.1 Equipment and acquisition technique All examinations will be performed with a \[Mindray Nuewa i9 / equivalent\] ultrasound system equipped with a transabdominal volumetric probe (5-8 MHz) and with the Smart Pelvic™ automated pelvic floor module, following the AIUM/IUGA practice parameter for urogynaecological ultrasound. Acquisition protocol, identical at T0, T1 and T2: 1. The woman is placed in the lithotomy position with hips flexed and abducted, after voiding (empty bladder) and, whenever possible, after bowel emptying. 2. The volumetric probe, covered with a sterile glove and coupling gel, is applied translabially on the midsagittal plane, with minimal pressure to avoid distortion of the hiatus. 3. Acquisition settings: high-frequency, high-quality volume mode, acquisition angle up to 85°, with the symphysis pubis and the anorectal angle both included in the volume. 4. Standardised verbal instructions are given and the manoeuvres rehearsed before acquisition: \- Rest: static 3D volume acquired during a period of quiet breathing without contraction. * Maximum pelvic floor contraction: 4D volume acquired over a contraction lasting \> 6 s. * Maximum Valsalva manoeuvre: 4D volume acquired over a Valsalva lasting \> 6 s, without concomitant levator co-activation being prompted; the manoeuvre is repeated up to three times and the volume with the greatest hiatal distension is retained. 5. Three volumes (one per manoeuvre) are stored for each assessment; all volumes are archived in anonymised form with a study identifier. Volumes are considered adequate if the symphysis pubis, the puborectalis sling and the anorectal angle are simultaneously identifiable and there is no motion artefact. Inadequate volumes are re-acquired during the same session. 6.2 Offline measurement Analysis is performed offline, on anonymised volumes presented in random order, by operators blinded to labour outcome, to postpartum findings and to each other's measurements. Primary (automated) measurement. Smart Pelvic™ automatically aligns the volume so that the posterior-inferior margin of the symphysis pubis and the anorectal angle lie on the same horizontal line, renders the plane of minimal hiatal dimensions (C-plane), segments the hiatal contour and computes the hiatal area. The automated software output is the primary measurement for all analyses, on the basis of the agreement with manual tracing previously demonstrated in pregnant women by Resta et al. (reference 11). Manual measurement. In all women at T0, and in a random subset of at least 25 % of postpartum volumes, the plane of minimal hiatal dimensions is identified manually from the C-plane view and the hiatal contour traced along the most medial aspect of the puborectalis muscle, as described by Dietz et al., by a sonographer blinded to the automated result, at least 2 weeks after the automated analysis. Variables recorded at each manoeuvre and each timepoint: • Levator hiatus area, LHA (cm²); • Anteroposterior hiatal diameter (cm) and left-right transverse diameter (cm); • Levator-urethra gap and TUI assessment of the puborectalis insertion (postpartum assessments; avulsion defined as an abnormal insertion in all three central slices of the TUI set on maximum contraction, uni- or bilateral); • Time (s) required to complete the measurement with each method. Derived variables: • Hiatal distensibility: LHAValsalva - LHArest (cm²) and (LHAValsalva - LHArest)/LHArest × 100 (%); • AP diameter dynamic change from rest to Valsalva (cm and %); • Contractility: LHArest - LHAcontraction (cm²) and percentage change; • Levator co-activation: LHAValsalva \< LHArest (or AP diameter on Valsalva smaller than at rest), recorded as a binary variable; • Hiatal ballooning on Valsalva, categorised according to Dietz: normal \< 25 cm², mild 25-29.9 cm², moderate 30-34.9 cm², marked 35-39.9 cm², severe ≥ 40 cm². 6.3 Reproducibility sub-study In the first 30 consecutive participants: * Intraobserver agreement: two separate volume acquisitions by the same operator at the beginning and end of the same session, each analysed automatically. * Interobserver agreement: an independent acquisition and analysis by a second sonographer, blinded to the first operator's results. * Intermethod agreement: automated versus manual analysis of the same volumes. Agreement will be expressed as intraclass correlation coefficients (ICC, two-way mixed, absolute agreement, single measures) with 95 % confidence intervals, and displayed as Bland-Altman plots with mean difference and 95 % limits of agreement. 7\. Labour management and outcome definitions Induction and labour are managed according to the local departmental protocol; the research team does not influence clinical decisions. Recorded induction data: indication for IOL; Bishop score and cervical length before induction; method of cervical ripening (dinoprostone vaginal insert or gel, misoprostol, single- or double-balloon catheter, amniotomy alone); number and timing of doses; time of amniotomy; use, timing and maximum dose of oxytocin; type and timing of analgesia (epidural, remifentanil, none). Labour outcome definitions (to be applied uniformly): * Start of the active phase of the first stage: first documented cervical dilatation ≥ 5 cm with regular painful contractions (≥ 3 per 10 min). The time of the first assessment at which this definition is met is used; the interval since the previous vaginal examination is recorded so that interval-censoring can be handled in sensitivity analyses. * Duration of the active first stage: interval (min) from the start of the active phase to confirmed full cervical dilatation. * Total induction-to-delivery interval: interval (min) from the administration of the first ripening agent to delivery. * Full dilatation: first vaginal examination documenting 10 cm dilatation. Examinations are performed at least every 2 h during the active phase and hourly at full dilatation, per departmental protocol. * Passive second stage: interval (min) from full dilatation to the onset of active maternal pushing. * Active (pushing) second stage: interval (min) from the onset of active pushing to delivery of the fetus - primary outcome. * Total second stage: full dilatation to delivery. * Prolonged active second stage: ≥ 60 min in nulliparous women and ≥ 30 min in parous women (sensitivity analysis using the 90th centile of the cohort). * Mode of delivery: spontaneous vaginal, operative vaginal (vacuum or forceps, with indication), intrapartum caesarean section (with indication: failure to progress in the first stage, failure to progress in the second stage, suspected fetal compromise, other). * Failed induction: caesarean section performed before reaching the active phase of the first stage. Other intrapartum and neonatal variables: duration of ruptured membranes; amniotic fluid characteristics; maternal position and pushing technique in the second stage; episiotomy (type); perineal trauma (intact, first-, second-, third- \[3a/3b/3c\] or fourth-degree tear); estimated blood loss and postpartum haemorrhage ≥ 500/1000 mL; birthweight; head circumference at birth; occiput position at delivery; Apgar scores; umbilical artery pH; neonatal unit admission. 8\. Maternal and fetal characteristics Collected at T0 unless otherwise stated: maternal age; height; pre-pregnancy weight and BMI; weight and BMI at the time of the scan; gestational weight gain; self-reported ethnicity (categorised per \[local standard classification\]); parity and obstetric history including previous vaginal deliveries, previous operative vaginal delivery and previous perineal trauma; smoking status; chronic and gestational medical conditions (chronic hypertension, hypertensive disorders of pregnancy, pre-gestational and gestational diabetes, connective tissue disorders); pelvic floor muscle training during pregnancy (yes/no, supervised/unsupervised, weeks); perineal massage; constipation; occupation involving heavy lifting; family history of pelvic organ prolapse; gestational age at scan and at delivery; ultrasound fetal biometry within 7 days of the scan (biparietal diameter, head circumference, abdominal circumference, femur length, estimated fetal weight and centile); occiput position and head-perineum distance at the time of the antepartum scan. 9\. Postpartum outcomes and questionnaires At T1 and T2, in addition to the imaging variables described in section 6.2: * Change in LHA from T0 (ΔLHA) at each manoeuvre; * Hiatal ballooning category on Valsalva; * Levator ani avulsion (unilateral/bilateral) on TUI; * Perineal healing, wound complications, need for readmission; * Breastfeeding status, resumption of menses, and referral to pelvic floor rehabilitation. Validated questionnaires (in the locally validated language version) administered at T0 and T2: * Pelvic Floor Distress Inventory (PFDI-20) with its POPDI-6, CRADI-8 and UDI-6 subscales; * Pelvic Floor Impact Questionnaire (PFIQ-7); * International Consultation on Incontinence Questionnaire - Urinary Incontinence Short Form (ICIQ-UI SF); * Wexner score for anal incontinence. 10. Sample size The primary analysis is the association between antepartum LHA on Valsalva and the duration of the active second stage among women delivering vaginally, analysed as a correlation and by multivariable linear regression on log-transformed duration. Assuming a two-sided α of 0.05 and 80 % power, 85 women delivering vaginally are required to detect a correlation coefficient of ρ = 0.30, the smallest association considered clinically relevant on the basis of previously reported associations between hiatal dimensions and second-stage duration. Allowing for: * an intrapartum caesarean section rate of approximately 20-25 % in an induced population, * approximately 10 % of women not reaching the active pushing phase or having non-evaluable labour timings, * and approximately 10 % of non-evaluable or missing antepartum volumes, a total of 140 women will be recruited. This sample also provides approximately 10 events per candidate predictor for a multivariable model containing up to 8 covariates in the analysis of second-stage duration. For the longitudinal postpartum analyses, an anticipated attrition of 25 % at one month leaves approximately 105 women with complete T0-T1-T2 imaging, which provides \> 90 % power to detect a within-woman change of 0.4 SD in LHA on Valsalva between timepoints using a mixed-effects model. Sample size for the reproducibility sub-study (n = 30, three volumes each) allows an expected ICC of 0.90 to be estimated with a lower 95 % confidence limit above 0.80. 11\. Statistical analysis A detailed statistical analysis plan will be finalised and dated before database lock. Descriptive analysis. Continuous variables will be summarised as mean ± SD or median (interquartile range) according to distribution (Shapiro-Wilk test); categorical variables as counts and percentages. Participant flow will be reported in a STROBE diagram. Primary analysis. The association between LHA on Valsalva and the duration of the active second stage will be assessed by Pearson or Spearman correlation as appropriate, and by multivariable linear regression on the natural logarithm of duration, adjusted for parity, maternal age, BMI, ethnicity, epidural analgesia, oxytocin use, gestational age at delivery and neonatal head circumference. Regression coefficients will be reported as the percentage change in duration per 1 cm² increase in LHA, with 95 % confidence intervals. Time-to-event analysis. Because women delivering by caesarean section do not complete the second stage, the duration of the first and second stages will additionally be analysed by Kaplan-Meier and Cox proportional hazards models, with caesarean section treated as a censoring event and, in a competing-risks sensitivity analysis, as a competing event (Fine-Gray subdistribution hazards). Hiatal variables will be entered as continuous predictors and, secondarily, by tertiles of the cohort distribution. Mode of delivery. Multivariable logistic regression for intrapartum caesarean section and for operative vaginal delivery, with the same covariate set. Discrimination will be expressed by the area under the ROC curve with 95 % CI, and calibration by calibration plots and the Hosmer-Lemeshow test; optimism-corrected estimates will be obtained by bootstrap internal validation (1000 replications). Longitudinal analysis. Changes in LHA and AP diameter across T0, T1 and T2 will be analysed with linear mixed-effects models including a random intercept per woman, with timepoint, manoeuvre, mode of delivery and their interaction as fixed effects. Predictors of hiatal ballooning and of levator avulsion at T2 will be examined by logistic regression. Reproducibility. ICCs with 95 % CI and Bland-Altman plots as specified in section 6.3; measurement times compared with the paired t test or Wilcoxon signed-rank test. Missing data. The pattern and extent of missingness will be described. Under a missing-at-random assumption, multiple imputation by chained equations (20 imputations) will be used for covariates in the multivariable models; complete-case analysis will be reported as a sensitivity analysis. Outcome data will not be imputed for the primary analysis. Subgroup analyses (pre-specified, interpreted as exploratory): nulliparous versus parous women; with versus without epidural analgesia; ethnicity groups with sufficient numbers; method of induction. A two-sided p value \< 0.05 will be considered statistically significant. No formal adjustment for multiplicity will be applied to secondary outcomes, which will be interpreted as hypothesis-generating. Analyses will be performed with \[SPSS v.30, MedCalc, R v.4.\] software.