- Top of page
Fetal head station and rotation in labor are traditionally determined by digital examination. Such evaluation has, however, been demonstrated to be inaccurate and poorly reproducible[1-4]. Intrapartum transperineal ultrasonography (ITU) has been suggested as an objective and reliable method for assessing both fetal head descent and internal rotation[5-11]. Although many sonographic parameters have been studied, very little is known about the longitudinal changes of these ITU indices during the second stage of labor[12-14].
The aim of our study was twofold: first, to evaluate the longitudinal changes of two ITU parameters by three dimensional (3D) ultrasound, namely the angle of progression (AoP) and the midline angle (MLA), during the active second stage of labor; and, second, to assess whether measurements of these sonographic parameters at different time points are associated with the mode of delivery.
- Top of page
A series of ultrasound volumes were prospectively acquired from women at the beginning of the active second stage of labor. The study group included a non-consecutive series of nulliparous women with uncomplicated singleton pregnancies at term gestation (37 weeks or more), with fetuses in cephalic presentation, attending the labor ward of our University hospital from November 2010 to November 2011. Patient enrollment was carried out when one of the investigators, with at least 3 years of experience in obstetric ultrasound (T.G., G.P., A.Y., E.M., M.N. and F.D.M.), was available in the labor ward. The obstetrician performing the ultrasound examination was present in the labor ward exclusively for this aim and was blinded to the clinical examination of the women.
Patients were excluded from the study if they underwent Cesarean delivery during the first stage of labor or if Cesarean section or instrumental vaginal delivery was performed in the second stage purely because of an abnormal fetal heart trace.
Ultrasound was performed using a portable machine (Voluson i; GE Medical Systems, Zipf, Austria) equipped with a volumetric probe. Ultrasound volumes were acquired transperineally, as previously described, at the beginning of the active second stage (T1), and every 20 min thereafter (T2, T3, T4, T5 and T6) until delivery. All volumes were transferred to a PC equipped with dedicated software (4D view 9.0 and Sono-VCAD labor; GE Medical Systems) for offline analysis following delivery.
All volumes were anonymized and analyzed after delivery in the multiplanar mode by an operator blinded to the labor outcome. In accordance with the acquisition technique, the sagittal plane was displayed on Panel A, and the axial and coronal planes on Panels B and C, respectively. Each volume was processed using the static-Volume Contrast Imaging (VCI) algorithm (GE Medical Systems) in order to improve image resolution in the reconstructed planes.
Volume alignment was obtained using the symphysis pubis and the urethra as reference points, as previously described. Subsequently, using Sono-VCAD labor software, the AoP (defined as the angle between the longitudinal axis of the pubic bone and a line joining the lowest edge of the pubis to the lowest convexity of the fetal skull) was measured on the mid-sagittal plane. Then, switching from the sagittal plane to the axial plane (Plane B), the MLA was calculated (defined as the angle between the anteroposterior axis of the maternal pelvis and the head midline).
All sonographic parameters obtained at different time intervals were compared between patients who underwent spontaneous vaginal delivery (Group A) and those who underwent an operative delivery (instrumental or Cesarean delivery; Group B). In our unit instrumental vaginal delivery is exclusively performed by vacuum extraction when the clinically assessed fetal head station is equal to or lower than +2 cm from the ischial spines.
Mean, SD and frequencies were used as descriptive statistics. The modes of delivery were compared using the Kruskal–Wallis test and Fisher's exact test. Receiver–operating characteristics (ROC) curves were calculated in order to estimate the accuracy of AoP and MLA in predicting the mode of delivery. The area under the ROC curve (AUC) was computed together with the standard error (SE) and the 95% CI. The best cut-off was calculated using a maximum-likelihood ratio method.
Univariable and stepwise forward multivariable logistic regression analyses were performed by pooling the overall set of ITU scans acquired at all time points. These analyses were performed taking into account the ultrasonographic parameters (AoP and MLA), as well as other maternal and intrapartum variables (namely gestational age at delivery, epidural analgesia, maternal age and body mass index, and oxytocin administration), in order to identify their role in the prediction of the mode of delivery. Multivariable analysis was initially carried out including AoP and MLA only in order to investigate their independence from each other and was then performed taking into account all of the variables in order to identify the potential confounding factors. Data were analyzed using SPSS (version 13.0 for Windows), and two-tailed P-values of less than 0.05 were considered statistically significant.
- Top of page
Overall, 76 women were enrolled into the study. Of these, three were excluded because of operative delivery for an abnormal fetal heart trace and two were excluded because of Cesarean delivery in the first stage of labor. In the remaining 71 women, a total of 174 ITU scans were performed. Spontaneous vaginal delivery occurred in 58 (81.7%) women (Group A), whereas vacuum extraction and Cesarean section (Group B) were performed in eight (11.3%) and five (7.0%) women, respectively. The indications for Cesarean or vacuum delivery were poor progress (n = 6), maternal exhaustion (n = 2), or a combination of more than one of the previous indications with or without suspicious or abnormal fetal heart trace (n = 5). Oxytocin was administered to 46 (64.8%) women, and epidural analgesia was given to 41 (57.7%) women. The characteristics of the patients are displayed in Table 1.
Table 1. Characteristics of 71 women in active second stage of labor, according to mode of delivery
|Characteristic||Spontaneous vaginal delivery (Group A; n = 58)||Operative delivery (Group B; n = 13)||P|
|Maternal age (years)||30.7 ± 4.9||33.2 ± 4.5||0.078a|
|Gestational age (weeks)||38.9 ± 1.3||39.1 ± 1.2||0.490a|
|Body mass index (kg/m2)||27.1 ± 3.9||28.7 ± 3.6||0.176a|
|Oxytocin administration||35 (60.3)||11 (84.6)||0.098b|
|Epidural analgesia||31 (53.4)||10 (76.9)||0.121b|
|Occiput posterior position at delivery||4 (6.9)||3 (23.1)||0.005b|
|Second stage duration (min)||84.3 ± 59.5||49.6 ± 31.9||0.062a|
Women who underwent a spontaneous vaginal delivery (Group A) had a significantly higher AoP at the beginning of the active second stage (T1: 140.0 ± 20.2° vs 122.9 ± 16.7°; P = 0.010) and after 20 min (T2: 149.7 ± 20.7° vs 126.9 ± 17.5°; P = 0.006), compared with women who underwent operative delivery (Group B), whereas this difference was not maintained in the subsequent time intervals (Table 2). As shown by the ROC curves, AoP yielded a significant prediction of operative delivery at the first two time points with an AUC of 0.731 ± 0.077 (values equal to or less than 108°; sensitivity: 5/13, 38.5%; specificity: 58/58, 100%) and 0.785 ± 0.080 (cut-off range: 125–127°; sensitivity: 6/10, 60.0%; specificity: 37/42, 88.1%) at T1 and T2, respectively (Figure 1a and Table 3).
Table 2. Angle of progression (AoP) and midline angle (MLA) at different time points in 71 women in active second stage of labor according to mode of delivery
|Time point||AoP (°)||MLA (°)|
|Spontaneous vaginal delivery (Group A)||Operative delivery (Group B)||Pa||Spontaneous vaginal delivery (Group A)||Operative delivery (Group B)||Pa|
|T1 (n = 71; Group A, n = 58; Group B, n = 13)||140.0 ± 20.2||122.9 ± 16.7||0.010||36.2 ± 22.1||40.2 ± 22.3||0.364|
|T2 (n = 52; Group A, n = 42; Group B, n = 10)||149.7 ± 20.7||126.9 ± 17.5||0.006||28.9 ± 18.7||41.3 ± 23.0||0.097|
|T3 (n = 27; Group A, n = 19; Group B, n = 8)||149.3 ± 16.2||136.9 ± 18.8||0.084||21.2 ± 11.7||40.8 ± 27.9||0.043|
|T4 (n = 15; Group A, n = 10; Group B, n = 5)||148.2 ± 15.7||138.2 ± 26.2||0.219||18.2 ± 15.0||47.4 ± 29.6||0.020|
|T5 (n = 7; Group A, n = 3; Group B, n = 4)||157.8 ± 36.6||129.3 ± 11.9||0.157||18.25 ± 6.0||34.7 ± 4.2||0.034|
|T6 (n = 2; Group A, n = 1; Group B, n = 1)||147||127||—||13||39||—|
Table 3. Area under the receiver–operating characteristics curves (AUC) for angle of progression (AoP) and midline angle (MLA) at different scan intervals in the prediction of operative delivery
| ||AoP (AUC)||MLA (AUC)|
|Time point||Mean ± SE||95% CI||Mean ± SE||95% CI|
|T1 (n = 71)||0.731 ± 0.077||0.580–0.882||0.581 ± 0.086||0.413–0.749|
|T2 (n = 52)||0.785 ± 0.080||0.627–0.942||0.670 ± 0.091||0.492–0.848|
|T3 (n = 27)||0.714 ± 0.121||0.477–0.950||0.750 ± 0.103||0.548–0.952|
|T4 (n = 15)||0.700 ± 0.175||0.357–1.000||0.880 ± 0.089||0.706–1.000|
|T5 (n = 7)||0.833 ± 0.173||0.493–1.000||1.000 ± 0.000||1.000–1.000|
For MLA, Group A showed comparable measurements to those of Group B at T1 and T2, and a significantly narrower MLA at T3 (21.2 ± 11.7° vs 40.8 ± 27.9°; P = 0.043), T4 (18.2 ± 15.0° vs 47.4 ± 29.6°; P = 0.020) and T5 (18.25 ± 6.0° vs 34.7 ± 4.2°; P = 0.034) (Table 2). As shown by ROC curves, MLA yielded significant prediction of operative delivery from 40 min onwards, with an AUC of 0.750 ± 0.103 (cut-off range: 18–19°; sensitivity: 7/8, 87.5%; specificity: 10/19, 52.6%), 0.880 ± 0.089 (values equal to or greater than 23°; sensitivity: 5/5, 100%; specificity: 7/10, 70.0%) and 1.000 ± 0.000 (cut-off range: 27–29°; sensitivity: 3/3, 100%; specificity: 4/4, 100%) at T3, T4 and T5, respectively (Figure 1b and Table 3).
The results of univariable and multivariable logistic regression analyses are shown in Table 4. All variables, except for gestational age, were significantly related, in univariable analyses, to the mode of delivery. The multivariable analysis when considering only ultrasound parameters showed that both AoP and MLA were independently associated with the mode of delivery. However, after taking all variables into account in the analysis, the AoP, maternal age and the use of epidural analgesia were the only independent predictors of mode of delivery (Table 4). In particular, the odds of operative delivery were found to decrease by 5.4% for a rise of 1° of AoP, to increase by more than threefold in women with epidural analgesia and to increase by 13.8% for each additional year of maternal age.
Table 4. Summary of univariable and stepwise multivariable logistic regression analyses to identify independent intrapartum transperineal ultrasonography (ITU) parameters that have a significant association with mode of delivery (operative vs spontaneous vaginal delivery)
| || ||Multivariable analyses|
| ||Univariable analyses||ITU parameters only||ITU, maternal and intrapartum variables|
|Variable||Odds ratio (95% CI)||P||Odds ratio (95% CI)||P||Odds ratio (95% CI)||P|
|Angle of progression (°)||0.952 (0.930–0.975)||< 0.001||0.955 (0.933–0.978)||< 0.001||0.946 (0.923–0.970)||< 0.001|
|Midline angle (°)||1.023 (1.007–1.040)||0.005||1.018 (1.000–1.035)||0.047||—a|| |
|Epidural analgesia||2.903 (1.646–5.120)||< 0.001||—|| ||3.586 (1.278–10.064)||0.015|
|Maternal age (years)||1.133 (1.068–1.202)||< 0.001||—|| ||1.138 (1.021–1.267)||0.019|
|Maternal BMI (kg/m2)||1.113 (1.045–1.187)||0.001||—|| ||—a|| |
|Oxytocin administration||3.614 (1.884–6.934)||< 0.001||—|| ||—a|| |
|Gestational age (weeks)||1.079 (0.883–1.318)||0.458||—|| ||—a|| |
- Top of page
Our study provides original data on the use of ITU during the active second stage of labor. We showed that both fetal head descent and rotation, as evaluated by intrapartum ultrasound, differed significantly at various phases according to the mode of delivery.
Interestingly we noticed that in the early second stage the AoP appeared to be significantly smaller among those patients who underwent instrumental or Cesarean delivery in comparison with those who experienced a spontaneous vaginal delivery. This difference had lost its statistical significance after 40 min from the beginning of the active second stage, probably because of the progressive reduction in the number of patients at that advanced phase of labor.
On the other hand, the degree of fetal head rotation, as assessed by the MLA, was initially comparable between spontaneous and operative delivery groups, whereas from 40 min onward a smaller angle, indicating the occurrence of internal rotation, was observed among those who experienced a spontaneous vaginal delivery. This seems to suggest that a delayed fetal head rotation is a clinically plausible cause of poor fetal head descent, leading to operative vaginal or Cesarean delivery in the late second stage. This is consistent with cases of persistent failure of fetal head rotation, for example as in deep transverse arrest, in which the fetal head initially descends in the pelvis but fails to complete its rotation in a more advanced phase of labor[18, 19].
Fetal head descent and internal rotation are classically assessed by digital examination. However, clinical evaluation of such findings is commonly reported to be inaccurate and poorly reproducible[1-4], particularly during the second stage of labor because of the development of caput succedaneum or cranial molding. Owing to these limitations, the definition of abnormal progression of fetal head in the second stage remains unclear. The second stage is generally described as abnormal when its duration exceeds that which is expected in accordance with parity or epidural administration[21-25] and the diagnosis of non-progressing labor in the second stage remains exclusively based on the subjective digital perception of a non-descending fetal head. Furthermore, digital examination is still the recommended method to define fetal head station and rotation before instrumental vaginal delivery[16, 26].
In order to increase the objectivity and reliability of the determination of fetal head descent and rotation, the complementary use of ultrasound has been recently proposed with promising results[5, 7-9, 12-14]. Among the suggested parameters, the AoP (also known as the angle of descent) seems to be the most reproducible one to evaluate fetal head station in the sagittal plane[11, 15, 27], whereas the MLA is the only sonographic index of the fetal head rotation in the axial plane. However, very little is known about how these parameters change longitudinally during the second stage of labor. Labor is a dynamic process, and decision making is better based upon time changes of the parameters of interest rather than upon a single evaluation. We have demonstrated that poor fetal head descent (as assessed by the AoP) seems to be an early finding in cases with a higher risk of operative delivery, whereas slow head rotation (as assessed by the MLA) seems to be a late finding.
At multiple logistic regression, both AoP and MLA were independently associated with the mode of delivery when only ultrasound parameters were included. However, when clinical variables were also taken into account the AoP remained the only sonographic predictor of labor outcome, whereas the MLA was replaced with epidural analgesia and maternal age. This does not seem unexpected, as epidural analgesia may affect fetal head rotation and in the multivariable analysis it was found to have a stronger association with the mode of delivery than the MLA itself.
It is important to point out that, as we sought to compare the sonographic indices of fetal head descent among fetuses who will progress to spontaneous vaginal delivery vs those who will be delivered by vacuum or Caesarean section because of poor progress, we opted to exclude women for whom the decision to expedite delivery was only indicated by an abnormal fetal heart trace in the absence of evidence of dystocia. The reason for this is that while ITU may predict an operative delivery as a result of mechanical causes, it is not expected to be useful in predicting cases where obstetric intervention is required as a result of suspected pure fetal hypoxia.
The major strength of this study is that it provides original data on the fetal head progression in the second stage of labor using an objective tool such as 3D ultrasound. However, we do acknowledge some limitations. The first is the limited number of patients who underwent operative delivery, which did not permit the separate analysis of data from patients who underwent vacuum extraction and those who had Cesarean delivery. Furthermore, as all vacuum extractions were successful, we cannot offer data on the sonographic prediction of the instrumental delivery success as previously proposed by some[9, 14].
A second limitation is that in our study we chose to evaluate sonographic data in a resting phase to avoid the interference of maternal pushing on the volume acquisition. However, as previously reported, we strongly believe that uterine contractions and active maternal pushing may significantly affect the sonographic measurements in the second stage of labor and that dynamic rather than static assessments may more accurately predict labor outcome. In addition, it should be pointed out that we performed the logistic regression analysis using the data for all women at all time points in order to include all available data from the entire duration of the active second stage. Consequently, most women were included multiple times within this analysis (i.e. once for each set of measurements obtained) and a possible overestimation of the statistical significance of any associations found cannot therefore be excluded. The final limitation of our study is that the fetal occiput position was not sonographically determined at the beginning of the second stage and we did not separately evaluate sonographic changes in the subset of fetuses presenting with occiput posterior, who are thought to follow different paths of fetal head descent and rotation. We are aware of this and a specific study on fetal head progression and rotation in the second stage among fetuses persisting in the occiput posterior position is underway.
In conclusion, poor fetal head descent or delayed fetal head rotation may be accurately documented in the second stage of labor by serial ITU. The latter may therefore play a role in the prediction of mode of delivery.