Can sonographic depiction of fetal head position prior to or at the onset of labor predict mode of delivery?

Given the irregular shape of the maternal pelvis and the relatively large dimensions of the fetal head at term, accommodation or adaptation of portions of the fetal head to various maternal pelvic planes is required for successful vaginal birth. Positional changes in the presenting part that are necessary to enable the fetus to navigate the pelvic canal are incorporated in the cardinal movements of labor and include: engagement, descent, flexion, internal rotation, extension, external rotation and expulsion1, 2. During labor, these movements, although sequential, often exhibit considerable temporal overlap. For example, engagement is associated with concurrent flexion and descent of the fetal head. Additional adaptive measures include intrapartum moulding of the fetal head: changes in the fetal head shape due to external compressive forces, resulting in a shortened suboccipitobregmatic diameter and a lengthened mentovertical diameter1, 2. Fetal descent in particular is essential for successful (non-operative) vaginal birth and hence serves as a major component in the clinical assessment of progress of labor. Lack of descent is often a result of malrotation of the fetal head, resulting in relative cephalopelvic disproportion. Fetal descent is evaluated clinically by digital vaginal assessment in relation to the maternal ischial spines in a highly subjective and frequently inaccurate manner1-3 and is often further compromised by the presence of moulding of the fetal head1, 2. Ultrasound assessment of labor has been applied increasingly during recent years. Clinical parameters/cardinal movements of labor amenable to intrapartum sonographic assessment include fetal head position4-15, engagement16, 17, descent and internal rotation18-25. Transperineal/translabial ultrasound has been utilized to enhance depiction of descent and internal rotation of the fetal head, and determine the following parameters: head direction, angle of the middle line, progression distance, head–perineum distance and angle of progression18-36. Intrapartum sonographic assessment of fetal descent is promising in that it potentially offers a more objective assessment compared with digital vaginal assessment3. Although clinical application remains to be determined, recently, Bamberg et al. reported real-time open magnetic resonance imaging, in the mid-sagittal plane, of extension of the fetal head in the late second stage of labor37. Of note, sonography may also depict moulding of the fetal head1, 38. In addition, intrapartum sonographic findings of asynclitism (lateral deflection of the sagittal suture), which, when severe, is an additional etiology of cephalopelvic disproportion, have been reported recently1, 2, 39, 40. Intrapartum transabdominal sonographic determination of fetal head position is performed with relative ease. Transverse suprapubic placement of the curvilinear transducer yields an axial image of the fetal head. Precise fetal head position is determined following identification of midline intracranial structures (cavum septi pellucidi, falx cerebri, thalami and cerebellar hemispheres) and anterior or posterior cranial structures (orbits, nasal bridge and cervical spine). While not all of the intracranial landmarks are visualized with each of the various fetal head positions, each sonographic assessment will visualize sufficient structures to define the head position. In most cases (other than direct occiput anterior and direct occiput posterior positions), the well-depicted midline falx cerebri, thalami and cavum septi pellucidi serve to define the fetal head position. When the occiput is directly anterior and when the occiput is directly posterior, the cervical spine and fetal orbits, respectively, will define the precise fetal position. Accordingly, fetal head position is assigned to one of the traditional eight positions: occiput anterior, occiput posterior and the six respective lateral (right or left) occiput anterior, occiput transverse and occiput posterior positions, similar to the traditionally assigned fetal head positions determined by digital vaginal examination1, 2, 5, 6. Determination of fetal head position by transverse suprapubic transabdominal sonographic assessment requires minimal time, and is usually completed without difficulty in less than 2 minutes7, 10. In addition, this relatively simple assessment has favorable interobserver agreement rates; within 15° in nearly 90% of cases, and within 30° in all cases9. Early studies of intrapartum sonographic assessment of fetal head position indicated uniformly the superiority of transabdominal ultrasound over digital vaginal examination in depicting precise fetal head position5-15. The precision of sonographically depicted fetal head position compared with digital vaginal determination has been confirmed in independent studies in both the active5, 7-10 and second stages of labor6, 11-15. In fact, the low reported rates of concurrence between sonographic and digital vaginal fetal head position assessment (24–31.3% in the active stage of labor and 65–70.4% in the second stage of labor) appear to reflect the low precision of digital vaginal assessment5, 6, 8, 11. Inaccuracies of 180° in fetal head position assignment by digital vaginal examination in comparison with sonographic assessment (suggesting correct recognition of the fetal sagittal suture yet incorrect designation of the anterior and posterior fontanels) of 15.7% and 8.9% in the active and second stages of labor, respectively, have been reported5, 6. It is noteworthy that intrapartum sonographic assessments of fetal head position confirm the results of the time-honored classic röentgenologic studies of Caldwell et al.41, 42 and the subsequent study of Calkins43, published in the 1930s. Thus, intrapartum ultrasound demonstrates that the fetal head usually engages in the occiput transverse position ( ± 45°), more commonly in the left than in the right position (56.8% vs 34.5%), consistent with the data of Caldwell et al. (58.5% vs 40.5%)5, 41. Interestingly, similar to the studies of Caldwell et al., in which the greater frequency of left positions was attributed to larger numbers of left occiput transverse (32%) and left occiput posterior (6%) positions, these specific positions were also those most commonly noted at transabdominal intrapartum sonography (23.5% and 17.6%, respectively)5, 41. Furthermore, intrapartum sonography in the second stage of labor showed, similar to the findings of Calkins43, that 53.6% of fetuses were in an occiput anterior position (direct occiput anterior (14.3%), left occiput anterior (28.6%) or right occiput anterior (10.7%))6, and, confirming the finding by Caldwell et al., that left-oriented occiput positions are more common than right ones6, 41. Sonographic assessment of fetal head position bypasses the (traditionally accepted) drawbacks associated with digital vaginal assessment, namely the subjective (and operator experience-dependent) aspects of physical examination, limited cervical dilatation and effacement early in labor, and the presence of caput succedaneum (often noted especially in association with obstructed labor). An additional advantage of sonographic assessment of fetal head position is that sonography avoids the increased risk of ascending infection inherently associated with vaginal digital assessment in the presence of ruptured fetal membranes. The objective nature of imaging assessment, and convincing independent multisource data regarding the clearly superior performance of intrapartum sonographic assessment in comparison with digital vaginal assessment of fetal head position, strongly support that sonographic assessment should become the gold standard of this important clinical parameter. Application of intrapartum sonography has enhanced our knowledge of the mechanisms of the cardinal movements of labor, particularly internal rotation. Classic textbook teachings suggest that, following occiput transverse ( ± 45°) head position at engagement (passage of the biparietal diameter through the plane of the pelvic inlet), descent of the fetal head occurs with concurrent rotation to an occiput anterior (or, less frequently, posterior) position at the midplane level (maternal ischial spines)1, 2. In contrast, intrapartum sonography suggests that this sequence of events may not always unfold4, 7, 44-46. Four separate studies specifically address internal rotation. Gardberg et al. sonographically assessed 408 women in labor after 37 weeks' gestation, following the initial sonographic assessment of fetal head position4. Of note, 57% of the women were in spontaneous labor and 43% had been induced4. The authors noted that most (68%) persistent occiput posterior positions develop through a malrotation during labor, only 32% of persistent cases being occiput posterior at the onset of labor4. Akmal et al., in a cross-sectional study of 918 women in active labor with singleton cephalic presenting fetuses between 37 and 42 weeks' gestation, noted that a majority of occiput posterior-positioned fetuses rotate to the anterior position, even at 10 cm of cervical dilatation44. However, the vast majority of occiput posterior-positioned fetuses were a consequence of persistence of this position during labor rather than a malrotation from an initial occiput anterior or posterior position44. Lieberman et al. studied 1562 nulliparous women during labor (698 in spontaneous labor and 864 undergoing induction) with serial sonographic assessments of fetal head position45. They noted that changes in fetal head position are common during labor, with 36% of women having an occiput posterior fetus on at least one sonographic assessment. Of fetuses that were occiput posterior-positioned late in labor, only 20.7% were occiput posterior at delivery45. Therefore, nearly 80% of fetuses in an occiput posterior position very late in labor rotated to a different position before delivery. Conversely, even fetuses in an anterior occiput position late in labor had a small (5.4%) risk of becoming occiput posterior by the time of delivery. Most fetuses in the occiput transverse fetal position, even late in labor, rotated to a different position, with only 8.3% remaining occiput transverse at delivery. Furthermore, of the 190 women who delivered an occiput posterior fetus, 52% (n = 99) were never occiput posterior on ultrasound examination, and of the 1246 women who delivered an occiput anterior fetus, 33% (n = 413) had an occiput posterior fetus detected by sonography at some time during labor45. Peregrine et al. studied 289 women immediately prior to induction of labor and noted that of 97 women (36%) with occiput posterior-positioned fetuses at induction of labor at or above 36 weeks' gestation, only eight were occiput posterior at delivery46. Of 25 patients with occiput posterior-positioned fetuses at delivery, 68% occurred due to a malrotation from a non-occiput posterior position during labor46. While these initial data relating to sonographic assessment of intrapartum fetal head position during labor, and specifically internal rotation, are not always in agreement, it is clear that intrapartum sonography has enabled further understanding of this process, which appears more complex than originally described. The concept of whether sonographic evaluation of the second stage of labor improves the assessment of progress of labor and its outcome was recently addressed by Drs Yeo and Romero in an Editorial/Opinion in this Journal47. These authors concluded that second-stage sonographic assessment of labor can play a role by: 1) identifying the precise fetal position; 2) noting the presence and extent of caput; 3) assessing the fetal station; and 4) serial ultrasound documentation of the degree of progress during the second stage of labor47. The incidence of Cesarean delivery is increased among patients with occiput posterior-positioned fetuses48, 49. In addition, in comparison with neonates delivered in the occiput anterior position, persistent occiput posterior position at delivery is associated with increased risks of adverse neonatal outcome, including: 5-min Apgar score < 7 (odds ratio (OR), 1.50; 95% CI, 1.17–1.91), acidemic umbilical cord gases (OR, 2.05; 95% CI, 1.52–2.77), meconium-stained amniotic fluid (OR, 1.29; 95% CI, 1.17–1.42), birth trauma (OR, 1.77; 95% CI, 1.22–2.57), admission to neonatal intensive care unit (OR, 1.57; 95% CI, 1.28–1.92) and longer neonatal hospital stay (OR, 2.69; 95% CI, 2.22–3.25)50. Akmal et al., in 2004, assessed whether intrapartum sonographic depiction of fetal head position in early labor could predict the likelihood of Cesarean delivery10. They assessed 601 patients in active labor at term with cephalic presenting fetuses and with cervical dilation between 3 and 5 cm. Delivery was vaginal in 514 (86%) cases and by Cesarean in 87 (14%)10. Fetal head position was posterior in 209 (35%) cases and in this group the incidence of Cesarean delivery was 19% (40 patients), compared with 11% (17 of 392) in the non-occiput posterior position group. Multiple regression analysis revealed that significant independent contribution to the prediction of Cesarean delivery was provided by: maternal age (OR, 1.1; 95% CI, 1.0–1.2), African-Caribbean origin (OR, 2.4; 95% CI, 1.2–4.6), height (OR, 0.93; 95% CI, 0.89–0.97), parity (OR, 0.2; 95% CI, 0.1–0.4), type of labor (spontaneous vs induction) (OR, 2.2; 95% CI, 1.3–3.8), gestational age (OR, 1.4; 95% CI, 1.1–1.7), fetal head descent (OR, 0.6; 95% CI, 0.4–0.9), occiput posterior position (OR, 2.2; 95% CI, 1.3–3.7) and male gender (OR, 2.0; 95% CI, 1.2–3.5). These results led the authors to conclude that the risk of Cesarean delivery can be estimated during the early stages of labor utilizing the sonographically determined fetal head position in addition to traditional maternal, fetal and labor-related characteristics10. In the systematic review published in this issue of the Journal, the authors attempt to determine whether sonographic assessment of fetal head position (or, in their words, ‘assessment of occipital position’) at the onset of spontaneous labor or induction may predict the mode of delivery (vaginal vs Cesarean) in a large group of patients51. This objective is altruistic in that Cesarean delivery, if deemed necessary (and performed) relatively early, may circumvent or at least mitigate maternal morbidities associated with Cesarean delivery performed later in labor, following arrest of labor or failed attempt at vaginal birth52, 53. The authors reviewed the available relevant literature and selected for inclusion in their meta-analysis 11 studies4, 8, 10, 45, 46, 54-59 which, published between 1999 and 2012, reflect data accumulated throughout the last 14 years51. Inclusion criteria were: availability of numerical data allowing construction of 2 × 2 tables cross-classifying fetal head position prior to, or at the beginning of, labor and the occurrence of Cesarean delivery or position of the fetal head at delivery51. Regarding two studies in which published data were initially insufficient, i.e. not meeting inclusion criteria for their analysis, the researchers contacted the authors who provided the required missing data, thus allowing their inclusion8, 45. Overall, the authors analyzed data pertaining to 5053 patients, with sample sizes of included studies ranging between 128 and 1562 patients51. Despite the authors' efforts to include appropriately designed studies to assess their objective, the 11 studies assessed were neither uniform nor homogeneous. Nine of the 11 studies included patients who underwent induction of labor subsequent to the sonographic evaluation4, 8, 10, 45, 46, 54, 56, 57, 59, thus incurring the inherent increased risk for fetal malposition and Cesarean delivery associated with induction of labor60-65. Five studies included patients in spontaneous labor and patients undergoing induction of labor4, 8, 10, 45, 56, four studies exclusively addressed patients undergoing induction of labor46, 54, 57, 59 and only two studies were limited to inclusion of patients in spontaneous labor55, 58. Overall, fewer than half of the patients (n = 2116, 41.87%) in the cumulative data assessed in the meta-analysis underwent true intrapartum sonographic assessment of fetal head position, i.e. were assessed during spontaneous labor. Significantly, data were not dichotomized uniformly into groups of nulliparous vs parous patients throughout the 11 studies. Although parity was often recorded, in the majority of the studies included in the systematic review, data regarding obstetric outcomes of nulliparous and parous patients were combined. This point is paramount in attempting to predict the mode of delivery in that nulliparous and parous patients represent two distinctly different groups of patients, in which the latter have previously exhibited an adequate pelvis and cervical and pelvic tissues present less resistance to uterine contractions10. Accordingly, parity is a strong independent predictor of mode of delivery and, indeed, Akmal et al. noted a nearly seven-fold difference between nulliparous and parous women in the incidence of Cesarean delivery10. In fact, regretfully, only one study exclusively addressed nulliparous patients (the most important group of patients in determining risk factors for pelvic inlet and midpelvic arrest of labor disorders)45. In contrast, appropriate separation of nulliparous vs parous patients has been reported recently by Levy et al., in their study investigating whether angle of progression as measured in pregnant women before the onset of labor could predict mode of delivery36. Not surprisingly, results differed between the two groups. While among non-laboring nulliparous women a narrow angle of progression (< 95°) was associated with a significantly increased risk of Cesarean delivery, most parous women with such measurements delivered vaginally36. There was additionally a lack of uniformity in the systematic review51 regarding status of the fetal membranes. In fact, one of the included studies addressed prediction of mode of delivery according to sonographic depiction of fetal head position exclusively among patients with prelabor rupture of membranes after 37 weeks' gestation56. Similarly, in the study by Souka et al., only women with ruptured fetal membranes were assessed8. In the majority of the remaining studies, the status of fetal membranes was not provided. Clearly, this parameter may also affect mode of delivery. Furthermore, as the authors of the current systematic review note51, clinicians were not always blinded to sonographic data, potentially introducing biased/skewed data, in that Cesarean delivery may have been influenced by intrapartum sonographic evidence of a malpositioned fetus. Given the heterogeneity of the data assessed, which include a relatively large number of patients who underwent subsequent induction of labor, it is not surprising that the incidence of Cesarean delivery varied considerably, ranging between 8.3% and 32% in the 11 differently designed studies51. In addition, as the authors themselves acknowledge, despite the intention of prediction of Cesarean delivery (resulting from fetal malposition-related relative cephalopelvic disproportion), indications for Cesarean were not available. Given the sonographically depicted dynamic process of labor (specifically internal rotation) discussed earlier, I am unconvinced regarding computations provided in the systematic review suggesting that sonographic assessment of fetal head position before or at the beginning of labor may predict position of the fetal head at delivery51. Notwithstanding the limitations of this systematic review51, the authors' conclusion that sonographic assessment of fetal head position at the onset of (spontaneous or induced) labor cannot predict the subsequent mode of delivery is not surprising. Labor is a complex process, dependent on the interactions between three variables: the powers (forces generated by uterine musculature), the passenger (fetus) and the passage (bony maternal pelvis and resistance provided by soft tissues)2. Fetal head position assessed sonographically at the onset of labor is but one factor, and represents an assessment at a single point in time, of what naturally is a lengthy process. As the authors note, numerous confounding factors may affect labor. These include (but are not limited to), maternal height, weight, parity, type of maternal pelvis, maternal gestational diabetes, spontaneous vs induced (or augmented) labor, quality of maternal contractions, gestational age, fetal weight (specifically macrosomia), status of fetal membranes (intact or ruptured), presence of neuroaxial anesthesia and possible non-reassuring fetal status. In summary, due to the inherent weaknesses of this systematic review51, in my opinion the conclusion, although probably correct, remains unsubstantiated. I believe that it is an oversimplification to consider that a single isolated ‘snapshot’ sonographic assessment of fetal head position in early spontaneous labor (or prior to induction of labor) would suffice to predict the subsequent mode of delivery. Furthermore, data gleaned from intrapartum sonography support the view that changes in fetal head position are common throughout labor. In fact, fetal head position at delivery is often not determined until very close to delivery45. Given current knowledge resulting from intrapartum sonography data pertaining to internal rotation of the fetal head, I doubt that mid-pelvic arrest of labor can be predicted based upon sonographic determination of fetal head position at the onset of labor (either prior to or immediately after engagement). If physical examination of the fetal head position (albeit with limited accuracy), cervical status and fetal head station prior to or at the onset of labor cannot precisely predict the mode of delivery, why should we expect ultrasound depiction of fetal head position alone (at the onset of labor) to perform better? Although I suspect the results might be similar, a preferable approach to assess whether sonographic imaging of fetal head position in early labor may assist in the prediction of mode of delivery could be evaluation of this parameter (in both nulliparous and parous patients, separately) prospectively, only in the presence of well-established, spontaneous active labor, thus mitigating the increased risk of Cesarean delivery inherently associated with induction of labor60-65. I believe that, despite recent advances, the full potential of intrapartum sonography has not yet been realized. In predicting mode of delivery it clearly offers the opportunity to enhance the clinical management of labor. Further refinement of intrapartum sonographic capabilities (especially regarding less complex assessment of fetal descent) is required. Although utilization of the angle of progression and other geometric models (which, indirectly, include a pelvic structure, the fibrocartilaginous symphysis pubis) to assess fetal station appear promising, direct incorporation of an osseous landmark, such as the midpelvic ischial spine (utilized in the highly subjective digital assessment), in the sonographic assessment of fetal descent remains to be developed. In addition, detailed pelvimetry (potentially three-dimensional intrapartum sonographic pelvimetry) may further enhance the sonographic prediction of subsequent mode of delivery. Expecting such outcome data from a single sonographic assessment of fetal head position prior to induction or at the onset of spontaneous labor is, in my opinion, placing the bar too high and at this stage appears unrealistic.