Is routine X-ray pelvimetry of value to decide on mode of delivery for women with labor dystocia?

Background: X-ray pelvimetry is used for evaluation of pelvic inlet generally to diagnose cephalopelvic disproportion (CPD) or contracted inlet. Cesarean section delivery (C/S) is often performed for labor dystocia without CPD or contracted inlet. Here, we examined whether X-ray pelvimetry is useful to decide on mode of delivery in women with dystocia. Method: A total of 1118 pregnant women received X-ray pelvimetry before or during labor. Overall, 205 women with cesarean deliveries for indications except for dystocia were excluded. The remaining 913 women undergoing induction/augmentation were retrospectively investigated. Obstetrical and maternal variables were analyzed by univariate, multivariate or ROC analysis. Results: Among 913 women, 37 (4%), including three with contracted inlet and seven with CPD, gave birth by C/S, whereas 876 gave birth by vaginal delivery. Low maternal height, older age, small obstetrical conjugate, large weight and infant head size were signi�cantly associated with risk of C/S for dystocia. Multivariate analysis revealed that the obstetrical conjugate was an independent variable for risk of C/S. The area under the ROC curve (AUC) and the optimal cut-off values, respectively, were as follows:


Background
X-ray pelvimetry is a digital examination to evaluate the pelvic cavity of pregnant women and is often used for the diagnosis of contracted pelvic inlet and cephalopelvic disproportion (CPD).Women who are short in stature are occasionally subjects of pelvimetry since low maternal height is assumed to correlate with contracted pelvic inlet.In East Asia, pregnant women with low height less than 150 cm are observed, and such women are evaluated by pelvimetry as a screening test for contracted inlet.Additionally, unengaged fetal head during labor regardless of effective uterine contraction suggests CPD, and pelvimetry is considered as an objective mean precisely to distinguish CPD.Such women often receive labor induction or augmentation by oxytocin after excluding CPD.In terms of CPD diagnosis, X-ray pelvimetry is very useful because of the simple and rapid examination involved.
The guidelines of maternal-fetal medicine and/or the attached documents of pharmaceutical agents for labor induction/augmentation often show that X-ray pelvimetry is recommended to negate contracted inlet prior to induction/augmentation (1).However, a Cochran review in 2017 reported that routine X-ray pelvimetry leads to the increased rate of cesarean delivery but does not improve other perinatal outcomes (2).Dystocia sometimes occurs during labor due to various mechanisms, even when CPD or contracted inlet are negated.In such cases, cesarean delivery for dystocia is often performed following a trial of vaginal delivery, including pharmaceutical or mechanical induction/augmentation.The cesarean birth rate increased for a decade starting in 2000 and has remained unchanged from 2010 at 32-33% in the United States (3).More than 85% of cesarean deliveries are performed for four major reasons: nonreassuring fetal status, fetal presentation, prior cesarean operation and dystocia.When repeat cesarean deliveries are included, approximately 60% of all cesarean deliveries are attributable to the diagnosis of abnormal labor (4).Thus, the diagnostic criteria of the arrested labor have been revised to prevent unnecessary rst cesarean deliveries for dystocia in the United State (5).It is important for reducing the cesarean birth rate to understand the risk factors of cesarean delivery for dystocia.
We have perform routine X-ray pelvimetry to evaluate the pelvic cavities of all pregnant women prior to pharmacological induction/augmentation.In this study, we retrospectively investigated their obstetrical variables, including evaluation of the pelvic inlet (obstetrical conjugate and transverse diameter) to determine whether routine X-ray pelvimetry can identify women at risk of cesarean delivery for dystocia.
Our data indicated that a small obstetrical conjugate was an independent warning sign associated with cesarean delivery for dystocia, while low maternal height could substitute for the conjugate, suggesting routine X-ray pelvimetry is not necessary for deciding on mode of delivery except for the negation of CPD.

Methods
A total of 1118 pregnant women had received X-ray pelvimetry at the end of pregnancy from 2014-2017 in the Tokyo Metropolitan Ohtsuka Hospital.The X-ray pelvimetry was performed routinely prior to pharmaceutic and/or mechanical methods of labor induction/augmentation to negate contracted inlet or CPD.
X-ray pelvimetry was performed by the Guthmann and Martius methods, which are standard for the evaluation of the pelvic inlet.The obstetrical conjugate and transverse diameter of all cases were measured by a specialist in maternal-fetal medicine.CPD was diagnosed due to unengaged fetal head, suspected large for date infant and maternal contracted inlet.Maternal contracted inlet was diagnosed by the de nition of the Japanese Society of Obstetrics and Gynecology (JSOG) as follows: the obstetrical conjugate measures 10.5 cm or more for normal, 9.5-10.5 cm for relatively contracted, and less than 9.5 cm for contracted inlet; and the transverse diameter measures 11.5 cm or more for normal, 10.5-11.5 cm for relatively contracted, and less than 10.5 cm for contracted inlet.
Patient backgrounds and obstetrical and infant variables were investigated retrospectively.Comparisons in these variables between cesarean versus vaginal delivery groups were made by the Mann-Whitney U test or Fisher's exact test.The independent risk factor for cesarean birth was analyzed by multivariate regression analysis.P value < 0.01 was de ned as a signi cant difference.Areas under the ROC curve (AUCs) and cut-off values of each variable for cesarean delivery were calculated by ROC analysis.This study was approved by the Ethics Committee of our hospital, and decisions of enrolled pregnant women to be included in this study were expressed by the opt-out approach.

Comparison of maternal and obstetrical variables between the C/S (for dystocia) group and the vaginal delivery group
To determine which variables are related to risk of C/S for labor dystocia, we focused on women with cesarean delivery for dystocia and then compared them to those with vaginal delivery.We rst excluded 205 women who underwent cesarean section (C/S) because of indications except for dystocia; nonreassuring fetal status, twin pregnancy, maternal hypertensive disorder, breech presentation, threated uterine rupture after previous uterine surgery, and intrauterine infection.The remaining 913 women were enrolled in this study.Overall, 37 (4%) of 913 women underwent C/S for dystocia, while 876 women gave birth by vaginal delivery (summarized in Fig. 1).Among them, seven cases were diagnosed as CPD according to the clinical situation.
Next, we compared between the C/S for dystocia vs. vaginal delivery groups in various obstetrical variables to identify a possible warning sign of C/S for dystocia (Table 1).Low maternal height, older age, small obstetrical conjugate, large weight and head size of infant were associated with risk of C/S for dystocia with signi cant differences (Mann-Whitney U test; p < 0.01).The differences in the transverse diameter, gestational week of delivery and fetal biparietal diameter (BPD) between the two groups were marginally signi cant, with p values from 0.01-0.05.The other variables did not differ between the two groups.Among the variables to show signi cant associations with C/S for dystocia, maternal height and the obstetrical conjugate were the only personal maternal characteristics that did not depend on the condition of the birth canal, uterine contractions or the newborn infant.Maternal height in the C/S group (154 ± 5.77 cm) was clearly shorter than that in the vaginal group (158 ± 5.39 cm), and the obstetrical conjugate in the C/S group (11.6 ± 0.90 cm) was clearly smaller than that in the vaginal group (12.2 ± 0.77 cm).One woman was shorter in stature than 150 cm (height = 142 cm), and she was diagnosed as contracted inlet (conjugate = 9.3 cm) and underwent elective C/S.Eight women were diagnosed with relatively contracted inlets by the obstetrical conjugate, and seven gave birth vaginally, whereas one gave birth by C/S.Among 37 women with C/S, only seven were diagnosed as CPD by abnormal labor conditions, including fetal head unengagement.Of seven with CPD, three were diagnosed as contracted inlet and one as relatively contracted inlet by the transverse diameter.The remaining 29 women underwent C/S for dystocia under conditions that did not include either contracted inlet or CPD.Multivariate analysis of obstetrical variables to identify women with risk of C/S for dystocia Older pregnant women are likely to undergo C/S due to not only their cervical resistance, uterine dysfunction or both but also maternal exhaustion.Indeed, maternal age in the C/S group (37 ± 5.5) was four years older than that in the vaginal delivery group (33 ± 5.4) in this study.Infant head size and birth weight in the C/S group were also greater than those in the vaginal delivery group, although these variables might be in uenced by gestational week of delivery.Thus, there were several confounding variables in the univariate analysis shown in Table 1.We then tried to identify independent variables to be associated with C/S for dystocia by logistic regression analysis (Table 2).Eight variables that showed signi cant differences (p < 0.05) between the C/S and vaginal delivery groups by univariate analysis were evaluated.The multivariate analysis revealed that the obstetrical conjugate was just an independent variable for C/S.The obstetrical conjugate was related inversely to the incidence of C/S for dystocia, with an odds ratio of 0.32 (95%CI: 0.14-0.76,p = 0.0098), while the transverse diameter did not show a signi cant inverse correlation.Maternal age and height, infant weight and head size were not independent variables associated with the risk of C/S for dystocia.Univariate and multivariate analyses revealed that maternal height and the obstetrical conjugate are important variables associated with the risk of C/S for dystocia.Correlation of the obstetrical conjugate to maternal height A correlation between maternal height and the obstetrical conjugate in Asian women has not been clari ed, although the two maternal physical characteristics are likely to correlate to each other.We then plotted the two variables of 913 women, as shown in Fig. 2. The open and closed dots indicate women with vaginal delivery (n = 876) and cesarean delivery (n = 37), respectively (Fig. 2A).The dots show the signi cant correlation between maternal height and the conjugate (R = 0.483, p < 0.001).Among them, women with C/S delivery were then plotted in Fig. 2B, and the correlation between maternal height and the conjugate was also observed (R = 0.661, p < 0.001).By comparison between the open and closed dots, the closed dots were likely to be plotted at a lower height and a smaller conjugate than the open dots, corresponding to our data in Table 1.These data indicated that maternal height was correlated with the obstetrical conjugate regardless of the mode of delivery.

ROC curve analysis to evaluate maternal variables for discrimination of mode of delivery
Our multivariate analysis revealed that the obstetrical conjugate was an independent variable associated with risk of C/S for dystocia.We also demonstrated the correlation between the conjugate and maternal height.We then tried to examine whether alternative maternal variables could be warning signs of C/S for dystocia compared with the obstetrical conjugate by ROC curve analysis (Fig. 3).We considered that ROC curve analysis can evaluate various maternal variables as measures that either con rm the presence of risk of C/S or rule out the risk.ROC curves of the maternal variables height, weight before pregnancy, the obstetrical conjugate and transverse diameter were calculated by plotting the sensitivity versus 1speci city of each threshold value.These ROC curves provided areas under the ROC curve (AUCs) and the optimal cut-off value of each variable and enabled us to compare the diagnostic abilities of the four variables to discriminate the possible risk of C/S in women with dystocia.The AUCs and the optimal cutoff values, respectively, were as follows: maternal height (bold line): 0.70 and 155.5 cm (odds ratio = 4.33), maternal weight before pregnancy (dotted thin line): 0.55 and 49.7 kg (odds ratio = 1.98), obstetrical conjugate (bold dotted line): 0.68 and 11.7 cm (odds ratio = 4.27), and transverse diameter (thin line): 0.59 and 11.4 cm (odds ratio = 1.82) (Fig. 3).The lines for maternal height and the obstetrical conjugate overlapped each other, and these AUCs (0.70 vs. 0.68) were clearly higher than those of maternal weight and transverse diameter.The odds ratios of C/S at the optimal cut-off value were 4.33 for maternal height and 4.27 for the obstetrical conjugate, indicating that maternal height was comparable to the obstetrical conjugate as a measure to discriminate the maternal risk of C/S.

Discussion
In this study, we investigated the association between measures (the obstetrical conjugate and transverse diameter of pelvic inlet) of X-ray pelvimetry and cesarean delivery for labor dystocia using data from 913 women in labor.Our data indicated that the obstetrical conjugate was an independent variable correlating inversely with the incidence of C/S for dystocia but that maternal height could be a comparable variable to the conjugate in terms of the discrimination of the risk of C/S.
X-ray pelvimetry is a useful measure to evaluate the inlet to exclude CPD or contracted inlet and is recommended prior to labor induction/augmentation to negate the contracted inlet according to a guideline of maternal-fetal medicine in Japan (1).Thus, we have routinely performed pelvimetry on all women who are to undergo induction or augmentation for various reasons.However, among 913 women, contracted inlet was only observed in four women (4/913 = 0.4%).Conversely, a Cochran systematic review demonstrated that women who receive X-ray pelvimetry are likely to undergo C/S.Dystocia is not only one of the major reasons for C/S but is also associated with pelvic disproportion, which is evaluated by pelvimetry.Then, we took an interest in the clinical question of whether measurements obtained from X-ray pelvimetry are associated with the incidence of C/S and in uence our decision of mode of delivery in the cases of dystocia.Our database of more than one thousand measures obtained from routine X-ray pelvimetry was valuable to clarify this question.The Cochran review performed in 2017 identi ed ve studies with a total of 1159 women, and these studies compared X-ray pelvimetry (n = 582) versus no pelvimetry (n = 577) (2).The number of women who received X-ray pelvimetry in our study was greater than that in the Cochran review.
In this study, we excluded women undergoing C/S due to nonreassuring fetal status, fetal presentation or prior uterine operation from the subject of analysis.All of the remaining 913 women received X-ray pelvimetry prior to labor induction and/or augmentation and delivered a baby by either cesarean or vaginal deliveries.Overall, 37 women with C/S for dystocia and 876 women with vaginal delivery were compared.Univariate analysis revealed that maternal height and age, infant weight and head size, and the obstetrical conjugate were associated with risk of C/S, with signi cant differences.Regarding measures of X-ray pelvimetry, a signi cant difference (p < 0.01) was found in the obstetrical conjugate, but not the transverse diameter.These obstetrical variables are likely to confound each other; thereby, multivariate analysis was needed to identify an independent variable.Multivariate analysis revealed that the obstetrical conjugate was the only independent variable associated with C/S for dystocia.Taken together with data from the 2017 Cochran review, measures of the obstetrical conjugate might in uence us to make decisions of mode of delivery.In our data, although the obstetrical conjugate identi ed a woman with contracted inlet and eight women with relatively contracted inlet, only two of them underwent C/S, suggesting that the obstetrical conjugate was not involved in the decision of mode of delivery.In other words, our statistical analysis indicates that a larger obstetrical conjugate resulted in a reduction of risk of C/S for dystocia (Odds ratio = 0.32).
Previous studies (6, 7) and a systematic review (2) have demonstrated that X-ray pelvimetry does not improve perinatal outcome and that routine X-ray pelvimetry is not useful to women with normal presentation.Regarding women with previous C/S or breech presentation, some studies have reported the usefulness of X-ray pelvimetry (8, 9).Our study did not address the usefulness of X-ray pelvimetry from the viewpoint of perinatal outcome, and thereby, we did not compare the outcome with the no-X-ray pelvimetry group.
In this study, the transverse diameter identi ed three cases of contracted inlet and one relatively contracted inlet, and all underwent C/S with CPD, while the obstetrical conjugate identi ed only two women with C/S among nine with contracted or relatively contracted inlet.The transverse diameter was more signi cant than the obstetrical conjugate in terms of the diagnostic measure of contracted inlet.
Nevertheless, the abnormality detected by X-ray pelvimetry rarely occurred in our population (contracted inlet: 4/913, relatively contracted: 9/913, and CPD: 7/913 women).Several studies have reported that maternal height with or without a measure of pelvimetry can be a variable to predict dystocia (10) or CPD (11).Maternal height is also reported to correlate to the obstetrical conjugate (12).Indeed, our data clearly indicated the correlation between maternal height and the obstetrical conjugate, regardless of mode of delivery.Since the obstetrical conjugate was associated with risk of C/S for dystocia, maternal height in our population was thought to be an alternative variable for the conjugate.Our ROC analysis demonstrated the hypothesis that maternal height is comparable to the obstetrical conjugate to discriminate maternal risk of C/S for dystocia.Numerous previous studies have demonstrated that maternal height is inversely related to risk of not only C/S (13-17) but also labor dystocia (18-21).
Taken together with our ROC curve, maternal short stature is su cient to be an independent variable associated with the decision of mode of delivery.
In contrast, the transverse diameter did not show high diagnostic capacity.Interestingly, the 11.4 cm cutoff value for the transverse diameter on the ROC curve was concordant with the upper criterion (11.5 cm) of a relatively contracted inlet, while 11.7 cm for the obstetrical conjugate was quite different from the upper criterion (10.5 cm) of a relatively contracted inlet.

Conclusion
X-ray pelvimetry is an important measurement to evaluate the pelvic inlet, especially prior to labor induction/augmentation.However, contracted inlet is a very rare event (approximately 1%) in Asian women, and maternal height can be comparable to measures of X-ray pelvimetry.Our data indicated that routine X-ray pelvimetry is not of value to decide on mode of delivery for women with dystocia.

Declarations
Ethics approval and consent to participate and publication:

Figures
Figures

Figure 1 Subject
Figure 1

Table 1
Comparison between C/S for dystocia vs. vaginal delivery groups *: Median and SD **: p value by Mann-Whitney U test

Table 2
Multivariate analysis of risk factors for C/S **: p value by Logistic regression analysis