SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

Objective To investigate clinical and ultrasonographic predictors of outcome of external cephalic version at term.

Design Prospective observational study.

Setting University obstetric unit.

Population All external cephalic versions performed over two years (n=243).

Methods Nineteen different clinical and ultrasonographic variables were recorded before each procedure. The ability of each of the 19 variables to predict the success or failure of external cephalic version was assessed by univariate analysis. The study population was then divided into two subgroups of 129 and 114 patients by random allocation using computer generated numbers. Logistic regression was performed in each subgroup to assess the relative importance and independence of the important variables. The derived regression models were then applied to the other subgroup of patients to assess accuracy and reproducibility.

Results The overall success rate of the procedure was 69.5%. Both regression models identified the same three variables as independent predictors of failed versions: 1. presenting part engaged; 2. difficult to palpate the fetal head, and 3. a tense uterus on palpation. The two models correctly predicted 75.2% and 84.2% of outcomes in the other subgroup. If uterine tone, which was assessed after administration of tocolytic, was excluded from the analysis, the other two factors remained in the models, with the addition of nulliparity as a significant predictor of failed external cephalic version. The chance of success of external cephalic version in the original 243 women was found to be < 20% if two of these variables were present, 0% if all three were present, and 94% if none were present.

Conclusions The outcome of external cephalic version can be predicted by easily available clinical parameters.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

External cephalic version, performed after 36 or 37 weeks of gestation, effectively reduces the incidence of breech deliveries and the number of caesarean sections1. Various clinical and ultrasonographic variables, including parity, placental location, position of the fetal spine, estimated fetal weight, amniotic fluid volume, and type of breech have been reported to be predictive of successful external cephalic version (Table 1)2–12. On the other hand, spontaneous version tends to occur in those with nonengagement of the presenting part if external cephalic version has not been performed13. Knowledge of these predictors allows us to provide more specific and individualised counselling of patients regarding treatment options for breech presentation at term. The aim of this prospective study was to determine the predictive value of a number of factors with regard to predicting the outcome of external cephalic version.

Table 1.  Commonly reported factors affecting the outcome of external cephalic version (ECV). S = the factor was significantly associated with the outcome of ECV, either positively or negatively; NS = the factor was not found to be related to the outcome of ECV; —= the factor was not examined in that study; FW = fetal weight
 No. of ECVParityObesity/BMI/Maternal weightPlacental siteType of breechPosition of fetal spineAmniotic fluid volumeEngagement/stationEstimate FW
  1. *Studies which included pregnancies less than 36 weeks of gestation

Brocks et al.231SSS     
Donald et al.365NS NSSSNSNS 
Ferguson 2nd et al.4158S sSSS NS
Fortunato et al.567NSSNSNSSSS 
*Hellstrom et al.6300SNSNSS S  
Hofmeyr et al.780S SNS NS  
*Kornman et al.8133SS      
Morrison et al.9304  S     
Newman et al.10108SNSS   SS
*Norde et al.11100S S     
Shalev et al.1255NSNSNSNSNSS S

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

This prospective study was undertaken over a two-year period in a university teaching hospital in which external cephalic version has been adopted as the preferred treatment option for the management of breech presentation at term. The procedure was offered to all women with breech presentation at 36 weeks of gestation or beyond where no contraindication existed. We have described the external cephalic version procedure, including contraindications, as performed in our unit14. Briefly, a reactive cardiotocogram was a pre-requisite and tocolytic agents were used except where contraindicated. The standard tocolytic used during the study period was 10 μg of intravenous bolus of hexaprenaline. External cephalic version was only performed by experienced operators and attempts were limited to a maximum duration of 5 minutes. Ultrasound examination was repeated after the procedure to confirm the fetal presentation and to exclude fetal bradycardia. Cardiotocography was also repeated.

Nineteen clinical and ultrasonographic factors considered likely to influence the outcome of external cephalic version were prospectively recorded before external cephalic version and administration of tocolytics (Tables 2 and 3), with the exception of uterine tone which was assessed after administration of tocolytics. The assessment of uterine contraction before external cephalic version and position of fetal spine were introduced in the second half of the study period. Most of these variables are widely used in clinical practice and require no further explanation. Presence or absence of uterine contraction was based on cardiotocogram. Women were considered in labour if they had painful regular uterine contraction every 3 minutes or more before the procedure. Clinical impressions were subjective assessments by operators and were classified as in Table 3. Body mass index (the twentieth variable) was not recorded before the procedure but was derived from maternal weight and height.

Table 2.  Details of continuous variables recorded prospectively before external cephalic version. Values are given as mean [SD]
 Outcome of external cephalic version
Variables recordedSuccessful (n= 169)Failed (n= 74)P
  1. *Calculated based on the Shepard formula: EFW = 10 exp(AC × 0.046 - BPD × AC × 0.002646 + BPD × 0.166 + 1.25081)

Clinical and historical   
Maternal weight (kg)64.2 [8.2]64.7 [7.9]0.693
Maternal height (cm)156.8 [5.2]157.9 [5.4]0.109
Symphysis fundal height (cm)36.4 [2.2]36.1 [2.0]0.232
Body mass indexf26.1 [3.1]25.9 [2.8]0.603
Ultrasonographic   
Biparietal diameter [BPD] (cm)9.1 [0.4]9.1 [0.4]0.337
Femur length (cm)6.9 [0.3]6.9 [0.3]0.873
Abdominal circumference [AC] (cm)32.8 [2.1]32.0 [2.3]0.007
Estimated fetal weight [EFW]*(g)3072 [487]2917 [485]0.023
Four quadrant amniotic fluid index (cm)11.9 [4.0]9.6 [3.3]0.000
Table 3.  Details of categorical variables recorded prospectively before external cephalic version (ECV). Values are given as n(%)
Variables recordedECV successfulECV failedP
Clinical and historical   
Parity   
Primigravida73 (43-2)55 (74-3)0-000
Multipara96 (56-8)19 (25-7) 
Uterine contraction present   
Yes36 (46-8)22 (56-4) 
No41 (53-2)17 (43-6)0-326
In labour at time of ECV   
Yes1 (0-0)2 (2-7)0-170
No168 (100)72 (97-3) 
Ultrasonographic   
Type of breech   
Extended (both legs)62(36-7)35(47-3) 
Any other types107(63-3)39(52-7)0-120
Location of placenta   
Anterior38 (22-5)17 (23-0) 
Posterior28 (16-6)11 (14-9) 
Lateral23 (13-6)5 (6-8) 
Fundal20 (11-8)14 (18-9) 
Cornual60 (35-5)27 (36-5)0-398
Position of fetal spine   
Anterior or posterior25 (32-5)15 (38-5) 
Lateral52 (67-5)24 (61-5)0-521
Clinical examination   
Engagement of fetal part   
Yes14 (8-3)35 (47-3) 
No155 (91-7)39 (52-7)0-000
Obesity of abdomen   
Fat15 (8-9)7 (9-5) 
Thin26 (15-4)4 (5-4) 
Unremarkable128 (75-7)63 (85-1)0-093
Abdominal wall   
Muscular13 (7-7)23 (31-1)0-000
Unremarkable156 (92-3)51 (68-9) 
Uterine tone   
Relaxed110 (65-1)11 (14-9) 
Tense15 (8-9)47 (63-5) 
Unremarkable44 (26-0)16 (21-6)0-000
How easy to palpate fetal head   
Easy115(68-0)14(18-9) 
Difficult15 (8-9)32 (43-2) 
Unremarkable39 (23-1)28 (37-8)0-000

The association of each of these variables with the outcome of external cephalic version was assessed by univariate analysis to identify potential predictors for external cephalic version outcome. The study population was then divided into two groups based on computer generated random numbers. From each group, further analysis by logistic regression was performed to investigate the relative importance of the variables in predicting the outcome of external cephalic version. Dummy variables were created for analysis of categorical variables consisting of more than two categories. The derived regression model was then tested in the other group of patients to assess the accuracy and reproducibility of the model.

The Statistical Package for Social Sciences (SPSS Inc, Illinois, USA) was used for analysis. For univariate analysis of continuous variables, a t test was used, while the χ2 test was used for the analysis of categorical variables. A value of P < 0.05 was considered statistically significant. For logistic regression, forward stepwise selection was used.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

During the two-year study period 243 external cephalic versions were performed. The age of women ranged from 17 to 43 with a mean of 28.5. There were 128 nulliparous and 115 multiparous pregnancies. Gestational age at the time of external cephalic version ranged from 36 to 43 weeks, with a mean of 38 weeks. Three women were in early labour at the time of version. Three women had a previous lower segment caesarean section. Tocolytic agents were used in all but eight women.

The 243 women were randomly divided into two groups of 129 and 114 women. The overall success rate of the procedure was 69.5% (169 out of 243 attempts). Eight variables were found to be significantly associated with the outcome of external cephalic version (Tables 2 and 3).

All of the eight variables, together with those with a P value of < 0.2 based on univariate analysis, were examined by logistic regression modelling in each group of women. In addition, placental site and position of fetal spine were also included in the modelling process because they have been found to be important predictors by previous researchers2–5. Regression models derived from the two subgroups consisted of the same variables: 1. whether the presenting part was engaged, 2. whether the fetal head was difficult to feel, and 3. whether the uterus was tense on palpation; although the regression coefficients and constants were different (Table 4). The results of applying the regression models to the opposite group of women are summarised in Table 4. The weighted average of the overall accuracy of predicting success or failure of external cephalic version for the two models was 79.4%.

Table 4.  The study group was divided into two subgroups and multiple regression models were derived from both subgroups to predict successful ECV. Both models consisted of the same variables and the coefficients were reported. This table also shows the result of applying these models to the other subgroups in predicting the outcome of ECV. Values are given as coefficients of the regression model [SE] or n (%)
  Coefficients of the regression modelOutcome of ECV correctly predicted by regression models derived from the other subgroup of data
  ConstantBreech nonengagedFetal head not difficult to palpateUterus not tense on palpationSuccessful versionsFailed versionsBoth successful and failed versions
Subgroup 1n= 129−3.0902.328 [0.603]1.738 [0.582]2.049 [0.58]78(87.6)19(47.5)97(75.2)
Subgroup 2n = 114−4.5322.343 [0.783]2.904 [0.738]2.038 [0.646]73(91.3)23(67.6)96(84.2)

If uterine tone was excluded from analysis, the resultant regression model consisted of three independent variables (i.e. engagement, how easy it was to feel the fetal head and parity). Table 5 shows the failure rate of external cephalic version in the presence of different combinations of these three independent predictors when applied to the original 243 patients.

Table 5.  Outcome of ECV in the presence of different combinations of three significant predictors among 243 ECVs. += presence of a characteristic; –= absence of a characteristic. Values are given as n or n (%)
Combination of patients' characteristics  
Presenting part engagedFetal head difficult to palpateNulliparityNo. of casesNo. of failed ECV
835 (6-0)
+4935 (71-4)
+4732 (68-1)
+12855 (43-0)
++1414 (100)
++3326 (96-3)
++3026 (86-7)
+++1313 (100)+

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

The ability to predict correctly the outcome of either a disease or a procedure is of paramount importance in all branches of medicine. It facilitates the balancing of the relative risks and benefits of available, alternative treatments by the physician, and allows individualised counselling.

Various parameters have been reported in the medical literature to be predictive of the outcome of external cephalic version (Table 1). However, most of these reports were limited by small sample size, with conflicting results. Univariate analysis predominated in these reports, but this does not exclude the possibility of confounding effect between the variables.

In this prospective study, we included most of the variables which have been reported in the past to be significant predictors of the outcome of external cephalic version, together with other variables, which from our own experience, were considered potential predictors. We have reported previously that the use of tocolytic agents significantly increases the success rate of external cephalic version14–15. Since then, we have used tocolytic agents in all external cephalic version procedures in our unit unless contraindicated; therefore we did not include the use of tocolytic agents as a variable in our prediction model.

We noticed that many of the variables tested are clinical characteristics which might have been associated with significant inter-observer variation. Therefore, we divided our study population into two subgroups by random allocation, and tested the regression model developed in one subgroup on the other subgroup. If the two models had differed significantly, it would have implied either the clinical assessments were not reproducible, or additional important predictors had not been included, leading to unreliable prediction modelling. We found that the significant variables in the two regression models were the same, and the models give reasonably accurate predictions when applied to data from the other subgroup. This finding confirms that the outcome of external cephalic version is predictable, and that the model derived is reliable and reproducible.

The results support our original hypothesis that there is significant interaction between many of the variables assessed. Although eight factors were found to be significantly associated with the outcome of external cephalic version, only three remained as independent variables when examined by regression model.

The ease with which the fetal head could be palpated clinically was found to be one of the most significant factors in determining the outcome of external cephalic version. This observation has not been described before. In fact this association is logical. It should be easier to apply pressure to the baby if both poles of the baby are easily identified; any attempt to turn a baby would be less effective with pressure applied to one pole only.

Engagement of the presenting part has been reported to be an important variable influencing the outcome of external cephalic version5,10, although this has been disputed by others3. Our results showed that engagement was the most significant independent predictor for failure of external cephalic version.

A tense uterus on palpation after a standard dose of tocolytic was also found to be significantly associated with failure of external cephalic version. We have shown that tocolytic agents, which induce uterine relaxation, increase the success rate of external cephalic version14–15. Our results suggest that some women may still have a tense uterus after the administration of a standard dose of tocolytic agent. A tense uterus limits the mobility of the baby and therefore decreases the chance of successful external cephalic version. It may be, therefore, worthwhile to consider increasing the dose of tocolytic agent, in the absence of significant side effects, in women with a tense uterus after a standard dose. This might improve the chance of success of external cephalic version.

Because most of our external cephalic versions were assisted by tocolysis, we thought that it was more appropriate to assess uterine tone after tocolytic administration. By doing so, however, this parameter was not available during the initial counselling for external cephalic version. Therefore we repeated the analysis without this variable. Parity, which is closely related to uterine tone and is the commonest reported predictor for successful external cephalic version, became the third independent predictor.

Various other clinical factors have been reported to be predictive of the outcome of external cephalic version, including placental site2,4,7,9, position of fetal spine3,5, attitude of the fetal lower limbs3,4 and maternal obesity2,5,8. None of these was found to be an important factor in our study when other clinical variables were controlled.

We conclude that the immediate outcome of external cephalic version can be predicted from clinical variables which are easily available. Assessment of the three clinical variables mentioned in Table 5 enable us to give more precise counselling. In the absence of any of these factors the chance of a failed external cephalic version was only 6%, compared with more than 80% when more than one of the factors are present. This estimation would certain facilitate decision-making for both obstetricians and women. On the other hand, those women consented for external cephalic version with a tense uterus, even after a standard dose of tocolytic, may benefit from a further dose of tocolytic.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References