Vaginal Birth After Cesarean Delivery:

An Admission Scoring System

BRUCE L. FLAMM, MD, AND ANN M. GEIGER, PhD
Objective: To develop a scoring system to predict the likelihood of vaginal birth in patients undergoing a trial of labor
after previous cesarean delivery using factors known at the
time of hospital admission.
Methods: Trial of labor was attempted in 5022 patients
who were assigned randomly to score derivation and score
testing groups. Multivariate logistic regression modeling
was used in the score derivation group to develop a predictive scoring system for vaginal birth. The scoring system was
then applied to the testing group to evaluate its predictive
ability.
Results: Five variables significantly affected the mode of
birth and were incorporated into a weighted scoring system.
Rates of successful vaginal birth after cesarean ranged from
49% in patients scoring 0 2 to 95% in patients scoring 8 10.
Increasing score was associated linearly with increasing
probability of vaginal birth after cesarean.
Conclusion: Increasing scores correlate with increasing
probability of vaginal birth after cesarean. The admission
vaginal birth after cesarean scoring system may be useful in
counseling patients regarding the option of vaginal birth or
repeat cesarean delivery. This information could be particularly valuable for the patient who opts for trial of labor but
has second thoughts about her mode of birth when labor
begins. (Obstet Gynecol 1997;90:90710. 1997 by The
American College of Obstetricians and Gynecologists.)

For more than a decade, cesarean delivery has been the

most frequently performed major operation in the
United States. Cesarean rates climbed sharply during
the 1980s, and almost half of this increase was due to
repeat cesarean deliveries.1 Large multicenter studies2,3
have demonstrated clearly that vaginal birth after cesarean (VBAC) is a safe option. However, it has become
increasingly apparent that complications are more
likely to occur in those women who have an unsuccessful trial of labor.4 Therefore, we used our ten-hospital
From the Departments of Obstetrics and Gynecology and Research
and Evaluation, Kaiser Permanente Medical Centers, Southern California Region, Riverside, California.
Supported in part by a grant from the Kaiser Foundation Research
Institute (Garfield Memorial Fund grant no. 101-9042).

VOL. 90, NO. 6, DECEMBER 1997

trial of labor database in an attempt to develop a scoring

system to predict which patients would be more likely
to deliver vaginally.

Materials and Methods

Between January 1990 and December 1992, all pregnant
women with previous cesareans were studied prospectively at all ten Southern California Kaiser Permanente
Hospitals (Anaheim, Bellflower, Fontana, Harbor City,
Los Angeles, Panorama City, Riverside, San Diego,
West Los Angeles, and Woodland Hills.) A trained,
full-time research associate rotated continuously among
the participating hospitals to supervise data collection
and computer data entry. Each patients chart was
reviewed fully. Details of the study methodology have
been published previously.5 Five thousand twenty-two
of 7229 patients with previous cesarean deliveries (69%)
had trials of labor. Details of the outcomes of these 5022
trials of labor also have been published.5 The computerized database for these patients was used to develop
a predictive scoring system. Seventeen women were
excluded because they opted for repeat cesarean on
admission, and two were excluded due to incomplete
data; hence, the final sample size was 5003.
The study database included information about a
number of possible predictors of successful trial of labor
that are available at the time of hospital admission for
delivery. Historic factors included maternal age, obstetric history, and reason for first cesarean. Prenatal factors
included weight gain, initial 1-hour postglucose screen,
and confirmed gestational diabetes. Intrapartum factors
included cervical dilation, cervical effacement, and status of membranes (all at the time of admission). Age
was used initially in the analysis as a continuous
variable but was reclassified as a categoric variable
when a threshold effect was observed. Cervical dilation
and effacement and 1-hour post glucose screen were not
normally distributed, so they were classified into categories based on clinically significant cut-points.

0029-7844/97/$17.00
PII S0029-7844(97)00531-0

907

For the analysis, the 5003 women who attempted

trials of labor were each assigned a computer-generated
random number. The database was then sorted in
ascending order by random number. The first 2502
women on the list were assigned to the score development group and the last 2501 women were assigned to
the score testing group. Data from the score development group was used in a multivariate logistic regression model to develop a predictive scoring system. The
relationship between trial of labor outcome and possible predictors in the score development group was first
evaluated using x2 for categoric variables and Student t
test for continuous variables. Predictors significant at
P , .05 were then entered into one of three logistic
regression models, one each for historic, intrapartum,
and perinatal factors. Predictors significant at P , .05 in
any of the three models were entered into the final
logistic regression model. Missing values for one or
more factors led to the exclusion of 88 women in the
final model. No statistically significant (P , .05) interactions were found in either the initial or final models.
In the score development group, 38% of patients were
white, 37% Hispanic, 15% black, and 10% from other
ethnic groups. In the score testing group, 38% of patients were white, 38% Hispanic, 14% black, and 10%
from other ethnic groups. There was no statistically
significant difference between the score developing and
score testing groups with regard to race or ethnicity.
The final logistic regression model was used to develop the score. Points were assigned to each of the
predictors in the model based on the b coefficient. The
higher the score, the more likely a successful trial of
labor. The highest possible score was 10. Scores of 2 or
less or 8 or more were combined due to small numbers.
The performance of the score was assessed in the score
testing group. The score for each woman attempting
trial of labor was determined by assigning points for
each factor present in the woman, then adding to obtain
a total. Missing values for one or more factors led to the
exclusion of 96 women from the examination of score
performance.

Results
The score development and score testing groups exhibited similar characteristics, indicating that the randomization process was successful (Table 1). Multivariate
logistic regression modeling was used in the score
development group to evaluate variables for inclusion
in the VBAC scoring system. Women with successful
vaginal births after trial of labor were younger and
more likely to have had a prior vaginal delivery. At
admission they were also more likely to have cervical
effacement greater than 75% and cervical dilation of

908 Flamm and Geiger

VBAC Score

Table 1. Characteristics of Score Development and Testing

Groups by Mode of Birth
Score development
group (n 5 2502)

Score testing group

(n 5 2501)

Failed
Failed
TOL
TOL
VBAC
VBAC
(n 5 1863) (n 5 639) (n 5 1883) (n 5 618)
Characteristic
Age under 40
Vaginal birth history
Before and after first
cesarean
After first cesarean
Before first cesarean
None
Reason other than FTP
for first cesarean
delivery
Cervical effacement at
admission
.75%
2575%
,25%
Cervical dilation 4 cm
or more at
admission

1820 97.7 607 95.0 1847 98.1 598 96.8

63
293
262
1236
1081

3.4

0.5

68

3.6

1.1

15.8 30 4.7 315 16.8 27 4.4

14.1 71 11.1 256 13.7 70 11.4
66.7 534 83.7 1236 65.9 512 83.1
59.8 266 42.2 1076 58.9 266 43.5

1224 66.5 288 45.8 1260 68.4 296 49.2

493 26.8 228 36.2 463 25.1 191 31.7
123 6.7 113 18.0 119 6.5 115 19.1
710 38.4 102 16.1 705 37.8 115 18.9

VBAC 5 vaginal birth after cesarean; TOL 5 trial of labor; FTP 5

failure to progress.
Denominator term used to calculate percentages varies across variables due to missing information.

4 cm or more. Failure to progress was a more common

reason for the first cesarean among women who experienced failed trials of labor. The number of prior
cesareans, weight gain during pregnancy, initial 1-hour
postglucose screen, the presence of confirmed gestational diabetes, and the status of membranes at admission were not significantly associated with trial of labor
success or failure. The five significant variables were
included in the final logistic regression model as predictors of a successful trial of labor (Table 2). History of
vaginal birth both before and after the first cesarean was
so highly correlated with subsequent successful VBAC
that it accounted for four points in the final score.
However, such a history was present in less than 3% of
all patients and thus less than 3% of patients could
potentially achieve a score of 9 or 10. Moreover, approximately 70% of patients had no history of vaginal
delivery and hence could not achieve a total score
greater than 6 regardless of other factors.
Rates of successful VBAC ranged from 49% in patients scoring 0 2 to 95% in patients scoring 8 10 (Table
3). Increasing score was associated in a linear fashion
with increasing probability of VBAC. In the score testing group, 82% of patients had scores of 4 or greater,
and at least two-thirds of these patients delivered
vaginally.

Obstetrics & Gynecology

Table 2. Admission Characteristics and Assigned Score

Points Predicting Successful Vaginal Birth After
Cesarean (Score Development Group)
Characteristic
Age under 40
Vaginal birth history
Before and after
first cesarean
After first cesarean
Before first cesarean
None
Reason other than
FTP for first
cesarean
Cervical effacement at
admission
.75%
25%75%
,25%
Cervical dilation 4 cm
or more at
admission

b
coefficient

Odds
ratio

95% CI

Score
points*

0.95

2.58

1.55, 4.3

2.21

9.11

2.18, 38.04

1.22
0.43
referent

3.39
1.53

2.25, 5.11
1.12, 2.10

2
1
0

0.66

1.93

1.58, 2.35

1.00
0.58
referent

2.72
1.79

2.00, 3.71
1.31, 2.44

2
1
0

0.77

2.16

1.66, 2.82

CI 5 confidence interval; FTP 5 failure to progress.

* Select one value, which may be zero, from each of the five
categories.

Discussion
Over the past generation the 1916 dictum, once a
cesarean, always a cesarean, has been laid to rest and
the option of VBAC has been accepted widely. Large
multicenter studies23 have demonstrated that trial of
labor after previous cesarean is a reasonable option for
many women. However, these same studies indicate
that at least 25% of women who attempt VBAC will
have an unsuccessful trial of labor. This is a source of
great emotional distress for some women who think
that they have in some way failed. Furthermore, patients who fail a trial of labor appear to be at greater
risk for complications than women who have a successTable 3. Performance of Admission Score
Score development group
Score
0 to 2
3
4
5
6
7
8 to 10
Total

Score testing group

Number
%
%
Number
%
%
with score VBAC Cesarean with score VBAC Cesarean
120
346
605
664
354
183
142
2414

41.7
59.2
64.3
79.1
87.6
93.4
99.3
74.2

58.3
40.8
35.7
20.9
12.4
6.6
0.7
25.8

VBAC 5 vaginal birth after cesarean.

VOL. 90, NO. 6, DECEMBER 1997

114
329
595
660
360
189
158
2405

49.1
59.9
66.7
77.0
88.6
92.6
94.9
74.9

50.9
40.1
33.3
23.0
11.4
7.4
5.1
25.1

ful VBAC or women who choose elective repeat cesarean.4 Clearly, a scoring system that could predict which
women would be more likely to have successful vaginal
births would be useful.
Several groups have attempted to develop VBAC
scoring systems. Troyer and Parisi6 analyzed 264 labors
in patients with prior cesareans and identified four
groups with VBAC rates ranging from 46% to 92%.
Consistent with our findings, they noted that both
previous dysfunctional labor and lack of prior vaginal
birth were associated with decreasing probability of
VBAC. Learman and associates7 studied admission
characteristics in 175 consecutive patients who underwent trial of labor and found no combination of variables capable of correctly classifying which patients
would require repeat cesarean delivery. Because 85% of
their study group delivered vaginally they concluded
that all eligible patients should be encouraged to undergo trial of labor. They did point out that the small
sample size in some of their subgroups limited their
conclusions and suggested that a larger study was
needed. Our sample size of more than 5000 patients
allowed us to evaluate many parameters and to split
our sample into score-development and score-testing
groups, each containing approximately 2500 patients.
It is important to point out that very few patients will
achieve high scores using this scoring system. For
example, only 6% of patients achieved a score of 8 or
more. It must also be emphasized that even in the group
with the lowest scores (0 2), about half of the patients
delivered vaginally. Some patients may view a 49%
chance of vaginal birth as quite favorable whereas
others might find it untenable. In either case, the
information would be helpful in their decision-making
process.
As in several other types of scoring systems, points
are assigned from five categories, and the maximum
possible score is 10. However, in contrast to obstetric
scoring systems used for other reasons, scores in the
middle range are favorable and high scores are rarely
achieved. For example, the modified biophysical profile
has a maximum score of ten and this score is frequently
achieved. A score of 4 on a biophysical profile would be
ominous, whereas a score of 4 on the admission VBAC
scale would be favorable, implying a 67% probability of
vaginal birth.
An important point is that this scoring system was
developed and tested using variables present at the
time of hospital admission for labor. Application of this
scoring system before the onset of labor would not be
valid. Cervical dilation and effacement often change
dramatically between the last prenatal examination and
the time of admission.
A study of this type has limitations. It may not be

Flamm and Geiger

VBAC Score

909

possible to generalize the study findings to all patient

populations and obstetric practices. Clearly, nonclinical
factors influence the decision to perform a cesarean
delivery.8,9 This study took place in ten hospitals operated by a large health maintenance organization and the
majority of patients were working-class, middle income
people. The majority of obstetricians at these hospitals
share a similar philosophy about childbirth and have
been familiar with the option of VBAC for many years.
Furthermore, certified nurse-midwives practice at all of
these medical centers and attended many of the births.
Patients with similar scores could have lower probabilities of vaginal birth in other settings.
Patients with prior cesarean for failure to progress
or dysfunctional labor represent a broad spectrum of
diagnoses ranging from initial cesarean in latent phase
with a closed cervix to initial cesarean at advanced
dilation with no progress despite many hours of welldocumented adequate labor. Because the goal of this
study was to develop a predictive score using only
parameters readily available at the time of hospital
admission, we did not delve into the vagaries of prior
dysfunctional labors. We are now conducting a multicenter study to investigate whether variables in these
labors are predictive of outcome of subsequent VBAC
attempts. This may be helpful, in combination with the
admission VBAC score, in counseling patients when
complete hospital records from the primary cesarean
delivery are available.
As the 20th century comes to a close, management of
the patient with a prior cesarean delivery remains
controversial.10 In particular, the patient who arrives at
the hospital for a planned trial of labor but then
vacillates in her decision presents a dilemma for the
attending physician. It is not unusual for a patient to
arrive in early labor excited about the prospect of
vaginal birth and have second thoughts as her contractions intensify. How should such patients be counseled?
There is no simple answer, but the admission VBAC
scoring system will provide information that may be
helpful to the physician and the patient. The scoring
system would be more useful if patients with a low
score had a dismal outlook for successful trial of labor.
With the ideal VBAC scoring system, patients with low
scores would all have a failed trial of labor and patients
with high scores would all have a successful VBAC.
Although scoring systems are probably never ideal they
can still be useful. For example, we used this scoring
system to identify more than 100 patients who had less

910 Flamm and Geiger

VBAC Score

than a 50% chance of VBAC. Some patients would not

want to continue with a planned trial of labor given
these odds. We also identified hundreds of women who
had at least a 90% chance of VBAC. This information
could be reassuring for a patient who becomes discouraged during a planned trial of labor.

References
1. Taffel S, Placek P, Liss T. Trends in the United States cesarean
section rate and reasons for the 1980 1985 rise. Am J Public Health
1987;77:9559.
2. Flamm BL, Newman LA, Thomas SJ, Fallon D, Yoshida MM.
Vaginal birth after cesarean delivery: Results of a 5-year multicenter collaborative study. Obstet Gynecol 1990;76:750 4.
3. Phelan JP, Clark SL, Diaz F, Paul RH. Vaginal birth after cesarean.
Am J Obstet Gynecol 1987;157:1510 5.
4. McMahon MJ, Luther ER, Bowes WA, Olshan AF. Comparison of
trial of labor with an elective second cesarean section. N Engl
J Med 1996;335:689 95.
5. Flamm BL, Goings JR, Yunbao L, Wolde-Tsadik G. Elective repeat
cesarean delivery versus trial of labor: A prospective multicenter
study. Obstet Gynecol 1994;83:92732.
6. Troyer LR, Parisi VM. Obstetric parameters affecting success in a
trial of labor: Designation of a scoring system. Obstet Gynecol
1992;167:1099 104.
7. Learman LA, Evertson LR, Shiboski S. Predictors of repeat cesarean delivery after trial of labor: Do any exist? J Am Coll Surg
1996;182:257 62.
8. Anderson GM, Lomas J. Explaining variations in cesarean section
rates: Patients, facilities or policies? Can Med Assoc J 1985;132:
253 6.
9. Stafford RS. The impact of nonclinical factors on repeat cesarean
section. JAMA 1991;265:59 63.
10. Flamm BL. Once a cesarean, always a controversy. Obstet Gynecol
1997;90:3125.

Address reprint requests to:

Bruce L. Flamm, MD
Department of Obstetrics and Gynecology and Research and
Evaluation
Kaiser Permanente Medical Centers, Southern California
Regions
10800 Magnolia Avenue
Riverside, CA 92505
E-mail: bruce.flamm@kp.org

Received June 18, 1997.

Received in revised form August 25, 1997.
Accepted September 4, 1997.
Copyright 1997 by The American College of Obstetricians and
Gynecologists. Published by Elsevier Science Inc.

Obstetrics & Gynecology