102 Santoso.

Med J Indones

Budi Iman Santoso Assessment (BISA): a model for predicting levator ani
injury after vaginal delivery
Budi I. Santoso
Department of Obstetrics and Gynecology, Faculty of Medicine, Universitas Indonesia, Jakarta, Indonesia

Abstrak
Latar belakang: Belum ada usaha maupun penelitian yang mampu memadukan berbagai faktor risiko untuk memprediksi
terjadinya kerusakan otot levator ani akibat persalinan pervaginam. Penelitian ini bertujuan untuk mengetahui indeks yang
dapat digunakan untuk memprediksikan kerusakan levator ani pada persalinan pervaginam.
Metode: Penelitian kohort prospektif di dua rumah sakit di Jakarta tahun 2010-2011. Kriteria subjek adalah wanita hamil
nulipara tanpa kerusakan levator ani saat hamil dan melahirkan pervaginam. Kerusakan levator ani diukur dengan USG 4
dimensi saat hamil dan tiga bulan pasca melahirkan. Variabel yang diteliti adalah usia, indeks masa tubuh, cara persalinan
pervaginam, berat badan bayi lahir, episiotomi, robekan perineum, dan lamanya kala 2. Model prediksi dianalisis dengan
analisis regresi logistik.
Hasil: Sebanyak 182 subjek direkrut dengan 124 subjek memenuhi kriteria dan 104 subjek dapat dianalisis. Insiden
kerusakan levator ani pada tiga bulan adalah sebesar 15,4% (IK 95%: 8,6-23%). Diperoleh dua model prediksi. Model
prediksi pertama terdiri dari berat bayi (OR= 5,36 IK 95%: 1,08-26,59), episiotomi (OR= 5,41 IK 95%: 0,94-31,18), dan
lama kala dua (OR= 15,27 IK 95%: 3,15-73,96). Model prediksi kedua terdiri dari lama kala dua (OR= 9,51 IK 95%: 1,2368,10) dan robekan perineum (OR= 142,70 IK 95%: 14,13-1440,78).
Kesimpulan: Variabel yang dapat memprediksikan kerusakan levator ani adalah berat bayi, episiotomi, dan kala dua pada
model 1 dan lama kala dua serta robekan perineum pada model 2. (Med J Indones. 2012;21:102-7)

Abstract
Background: There have been no attempts or studies to integrate various risk factors that can be utilized to predict levator
ani injury caused by vaginal delivery. This study was aimed to establish an index measurement system by using various risk
factors for predicting levator ani injury in vaginal delivery.
Methods: A prospective cohort was conducted at two hospitals in Jakarta between 2010 and 2011. The subjects were
nulipara pregnant women without levator ani injury during pregnancy and vaginal birth. Levator ani injury was evaluated
using 4D USG during pregnancy and three months after delivery. The variables studied were age, body mass index, mode of
delivery, fetal birth weight, episiotomy, perineum rupture and duration of second stage labor. Prediction model was analyzed
using logistic regression analysis.
Results: There were 182 recruited subjects of which 124 subjects were eligible and only 104 subjects could be analyzed.
Incidence of levator ani injury at three months after delivery was 15.4% (95% CI: 8.6-23%). Two prediction models were
obtained. The first consisted of fetal birth weight (OR= 5.36, 95% CI: 1.08-26.59), episiotomy (OR= 5.41, 95% CI: 0.9431.18), and duration of second stage labor (OR= 15.27, 95% CI: 3.15-73.96). The second model consisted of duration of
second stage labor (OR= 9.51, 95% CI: 1.23-68.10) and perineum rupture (OR= 142.70, 95% CI: 14.13-1440.78).
Conclusion: Fetal birth weight, episiotomy and duration of second stage labor could predict levator ani injury for model 1;
while the variables of prediction for model 2 were duration of second stage labor and perineum rupture. (Med J Indones.
2012;21:102-7)
Keywords: Levator ani, prediction model

Levator ani muscle is one of important components of

pelvic floor structure. Levator ani damage may result
in impaired pelvic floor function, known as pelvic floor
dysfunction. It includes various symptoms that could
reduce quality of life such as urinary incontinence,
fecal incontinence, pelvic organ prolapse and sexual
dysfunction.1,2 Obstetricians and gynecologists have
assumed that the natural mode of delivery or vaginal
childbirth may contribute to pelvic floor dysfunction;
particularly the levator ani damage.3 Prevalence of
levator ani damage at 3-months after delivery as has
been revealed by Dietz4 is 15-30% in women with
vaginal delivery.
Correspondence email to: budi.imansantoso@gmail.com

However, until now, the clinical relevance of the

association between levator ani damage and developing
symptoms of pelvic floor dysfunction is still vague.
Dietz4 even demonstrated that there were many women
with levator ani damage who did not have symptoms
of pelvic floor dysfunction. Moreover, the concern on
pelvic floor dysfunction which leads to the selection
Caesarean birth seems to be over-worried, since the
Caesarian birth actually could only prevent 1 of 7
women from experiencing levator ani damage due to
vaginal birth.5 It should be noted that the mortality risk
of Caesarean birth increases up to five times compared
to vaginal birth. Moreover, there are various risk factors

BISA: predicting levator ani injury 103

Vol. 21, No. 2, May 2012

for levator ani injury, i.e. demographic and obstetric risk

factors.5 Demographic risk factors include maternal age,
race, parity and body mass index (BMI).1,6,7 Obstetric
risk factors include age at first delivery, mode of delivery,
second stage period, fetal birth weight, episiotomy, and
perineum rupture.3,4,6
Until now, there have been no attempts to integrate
various risk factors that can be utilized to predict levator
ani injury caused by vaginal delivery. Therefore, a scoring
system that could predict the occurrence of levator ani
injury is required. When the scoring system suggests
a low risk of levator ani injury, then the patient could
be convinced to choose vaginal birth without any fear
of the developing pelvic floor dysfunction. Therefore,
the aim of the present study was to establish an index
measurement system by using various risk factors in
predicting levator ani injury in vaginal delivery, which
were represented in a comprehensive and integrated
dynamic system model.
METHODS
The study was a prospective observational cohort study
using a quantitative approach, which was followed by
dynamic system assessment. This study was approved
by The Ethical Committee of Faculty of Medicine,
Universitas Indonesia, ref:200/PT02.FK/Etik/2010.
There were three steps of assessment, i.e. the descriptive
analysis, development of dynamic system model,
and model simulation. Samples were collected using
consecutive sampling at Cipto Mangunkusumo Hospital
and YPK (Yayasan Pemeliharaan Kesehatan) Hospital
between June 2010 and December 2011. Subjects were
nulipara women who had their vaginal birth at maternal
age of 37 weeks or more. The inclusion criteria were
nulipara women who had planned to have vaginal birth
with maternal age of 37 weeks or more, in healthy
condition and could have normal delivery. The exclusion
criteria were subjects with complications of pregnancy
such as antepartum bleeding, was not able to deliver
the baby at Cipto Mangunkusumo Hospital or YPK
Hospital, subjects with pregnancy and comorbidities,
such as preeclampsia, eclampsia, placenta previa and
solution placenta, and subjects who already had levator
ani avulsion prior to the delivery.
Data collection was performed through two steps, i.e.
recruitment of subjects and examinations for evaluating
levator ani injury. Nulipara women at maternal age of 37
weeks or more who were candidates for study subject will
be drawn from pregnant women who had their routine
medical visit at the Cipto Mangunkusumo Hospital or
YPK Hospital and who had completed their examination
according to the available protocol. The eligible

subjects got explanation about the study conducted, and

on their approval, they were asked to sign the informed
consent form. BMI during pregnancy was calculated
before pregnancy and classified according to maternal
weight gain at near delivery. BMI classification was
applied from table by Arisman.8 Antenatal examination,
abdominal and pelvic floor ultrasound were performed
as the initial examination to detect levator ani injury.
Afterwards, when the subjects had their delivery, various
parameter of levator ani risk factors were recorded (such
as fetal birth weight, the duration of second stage labor,
episiotomy). Moreover, the subjects were asked to have
another visit for ultrasound examination at 6 weeks and
3 months later to establish the diagnosis of levator ani
injury. All examinations were performed in blind method,
i.e. the investigator did not aware about the obstetric
data of the subjects. Data was processed using SPSS
software version 11.0 and two steps were conducted,
that were making description and explanation about the
observed process, including the calculation and analysis
of risk factors, which would be utilized for the model
development.
RESULTS
Between June 2010 and December 2011, there were
182 subjects who wanted to participate in the study.
About 53 subjects were excluded due to abdominal
delivery (29.1%), twenty subjects left the study at the
6-weeks avulsion examination (11.0%), and 5 subjects
were excluded due to obvious levator ani avulsion on
the antenatal examination (2.7%). Characteristics of
subjects based on demographic factors are given in
table 1 and characteristics of subjects based on obstetric
factors could be seen in table 2.
Of 182 patients enrolled in the study, there were only
104 patients who were eligible for the analysis. There
was no significant difference in age and BMI during
Table 1. Subject characteristics based on demographic factors
Variables
Age (years)

Education level
BMI during pregnancy

Category

20 23

25

24.03

24 27

41

39.42

28 31

28

26.92

32 35

7.69

36 max

1.92

High school University

79

76.0

Elementary Junior High

25

24.0

Low

48

46.2

Normal

41

39.4

High

11

10.6

Obesity

3.8

104 Santoso.

Med J Indones

Table 2. Subject characteristics based on obstetric factors

Variables

Category

Mode of delivery

Vacuum
Spontaneous

17
87

16.3
83.7

Perineum tear

Grade IV
Grade III
Grade II

1
18
81

1.0
17.3
77.9

Episiotomy
Duration of second stage labor
Fetal birth weight

Grade I

3.8

Yes
No
65 minutes
< 65 minutes

40
64
26
78

38.5
61.5
25.0
75.0

3325 g

23

22.1

< 3325 g

81

77.9

pregnancy between data that could be analyzed (104

cases) with could not be analyzed due to loss to follow
up (20 cases) and the statistic results showing similar
subject characteristics among the patients, i.e. age (p =
0.448), education (p = 0.687) and pregnancy BMI (p =
0.791). Most subjects (n = 41, 39.42%) were at 24-27
years of age and the education level was high school
to university in 79 (76%) subjects. Most subjects had
BMI as low, found in 48 subjects (46.2%). Incidence
of avulsion at 6 weeks and 3 months were 11.5%
(95% CI: 5.7-18.5%) and 15.4% (95% CI: 8.6-23.0%)
respectively. There was one subject who demonstrated
avulsion at the six weeks had but become normal at
the three months period. Moreover, there were four
subjects who had no avulsion at the six weeks but had
avulsion at three months.
On bivariate analysis, perineum rupture (p < 0.001),
episiotomy (p < 0.001), duration of second stage
labor (p < 0.001), and fetal birth weight (p = 0.003)
demonstrated significant association with avulsion at
three months period after delivery. Meanwhile, BMI
(p = 0.144) and mode of delivery (p = 0.208) were not
associated with avulsion at three months period after
delivery.
The numerical variables including the fetal birth weight
and long duration of second stage periods appeared to
have significant correlation with levator ani injury or
avulsion at the third month. By using receiver operating

characteristic (ROC) curve, an optimal cut-off point

was found of 3325 g for the fetal birth weight and
65 minutes for the duration of second stage labor.
Independent variables that fulfilled the criteria for
multivariate analysis were BMI, mode of delivery,
perineum rupture, episiotomy, duration of second
stage labor and fetal birth weight. There were
autocorrelation between perineum rupture and
episiotomy. Therefore, this study developed two
models, i.e. the first model without including perineum
rupture (model 1) and the second model without
including episiotomy (model 2). On multivariate
analysis, stepwise logistic regression was performed
with independent variables of BMI, mode of delivery,
duration of second stage labor and fetal birth weight.
On the third step, the variable of episiotomy was found
statistically not significant (p = 0.059); however, it was
clinically significant since the odds ratio of episiotomy
(OR = 5.41) was greater than the minimal odds ratio,
which was considered as significant (OR = 3.0). Other
variables (duration of second stage labor and fetal birth
weight) were significant both statistically and clinically.
The third step was decided as the final model for the
first model (model 1), which was further transformed
into a score system.
From the logistic regression analysis in model 1, the
result showed that
Y (avulsion at the third month) = -4.64 + (1.69)*episiotomy + (2.73)*second
stage labor + (1.68)*fetal birth weight

According to the model, the log of the odds of pregnant

woman got avulsion at the third month was positively
related with episiotomy, second stage labor and fetal
birth weight (Table 3).
Afterward, the probability based on fetal birth weight,
episiotomy and second stage labor were classified into
3 categories, i.e. low, moderate and high probabilities.
The probability was considered low when it reached
0.96 percentages, 4.92 percentages, 4.96 percentages,
12.85 percentages and 21.87 percentages. It was
considered moderate at 44.15 percentages and 44.37
percentages and the probability was considered high at
81.05 percentages (Table 4).

Table 3. Logistic regression analysis for model 1

Predictor
Constant
Episiotomy

-4.64
1.69

SE
0.9
0.89

Wald
26.66
3.57

df
1
1

p
< 0.001
0.059

OR
0.01
5.41

Second stage labor

2.73

0.81

11.46

0.001

15.27

Fetal weight

1.68

0.82

4.22

0.040

5.36

Hosmer & Lemeshow p = 0.32

BISA: predicting levator ani injury 105

Vol. 21, No. 2, May 2012

Table 4. Probability of avulsion based on logistic regression equation in model 1

Variable

Fetal weight 3325 g

Yes

No

Case1
Case 2

Case 4

Case 5

Case 6

No

Yes

0.96

4.92

4.96

12.85

21.87

44.15

44.37

81.05

Probability
(%)

No

Case 7
Case 8

Yes

Case 3

2nd stage
65 minutes

Episiotomy

Table 5. Logistic regression analysis for model 2

Predictor

SE

Wald

df

OR

Constant

-5.22

1.19

19.29

< 0.001

0.01

Second stage labor

2.21

1.02

4.67

0.031

9.51

Perineum rupture

4.96

1.18

17.68

< 0.001

142.70

Hosmer & Lemeshow p = 0.653

Table 6. Probability of avulsion based on logistic regression equation on model 2
Variable

2nd stage 65 minutes

Yes

Case 1
Case 2

Yes

Case 3
Case 4

No

Perineum Rupture

In the second model without including episiotomy

(model 2), the stepwise multivariate analysis of
regression logistic was presented with independent
variables of BMI, mode of delivery, perineum rupture,
second stage labor and fetal birth weight. Based on
clinical consideration, which is in keeping with the
research proposal, i.e. the minimal odds ratio considered
as significant was three; therefore, the fourth step was
considered as the final model. The result for model 2
showed that
Y (avulsion at the third month) = -5.22 + (2.21)* second stage labor
+ (4.96)*perineum rupture

From model 2, we found that if woman had longer second

stage labor and perineum rupture, the more likely it is
that a woman got avulsion at the third month (Table 5).
Subsequently, the probability based on second stage
labor and perineum rupture was classified into 3

No

Probability
(%)

0.54

4.72

43.56

87.60

categories, i.e. low, moderate and high probabilities.

The probability was low when it reached the probability
of 0.54 percentages and 4.72 percentages. Moderate
probability was considered when the probability was
43.56 percentages and it was high probability when the
probability was 87.60 percentages (Table 6).
DISCUSSION
The study demonstrated that of 182 subjects who were
willing to participate in the study, only 104 (57.12%)
subjects were eligible for the analysis. The reasons were
having abdominal delivery in 53 subjects (29.1%), left
the study in twenty subjects (11.0%) and detected levator
ani avulsion during antenatal examination in 5 subjects
(2.7%). Similar results were also found by Chan et al9
who studied about the prevalence of levator ani injury in
primipara women in China. Of 339 subjects, there were
only 201 (59.3%) subjects who could be analyzed, about

106 Santoso.
62 (18.3%) subjects had assisted surgical vaginal delivery
by vacuum and forceps, fourteen (4.1%) subjects had
elective abdominal delivery and 62 (18.3%) subjects had
emergency abdominal delivery.
The subject characteristics revealed that the major
age range was at 24-27 years, which was found in 41
(39.42%) subjects. While the subject age in a study
conducted by Chan et al9 was 30.6 (+ 3.9) years. There
was no significant difference regarding the subject
characteristic between the subjects that could be
analyzed and the drop-out subject. The present study
indicated that most subjects had low BMI in pregnancy.
There were only 11 (10.6%) subjects with high BMI
and only 4 (3.8%) subjects with obesity.
There was one subject who experienced avulsion on the
6-weeks examination but demonstrated normal result at
the 3-months examination. It may occur since the pelvic
floor innervations had been restored. Furthermore, there
were 4 subjects who had no avulsion at the 6-week
examination but developed avulsion at the 3-months
examination. Such case probably occurs due to
persistent damage of pelvic floor nerves. Our findings
are consistent with the results reported by Snooks et
al10 as well as by Dietz and Lanzarone11 that one third of
women who had vaginal delivery would developed avulsion
of the fascia layer that supported the pelvic floor muscles
within 3 months after the delivery.
The bivariate analysis in this study demonstrated a
significant correlation between perineum rupture and
the occurrence of levator ani injury with OR of 235.20
(95% CI: 25.54 - 2166.28). Episiotomy also had effect
on levator ani injury with OR of 14.93 (95% CI:
3.15 - 70.73). Moreover, obesity also had a clinically
significant correlation to levator ani injury with OR of
2.58 (95% CI: 0.70 - 9.53). Other investigators have also
reported similar results. Dietz4 suggests that episiotomy
is biomechanical risk factors in the development
of levator ani injury and it is not a protective factor.
Moalli et al7 demonstrates that episiotomy and vaginal
laceration/ perineum rupture are the risk factors of
levator ani injury at the first delivery.
Carroli and Belizan12 in Cochrane Review 2009
indicates that episiotomy on indication has
significantly involved less trauma to pelvic floor
compared to routine episiotomy. Dietz et al13 found
that women with lesser BMI had greater risk for
levator ani injury. Considering that most Indonesian
people have low socio-economic and education level,
which may result in low BMI, it could be assumed
that Indonesian people are likely to carry high risk for
levator ani injury.

Med J Indones

In our study, we found that fetal birth weight of 3325

g and duration of second stage labor of 65 minutes
appeared to have significant correlation with levator injury
or avulsion at the third month. Kearney et al14 showed that
the duration of second stage labor of 78 minutes was the
risk factor for levator ani injury. Lavy et al15 indicated that
there was a positive correlation between high fetal birth
weight and levator ani injury after delivery. However, there
has been no data about the exact fetal birth weight that may
cause levator ani injury. Most experts suggest that fetal
birth weight of 4.5 kg may cause levator ani injury, while
our study showed that the fetal birth weight of 3325 g
may cause the injury.

In the multivariate analysis, all variables of bivariate
analysis were included in the multivariate analysis since
all had p < 0.25, i.e the BMI during pregnancy, mode
of delivery, perineum rupture, episiotomy, duration of
second stage labor and fetal birth weight. After several
steps, we found a correlation between perineum rupture
and episiotomy.
In model 1, after performing several multivariate
analysis on levator ani injury in 3 months period, we
found that episiotomy and duration of second stage
labor longer than 65 minutes and fetal birth weight over
3325 g could be utilized for the scoring system. After
obtaining the regression equation, the subject probability
was subsequently calculated by certain score. Following
the calculation of subject probability with certain score
that could determine the prognosis of levator ani injury,
a scoring card was developed, which could be utilized
for daily practice. Model 1 could be applied at all health
care level since almost every health personnel are able to
evaluate second stage labor and episiotomy; while model
2 could only be applied if the medical personnel are able
to evaluate perineum rupture. Therefore, trainings about
assessment of perineum rupture after delivery for medical
personnel are essentially needed. The study limitation
includes the lesser sample of the expected. However, it
did not affect the final result of our study considering
that the odds ratio for almost all variables were > 3.
An index measurement system model has been developed
to represent the role of demographic risk factor (body
mass index) and obstetric risk factors (maternal age
at delivery, mode of assisted vaginal delivery such as
vacuum extraction, duration of second stage, fetal birth,
episiotomy and perineum rupture) and the association
with levator ani injury at vaginal delivery. However,
application of the scoring system in daily practice still
requires some case examples so that we could validate
the scoring system with the true fact. Further studies
with larger sample size are extremely needed before we
could determine any policy.

BISA: predicting levator ani injury 107

Vol. 21, No. 2, May 2012

Modified and applied policy to decrease the

incidence of levator ani injury should be determined
as policy of preventive strategies, which include
the policy about pelvic floor exercise, preventing
perineum damage, providing training on grade III
perineum tear for general physicians and midwives,
establishing guidelines on performing appropriate
episiotomy according to the indication or even
policy on the elective Caesarian delivery, which is
still controversial. Current data and further studies
should be considered in those policies. Further studies
that support the preventive strategies policies are
suggested to be performed on all sectors involving
government institutions such as Department of
Health, Department of National Education in
collaboration with the Collegiums of Obstetrics and
Gynecology and Indonesian Midwives Association.
REFERENCES
1.

MacDorman M, Menacker F, Declercq E. Cesarean birth

in the United States: epidemiology, trends, and outcomes.
Clin Perinatol. 2008;35:293-307.
2. American College of Obstetricians and Gynecologists.
ACOG Committee Opinion Number 394, December 2007.
Cesarean delivery on maternal request. Obstet Gynecol.
2007;110:1501-4.
3. Dietz HP. Levator trauma in labor: a challenge for
obstetricians, surgeons and sonologists. Ultrasound Obstet
Gynecol. 2007;29:368-71.

4.
5.
6.
7.
8.
9.

10.
11.
12.
13.
14.
15.

Dietz HP. Quantification of major morphological

abnormalities of the levator ani. Ultrasound Obstet Gynecol.
2007;29:329-34.
Lukacz ES, Lawrence JM, Contreras R, Nager CW, Luber
KM. Parity, mode of delivery, and pelvic floor disorders.
Obstet Gynecol. 2006;107:1253-60.
Sze EHM, Sherrad GB, Dolezal JM. Pregnancy, labor, delivery
and pelvic organ prolapse. Obstet Gynecol. 2002;100:981-6.
Moalli PA, Ivy SJ, Meyn LA, Zyczynski HM. Risk factors
associated with pelvic floor disorders in women undergoing
surgical repair. Obstet Gynecol. 2003;101:869-74.
Arisman MB. Gizi dalam daur kehidupan: buku ajar ilmu
gizi. Jakarta: EGC; 2004. Indonesian.
Chan SSC, Cehun RY, Yiu AK, Lee LL, Pang AW, Choy
K, et al. Prevalence of levator ani muscle injury in Chinese
primiparous women after first delivery. Ultrasound Obstet
Gynecol. In press Jan 2012.
Snooks SJ, Swash M. The innervation of the muscles of
continence. Ann R Coll Surg Engl. 1986;68:45-9.
Dietz HP, Lanzarone V. Levator trauma after vaginal
delivery. Obstet Gynecol. 2005;106(4):707-12.
Carroli G, Belizan J. Episiotomy for vaginal birth. Cochrane
Db Syst Rev. 2000;(2):CD000081.
Dietz HP, Shek C, Clarke B. Biometry of the pubovisceral
muscle and levator hiatus by 3D pelvic floor ultrasound.
Ultrasound Obstet Gynecol. 2005;25:580-5.
Kearney R, Miller JM, Ashton-Miller JA, DeLancey JO.
Obstetric factors associated with levator ani muscle injury
after vaginal birth. Obstet Gynecol. 2006;107:144-9.
Lavy Y, Sand PK, Kaniel CI, Celnikier DH. Can pelvic floor
injury secondary to delivery be prevented? Int Urogynecol
J. In press Aug 2011.