Risk factors for perineal injury during delivery

L. M. Christianson, MD, V. E. Bovbjerg, PhD, E. C. McDavitt, RN, and K. L. Hullfish, MD

Charlottesville, Va

OBJECTIVE: We sought to identify risk factors for anal sphincter injury during vaginal delivery.
STUDY DESIGN: This was a retrospective, case-control study. We reviewed 2078 records of vaginal deliver-
ies within a 2-year period from May 1, 1999, through April 30, 2001. Cases (n = 91) during the study period
were defined as parturients who had documentation of greater than a second-degree perineal injury. Control
subjects (n = 176), who were identified with the use of a blinded protocol, included women who were deliv-
ered vaginally with less than or equal to a second-degree perineal injury. For each patient, we reviewed med-
ical and obstetrics records for the following characteristics: maternal age, race, weight, gestational age,
parity, tobacco use, duration of first and second stages of labor, use of oxytocin, use of forceps or vacuum,
infant birth weight, epidural use, and episiotomy use.
RESULTS: Of the 2078 deliveries that were reviewed, we discovered 91 cases (4.4%) of documented anal
sphincter injury. The mean maternal age of our sample was 24.9 ± 5.9 years). Nearly two thirds (63.2%) were
white; 26.7% were black, and 10.1% were of other racial backgrounds. Forceps were used in 51.6% of deliv-
eries that resulted in tears (cases), compared to 8.6% of deliveries without significant tears (control subjects,
P < .05). Using cases and control subjects with complete data (cases, 82; control subjects, 144), delivery
with forceps was associated with a 10-fold increased risk of perineal injury (odds ratio, 10.8; 95% CI, 5.2-
22.3) compared to noninstrumented deliveries. The association was similar after adjustment for age, race,
parity, mode of delivery, tobacco use, episiotomy, duration of labor (stages 1 and 2), infant birth weight,
epidural, and oxytocin use (odds ratio, 11.9; 95% CI, 4.7-30.4). Nulliparous women were at increased risk for
tears (adjusted odds ratio, 10.0; 95% CI, 3.0-33.3) compared with multiparous patients, but parity did not re-
duce the association between forceps-assisted deliveries and anal sphincter injuries. Increasing fetal weight
was also a risk factor in both unadjusted and adjusted analyses. The performance of a midline episiotomy
was associated with an increased risk of anal sphincter tear compared with delivery without an episiotomy in
the univariate analysis (odds ratio, 4.9; 95% CI, 2.5-9.6), but this association was reduced in the adjusted
analysis (odds ratio, 2.5; 95% CI, 1.0-6.0). The increased duration of both the first and second stages of
labor increased injury risk in the unadjusted, but not adjusted, analysis. No significant association was ob-
served between case status and the use of oxytocin or epidural anesthesia. Greater, but not significant, in-
creased risk was associated with maternal indications for operative delivery compared with fetal indications.
CONCLUSION: Our results are consistent with recent reports that identify forceps delivery and nulliparity as
risk factors for recognized anal sphincter injury at the time of vaginal delivery. Further investigation should
focus on the determination of whether the association of injury to instrumentation is causal or, in fact, modifi-
able. Because of the established association between sphincteric muscular damage and anal incontinence,
patients should be counseled about the risk of anal sphincter injury when operative vaginal delivery is con-
templated. Such patients should be followed closely in the postpartum setting to assess for the development
of potential anorectal complaints. (Am J Obstet Gynecol 2003;189:255-60.)

Key words: Anal sphincter injury, perineal injury, laceration, incontinence

Damage to the perineum during vaginal delivery and
its prevention have been the subject of frequent specula-
tion regarding delivery technique over the years, with lit-
tle conclusion and much controversy. It appears that
From the Departments of Obstetrics/Gynecology and Health Evaluation
Sciences, University of Virginia.
Presented at the Twenty-Third Annual Meeting of the American UroGy-
necologic Society, San Francisco, California, October 6-18, 2003.
Reprint requests: Lisa M. Christianson MD, 302 Spring St, Char-
lottesville, VA 22903.
© 2003, Mosby, Inc. All rights reserved.
0002-9378/2003 $30.00 + 0
doi:10.1067/mob.2003.547
damage to the anal sphincter complex is the main con-
tributing factor to anal incontinence and urgency symp-
toms1-3 and that this damage can occur either visibly
(third- and fourth-degree lacerations) or as an occult in-
jury.3,4 After the primary repair of third- or fourth-degree
lacerations, ≤85% of women have persistent sphincter de-
fects,1 and 10% to 50% of women with sphincter injuries
have anorectal complaints1-3,5-6 ≤2 years after delivery,7
including fecal and flatus incontinence. In addition,
sphincter lacerations are associated with increased pain
in the postpartum period7-8 and the potential complica-
tions of wound breakdown, infection, repair revision, and
formation of rectovaginal fistulae.9

255
256 Christianson et al
Much of the research that has been performed on per-
ineal trauma that is associated with childbirth has been
performed outside the United States,1-7,10-15 with some
notable exceptions.9,12,16-19 We sought to identify risk fac-
tors for anal sphincter injury in the setting of a university-
based hospital in the southeastern United States. Many
European practitioners use fewer forceps and more vacu-
ums for their operative deliveries and prefer mediolateral
episiotomies to midline episiotomies. Some characteris-
tics that make our situation potentially different include
more forceps deliveries and the more frequent use of mid-
line rather than mediolateral episiotomies. Also, all deliv-
eries were performed by either attending physicians or
house staff in the Departments of Obstetrics and Gynecol-
ogy and Family Medicine, with no midwife support staff.
Methods
Setting. This study was a retrospective case-control
study that used a 1:2 ratio of cases to control subjects. To
examine the entire range of delivery experience, cases
and control subjects were drawn from all vaginal deliver-
ies of >18 weeks gestation at the University of Virginia
Health System from May 1999 to April 2001. The project
was approved by the University of Virginia Institutional
Review Board for human research.
Participants. During the study period, there were
2095 vaginal births at ≥18 weeks of gestation. Patients
for whom records were unavailable or cases that in-
volved breech or twin deliveries were excluded, which
left 2078 charts for review. All cases (n = 91), defined by
perineal lacerations greater than second degree in-
curred during vaginal delivery, were identified. These
included any visible interruption of the capsule of the
external anal sphincter with exposure or damage to the
sphincter muscle (“partial” third degree) and third-de-
gree and fourth-degree tears. Cases were eliminated
from the analysis if they had missing values on crucial
predictors or adjustment variables, which left 82 cases
for analysis. Control subjects (n = 176) were selected
randomly from among women who were delivered vagi-
nally during the study period, but who did not incur
greater than a second-degree laceration. Delivery dates
of control subjects were frequency matched to within 2
months of cases. After the removal of the control sub-
jects with missing values, there were 144 control sub-
jects available for analysis.
Data collection and definition. Data were transcribed
from medical records directly to a computerized data
entry system. Data elements included date and time of de-
livery, race (white, black, or other ancestry), maternal
age, parity, mode of delivery (uninstrumented, forceps,
or vacuum), degree of laceration (none through fourth
degree), presentation and position, episiotomy perfor-
mance (yes/no), tobacco use (yes/no), duration of first
and second stages of labor (minutes), fetal weight
July 2003
Am J Obstet Gynecol
(grams), and oxytocin use (yes/no). Missing data for
fetal weight (0 cases, 2 control subjects), first stage of
labor (9 cases, 30 control subjects), or second stage of
labor (6 cases, 23 control subjects) resulted in the final
data set of 82 cases and 144 control subjects.
Data analysis. Cases and control subjects were de-
scribed on mode of delivery and other predictor vari-
ables, and the distribution of those variables was
compared with the use of chi-squared tests (likelihood
ratio) for categoric predictors (eg, episiotomy) and t-
tests for continuous predictors (eg, maternal age). Un-
conditional logistic regression was used to assess the
relative odds of laceration during delivery among deliv-
eries in which forceps or vacuum was used, compared
with uninstrumented deliveries. Other predictors in-
cluded infant birth weight, maternal age, race, parity,
epidural and oxytocin use, episiotomy performance,
and tobacco use. Unadjusted models estimated the
crude association between predictors and case-control
status; multivariate models simultaneously adjusted for
all predictor variables and for the duration of the first
and second stages of labor. Odds ratios and 95% CIs
were calculated for each predictor. To assess the effect
of the elimination of cases and control subjects with
missing data, unadjusted models that used all available
data from the 91 cases and 176 control subjects were
conducted and compared with unadjusted models with
the use of the restricted data set.
Results
In the 2078 vaginal deliveries that were reviewed dur-
ing the study period, the overall incidence of lacerations
greater than second degree was 4.4% (n = 91 lacera-
tions). Forty of the lacerations (43.9%) were partial third
degree; 34 of the lacerations (37.4%) were third degree,
and 17 of the lacerations (18.7%) were fourth degree.
The mean maternal age of our sample was 24.9 ± 5.9
years. Nearly two thirds of the sample (63.2%) were
white; 26.7% were black, and 10.1% were of other racial
background. For other characteristics of the study popu-
lation, please refer to Table I.
We estimated the rate of forceps use in our underlying
patient population at approximately 9%. Potential con-
trol subjects constituted 95.6% of the deliveries in the
population, and forceps were used in 13 of 172 control
subjects; cases constituted 4.4% of deliveries, and forceps
were used in 41 of 91 cases.
Cases and control subjects differed significantly on
most predictor variables. Cases were more likely to have
had deliveries that were aided by forceps, been nullipa-
rous, delivered heavier infants, labored longer, been of
other than black ancestry, and had deliveries that in-
volved epidural anesthesia, oxytocin, or an episiotomy
(Table II). Cases and control subjects did not differ sig-
nificantly in maternal age or tobacco use.
Volume 189, Number 1
Am J Obstet Gynecol
Fig 1. Effect of mode of delivery on sphincter lacerations.
Fig 1 illustrates the association between forceps-assisted
delivery and perineal lacerations. More than 80% of the
control subjects, but fewer than one half of the cases,
were delivered spontaneously without instruments; one
half of the cases, but fewer than 10% of the control sub-
jects, were delivered with the use of forceps. The case-
control difference in vacuum use was less pronounced,
partly because of its relative infrequency overall.
In both unadjusted and adjusted models, forceps
delivery was associated with a substantially increased
risk of perineal tears (Table III). The effect was inde-
pendent of other situational predictors of lacerations,
including nulliparity, duration of labor, and fetal
weight. Epidural and oxytocin use, predictors of tears
in unadjusted analyses, were not significant after the
adjustment. Although episiotomy remained associated
with an increased risk, the association was reduced
substantially in the adjusted model. Greater parity was
associated strongly with reduced risk in both adjusted
and unadjusted analyses (Fig 2). Ancestry other than
white or black was associated with increased risk in
both models, and neither age nor tobacco use were as-
sociated significantly with increased risk of perineal
tears. As expected, in unadjusted analyses, tears were
significantly more likely with increased fetal weight
and duration of the first and second stages of labor. In
multivariate models, fetal birth weight remained a sig-
nificant predictor of tears, although duration of labor
was reduced to near unity and no longer significant.
Assuming a linear association between birth weight
and log odds of tears in adjusted analyses, a 100-g in-
crease in birth weight was associated with a 9% in-
creased risk of tear.
Among the 41 cases who were delivered with forceps,
there were 30 episiotomies (73.2%); among the 13 con-
trol subjects who were delivered with forceps, there were
10 episiotomies (76.9%). Interaction between operative
delivery and episiotomy use was not significant in ad-
justed models of case-control status, although the power
to detect clinically significant interactions was limited.
Among the 41 cases in which forceps were used, in 24 de-
Christianson et al 257
Fig 2. Effect of vaginal parity on sphincter lacerations. The closed
diamonds indicate the cases; the gray squares indicate the control
subjects.
liveries (58.5%) the record suggested that fetal heart
rate was the indication. In 15 cases (36.6%), the mater-
nal indications were cited, and in 2 cases (4.9%) no indi-
cation was listed. Among 13 control subjects who were
delivered with forceps, 10 control subjects (79.9%) had
fetal heart rate as the indication, and 3 control subjects
(23.1%) had maternal reasons. Compared with fetal
heart rate indications, forceps use because of maternal
reasons trended towards an increased risk of anal sphinc-
ter injury; however, these differences did not reach sta-
tistical significance.
Unadjusted odds ratios for mode of delivery with all
available cases and control subjects (odds ratioforceps,
12.5; 95% CI, 6.3-24.8; odds ratiovacuum, 3.0; 95% CI, 1.0-
9.2) did not differ substantially from those odds ratios
that were obtained with the use of the restricted data set.
Comment
Perineal tears that result in greater than second-degree
lacerations are uncommon. The incidence of sphincter
lacerations has been found in previous studies to range
from 0.6% to 6.0%.1,3,10-12,20 Risk factors that were found
previously include operative vaginal delivery (forceps,
vacuum, or the combination), fundal pressure, nullipar-
ity, high birth weight (usually >4000 g), episiotomy, fetal
positions other than vertex and occiput anterior (breech,
occiput posterior), shoulder dystocia, prolonged second
stage, epidural anesthesia, and Indian and Asian de-
scent.1-2,5-6,11-12,17-19,21 Instrumented vaginal delivery and
prolonged second stage were also found to be associated
with occult sphincter lacerations (documented by en-
dosonography).3,4 Combs et al9 found that, among in-
strumented deliveries, the sphincter laceration rate was
30%. The combination of operative delivery and epi-
siotomy has been associated with increased risk of third-
and fourth-degree lacerations, with forceps more so than
vacuum.9,18
Comparing the risk of anal sphincter injury and/or
anorectal complaints with the use of forceps versus vac-
258 Christianson et al July 2003
Am J Obstet Gynecol

Table I. Characteristics of study population

Characteristic Total no. of patients (%) No. of cases (%) No. of control subjects (%)

Nulliparous 161 (60.3) 84 (92.3) 76 (43.2)

Multiparous 106 (39.7) 7 (7.7) 100 (56.8)
Diabetes mellitus 9 (3.4) 4 (4.3) 5 (2.9)
Pitocin 128 (47.9) 53 (57.6) 75 (42.9)
Epidural 197 (73.8) 76 (82.6) 121 (69.1)
Infant weight
<4000 g 247 (89.8) 81 (88.0) 166 (94.9)
>4000 g 20 (7.5) 12 (13.0) 8 (4.6)
Shoulder dystocia 10 (3.7) 5 (5.4) 5 (2.9)
Tobacco use 65 (24.3) 17 (18.5) 48 (27.4)
Race
White 169 (63.3) 57 (62.0) 112 (64.0)
Black 71 (26.6) 17 (18.5) 54 (30.9)
Other 27 (10.1) 17 (18.5) 10 (5.7)
Insurance status
Private 112 (41.9) 48 (52.2) 64 (36.6)
Medicaid 103 (38.6) 33 (35.9) 70 (40.0)
None 36 (13.5) 11 (12.0) 25 (14.3)

Table II. Comparison of cases and control subjects* ered by vacuum, but more facial injuries in the forceps
group.
Cases Control subjects
Predictor (n = 82) (n = 144) The role of episiotomy in sphincter lacerations has re-
mained controversial. Several studies cited midline epi-
Delivery mode (%)† siotomy as a risk factor that is associated with third- and
Vaginal 43.9 85.4
Forceps 50.0 9.0
fourth-degree tears,1,2,8-12,19,22 although others have not
Vacuum 6.1 5.6 found this association.20 No benefits have been shown
Episiotomy (%)† 37.8 11.1 from the routine use of episiotomy compared with re-
Infant birth weight (g)† 3418 ± 623 3175 ± 660
Labor duration (min)
stricted use.8,14,15,22 There have been mixed findings con-
Stage 1† 626 ± 320 470 ± 278 cerning the protective use of mediolateral episiotomy,
Stage 2† 101 ± 77 51 ± 58 with a recent prospective study by Donnelly et al4 and
Epidural use (%)† 84.1 72.9
Pitocin use (%)† 59.8 42.4
Eason et al22 finding no protective effects, and an obser-
Parity (%)† vational study by Leeuw et al10 finding some protection.
0 91.5 41.7 Interestingly, both randomized prospective trials on epi-
1 6.1 33.3
≥2 2.4 25.0
siotomy, the West Berkshire trial and the Argentine trial,
Maternal age (y) 24.2 ± 6.9 25.4 ± 5.3 used only mediolateral episiotomies.14,15 Klein et al8 per-
Race (%)† formed a randomized trial using midline episiotomy and
White 62.2 65.3
Black 20.7 28.5
found significantly less third- and fourth-degree tears
Other 17.1 6.3 with the restrictive use of midline episiotomy and less
Tobacco use (%) 19.5 29.2 pain and sexual dysfunction among both women with in-
tact perinei and spontaneous lacerations, as compared
*Percentages for categoric data, means ± SD for continuous
variables. with episiotomy. A recent study by Signorello et al16
†P < .05, for difference in proportions or means between cases found that episiotomy was associated significantly with
and control subjects. fecal and flatal incontinence at 3 and 6 months after de-
livery, compared with women who were allowed to tear
spontaneously. This effect was seen even when the epi-
uum has yielded mixed results. Sultan et al,3 MacArthur siotomy did not visibly extend into the anal sphincter
et al,5 and Combs et al9 suggested increased risk with for- complex.
ceps but not vacuum; Handa et al20 suggested that vac- The goal of the present study was to identify risk factors
uum carries more risk. Others,9,10 including a for third- and fourth-degree lacerations, by taking into ac-
meta-analysis by Easonet al,22 have implicated both forms count other factors that are associated with birth that po-
of operative delivery, with forceps carrying a greater risk tentially may have confounded the association. We hoped
than vacuum. A prospective, randomized trial by Johan- to discover potentially modifiable risk factors, to strive to-
son et al13 showed significantly less perineal trauma in ward future primary prevention efforts.
the vacuum-extraction group. In this study, there were Our study found an incidence of third- and fourth-de-
more cephalohematomas in the infants who were deliv- gree lacerations of 4.4%, which is within the range previ-
Volume 189, Number 1 Christianson et al 259
Am J Obstet Gynecol

Table III. Association of delivery mode to risk of perineal tears

Unadjusted Adjusted*

Predictor Odds ratio 95% CI Odds ratio 95% CI

Delivery mode
Vaginal 1.0 Reference 1.0 Reference
Forceps 10.8 5.2-22.3 11.9 4.7-30.4
Vacuum 2.1 0.7-6.9 1.1 0.3-4.8
Episiotomy 4.9 2.5-9.6 2.5 1.0-6.0
Epidural use 2.0 1.0-4.0 0.8 0.3-4.8
Pitocin use 2.0 1.2-3.5 1.8 0.8-4.1
Parity
0 1.0 Reference 1.0 Reference
1 0.08 0.03-0.22 0.10 0.03-0.33
≥2 0.04 0.01-0.19 0.06 0.01-0.41
Maternal age (y) 1.0 0.9-1.0 1.0 0.9-1.1
Race
White 1.0 Reference 1.0 Reference
Black 0.8 0.4-1.5 0.8 0.3-2.2
Other 2.7 1.2-7.1 4.0 1.1-14.3
Tobacco use 0.6 0.3-1.1 0.8 0.3-2.1

Cases, 82; control subjects, 144.

*Adjusted for the other variables in the Table and for fetal weight (g) and duration (min) of stage 1 and stage 2 labor.

ously reported. Our forceps delivery rate was 9%, within
the range of 8.2% to 12.9% that was found previously by
Learman.23 Also in concordance with previous investiga-
tions, we found that forceps use, nulliparity, and increased
infant birth weight were risk factors for anal sphincter in-
jury during vaginal delivery. It would appear that fetal indi-
cations (bradycardia or other heart rate abnormalities) did
not increase the risk of sphincter injury during operative
deliveries, as might be suspected. This finding agrees with
a previous study by Helwig et al.17 The use of midline epi-
siotomy was a risk factor only before they controlled for the
other variables that were associated with the delivery. Fac-
tors that were not associated with increased risk included
epidural use, oxytocin, vacuum-assisted vaginal delivery,
duration of first and second stages of labor, maternal age,
and tobacco use. The number of vacuum deliveries in the
study was small, and a larger sample may have yielded sig-
nificant results in this category.
Some potential limitations of this study include gener-
alizability (university-based hospital, residents, no mid-
wives), the small number of vacuum deliveries in the
sample (consistent with trends in the southern United
States found by Learman23), and the complete lack of
mediolateral episiotomies. Inherent limitations in a ret-
rospective analysis include the inability to control data
quality, because all the data were from medical records,
and the potential inconsistency in record keeping across
time and across clinicians. In addition, we were unable to
study other potential risk factors for anal sphincter injury,
such as perineal body length and station of the fetal ver-
tex at which operative deliveries were performed, be-
cause they are not recorded consistently in the delivery
record. It is therefore difficult to demonstrate causal as-
sociations, because some relevant predictors of laceration
may not have been available in the records.
The strengths of our study include a large sample size,
which is one of the few US studies that have been per-
formed to date, that brings with it a slightly different prac-
tice model. We concluded that obstetric decision-making
in vaginal deliveries should be approached from a view-
point of reducing modifiable risk factors for perineal in-
jury. This process may include the assessment of what risk
factors are present already (ie, nulliparity or an epi-
siotomy cut earlier in the delivery process) and basing de-
cisions on the presence of these factors. Examples may
include a consideration of vacuum- versus forceps-as-
sisted delivery, whether operative vaginal delivery is
deemed necessary in the situation, whether to cut an epi-
siotomy, and if so, what type. Given the potential sequelae
of anal sphincter injury, we believe an informed consent
process should be considered before episiotomy or oper-
ative delivery is performed.
Future studies may be improved by more systematic
documentation, which includes attention to detail in dic-
tation and delivery records, and the recording of all vari-
ables that are associated with the delivery, such as infant
weight, head position, fetal station, etc. A computerized
labor and delivery database of this information, such as
those mentioned in European studies, certainly would fa-
cilitate future studies and provide the ability to follow
trends. These potential risk factors should also be ad-
dressed with a prospective study, because the decision of
which potential risk factors to include could be broad-
ened to include those risk factors that have not yet been
studied extensively, such as perineal body length and ma-
ternal height and weight. All desired information would
260 Christianson et al
be obtained at the time of delivery. Finally, future investi-
gations could be enhanced by obtaining follow-up data
on maternal symptoms after delivery, which should in-
clude fecal or flatal incontinence, pain, and sexual dys-
function.
REFERENCES
1. Sultan AH, Kamm MA, Hudson CN, Bartram CI. Third degree
obstetric anal sphincter tears: risk factors and outcome of pri-
mary repair. BMJ 1994;308:887-91.
2. Pregazzi R, Sartore A, Bortoli P, Grimaldi E, Ricci G, Guaschino
S. Immediate postpartum perineal examination as a predictor
of puerperal pelvic floor dysfunction. Obstet Gynecol
2002;99:581-4.
3. Sultan AH, Kamm MA, Hudson CN, Thomas JM, Bartram CI.
Anal-sphincter disruption during vaginal delivery. N Engl J Med
1993;329:1905-11.
4. Donnelly V, Fynes M, Campbell D, Johnson H, Oconnell PR,
Oherlihy C. Obstetric events leading to anal sphincter damage.
Obstet Gynecol 1998;92:955-61.
5. MacArthur C, Glazener CMA, Wilson PD, Herbison GP, Gee H,
Lang GD, et al. Obstetric practice and faecal incontinence three
months after delivery. BJOG 2001;108:678-83.
6. Sorensen SM, Bondesen H, Istre O, Vilmann P. Perineal rupture
following vaginal delivery: long-term consequences. Acta Obstet
Gynecol Scand 1988;67:315-8.
7. Haadem K, Dahlstrom JA, Ling L, Ohrlander S. Anal sphincter
function after vaginal delivery rupture. Obstet Gynecol
1987;70:53-6.
8. Klein MC, Gauthier RJ, Robbins JM, Kaczorowsky J, Jorgensen
SH, Franco E. Relationship of episiotomy to perineal trauma
and morbidity, sexual dysfunction and pelvic floor relaxation.
Am J Obstet Gynecol 1994;171:591-8.
9. Combs A, Robertson P, Laros R. Risk factors for third-degree and
fourth-degree perineal lacerations in forceps and vacuum deliv-
eries. Am J Obstet Gynecol 1990;163:100-4.
July 2003
Am J Obstet Gynecol
10. Leeuw JW, Struijk PC, Vierhout ME, Wallenburg HCS. Risk fac-
tors for 3rd degree perineal ruptures during delivery. Br J Obstet
Gynaecol 2001;108:383-7.
11. Jander C, Lyrenas S. Third and fourth degree perineal tears.
Acta Obstet Gynecol Scand 2001;80:229-34.
12. Zetterstrom J, Lopez A, Anzen B, Norman M, Holmstrom B,
Mellgren A. Anal sphincter tears at vaginal delivery: risk factors
and clinical outcome of primary repair. Obstet Gynecol
1999;94:21-8.
13. Johanson RB, Rice C, Doyle M. A randomised prospective study
comparing the new vacuum extractor policy with forceps deliv-
ery. BJOG 1993;100:524-30.
14. Argentine Episiotomy Trial Collaborative Group. Routine vs se-
lective episiotomy: a randomized controlled trial. Lancet
1993;342:1517-8.
15. Sleep J, Grant A, Garcia J, Elbourne D, Spencer J, Chalmers I.
West Berkshire perineal management trial. BMJ 1984;289:587-
90.
16. Signorello LB, Harlow BL, Chekos AK, Repke JT. Midline epi-
siotomy and anal incontinence: retrospective cohort study. BMJ
2000;320:86-90.
17. Helwig J, Thorp J, Bowes W. Does midline episiotomy increase
the risk of third and fourth-degree lacerations in operative vagi-
nal deliveries? Obstet Gynecol 1993;82:276-9.
18. Peleg D, Kennedy C, Merrill D, Zlatnik F. Risk of repetition of a
sever perineal laceration. Obstet Gynecol 1999;93:1021-4.
19. Green JR, Soohoo SL. Factors associated with rectal injury in
spontaneous deliveries. Obstet Gynecol 1989;73:732-8.
20. Handa VL, Danielsen BH, Gilbert WM. Obstetric anal sphincter
lacerations. Obstet Gynecol 2001;98:225-30.
21. Handa VL, Harris TA, Ostergard DR. Protecting the pelvic floor:
obstetric management to prevent incontinence and pelvic organ
prolapse. Obstet Gynecol 1996;88:470-8.
22. Eason E, Labrecque M, Wells G, Feldman P. Preventing perineal
trauma during childbirth: a systematic review. Obstet Gynecol
2000;95:464-71.
23. Learman LA. Regional differences in operative obstetrics: a look
to the south. Obstet Gynecol 1998;92:514-9.