Risk factors for anal sphincter tears: the importance of maternal position at birth
Dr K Gottvall, Department of Women and Child Health, Division of Reproductive and Perinatal Health Care, Karolinska Institutet, Retzius väg 13 A-B, SE-171 77 Stockholm, Sweden. Email firstname.lastname@example.org
Objective To assess the role of birth position in the occurrence of anal sphincter tears (AST).
Design Observational cohort study.
Setting South Hospital in Stockholm, a teaching hospital with around 5700 births per year.
Population Among all 19 151 women who gave birth at the South Hospital during the study period 2002–05, 12 782 women met the inclusion criteria of noninstrumental, vaginal deliveries.
Methods Data on birth position and other obstetric factors were analysed in relation to occurrence of AST.
Main outcome measure Third- and fourth-degree AST.
Results AST occurred in 449 women (3.5%). The trauma was more frequent in primiparous (5.8%) than in multiparous women (1.7%). The highest proportion of AST was found among women who gave birth in lithotomy position (6.9%), followed by squatting position (6.4%). Logistic regression analyses showed that lithotomy (adjusted OR 2.02, 95% CI 1.58–2.59) and squatting positions (adjusted OR 2.05, 95% CI 1.09–3.82) were associated with a significantly increased risk for AST. Other major risk factors for anal sphincter trauma were primiparity (adjusted OR 3.29, 95% CI 2.55–4.25), prolonged second stage of labour >1 hour (adjusted OR 1.52, 95% CI 1.11–2.10), infant birthweight more than 4 kg (adjusted OR 2.12, 95% CI 1.64–2.72) and large infant head circumference (adjusted OR 1.57, 95% CI 1.23–1.99).
Conclusion Lithotomy and squatting position at birth were associated with an increased risk for AST also after control for other risk factors.
Anal sphincter tears (AST) are a serious complication of vaginal delivery. Injuries to the anal sphincter may cause considerable pain and reduced quality of life in affected women, and they are a major cause of anal and faecal incontinence in healthy women.1–3 The degree of postnatal morbidity for women is directly related to the degree of genital tract trauma.4 Short-term problems include blood loss, need for suturing and perineal pain. Long-term problems may include pain and various functional problems, such as bowel, urinary and sexual dysfunction.5–7 In Sweden, the incidence of AST has increased from 1.7% in 1990 to 4.2% in 2004.8 The main reasons for this increase are unknown, but there is one national study in progress to explore this question.
Factors that have consistently been associated with increased risk of AST are primiparity, abnormal fetal presentation, midline episiotomy, shoulder dystocia, length of second stage of labour, instrumental deliveries, large head circumference and infant birthweight >4 kg and postmaturity.3,9–12 Less is known about the birth position during second stage of labour and the risk for AST.13,14
In many societies, midwives and physicians have influenced women to give birth in a recumbent position, sometimes with their legs up in stirrups to facilitate different obstetric manoeuvres. In Sweden, midwives are involved in all hospital births; they handle uncomplicated deliveries by themselves, while obstetricians are present and responsible for complicated childbirths, for example breech deliveries and vacuum extraction.15 Midwives are trained to guard the perineum during delivery of the fetal head to prevent or reduce injuries to the perineal tissues, independently of whether or not an obstetrician is present. However, management techniques to reduce perineal trauma are often based on empirical experiences, rather than research-based evidence.
Most trials investigating the effects of birth position have categorised the position as upright or horizontal. Upright position usually includes sitting, kneeling, all four, squatting, birth stool and standing, whereas horizontal position usually includes recumbent, semirecumbent, supine and lithotomy position.6,16–19 Physiological advantages with upright position that have been proposed (e.g. in a Cochrane review) include the effect of gravity, more efficient uterine contraction and less compression on vena cava.20,21 Previous studies have shown that upright position may lead to less painful delivery19,22 and a reduction of abnormal fetal heart rate pattern,16 shorter second stage of labour23 and less use of vacuum and forceps, as well as episiotomy.14 However, upright position has been associated with oedema, increased blood loss more labia and second-degree perineal tears.14,16,19,20,24,25 Thus, studies have reported inconsistent findings on the impact of different birth positions per se on the risk of AST. Since most studies have grouped different birth positions because of small sample sizes, a larger study is needed to enable analyses of specific birth positions. In this study, we have been able to include a large population, which makes it possible to analyse the risk for AST in relation to different specific birth positions.
The main objective of our study was to assess the role of maternal position in the second stage of labour as a risk factor for AST.
The study was carried out at South Hospital, Stockholm, Sweden, one of the city’s major hospitals with 6000 deliveries per year (2006). The population of the catchment area corresponds to the national average with regard to ethnic and socio-demographic distribution. As in all major obstetric clinics in Sweden, all women were admitted independently of risk status of the delivery, since specialist services including neonatal care were available. We collected data from the main hospital computerised database from April 2002, when the database was introduced, to 31 December 2005. Over the 4-year study period (45 month), 19 157 women gave birth at South hospital. Multiple pregnancies (1.9%), caesarean (22.1%) and instrumental vaginal deliveries (vacuum and forceps [9.6%]) were excluded. The latter were excluded, since they are strongly related to specific birth positions, especially lithotomy position. Vaginal breech deliveries were not excluded.
Collection of data
According to Swedish routines, all uncomplicated deliveries were managed by trained midwives. An obstetrician was called upon whenever needed, for example when a sphincter tear or other complication was suspected. The midwife in charge was responsible for entering and collecting information on pregnancy, labour, delivery and the puerperium continuously, on all births by means of bedside computers. Registration was completed before maternal transfer to home or to the postnatal unit. The obstetrician was responsible for recording any consultations, medical interventions and diagnoses during the woman’s hospital stay.
We collected information on the following maternal and obstetric factors: parity, maternal age, length of gestation, induction of labour, augmentation of labour, epidural anaesthesia, no pain relief (pharmacological or nonpharmacological), episiotomy (mediolateral or midline), length of second stage of labour and maternal position. Maternal position was defined as position of the woman in late second stage of labour, at the time of birth. Table 1 shows the classification of maternal positions based on standard references14,26 and applied by midwives. Information on infant-related factors was: non-occipitoanterior position including breech position, Apgar score less than 7 at 1 and 5 minutes, asphyxia, infant birthweight and head circumference.
Table 1. Classification of maternal positions at birth applied by midwives
|Sitting||Sitting up at about 60° angle with the delivery bed. The legs may be out straight or bent with the feet flat on the bed|
|The lithotomy||The woman lies on her back in a semirecumbent posture with her trunk raised less than 45° to reduce the aortocaval compression, with legs in stirrups. The position is used for both spontaneous and particularly for assisted vaginal deliveries|
|Supine or semirecumbent||The woman lies flat on her back or with her trunk slightly raised about 45° or less. Her legs may be out straight or bent with her feet flat on the bed|
|Kneeling/all-fours||Kneeling on bed or floor, leaning forward on the palms of her hands or into the back of the bed or other support|
|Lateral recumbent||The woman lies on side with both hips and knees flexed and often with a pillow between her legs, or with her upper leg raised and supported while pushing|
|Squatting||Often termed as the most natural position and is often used by women if left alone to choose their own position for birth|
|Birth chair||Sitting upright on a birthing stool supported by the partner (sitting behind)|
|Standing||Supported by the husband/partner or leaning on a counter/furniture|
The main outcome variable was rupture of the anal sphincter, diagnosed as third- and fourth-degree tears. Third degree was defined as lacerations, ruptures or tears extending through the skin, mucous membrane, perineal body and involving the anal sphincter (partial or complete). Fourth-degree laceration was defined as lacerations extending through the anal and rectal mucosa to expose the lumen of the rectum. The classification was based on International Classification of diseases (ICD–10 codes O702A, O702B, O703).27 In the analyses, we collapsed third- and fourth-degree lacerations as outcome variable.
The study was approved by the Regional Research and Ethics Committee in Stockholm (Ref. no 2006/1:11).
Prevalence of AST was calculated as proportions of women with AST among all women in the study population, with 95% CI. We analysed risk factors for AST related to maternal, obstetric and perinatal factors. Crude odds ratios with 95% CI were calculated in all women and stratified by parity. Since maternal position was related to several known risk factors for AST, we performed stepwise logistic regression analyses to obtain odds ratios with 95% CIs for different birth position, adjusting for known risk factors as shown in Table 2. Since most women gave birth in a sitting position (45%), these were treated as reference group. Statistical analysis was performed using SPSS version 13.0.28
Table 2. Prevalence of AST distributed by women’s position at birth in primiparas and multiparas, presented as proportions with 95% CI
|Total||12 782|| ||5600|| ||7182|| |
We analysed each birth position separately. Only kneeling and standing on all-fours were collapsed, since few women (n = 6) used the latter position.
Among the 12 782 women who fulfilled the inclusion criteria, the proportion of primiparous women was 43.8% and multiparous women was 56.2%. Anal sphincter injury occurred in 3.5% (n = 449) of the deliveries; most of them (84.6%) were third-degree tears (partial, n = 325; or complete, n = 55) and 15.4% fourth-degree tears (n = 69). Among primiparous women, 5.8% of the women had an anal sphincter injury, while among multiparous women the proportion was 1.7%.
The most common birth position was sitting (42.0% of all women), followed by lithotomy and kneeling position (21.1 and 11.6%). Supine, lateral- and semirecumbent positions were used by 8.5, 8.0 and 2.7% of the women. Very few women gave birth in standing, birthing stool and squatting positions (1.2, 1.5 and 1.8%, respectively).
Table 2 presents the prevalence of AST in different birth positions distributed by parity. The highest occurrence of AST was found among women who gave birth in lithotomy position (6.9%), followed by squatting (6.4%), and the prevalence was higher in primiparous women than in multiparous women. The lowest AST rate among primiparous women occurred in the lateral recumbent position.
Maternal, obstetric and infant factors associated with AST are shown in Table 3. Women less than 35 years, those who obtained epidural anaesthesia and had a prolonged second stage of labour, had a higher proportion of AST. Also among women who gave birth to infants with birthweight over 4 kg and neonatal head circumference 35 cm or more the proportion of AST were increased. Postmaturity (≥42 weeks) and augmentation of labour with oxytocin were also associated with a higher proportion of AST in primiparous women but not in multiparous women. Women who received no pain relief at all (pharmacological or nonpharmacological) had a lower rate of AST. There was no association between induction of labour, episiotomy, non-occipitoanterior fetal head position, Apgar score and the occurrence of AST. None of the women whose infant had an Apgar score less than 7 at 5 or 10 minutes had an AST.
Table 3. Factors associated with AST in the study population, presented as crude odds ratios with 95% CI
|Maternal age (years)|
|<35||9150||358||3.9||1.0|| ||4708||279||5.9||1.0|| ||4442||79||1.8||1.0|| |
|Length of gestation (weeks)|
|<42||12 146||412||3.4||1.0|| ||5289||298||5.6||1.0|| ||6857||114||1.7|| |
|Induction of labour|
|No||11 803||413||3.5||1.0|| ||5053||296||5.9||1.0|| ||6750||117||1.7||1.0|| |
|Augmentation of labour|
|No||10 366||319||3.1||1.0|| ||4090||214||5.2||1.0|| ||6276||105||1.7||1.0|| |
|No||8622||253||2.9||1.0|| ||2978||172||5.8||1.0|| ||5644||81||1.4||1.0|| |
|No pain relief|
|No||11 684||432||3.7||1.0|| ||5335||316||5.9||1.0|| ||6349||116||1.8||1.0|| |
|Length of second stage (minutes)|
|≤60||11 521||372||3.4||1.0|| ||4831||265||5.5||1.0|| ||6690||107||1.6||1.0|| |
|Missing||551||23|| ||192||16|| ||359||7|| |
|Episiotomy, all types|
|No||12 112||419||3.5||1.0|| ||5082||301||5.9||1.0|| ||7030||118||1.7||1.0|| |
|No||11 715||421||3.5||1.0|| ||5097||303||5.9||1.0|| ||7039||118||1.7||1.0|| |
|No||12 758||447||3.5||1.0|| ||5585||324||5.8||1.0|| ||7173||123||1.7||1.0|| |
|No||12 299||436||3.5||1.0|| ||5381||316||5.9||1.0|| ||6918||120||1.7||1.0|| |
|Apgar score <7 at 1 minutes|
|No||12 363||432||3.5||1.0|| ||5413||310||5.7||1.0|| ||6950||121||1.7||1.0|| |
|Missing||72||0|| ||14||0|| ||58||0|| |
|Apgar score <7 at 5 minutes|
|Yes||83||0||0|| ||48||0|| ||35||0|| |
|No||12 626||448||3.6|| ||5536||325||5.9||1.0|| ||7090||123||1.7||1.0|| |
|Missing||73||1|| ||16||1|| ||57||0|| |
|No||12 753||446||3.5||1.0|| ||5579||323||5.8||1.0|| ||7174||123||1.7|| |
|≤4000||10 261||299||2.9||1.0|| ||4833||230||4.8||1.0|| ||5428||69||1.3||1.0|| |
|Missing||19||449|| ||0|| ||9||0|| |
|Infants head circumference (cm)|
|≥35||8140||221||2.7||1.0|| ||3870||172||4.4||1.0|| ||4270||49||1.1||1.0|| |
|Missing||873||30|| ||360||21|| ||513||9|| |
We also analysed the association between the factors presented in Table 3 and different birth positions as the outcome variable (not in table). Lithotomy position was associated with postmaturity (OR 1.9, 95% CI 1.6–2.2), induction of labour (OR 1.6, 95% CI 1.4–1.9), use of epidural anaesthesia (OR 2.7, 95% CI 2.5–3.0), a prolonged second stage of labour (OR 2.0, 95% CI 1.7–2.3), episiotomy (all types) (OR 4.2, 95% CI 3.5–4.9), low Apgar score (less than 7) at 1 minute (OR 1.8, 95% CI 1.4–2.2) and infant birthweight more than 4000 g (OR 1.2, 95% CI 1.1–1.3). In the squatting position, fewer women had an epidural (OR 0.7, 95% CI 0.5–0.9) and episiotomy (OR 0.3, 95% CI 0.1–0.8). More women in squatting position gave birth without any kind of pain relief (OR 2.2, 95% CI 1.6–3.1). No other birth position was associated with any of the risk factors as shown in Table 2.
We conducted forward stepwise logistic regression analysis adjusting for different types of risk factors. The final model (Table 4) indicated that lithotomy and squatting positions were independently associated with a double risk for AST (lithotomy: adjusted OR 2.1, 95% CI 1.6–2.6; squatting: adjusted OR 2.0, 95% CI 1.1–3.8). Supine, semirecumbent, lateral recumbent and kneeling positions showed a tendency to decreased risk for AST, although not statistically significant. Other risk factors for AST were primiparity (adjusted OR 3.5, 95% CI 2.8–4.5), prolonged second stage of labour (>1 hour) (adjusted OR 1.5, 95% CI 1.1–2.1), infant birthweight more than 4000 g (adjusted OR 2.2, 95% CI 1.7–2.8) and infant head circumference more than 35 cm (adjusted OR 1.6, 95% CI 1.3–2.0).
Table 4. Logistic regression for AST in spontaneous vaginal deliveries presented as adjusted odds ratios with 95% CI
|Maternal age ≥ 35 years||0.87||0.67–1.13|
|Length of gestation ≥ 42 weeks||1.14||0.78–1.69|
|Induction of labour||0.81||0.55–1.20|
|No pain relief||0.66||0.38–1.13|
|Augmentation of labour||1.14||0.90–1.45|
|Second stage of labour >1 h||1.52||1.11–2.10|
|One minute Apgar score <7||1.43||0.84–2.43|
|Birthweight ≥ 4000 g||2.12||1.65–2.73|
|Head circumference ≥35 cm||1.56||1.23–1.98|
When performing multivariate analysis separately for primiparous and multiparous (not in table), the increased risk for AST in squatting position disappeared in multiparas (OR 0.6, 95% CI 0.1–4.3).
Our main finding was that women who used a lithotomy or squatting position in the second stage of labour had a higher risk of AST compared with women using other positions. The high proportion of AST in women with a lithotomy position could not be explained by a more complicated birth, since we controlled for already known risk factors, such as parity, gestational age, birthweight and other obstetric characteristics. The fact that the prevalence of AST was lower than the national average can be explained by our exclusion of instrumental deliveries that have a higher rate of AST.
Zetterstrom et al.3 found an association between nonupright (sitting and lithotomy) position and AST in univariate analysis (sitting: OR 2.2, 95% CI 1.3–3.9; lithotomy: OR 4.6, 95% CI 1.9–11.1). The association disappeared, however, in multivariate analysis, perhaps due to the fact that instrumental deliveries were included in the study, and the risk of birth position was limited when these were adjusted for.
While not reporting the occurrence of AST specifically, Albers et.al.29 found a lower rate of intact perineum in lithotomy position (P = 0.05). Soong and Barnes21 found a moderately increased risk of perineal trauma in lithotomy position, although the numbers were probably too small to reach statistical significance.
Similar to our findings, Shorten et al.30 reported that squatting position was associated with an increased proportion of AST (3.2 compared with 0.9% in the total study population). Other studies analysing squatting position16,31,32 have not had the power to assess a possible increased risk of AST. In a systematic review, Gupta and Hofmeyr14 reported more second-degree lacerations in squatting position but no differences in third- and fourth-degree tears because of too few studies and lack of power.
Several factors have been put forward that might explain the association between squatting position and risk of AST: first, the gravitational forces might cause a high pressure on the pelvic floor with increased risk of rupture, although it has been pointed out that contraction activity is the important factor in determining pressure.33 Second, this position makes the pelvic floor less accessible for the midwife, and the application to guard the perineum is thus more difficult. It is less obvious why lithotomy position should be associated with increased risk. Albers et al.29 suggested that lithotomy position may increase pressure sensations in the perineal area and decrease the woman’s ability to moderate the tempo of her own pushing efforts. The possibility of confounding by large infant should be considered, since lithotomy position is often suggested for women with large infants, and these might be at greater risk for AST. We did analyse the occurrence of AST in women with infants more than 4000 g who used lithotomy position, but there was no increased risk for AST in these women. Furthermore, we did control for birthweight in the logistic regression analyses.
The squatting position is often used if women are left alone to choose their own position,34 although western women have difficulties to adopt the position for a longer time, which may be a reason why few women birthed in this position.
Contrary to our findings, Gareberg et al.17 found that standing position was associated with increased occurrence of anal sphincter lacerations than sitting. Albers showed that sitting position as well as delivery of the fetal head between uterine contractions are protective for genital tract trauma (not specifically AST). We had no record on the latter practice in our data, since it is uncommon in Sweden.35
Most episiotomies were mediolateral and only a few were midline episiotomies. Because of the small numbers of the latter, we merged them and treated all episiotomies as one group in the multivariate analysis. Previous studies have reported contradictory results on the relation between episiotomy and AST. In our study, the relationship between episiotomy and AST was not statistically significant. One reason for this is probably the low proportion of episiotomies in the study population.
Some limitations with our study should be mentioned: we have no record on how long the women adopted the position noted in the files. The instruction is to record the position in late second stage and in the moment of birth (head crowned), and there is reason to believe that this is also the moment when the risk of perineal injury is greatest. Aspects of perineal management, such as guarding (protection of perineum, hands on), were not recorded, although these do have an impact on perineal outcome.36 As already pointed out, the effect of position at birth on the occurrence of AST could be mediated by perineal management. Thus, a full understanding of the pathways leading to perineal injury would need better recordings by midwives of all events during labour and birth. Our data were based on routine recordings by the midwife, with an inherent risk of underreporting or misreporting. However, we have no reason to believe that the occurrence of AST would be in association with any particular type of misclassification of procedures during labour and birth.
Some of the limitations of this study are inherent in observational studies. As already mentioned, results of a number of randomised controlled trials (RCTs) have not shown any increased risk of AST by position at birth. However, while RCTs in general are the best way of analysing effects of interventions, the RCTs performed have been small and not had the power to assess occurrence of AST.
We found that lithotomy and squatting positions may increase the risk of third- and fourth-degree perineal tears. These findings contribute to growing evidence that birth position may affect severe perineal trauma, although the current evidence on the overall effects of various delivery positions is inconclusive. Continuing audit of procedures performed as well as cases of trauma should be encouraged to minimise the occurrence of severe birth trauma.