Episiotomy rate in Vietnamese-born women in Australia: support for a change in obstetric practice in Viet Nam

Abstract

Objective

To describe the use of episiotomy among Vietnamese-born women in Australia, including risk factors for, and pregnancy outcomes associated with, episiotomy.

Methods

This population-based, retrospective cohort study included data on 598 305 singleton, term (i.e. ≥ 37 weeks’ gestation) and vertex-presenting vaginal births between 2001 and 2010. Data were obtained from linked, validated, population-level birth and hospitalization data sets. Contingency tables and multivariate analysis were used to compare risk factors and pregnancy outcomes in women who did or did not have an episiotomy.

Findings

The episiotomy rate in 12 208 Vietnamese-born women was 29.9%, compared with 15.1% in Australian-born women. Among Vietnamese-born women, those who had an episiotomy were significantly more likely than those who did not to be primiparous, give birth in a private hospital, have induced labour or undergo instrumental delivery. In these women, having an episiotomy was associated with postpartum haemorrhage (adjusted odds ratio, aOR: 1.26; 95% confidence interval, CI: 1.08–1.46) and postnatal hospitalization for more than 4 days (aOR: 1.14; 95% CI: 1.00–1.29). Among multiparous women only, episiotomy was positively associated with a third- or fourth-degree perineal tear (aOR: 2.00; 95% CI: 1.31–3.06); in contrast, among primiparous women the association was negative (aOR: 0.47; 95% CI: 0.37–0.60).

Conclusion

Episiotomy was performed in far fewer Vietnamese-born women giving birth in Australia than in Viet Nam, where more than 85% undergo the procedure, and was not associated with adverse outcomes. A lower episiotomy rate should be achievable in Viet Nam.

Introduction

Episiotomy, which involves surgical incision to increase the diameter of the vaginal outlet to facilitate a baby’s birth, is one of the most common medical procedures experienced by women throughout the world.¹ Internationally, there is a wide variation in episiotomy practice, ranging from routine use in all births to use only when clinically indicated (i.e. restrictive or selective use).^1,2 A systematic review of randomized controlled trials showed that restrictive episiotomy has several advantages over routine episiotomy: less posterior perineal trauma, less suturing and fewer healing complications.¹ Although the restrictive approach was associated with an increased risk of anterior perineal trauma, there was no difference between the two approaches in severe vaginal or perineal trauma or in most pain measures. As a result, the World Health Organization recommended that episiotomy be performed only for a strictly limited number of indications.³ Nevertheless, the procedure is still used routinely in Viet Nam, although the reason for this is unclear. As in many but not all Asian countries, over 85% of Vietnamese women who give birth vaginally, including almost 100% of primiparae, have an episiotomy.^4–6 Since none of the randomized trials on episiotomy involved Asian women,¹ Vietnamese clinicians question whether the results of these trials are generalizable to Vietnamese women.

We hypothesized that a study of episiotomy among Vietnamese women would help Vietnamese clinicians make informed decisions about clinical practice and, thereby, achieve the best pregnancy outcomes. The availability of population-based data in Australia provided an opportunity to examine this issue. Australia has a multicultural population and Viet Nam is the fifth most common country of birth for women having babies in Australia.⁷ Furthermore, episiotomy is used on a restricted basis in the country.^8,9 Therefore, the aim of this study was to describe the use of episiotomy among Vietnamese-born women giving birth in Australia. Specific objectives were to examine both risk factors for, and pregnancy outcomes associated with, episiotomy.

Methods

This population-based, retrospective cohort study included all births in the Australian state of New South Wales between 1 January 2001 and 31 December 2010 that were singleton, took place at full term (i.e. at 37 weeks’ gestation or more) and involved vaginal delivery of a fetus with a vertex presentation. New South Wales is the most populous Australian state and contains approximately one third of the country’s population. Data were obtained from two linked, validated population health data sets: the New South Wales Perinatal Data Collection and the New South Wales Admitted Patients Data Collection. The Perinatal Data Collection, whose contents are hereafter referred to as “birth records”, is a statutory, population-based surveillance system containing routinely collected perinatal data that include information on all births taking place at 20 weeks’ gestation or more or resulting in a baby that weighs at least 400 g. Information on maternal characteristics, pregnancy, labour, delivery and infant outcomes were collected using a standardized data collection form with check boxes that are completed by the attending midwife or doctor. The Admitted Patients Data Collection, whose contents are hereafter referred to as “hospital records”, is a census of all inpatient discharges from both public and private hospitals in New South Wales. For each hospitalized patient, diagnoses and procedures are coded from medical records using the Australian modification of the International statistical classification of diseases and related health problems, tenth revision, and the affiliated Australian classification of health interventions.¹⁰

As Australia does not have a unique registration number for citizens, the separate data sets were linked using probabilistic linkage methods by the Centre for Health Record Linkage.^11,12 This process involves blocking and matching combinations of selected variables, such as name, date of birth, address and hospital, and assigning a probability weight to the match. Linkage was performed independently of the research. The Centre for Health Record Linkage undertakes quality assurance for any data linkage and assesses linkage quality by manually reviewing personal identifiers for a sample of the records analysed for linkage.¹² For this project, the reported linkage quality was less than 1 missed link per 1000 and less than 3 false-positive links per 1000. Researchers were provided with anonymized data. Approval for the record linkage and analysis for this study was obtained from the New South Wales Population Health and Health Services Research Ethics Committee.

Explanatory and outcome variables were prespecified and defined according to the data available in the two data sets and data that had been assessed in previous validation studies.^13–16 Episiotomy was treated both as an outcome variable for factors associated with episiotomy use and as an exposure variable for third- or fourth-degree perineal tears, postpartum haemorrhage and prolonged postnatal hospitalization. A third-degree perineal tear is defined as partial or complete disruption of the anal sphincter muscles, which may involve either the external or internal anal sphincter muscles or both, and a fourth-degree tear, as disruption of the anal sphincter muscles involving a breach of the rectal mucosa. Episiotomy use was identified from birth or hospital records, with an ascertainment rate of 91% and a positive predictive value of 98%.¹⁵ The occurrence of third- or fourth-degree perineal tears was identified from either diagnostic or procedural (i.e. repair) codes in hospital records (ascertainment rate: 94%, positive predictive value: 100%),¹⁵ as was the occurrence of postpartum haemorrhage (ascertainment rate: 74%, positive predictive value: 84%).¹⁶ The length of the mother’s hospital stay after birth was defined as the number of days between the date of the baby’s birth and the discharge date and prolonged hospitalization was considered to be any stay longer than 4 days.

Information about the mother’s country of birth (i.e. Viet Nam or Australia) and about the explanatory variables of tobacco smoking, type of obstetric care (i.e. public or private) and maternal hypertension were obtained from either birth or hospital records.^14,16 Data on the presence of diabetes were obtained from hospital records alone.¹³ Birth records alone provided details of other explanatory variables, including: maternal age; parity; rural or urban residence; initiation of antenatal care by 12 weeks’ gestation; onset of labour (i.e. spontaneous or induced); epidural analgesia, including combined spinal and epidural anaesthesia; instrumental birth involving forceps or vacuum suction; and birth weight (i.e. < 3800 or ≥ 3800 g). Among Vietnamese-born women giving birth in Australia, a birth weight of 3800 g equates to approximately the 90th birth weight percentile for babies born at full term. In Viet Nam, this birth weight equates to the 95th percentile. In the absence of information on fetal distress, which is a potential explanatory factor for episiotomy, we used an Apgar score of less than 4 one minute after birth.

Statistical analysis

The analysis was conducted in three stages. First, we compared episiotomy rates and trends in Vietnamese-born and Australian-born women. Second, we assessed explanatory factors for episiotomy. Finally, we explored the association between episiotomy and the outcomes of postpartum haemorrhage, prolonged hospitalization and third- or fourth-degree perineal tears. Explanatory factors for episiotomy were examined using contingency table and multivariate logistic regression analysis. Multivariate logistic regression was also used to examine associations between episiotomy and the outcomes of postpartum haemorrhage, prolonged hospitalization and third- or fourth-degree perineal tears. Since the association between episiotomy and third- or fourth-degree perineal tears was influenced by parity, primiparous and multiparous women were modelled separately. All potential explanatory factors suggested by univariate analysis and possible confounders were included in the multivariate adjusted models. Less than 0.1% of data were missing for any variable and records with missing data were excluded from all logistic regression models. The results of logistic regression analysis are presented as odds ratios (ORs) and 95% confidence intervals (95% CIs). All analyses were carried out using SAS 9.1 (SAS Institute Inc., Cary, United States of America).

Results

Between 2001 and 2010, there were 598 305 vaginal births of single, vertex-presenting infants at term among women of all ethnicities in New South Wales, including 100 910 women (16.9%) who had an episiotomy. The rate of episiotomy among Australian-born women was 15.1% (i.e. 62 349 of 413 047 women), compared with 29.9% (i.e. 3646 of 12 208 women) among Vietnamese-born women (P < 0.0001). The rate differed by parity in both groups: in Australian-born women, it was 27% among primiparae and 7% among multiparae; in Vietnamese-born women, the corresponding rates in the two groups were 48% and 17%, respectively.

Table 1 lists the maternal and labour factors associated with the performance of an episiotomy among Vietnamese-born mothers. Multivariate analyses showed that a woman who had an episiotomy was more likely than one who did not to be primiparous, to have received private obstetric care and to have had an induced labour or a forceps- or vacuum-assisted delivery. Living in a rural locality and being a smoker during pregnancy were associated with a decreased likelihood of an episiotomy. Although epidural analgesia was associated with episiotomy in the univariate analysis (data not shown), the relationship was confounded by instrumental birth and private care, and epidural analgesia was not independently associated with episiotomy in the multivariate analysis. A birth weight of 3800 g or more was not associated with episiotomy and increasing the birth weight cut-off did not change this finding. Only 380 (3.3%) Vietnamese-born women had infants that weighed 4000 g or more, including 28 (0.2%) who had infants weighing 4500 g or more.

Table 1. Factors associated with episiotomy in Vietnamese-born women, New South Wales, Australia, 2001–2010.

Factor	No. (%) of women who had an episiotomy (n = 3646)	No. (%) of women who did not have an episiotomy (n = 8562)	Risk of episiotomy, aOR (95% CI)a,b
Maternal age ≥ 35 years	638 (17.5)	1721 (20.1)	1.09 (0.97–1.23)
Parity
Primiparous	2391 (65.6)	2601 (30.4)	3.36 (3.06–3.69)
Multiparous	1255 (34.4)	5961 (69.6)	1.00 (NA)c
Rural residence	71 (2.0)	233 (2.7)	0.60 (0.45–0.82)
Type of health care
Public	2568 (70.4)	7344 (85.8)	1.00 (NA)c
Private	1078 (29.6)	1218 (14.2)	2.07 (1.85–2.32)
Diabetesd	447 (12.3)	1040 (12.2)	1.02 (0.89–1.17)
Hypertensive disorders of pregnancye	125 (3.4)	188 (2.2)	1.04 (0.77–1.36)
Smoker during pregnancy	76 (2.1)	240 (2.8)	0.63 (0.47–0.85)
Induction (versus spontaneous labour)	751 (20.6)	1129 (13.2)	1.26 (1.12–1.42)
Antenatal visit by week 12 of gestation	1741 (48.5)	3562 (42.1)	1.03 (0.94–1.13)
Epidural analgesia	557 (15.3)	529 (6.2)	0.85 (0.73–1.00)
Mode of delivery
Forceps	350 (9.6)	82 (1.0)	8.27 (6.36–10.77)
Vacuum-assisted	914 (25.1)	478 (5.6)	3.87 (3.40–4.41)
Spontaneous vaginal birth	2382 (65.3)	8002 (93.5)	1.00 (NA)c
Birth weight ≥ 3800 g	303 (8.3)	711 (8.3)	1.12 (0.95–1.31)
Apgar score at 1 min < 4	56 (1.5)	104 (1.2)	0.96 (0.66–1.41)

aOR, adjusted odds ratio; CI, confidence interval; NA, not applicable.

^a All factors listed in the table were taken into account for the calculation of the aOR.

^b The reference group comprises women without the relevant characteristic, unless otherwise stated.

^c Reference group.

^d Diabetes includes pre-existing and gestational diabetes.

^e Hypertensive disorders of pregnancy include chronic hypertension, gestational hypertension, pre-eclampsia and eclampsia.

The overall rate of postpartum haemorrhage, a hospital stay longer than 4 days and third- or fourth-degree perineal tears among Vietnamese-born mothers was 8.4%, 14.2% and 4.1%, respectively. The postpartum haemorrhage rate was greater in women who had an episiotomy than in those who did not (10.6% versus 7.4%, respectively), as was the rate of prolonged hospitalization (22.6% versus 10.6%, respectively). Among multiparous women, episiotomy was associated with third- or fourth-degree perineal tears: the rate was 3.2% in women who had an episiotomy and 1.3% in those who did not (Table 2). In contrast, among primiparous women, 6.1% of those who had an episiotomy had a third- or fourth-degree perineal tear compared with 9.0% of those who did not have an episiotomy. This reduction in risk among primiparous women persisted after adjustment for other explanatory factors for third- or fourth-degree perineal tears (Table 2).

Table 2. Perinatal outcomes in Vietnamese-born women, by episiotomy, New South Wales, Australia, 2001–2010.

Perinatal outcome	No. (%) of women who had an episiotomy (n = 3646)	No. (%) of women who did not have an episiotomy (n = 8562)	Risk of adverse perinatal outcome
			cOR (95% CI)	aOR (95% CI)a
Postpartum haemorrhage	387 (10.6)	635 (7.4)	1.48 (1.30–1.69)	1.26 (1.08–1.46)
Postnatal hospitalization for > 4 days	822 (22.6)	908 (10.6)	2.45 (2.21–2.72)	1.14 (1.00–1.29)
Third- or fourth-degree perineal tear
Primiparae	146 (6.1)	233 (9.0)	0.66 (0.53–0.82)	0.47 (0.37–0.60)
Multiparae	40 (3.2)	77 (1.3)	2.52 (1.71–3.70)	2.00 (1.31–3.06)

aOR, adjusted odds ratio; CI, confidence interval; cOR, crude odds ratio.

^a The following factors were taken into account for the calculation of the aOR: maternal age, parity (for postpartum haemorrhage and postnatal hospitalization), private or public care, diabetes, hypertension, induction of labour, epidural analgesia, mode of delivery and birth weight.

Discussion

This study shows that episiotomy was performed twice as often in Vietnamese-born women living in Australia than in Australian-born women: the episiotomy rate was approximately 30% in Vietnamese-born women and 15% in Australian-born women. Nevertheless, the rate in Vietnamese-born women was markedly lower than that in Viet Nam. In Australia, where the use of episiotomy is restricted, episiotomy among Vietnamese women was associated with an increased risk of postpartum haemorrhage and of prolonged hospitalization and, in primiparous women only, with a decreased risk of a third- or fourth-degree perineal tear. If these findings are generalizable to Vietnamese women giving birth in Viet Nam, a reduction in the episiotomy rate in that country may be achieved without increasing the number of adverse outcomes.

The factors that influence decision-making around episiotomy in an environment where use of the procedure is restricted may be related to preconceptions about high-risk subgroups, in addition to clinical indicators or the risk of impending perineal trauma. It has been reported that Asian women are at an increased risk of both episiotomy and perineal trauma^17–19 and that the risk remains elevated after adjustment for explanatory factors such as parity and instrumental delivery.^17–19 Anecdotally, there is a perception that Asian women are more likely to experience severe perineal trauma because of physiological differences such as a shorter perineum.^20–22 However, this perception is not supported by a study from Hong Kong Special Administrative Region, China, that involved measuring the perineal length during labour.⁴ Lai et al. reported that the mean ± standard deviation perineal length of Chinese women was 38.8 ± 7.9 mm, which was comparable to the perineal lengths found using similar methods in other countries, including Israel (mean ± standard deviation: 40.2 ± 10.7 mm), Turkey (mean ± standard deviation: 36.6 ± 5.2 mm) and the United States of America (mean ± standard deviation: 39 ± 7.0 mm).^4,20,23,24 However, this does not rule out the possibility that there are other ethnic differences in the functional and morphological characteristics of the pelvic floor.^25,26

The explanatory factors for episiotomy among Vietnamese-born women observed in our study (i.e. nulliparity, private obstetric care, induction of labour and instrumental delivery) are consistent with published risk factors.^19,27–30 However, there was no association with an infant’s birth weight exceeding the 90th percentile of 3800 g in these women. In Caucasian populations, a birth weight of 4000 g or more has been identified as a predictive factor for episiotomy, as has a birth weight of 4500 g or more.^27–29 Few infants of Vietnamese women reached these sizes and the few that did were no more likely to be born with the aid of an episiotomy than smaller infants. Similarly, an Apgar score less than 4 one minute after birth was not an independent explanatory factor for episiotomy, which suggests that fetal distress was unlikely to have had a major influence on the use of episiotomy in this population.

The relationship between episiotomy and severe perineal trauma was complex. Although the risk of a third- or fourth-degree perineal tear was associated with episiotomy, it was influenced by parity: the risk was decreased in primiparae and increased in multiparae. However, the burden of disease fell largely on primiparae: 78% of third- or fourth-degree perineal tears occurred among women having their first baby. As in our study, previous observational studies have reported that episiotomy is associated with fewer third- or fourth-degree perineal tears among primiparous women but not among multiparous women.^31–34 In contrast, a systematic review of randomized trials of restrictive versus routine episiotomy showed a reduced risk of severe perineal trauma in both primiparae and multiparae, although the effect did not reach statistical significance among multiparae.¹ A key difference between observational and experimental studies is the selection of patients, which may explain the observed difference in the effect of episiotomy. In Australia, an effort is made to avoid episiotomy, especially among multiparae, in whom the procedure is used only when there is a high risk of a third- or fourth-degree perineal tear, the birth of a compromised fetus has to be accelerated or there is impending perineal trauma. In contrast, among primiparae, more episiotomies may be performed specifically to prevent perineal tears.⁸

Our population-based study benefitted from the availability of reliably collected, labour and delivery data.^13–16 However, these data did not include detailed clinical information on the type of episiotomy performed, other clinical risk factors (e.g. the duration of the second stage of labour), the accoucheur or the outcomes of perineal pain and dyspareunia. Furthermore, Vietnamese-born women giving birth in Australia may differ from those giving birth in Viet Nam and obstetric practice in the two courtiers is likely to be different; both of these factors may restrict the generalizability of our findings to Viet Nam. Nevertheless, our study suggests that a much lower episiotomy rate can be achieved in Vietnamese and other Asian women without increasing the number of adverse outcomes.

We hope that the results of this study will be helpful in bringing about a change in obstetric practice in Viet Nam. A maternity hospital in Hong Kong SAR reported successfully decreasing the episiotomy rate from 73% in 2003 to 27% in 2008 without increasing the rate of third- or fourth-degree perineal tears.⁴ However, the strategies implemented to achieve this outcome were not reported. Similarly a randomized controlled trial of a multifaceted intervention in South America reduced the episiotomy rate from 41% to 30%;³⁵ in this trial, the intervention included: the involvement of selected opinion leaders; interactive workshops; skills training in the use of episiotomy for birth attendants; individual evidence-based education on episiotomy for birth attendants by trained health-care professionals; reminders; and feedback. Other activities planned to facilitate practice change in Viet Nam to reduce the episiotomy rate include: an audit of current episiotomy practice and outcomes based on delivery room registration of the episiotomies that have been performed; a survey of maternity care providers’ knowledge of and attitudes towards episiotomy; and an assessment of health service resources currently used (e.g. the number of staff skilled in conducting and repairing episiotomies, equipment costs and the length of hospitalization after episiotomy). This information will be used to develop a programme for improving practice. In conclusion, a more restrictive episiotomy policy for Vietnamese women may achieve the benefits seen in other populations without any increase in harm.