The New Zealand Guidelines Group recommended that women without additional risk factors with a previous caesarean section should be offered a vaginal birth after caesarean section (VBAC).1 However, a recent large cohort study has shown that an elective repeat caesarean section (RCS) significantly reduced the risk of fetal death or infant death compared to those women who had a VBAC.2 An earlier survey of Australian and New Zealand (NZ) obstetricians found that 96% agreed that an option of VBAC should be offered although only 40% agreed that it was the safest option for the woman.3 Our study from a large tertiary hospital in a low socioeconomic area is important because an uncertainty in management has persisted. There has been one other recent NZ study4 which found a lower VBAC rate in Asians and those with a high Body Mass Index (BMI). However, they did not include other variables which include previous vaginal deliveries, augmentation epidural analgesia and maternal morbidity. An understanding of local factors that may affect women's chances of a successful VBAC will assist informed consent and counselling. The aims of this study were to determine the VBAC rate and evaluate the importance of ethnicity, body mass index (BMI), parity, previous vaginal delivery and the conduct of labour as factors that may determine the success of a VBAC. Serious maternal and perinatal outcomes were also determined. Methods Study design—A retrospective study was conducted at MMH, a tertiary referral centre in South Auckland, New Zealand. Ethics approval was obtained from the Northern X Ethics Committee (NTX/10/EXP/003). Included in the study were women who delivered a singleton baby in 2008 and 2009, with at least one previous lower segment caesarean section who chose to have a VBAC and were assessed eligible by an obstetrician for a TOL. TOL is defined as a planned labour after a previous caesarean section delivery in an informed woman. A previous VBAC is defined as a woman who had a vaginal delivery after a caesarean section and a previous NVD were those who had a vaginal delivery prior to a caesarean section. Exclusions were an elective caesarean delivery due to the woman's personal choice, non-vertex presentation, gestation of less than 24 completed weeks, classical caesarean section, severe pre-eclampsia, placenta praevia, maternal medical condition necessitating urgent delivery and multiple gestation. Severe pre-eclampsia is the presence of one of the following symptoms or signs: systolic BP ≥ 160 mmHg or diastolic BP of ≥ 110 mmHg; proteinuria> 5g in a 24-hour collection; pulmonary oedema or cyanosis; oliguria (< 400 ml in 24 h); persistent headaches; epigastric pain and/or impaired liver function; thrombocytopenia; oligohydramnios; decreased fetal growth; or placental abruption. Data sources—Data was collected from the Delivery Suite Birth Register, electronic hospital databases (Healthware, Concerto and Patient Information System) and selected case notes by a single data collector (AMvdM). Data on women with previous caesarean sections were extracted from Healthware and cross-checked with the data from the Birth Register. Data on age, weight, height, ethnicity, obstetric and medical history, intrapartum events and neonatal outcomes were obtained from the relevant databases and medical records where the data was absent in Healthware. Indications for previous caesarean delivery, previous uterine incision (if available), history of previous vaginal delivery (defined as a delivery preceding a previous caesarean section)5,6 or previous VBAC (defined as any vaginal delivery after a caesarean section) and where there was a presence of maternal disease (hypertension, diabetes, asthma, epilepsy, renal disease, thyroid disease or collagen vascular conditions) was recorded from the Healthware and Concerto databases. Obstetric factors related to the present pregnancy such as spontaneous labour, induction, augmentation, epidural use and VBAC success, were also recorded in a Microsoft Excel 2007 spreadsheet. Neonatal factors including the birth weight, gestational age, gender, 5-minute Apgar score were also documented. Neonatal and maternal complications were also collected from Healthware and Concerto. There were 16,051 deliveries in Counties Manukau DHB in 2008 and 2009 and all women with a previous caesarean section were referred to MMH for delivery. Of the 16,051 women, 1543 (9.6%) had one or more previous caesarean section deliveries. Women were excluded due to an elective caesarean delivery (679), of which was usually due to the woman's preference after discussion with a specialist, non-vertex presentation (33), unclear plan (6), medical conditions (6), placenta praevia (5), multiple pregnancy (4), and classical incision (4). After exclusions, 806 (52.2%) women were eligible for analysis. Statistical analysis—Analysis was performed using SAS v9.1 (SAS Institute, Cary, NC) and odds ratios (OR) was estimated using logistic regression. Statistical analysis was conducted using Chi-squared for categorical variables and Student's t-test for normally distributed continuous variables. Factors potentially affecting VBAC success were initially analysed in univariable analysis. Following this, multivariable logistic regression was performed to evaluate the likelihood of successful VBAC during a subsequent trial of labor controlling for gender, ethnicity, previous vaginal delivery, previous VBAC, maternal BMI, induction of labour and epidural use. Parity was not included in the multivariable model because of its close relationship to previous vaginal delivery and previous VBAC. Augmentation was excluded from the multivariable model due to its close association with IOL, which was found to be the stronger of the two variables in relation to VBAC. Confidence intervals and p values of ≤0.05 were considered statistically significant. Results Of the 806 women who planned a vaginal delivery, 592 (73.4%) had a vaginal delivery and the remaining 214 (26.6%) had a RCS delivery, albeit an emergency one. The factors associated with a successful or failed VBAC at the univariable level were parity, previous vaginal delivery, previous VBAC, induction of labour, augmentation and epidural (Table 1). Table 1. Variables associated with a VBAC or a RCS delivery Variables VBAC (n=592) RCS (n=214) OR (95%CI) P value Age (mean)17–44 years 30.2±5.3 30.1±5.7 30.4±5.2 p=0.71, p=0.48 Sex (n/%) Male (409/50.7) Female (397/49.3) 292 (50.7) 300 (49.3) 109 (50.9) 105 (49.1) 0.99 (0.72–1.35) 1.00 p=0.95 Ethnicity (n/%) missing 11 European (164/20.3) Maori (162/20.1) Pacific (355/44)∗ Other (114/14.1)** 120 (20.6) 128 (22.0) 254 (43.6) 81 (13.9) 44 (20.8) 34 (16.0) 101 (47.6) 33 (15.6) 1.00 1.38 (0.83–2.30) 0.92 (0.61–1.40) 0.90 (0.53–1.53) p=0.30 Parity (n/%) missing=0 1 (392/48.6) 2 (172/21.3) 3+(242/30) 241 (40.71) 142 (24) 209 (35.30) 151 (70.6) 30 (14) 33 (15.4) 1.00 2.97 (1.90–4.62) 3.97 (2.61–6.04) p<0.0001 Gestation (n/%)missing=0 ≤37 38 39 40 41 ≥42 63 (10.6) 54 (9.1) 98 (16.6) 198 (33.4) 121 (20.4) 58 (9.8) 20 (9.3) 24 (11.2) 39 (18.2) 50 (23.4) 51 (23.8) 30 (14) 0.80 (0.44–1.44) 0.57 (0.32–1.01) 0.64 (0.39–1.03) 1.00 0.60 (0.38–0.94) 0.49 (0.29–0.84) p=0.08 Previous NVD (n/%) No(417/51.7) Yes(389/48.3) 257(43.4) 335(56.6) 160 (74.8) 54 (25.2) 1.00 3.86 (2.73–5.47) p<0.0001 Previous VBAC (n/%)missing=3 No(521/64.7) Yes(282/35) 332 (56.4) 257 (43.6) 189 (88.3) 25 (11.7) 1.00 5.85 (3.74–9.16) p<0.0001 Induction of labour (n/%)missing=1 Spontaneous(654/81.1) IOL(151/18.7) 496 (83.8) 96 (16.2) 158 (74.2) 55 (25.8) 1.00 0.56 (0.38,0.81) p=0.002 Maternal BMI (n/%)missing=116 <25(167/24.2) 25–29.99(161/23.3) 30+(362/52.5) 131 (26.1) 112 (22.3) 259 (51.6) 36 (19.2) 49 (26.1) 103 (54.7) 1.00 0.69 (0.45–1.07) 0.63 (0.38–1.03) p=0.15 Augmentation (n/%)missing=7 No(652/80.9) Yes(147/18.2) 492 (83.1) 97 (16.4) 160 (74.8) 50 (23.4) 1.00 0.63 (0.43–0.93) p=0.019 Epidural (n/%) No (510/63.3) Yes (296/36.) 451 (76.2) 141 (23.8) 59 (27.6) 155 (72.4) 1.00 0.12 (0.08–0.17) p<0.0001 ∗ Tongan, Samoan, Fijian, Cook Island Maori, Tokelauan, Niuean, and Other Pacific Islander. ** Other Asian , Indian , Vietnamese , Sri Lankan , Chinese and Other. Of the 389 (48.2%) women who had a previous vaginal delivery, 335 (86%) had a vaginal delivery again. Of the 417 (51.7%) women who did not have a previous vaginal delivery, 257 (61.6%) had a vaginal delivery. Of the 282 (35.0%) women who had a previous VBAC, 257 (91.1%) delivered vaginally again whilst 332 (63.7%) of the women who had not had a VBAC before had a vaginal delivery. Thus the increased Odds Ratio associated with having another successful VBAC was 5.85 (95%CI 3.74–9.16). In multivariable analysis several factors remained significantly associated with the successful completion of a VBAC (Table 2). A previous VBAC was still associated with an increased VBAC success after adjustment for confounding factors (OR 3.69; 95%CI 1.83–7.43). Whilst a similar effect was seen in univariable analyses for a previous vaginal delivery, this did not remain significantly associated with a successful VBAC in multivariable analysis. After adjusting for confounding factors the Odds Ratios in the overweight BMI category was 0.44 (95%CI 0.25–0.79) and obese BMI category was 0.43 (95%CI 0.24–0.78). In 116(14%) eligible women data on height, weight, and BMI was not available. In women of single parity (P1) however, VBAC success was reduced after multivariate analysis in overweight (OR 0.47; 95%CI 0.24–0.91) and obese women(OR 0.43;95%CI 0.22–0.86). Spontaneous labour was not associated with greater VBAC success (OR 0.81; 95%CI 0.50–1.31) and this was also found in P1 (OR 1.09; 95%CI 0.58–2.02). Labour augmentation statistically reduced VBAC success in the univariable analysis (OR 0.63; 95%CI 0.43–0.93). Epidural anaesthesia was associated with a reduced VBAC success (OR 0.18; 95%CI 0.12–0.28) and this was also found in P1 (OR 0.20; 95%CI 0.12–0.34). No effect was seen in relation to ethnicity in univariable analysis or after controlling for other factors associated with VBAC. Those with an unsuccessful VBAC were more likely to deliver a baby with a birth weight ≤ 3000g (OR 0.65; 95%CI 0.44–0.97) or ≥4001g (OR 0.47; 95%CI 0.32–0.69). Of the 592 women who had a VBAC, 29 (4.9%) had a postpartum haemorrhage (PPH) of>1000ml, 18 (3.0%) had a 3rd or 4th degree vaginal tear, 12 (2.0%) had a manual removal of placenta (MROP) and 2(0.3%) women had a peripartum hysterectomy due to PPH. Of the 214 women who had an emergency caesarean section delivery, 19 (8.9%) had an estimated blood loss of>1000ml and none had a hysterectomy. Infection morbidity was similar in the two groups (0.3% of VBAC and 1% of RCS). None of these outcomes was statistically significant. There was no uterine rupture in this study, which was not adequately powered to investigate this outcome. Most of the repeat caesarean sections (105, 48%) were performed for failure to progress either in the first or second stage of labour, followed by fetal distress (56, 26%) (Table 3). Table 2. Multivariate logistic regression of variables for all women and women of single parity Multivariable regression All subjects OR (95%CI) Para 1 OR (95%CI) Ethnicity (missing=11) European Maori Pacific Other P=0.23 1.00 1.21 (0.64–2.26) 0.72 (0.40–1.28) 1.10 (0.58–2.08) p=0.21 1.00 1.36 (0.63–2.95) 0.68 (0.34–1.36) 1.19 (0.59–2.41) Previous vaginal delivery No Yes p=0.35 1.00 1.33 (0.74–2.41) Previous VBAC (missing=3) No Yes p=0.0003 1.00 3.69 (1.83–7.43) Maternal BMI (missing=116) <25 25–29.99 30+ p=0.091 1.00 0.44 (0.25–0.79) 0.43 (0.24–0.78) p=0.028 1.00 0.47 (0.24–0.91) 0.43 (0.22–0.86) Induction of labour (missing=1) Spontaneous Induction p=0.39 1.00 0.81 (0.50–1.31) p=0.79 1.00 1.09 (0.58–2.02) Epidural No Yes p<0.0001 1.00 0.18 (0.12–0.28) p<0.0001 1.00 0.20 (0.12–0.34) Table 3. Indications for repeat caesarean section Indications for repeat caesarean section N (%) Failure to progress, 1st stage Fetal distress Failure to progress, 2nd stage Failed induction of labour Antepartum haemorrhage Failed instrumental delivery Other Not specified in case notes 85 (39.2) 56 (25.8) 20 (9.2) 9 (4.2) 5 (2.3) 5 (2.3) 7 (3.2) 27 (12.4) Total 217 (100) Discussion The number of caesarean section births in NZ has been increasing in the last two decades and in 2009, 15.3% of all first deliveries in NZ were by caesarean section.6 MMH has, for many years, had a lower caesarean section rate compared to other referral hospitals in NZ8 and this may be because about 57% of all deliveries annually were to women of Maori and Pacific ethnicities who had higher parities than other groups.9 It may also be due to a different clinical practice at MMH.10 But whereas the cyclical debate on the reasons for the increase in caesarean section deliveries has continued, what has come to the fore is the discussion on the subsequent management of these women given the recent publication of the Australian trial by Crowther et al showing a significantly higher maternal and neonatal morbidity in women having a VBAC compared to those who have elective RCS.2 They found that 53% of the 2323 women with a previous caesarean section elected to have a TOL and only 57% of them achieved a VBAC. In our study, 806 (52.2%) of women agreed to a TOL and 592 (73.5%) achieved a VBAC. The Auckland Hospital study1 cohort were women without normal deliveries therefore a TOL and VBAC rate cannot be compared with ours. The MMH VBAC rate of 73% lies within the 60–80% VBAC rate of studies as reported in the NZ guidelines although reported VBAC rates reported in the literature have ranged between 25 to 93%.5,12,13 We found the VBAC rates for women having a first delivery after a caesarean section and for those who had a previous VBAC were 64% and 91% respectively. These findings were similar to those of an American study by Elkousy et al6 which found a VBAC rate for women of single and multi-parity were 65% and 83–94% respectively. There was no difference in VBAC success rates between Pacific, Maori and European groups, which is similar to the findings of a the Auckland Hospital Study.1 American studies had shown women of Caucasian ethnicity had a better chance of a VBAC compared to other ethnic groups.5,12,14,15 Increasing parity, previous vaginal delivery and a previous VBAC have been shown to be strong predictors for a successful VBAC7,12–14,16,17 and this was also evident from our findings. The increasing VBAC rate with the increase in parity is most likely an association with an increasing number of previous vaginal deliveries, whether they were before or after the original caesarean section. Whereas a previous vaginal delivery and a previous VBAC significantly increased the chances of a VBAC, the difference remains significant only for the latter group on multivariable analysis. Landon et al12 found in their large study, a significantly higher VBAC rate of 87% in women with a previous vaginal birth compared with 61% in women without such a history. Although they did not find a VBAC rate difference between those who delivered vaginally before or after a caesarean section, a smaller study17 confirmed the difference found in our study. The likelihood of a VBAC is less likely in women with increased BMI and was significant in women of single parity who had not had a vaginal delivery. This confirms similar findings in other studies5,12,19,20 and the Auckland Hospital study.1 Data for the calculation of BMI came from the booking weight and height and 14% of that data was missing. Although this may not be the most reliable data, a significant effect in the multivariable analysis suggests that this effect is real. In the conduct of labour, we found that induction of labour was not statistically significant but augmentation of labour and an epidural analgesia were found to be significant factors in reducing the chances of a VBAC compared to those who did not have labour augmentation or an epidural analgesia. These findings have also been found in other studies.2,16,12,22 The reason for this could be that once the labour is dysfunctional requiring augmentation and an epidural, then the chances of a VBAC diminishes. Our results demonstrated a reduction in VBAC success in babies that ended up weighing less than 3001g or more than 4001g (p=0.0005). In a study by Peaceman et al,23 only 38% had a VBAC if the birth weight exceeded the preceding pregnancy birth weight by more than 500 grams. Elkousy et al5 demonstrated a systematic decline in VBAC success with increasing birth weight (68% <4000g; 52% 4000–4249g; 45% 4250–4500g and 38%>4500g). An explanation as to why babies of less than 3000g were more likely to be delivered by caesarean section could be because of prematurity, intrauterine growth restriction and the increase likelihood of fetal distress and malpresentation that accompany these scenarios. The indications for RCS had similar percentages to the background rate of caesarean sections for failure to progress and fetal distress at MMH.6 There was no uterine rupture in our study, which was not adequately powered to investigate the incidence which has been estimated to be 0.2–1.5%.1 The considerable maternal morbidity is a concern in those who had a VBAC and these included PPH (5%), 3rd/4th degree vaginal tears (3%), MROP (2%) and the two women who had hysterectomies. The Crowther et al study2 confirmed the high rate of PPH in the VBAC group but it seems that there was a higher total morbidity rate (11%) in our study. Thirty women (5%) needed procedures in the operating theatre to have either a vaginal repair or MROP. The Crowther et al study2 did not mention MROP and had only two cases of “perineal haematomas”. Neonatal morbidity defined as Apgar scores ≤7 was higher in the caesarean group and may be explained by the fact that these were not elective caesarean section procedures. The Crowther et al study2 found significant neonatal morbidity but there was no difference in Apgar scores at 5 minutes between the planned VBAC and elective caesarean section groups. The main limitation of our study is its retrospective nature, which meant that some of the variables could not be collected due to the incompleteness of the database and case notes. For example, 14% of the BMI data was not available. The strength of our study was in the large number of women. Our study's findings, would contribute to the continuing discussion on the management of women with previous caesarean sections in the NZ context. We have identified factors that affect the VBAC rate in our setting and a prospective study may identify or clarify more. The significant maternal morbidity in the VBAC group is a concern and we will need to address these if a TOL and VBAC are to be made safer and continue to be offered as a viable option to women.

The New Zealand Guidelines Group recommended that women without additional risk factors with a previous caesarean section should be offered a vaginal birth after caesarean section (VBAC).1 However, a recent large cohort study has shown that an elective repeat caesarean section (RCS) significantly reduced the risk of fetal death or infant death compared to those women who had a VBAC.2 An earlier survey of Australian and New Zealand (NZ) obstetricians found that 96% agreed that an option of VBAC should be offered although only 40% agreed that it was the safest option for the woman.3 Our study from a large tertiary hospital in a low socioeconomic area is important because an uncertainty in management has persisted. There has been one other recent NZ study4 which found a lower VBAC rate in Asians and those with a high Body Mass Index (BMI). However, they did not include other variables which include previous vaginal deliveries, augmentation epidural analgesia and maternal morbidity. An understanding of local factors that may affect women's chances of a successful VBAC will assist informed consent and counselling. The aims of this study were to determine the VBAC rate and evaluate the importance of ethnicity, body mass index (BMI), parity, previous vaginal delivery and the conduct of labour as factors that may determine the success of a VBAC. Serious maternal and perinatal outcomes were also determined. Methods Study design—A retrospective study was conducted at MMH, a tertiary referral centre in South Auckland, New Zealand. Ethics approval was obtained from the Northern X Ethics Committee (NTX/10/EXP/003). Included in the study were women who delivered a singleton baby in 2008 and 2009, with at least one previous lower segment caesarean section who chose to have a VBAC and were assessed eligible by an obstetrician for a TOL. TOL is defined as a planned labour after a previous caesarean section delivery in an informed woman. A previous VBAC is defined as a woman who had a vaginal delivery after a caesarean section and a previous NVD were those who had a vaginal delivery prior to a caesarean section. Exclusions were an elective caesarean delivery due to the woman's personal choice, non-vertex presentation, gestation of less than 24 completed weeks, classical caesarean section, severe pre-eclampsia, placenta praevia, maternal medical condition necessitating urgent delivery and multiple gestation. Severe pre-eclampsia is the presence of one of the following symptoms or signs: systolic BP ≥ 160 mmHg or diastolic BP of ≥ 110 mmHg; proteinuria> 5g in a 24-hour collection; pulmonary oedema or cyanosis; oliguria (< 400 ml in 24 h); persistent headaches; epigastric pain and/or impaired liver function; thrombocytopenia; oligohydramnios; decreased fetal growth; or placental abruption. Data sources—Data was collected from the Delivery Suite Birth Register, electronic hospital databases (Healthware, Concerto and Patient Information System) and selected case notes by a single data collector (AMvdM). Data on women with previous caesarean sections were extracted from Healthware and cross-checked with the data from the Birth Register. Data on age, weight, height, ethnicity, obstetric and medical history, intrapartum events and neonatal outcomes were obtained from the relevant databases and medical records where the data was absent in Healthware. Indications for previous caesarean delivery, previous uterine incision (if available), history of previous vaginal delivery (defined as a delivery preceding a previous caesarean section)5,6 or previous VBAC (defined as any vaginal delivery after a caesarean section) and where there was a presence of maternal disease (hypertension, diabetes, asthma, epilepsy, renal disease, thyroid disease or collagen vascular conditions) was recorded from the Healthware and Concerto databases. Obstetric factors related to the present pregnancy such as spontaneous labour, induction, augmentation, epidural use and VBAC success, were also recorded in a Microsoft Excel 2007 spreadsheet. Neonatal factors including the birth weight, gestational age, gender, 5-minute Apgar score were also documented. Neonatal and maternal complications were also collected from Healthware and Concerto. There were 16,051 deliveries in Counties Manukau DHB in 2008 and 2009 and all women with a previous caesarean section were referred to MMH for delivery. Of the 16,051 women, 1543 (9.6%) had one or more previous caesarean section deliveries. Women were excluded due to an elective caesarean delivery (679), of which was usually due to the woman's preference after discussion with a specialist, non-vertex presentation (33), unclear plan (6), medical conditions (6), placenta praevia (5), multiple pregnancy (4), and classical incision (4). After exclusions, 806 (52.2%) women were eligible for analysis. Statistical analysis—Analysis was performed using SAS v9.1 (SAS Institute, Cary, NC) and odds ratios (OR) was estimated using logistic regression. Statistical analysis was conducted using Chi-squared for categorical variables and Student's t-test for normally distributed continuous variables. Factors potentially affecting VBAC success were initially analysed in univariable analysis. Following this, multivariable logistic regression was performed to evaluate the likelihood of successful VBAC during a subsequent trial of labor controlling for gender, ethnicity, previous vaginal delivery, previous VBAC, maternal BMI, induction of labour and epidural use. Parity was not included in the multivariable model because of its close relationship to previous vaginal delivery and previous VBAC. Augmentation was excluded from the multivariable model due to its close association with IOL, which was found to be the stronger of the two variables in relation to VBAC. Confidence intervals and p values of ≤0.05 were considered statistically significant. Results Of the 806 women who planned a vaginal delivery, 592 (73.4%) had a vaginal delivery and the remaining 214 (26.6%) had a RCS delivery, albeit an emergency one. The factors associated with a successful or failed VBAC at the univariable level were parity, previous vaginal delivery, previous VBAC, induction of labour, augmentation and epidural (Table 1). Table 1. Variables associated with a VBAC or a RCS delivery Variables VBAC (n=592) RCS (n=214) OR (95%CI) P value Age (mean)17–44 years 30.2±5.3 30.1±5.7 30.4±5.2 p=0.71, p=0.48 Sex (n/%) Male (409/50.7) Female (397/49.3) 292 (50.7) 300 (49.3) 109 (50.9) 105 (49.1) 0.99 (0.72–1.35) 1.00 p=0.95 Ethnicity (n/%) missing 11 European (164/20.3) Maori (162/20.1) Pacific (355/44)∗ Other (114/14.1)** 120 (20.6) 128 (22.0) 254 (43.6) 81 (13.9) 44 (20.8) 34 (16.0) 101 (47.6) 33 (15.6) 1.00 1.38 (0.83–2.30) 0.92 (0.61–1.40) 0.90 (0.53–1.53) p=0.30 Parity (n/%) missing=0 1 (392/48.6) 2 (172/21.3) 3+(242/30) 241 (40.71) 142 (24) 209 (35.30) 151 (70.6) 30 (14) 33 (15.4) 1.00 2.97 (1.90–4.62) 3.97 (2.61–6.04) p<0.0001 Gestation (n/%)missing=0 ≤37 38 39 40 41 ≥42 63 (10.6) 54 (9.1) 98 (16.6) 198 (33.4) 121 (20.4) 58 (9.8) 20 (9.3) 24 (11.2) 39 (18.2) 50 (23.4) 51 (23.8) 30 (14) 0.80 (0.44–1.44) 0.57 (0.32–1.01) 0.64 (0.39–1.03) 1.00 0.60 (0.38–0.94) 0.49 (0.29–0.84) p=0.08 Previous NVD (n/%) No(417/51.7) Yes(389/48.3) 257(43.4) 335(56.6) 160 (74.8) 54 (25.2) 1.00 3.86 (2.73–5.47) p<0.0001 Previous VBAC (n/%)missing=3 No(521/64.7) Yes(282/35) 332 (56.4) 257 (43.6) 189 (88.3) 25 (11.7) 1.00 5.85 (3.74–9.16) p<0.0001 Induction of labour (n/%)missing=1 Spontaneous(654/81.1) IOL(151/18.7) 496 (83.8) 96 (16.2) 158 (74.2) 55 (25.8) 1.00 0.56 (0.38,0.81) p=0.002 Maternal BMI (n/%)missing=116 <25(167/24.2) 25–29.99(161/23.3) 30+(362/52.5) 131 (26.1) 112 (22.3) 259 (51.6) 36 (19.2) 49 (26.1) 103 (54.7) 1.00 0.69 (0.45–1.07) 0.63 (0.38–1.03) p=0.15 Augmentation (n/%)missing=7 No(652/80.9) Yes(147/18.2) 492 (83.1) 97 (16.4) 160 (74.8) 50 (23.4) 1.00 0.63 (0.43–0.93) p=0.019 Epidural (n/%) No (510/63.3) Yes (296/36.) 451 (76.2) 141 (23.8) 59 (27.6) 155 (72.4) 1.00 0.12 (0.08–0.17) p<0.0001 ∗ Tongan, Samoan, Fijian, Cook Island Maori, Tokelauan, Niuean, and Other Pacific Islander. ** Other Asian , Indian , Vietnamese , Sri Lankan , Chinese and Other. Of the 389 (48.2%) women who had a previous vaginal delivery, 335 (86%) had a vaginal delivery again. Of the 417 (51.7%) women who did not have a previous vaginal delivery, 257 (61.6%) had a vaginal delivery. Of the 282 (35.0%) women who had a previous VBAC, 257 (91.1%) delivered vaginally again whilst 332 (63.7%) of the women who had not had a VBAC before had a vaginal delivery. Thus the increased Odds Ratio associated with having another successful VBAC was 5.85 (95%CI 3.74–9.16). In multivariable analysis several factors remained significantly associated with the successful completion of a VBAC (Table 2). A previous VBAC was still associated with an increased VBAC success after adjustment for confounding factors (OR 3.69; 95%CI 1.83–7.43). Whilst a similar effect was seen in univariable analyses for a previous vaginal delivery, this did not remain significantly associated with a successful VBAC in multivariable analysis. After adjusting for confounding factors the Odds Ratios in the overweight BMI category was 0.44 (95%CI 0.25–0.79) and obese BMI category was 0.43 (95%CI 0.24–0.78). In 116(14%) eligible women data on height, weight, and BMI was not available. In women of single parity (P1) however, VBAC success was reduced after multivariate analysis in overweight (OR 0.47; 95%CI 0.24–0.91) and obese women(OR 0.43;95%CI 0.22–0.86). Spontaneous labour was not associated with greater VBAC success (OR 0.81; 95%CI 0.50–1.31) and this was also found in P1 (OR 1.09; 95%CI 0.58–2.02). Labour augmentation statistically reduced VBAC success in the univariable analysis (OR 0.63; 95%CI 0.43–0.93). Epidural anaesthesia was associated with a reduced VBAC success (OR 0.18; 95%CI 0.12–0.28) and this was also found in P1 (OR 0.20; 95%CI 0.12–0.34). No effect was seen in relation to ethnicity in univariable analysis or after controlling for other factors associated with VBAC. Those with an unsuccessful VBAC were more likely to deliver a baby with a birth weight ≤ 3000g (OR 0.65; 95%CI 0.44–0.97) or ≥4001g (OR 0.47; 95%CI 0.32–0.69). Of the 592 women who had a VBAC, 29 (4.9%) had a postpartum haemorrhage (PPH) of>1000ml, 18 (3.0%) had a 3rd or 4th degree vaginal tear, 12 (2.0%) had a manual removal of placenta (MROP) and 2(0.3%) women had a peripartum hysterectomy due to PPH. Of the 214 women who had an emergency caesarean section delivery, 19 (8.9%) had an estimated blood loss of>1000ml and none had a hysterectomy. Infection morbidity was similar in the two groups (0.3% of VBAC and 1% of RCS). None of these outcomes was statistically significant. There was no uterine rupture in this study, which was not adequately powered to investigate this outcome. Most of the repeat caesarean sections (105, 48%) were performed for failure to progress either in the first or second stage of labour, followed by fetal distress (56, 26%) (Table 3). Table 2. Multivariate logistic regression of variables for all women and women of single parity Multivariable regression All subjects OR (95%CI) Para 1 OR (95%CI) Ethnicity (missing=11) European Maori Pacific Other P=0.23 1.00 1.21 (0.64–2.26) 0.72 (0.40–1.28) 1.10 (0.58–2.08) p=0.21 1.00 1.36 (0.63–2.95) 0.68 (0.34–1.36) 1.19 (0.59–2.41) Previous vaginal delivery No Yes p=0.35 1.00 1.33 (0.74–2.41) Previous VBAC (missing=3) No Yes p=0.0003 1.00 3.69 (1.83–7.43) Maternal BMI (missing=116) <25 25–29.99 30+ p=0.091 1.00 0.44 (0.25–0.79) 0.43 (0.24–0.78) p=0.028 1.00 0.47 (0.24–0.91) 0.43 (0.22–0.86) Induction of labour (missing=1) Spontaneous Induction p=0.39 1.00 0.81 (0.50–1.31) p=0.79 1.00 1.09 (0.58–2.02) Epidural No Yes p<0.0001 1.00 0.18 (0.12–0.28) p<0.0001 1.00 0.20 (0.12–0.34) Table 3. Indications for repeat caesarean section Indications for repeat caesarean section N (%) Failure to progress, 1st stage Fetal distress Failure to progress, 2nd stage Failed induction of labour Antepartum haemorrhage Failed instrumental delivery Other Not specified in case notes 85 (39.2) 56 (25.8) 20 (9.2) 9 (4.2) 5 (2.3) 5 (2.3) 7 (3.2) 27 (12.4) Total 217 (100) Discussion The number of caesarean section births in NZ has been increasing in the last two decades and in 2009, 15.3% of all first deliveries in NZ were by caesarean section.6 MMH has, for many years, had a lower caesarean section rate compared to other referral hospitals in NZ8 and this may be because about 57% of all deliveries annually were to women of Maori and Pacific ethnicities who had higher parities than other groups.9 It may also be due to a different clinical practice at MMH.10 But whereas the cyclical debate on the reasons for the increase in caesarean section deliveries has continued, what has come to the fore is the discussion on the subsequent management of these women given the recent publication of the Australian trial by Crowther et al showing a significantly higher maternal and neonatal morbidity in women having a VBAC compared to those who have elective RCS.2 They found that 53% of the 2323 women with a previous caesarean section elected to have a TOL and only 57% of them achieved a VBAC. In our study, 806 (52.2%) of women agreed to a TOL and 592 (73.5%) achieved a VBAC. The Auckland Hospital study1 cohort were women without normal deliveries therefore a TOL and VBAC rate cannot be compared with ours. The MMH VBAC rate of 73% lies within the 60–80% VBAC rate of studies as reported in the NZ guidelines although reported VBAC rates reported in the literature have ranged between 25 to 93%.5,12,13 We found the VBAC rates for women having a first delivery after a caesarean section and for those who had a previous VBAC were 64% and 91% respectively. These findings were similar to those of an American study by Elkousy et al6 which found a VBAC rate for women of single and multi-parity were 65% and 83–94% respectively. There was no difference in VBAC success rates between Pacific, Maori and European groups, which is similar to the findings of a the Auckland Hospital Study.1 American studies had shown women of Caucasian ethnicity had a better chance of a VBAC compared to other ethnic groups.5,12,14,15 Increasing parity, previous vaginal delivery and a previous VBAC have been shown to be strong predictors for a successful VBAC7,12–14,16,17 and this was also evident from our findings. The increasing VBAC rate with the increase in parity is most likely an association with an increasing number of previous vaginal deliveries, whether they were before or after the original caesarean section. Whereas a previous vaginal delivery and a previous VBAC significantly increased the chances of a VBAC, the difference remains significant only for the latter group on multivariable analysis. Landon et al12 found in their large study, a significantly higher VBAC rate of 87% in women with a previous vaginal birth compared with 61% in women without such a history. Although they did not find a VBAC rate difference between those who delivered vaginally before or after a caesarean section, a smaller study17 confirmed the difference found in our study. The likelihood of a VBAC is less likely in women with increased BMI and was significant in women of single parity who had not had a vaginal delivery. This confirms similar findings in other studies5,12,19,20 and the Auckland Hospital study.1 Data for the calculation of BMI came from the booking weight and height and 14% of that data was missing. Although this may not be the most reliable data, a significant effect in the multivariable analysis suggests that this effect is real. In the conduct of labour, we found that induction of labour was not statistically significant but augmentation of labour and an epidural analgesia were found to be significant factors in reducing the chances of a VBAC compared to those who did not have labour augmentation or an epidural analgesia. These findings have also been found in other studies.2,16,12,22 The reason for this could be that once the labour is dysfunctional requiring augmentation and an epidural, then the chances of a VBAC diminishes. Our results demonstrated a reduction in VBAC success in babies that ended up weighing less than 3001g or more than 4001g (p=0.0005). In a study by Peaceman et al,23 only 38% had a VBAC if the birth weight exceeded the preceding pregnancy birth weight by more than 500 grams. Elkousy et al5 demonstrated a systematic decline in VBAC success with increasing birth weight (68% <4000g; 52% 4000–4249g; 45% 4250–4500g and 38%>4500g). An explanation as to why babies of less than 3000g were more likely to be delivered by caesarean section could be because of prematurity, intrauterine growth restriction and the increase likelihood of fetal distress and malpresentation that accompany these scenarios. The indications for RCS had similar percentages to the background rate of caesarean sections for failure to progress and fetal distress at MMH.6 There was no uterine rupture in our study, which was not adequately powered to investigate the incidence which has been estimated to be 0.2–1.5%.1 The considerable maternal morbidity is a concern in those who had a VBAC and these included PPH (5%), 3rd/4th degree vaginal tears (3%), MROP (2%) and the two women who had hysterectomies. The Crowther et al study2 confirmed the high rate of PPH in the VBAC group but it seems that there was a higher total morbidity rate (11%) in our study. Thirty women (5%) needed procedures in the operating theatre to have either a vaginal repair or MROP. The Crowther et al study2 did not mention MROP and had only two cases of “perineal haematomas”. Neonatal morbidity defined as Apgar scores ≤7 was higher in the caesarean group and may be explained by the fact that these were not elective caesarean section procedures. The Crowther et al study2 found significant neonatal morbidity but there was no difference in Apgar scores at 5 minutes between the planned VBAC and elective caesarean section groups. The main limitation of our study is its retrospective nature, which meant that some of the variables could not be collected due to the incompleteness of the database and case notes. For example, 14% of the BMI data was not available. The strength of our study was in the large number of women. Our study's findings, would contribute to the continuing discussion on the management of women with previous caesarean sections in the NZ context. We have identified factors that affect the VBAC rate in our setting and a prospective study may identify or clarify more. The significant maternal morbidity in the VBAC group is a concern and we will need to address these if a TOL and VBAC are to be made safer and continue to be offered as a viable option to women.