This study employed a descriptive study design utilising a retrospective record review and analysed Maternity Case Records. This approach enabled structured analysis of routinely collected clinical data from maternity case records without altering patient care.

The study was conducted at Zithulele Hospital, a 147-bed rural district hospital situated in the King Sabata Dalindyebo subdistrict of the OR Tambo District, Eastern Cape, South Africa. The hospital provides 24-h maternity care with on-site emergency surgical capacity and emergency blood transfusion services. Resource constraints include staff shortages, limited access to senior clinicians and emergency transport delays, which reflect the realities of many South African rural, district hospitals. The hospital conducts approximately 2100 deliveries annually, with a CD rate of 22% – 25%.¹⁶

Zithulele Hospital serves a predominantly rural, low-income population and its obstetric service is carried out by medical officers and/or community-service doctors and midwives, with one family physician overseeing clinical services. Access to hospital remains difficult and expensive for a large proportion of the served population with some reasons being poorly developed road infrastructure and limited household income. This setting provided a unique opportunity to assess delivery outcomes and VBAC feasibility within a rural district health system.

The study population is made up of records of pregnant women arriving at hospital from 01 January to 30 September 2021 and having had a previous CD in a previous pregnancy (Robson score 5).²

Inclusion criteria comprised: (1) women with a singleton pregnancy and at least one prior CD; (2) delivery at Zithulele Hospital (vaginally or by CD); and (3) live foetus upon arrival to hospital. Exclusion criteria comprised of: (1) women with absolute contraindications to vaginal delivery in the delivery being assessed (e.g. transverse lie, major placenta praevia, etc.); and (2) incomplete or missing delivery records.

A minimum sample size of 179 records was calculated to provide a representative estimate of outcomes with a 95% confidence level and 5% margin of error. To account for potential exclusions, 20 additional records were reviewed, giving a final dataset of 199 records, of which 195 met inclusion criteria.¹⁷ All eligible records meeting the criteria during the study period were included.

Data were extracted retrospectively from maternity case records using a standardised data collection form utilising KoboToolbox software (www.kobotoolbox.org). The collection form captured variables across various domains:

Data were captured by the primary investigator into a secure, password-protected spreadsheet and de-identified prior to analysis. Data were exported into Microsoft Excel (2023) for cleaning and analysis. Range and consistency checks were performed to identify outliers, missing values or implausible entries.

Descriptive statistics were calculated for demographic and clinical characteristics. Continuous variables were summarised as means, medians, standard deviations (s.d.) and interquartile ranges (IQR) as appropriate, while categorical variables were expressed as frequencies and percentages.

Comparative analyses were performed to identify factors associated with successful VBAC. Analytical functions were applied including the Chi-square tests for categorical variables (e.g. previous vaginal delivery, number of prior CD, comorbidities). Binary logistic regression (univariate) was used to assess associations between VBAC success and independent predictors, including parity, stage of labour on admission and presence of previous vaginal delivery. A p-value < 0.05 was considered statistically significant for all analyses.

As this was a retrospective record review using anonymised data, a waiver of informed consent was granted. All data were handled confidentially, stored securely on password-protected devices and used solely for research purposes.

Ethical approval was obtained from the Walter Sisulu University Faculty of Health Sciences Research Ethics and Biosafety Committee (Reference: 228/2024) and the Eastern Cape Department of Health Research and Ethics Committee. Institutional permission was granted by the Zithulele Hospital Chief Executive Officer.

Among women with previous CD, 46% (n = 90) achieved successful VBAC, while 54% (n = 105) delivered by CD (Table 1). Of those with successful VBAC, 79% (n = 71) were unassisted vaginal deliveries and 21% (n = 19) required vacuum assistance. All women with two-or-more-previous-CD (18%, n = 35) delivered by repeat CD, either elective or emergency, depending on the onset of labour (considered to be emergency CD if arriving to hospital in established labour).

Most women in this study with a single-previous-CD (99%, n = 158) underwent a TOLAC in hospital, with only two women from this cohort having an elective CD (refusal of TOLAC and maternal intellectual impairment). Overall, 57% (n = 90) of patients who underwent TOLAC achieved a successful VBAC.

Of the 90 successful VBACs, the majority (74%, n = 67) were attended solely by midwives, while the remaining 26% (n = 23) were attended by doctors – most of which were vacuum-assisted deliveries (Figure 1).

Among the 105 mothers who delivered by CD in the study population, 20% (n = 21) were elective and 80% (n = 84) were emergency CD performed after failed TOLAC or for multiple-previous-CD presenting in labour. Elective CD were mainly for multiple prior CD or refusal of TOLAC. Pregnancies with absolute indications (e.g. placenta praevia, footling breech) had been excluded during sampling.

The leading indications for emergency CD are tabulated in Table 2.

Women with a previous successful VBAC and those presenting in more advanced labour were significantly more likely to achieve VBAC (Table 3). Mean birthweight did not differ significantly between groups.

A total of 26 women (13%) experienced post-delivery complications (Table 4), the majority (69%, n = 18) occurring in the CD group. Puerperal sepsis (38%, n = 10) was seen exclusively after CD. Adhesions were unique to repeat CD, while perineal tears occurred only in the vacuum-assisted VBAC group. One partial uterine rupture was suspected and confirmed at emergency CD. Only 4% (n = 8) required blood transfusion; no maternal deaths occurred.

Thirteen neonates (6.7%) required admission to the neonatal unit, most commonly after emergency CD (Table 5). There were no stillbirths or early neonatal deaths recorded during the study period.

Because of the increasing risk of uterine rupture in women with higher-order CD, women undergoing their third CD are routinely offered bilateral tubal ligation (BTL) after informing them of the increased risks in any future pregnancy. Of the women counselled for BTL at their third CD, 34% (n = 12) declined the procedure despite extensive counselling.

A majority (89%, n = 175) presented before or during early labour (≤ 3 cm dilated), which is consistent with health system messaging promoting proactive early arrival to hospital because of long travel distances and limited infrastructure (especially after hours).

Of the total women considered in the study period, those with a previous successful VBAC (15%, n = 29) achieved a 72% VBAC success rate, compared with only 37% success among those whose last delivery was by CD.

This study demonstrates that women with scarred uteri (Robson score 5) contribute substantially to the overall CD rate at Zithulele Hospital, accounting for 25% of all CD deliveries during the study period. The VBAC success rate was 46%, which compares favourably with published South African data from tertiary hospitals in Pretoria (36%) and Mpumalanga (42%).^12,13 However, optimal VBAC success rates internationally are quoted at between 60% and 80%.¹⁸

Similar patterns have been reported in other rural district hospital settings; for example, a study from KwaZulu-Natal found that 35% of CS occurred in women with previous caesarean sections (Robson group 5), with an improved VBAC success rate of 58% among women undergoing a TOLAC. Together, these findings highlight the significant contribution of women with scarred uteri to overall CS rates while demonstrating that successful VBAC is achievable in rural district hospital settings.¹⁹

All eligible women (single-previous-CD; no other contra-indications) were offered a TOLAC, and a majority of successful VBACs were attended by midwives. Factors significantly associated with VBAC success were advanced cervical dilatation at presentation (p < 0.001) and a previously successful VBAC (p = 0.001). Maternal complications occurred in 13% of cases, predominantly in the CD group, while neonatal complications occurred in approximately 7%, with no perinatal deaths.

The finding that one quarter of all CD were in women with scarred uteri aligns with global data showing that repeat CD among women with previous scars (Robson score 5) is the single largest contributor to rising CD rates worldwide.^1,2 Betran et al. and the WHO 2015 report both noted that once a CD is performed, the probability of repeat CD rises sharply, perpetuating a ‘cascade’ of surgical deliveries.^1,2 This observation has particular relevance for South Africa, where national data show persistently high CD rates in both public and private sectors.^6,20

The implications are significant for rural district hospitals, where operating theatre capacity and appropriately skilled surgical staff are limited. Repeat CD not only increases the clinical workload, but also exposes women to cumulative risks such as abnormal placentation, haemorrhage and surgical adhesions.^4,21

The VBAC success rate of 46% in this study compares more favourably to those reported at South African tertiary referral hospitals (36% – 42%), and somewhat below that of an isolated obstetric study done at a rural district hospital in South Africa.^12,13,19 This may reflect a less complicated patient cohort at district level, as more complex or high-risk pregnancies are referred elsewhere. The finding reinforces that VBAC is both safe and feasible in rural settings when appropriately selected, supervised and supported by clear protocols with access to emergency surgical intervention (including total abdominal hysterectomy) as needed.

This study also confirms the important role of midwives because nearly three-quarters of successful VBACs were attended solely by midwives once decision was made for TOLAC. This illustrates the critical contribution of nurse-led intrapartum care in district-level health facilities in South Africa. Furthermore, the high rate of women agreeable to TOLAC in this study also demonstrates its acceptability and feasibility in a predominantly midwife-managed rural maternity service.

Thus, the study findings align with the RCOG 2015 among other guidelines, which recommend offering VBAC to women with a single previous lower-segment CD and no contraindications.²²

A large proportion (79%) of CD in this study were indicated for reasons directly related to the uterine scar – including multiple-previous-CD and short inter-pregnancy interval – both of which are well-recognised risk factors for uterine rupture.⁴ Furthermore, restricted options for safe induction or augmentation in women with scarred uteri meant that poor labour progress often resulted in repeat CD, again illustrating the Caesars-beget-Caesars phenomenon.⁴ These findings are consistent with international evidence and emphasise how structural and clinical limitations within rural hospitals may further perpetuate repeat surgical deliveries.⁶

The strong association between advanced cervical dilatation at presentation and successful VBAC is consistent with existing evidence.^12,13 Women arriving in established labour are more likely to achieve vaginal delivery, whereas those admitted in latent or early labour are more susceptible to interventions possibly leading to CD.

Similarly, a previous successful VBAC was a robust predictor of VBAC success (p = 0.001). This supports international evidence showing that a prior vaginal birth, especially a previous VBAC, significantly increases the likelihood of subsequent VBAC.^12,22 These findings provide valuable guidance for risk stratification and counselling at the district level.

In contrast, birthweight and parity were not significantly associated with VBAC success in this cohort, differing from other studies that link lower birthweights and higher parity to increased VBAC rates.¹² The relative homogeneity of the study population and smaller sample size may explain this difference.

The overall maternal complication rate (13%) was low and predominantly associated with CD, echoing national maternal mortality reports showing higher morbidity in surgical deliveries.^6,20 Puerperal sepsis was the most frequent complication, confined to the CD group, underscoring the unique additional risks post CD delivery. Adhesions encountered during repeat CD procedures highlight the cumulative risk of multiple surgeries, particularly where junior, inexperienced doctors perform most operations.

The incidence of uterine rupture (0.5%) was low and comparable to published global rates for optimally managed TOLAC (0.5% – 0.7%).^4,22 Importantly, there were no maternal or neonatal deaths, supporting the safety of VBAC when conducted according to published guidelines and with access to appropriately skilled clinicians. Neonatal complications were mild and evenly distributed, with no difference in Apgar outcomes between VBAC and CD groups.

A notable observation was that one-third of women eligible for BTL declined the procedure despite counselling on the risks of higher-order CD. This suggests that preserving fertility and having larger families remains a very strong cultural priority in this rural setting. However, this has important public health implications, as each successive CD increases the risk of placenta accreta and massive haemorrhage as well as uterine rupture in further pregnancies.^4,21,23

A key strength of this study is that it provides real-world evidence from a rural district hospital, a setting underrepresented in South African obstetric literature. The dataset captures all women with previous CD who completed delivery at Zithulele over a defined period, ensuring comprehensive inclusion and minimising sampling bias.

However, as a retrospective record review, the study relied on the accuracy and completeness of clinical documentation. Potential underreporting of minor complications and missing data on certain variables (e.g. duration of labour, exact indication for prior CD) may have limited further subgroup analysis and may be a potential source of bias in the study. Additionally, the sample size, though adequate for descriptive analysis, restricted the ability to perform multivariate modelling of predictors.

Despite these limitations, the findings remain generalisable to other rural hospitals in South Africa with similar service structures and patient demographics.

This study provides actionable insights for district-level obstetric care: