A before and after study of the impact on obstetric and perinatal outcomes following the introduction of an educational package of fetal heart rate monitoring education coupled with umbilical artery lactate sampling in a low resource setting labor ward in South Africa

Both rising cesarean section rates and high perinatal mortality are issues of global concern. We demonstrate using a simple method of testing umbilical artery lactate at birth, that it is possible to reduce cesarean section rates without worsening neonatal morbidity in a low resource setting.

The rates of cesarean section (CS) are increasing globally [1], although both the total rates and the rate of rise vary considerably between low, middle, and high income countries [2]. While there is little agreement about the optimal rate of CS, there is some evidence that increasing rates above a certain threshold may not result in a corresponding improvement in maternal and perinatal mortality [3]. Moreover, CS, while a life-saving procedure that all women must have access to when required, is also associated with potential adverse maternal outcomes in both the index and future pregnancies [4]. Balancing the indication for CS with its potential risks is a critical clinical and public health challenge.

Abnormal fetal heart rate tracing or suspected fetal distress with concern for consequent hypoxia and acidemia remains one of the most common indications for the first CS (with rates cited between 10 and 32%) [5‐8]. However, the rates of pathological acidemia in neonates generally is far lower than this level, and although there are documented benefits, the use of electronic fetal heart rate monitoring in labour has not been shown to improve neonatal mortality [9]. It is important to consider fetal physiology in the intrapartum period and cardiotocograph (CTG) interpretation in great detail given the need to balance the role that intrapartum hypoxia plays in global perinatal deaths, with the role that potentially unnecessary CS play in maternal morbidity and mortality.

Objective measurements of intrapartum fetal hypoxia (such as pH or lactate sampled from the fetal scalp) may improve upon the known limited specificity of CTG and reduce the rate of unnecessary cesarean sections [10, 11]. The cost of the equipment to do this, alongside high rates of maternal HIV infection makes this prohibitive in some units around the world. Universal umbilical cord-gas sampling after delivery has been suggested as a method to improve the specificity of the CTG [12]. Paired cord arterial and venous gas sampling is both time and resource intensive however, particularly outside of high-income country settings. To that end, it is possible to introduce inexpensive bedside point of care umbilical artery (UA) lactate testing, which provides clinicians feedback on intrapartum care and potentially alters practice for future cases, which may partly mitigate the poor specificity of the CTG. Therefore, we conducted a before and after study in a tertiary labor ward in Pretoria, South Africa, to determine the impact of introducing UA lactate sampling, in conjunction with education on CTG interpretation and fetal and lactate physiology, on maternal and perinatal outcomes.

Kalafong Hospital is a tertiary hospital in Pretoria, South Africa, affiliated with the University of Pretoria. In 2014, there were 6499 births in the unit. During the period 3rd March to the 12th November 2014 we conducted a before and after study to determine the impact of umbilical artery lactate sampling, combined with an educational program on fetal heart rate monitoring, on maternal and neonatal outcomes.

The study groups included prospectively collected samples of umbilical artery lactate obtained by the primary researcher (EA), or a doctor in the unit trained by the primary researcher. We have presented the specifics of UA sampling in the unit, and the data points collected previously [13]. Briefly, after maternal written consent was obtained, the umbilical cord was clamped and cut within 1 minute of birth. A small arterial blood sample (< 0.5uL) was then taken from a double clamped segment of the remaining umbilical cord, prior to delivery of the placenta. The lactate level was measured (mmol/L) on a Roche Accutrend PlusTM© hand-held lactate meter (Rotkreuz, Switzerland). The coefficient of variation (CV) for this lactate meter is 1.8–3% [14]. All women aiming to have a vaginal birth were eligible to be consented for the trial, regardless of their ultimate birth mode.

During the period 3rd March to 18th of July, 2014, a prospective cohort of women were consented to be enrolled in the study. As this was a blinded cohort, this was a convenience sample, obtained within the capacity of a single researcher (EA). The results of the lactate sampling in this cohort were blinded to the clinicians (midwives, registrars, and consultants) managing the intrapartum care of women. This cohort served as the baseline for the unit, the results of which have been published previously [13]. Before the completion of the cohort, all midwives and registrars in the unit underwent a training course in fetal physiology, lactate physiology and cardiotocograph (CTG) interpretation, conducted by the primary researcher (EA). The course consisted of a pre-test, a series of didactic lectures, interactive application of knowledge with CTG examples, and a post-test. Subsequently, from 21st July to 12th November 2014, lactate sampling continued, with the results made immediately available to the clinicians managing the intrapartum care of women in the labor ward. Furthermore, all lactate samples obtained in the previous 24 h were reviewed at the daily morning labour ward meetings of all medical staff (interns, registrars, and consultants), and twice weekly, an in-depth review of interesting or challenging cases was facilitated by the primary researcher (EA). Daily reviews involved correlating CTG interpretation pre-birth with neonatal outcomes, including the UA lactate. During the two study periods, basic data on all births (regardless of recruitment to the study) were collected, including number of births, mode of birth, and admissions to the neonatal nursery (all levels).

Three hundred ninety-seven samples were collected in the pre-intervention period from 2196 planned vaginal births and 2436 total births. Five hundred ninety-seven samples were collected in the post-intervention period from 1978 planned vaginal births and 2232 total births. After exclusion of 54 twin pregnancies, a total of 936 samples were analysed (pre n = 374 and post n = 562).

This study explored the impact of an educational program coupled with the introduction of UA lactate sampling in a low-resource setting on intrapartum outcomes. Previous studies have been based in high-income settings [16] and it was not clear whether the previously observed benefits in earlier studies would translate to a lower resource centre. While there was no significant difference in the cord lactate results between the two groups, we found a reduction in the CS rate without worsening any of neonatal outcomes. The largest contributor to this reduction was in CS for suspected fetal compromise. A significant reduction post-intervention was observed in CS, emergency CS, and neonatal admissions for all women regardless of whether they were recruited to the study for lactate measurement or not. While these overall results are not adjusted for confounding factors, they are reassuring; the falling CS rate has not worsened fetal and neonatal outcomes, and, based on the reduction in overall neonatal unit admission rate, may have improved them. The impact of an educational program on fetal physiology and intrapartum heart rate monitoring demonstrated benefits that were translated to all women in the labour ward and we speculate this should be sustained when coupled with the introduction of universal cord lactate sampling for ongoing clinician outcome feedback.

There is evidence that the lack of appropriate interpretation of fetal monitoring in South Africa contributes to the unacceptably high intrapartum and early neonatal mortality rates [17, 18]. Equally, there is evidence that the process of quality of care audit and the feedback to clinicians of the UA lactate results may improve maternal and neonatal outcomes [12, 19]. Given the reduction in suspected fetal compromise without an increase in adverse neonatal outcomes that occurred following our intervention, we hypothesise that the objective feedback available to clinicians from the UA lactate result highlighted recognition on CTG interpretation of the “true” acidemic cases. This theory may be supported by the trend to less neonatal nursery admissions in the study group and the significantly less admissions overall in the post-intervention study period, regardless of trial recruitment status. In addition, we observed an increase in delayed labour progress in the post-intervention group without worse neonatal outcomes; it may be that less early interventions for suspected fetal distress allowed other potential complications of longer labours labor to become apparent.

Interestingly, there were less maternal problems in the post intervention period, which plausibly could have resulted in fewer fetuses at risk of compromise. There were, however, significantly less intrapartum transfers from referral facilities in the post-intervention period. The nature of an intrapartum transfer indicates the women are more likely to have maternal related pregnancy problems, necessitating up-referral to a tertiary unit [20, 21], and this may have accounted for some of the reduction in maternal complications in the post-intervention period.

There was no significant change in the elective CS rate, with the reduction in our study almost exclusively intrapartum non-elective CS, which remained significant when adjusted for primiparity. This is where we believe the strength of this intervention lies – in preventing the first CS, particular considering evidence such as that of Villar et al, demonstrating that intrapartum cesarean section in a sample of nearly 100,000 deliveries in 8 Latin American countries was associated with an increased risk in several maternal and neonatal morbidity and mortality [22]. The Robson classification for cesarean section is widely used to monitor CS rates internationally [23], with a particular focus on reducing cesarean section in nulliparous women (group 1 and 2 in the Robson classification) and those with a previous cesarean section, (group 5), with the latter a clear consequence of the former [1]. Although heterogeneous in the quality of included studies, a recent meta-analysis has demonstrated that quality of care audit using the Robson classification system to monitor cesarean section rates can result in a reduction in CS [24]. In our study we introduced a quality of care audit process and provided instant feedback to the clinicians in the unit, with a reduction in CS in women in groups 1 and 2 of the Robson classification.

Reduction in CS is of particular importance in South Africa, as women in this country are more likely to die following a CS compared with a vaginal birth [25]. This is particularly concerning as 22·6% of deliveries in South Africa are by CS, and in provincial tertiary hospitals, as is the unit in this study, the rate is 35·2% [26]. Moreover, maternal deaths related to CS have been increasing, and disturbingly, large numbers of these are due to perioperative haemorrhage [27], with surgical trauma being one of, if not the most common, reason for death [25, 28]. The reduction in CS in our study (both in the included cases and overall during the study period) has the potential for immediate and downstream effects for future pregnancies, as well as long-term outcomes for women and their families.

Finally, testing for potential fetal acidemia with lactate on a handheld, point of care device has several advantages. Firstly; it may be more feasible (both economically and in ease of use) than other measures (such as paired arterial and venous cord gas sampling), especially in the limited resource setting. Secondly, the hand-held device correlates well with laboratory analysers, adding validity to the process of point of care testing [29, 30].

The introduction of UA lactate has the ability to significantly reduce the CS rate without increasing adverse neonatal outcomes in a low-resource setting. Further research to expand the body of evidence on the use of UA lactate sampling in low and middle-income countries is needed.