Women are suffering high levels of morbidity as a result of a prolonged labor [1–5]. The most common solution for a dystocic delivery is a caesarean section. This method of delivery raises concerns not only regarding risks of complications for the current delivery, but also for future deliveries [6–8]. A more restrictive approach is therefore desirable, especially in deliveries of nulliparous women.

For labor to end successfully, the uterus needs to produce strong, coordinated and effective contractions. Under exertion, lactic acid is produced by glycolysis in all human cells. Glycolysis mainly occurs under hypoxic conditions; but the uterus is highly glycolytic and produces lactate even under normoxic conditions [9–11]. Repeated transient hypoxia is a normal feature of labor as the uterine vessels are occluded during each contraction [9–11]. Lactate will be produced and the levels of pH in the tissue will decrease. Decreasing pH leads to intracellular acidification and an inhibition of the Ca²⁺ channels in the myometrial cells [12]. A decreased inflow of Ca²⁺ into the muscle means that the contraction will be weaker and consequently less effective [9, 10]. New studies show a significant correlation between lactate production, hypoxia and the effect of oxytocin [13–15]. This is an important interaction if labor is to end normally. The level of lactate has been recognized as a factor in uterine activity and pathophysiological processes during labor [13–15].

A partogram should be used to facilitate early detection of poor progress of labor [16–20]. If poor progress is confirmed, oxytocin is suggested for augmentation. Indeed, responses to the administration of oxytocin are variable, and many labors diagnosed as dystocic will be long, painful and unsafe, as the use of oxytocin does not guarantee vaginal delivery, i.e. labor dystocia occurs despite adequate stimulation with oxytocin. A reasonable assumption is that increased knowledge of uterine activity and pathophysiological processes will be the key to the ability to improve the treatment of dysfunctional labors [13–15].

A close correlation has been described between lactate produced by the uterine unit and the level of lactate in amniotic fluid (AFL) [21]. High levels of AFL are shown to be over-represented in dystocic deliveries compared to deliveries with a normal progress of labor [21–25]. A new, non-invasive method has been developed that makes it possible to detect the AFL level in a small sample of amniotic fluid collected from the vaginal pouch and analysed at the bedside in the delivery room. Levels of AFL can be measured immediately and an answer provided within 15 seconds [21–25].

The aim of this study was to evaluate whether the AFL-level, analysed in a sample of amniotic fluid collected at arrested labor when oxytocin was needed, could predict labor outcome in nulliparous deliveries.

A prospective multicentre trial was performed in delivery wards in Sweden, Switzerland, France and Tanzania between 2010 and 2014. Frequencies of included deliveries and operative interventions per country are presented in Figs 1 and 2.

A total of 3000 deliveries were included in the study. However, 703 mostly African women were subsequently excluded due to failure to meet the inclusion criteria. This was something detected after enrolment, and most common reason was that they were not primiparas or was not completely healthy (Fig 3, flow chart). Table 1 describes demographic data for the entire study population according to nationality. The African women were younger and less educated than the other participants; their pregnancies were shorter, and African new-borns were smaller at delivery.

Oxytocin was used in 66% (1515/2297) of all deliveries in the study. Table 2 presents delivery characteristics in the augmented deliveries where an AF sample was collected within 30 minute before oxytocin was administered (n = 638). In 877 deliveries, a sample of AF was not collected within 30 minutes, and they were excluded from the calculation.

In the augmented deliveries the spontaneous vaginal delivery rate was 69.4% (443/638). 75.2% (480/638) of them were diagnosed with labor dystocia according to the partogram, and 64.2% (308/480) of deliveries with labor dystocia arrested at the first stage of labor. 23.8% (152/638) of the augmented deliveries had an active period of labor for more than 12 hours.

Of the new-borns, 2.2% (14/638) had an Apgar score of < 7, at 5 minutes after delivery, and most affected new-borns were seen in Tanzania (3.7% with a low apgar score). One maternal and three fetal deaths occurred in the study, all of them in the Tanzanian group.

Totally, 5418 samples of AF were collected in the study. 1854 of them at the first vaginal examination after the woman attended delivery ward, 1515 prior to augmentation with oxytocin (638 of them within 30 minutes), and 2049 at the time of delivery. Deliveries with AFL ≥ 10.1 mmol/l correlated with an increased frequency of operative intervention (p < 0.001) and with an active time of delivery >12h (p = 0.04). Significantly more epidurals were used in the high AFL group (p = 0.007), and the group also showed a higher incidence of post-partum fever (>38°C, p = 0.01) and post-partum hemorrhage >1.5L (p = 0.04, Table 3).

According to study design AFL-values within the 30 minutes before augmentation were used in our prediction of labor outcome. The sensitivity for a cesarean section according to high/low AFL values was 39%, the specificity 90%, the PPV 37%, and the NPV was 91%. The overall percentage of correct predictions when AFL was used was 84% (Table 3).

High maternal age, gestational age > 41w, occiput posterior presentation of the fetus, the use of EDA, being delivered in Switzerland and an AFL level ≥10.1 mmol/l at the time of augmentation were all associated with an increased likelihood of operative intervention (Table 4).

No significant association was indicated between the levels of education, a long latent phase, or labor dystocia according to the partogram.

High AFL value (≥10.1 mmol/l) and occiput posterior presentation of the fetus had the strongest association with operative intervention after adjusting for the other factors (Table 4). No significant interactions between AFL and each of the other eight factors were detected (p > 0.48 for all test of interactions), which implies that the sensitivity for operative intervention among women with a high AFL value (≥ 10.1 mmol/l) was 4.5 times higher compared to women with a low AFL value, irrespective of the levels of the other factors.

This large observational study of healthy nulliparous women from different countries and settings presents a new way of describing labor dystocia. The lactate values in amniotic fluid analysed just before augmentation with oxytocin provides important information about the uterus. Low levels of AFL may support the decision to continue a prolonged vaginal labor by augmentation with oxytocin as the uterus appears to be receptive to augumentation, whereas high levels of AFL indicate a higher risk not only of cesarean section, but also of post-partum complications.

Labor dystocia is a leading indication for cesarean section worldwide [26–31]. Several publications have linked a number of well-known risk factors for an operative intervention among otherwise low-risk women [29–31]. In this work we have studied some of these risk factors, as well as their relation to the metabolic status of the uterus presented as the AFL value. Our results show that AFL is a new predictor of labor outcome in arrested labors. If a high AFL level is present when augmentation starts, the likelihood of an operative intervention increases 4.5 times irrespective of the levels of the other known risk factors (Table 4).

Another very important observation of this study is that the partogram, recommended by the WHO, did not predict labor outcome. More than half of all deliveries in this study were diagnosed as having labor dystocia according to the partogram, whilst over 60% of them ended in a spontaneously vaginal delivery. Our reflection is that deliveries among primiparas may be different from how they were described by Friedman and his co-workers who created the partogram in the 1950s [18, 19]. Our message is that whilst the partogram is in every way an important tool for keeping records of activity during childbirth, it is not a useful predictor of delivery outcome [16–20].

A study of AFL was just recently performed at Maternity Hospital, Dublin, Ireland, and presented by Murphy et al [32]. In the Dublin study, the first sample of AFL, when a woman attended the delivery ward, was analysed. They concluded that primarily, high levels of lactate in amniotic fluid in spontaneously laboring; single cephalic, nulliparous womans deliveries is a new and independent predictor of labor disorder and cesarean section [32].

Strength of this project is the understanding why some dystocic deliveries, despite adequate stimulation with oxytocin, do not reach a spontaneous vaginal delivery. It is also a strength that clinics from various settings are included, which makes it possible to confirm that the method of AFL is easy to use, irrespective of geography, culture or clinic size. Some limitations of the study should be noted. As it was not feasible to change clinical guidelines simply for this observational study. Different methods of managing arrested labors were seen in the clinics included, even though the same inclusion criteria were used for the study. The number of deliveries eligible to be included was also limited, due to the strict inclusion criteria: all women included had to be healthy nullipara with a spontaneous onset of labor, in order to avoid bias and interference factors.

The main question in this work is whether the introduction of this new method could help obstetricians/midwives to make a more informed decision in the delivery room. When no information about AFL is available, a good guess would be that arrested delivery will end in a vaginal delivery after a proper administration of oxytocin. If that assumption was correct, the overall percentage of correct classifications of labor outcomes in this study would be 69.4% (Table 4). The sensitivity for an operative delivery would be 0% and the specificity 100%. If the measure of AFL had been used together with the partogram, the overall percentage of correct classifications is better (84%). 39% of the women who had an operative intervention at a later stage would have been identified earlier in the process through their raised AFL levels, hopefully with a higher possibility of avoiding a long and painful parturition. 39% can be perceived as a rather low figure, but in clinical obstetrics of today’s, this prediction cannot be made at all.

The predictive values of AFL may be of more interest for the individual obstetrician/midwife. The NPV imply that more than 90% of cases with a low AFL will have a vaginal delivery. This represents important information. We anticipate that the number of unnecessary cesarean sections in this group can be reduce, as important information about the uterine receptivity to augmentation hopefully lead the to a more individually adapted use of oxytocin and to more normal deliveries. The NPV imply that 37% of deliveries with high AFL would end in an operative intervention. This is also important information for the staff in charge. This is a small group, approximately 15% of the material, but shown to have a higher frequency of complications, probably due to the prolonged time of delivery. This is probably the group described earlier where labor dystocia occurs despite adequate stimulation with oxytocin. By using this new method, very long deliveries can be avoided, caesarean sections that will be carried out can be made at an earlier stage, and postpartum complications will hopefully be reduced. The study has confirmed a new means of classifying dystocic deliveries into an early identification of a normal group with a good receptivity of oxytocin and a possibility of a vaginal delivery after an adequate use of oxytocin, and a group where problems are likely to arise. This is valuable information, especially in countries where referral to a higher level of care may be needed.

One important factor to be taken into consideration is how common an operative intervention is in a specific population. This is problematic as little or no common consensus exists, what labor dystocia really is and how to handle a dystocic delivery. Operative interventions in this study are more common in Switzerland; the predictive value of AFL might therefore be better in this subgroup than among the Tanzanian women, where fewer deliveries end in an operative intervention. On the other hand, more complications such as post-partum infections and post-partum hemorrhages are seen among Tanzanian women in the study. One maternal and three fetal deaths occurred in the Tanzanian group, all of them after extremely long deliveries with high AFL values. AFL is likely to be an even more useful measurement in these settings. The question becomes, whether we are using the correct operative delivery criteria when handling a dystocic delivery. Hopefully this study can help us to reach a new kind of consensus with a reduction in the number of unnecessary operative interventions.

“In conclusion” we have found that the AFL values collected within 30 minutes of augmentation with oxytocin are an early predictor of labor outcome in dysfunctional primiparous deliveries. Given that today many deliveries are unnecessarily long with risks of increased perinatal morbidity; our findings have important implications for public health.