Maternal plasma levels of oxytocin during physiological childbirth – a systematic review with implications for uterine contractions and central actions of oxytocin

A comprehensive and systematic search was done in the following databases: PubMed, CINAHL and PsycINFO in October 2015, to identify studies or existing systematic reviews. We also hand-searched the reference lists of all eligible publications for references to other possibly relevant studies. The search string is available in the Additional file 1.

A total of 4039 publications were found after removal of duplicates. Assessment of titles and abstracts were divided in six groups of authors (KUM & AEB; AK& BK; MB & LL; EH & CMM; FLL & ZP; AD & SD) and assessed independently by the authors in pairs. The assessment was based on the inclusion criteria, i.e. women without medical complications during pregnancy or labour. In case of disagreement a third reviewer (KUM) was involved. This process ended in exclusion of 3973 papers. The full text of the remaining papers (n = 66) was read by a team of two researchers (KUM and AEB). Three further studies were added from reference lists giving 69 papers which were screened for eligibility. Of these studies 49 were excluded because they did not fulfil the inclusion criteria. This process led to a final number of 20 publications to be included for data extraction. For overview of study selection see the flow diagram in Fig. 1.

A combined analysis of all data was not possible because the studies differed substantially as to the type of biochemical analysis used to determine oxytocin levels as to when during labour, as well as to the timing and the frequency of blood sampling (single samples, repeated samples with different time intervals including months, weeks, hours, minutes or seconds). We therefore created a narrative synthesis for each article. From these narrative results, oxytocin levels during labour were extracted and then summarized to allow characterization of oxytocin levels during the different time periods of birth, and also comparison between groups.

Some of the included studies also contained information about plasma levels of oxytocin during basal conditions (non-pregnant women) and during pregnancy. These data were included to allow comparison between oxytocin levels in these situations and plasma levels of oxytocin during birth. Information about oxytocin levels in cerebrospinal fluid (CSF) during pregnancy and birth was also included. Data on oxytocin levels following bolus injections or infusions of synthetic oxytocin described in the included studies are also presented. Finally, data on the association between uterine contractions and plasma levels of oxytocin were included. In some studies, oxytocin levels were measured both before and after medical interventions. In all except one of the included studies, radioimmunoassay (RIA) was used to determine oxytocin levels [12].

The levels of oxytocin in the 20 selected papers were presented in different measures/units, e.g. μU/mL, pg/mL or pmol/L (pM). In order to facilitate interpretation of the results, Table 1 shows conversion between international units (IU), weight units (g) and molar units (M). All samples were from peripheral veins, except one study that simultaneously also collected blood from the jugular vein.

Data from the included articles has been reported in Tables 1, 2, 3, 4, 5, 6 and Figs. 2, 3a, b, and 4, as described above.

Based on a systematic review on plasma levels oxytocin in physiological labour and birth, we summarize the data as follows: Basal oxytocin levels rise during pregnancy. Oxytocin levels increase during labour. Pulses of oxytocin occur with increasing frequency, amplitude, and duration from the end of pregnancy towards the end of the second stage of labour. The maximal frequency of pulses is 3 per 10 min. A pronounced (four-fold) rise of oxytocin occurs in connection with birth. Oxytocin is also released after birth during the third phase of labour in connection with expulsion of the placenta. Oxytocin pulses are not directly in time associated with uterine contractions. Infusion of synthetic oxytocin at a rate of 4–9 mU/minute gives rise to oxytocin levels equivalent to levels during physiological labour. Oxytocin levels double in response to doubling of the infused dose. Administration of 4–8 mU of synthetic oxytocin as an intravenous bolus injection gives rise to peaks of oxytocin that are equivalent in size to peaks that occur during physiological birth and stimulate uterine contractions. Oxytocin levels in CSF rise during labour and birth.

The studies that are described and analysed in this systematic review are mainly of explorative research design, and may at first sight seem less sophisticated, compared to contemporary randomized controlled studies. However, these studies are of high quality. It is important to realise that measuring oxytocin levels in repeated samples during the long duration of labour and birth is extremely complex, not only because of the intrinsic challenges of sampling blood in a labouring woman, but also because of ethics, expense, and the complex bio-analytic methods that are required.

The design and content of the studies vary considerably, and therefore the studies are difficult to compare in a systematic way, and also a standard quality assessment of the articles was not possible to carry out. The studies included single oxytocin levels obtained during labour, as well as sequences of oxytocin levels obtained during pregnancy and labour, with time intervals varying from weeks to minutes, even seconds. In addition, the included studies used different techniques for oxytocin measurement, giving numerical differences in oxytocin levels, and even different effect patterns. Some of these differences are related to the technology and assays available at the time of the study. Generally, more recent studies have used more accurate measurement tools, as discussed below. Because of these differences, data from articles using different methods cannot be collated.

The first methods used to determine oxytocin levels in blood were based on biological activity. For example, “oxytocin equivalent activity” measured the effect of oxytocin to stimulate milk ejection in lactating rabbits [13]. Such techniques are complicated and relatively unspecific and insensitive. Subsequently, immunological techniques were developed such as radioimmunoassay (RIA), which measures the propensity of the oxytocin in plasma samples to influence the binding between antibodies to oxytocin and radioactive oxytocin. With the development of RIA, increasingly smaller amounts of oxytocin could be measured. However, these techniques are not completely specific, and it is possible that other substances in plasma might influence the binding between oxytocin and its antibodies, and affect the results obtained. Extraction of the plasma samples prior to the assays was introduced to reduce this risk and increase the specificity of the RIA. Another factor that may influence the results is the specificity of the antibodies that react with oxytocin in the radioimmunoassay. This may explain the variation in the included studies, as to basal levels of oxytocin obtained with different assays, but also in the patterns of effect [31].

Enzyme-linked immunosorbent assay (ELISA, EIA) is another type of immunological method used to measure oxytocin levels. Studies using ELISA have obtained very high oxytocin levels, along with different effect patterns, compared to studies using RIA [33]. For example, some studies that used ELISA have found no rise of oxytocin levels during pregnancy [34]. In addition, the validity and reliability of ELISA as a measure for oxytocin levels have been seriously questioned [35].

The timing and frequency of collection of plasma samples is also very important. Even with a specific and sensitive technique for analysis of oxytocin, the short-duration oxytocin pulses in labour require special consideration. Repeated measurements during a limited period of time (sampling at 30 s or 1 min intervals) will allow detection of peaks [31].

The time period during which sampling is performed is also of importance. For example, sampling at several time points during pregnancy is obviously necessary to detect a gestational rise. Likewise, the clear-cut rise of maternal oxytocin levels in connection with birth can only be identified if repeated samples are collected over the duration of labour.

There is no data suggesting that physiological labour onset occurs as a consequence of a sudden rise of oxytocin levels. Several of the included studies, however, show a gradual rise of oxytocin levels during physiological labour. During the first stage of labour, oxytocin levels are twice as high as before labour onset, as demonstrated in Figs. 2 and 3a. The oxytocin pulses are very short-lasting and may reach quite high concentrations (Fig. 4). As labour progresses, these pulses increase in size and frequency, reaching maximal frequency (3 pulses per 10 min) just before the baby is born (Table 4 and Figs. 2, and 4).

Some of the oxytocin peaks occurring during labour may be spontaneous and of central origin, whereas some of them are likely to be induced in response to activation of the Ferguson reflex. This reflex is stimulated when uterine contractions press the head of the foetus down against the cervix and vaginal wall. Afferent sensory nerve fibres are activated, and send impulses via the spinal cord to the SON and PVN, resulting in the release of oxytocin into both the brain and circulation. This causes a feed-forward effect that accelerates oxytocin release, known as the Ferguson reflex [36]. The maximal expression of this reflex during birth corresponds to the 3- to 4-fold rise of oxytocin levels observed during, and immediately after, birth of the baby (Table 4 and Fig. 2).

The uterine muscles are highly sensitive to oxytocin in labour and birth due to upregulation of the sensitivity/concentration of oxytocin receptors by high oestrogen elevations in late pregnancy [37]. In addition, oxytocin release in labour promotes prostaglandin release, further strengthening uterine contractions and labour progress [23, 37]. As little as 4–8 mU given as bolus dose stimulated uterine contractions just before labour onset [31].

It may seem surprising that studies consistently found no temporal connection between uterine contractions and oxytocin peaks, even when sampling was very frequent. This may be explained by the involvement of the autonomic nervous system (ANS) in labour, and especially the parasympathetic nervous system (PSNS) branch. The uterus is innervated by both PSNS and SNS branches of the ANS, with powerful effects on labour and birth. Parasympathetic activation promotes uterine contractility and increases circulation to the uterus and baby. The role of the SNS is more complicated. Sympathetic activation (e.g. from anxiety or fear) may trigger ineffective contractions and inhibit uterine circulation, or may in some cases give rise to long-lasting and more painful contractions [38].

Oxytocin in the brain activates the PSNS. Some oxytocinergic fibres from the PVN reach the parasympathetic networks (plexa) in the lumbosacral region of the spinal cord, where they connect with neurons that stimulate uterine contractions [39]. Thus neurogenic oxytocin may influence and amplify the effects of circulating oxytocin, as the release of oxytocin into the circulation and the activation of the PSNS from the PVN are coordinated.

This two-fold effect of oxytocin on uterine contractions, via circulating oxytocin and via the PSNS, may explain the unexpected finding that the peaks of oxytocin seen during normal labour are not directly in time associated with uterine contractions [16, 19‐21].

As discussed in the introduction, oxytocin induces many positive effects by its central actions within the brain. These include stimulation of friendly social interaction, enhanced wellbeing and positive mood, and reduced anxiety, pain and stress [4]. Some of these effects also occur during labour and birth as a consequence of oxytocin released in the brain. Oxytocin is released from dendrites and cell bodies of the magnocellular neurons of the SON and PVN during physiological labour and birth. In addition oxytocin is released from nerves from the parvocellular neurons in the PVN. In this way oxytocin reaches important regulatory areas in the brain, for example areas involved in regulation of social behaviour and fear, pain, stress and wellbeing. Obviously the central effects of oxytocin are not induced by the oxytocin released into circulation but directly from nerves within the brain. In fact oxytocin in the circulation does not pass into the brain because of the blood-brain barrier [4, 40‐42].

While the use of synthetic oxytocin was not initially the subject of this review, some of the included studies measured oxytocin levels when synthetic oxytocin was infused. In view of the widespread use of synthetic oxytocin for induction and augmentation of labour, we decided to include this information.

Infusions of synthetic oxytocin for induction or augmentation of labour are generally commenced at a low rate of around 1–3 mU/min, with the dose either increased step-wise (arithmetic) or doubled (geometric) at 15–40 min intervals, although protocols vary widely. Generally the maximum dose of synthetic oxytocin does not exceed 32 mU /min [23, 48].

One of the included studies found that, when synthetic oxytocin was infused in doses up to 9 mU/min, maternal plasma levels of oxytocin were similar to those observed during physiological labour [23] whereas doses around 10–16 mU/min gave rise to levels that were approximately double those of physiological labour (Fig. 3a & b) [23]. This implies that the concentrations obtained following infusion rates over 16 mU/min exceeds physiological levels. A steady state concentration was found by 40 min after a change in infusion rate [24]. It is of interest that administration of a bolus as low as 4–8 mU of synthetic oxytocin gave rise to uterine contractions in women at term, with an elevation of oxytocin levels comparable to levels obtained during a physiological pulse in labour [31].

In contrast to the episodic, narrow peaks of oxytocin in physiological labour, administration of synthetic oxytocin produced flat oxytocin levels in maternal blood. These flat levels may influence the pattern of uterine contractions and contribute to the hyperstimulation that can result from high doses of synthetic oxytocin [48]. Excessive uterine activity may also compromise fetal blood supply, causing hypoxia.

In addition, prolonged exposure to synthetic oxytocin may eventually lead to reduced contractility of the uterine muscles due to desensitization of oxytocin receptors [49]. The decreased efficacy of oxytocin may also increase the risk of postpartum haemorrhage [50]. In addition, the high levels of lactate that accumulate in uterine muscle [51], and hypoxia due to excessive uterine contractions, may also contribute to these risks.

The pharmacokinetics of oxytocin during labour are surprisingly poorly studied. As studied in non-pregnant women, synthetic oxytocin is degraded according to a two-compartment model [52]. However in labour, the levels of oxytocinases, the enzymes that degrade oxytocin, increase 10- to 20-fold, which may also influence the degradation of oxytocin and thereby the half-life and metabolic clearance rate of oxytocin [19, 24, 53].

Following on from the findings of our systematic review, an important question is: “How can the administration of synthetic oxytocin in labour and birth imitate physiological patterns and effects, as far as possible?”

From this perspective, a more physiological method of administering synthetic oxytocin would be to use a pulsatile infusion pattern, which is more aligned with labour physiology. Several studies have found that the total dose requirement of oxytocin is 20 to 60% reduced with pulsatile administration compared to continuous infusions. In addition, there is generally less hyperstimulation, and equivalent outcomes for mothers and babies such as caesarean rates and outcomes for the baby [54‐61]. In addition, a recent in vitro study has suggested that the pulsatile administration pattern of oxytocin may maintain myometrial receptor sensitivity [62] and therfore maintain uterine contractility.

Another possible conclusion from the recent pharmacokinetic data, which show a relatively longer half-life of oxytocin than currently assumed [52], would be that doses should not be increased with too short intervals. Intervals of 30–40 min rather than 15 min would be consistent with these understandings. Too-short dosing intervals could lead to unnecessarily high cumulative exposures, as levels are increased before the full effects of the previous doses are achieved. In order to achieve the lowest dose exposure, it makes sense to carefully titrate the dose against effect, taking into account this pharmacokinetic data. Arithmetical vs geometric dose increases would be more consistent with low dose exposure.

These hypotheses and models need to be thoroughly studied in relation to the use of synthetic oxytocin in labour, especially given its widespread use.