Intrapartum uterine activity and neonatal outcomes: a systematic review

Background Increased uterine activity (UA) may not allow adequate recovery time for foetal oxygenation. Methods The aim of the study was to determine if increased UA during labour is associated with an increased risk of either short- or long-term neurological injury in term neonates, or with neonatal proxy measures of intrapartum hypoxia-ischemia. MEDLINE, CINAHL, and ClinicalTrials.gov were searched using the following terms: uterine activity, excessive uterine activity, XSUA, uterine hyperstimulation, and tachysystole. Any study that analysed the relationship between UA during term labour and neurological outcomes/selected proxy neurological outcomes was eligible for inclusion. Outcomes from individual studies were reported in tables and presented descriptively with odds ratios (OR) and 95% confidence intervals (CI) for dichotomous outcomes and means with standard deviations for continuous outcomes. Where group numbers were provided, ORs and their CIs were calculated according to Altman. Main results Twelve studies met the inclusion criteria. Seven studies featured umbilical artery pH as an individual outcome. Umbilical artery base excess and Apgar scores were both reported as individual outcomes in four studies. No study examined long term neurodevelopmental outcomes and only one study reported on encephalopathy as an outcome. The evidence for a relationship between UA and adverse infant outcomes was inconsistent. The reported estimated effect size varied from non-existent to clinically significant. Conclusions There is some evidence that increased UA may be a non-specific predictor of depressed neurological function in the newborn, but it is inconsistent and insufficient to support the conclusion that an association generally exists.


Background
The rationale for the proposed link between excessive uterine activity (UA) and fetal hypoxia is based on physiological studies of the haemodynamic changes that occur in the utero-placental and fetal circulations during a contraction. As a uterine contraction progresses, the amount of oxygenated blood delivered to the placenta decreases. Thus, the lowest fetal oxygen saturation percentage values are found towards the end of a contraction and take some time to recover. Increased contraction frequencies may not allow adequate recovery time and may result in progressive reductions in fetal oxygen levels [1][2][3].
There is evidence that tocolytic medications may improve fetal heart rate (FHR) abnormalities when used while emergency delivery is pending [4]. Compared to emergent delivery, tocolysis for fetal distress may improve umbilical artery (UmA) base excess (BE) values and reduce neonatal intensive care unit admission at the expense of increasing caesarean delivery [5]. A 2018 Cochrane review found several studies which showed an improvement in fetal wellbeing in response to tocolysis but concluded that, given the small sample sizes involved, "the clinical significance is unclear" [6].
Current guidelines are based on the opinion that when there is an abnormal FHR pattern in association with frequent uterine contractions, increased UA may cause fetal hypoxia [7]. The 2014 American College of Obstetricians and Gynaecologists (ACOG) guidelines, Neonatal Encephalopathy and Neurologic Outcome, define tachysystole (TS) as more than five contractions in 10 min averaged over a 30-min window, and state that TS should be treated whenever it is associated with recurrent FHR decelerations or if oxytocin is being administered [8]. The UK National Institute for Health and Care Excellence guidelines on labour management state that where the FHR pattern is suspicious or pathological any uterine hyperstimulation should be corrected [9]. The authors are unaware of any existing guidelines which address TS in unaugmented labour without FHR abnormalities.
The aim of this systematic review is to determine if increased UA during labour is associated with an increased likelihood of either short-or long-term neurological injury in term neonates, or with proxy neonatal measures of intrapartum hypoxia-ischemia.

Methods
The study protocol was registered with PROSPERO (CRD4201705258) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRIS MA) statement.

Search strategy
We searched MEDLINE, CINAHL, and ClinicalTrials.gov using terms related to UA ("uterine activity", "excessive uterine activity", "XSUA", "uterine hyperstimulation", and "tachysystole"). Due to concern about the impact of changes in clinical practice over time, the search was limited to articles published after January 1 st , 1996. Reference lists of relevant papers were checked to identify further papers for consideration. The results were de-duplicated using COVIDENCE (Covidence systematic review software, Veritas Health Innovation Ltd., Melbourne, Australia). The search was repeated prior to final analysis.

Eligibility criteria
Eligibility criteria were pre-determined. Based on the anticipated heterogeneity and small size of the evidence base, we opted for permissive eligibility criteria. Any study that analysed the relationship between UA during term labour and neurological outcomes or selected proxy outcomes was eligible. The proxy outcomes were selected because they are commonly used indicators of intrapartum hypoxia-ischemia [10]. Based on incomplete reporting, conference abstracts were excluded. Studies retrieved were screened by one author and potentially eligible studies were then assessed by two team members before a final inclusion decision.

Data extraction and evidence summary
A standardised form was used to extract data. Missing data was requested from authors and included if returned. Outcomes from individual studies were reported in tables and presented descriptively with odds ratios (OR) and 95% confidence intervals (CI) for dichotomous outcomes and means with standard deviations (SD) for continuous outcomes where available. Where group numbers were provided, ORs and their CIs were calculated according to Altman [11]. Results adjusted based on multivariate analysis were included as reported. A post-hoc analysis of the relationship between uterine activity and both FHR patterns and delivery mode was also performed.

Quality assessment
The risk of bias was independently assessed by both reviewers using The Scottish Intercollegiate Guideline Network (SIGN) critical appraisal checklists but was not used to exclude any results.

Search results and study selection
Our search retrieved 1777 citations from MEDLINE, CINAHL. (Table S1) Four additional articles were identified after review of references. The search result details are specified in supplementary material. A search of ClinicalTrials.gov did not reveal any unpublished but otherwise eligible studies. Two conference abstracts with possibly eligible results were identified but not included owing to the incomplete nature of the reports [12,13]. After screening of abstracts, 21 articles were selected for full-text review. Of these, nine were excluded [3,[14][15][16][17][18][19][20][21]. (Fig. 1).
Eleven studies featured contraction rate as an individual exposure, with one [27] reporting rate only as part of a combined exposure with contraction duration. All studies apart from Bakker et al. reported contraction rate as either a dichotomised or categorised variable. Two studies [29,30] included contraction rate as a continuous variable. All studies reported dichotomised outcomes only apart from one study [27] which reported UmA pH and BE as continuous measures.
Six of the ten included studies published after 2008 adhered precisely or closely to the ACOG 2008 definition of TS i.e. more than five contractions in 10 min, averaged over a 30-min window. Of the four that did  [23,25,32,33].

Risk of bias within studies
Of the twelve included studies, three were classed as at high risk of bias, six were classed as at medium risk, and three were classed as at low risk. (Tables S3-S14).

Synthesis of results
The results of the individual studies are summarized in Table S15.

Neurodevelopmental outcome
No studies were found which reported neurodevelopmental outcomes in relation to UA.

Umbilical artery pH
Seven studies featured UmA pH as an individual outcome. Three rejected and four retained the null hypothesis that UA is not associated with UmA pH levels. Jonsson et al. compared cases with UmA pH < 7.05 to controls with pH ≥ 7.05 and 5-min Apgar scores ≥5 [24]. TTS in the last 2 hours before delivery was found to be associated with the likelihood of an adverse outcome

Composite outcomes
Two studies featured composite outcomes which were composed of eligible outcomes but could not be separated based on available data. Smith et al. compared cases with an UmA BE ≤10 mmol/L or a 5-min Apgar score ≤ 6 to controls without those characteristics. Using automated analysis of external tocography recordings from up to 4 h before delivery, they found that TS regardless of duration was not more common in cases than controls (

Main finding
The evidence for a relationship between UA and adverse infant outcomes is inconsistent. The reported effect sizes vary from small and not statistically significant to highly clinically significant.

Strengths and limitations Individual studies
Only one study reported direct measures of neurological function in neonates and no study reported long-term neurological outcomes. These interactions are inherently difficult to assess because adverse outcomes of these types are rare and often require long-term follow-up.
As pre-specified in our protocol, NICU admission was excluded as an outcome. Admission to the NICU was reported either as an individual or as part of a composite outcome in three of the included studies [22,26,31]. The group of babies admitted to NICU is heterogenous and likely significantly composed of babies conditions not associated with neurological function.
One study explicitly excluded neonates with encephalopathy, potentially introducing significant selection bias.
Selection criteria for studies of intrapartum monitoring should allow for the inclusion of labours with negative outcomes including HIE and fetal death. The remaining studies reported proxy outcomes used as measures of intrapartum hypoxia-ischemia i.e. umbilical cord gas values or Apgar scores. These measures indicate an increased risk of neurological injury but are acknowledged to be decidedly non-specific predictors of long-term outcome [34].
None of the included articles reported sample size calculations. Since the relevant adverse outcomes are rare and assuming modest estimated effect sizes, several of the included studies were underpowered for the variables of interest to this review.
When trying to establish a causative relationship, controlling for confounding variables through multivariate analysis should be based on a model of the mechanisms of the purported relationships between the exposure, the outcome, and the potential confounder [35]. Oxytocin and nulliparity are both associated with increased UA [14,36]. Therefore, in a study of UA and neonatal outcomes, it is not usually appropriate to control for these factors since, according to the hypothesis, they may be causally related to the outcome via their effect on uterine activity. Some included studies e.g. Heuser et al. reported multivariate analyses which included these variables.
Chorioamnionitis, oligohydramnios, pre-eclampsia and the duration of labour have all been associated with both increased UA and adverse neonatal outcomes [22,26,37,38]. For the most part, the connection between these factors and neonatal outcomes is not plausibly via their effect on tachysystole. Therefore, it may be appropriate to control for these factors when analysing the influence of UA on neonatal outcomes. They were not considered in the included studies.

Review level
The permissive eligibility criteria for this review allowed a comprehensive survey of the literature. Due to the significant heterogeneity of the studies in terms of overall design, as well as the types and definitions of exposures and outcomes, metanalysis was precluded by this approach. However, given the disparate nature and small size of the evidence base, valid metanalysis would not have been possible even with a narrowly focused review.
The assessment of bias in the individual studies was limited. Current tools for assessing bias in noninterventional studies are not robust [39].
Publication bias is common in systematic reviews and may particularly affect observational studies [40]. Since publication bias is difficult to assess accurately and owing to the heterogeneity of the included studies, no formal assessment was attempted. Study registration and replication have been proposed as solutions to publication bias and selective reporting in observational studies [41,42]. None of the included trials had protocols registered at ClinicalTrials.gov. No replication studies were found.

Interpretation
Most commonly UA was defined in terms of rate only, as a dichotomous variable and without regard to labour progress. This model is partly based on unavoidable limitations of external tocography but is nonetheless a reductive view of UA which may obscure the effect of increased UA on neonatal outcomes.
The aim of this review was not to assess the impact of UA levels on FHR patterns. However, it is important to consider how interventions for FHR abnormalities might affect the relationship between UA and neonatal outcomes. Three included studies reported increased rates of FHR pattern abnormalities in labours with TS and another [22] reported worsening of FHR traces to be temporally related to TS. In Bofill et al. labours with TS were more likely to result in caesarean delivery for FHR abnormalities. If increased UA leads to fetal distress, interventions for fetal distress such as caesarean delivery might lessen the impact of increased UA and the observed effect on outcome could be weakened. Therefore, differences in the management of TS and/or fetal distress are among the possible explanations for the disparity in reported results. Future studies should report on delivery methods and their indications, as well as any intrauterine resuscitation administered so that these measures can be taken into account.
As an individual marker, it is unlikely that UA can accurately predict the condition of a fetus after delivery. It is possible that increased UA is not on its own typically sufficient to produce significant fetal hypoxia-ischemia but that, in concert with other factors such as placental insufficiency or prolonged labour, it may contribute to adverse neonatal outcomes.
Based on current evidence, which is limited, tachysystole is common and mostly does not result in neonatal complications. There is inconsistent evidence to support the hypothesis that increased UA is associated with neonatal markers of intrapartum hypoxiaischemia and depressed neurological function in the newborn.