Vous êtes sur la page 1sur 8

DOI: 10.1111/1471-0528.

15448 Intrapartum care


www.bjog.org

Admission cardiotocography versus intermittent


auscultation of the fetal heart in low-risk
pregnancy during evaluation for possible labour
admission – a multicentre randomised trial:
the ADCAR trial
V Smith,a C Begley,a J Newell,b S Higgins,c DJ Murphy,d,e MJ White,f JJ Morrison,g S Canny,g
D O’Donovan,h D Devanei,j
a
School of Nursing and Midwifery, Trinity College Dublin, Dublin, Ireland b School of Mathematics, Statistics and Applied Mathematics,
National University of Ireland, Galway, Ireland c Department of Obstetrics and Gynaecology, National Maternity Hospital, Dublin, Ireland
d
Department of Obstetrics, School of Medicine, Trinity College Dublin, Dublin, Ireland e Department of Obstetrics and Gynaecology, Coombe
Women and Infants University Hospital, Dublin 8, Ireland f Department of Neonatology/Paediatrics, Coombe Women and Infants University
Hospital, Dublin 8, Ireland g Department of Obstetrics and Gynaecology, University College Hospital Galway, Galway, Ireland h Department of
Neonatology/Paediatrics, University College Hospital Galway, Galway, Ireland i School of Nursing and Midwifery, National University of Ireland,
Galway, Ireland j Health Research Board, Trials Methodology Research Network, National University of Ireland, Galway, Ireland
Correspondence: V. Smith, School of Nursing and Midwifery, Trinity College Dublin, 24 D’Olier Street, Dublin 2, Ireland.
Email: smithv1@tcd.ie

Accepted 2 August 2018. Published Online 19 September 2018.

Objective To assess the effect of admission cardiotocography Results Based on 3034 women (1513 and 1521 randomised to IA
(ACTG) versus intermittent auscultation (IA) of the fetal heart and ACTG, respectively), there was no statistical difference
(FH) in low-risk pregnancy during assessment for possible labour between the groups in caesarean section [130 (8.6%) and 105
on caesarean section rates. (6.9%) for IA and ACTG groups, respectively; relative risk (RR)
1.24; 95% CI 0.97–1.58], or in any other outcome except for use
Design A parallel multicentre randomised trial.
of continuous CTG during labour, which was lower in the IA
Setting Three maternity units in the Republic of Ireland. group (RR 0.90, 95% CI 0.86–0.93).
Population Healthy, low-risk pregnant women, at term and Conclusion Our study demonstrates no differences in obstetric or
≥ 18 years old, who provided written informed consent. neonatal outcomes between IA and ACTG for women with
possible labour onset, other than an increased risk for continuous
Methods Women were randomised to receive IA of the FH or
CTG in women receiving ACTG.
20 minutes ACTG on admission for possible labour onset, using
remote telephone randomisation. Both groups received IA Keywords Admission cardiotocography, cardiotocography, fetal
during labour, with conversion to continuous CTG as clinically assessment, intermittent auscultation, labour admission test.
indicated.
Tweetable abstract No differences in outcomes between
Main outcome measures Caesarean section (primary outcome), intermittent auscultation and admission cardiotocography for
obstetric interventions (e.g. continuous CTG during labour, fetal women with possible labour onset.
blood sampling, augmentation of labour) and neonatal morbidity
Linked article This article is commented on by AM Vintzileos,
(e.g. metabolic acidosis, admission to the neonatal intensive care
p. 122 in this issue. To view this mini commentary visit https://
unit, neonatal death).
doi.org/10.1111/1471-0528.15445.

Please cite this paper as: Smith V, Begley C, Newell J, Higgins S, Murphy DJ, White MJ, Morrison JJ, Canny S, O’Donovan D, Devane D. Admission
cardiotocography versus intermittent auscultation of the fetal heart in low-risk pregnancy during evaluation for possible labour admission – a multicentre
randomised trial: the ADCAR trial. BJOG 2019;126:114–121.

Trial registration: http://www.isrctn.com/(ISRCTN96340041).

114 ª 2018 Royal College of Obstetricians and Gynaecologists


A randomised trial of ACTG versus IA on labour ward admission

to understand study information and give written informed


Introduction consent. Women were excluded if any of the following con-
Monitoring of the fetal heart rate (FHR) throughout preg- ditions applied to them: previous caesarean section, pre-
nancy, labour and birth is central to the assessment of fetal eclampsia, diabetes, gestational age greater than 40+6 or less
wellbeing to optimise fetal and neonatal outcomes. A pre- than 37+0 weeks, undergoing induction of labour, rupture
cursor to FHR monitoring during labour is an assessment of membranes for longer than 24 hours, temperature
of the FHR on admission to the labour ward or labour > 37.5°C on admission, body mass index > 35 kg/m2 at
assessment room in women with signs and symptoms of antenatal booking, maternal age > 40 or < 18 years,
possible labour onset. This is usually performed by either assisted conception in this pregnancy, previous stillbirth or
intermittent auscultation (IA), which involves listening to neonatal death, small-for-gestational-age (< 10th centile),
and counting the fetal heart (FH) for one or more minutes, oligiohydramnios, polyhydramnios, abnormal Doppler
or admission cardiotocography (ACTG), a screening test artery velocimetry, multiple pregnancy, breech presentation,
consisting of an approximately 20-minute electronic reduced fetal movements on two or more occasions, meco-
recording of the FHR and uterine activity. The rationale nium stained liquor, fetal congenital abnormality, or seri-
for labour admission assessment of the FHR is to try to ous maternal medical disease (cardiac or vascular disease,
identify those babies at greater risk of intrapartum fetal epilepsy, liver disease, hyperthyroidism, or abnormal renal
compromise, and hence, who might benefit from continu- function).
ous cardiotocography (cCTG) during labour.1 However,
there is evidence of a lack of benefit supporting the use of Enrolment procedures
ACTG in low-risk pregnancy with many guidelines recom- A two-stage process of consent was used. Women attending
mending the use of IA in the absence of risk factors for the antenatal clinic between 32 and 40 weeks of gestation
continuous monitoring of fetal wellbeing.2,3 However, pre- were given written information about the study and asked
vious trials informing these guidelines have evaluated the to consider participation before their admission to the hos-
effects of ACTG on women in the context of diagnosed pital with signs and symptoms of possible labour onset. A
labour, either spontaneous onset or induced labour.4–7 Rec- sticker showing ‘ADCAR study information given’ was
ommendations have called for studies to evaluate the placed on the front cover of their maternity notes high-
effects of ACTG on women admitted with signs and symp- lighting that they had received the study information.
toms of possible labour onset, when ACTG is applied, gen- Women attending the labour ward/labour assessment unit
erally, and before a formal diagnosis (e.g. following an between 37+0 and 40+6 weeks of gestation with signs and
assessment of cervical dilatation and effacement on vaginal symptoms of possible labour onset (e.g. regular uterine
examination) of spontaneous or induced labour.8,9 contractions, rupture of membranes, maternal report of
The objective of this paper is to report the findings of a suspected labour), and who had received the study infor-
multicentre randomised trial that compared ACTG with IA mation, were screened for eligibility to participate in the
of the FHR in low-risk healthy pregnant women presenting trial using a trial screening and register form. Screening
with signs and symptoms of possible labour onset, and was conducted by the admitting midwives, on a 24-hour
before confirmation of established labour, on rates of cae- basis, as women presented. Women who were screened eli-
sarean section, obstetric intervention and neonatal morbid- gible were invited to participate in the trial and sign the
ity. The trial is registered with http://www.isrctn.com/ study consent form. Following consent, participating
ISRCTN96340041. women were randomised to receive ACTG or IA (allocation
ratio of 1:1) by the admitting midwife using a remote cen-
tral telephone randomisation service (ALEA, a software
Methods
package developed to support randomisation in healthcare
Design and participants research in Europe). Randomisation occurred before a for-
A two-group, parallel, randomised trial was undertaken in mal diagnosis of labour; eligible women with signs and
three maternity units in the Republic of Ireland between symptoms of possible labour onset were entered into the
May 2008 and June 2012. Two units were tertiary centres trial irrespective of whether they were subsequently found
with approximately 8000 and 3000 births per annum, and to be in established labour or not, or whether their admis-
the other unit was a regional centre with an annual birth sion was for nonreassuring FHR following assessment. The
rate of approximately 4000. Participants were healthy low- randomisation sequence was computer generated, using
risk pregnant women attending any of the three study sites block randomisation, with random block sizes of two or
with signs and symptoms of possible labour onset. Eligibil- four, and stratification by study site. Due to the nature of
ity criteria included gestation between 37+0 and 40+6 com- the intervention, blinding was not possible at participant
pleted weeks of pregnancy, age ≥ 18 years, and an ability and clinician level; however, the biostatistician performing

ª 2018 Royal College of Obstetricians and Gynaecologists 115


Smith et al.

the analysis was blinded to group allocation during the the antenatal ward to await further signs of possible labour
analysis process. onset. On readmission to the labour ward or labour assess-
ment room with further signs and symptoms of labour,
Intervention group women received ACTG with subsequent care in accordance
Women randomised to the intervention received IA of the with care as described above.
FH, performed by the admitting midwife, on admission for
possible labour onset, using either a Pinard stethoscope or Outcome measures
a hand-held Doppler ultrasound device. Although the cho- Although a core outcome set for intrapartum fetal assess-
sen method of IA was not recorded routinely, the likeli- ment is registered with the Core Outcome Measures in
hood was that a Doppler device was used in the majority Effectiveness Trials (COMET) database (http://www.comet-
of cases as this is the more commonly adopted method for initiative.org/studies/details/741?result=true), this core out-
IA in the trial settings. IA included abdominal palpation of come set is not yet developed. In the absence of a core out-
uterine contractions and assessment of the FHR for at least come set, our outcome measures were as follows. The
60 seconds after a uterine contraction. Conversion to CTG primary outcome measure was overall incidence of cae-
occurred where IA revealed a baseline FHR of < 110 beats/ sarean section. Maternal secondary outcome measures were
minute or > 160 beats/minute or any decelerations in the cCTG during labour (defined as cCTG for > 75% of labour
FH; or if any other risk factors developed during admis- duration), fetal blood sampling, augmentation of labour
sion, which warranted cCTG, or the clinician caring for the with oxytocin, augmentation of labour with artificial rup-
woman had any other cause for concern. Following IA, ture of membranes, duration of labour, epidural analgesia,
where the above interpretation criteria were met, subse- pethidine analgesia and mode of birth. Neonatal outcome
quent care during labour was in accordance with standard measures were metabolic acidosis (defined as an umbilical
care; that is, monitoring of the FHR by IA for 1 minute at artery blood pH < 7.0, and a base deficit of > 12.0 mmol/
least every 15 minutes in the first stage and at least every l),1 hypoxic ischaemic encephalopathy, seizures (either
5 minutes in the second stage of labour or conversion to apparent clinically or detected by electroencephalographic
cCTG if clinically indicated. Women assessed as not being recordings), neonatal Apgar scores at 1 and 5 minutes,
in established labour following randomisation were either admission to the neonatal intensive care unit (NICU),
discharged home or admitted to the antenatal ward to length of stay in NICU (in days), neonatal death, stillbirth,
await further signs and symptoms of labour. On readmis- intracranial haemorrhage (as determined by cranial ultra-
sion to the labour ward or labour assessment room, with sound/magnetic resonance imaging report), meconium
further signs and symptoms of labour, and where the trial aspiration (as determined on X-ray) and any neonatal
screening and register form criteria for low-risk continued resuscitation. The following baseline characteristics were
to be met, women received IA of the FHR, with subsequent also measured: maternal age, gestation at birth, parity,
care in accordance with the care as described above. neonatal birthweight and neonatal congenital abnormality
(undetected before birth).
Control group Public and patient involvement
Women randomised to the control group received 20-min- A maternity service user was a member of the Trial Steer-
utes of ACTG on admission to the labour ward or labour ing Committee. Pregnant women were involved as partici-
assessment room, as was standard care in the three partici- pants in the trial.
pating hospitals. If the baseline FHR was between 110 and
160 beats/minute, baseline variability was >10 beats/min- Sample size assumptions and estimates
ute, there were more than two accelerations present, and A retrospective clinical audit of 600 women was conducted
decelerations were absent then the tracing was classified as in two hospitals in 2007 to determine the number of
normal and discontinued, and the findings were docu- women who would potentially be eligible to take part in the
mented. The uterine contraction pattern was also assessed. study and to determine the incidence of caesarean section
If, following the ACTG, the above interpretation criteria in this low-risk maternity sample. The audit found that
were met, the woman’s care was in accordance with stan- 55% of women, across the two hospitals, met the eligibility
dard care; i.e. IA for monitoring of the FH during labour criteria, and the caesarean section rate was 5.2%. With 80%
or conversion to cCTG as warranted. If the interpretation power and a set at < 0.05, a sample size of 5776 women
criteria were not met on ACTG, CTG was continued and (2888 per group), allowing for a 10% attrition rate, was
interpretation of the tracing was used to inform the subse- required to detect a 30% reduction in the incidence of cae-
quent care of the woman, as was current practice. Women sarean section (i.e. from 5.2% with ACTG to 3.6% with IA)
assessed as not being in established labour following ran- allowing for a 10% attrition rate. Calculations were per-
domisation were either discharged home or admitted to formed using SAMPLE POWER.

116 ª 2018 Royal College of Obstetricians and Gynaecologists


A randomised trial of ACTG versus IA on labour ward admission

Data collection and statistical analysis Results


Data were extracted from the Maternity Information Sys-
tems in each hospital and from hard copy clinical notes as During the study period, 14 928 women who had received
required. Predesigned data extraction forms were devel- the study information were assessed for eligibility to partic-
oped, piloted and refined for collecting the relevant data. ipate in the study. Of these, 6771 (45%) were screened eli-
All data were entered into SPSS (Version 16) in prepara- gible, of whom 3045 (45%) consented and participated.
tion for analysis. For dichotomous data, relative risks (RR) Eleven (0.4%; six ACTG, five IA) women were lost to fol-
with 95% CI were calculated. For continuous data, mean low up, resulting in data from 3034 women available for
differences (MD) with 95% CI were calculated. For the pri- analysis. Table 1 presents participant characteristics by
mary outcome, incidence of caesarean section, an unad- group; no differences in baseline characteristics were found.
justed analysis, followed by a secondary exploratory Figure 1 illustrates the recruitment and randomisation
analysis adjusting for a mother’s age, parity and site (and process.
possible interactions), was performed. All comparisons for
secondary outcomes were unadjusted. All analyses adhered Primary outcomes
to ‘intention-to-treat’; however, as a number of women A total of 235 women had a caesarean section; 130 (8.6%)
were discharged home ‘not in labour’ following randomisa- in the IA group and 105 (6.9%) in the ACTG group. The
tion (see Table 1), and became ineligible subsequently, we results demonstrate no significant difference between the
performed a sensitivity analysis for the primary outcome groups in relation to caesarean section rates observed
excluding these women. (unadjusted RR 1.24, 95% CI 0.97–1.59, P = 0.08). The
primary reasons for caesarean section in the IA and ACTG
Ethical considerations groups, respectively, were caesarean section for abnormal
Ethical approval to conduct the study was granted by the FHR pattern (n = 84 and n = 71), emergency caesarean
Faculty of Health Sciences, Trinity College Dublin, and the section for reasons other than abnormal FHR pattern
Research Ethics Committees of each of the three study (n = 42 and n = 34), and elective caesarean section (n = 4,
sites. Participation was voluntary, and women could with- IA group only).
draw from the study at any time. Written informed consent
was sought from eligible women as close as possible to the Secondary outcomes
time of randomisation. All data were collected, processed There were no statistical differences in any of the secondary
and stored confidentially in accordance with the Data Pro- outcomes, other than the use of cCTG during labour,
tection (Amendment) Act (2003).10 The trial was funded which was significantly less in the IA group (RR 0.90, 95%
by an HRB-Ireland grant (ref: RP/2006/55) and supported CI 0.86–0.93), P < 0.00001). Table 2 presents the results
by an equipment grant from the Department of Health and for the secondary outcomes.
Children (Ireland). No fetal or neonatal deaths (up to discharge home) were
observed in the study. Ninety-six babies in total were

Table 1. Participant characteristics Total screened


14 928
Characteristic IA = 1513 ACTG =1521

Ineligible Declined Randomised


Maternal age at time 30.4 years 30.1 years 8157 3726 3045
of randomisation (mean)
Para 0 668 (44%) 654 (43%)
Para 1 533 (35%) 524 (35%) ACTG IA
Para 2+ 312 (21%) 343 (22%) 1527 1518
Ineligible post-randomisation 227 (15%) 249 (16%)
Spontaneous labour onset 1376 (91%) 1363 (91%)
Induced labour* 123 (8%) 149 (10%) Loss to follow up Loss to follow up
Not in labour caesarean section 14 (1%) 9 (1%) 6 5

*Two hundred and seventy-two women presenting with signs of


labour were randomised to the study, found not to be in labour, ACTG for analysis IA for analysis
1521 1513
and subsequently had an induction of labour for post-dates, or
other clinically relevant indications.
Figure 1. Randomisation and recruitment.

ª 2018 Royal College of Obstetricians and Gynaecologists 117


Smith et al.

Table 2. Secondary outcomes

Outcome IA* = 1513 ACTG** = 1521 RR (95% CI) (unadjusted)


n (%) n (%)
mean (SD) mean (SD)

Spontaneous vaginal birth 1075 (71%) 1113 (73%) 0.97 (0.93–1.01)


Instrumental birth (all types) 307 (20%) 303 (20%) 1.02 (0.88–1.17)
Instrumental birth (specific sub-type)
Ventouse—prolonged second stage 74 (5%) 74 (5%) 1.01 (0.73–1.38)
Forceps—prolonged second stage 28 (2%) 45 (3%) 0.63 (0.39–1.00)
Instrumental—abnormal FHR pattern 205 (14%) 184 (12%) 1.12 (0.93–1.35)
Acceleration of labour (ARM + oxytocin) 829 (55%) 864 (57%) 0.96 (0.91–1.03)
ARM only 573 (38%) 608 (40%) 0.95 (0.87–1.04)
Oxytocin only 256 (17%) 256 (17%) 1.01 (0.86–1.18)
Pethidine during labour 271 (18%) 277 (18%) 0.98 (0.85–1.14)
Epidural during labour 658 (44%) 627 (41%) 1.05 (0.97–1.15)
cCTG during labour 1093 (72%) 1224/1520 (86%) 0.90 (0.86–0.93)
FBS during labour 99 (7%) 93 (6%) 1.07 (0.81–1.41)
Length of first stage (hours) 3.1 (2.4) 3.2 (2.5) –0.10 (–0.27–0.7)
Length of second stage (hours) 0.5 (0.5) 0.5 (0.5) –
Length of third stage (hours) 0.1 (0.4) 0.1 (0.2) –
Length of maternal hospital stay (days) 2.2 (1.2) 2.2 (1.2) 0.00 (–0.09–0.09)
Gestation at birth (weeks)*** 39.9 (0.9) 39.9. (0.9) –
Birthweight (g) 3523 (431) 3554 (442) –
Meconium-stained liquor 252 (17%) 266 (18%) 0.95 (0.81–1.11)
Paediatrician present at birth 556 (37%) 551 (56%) 1.01 (0.92–1.11)
Neonatal resuscitation (any) 381 (25%) 405 (27%) 0.95 (0.84–1.07)
Apgar score ≤ 7 at 1 minute 91 (6%) 107/1519 (7%) 0.85 (0.65–1.12)
Apgar score ≤ 7 at 5 minutes 12 (1%) 9/1519 (1%) 1.34 (0.57–3.17)
Arterial pH ≤ 7.05 14/1120 (1%) 15/1070 (1%) 0.89 (0.43–1.84)
Arterial BE ≥ 12.0 19/1098 10/1043 1.80 (0.84–3.86)
Admission to NICU/SCBU 42 (3%) 54 (4%) 0.78 (0.53–1.16)
Length of neonatal hospital stay (days) 2.3 (1.3) 2.3 (1.8) –

ARM, articifial rupture of membranes; BE, base excess; FBS, fetal blood sampling; SCBU, special-care baby unit.
*Denominator in all cases 1513 unless otherwise indicated.
**Denominator in all cases 1521 unless otherwise indicated.
***Participant’s gestation at birth was extracted from hospital records and mean gestational age was calculated for each group.
Bold text used to highlight statistically significant result.

admitted to the NICU/special-care baby unit. Table 3 sum- A), 10.8% versus 8.7% (site B) and 5.4% versus 6.0% (site
marises reasons for admission and outcome details. C); significantly fewer caesarean sections occurred in the
ACTG group in site A (P = 0.046), but no differences were
Secondary exploratory analyses observed between the groups in site B (P = 0.27) or in site
A sensitivity analysis, excluding women who became ineli- C (P = 0.69).
gible post-randomisation, found no difference in caesarean
section between the groups (IA, n = 227, 15% and ACTG,
Discussion
n = 249, 16%; RR 1.19, 95% CI 0.86–1.65). To further
explore the difference between the groups in the primary Main findings
outcome, adjusting for mother’s age, parity and site, no The findings of our comparison of IA versus ACTG, in
significant difference between the groups in caesarean sec- over 3000 low-risk women presenting with possible labour
tion was found (adjusted RR 1.22; 95% CI 0.93–1.60, onset, provide no evidence of a significant difference in any
P = 0.16). There was no evidence of a significant site by of our measured primary and secondary outcomes other
allocation interaction in the adjusted models. The propor- than a decrease in use of cCTG associated with being ran-
tions of caesarean sections by allocation (IA versus ACTG, domised to IA. These findings, while offering, in the main,
respectively) and by study site were 9% versus 5.9% (site equivalence for choice of method, reflect current clinical

118 ª 2018 Royal College of Obstetricians and Gynaecologists


A randomised trial of ACTG versus IA on labour ward admission

Table 3. Summary details of babies admitted to NICU/SCBU

Reason IA ACTG Outcomes


N = 42 N = 54

Respiratory difficulties 11 13 All babies were discharged home


(poor respiratory effort, TTN, RDS) One baby receiving CPAP (ACTG) spent 8 days in NICU
Sepsis/infection 8 12 All babies were treated with intravenous antibiotics and discharged home
Two babies (ACTG) were diagnosed with meningitis and both
spent 17 days in NICU
One baby (ACTG) with congenital pneumonia spent 8 days in NICU
Meconium aspiration syndrome 2 3 All babies were discharged home
One baby (IA) was ventilated at birth and spent 10 days in NICU
Three babies (ACTG) spent 3, 5 and 10 days, respectively, in NICU
Jaundice 3 1 All babies required phototherapy
Observation (low Apgars, low temperature, 7 8 All babies were discharged home
grunting, etc.)
HIE I 3 1 All babies were discharged home within 1 week of birth
HIE II 1 1 Baby (IA) received anti-convulsive therapy 9 three doses and was
transferred to a national paediatric unit for follow-on care
Baby (ACTG) was ventilated and transferred
to a different NICU at 2 days
Pneumothorax 1 1 Baby (ACTG) developed Escherichia coli abscess; discharged home after
41 days in NICU
Cardiac 1 6 All babies were discharged home with plans for follow-up care
Trisomy 21 1 2 All babies were discharged home within 2 weeks
Other 4 6 Four babies (IA) and four babies (ACTG) were discharged home
One baby (ACTG) was queried with bowel obstruction and was transferred
to a national paediatric unit for further investigations
One baby (ACTG) diagnosed with subdural haematoma (ACTG) was
transferred to a national paediatric unit for follow-on care

HIE, hypoxic ischaemic encephalopathy; RDS, respiratory distress syndrome; TTN, transient tachypnoea of the newborn.

guideline recommendations for IA use during labour group. However, an important difference in our study is
admission in low-risk, healthy pregnant women2,3 and con- that women were randomised before a formal diagnosis of
cur with previous studies where significantly lower rates of labour. One hypothesis is that ACTG has a different effect
cCTG use during labour were experienced by women on mode of birth for women with and without a diagnosis
receiving IA.4,5 Although our finding of no significant dif- of labour. Another possible explanation might simply be a
ference in caesarean birth between IA and ACTG is similar ‘chance’ finding. Given that our study failed to achieve tar-
to findings in other studies,4–7 our study differs whereby a get recruitment, it is reasonable to suppose that had we
higher incidence of caesarean birth, albeit statistically non- succeeded in achieving our sample size estimate, the trend
significant, was observed in the IA group compared with towards caesarean birth with IA may have become more
the ACTG (8.6% and 6.9%, respectively). Previous trials balanced between groups, or even reversed. A further
report higher rates of caesarean sections in ACTG groups; intriguing finding in our study was the high rate of acceler-
for example, for IA versus ACTG, Mires et al.4 report cae- ation of labour in both the IA (55%) and ACTG (57%)
sarean rates of 3.6% and 5.1%,4 Mitchell7 reports rates of groups, the majority of which was by artificial rupture of
7.7% and 8.7%, and Cheyne et al.6 report rates of 6.2% membranes. This finding, rather than reflecting abnormally
and 8.9%. This finding in our study is surprising given the slow labours, per se, or study sample bias, is most likely
significantly increased use of cCTG with ACTG and the because of active management of labour practices,11 which
known association between cCTG during labour and higher are common in many maternity units in Ireland.
caesarean rates in low-risk women when compared with IA
of the FHR.1 Although we performed secondary explora- Strengths and limitations
tory analyses on the outcome of caesarean section, we Over half of women eligible to participate in the ADCAR
remain uncertain as to why there was a nonsignificant trial declined to do so. Although similar rates of eligible
trend towards higher rates of caesarean birth in the IA consenting women have been found in another large

ª 2018 Royal College of Obstetricians and Gynaecologists 119


Smith et al.

maternity care trial in Ireland,12 timing of consent (i.e. Interpretation


women presenting with signs of possible labour) and Although the evidence from previous randomised trials,
approaches to taking consent (e.g. staff preferences for and a systematic review, for IA use on labour admission, is
CTG use13) may have impacted on consent rates. Failing to compelling,9 conflicting findings from recent non-rando-
recruit a significant portion of eligible women has the mised studies continue to provide some clinicians with a
potential to introduce selection bias, and this needs to be rationale for ACTG use. Rahman et al. for example, in a
considered when interpreting the results of the study. study evaluating the role of ACTG, reported a low risk of
Failing to recruit our sample size estimate is a further fetal compromise during labour for women with a reactive
study limitation. Our sample size of 1521 and 1513 for the ACTG compared with women with equivocal and ominous
two groups achieves a power of 69.3% (based on the same ACTGs; 6.9%, 39.9% and 84.6%, respectively (P < 0.001).17
a priori sample size assumptions of a 0.05 to detect a 30% The incidence of thick meconium-stained liquor, admission
reduction in the incidence of caesarean section from 6.9% to NICU and neonatal mortality were also significantly
with ACTG identified in this study). Despite responsive higher in the ominous and equivocal groups compared
efforts to increase recruitment, including embedding a with the reactive ACTG group and operative birth for sus-
Study Within A Trial (SWAT) (https://www.qub.ac.uk/site pected fetal compromise was required in 5.5% (n = 8/145)
s/TheNorthernIrelandNetworkforTrialsMethodologyResearc of women in the reactive group compared with 27.8%
h/SWATSWARInformation/) to assess the effectiveness of a (n = 5/18) in the equivocal group and 84.6% (n = 11/13)
site visit on recruitment,14 extending the trial to a third site of women in the ominous group. Bhartiva et al. also report
(site C) and extending the time frame to conduct the trial, a significant correlation (P = 0.001) between nonreassuring
achieving monthly recruitment targets was challenging, ACTG and meconium-stained liquor in low-risk women,
with monthly rates rarely exceeding 50% of expected tar- although statistically insignificant correlations were found
gets. Achieving target recruitment is known to challenge for low Apgar scores, NICU admission and mortality.18
almost all triallists, with evidence to suggest that < 50% of Contrastingly, Akhavan et al., in their recent study of 425
trials will achieve their target rates within the planned time pregnant women of whom 142 (33.4%) had abnormal
frames.15,16 Nevertheless, recruiting less than the number of ACTGs, assert that an abnormal ACTG was predictive of
expected participants to our study does limit the interpre- caesarean section, intrauterine growth restriction and low
tation of the results. Reasons for low recruitment to this Apgar scores (< 7) at birth.19
study were multi-fold, and reflect evidence that highlights
clinician preference for CTG use due to fear of litigation
Conclusion
and a perceived view of CTG technology assisting staff with
workload burden and staff shortages.13 This trial demonstrates that outcomes were not different
Furthermore, given that the rationale for either IA or between IA and ACTG for women with signs of labour
ACTG use on labour admission is to assess fetal wellbeing, with the exception of increased cCTG use associated with
ideally a trial of this type would be powered to detect a dif- ACTG. In particular, our study presents evidence for IA
ference between the groups in a serious adverse neonatal compared with ACTG in healthy pregnant women with
outcome, such as neonatal mortality. However, to do so signs and symptoms of possible labour onset, before a for-
would require the recruitment of many thousands of mal diagnosis of established labour. Further research com-
women (> 50 000) to the trial, and considerable resources. paring IA and ACTG use before the diagnosis of
In the absence of a core set of outcome measures evaluat- spontaneous labour in selected samples of women present-
ing the effects of ACTG, we measured and reported a rela- ing with signs and symptoms of possible labour is required.
tively large number of prespecified outcomes. We Further research is also required comparing the use of IA
acknowledge that this increased multiple testing and there- or the use of ACTG in women with confirmed compared
fore the risk of a higher false-positive error rate. with unconfirmed labour during labour ward admission.
Acknowledging these limitations, our study provides evi-
dence for FHR assessment based on a considerably large Acknowledgements
sample of women and presents the third largest of five tri- We thank all of the midwives, and other staff, at the three
als examining IA and ACTG during labour admission.4–7 participating study sites for supporting and assisting with
Furthermore, it is the first trial, that we are aware of, which the conduct of the ADCAR trial. We thank also the thou-
evaluates IA and ACTG use in women presenting with sands of women who participated in the study.
signs and symptoms of possible labour onset, and before
confirmation of established labour, which is more likely to Disclosure of interests
accurately reflect how ACTG is applied in practice in many VS, DD, and CB declare being awarded a grant from the
settings. Health Research Board (HRB) Ireland (Grant ref: (ref: RP/

120 ª 2018 Royal College of Obstetricians and Gynaecologists


A randomised trial of ACTG versus IA on labour ward admission

2006/55) to conduct the study. All other authors disclose admission in labour in low risk obstetric population. BMJ
no known competing interests. Full disclosure of interests 2001;322:1457–60.
5 Impey L, Reynolds M, MacQuillan K, Gates S, Murphy J, Sheil O.
available to view online as supporting information. Admission cardiotocography: a randomised controlled trial. Lancet
2003;361:465–70.
Contribution to authorship 6 Cheyne H, Dunlop A, Shields N, Mathers AM. A randomised
VS managed the day-to-day running of the trial and controlled trial of admission electronic fetal monitoring in normal
drafted the manuscript. DD conceived the idea for the trial labour. Midwifery 2003;19:221–9.
7 Mitchell K. The effect of the labour electronic fetal monitoring
and reviewed drafts of the manuscript. CB contributed admission test on operative delivery in low-risk women: a
methodological expertise and reviewed drafts of the manu- randomised controlled trial. Evid Based Midwifery 2008;6:18–22.
script. JN performed the statistical analysis and provided 8 Talaulikar VS, Arulkumaran A. Labour admission test. Int J Infertil
content in this section of the manuscript. SH, DJM, MJW, Fetal Med 2011;2:89–94.
JJM, SC and DOD provided clinical advice and site support 9 Devane D, Lalor JG, Daly S, McGuire W, Cuthbert A, Smith
V. Cardiotocography versus intermittent auscultation of fetal
during protocol development and during the conduct of heart on admission to labour ward for assessment of fetal
the trial. All authors read, contributed important intellec- wellbeing. Cochrane Database Syst Rev, 2017, Issue 1. Art. No.:
tual content to, and approved the final draft of the manu- CD005122.
script before submission. 10 Government of Ireland, Data Protection Amendment Act, 2003.
[http://www.irishstatutebook.ie/eli/2003/act/6/enacted/en/html]. Accessed
5 April 2018.
Details of ethics approval 11 O’Driscoll K, Meagher D, Robson M. Active Management of Labour,
Ethics approval to conduct the study was granted by the 4th edn. St. Louis, MO: Mosby Ltd.; 2003.
Faculty of Health Sciences, Trinity College Dublin (6-Oct- 12 Begley C, Devane D, Clarke M. An Evaluation of Midwifery-led Care
2008; Ref C.A. 149), and the Research Ethics Committees in the Health Service Executive North Eastern Area – The Report of
the MidU Study. Dublin, Ireland: School of Nursing and Midwifery,
of each of the three study sites (16-Nov-2007).
Trinity College Dublin/Health Service Executive; 2009.
13 Smith V, Begley CM, Clarke M, Devane D. Professionals’ views of
Funding fetal monitoring during labour: a systematic review and thematic
The ADCAR trial was funded by the Health Research analysis. BMC Pregnancy Childbirth 2012;12:166.
Board Ireland (ref: RP/2006/55) and supported by an 14 Smith V, Clarke M, Begley C, Devane D. SWAT-1: the effectiveness
of a ‘site visit’ intervention on recruitment rates in a multicentre
equipment grant from the Department of Health and Chil- randomised trials. Trials 2015;16:1–7.
dren (Ireland). & 15 McDonald AM, Knight RC, Campbell MK, Entwistle VA, Grant AM,
Cook JA, et al. What influences recruitment to randomised
controlled trials? A review of trials funded by two UK funding
References agencies. Trials 2006;7:9.
16 Gardner H, Fraser C, MacLennan G, Treweek S. A protocol for a
1 Alfirevic Z, Devane D, Gyte GML, Cuthbert A. Continuous
systematic review of non-randomised evaluations of strategies to
cardiotocography (CTG) as a form of electronic fetal monitoring
improve participant recruitment to randomised controlled trials. Syst
(EFM) for fetal assessment during labour. Cochrane Database Syst
Rev 2016;5:131.
Rev, 2017, Issue 2. Art. No.: CD006066.
17 Rahman H, Renjhen P, Dutta S, Kar S. Admission cardiotocography:
2 Health Services Executive/Institute of Obstetricians and
its role in predicting foetal outcome in high-risk obstetric patients.
Gynaecologists. Clinical Practice Guideline: Intrapartum Fetal Heart
Australas Med J 2012;5:522–7.
Rate Monitoring. Dublin: Royal College of Physicians of Ireland and
18 Bhartiya V, Sharma R, Kumar A, Srivastava H. Admission
Health Service Executive; 2012.
cardiotocography: a predictor of neonatal outcome. J Obstet
3 National Institute for Health and Care Excellence (NICE). Intrapartum
Gynaecol India 2016;66(Suppl 1):321–9.
care: care of healthy women and their babies during childbirth –
19 Akhavan S, Lak P, Rahimi-Sharbaf F, Rahim Mohammadi S, Shirazi
Clinical Guideline 190, 2014, United Kingdom.
M. Admission test and pregnancy outcome. Iran J Med Sci
4 Mires G, Williams F, Howie P. Randomised controlled trial of
2017;42:362–8.
cardiotocography versus Doppler auscultation of fetal heart at

ª 2018 Royal College of Obstetricians and Gynaecologists 121