The trial protocol is reported in line with the Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines (Additional file 2) [32].
Aims and objectives
The aim of this trial is to assess the benefits and possible harms of two regimens of ACS;(i) dexamethasone phosphate 4x6 mg q12 h or (ii) betamethasone phosphate 4x2 mg q 12 h) compared to placebo, when given to women in the late preterm period (gestational age of 34 weeks 0 days to 36 weeks 5 days) with a high probability of preterm birth. The primary objectives are to compare the effect of each active ACS arm with placebo on a composite outcome of stillbirth, neonatal death, or use of respiratory support within 72 hours of life or prior to discharge from hospital, whichever is earlier. The secondary objectives are to compare the effects of each ACS regimen versus placebo on maternal and neonatal safety and healthcare utilization outcomes.
Trial design
ACTION III is a parallel-group, three-arm, individually randomized, double-blind, placebo-controlled trial of two ACS regimens, dexamethasone phosphate 4x6 mg q12 h regimen and betamethasone phosphate 4x2 mg q 12 h regimen, given to women with a high probability of preterm birth in the late preterm period to improve neonatal outcomes (Fig. 1).
Study setting
This is a multi-country, multi-centre trial that will be conducted in Bangladesh, India, Kenya, Nigeria, and Pakistan, in hospitals where the WHO ACS treatment criteria can reasonably be met [30]. Specifically, these include hospitals where gestational age assessment can be accurately undertaken, there high likelihood of preterm birth within 7 days of starting ACS therapy, capacity to recognize and rule out any clinical maternal infection, adequate childbirth care is available (including capacity to recognize and safely manage preterm labour and birth) and the preterm newborn can receive adequate care (including resuscitation, kangaroo mother care, thermal care, feeding support, infection treatment and respiratory support including safe use of oxygen and continuous positive airway pressure [CPAP] as needed) [30].
A total of 29 hospitals are currently recruiting into the trial. The hospitals are largely similar to those that participated in the WHO ACTION I trial [6]. Although these are hospitals with the capacity to manage women having preterm birth and provide care for preterm newborns with minimal out-referral, they do however experience human resource, and health system challenges that are common in LMICs. Trial activities will be facility-based, with hospital or community follow-up of recruited women and newborns after hospital discharge, to 28 completed days of life.
Participants
Women with singleton or multiple pregnancy at 34 weeks 0 days to 36 weeks 5 days, with at least one live fetus, and a high probability of late preterm birth will be included. High probability of late preterm birth (up to 36 weeks 6 days) is defined as birth expected between 12 hours and 7 days after randomization as a result of one of the following:
-
preterm labour with intact membranes, where preterm labour is defined as at least 6 regular contractions/hour and at least one of the following: cervix ≥ 3 cm dilated or 75% effaced; or
-
membranes rupture without preterm labour (preterm labor defined as above; or
-
planned delivery by induction of labour or caesarean section between 24 hours and 7 days, as deemed necessary by the provider. An induction must be scheduled to start by 36 weeks 5 days at the latest, whereas a caesarean section must be scheduled by 36 weeks 6 days at the latest.
In order to assess eligibility, a good-quality antenatal ultrasound with reliable gestational age estimation must be available. If a woman has not received an obstetric ultrasound of reasonable quality for gestational age estimation previously (at least two weeks prior to screening) during the current pregnancy, it must be performed as part of eligibility assessment during the screening process.
A woman is ineligible if she is expected to give birth in < 12 hours (i.e., if she had ruptured membranes with cervix dilated ≥ 3 cm or effaced ≥ 75%, or with more than 6 contractions per hour or both cervical changes and contractions as specified; or cervical dilation ≥ 8 cm with intact membranes) or; if there is evidence of non-reassuring fetal status or other clinical indication requiring immediate preterm delivery. She will be excluded if the obstetric care provider has a clinical suspicion or evidence of clinical chorioamnionitis or severe infection; if she has received any systemic corticosteroid in the last two weeks (outside of trial); or if no prior ultrasound assessment of gestational age is available and an immediate ultrasound examination is not possible. Other reasons a woman may be ineligible to participate include a major or lethal congenital fetal anomaly being identified, confirmed COVID infection deemed severe enough to require steroid treatment as per national standards of COVID treatment, is unwilling or unable to provide consent or assent (including due to active labour), or is currently participating in another clinical trial, or has previously participated in any ACTION trial or any other clinical indication where the treating clinician considers corticosteroids to be contraindicated.
Intervention and control
The intervention regimens are (a) dexamethasone phosphate 4x6 mg q12 h, or (b) betamethasone phosphate 4x2 mg q12 h. A single course of 6 mg intramuscular (IM) dexamethasone phosphate or 2 mg IM betamethasone phosphate is administered every 12 hours, to a total of four doses or until birth occurs, whichever occurs first. The control arm will receive four saline injections of the same volume at the same dosing intervals.
Outcomes
Primary outcomes. Stillbirth or neonatal death within 72 hours of birth or use of respiratory support within 72 hours of birth or prior to discharge from the hospital, whichever is earlier. Use of respiratory support is defined as any one of the following: (i) use of invasive mechanical ventilation, (ii) continuous use of CPAP for 12 hours or more with an FiO2 ≥ 0.4 at any time, or (iii) continuous use of supplementary oxygen for 24 hours or more with an FiO2 ≥ 0.4 at any time.
Secondary outcomes (also see additional file 3).
Newborn mortality and respiratory morbidity outcomes: stillbirth; neonatal death within 72 hours, 7 days, and 28 days of birth; resuscitation at birth (i.e., use of positive pressure ventilation for > 1 min at birth); severe respiratory distress within 72 hours of birth or prior to discharge from the hospital; use of respiratory support within 72 hours of birth or prior to discharge from the hospital (as defined above); death or mechanical ventilation or very high CPAP settings (≥ 8 cm water pressure and ≥ 0.7 FiO2) in the first 72 hours of birth; and cause-specific mortality.
Newborn safety outcomes: neonatal sepsis in the first 7 days of birth; hypoglycaemia in the first 36 hours after birth.
Newborn health service utilization outcomes: admission to neonatal care unit in the first three days after birth; duration of birth hospitalization; and any parenteral antibiotic use in the first seven days after birth.
Maternal safety outcomes: maternal death; possible maternal bacterial infection during hospital admission(s); chorioamnionitis; and postpartum endometritis.
Maternal health service utilization outcomes: duration of hospital stay; any therapeutic antibiotic use; and any antibiotic use.
Participant timeline
The participant timeline and follow-up process are summarized in Fig. 2. Screening, informed consent, and randomization will take place in study hospitals where the mother has presented and will give birth. Randomized women and their newborns will be followed up during the hospital stay and then to 28 days after birth. After randomization, study data will be recorded by trained research staff in participating hospitals. At the time of discharge, the woman will be advised to return to the study hospital or call the site investigators in the event of any adverse outcomes for her or her baby. All randomized participants (women and newborns) will have scheduled postpartum/postnatal follow-up visits conducted on day 7 and day 28.
Screening, informed consent, and recruitment to the trial
Pregnant women admitted to the antenatal, labour ward or emergency admission area in the third trimester at participating hospitals will be routinely evaluated on arrival by obstetric care physicians. Women with clinical features or indications suggestive of preterm birth in the late preterm period will be informed of the study. Study staff (including research or clinical staff trained in study procedures) will conduct formal screening using a standardized screening form. The screening population is women who are between 34 weeks 0 days and 36 weeks 5 days and who are expected to deliver between 12 hours and 7 days. Screening will consist of three key activities to assess women for eligibility and trial entry: assessment of gestational age (34 weeks 0 days to 36 weeks 5 days), assessment if the delivery is likely between 12 hours and 7 days, and assessment for the presence of any exclusion criteria.
Gestational age will be based on a best obstetric estimate combining information from the last menstrual period, regularity of cycles, and informed by the earliest ultrasound obtained in the pregnancy. Women will receive an ultrasound assessment for gestational age as part of this screening process if an obstetric ultrasound of reasonable quality for gestational age estimation has not been performed during the current pregnancy at least 2 weeks prior to the screening. In these screening ultrasounds, biometric measurement and gestational age assignment will be performed using INTERGROWTH-21st project biometry guidelines and fetal growth curves, respectively [33] (Additional file 3. Figure 1). Prior to enrolling patients, all study sites received standardized training in the ultrasound evaluation of fetal biometry in the third trimester. Throughout the study, images will be reviewed regularly for quality assurance.
If women are unable to complete the full screening and informed consent process (due to distress, or other reasons), they will not be recruited. All women will receive information about the trial in their language of choice via an information sheet. If willing to participate, the informed consent form will be signed by the participant and study staff.
Allocation sequence generation
Participants will be randomly assigned to either dexamethasone-4x6 mg or betamethasobe-4x2 mg ACS regimen, or placebo in a 1:1:1 allocation as per a computer-generated randomization sequence, in balanced permuted blocks. The randomization sequence will be generated by a researcher external to the ACTION III trial. The assignment schedule will be stored at WHO headquarters in Geneva, Switzerland. All sites will receive treatment packs according to the randomization sequence, assembled in special dispensers. Facility study team members at participating sites will remove and open the next pack from the dispenser for allocation as per a set standard operating procedure.
Allocation concealment and blinding
Allocation concealment will be achieved by having identical treatment packs across the three arms. At the time of randomization, study staff will take the next sequentially numbered pack from the box (Additional file 3, Fig. 2). The IM injections will be administered by the hospital staff nurses according to study procedures.
Participants, care providers, investigators, the WHO trial coordinating unit, and data collectors will be blinded to the group allocation. Each active drug will have its own saline placebo identically packaged to allow for blinding of the three arms. Betamethasone phosphate (4mg/ml as 1ml ampules) and its placebo (1ml normal saline) will be manufactured by Recipharm AB. Dexamethasone (4mg/ml as 1ml ampules) and its placebo (1ml normal saline) will be manufactured by Fresenius Kabi. All three arms will have identical packaging, appearance, labelling, and volumes to be administered.
Emergency unblinding procedures
The principal investigators at each site and a designated WHO project manager will receive the participants’ treatment codes in the form of separate sealed envelopes that contain the treatment allocation for each participant, in case the code for a participant’s treatment needs to be broken urgently.
Retention and follow-up procedures
The study team will request contact details (address, phone number, relatives) from randomized women, in order to facilitate communication at follow-up to 28 days after birth by home visit.
Data collectors will make every reasonable effort to follow the woman and her newborn for the entire study period. After randomization, outcomes occurring in the facility (prior to discharge) will be captured by research study staff working in participating hospitals. At the time of discharge, the woman will be advised to return to the study hospital or call the site investigators in the event of any adverse outcomes for her or her baby.
Data management
Data will be managed centrally by a data management team, supervised directly by WHO project managers. The study statistician will be responsible for the development of the statistical analysis plan and reporting to the DSMB. A web-based, Good Clinical Practice (GCP) compliant data management platform (Kamolo, Centro Rosarino Estudios Perinatales (CREP) will be used, and be overseen by the site data managers. All data will be collected in study centres on paper case report forms (CRFs). Quality control will be performed at each site, and a validation system will be built into the data entry and management system to ensure consistency, accuracy and completeness of the data collected.
Confidentiality
To ensure participant confidentiality, each participant will be identified by a unique ID number. The local trial register linking personal information and trial ID numbers, and all personal information of participants, will be kept separate from the CRFs. Trial documents will be kept securely under lock and key in the research offices and will not be accessible, other than to the researchers. Data will be entered by trial ID number in the password-protected data management system to which only trial staff will have access. The trial report will not contain the names of any participants, and after completion of the trial, the trial documents will be archived in accordance with institutional and national rules for clinical research archiving.
Statistical methods
Sample size
It is estimated that the prevalence of the primary composite outcome in the control arm will be between 10–12% based on data from ACTION I [6], ACTION II [12], and the ALPS study [9] on late preterm births. A reduction of 20% is the minimal change deemed acceptable in the composite outcome, in order to change practice. Assuming a 2.5% loss to follow-up, a sample size of 4,500 women per arm in the three-arm trial will have at least 80% power and α at 0.027 to account for multiple comparisons using Dunnett's method to detect a 20% reduction in the composite primary outcome.
Statistical methods for primary and secondary outcomes
The primary intention-to-treat analysis will be based on all participants (i.e. newborns of randomized women, and women) with outcome data available. Data from participants who withdraw their consent for their data to be used will be excluded from the analysis and considered lost to follow-up.
Comparative analyses between trial arms will consider multiplicity as both ACS arms use the same placebo arm as a comparator: confidence intervals for the intervention effect (e.g. risk ratios) will be computed to have a joint 95% coverage probability using Dunnett’s method [34].
The primary outcome and most secondary outcomes are binary variables. For these outcomes, the total number of observations, number of missing values and percentages will be reported per arm. Comparisons of outcomes between each intervention arm and the placebo arm, will be described using risk ratios. Risk ratios will be estimated by binomial generalised estimating equations with log links and robust standard errors to account for potential correlation of outcomes among babies born to the same mother. The primary analysis will include arm and site as fixed covariates, and a secondary analysis will also adjust for any baseline covariates for which there is an important imbalance at baseline. If the log binomial models fail to converge, then Poisson models with robust standard errors will be used.
For continuous outcomes (duration of birth hospitalization for the mothers and the babies), the number of participants, missing values, minimum, maximum, means, and standard deviations by arm will be reported. Comparisons of each ACS dose arm against the placebo arm will be described as mean differences. Duration of birth hospitalization (continuous neonatal outcome) for all babies will be compared between arms as mean differences estimated using mixed linear models, including a maternal random effect to account for potential correlation of outcomes among babies born to the same mother. The median duration and Kaplan Meier curves will also be reported by arm.
Interim analyses
A first interim analysis by the Data Safety Monitoring Board (DSMB) is tentatively planned at 60% recruitment completion, however reporting on safety criteria (including serious adverse events) will occur quarterly. At this first interim analysis, the DSMB will look at the performance of both active arms combined, versus placebo. This is to minimize unnecessary exposure of additional women to the placebo (in case the combined arms are better than placebo). If the analysis reveals that the combined ACS arms are superior to placebo with p < 0.001 (Peto’s rule), then the DSMB could recommend cessation of the placebo arm (after confirming that there is statistical evidence that at least one ACS arm is better than placebo with p < 0.025, and for safety p < 0.01 for mortality. However, recommendations after the results of an interim analysis will not be guided only by statistical considerations, but also by practical issues (adverse events, unanticipated costs), as well as clinical considerations or external new information.
In the event of both ACS regimens being superior to placebo, and based on a benefit-risk assessment, the DSMB could recommend a non-inferiority comparison between the two active arms..
Subgroup analyses
Pre- and post-randomization subgroup analyses will be conducted for the primary endpoint. Pre-randomization subgroups include different indications for enrolment (i.e. rupture of membranes, preterm labour with intact membranes, planned termination), gestational age at enrolment (< 34 weeks 6 days, 35 weeks 0 days to 35 weeks 6 days, > 36 weeks 0 days), study site, and single vs multiple births. Post-randomization subgroups include gestational age at birth (preterm (< 37 weeks) vs. not preterm (≥ 37 weeks)), interval from time of IMP administration (i.e., first dose) to birth (0 to 12 hours, > 12 to 24 hours, > 24 hours to 7 days, > 7 days), use of tocolytics post-randomization, appropriate size for gestational age vs small for gestational age, and mode of birth (vaginal vs caesarean section).
Statistical tests for effect modification by the different factors mentioned above will be performed. While post-randomization subgroup analyses are at risk of bias, in the current trial we believe there are good scientific reasons to investigate these subgroups as there are plausible reasons why the treatment effects could be different. Also, these subgroups are clinically important, and are explicitly considered in the latest update of WHO’s ACS recommendations [26]. Before conducting these post-randomization subgroup analyses, we will first examine whether the intervention has an effect on the stratifying variable.
Trial oversight
Monitoring procedures have been prepared in accordance with the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) harmonised tripartite guideline for Good Clinical Practice E6 (R2). Monitoring activities will be conducted overall, per site and per hospital. Standard operating procedures will be prepared for all monitoring activities and will govern all monitoring procedures. Monitoring will be intensive throughout the trial recruitment period and will be conducted by independent trial monitors, principal investigators and co-investigators, and WHO trial coordinating unit (TCU) comprising WHO staff from two WHO Departments (Maternal, Newborn, Child, and Adolescent Health and Ageing, and Sexual and Reproductive Health and Research). Day-to-day oversight will be done by the trial steering committee comprising TCU and the principal investigators at each site. A technical Trial Advisory Group (external independent scientists with expertise in the area of preterm birth) led by an independent chair will advise the trial steering committee.
A study DSMB will comprise five members, including an independent chair, a statistician, and three technical experts familiar with the intervention, maternal and newborn health care, and clinical trial methodology. The DSMB will monitor adverse events on an ongoing basis to look for emerging safety risks and advise the trial coordinating unit (TCU) accordingly.
Ethical considerations
The trial protocol was reviewed and approved by the WHO Ethics Review Committee. All participating sites received approval from the relevant institutional scientific and ethical review committees in the respective country (as well as required permissions from the relevant national regulatory authorities) (Additional file 6). Any modifications to the protocol which may impact the conduct of the study, a potential benefit of the study participants, or may affect their safety, including changes in study objectives, study design, study population, sample sizes, study procedures, or other significant aspects will require a formal amendment to the protocol. Such amendments will be agreed upon by study co-investigators and submitted to WHO ethical review committee and participating institutional ethical review committees prior to implementation.