Study design and participants

The STRIDER study was a Phase III clinical trial to quantify the effects of administration of sildenafil on pregnancy outcome in severe early‐onset IUGR. A total of 135 women with affected pregnancies were recruited and randomised to receive one of two treatment arms, sildenafil or placebo. The study received funding from the National Institute of Health and Care Research (NIHR) and Medical Research Council (MRC). It was co-sponsored by the University of Liverpool and Liverpool Women’s NHS Foundation Trust and co-ordinated through the Liverpool Clinical Trials Unit (LCTU, which is part of the Liverpool Trials Collaborative, UKCRC Registration 12).

All participants were recruited from one of the 19 STRIDER research sites located in the UK. All sites were leading obstetric units within the UK and successfully completed site feasibility during the green light phase of the trial set-up. Suitable collaborating sites and investigators were therefore assessed on the level of fetal medicine and neonatal service they provide and their ability to conduct the trial. Ahead of the trial starting at a site the Principal Investigators were required to agree to adhere to the good clinical practice (GCP) guidelines. In addition, all relevant regulatory and ethics approvals were required.

The study was designed as a randomised double-blind, placebo-controlled trial with sildenafil or placebo prescribed orally at a dose of 25 mg three times per day (Figure 1). This dosage regime was based on previous studies by the collaborators on the project. All participants recruited had a singleton pregnancy between 22⁺⁰ weeks’ gestation and 29⁺⁶ weeks’ gestation with a diagnosis of IUGR and had agreed to expectant management. For the purpose of study, IUGR was defined as a fetus with an estimated fetal weight or abdominal circumference below the 10th centile using local charts and absent or reversed end-diastolic flow in the umbilical artery on Doppler velocimetry.

FIGURE 1

STRIDER trial flow chart.

Gestational age was confirmed by first trimester ultrasound and in each case, the diagnosis of severe early-onset IUGR was confirmed by a fetal medicine expert having excluded fetal anatomical abnormalities. Following diagnosis and informed consent, a full history, measurements of maternal cardiovascular parameters (BP and pulse rate), fetal biometry and Doppler velocimetry were taken. Maternal venepuncture for angiogenic biomarkers was also performed.

All participants had further BP and pulse rate measurements and blood sampling 2 hours after receiving the first dose of the study drug. Subsequently, participants were followed up within 3–4 days and at weekly intervals thereafter, or earlier if clinically indicated. The remainder of clinical care was at the discretion of the local fetal medicine experts and included regular ultrasound assessment of growth and Doppler blood flow and antenatal cardiotocography.

Study medication was over-encapsulated (Sharp Clinical Services, Crickhowell, UK) to ensure that participants, clinicians and pharmacists were masked to the study drug.²⁶ Medication was dispensed in 10-day supplies with a new supply being provided weekly to ensure there was no period where medication was missed. Treatment ended at 31⁺⁶ weeks’ gestation or delivery, whichever came first. All participants were advised of the potential side effects.

Data on pregnancy outcome were collected prospectively from clinical maternity notes and entered onto a secure electronic case report form (eCRF) platform at research sites. Data quality and protocol compliance were monitored regularly by central and on-site monitoring methods.

The STRIDER UK study protocol v7.0 is available as Appendix 1

Start of study

Prior to the start of recruitment on 11 November 2014, the following approvals were obtained on the respective dates: Sponsorship – 9 October 2013, Research Ethics Committee (REC; North East – Newcastle and North Tyneside 2, Chair Dr Alasdair MacSween) – 20 March 2014, Clinical Trial Authority (CTA) – 18 July 2014 and Trial Greenlight Approval – 6 November 2014. All approvals for participating research sites were in place within 12 months of opening. The trial protocol (Appendix 1) was first registered on 31 July 2014, 4 months prior to the first participant being recruited. The first participant was recruited and randomised on 21 November 2014.

End-of-study definition

As defined in the protocol, the end of phase 1 of the STRIDER study was ‘when the last recruited woman/baby is discharged from hospital, or the baby has reached expected date of birth, whichever is later’. The last participant to be recruited to the study was on 6 July 2016 and the last infant to be discharged from hospital was on 13 February 2017. Therefore, the end of study for the STRIDER trial was reported as 13 February 2017.

Inclusion and exclusion criteria

The study inclusion criteria were as follows:

• singleton pregnancy with severe, early‐onset IUGR between 22⁺⁰ and 29⁺⁶ weeks’ gestation AND a clinical decision to manage expectantly
• IUGR defined as an estimated fetal weight < 10th centile OR abdominal circumference < 10th centile AND absent or reversed end-diastolic flow in the umbilical artery
• 16 years of age or older
• consent to take part in the trial

The exclusion criteria for the study were as follows:

• multiple pregnancy
• known or suspected structural or chromosomal fetal abnormality
• maternal illness (such as pre‐eclampsia) expected to require delivery for maternal reasons within 72 hours
• maternal wish not to have active management of the pregnancy, such as a decision to have a termination of pregnancy
• inability to give informed consent
• cocaine use in the current pregnancy
• contraindication to sildenafil therapy

  • known maternal cardiac disease
  • left ventricular outflow tract obstruction
  • concomitant treatment with nitrates, nitrate drugs for chest pains/heart problems including nitro-glycerine (glyceryl trinitrate, GTN), isosorbide dinitrate, isosorbide mononitrate
  • nitrates – some recreational drugs contain amyl nitrate (‘poppers’)
  • previous allergy to sildenafil, including hives, difficulty breathing, swelling of the face, lips or tongue

Sample size estimation

Internal audits of early‐onset IUGR cohorts revealed an average diagnosis to delivery interval of around 20 days with a SD of 11 days. In order to confirm that sildenafil could prolong pregnancy by 1 week (7 days), a total recruitment of 104 women (alpha 5%, power 90%) was required. Although loss to follow‐up was not anticipated, recruitment of 135 women was planned in order to account for any possible post‐randomisation withdrawal of consent or missing data.

The secondary hypothesis was that sildenafil will improve utero‐placental circulation and therefore delay the development of fetal cardiovascular changes (reduced short-term heart rate variability, deterioration of fetal Doppler indices) that lead to the indication for iatrogenic delivery. With a complete data set for approximately 100 participants, it was predicted that a clinically meaningful 20% difference in mean Doppler Pulsatility Index (PI) values of uterine arteries (0.86, SD 0.20), middle cerebral artery (2.21, SD 0.39) and ductus venosus (0.62, SD 0.22) would be detectable (alpha 5%, power > 80%).

Randomisation

Randomisation was performed using a web‐based randomisation service operating at the Clinical Trials Unit (CTU), British Columbia Women’s Hospital (Vancouver, Canada). Passwords and login details were provided to each STRIDER research site at the point of site ‘green light’ authorisation by the LCTU.

Treatments were allocated with equal probability by means of computer generated random permuted blocks of size two and four in equal proportions. The randomisation was stratified by two factors, the participating research sites and the gestational age at diagnosis: < 26⁺⁰ and ≥ 26⁺⁰ weeks of gestation.

As STRIDER was a double-blind placebo-controlled trial, both the participant and any clinical staff were blinded to the treatment allocation. It was a requirement that any unbinding that occurred during the running of the study was reported as a major protocol deviation. Unblinded participants would then be retained in the intention to treat (ITT) population, but removed from any per-protocol analyses.

Study end points

The primary end point for the study was the difference in length of gestation (days), defined as the time from estimated gestation until birth.

The secondary end points were divided into subgroups for fetal, infant and maternal safety and were as follows.

Statistical analysis

Participants’ groups for analysis were defined on an ITT basis. Unadjusted estimates with Kaplan–Meier estimates were presented and analysed with linear regression techniques, including the stratification factor as a main effect. The treatment effect was reported as the mean difference between groups. Statistical significance was determined as p = 0.05 or less and participants randomised before 26⁺⁰ weeks and at 26⁺⁰ weeks or later were included in the subgroup analyses.

For continuous data, the analysis of secondary end points matched the analysis for the primary end point. Binary data were compared across treatment groups using a chi-squared (χ²) test or Fisher’s exact test as appropriate and reported using RR with 95% confidence intervals (95% CI). All analyses were performed using the statistical software package, R (version 3.3.3).

A Statistical Analysis Plan (SAP) was not produced prior to the analysis being undertaken.

Quality control and data validation

The STRIDER study was subject to regular data checks and reviews as set out in the trial-specific Data Management Plan. The study was also subject to both central and on-site monitoring as set out in the trial-specific Monitoring Plan. Regular Central Monitoring Reports were produced and reviewed by the Trial Management Team. In addition, on-site monitoring visits were carried out for each research site following the hospital discharge of the first participant and surviving infant. All visits were completed and any outstanding issues identified were actioned and closed accordingly.

An Independent Trial Steering Committee (TSC – Chair Professor Alan Cameron) and Safety and Data Monitoring Committee (ISDMC – Chair Professor Ed Juszczak) were established to provide oversight for the study. These committees met prior to the study opening, twice yearly while it was running and then one final time at the end of the study to review and approve the results. No significant issues relating to the management of the study or the safety of the participants were escalated.

The STRIDER randomisation list was reviewed to ensure provision of the correct number of strata, adequate randomisation numbers per stratum, appropriate block sizes and treatment allocations, and balanced allocation of treatments for various cumulative totals. This was found to be accurate. In line with the regular safety and efficacy review of the data by the ISDMC, checks were carried out for omitted, or, out of sequence allocations and balanced in treatment allocations.

At the end of the trial, a multiple logistic regression model with treatment arm as response and baseline variables as explanatory variables was used to confirm whether the best-fitting (minimum AIC) model was the one with no explanatory variables – that is, the baseline variable was uninformative as to treatment allocation.

All statistical coding relating to the analysis of the STRIDER data deriving the primary outcome variable was reviewed by an independent statistician who performed checks to ensure that the number of participants from the database matched the number in the analyses. A random check of at least 10% of participants was also performed to ensure that the derived gestational time and birth date was correct.

Adverse events and compliance

Adverse events (AEs) and treatment adherence were assessed and recorded at weekly clinical visits from recruitment to delivery. Participants were encouraged to record any side effects or AEs, which were then reviewed and documented during each clinical visit. Adherence was assessed weekly during clinical review, with any temporary discontinuation in treatment being recorded. Treatment adherence was considered to be good if the reported intake of tablets was 90% or more of the total expected to have been taken between randomisation and the visit date.

Phase 2 – 2-year follow-up

All surviving children of mothers recruited to the STRIDER UK study were eligible and invited for follow-up. Nominated members of the core research team were responsible for accessing the original confidential trial data in order to determine the contact details of all potential participants. As part of the original STRIDER UK trial, an audit of all consent forms took place to ensure that all participants had consented to further contact in relation to future research. In addition, a check was also made to ensure that the infants of all such participants survived. This was carried out via a number of methods including a thorough audit of all SAEs (which detail all fetal and neonatal deaths) and a final check at the local research site on the child’s health status. All participants received newsletters which provided information on the trial and also give the option of opting out of any further contact. Any participant who contacted the trial management team and requested to opt out of future correspondence and participation in further research was removed from the list of potential participants for the follow-up study.

A study invitation pack was sent to all parents/carers of surviving children. This included an invitation letter, participant information sheet and informed consent form. Participants were asked to read the information carefully and discuss their child’s participation in the study with either close friends, family and/or a relevant health professional. A named contact, telephone number and e-mail address were included on all correspondence so that parents/carers were able to contact the research team and discuss their child’s participation further, if required. If they were interested in taking part, they were asked to contact the research team in order to give verbal consent to take part in the study and to arrange a convenient date and time for their child’s assessment. Those participants who did not contact the research team within 2 weeks of the invitation pack being sent were contacted by a member of the core research team. For those participants who wanted to take part, an assessment date and time was arranged. Once an assessment date and time had been confirmed, an assessment pack was sent to the participant, as detailed below who have given expressed verbal consent to take part in the study.

The assessment pack included confirmation of the (already agreed) assessment date, time and location, a map (if necessary), details of what will take place during the assessment, who will carry out the assessment and a study questionnaire pack.

Assessments took place at either a local outpatient facility or in the home. A preference was made to carry out all assessments in a controlled setting (i.e. clinical research setting), however, where necessary assessments also took place in the child’s home after the researcher had assessed the home environment for suitability. Where assessments were planned to take place in the child’s home, further information was provided to parents on how to prepare their home (e.g. clearing an open space on the floor, providing a small table/work surface and turning off any distractions such as the television and or radio). Every effort was made to ensure that assessment dates were booked to suit the needs of the participant; however, it was preferred that assessments took place in the morning as this is a time that children are often well rested, fed and able to concentrate.

Participants were given contact details should they need to cancel or rearrange their assessment. Furthermore, they received a reminder telephone call 3 days prior to the assessment date and a text reminder the day before.

All assessments were performed by a suitably trained senior research psychologist with expertise in developmental assessment techniques in young infants. This researcher was also suitably trained in the specific cardiovascular assessment techniques used within this study and blinded to treatment allocation for the main STRIDER UK study. Additional oversight and supervision were provided by co-investigator, Professor Brigitte Vollmer for neurodevelopmental assessment techniques and Dr Andrew Sharp for cardiovascular assessment procedures.

Initially, the researcher facilitated introductions between all present, following on from this the study was explained and the child and parent/carer were given the opportunity to ask any questions relating to their child’s participation. Once this was complete and informed written consent had been received, the research assessment began. The researcher explained to the parent/carer the importance of allowing the child to complete all activities independently. The parent/carer was allowed to remain in the room while the assessment took place; however, they were asked not to interrupt or assist the child with these tasks unless otherwise invited to by the researcher.

The questionnaires sent to parents/carers in the assessment pack were reviewed by the researcher and any missing items were highlighted and discussed. If the parent/carer had any questions relating to the completion of the questionnaires, this was addressed so that the parent/carer could complete them during the assessment. Other factors were determined from the core STRIDER UK data set and where necessary neonatal records including gestational age, birthweight, APGAR (Activity, Pulse, Grimace, Appearance, Respiration) scores, duration of ventilation, chronic lung disease, persistent arterial ductus, neonatal sepsis, retinopathy of prematurity, necrotising enterocolitis and focal brain injury.

Once this was complete, the researcher played with the child to build a rapport so the child was relaxed and at ease during the assessment activities. Once the researcher deemed that the child was ready, the formal assessments were introduced to the child in the order set out in Table 1. These included the Cognitive, Language and Motor Subscales of the Bayley Scales of Infant and Toddler Development – III (BSID-III);²⁷ Hempel’s Neurological Examination for Toddler Age²⁸ which was used to identify major neurological impairment (cerebral palsy; CP) and to detect subtle deviations from typical neurological and neuromotor function and a cardiovascular assessment which included brachial systolic BP and diastolic BP and arterial stiffness, assessed as aortic (central) augmentation index (AIx).

TABLE 1

STRIDER UK follow-up assessment schedule

The assessments took approximately 2.5 hours, excluding rest breaks (Table 1). Additional data were collected by the parental questionnaire and completed prior to the follow-up assessment. The assessment included well-established, valid and reliable standardised measures selected to provide a comprehensive neurodevelopmental and cardiovascular assessment together with the collection of core information on the child’s general health and well-being.

Once the assessments were complete, the researcher thanked the child and their parent/carer for their participation and parents/carers and children were given the opportunity to ask any further questions. They were advised that once the results of the assessment had been scored and interpreted, they would receive a report which would summarise the neurodevelopmental findings. Where specifically requested, a more detailed report was sent to the child’s GP and/or paediatrician who they would be able to contact if they should require any further information regarding these outcomes. The children were given a pack for participating in the study which included a certificate and a small gesture of thanks (e.g. age appropriate book) and parents/carers were given a voucher to the value of £10 to reimburse them for their time and any travel expenses.

Where potential participants cancelled or failed to attend follow-up appointments on more than three occasions, they were invited to participate remotely. All such participants received a follow-up questionnaire pack which included all questionnaires detailed as part of the main study in addition to the Ages and Stages Questionnaire-3 (in place of the BSID-III, neurodevelopmental assessment). A separate participant information sheet and consent form was also included in this pack, which had been modified to reflect the changes for remote participation in the study. This pack was also sent to potential participants who were hard-to-reach in the initial contact phase. Those participants invited to take part remotely who did not respond within 4 weeks of the remote participation pack being sent were contacted by telephone by a member of the core research team. This was to check whether the participant had received their pack and if they would like to discuss their potential participation further. During this telephone contact, participants interested in taking part were given the option to provide consent and complete the follow-up questionnaires via the paper-based method or verbally over the telephone.

Assessment methods