Walking the walk to include pregnant participants in non-obstetric clinical trials: Insights from the SNAP Trial

Despite several calls for greater inclusion of pregnant people in non-obstetric clinical trials, their systematic exclusion remains common practice (i.e. more than 95% of trials in recent estimates).^1,2 Excluding pregnant individuals from clinical trials may result in unintended consequences such as inadequate treatment of medical conditions in pregnancy, inappropriate dosing of medications, and investigational therapies being used off-label outside of the context of a clinical trial, risking adverse events in the absence of demonstrated efficacy.^2–4 Unique challenges in clinical trial design related to pregnancy have been identified as important barriers to inclusion of this population.^3,5 Our study group's experience of adapting the protocol of a randomized clinical trial to include pregnant participants may serve as an example for other research teams and, in turn, may contribute to increasing representation of pregnant participants in clinical trials.

The Staphylococcus aureus Network Adaptive Platform (SNAP) Trial is a multinational adaptive platform trial for S. aureus bloodstream infections. ⁶ This pragmatic trial includes participants across the lifespan with S. aureus grown from one or more blood culture and admitted to hospital at the anticipated time of eligibility assessment. ⁶ The SNAP Trial aims to improve treatment outcomes for patients with S. aureus bloodstream infections and the primary outcome is all cause mortality at 90 days after study entry. ⁶ Participants are randomized within one or more domains (i.e. “group of interventions with comparable modes of action”), functioning as sub-trials nested within a core protocol. ⁶ The adaptive design facilitates answering multiple questions through a single trial infrastructure, thereby optimizing resources, cost-effectiveness, and expediting results. ⁶ The initial domains are: (1) backbone antibiotic (options depend on antibiotic susceptibility of the S. aureus isolate); (2) adjunctive antibiotic (clindamycin vs. no clindamycin); and (3) early oral switch (early switch to oral antibiotics vs. continued intravenous antibiotics). ⁶

Colonization with S. aureus in peripartum patients is common and has been associated with invasive infections—such as skin and soft tissue infections, mastitis, pneumonia, and sepsis—as well as vertical transmission, and adverse neonatal outcomes.^7–10 Moreover, a prolonged course of intravenous antibiotics is often needed for patients with S. aureus bloodstream infections, including through peripherally inserted central lines, which are at increased risk of infectious and thromboembolic complications during pregnancy. ¹¹ S. aureus infections in pregnancy have also been linked to significant healthcare utilization and subsequent economic burden. ⁷ Importantly, sepsis is a leading cause of maternal mortality globally. ¹² As such, ongoing efforts to improve therapies for the management of this serious infection must not leave pregnant individuals nor their newborn infants behind.

We herein describe the steps undertaken to ensure participation of pregnant people in the SNAP Trial. At the outset of the trial planning stages, a specialized interdisciplinary working group, comprising clinician scientists with expertise in infectious diseases and other medical disorders in pregnancy, was formed to collaborate with the central trial steering committee and formulate pregnancy-specific recommendations for the trial methodology and conduct. The working group identified members for an independent pregnancy-specific Drug Safety and Monitoring Board. Subsequently, the SNAP pediatric and pregnancy working group designed a nested pregnancy domain and developed a pregnancy-specific appendix to the core protocol (Supplementary Appendix).

Eligibility criteria for the pregnancy domain aimed to be pragmatic and included a positive pregnancy test and/or patient-report of pregnancy. While scientifically complex, pregnant individuals do not meet the definition of a “vulnerable” population, and therefore, do not have a compromised ability to protect their interests or to provide informed, free, and ongoing consent. ¹³ As such, usual consent procedures for the core protocol remained applicable and a pregnancy-specific information sheet enclosing detailed information about the safety of drugs and planned follow-up was provided to eligible participants.

Investigational drugs of the core protocol were individually reviewed to examine whether they should be offered as options to eligible pregnant participants according to available human safety data. Given the wide range of investigational drugs with a well-established safety profile in pregnancy, it was decided that the rare options with no or scarce human pregnancy data (e.g. linezolid) would not be recommended initially. Additional pregnancy-specific monitoring was recommended for certain drugs (e.g. monitoring of liver enzymes with rifampicin) and medications with known fetal toxicities were excluded. Dosing of individual drugs was also reviewed to ensure that the highest end of the dosing range was planned, given the well described increased creatinine clearance in pregnancy. ¹⁴ All drugs offered as part of the backbone and adjunctive treatment domains of the core protocol were deemed to be acceptable options in pregnancy, and at similar dosing. Limitations only affected the early oral switch domain since certain drugs lacked data in human pregnancy.

Along with primary and secondary outcomes of the core protocol, pregnancy-specific endpoints were defined. These endpoints needed to balance the burden of added data collection and follow-up, with the need for essential pregnancy-specific information. We opted to evaluate pregnancy outcomes (i.e. gestational age at birth, spontaneous pregnancy loss, intrauterine fetal death, and stillbirth), pregnancy-specific safety outcomes (i.e. complications from central venous catheters requiring line removal, and catheter-associated venous thromboembolism), and neonatal outcomes (i.e. birth weight, neonatal intensive care unit admission, and neonatal mortality) according to scheduled study visits at 10–12, 24–28, 36–38 weeks of gestation, at delivery, and at 6 weeks postpartum.

The trial in its current form, with eligibility of pregnant participants, has been granted national operational funding and has successfully obtained regulatory approval with the trial sponsors, insurers and institutional review boards in Australia, New Zealand, Canada, Israel, and Singapore. Recruitment has commenced in in each of these countries with 515 enrolled participants to date (7 March 2023), including pregnant individuals (n < 10). We and others have demonstrated that with careful and diligent planning by a multidisciplinary team dedicated to considering pregnancy-related complexities of trial design, study protocols can be successfully adapted to accommodate pregnant participants. Accordingly, participation of pregnant individuals in non-obstetric clinical trials can and should be offered, particularly in trials addressing severe and potentially life-threatening conditions. An assumption of eligibility of pregnant individuals for clinical trials instead of a culture of exclusion must become the norm going forward.

Supplemental Material