A sponsor designs a multi-site clinical trial intended for regulatory submission in Switzerland, the EU, and the United States. The protocol includes pregnant participants. The regulatory requirements across the three jurisdictions sound similar: benefit-risk assessment, ethics committee review, informed consent, safety monitoring, liability coverage. They are not similar in the ways that matter. The points of divergence create exposure that a single-jurisdiction analysis cannot identify and that the assumption of harmonization may obscure until the divergence produces a consequence the sponsor did not anticipate.
1. The Illusion of Convergence
All three jurisdictions are ICH members. All three are moving toward inclusion of pregnant participants in clinical trials. All three are expected to adopt ICH E21 once it is finalized as a Step 4 document, with finalization anticipated in Q1 2028.1ICH, ‘E21 Guideline on Inclusion of Pregnant and Breastfeeding Individuals in Clinical Trials’ (Draft, Step 2, endorsed 14 May 2025). Finalization anticipated Q1 2028. The convergence narrative is appealing: a single international guideline producing a harmonized approach across the three regulatory systems through which most global pharmaceutical development proceeds. The narrative is also incomplete. Beneath the shared policy direction, the regulatory architectures diverge on questions that determine whether a protocol can actually enroll pregnant participants: what ethics committees require, how liability is allocated, whether insurance is available, what constitutes adequate consent, and what happens when an adverse event occurs. These are not peripheral implementation details. For many sponsors, they are the factors that determine whether inclusion of pregnant participants is practically feasible regardless of what the science and the guideline support.
A sponsor who designs a single global protocol on the assumption that ICH E21 harmonization will resolve cross-border complexity may discover (at the ethics committee stage, the insurance procurement stage, or the adverse event stage) that the jurisdictional differences the guideline does not address are precisely the ones that matter most. The underlying evidence gap (one analysis estimates that, when a drug’s safety in pregnancy is unknown at the time of regulatory approval, an average of 27 years passes before sufficient evidence emerges to rate that safety with confidence)2MP Adam, JE Polifka and JM Friedman (2011) 157C Am J Med Genet C 175, cited in G Guglielmi (2026) 650 Nature 286: 27-year evidence-latency estimate. is shared across jurisdictions. The regulatory mechanisms for closing it are not.
2. Similar Words, Different Machinery
The Swiss Humanforschungsgesetz (HFG) establishes additional conditions for research involving pregnant women under Art. 26, implemented through the Verordnung über klinische Versuche mit Ausnahme klinischer Versuche mit Medizinprodukten (KlinV).3Art. 26 HFG (SR 810.30); Art. 19 KlinV (SR 810.305), as amended 7 June 2024, in force since 1 November 2024. The EU Clinical Trials Regulation (CTR) sets pregnancy-specific conditions in Art. 33, supplementing the general conditions of Art. 28.4Regulation (EU) No 536/2014 [2014] OJ L158/1 (CTR), Art. 33, supplementing Art. 28. The US framework operates through 45 CFR 46, Subpart B, providing additional protections for pregnant women, human fetuses, and neonates.545 CFR 46, Subpart B (Additional Protections for Pregnant Women, Human Fetuses and Neonates Involved in Research). All three frameworks sound protective. The word “protective” performs different work in each.
The Swiss framework operates through cantonal ethics committees whose expertise requirements for pregnancy trial review are not defined in the HFG or the KlinV, an ambiguity examined in Insight 27. The EU framework operates through a Part I/Part II assessment architecture in CTIS, where scientific assessment is coordinated by the reporting Member State but ethical assessment proceeds independently in each concerned Member State, producing the variation analyzed in Insight 28. The US framework operates through institutionally autonomous IRBs applying the “minimal risk” threshold without operational definition for fetal exposure, within a dual FDA/HHS regulatory structure where guidance recommends compliance with provisions that would not otherwise apply, the architecture dissected in Insight 29.
The structural divergence means that a protocol approved under one framework may face objections under another, not because the science or the risk profile differs, but because the institutional expectations applied to the same regulatory concept differ. A benefit-risk assessment that satisfies a Swiss cantonal ethics committee may not satisfy an Italian Part II assessment, which may not satisfy a US IRB applying Subpart B’s prospect-of-direct-benefit analysis. A sponsor is unlikely to resolve this through protocol design alone. The resolution depends on engagement with each jurisdiction’s decision-makers, whose criteria are not publicly harmonized and may not be internally consistent.
3. One Protocol, Three Consent Standards
The informed consent requirements for pregnant participants diverge across jurisdictions in ways that a global consent form template may not accommodate. Swiss law introduced electronic consent for clinical trials through Art. 7c KlinV, effective since 1 November 2024, a procedural modernization that no other jurisdiction has implemented in the same form for pregnancy trials.6Art. 7c KlinV (n 3) (electronic consent), in force since 1 November 2024. Art. 29 CTR governs informed consent generally; Art. 33’s pregnancy-specific conditions apply alongside but do not themselves modify the consent regime. Implementation is subject to Member State-level institutional expectations that may add requirements beyond the regulation’s text. The US framework introduces the Subpart B requirement for biological father consent in specified circumstances, a provision that has no equivalent in either Swiss or EU law and that interacts with evolving reproductive rights jurisprudence across US states in ways the regulation does not address.745 CFR § 46.204(e); see Insight 29 for detailed analysis.
The divergence is not merely formal. The underlying conception of the pregnant participant’s autonomy (balanced against fetal interests, partner consent requirements, and institutional oversight expectations) differs across jurisdictions in ways that affect protocol design, site selection, and the legal defensibility of the consent obtained. A consent document that satisfies Swiss requirements may omit provisions the US framework may expect. A consent document that satisfies Subpart B’s father-consent provision may include elements that have no legal basis under EU or Swiss law and whose inclusion could create confusion about the participant’s rights in those jurisdictions. Whether a consent obtained under one jurisdiction’s standards satisfies another’s requirements is a question the regulatory frameworks do not coordinate; a sponsor operating across all three would typically need to resolve through document design, site-level adaptation, and legal analysis that the consent form itself is unlikely to contain.
4. Liability and Insurance: The Unharmonized Risk
The liability dimension exposes the critical divergence that ICH E21 cannot resolve. Swiss liability law operates through the HFG regime with mandatory insurance and a burden-of-proof debate that remains pending before the BAG (Federal Office of Public Health).8Art. 19–20 HFG; BAG, ‘Evaluation of Liability Regulation in the Swiss Human Research Act’ (asim, University Hospital Basel, 2015; updated 2018). See Insight 27, Section 3. EU liability operates through Art. 76 CTR, deferring compensation mechanisms to twenty-seven national regimes with divergent approaches to causation, burden of proof, and damage calculation.9Art. 76 CTR (n 4); see Insight 28, Section 5. US liability operates without a federal compensation scheme, through state tort law, jury determination, and the learned intermediary doctrine. The Task Force on Research Specific to Pregnant Women and Lactating Women (PRGLAC) recommended in 2018 reducing liability barriers, with the Vaccine Injury Compensation Program invoked as a model; that recommendation remains unimplemented.10PRGLAC, ‘Report to the Secretary of Health and Human Services’ (September 2018), Recommendation 7; see Insight 29, Section 5.
A clinical trial agreement governing a multi-jurisdictional pregnancy trial would typically need to allocate risks whose magnitude depends on which jurisdiction’s law applies to a particular claim, a determination that may itself be contested. The insurance architecture would need to respond to actuarial uncertainty that is jurisdiction-specific: an insurer comfortable with the Swiss liability framework may balk at the US litigation environment, and an insurer willing to cover EU trials may require exclusions for particular Member States whose liability regimes present elevated exposure. The contractual allocation cannot override the applicable mandatory liability regime; it can only attempt to distribute the financial consequences of a regime the parties cannot control. The interaction between the clinical trial agreement’s governing law clause and the mandatory application of local liability regimes (the overriding mandatory provisions, or lois de police, that apply irrespective of the law the parties choose) creates a tension that standard choice-of-law provisions may not resolve, particularly where an adverse fetal or neonatal outcome gives rise to claims in multiple jurisdictions simultaneously.
5. Three Frameworks, One Adverse Event
When an adverse event occurs in a pregnancy trial running across all three jurisdictions, the reporting obligations diverge in timeline, format, institutional recipient, and the definition of what constitutes a reportable event. Swiss safety reporting under the KlinV requires notification to Swissmedic. EU reporting of suspected unexpected serious adverse reactions (SUSARs) proceeds via EudraVigilance under Art. 42 CTR, with annual safety reporting under Art. 43 and Member State cooperation in safety assessment governed by Implementing Regulation (EU) 2022/20.11Art. 42–43 CTR (n 4); Commission Implementing Regulation (EU) 2022/20 of 7 January 2022 [2022] OJ L5/14. US Investigational New Drug (IND) safety reporting operates under 21 CFR 312.32, with Form FDA 3500A (MedWatch) as the reporting mechanism. Each framework must be satisfied independently, and the determination of whether a particular fetal outcome constitutes a reportable event may yield different answers under each framework’s criteria.
The standardized outcome definitions developed by an international consortium led by Anna David and published in 2022 offer a voluntary harmonizing framework for grading maternal and fetal adverse events diagnosable in utero in maternal–fetal-therapy trials.12RN Spencer and others, ‘Development of standard definitions and grading for Maternal and Fetal Adverse Event Terminology’ (2022) 42(1) Prenatal Diagnosis 15. Their adoption, however, does not guarantee compliance with any jurisdiction’s specific reporting requirements. A sponsor using these definitions as its primary framework would still need to map them onto each jurisdiction’s regulatory expectations, a mapping exercise that produces different results depending on the jurisdiction. Whether a particular adverse in-utero finding constitutes a SUSAR under the CTR, triggers an IND safety report under 21 CFR 312.32, and requires notification to Swissmedic under the KlinV are three separate determinations, each governed by its own criteria, timelines, and institutional expectations. The coordination burden falls largely on the sponsor.
6. ICH E21: The Harmonization That Leaves the Hard Questions Open
ICH E21 is expected to harmonize the scientific and methodological framework: development strategy, nonclinical data requirements, study design principles, informed consent approaches, safety monitoring expectations, and labeling considerations. These are significant contributions. E21 spans both pregnant and breastfeeding individuals; this article addresses the pregnancy dimension. The guideline establishes a shared vocabulary and a common set of expectations that did not exist before its drafting.13ICH (n 1). EMA characterized ICH E21 as a “change in paradigm” in its 4 June 2025 news announcement.
What ICH E21 will not harmonize: liability regimes, insurance requirements, ethics committee composition and expertise standards, compensation mechanisms, IRB or ethics committee interpretive approaches to “minimal risk” or “direct benefit,” and the institutional relationships that determine whether a protocol proceeds from approval to enrollment. These are not peripheral concerns. For many sponsors, they are the factors that determine whether inclusion of pregnant participants is practically feasible regardless of what the science and the guideline support. The gap between harmonized aspiration and unharmonized mechanism is where legal, regulatory, and commercial complexity concentrates, and where the assumption that ICH E21 resolves the cross-border challenge may prove most costly.
The implementation timelines compound the uncertainty. Swissmedic has committed to adopting ICH E21 but has not specified a transposition timeline. The EMA’s Committee for Medicinal Products for Human Use (CHMP) adopted the draft at Step 2b on 22 May 2025, but Member State-level integration depends on national institutional processes the EMA cannot direct. The FDA published its notice of availability for the draft in the Federal Register on 21 July 2025 but has not initiated the Subpart B reform that would align the underlying regulatory architecture with the guideline’s expectations.14FDA, ‘E21 Inclusion of Pregnant and Breastfeeding Women in Clinical Trials’ 90 Fed Reg 34279 (21 July 2025), Docket FDA-2025-D-1797. A sponsor submitting a trial in 2028 or 2029 may encounter jurisdictions at different stages of E21 integration: one that has fully adopted the guideline’s expectations, another that has acknowledged but not implemented them, and a third where the existing regulatory text and the incoming guideline point in different directions without resolution.
7. The Coordination That Protocols Cannot Achieve
A sponsor pursuing a multi-jurisdictional pregnancy trial would typically need to coordinate regulatory submissions, ethics committee engagement, insurance procurement, clinical trial agreements, consent frameworks, and safety reporting across three systems that share a direction but not a mechanism. That coordination involves judgments about risk tolerance, institutional relationships, and commercial priorities that no regulatory framework addresses, because each framework addresses only its own jurisdiction, and the interaction between jurisdictions is the sponsor’s problem to solve.
The clinical trial agreement for a multi-jurisdictional pregnancy trial would need to accomplish what the regulatory frameworks do not: allocate liability across regimes that differ fundamentally in their approach to research-related injury, establish insurance coverage across markets that cannot price the risk, coordinate consent requirements that reflect different conceptions of participant autonomy, and address safety reporting obligations that may require different determinations about the same clinical event. Whether a particular agreement structure can achieve this depends on the trial’s specific design, the participating institutions’ risk appetite, the available insurance products, and the governing law’s interaction with the mandatory provisions of each participating jurisdiction, none of which standard templates accommodate.
The regulatory direction is clear and, from a public health perspective, welcome. The mechanism is not. It remains jurisdiction-specific, incompletely defined, and unresolved on the questions sponsors face when deciding whether to include pregnant participants. The evidence gap is shared. The path to closing it is not.