Teratogenicity in Pharmacovigilance

Teratogenicity refers to the capacity of an exposure to cause structural developmental abnormalities in an embryo or fetus. In pharmacovigilance, however, reproductive safety is broader than congenital malformations alone. Medicinal-product exposure can potentially affect embryonic or fetal survival, growth, functional development and postnatal development, and the type of effect depends strongly on dose, exposure timing, maternal physiology and the biological mechanism involved.

Purpose and Scope

This article explains teratogenicity as one part of developmental and reproductive toxicity, then connects that biology to human pregnancy pharmacovigilance, regulatory evaluation and risk minimisation.

The key distinction is important:

• reproductive toxicity is the broad category, encompassing effects on fertility, pregnancy and development;
• embryo-fetal toxicity includes harmful effects on the developing conceptus, including death, growth disturbance and structural or functional developmental effects; and
• teratogenicity is used more specifically for induction of structural malformations, although terminology can vary in scientific literature.

Using these terms precisely prevents every adverse pregnancy outcome from being labelled a teratogenic effect.

Current Regulatory Framework

Non-clinical reproductive and developmental toxicity assessment for human pharmaceuticals is addressed by ICH S5(R3). Post-authorisation pregnancy pharmacovigilance in the EU is governed through the broader pharmacovigilance legislation and GVP framework.

As of 2026, GVP Product- or Population-Specific Considerations III: Pregnant and breastfeeding women and their children exposed in utero or via breast milk is adopted and legally effective from 9 February 2026. In addition, GVP Module XVI Addendum I — Risk minimisation measures for medicinal products with embryo-fetal risks became effective on 29 August 2025.

These documents serve different purposes. ICH S5(R3) addresses non-clinical detection and characterisation of reproductive/developmental toxicity. The GVP pregnancy chapter addresses pharmacovigilance in exposed pregnancies and children, while Module XVI Addendum I addresses when and how risk-minimisation measures for embryo-fetal risks should be designed.

Developmental Timing Determines the Type of Risk

Embryonic and fetal development is sequential. Exposure to the same agent can therefore have different consequences depending on when it occurs.

Before and Around Implantation

Very early development involves a small number of rapidly dividing cells. Severe injury can result in loss of the conceptus, while surviving embryos may sometimes recover because cell lineages are not yet fully committed. This is often described as an “all-or-none” period, but the phrase is a simplification rather than an absolute biological rule.

Organogenesis

During organogenesis, major organ systems are being established. This is the period in which disruption of developmental pathways can produce characteristic structural malformations. For many human organs, the highest structural susceptibility lies approximately during weeks 3–8 after conception, although individual organs have different critical windows.

Fetal Development

After major organogenesis, growth, maturation and functional development continue. Exposure can affect growth, central nervous system development, endocrine function, organ maturation or fetal survival even when no major structural malformation is produced.

For pharmacovigilance, “exposed during pregnancy” is therefore insufficient description. Gestational timing relative to treatment is a central piece of causal evidence.

Mechanisms of Developmental Toxicity

A medicinal product can disturb development through multiple mechanisms. Examples include:

• interference with DNA synthesis or cell division;
• disruption of signalling pathways controlling differentiation;
• oxidative or metabolic injury;
• vascular disruption or impaired placental function;
• endocrine effects;
• antagonism of vitamins or essential developmental substrates;
• receptor or enzyme effects shared by maternal and fetal tissues; and
• indirect maternal toxicity severe enough to affect the developing fetus.

A mechanistic explanation strengthens biological plausibility but is not always required before a human teratogenic risk can be recognised.

Placental Transfer Is Not a Binary Property

The placenta is an active biological interface rather than a simple barrier. Fetal exposure depends on molecular size, lipophilicity, ionisation, protein binding, transporters, maternal metabolism, placental metabolism and gestational age.

Absence of high fetal concentrations does not automatically imply absence of developmental risk, and measurable placental transfer does not by itself establish teratogenicity. Exposure must be interpreted together with pharmacology, toxicology and observed outcomes.

Sources of Evidence

No single evidence source is sufficient for every developmental-risk question. Assessment usually integrates non-clinical findings with human exposure data, pharmacology and epidemiology.

Non-Clinical Reproductive Toxicology

ICH S5(R3) provides the framework for detecting reproductive and developmental toxicity in pharmaceuticals. Relevant programmes may examine fertility, embryo-fetal development and pre-/postnatal development according to the product and clinical context.

Animal findings can identify developmental hazards and help characterise dose-exposure relationships, affected developmental endpoints and possible mechanisms. Translation to humans nevertheless requires judgement because species differ in placentation, metabolism, developmental timing, target biology and exposure.

A negative animal study also does not prove zero human risk, particularly where human exposure differs from the tested exposure or where a mechanism is species-specific.

Clinical Trials

Pregnant people are often absent or uncommon in pre-authorisation trials, so direct human pregnancy evidence can be sparse at approval. Accidental pregnancies, partner pregnancies where relevant, and follow-up of trial participants can nevertheless contribute information.

The small number of exposed pregnancies usually limits the ability to estimate rare outcomes precisely.

Spontaneous Pregnancy Reports

Post-authorisation reports can identify unusual or recurrent patterns, especially for severe or distinctive congenital anomalies. Their limitations are substantial: incomplete gestational timing, uncertain dose, missing maternal disease information, concomitant exposures, selective reporting and incomplete follow-up are common.

A cluster of similar outcomes can be signal-generating, but spontaneous reports do not provide a reliable denominator for estimating absolute teratogenic risk.

Pregnancy Registries and Other Observational Data

Pregnancy registries, birth-defect registries, healthcare databases and purpose-designed observational studies can provide more systematic evidence. Their usefulness depends on design quality, exposure ascertainment, comparator choice, outcome validation and control of confounding.

Confounding by indication is especially important. The maternal disease being treated may itself influence miscarriage, prematurity, fetal growth or congenital anomaly risk.

Background Risk Is Essential Context

Congenital anomalies and pregnancy loss occur in pregnancies without medicinal-product exposure. A reported abnormal outcome therefore cannot be attributed to a medicine merely because exposure preceded it.

Assessment requires comparison with the expected background pattern, while recognising that background risks vary by outcome, population, maternal age, disease, ascertainment method and other factors.

For rare specific malformations, the pattern of defects may be more informative than the total number of malformed pregnancies.

Evaluating an Individual Pregnancy Exposure

A clinically useful pregnancy report should capture enough information to assess both exposure and outcome. Relevant elements can include:

• maternal age and relevant medical history;
• indication and severity of maternal disease;
• medicinal product, dose, route and treatment dates;
• gestational age at exposure;
• concomitant medicines and other exposures;
• prenatal investigations;
• pregnancy outcome;
• gestational age at outcome;
• congenital anomalies with diagnostic detail;
• neonatal condition; and
• longer-term developmental follow-up where clinically relevant and obtainable.

The purpose is not to collect every conceivable variable. It is to capture the information needed to understand the developmental question.

Pregnancy Reports and ICSR Requirements

EU GVP distinguishes pregnancy exposure from a report containing a suspected adverse reaction. Pregnancy exposure with an abnormal outcome such as congenital anomaly, developmental delay, fetal death, spontaneous abortion or a serious suspected neonatal reaction can meet ICSR reporting requirements according to the applicable criteria.

Exposure reports with no adverse outcome are not automatically reportable as ICSRs merely because pregnancy occurred. However, such exposures should still be collected and evaluated where relevant, and specific requirements can apply when pregnancy surveillance is a condition of the marketing authorisation or part of an RMP—for example, for medicinal products with high teratogenic potential.

The exact current GVP requirements should be applied rather than a generic rule that “all pregnancies are serious cases.”

Signal Detection for Teratogenicity

Developmental signals can be difficult to recognise because outcomes are heterogeneous and individual malformations are rare. A useful signal may arise from:

• repeated reports of the same or biologically related malformation;
• a distinctive pattern of anomalies;
• consistency with non-clinical findings;
• exposure concentrated in a plausible developmental window;
• an epidemiological excess relative to a suitable comparator; or
• a coherent mechanism or class effect.

The signal-management principle remains the same as for other safety questions: a signal is a hypothesis requiring evaluation, not a confirmed causal association.

Causal Reasoning

Assessment should integrate:

• timing of exposure relative to the developmental window;
• strength and consistency of the phenotype;
• dose and exposure where known;
• maternal disease and other causes;
• concomitant medicinal products;
• non-clinical concordance;
• epidemiological evidence;
• mechanistic plausibility; and
• evidence against the association.

For structural anomalies, positive rechallenge is generally neither expected nor ethically useful. Causal reasoning therefore relies more heavily on timing, phenotype, replication across independent data and comparative evidence.

Quantifying Risk

Where adequate data exist, risk can be described using absolute and relative measures. Both matter.

A relative increase can sound large while the absolute risk remains small; conversely, a moderate relative increase in a common serious outcome may have substantial clinical importance. Precision and uncertainty should therefore accompany point estimates.

Spontaneous-report counts should not be converted into incidence estimates because the number of exposed pregnancies and completeness of reporting are generally unknown.

Risk Management and Risk Minimisation

When embryo-fetal risk is established or sufficiently plausible, the regulatory response should be proportionate to the seriousness of the potential harm, strength of evidence, treatment context and feasibility of preventing exposure.

Possible measures can include routine product-information warnings, contraception recommendations, pregnancy testing, restrictions on use, educational materials and—where justified—a structured pregnancy prevention programme (PPP).

GVP Module XVI Addendum I provides the current EU framework for risk-minimisation measures for medicinal products with embryo-fetal risks. A PPP is not an automatic consequence of every non-clinical developmental finding. The need for additional measures depends on the product's embryo-fetal risk, target population, treatment alternatives, exposure circumstances and ability of routine measures to minimise risk.

Pregnancy Prevention Programmes

Where a PPP is justified, its components should form a coherent risk-control system rather than a collection of documents. Depending on the authorised programme, controls can address:

• clear communication of embryo-fetal risk;
• identification of people who can become pregnant;
• pregnancy testing where required;
• effective contraception before, during and, where relevant, after treatment;
• prescribing or dispensing controls;
• educational materials for healthcare professionals and patients; and
• procedures for pregnancy exposure and follow-up.

The exact measures are product-specific and should be taken from the authorised RMP/product information rather than copied from another teratogenic medicine.

Evaluating Whether Risk Minimisation Works

Implementation alone does not establish effectiveness. For products with important embryo-fetal risks, evaluation may consider whether intended users understand and follow the measures and whether pregnancy exposures occur despite the controls.

Process indicators can show whether materials were distributed or tests were performed. Outcome indicators can provide information about actual exposure or pregnancy outcomes. The choice of measure and threshold should be justified for the specific programme rather than imposed as a universal percentage.

Periodic Benefit-Risk Evaluation

Pregnancy and developmental data can contribute to PSUR/PBRER evaluation even when individual exposure reports are not themselves reportable ICSRs. Aggregate assessment may consider:

• numbers and characteristics of prospectively followed exposures;
• pregnancy outcomes;
• patterns of congenital anomalies;
• observed versus expected findings where valid comparators exist;
• new non-clinical or epidemiological evidence;
• effectiveness of pregnancy-specific risk minimisation; and
• remaining uncertainty.

The PBRER should avoid pooling incomparable data in a way that creates a misleading crude rate.

Relationship With the RMP

Teratogenic or embryo-fetal risk can be represented in an RMP according to the strength and importance of the evidence. The scientific question is not whether pregnancy data are limited—the question is whether an identified risk, potential risk or safety-relevant area of missing information is important enough to influence risk-management planning.

New evidence may therefore lead to improved characterisation, reclassification, additional pharmacovigilance or risk-minimisation changes, but these are not automatic consequences of a fixed exposure count.

QPPV Oversight

The QPPV should have appropriate visibility of important pregnancy-safety issues, emerging developmental signals, significant failures of pregnancy-prevention measures and regulatory actions affecting the pharmacovigilance system. This does not require the QPPV to personally follow every pregnancy report or approve every questionnaire.

The oversight system should make material issues visible early enough for scientific and regulatory action.

Illustrative Scenario

Consider a medicine for a severe inflammatory disease with a developmental toxicity signal in animals. Early post-marketing human pregnancies are few and show no consistent malformation pattern. Later, a well-designed database study provides a more precise estimate that does not show a major increase in overall congenital malformations but remains underpowered for a rare specific defect.

An inappropriate conclusion would be either “the medicine is teratogenic because animal studies were positive” or “human risk has been excluded because the overall study was negative.”

A stronger assessment would integrate species relevance, human exposure timing, the study's precision, the specific endpoint still unresolved and the consequences of maternal disease. Risk minimisation and further data collection would then be proportionate to the residual uncertainty.

Potential Failure Modes

The following are illustrative failure modes, not published inspection findings.

Practical Review Checklist

1. Is teratogenicity distinguished from broader reproductive and embryo-fetal toxicity?
2. Is gestational timing characterised as accurately as possible?
3. Are maternal disease, concomitant medicines and other confounders considered?
4. Are non-clinical findings interpreted for human relevance rather than copied literally?
5. Are spontaneous reports used for signal detection rather than incidence estimation?
6. Are prospective and retrospective pregnancy data distinguished where relevant?
7. Are comparator/background rates appropriate for the outcome and population?
8. Are supporting and reassuring data both considered?
9. Are pregnancy ICSR requirements applied according to current GVP rather than a blanket rule?
10. Are RMP classifications and risk-minimisation measures proportionate to evidence and importance?
11. Is PPP effectiveness evaluated where applicable?
12. Are important developmental-safety changes reflected consistently in PSUR/PBRER, RMP and product information?

Key Takeaways

Teratogenicity is a specific form of developmental toxicity involving structural malformations; reproductive safety is broader and includes fetal death, growth and functional development.

Developmental risk depends strongly on exposure timing, dose and mechanism, so pregnancy reports require gestational context.

Human assessment integrates non-clinical findings, spontaneous reports, registries, epidemiology and clinical context. No single source is sufficient in every case.

Pregnancy exposure without an adverse outcome is not automatically a reportable ICSR in the EU, although specific surveillance requirements can apply for high-risk products.

Risk minimisation for embryo-fetal hazards should be product-specific and proportionate. A pregnancy prevention programme is a structured intervention used when justified, not a universal response to every reproductive-toxicity concern.

References

1. International Council for Harmonisation. ICH S5(R3): Detection of Reproductive and Developmental Toxicity for Human Pharmaceuticals. Step 4, 18 February 2020.
2. European Medicines Agency. GVP Product- or Population-Specific Considerations III: Pregnant and breastfeeding women and their children exposed in utero or via breast milk. EMA/653036/2019 Corr.; legal effective date 9 February 2026.
3. European Medicines Agency. GVP Module XVI Addendum I — Risk minimisation measures for medicinal products with embryo-fetal risks. EMA/608947/2021; effective 29 August 2025.
4. European Medicines Agency. GVP Module VI — Collection, management and submission of reports of suspected adverse reactions to medicinal products (Rev. 2). EMA/873138/2011 Rev. 2.
5. European Medicines Agency. GVP Module V — Risk management systems (Rev. 2). EMA/838713/2011 Rev. 2.
6. European Medicines Agency. GVP Module VII — Periodic safety update report (Rev. 1). EMA/816292/2011 Rev. 1.
7. European Union. Commission Implementing Regulation (EU) No 520/2012, as amended.

Regulatory Note

As of 9 September 2026, GVP Product- or Population-Specific Considerations III on pregnancy and breastfeeding is adopted and legally effective from 9 February 2026. GVP Module XVI Addendum I on embryo-fetal risk minimisation is effective from 29 August 2025. Product-specific labelling, RMP commitments and pregnancy-prevention measures should always be checked against the current authorised documents rather than inferred from general guidance.