GVP Product- or Population-Specific Considerations III: Pregnancy and Breastfeeding

Introduction

Pharmacovigilance during pregnancy and breastfeeding requires a wider clinical assessment than routine evaluation in the general adult population. The medicine may affect the woman, but exposure can also affect an embryo, fetus, neonate, infant or child. At the same time, withholding effective treatment may itself create risks for the woman and the developing or breastfed child.

This makes pregnancy and breastfeeding a particularly important example of why pharmacovigilance is not simply the identification of adverse events. The assessment must place exposure within the timing of pregnancy or lactation, the underlying disease, alternative treatments, background risks and the quality of the available evidence.

The current EMA guidance states that the benefit-risk evaluation in these populations contains additional elements compared with evaluation in the overall population, and that patients and healthcare professionals should be provided with information that supports therapeutic decision-making while uncertainty remains. citeturn0search12

1. Scope of the Guidance

The product- or population-specific consideration covers medicinal-product use during pregnancy and breastfeeding and the resulting exposure of children in utero or through breast milk.

The scope extends beyond confirmed adverse reactions. Pharmacovigilance must also address exposure information, pregnancy outcomes, neonatal and infant outcomes, and the circumstances in which exposure occurred.

The guidance therefore supplements the general pharmacovigilance framework rather than replacing the general requirements for case collection, signal management, risk management and aggregate safety evaluation.

2. Why Pregnancy Requires a Distinct Approach

Pregnancy changes both exposure and clinical context over time. The developmental stage at exposure can be as important as the medicine and dose themselves.

The consequences being evaluated may also occur at different stages. Some outcomes concern early pregnancy, others fetal development, pregnancy complications, delivery, neonatal adaptation or later infant development.

The assessment therefore needs enough temporal information to understand when exposure occurred and what outcome followed.

3. The Benefit-Risk Balance Has Two Patients in View

In pregnancy, the benefit-risk assessment must consider the woman and the developing child.

The medicine may provide important therapeutic benefit to the woman, while untreated maternal disease may itself increase risks to both the woman and fetus. The pharmacovigilance assessment should therefore not frame the question simply as whether medicine exposure creates fetal risk.

The clinically meaningful question is how the risks and benefits compare with the alternatives available in the particular circumstances.

4. Background Risk Is Essential

Many adverse pregnancy outcomes occur without medicine exposure.

Congenital anomalies, miscarriage, preterm birth and other outcomes have background frequencies that vary with maternal characteristics, disease, age, genetics, environmental factors and other exposures.

Consequently, observing an outcome after medicine exposure does not establish that the medicine caused it. Interpretation requires an appropriate understanding of the expected background risk and the characteristics of the exposed population.

5. Exposure Timing

Timing is one of the most important pieces of information in pregnancy-related pharmacovigilance.

The assessment may require the estimated gestational age, exposure dates, dose and duration, and the timing of relevant pregnancy or fetal outcomes. A medicine exposure before conception may need to be considered differently from exposure during a particular developmental period.

Incomplete timing information can substantially limit interpretation and should therefore be addressed through appropriate follow-up where feasible.

6. Pregnancy Exposure Does Not Equal Adverse Outcome

A pregnancy exposure report may be clinically important even when no adverse outcome has been reported.

Exposure information can contribute to understanding the frequency and circumstances of use and may support future assessment of pregnancy outcomes. Conversely, a reported pregnancy outcome following exposure requires careful assessment rather than automatic attribution to the medicine.

The pharmacovigilance system should therefore preserve the distinction between exposure and outcome.

7. Pregnancy Exposure Cases

A pregnancy exposure case may originate from a spontaneous report, healthcare professional contact, patient report, literature, a study, a registry or another organised source.

The source affects how the information is processed, but the clinical assessment still requires reconstruction of the exposure and pregnancy history.

Where follow-up is appropriate and possible, the organisation should seek information that materially improves the ability to evaluate the pregnancy and its outcome.

8. Prospective and Retrospective Information

Information collected prospectively during pregnancy may provide different evidential value from information obtained retrospectively after an outcome has occurred.

Retrospective reports can be affected by recall limitations and selective reporting, while prospective follow-up may provide more reliable exposure and outcome timing.

Neither category should be treated as automatically valid or invalid. The design and quality of the information determine its contribution to the assessment.

9. Congenital Anomalies

Congenital anomalies require particular care because their baseline frequency, classification and causes are complex.

The assessment should consider the specific anomaly or phenotype, timing of exposure, maternal disease, concomitant exposures, family history where relevant and other plausible explanations.

Broad statements about "birth defects" can obscure clinically important differences between outcomes and should not replace appropriate clinical characterisation.

10. Pregnancy Loss

Pregnancy loss can include different clinical outcomes with different background rates and potential causes.

The pharmacovigilance assessment should therefore identify the type and timing of the loss where possible and consider maternal disease, gestational age, concomitant exposures and other relevant factors.

A pregnancy loss following exposure is an important safety observation, but the temporal relationship alone is insufficient to establish causality.

11. Preterm Birth and Other Pregnancy Outcomes

Outcomes such as preterm birth, fetal growth abnormalities, hypertensive disorders or other pregnancy complications may be influenced by the underlying disease as well as treatment.

The assessment should therefore consider whether the outcome is associated with the indication, disease severity, maternal characteristics, treatment exposure or other factors.

Where comparative evidence exists, it may be particularly useful for separating treatment-associated risk from background and disease-associated risk.

12. Maternal Outcomes

Pharmacovigilance should not focus exclusively on fetal outcomes.

The medicine may affect maternal health during pregnancy, and pregnancy itself may alter the clinical presentation or management of an adverse reaction. The underlying disease may also change during pregnancy or postpartum.

A complete assessment therefore considers both maternal and child outcomes and the relationships between them.

13. The Risk of Untreated Disease

An assessment of medicine use during pregnancy must account for the potential consequences of untreated maternal disease.

Discontinuation or avoidance of an effective medicine may increase maternal disease activity, which may itself adversely affect pregnancy or fetal outcomes. This makes simplistic recommendations to avoid exposure potentially harmful.

The pharmacovigilance objective is to support an informed therapeutic decision rather than to minimise medicine exposure regardless of clinical context.

14. Breastfeeding Requires a Different Exposure Model

During breastfeeding, the child is exposed through breast milk rather than directly through maternal administration.

The assessment therefore needs to consider maternal dose, timing of administration, transfer into milk, infant exposure and the infant's ability to handle the medicine. The clinical relevance of exposure depends on the medicine and the characteristics of the infant.

The benefits of breastfeeding also form part of the benefit-risk assessment.

15. Effects on Breast Milk and Lactation

A medicine may affect breastfeeding independently of direct infant exposure.

Potential effects on milk production, composition or the ability to breastfeed may therefore be relevant. The assessment should distinguish these effects from adverse reactions occurring in the breastfed child.

This distinction can affect both clinical recommendations and the evidence required for risk characterisation.

16. Infant Outcomes During Breastfeeding

Potential outcomes in a breastfed infant should be evaluated in relation to exposure through milk and other possible causes.

Relevant information may include the timing of maternal treatment, dose, duration of breastfeeding, infant age, symptoms, growth and development, and any other infant medicines or exposures.

The absence of an observed infant reaction does not necessarily establish absence of pharmacological exposure, while an observed event does not establish causality.

17. Children Exposed In Utero or Through Breast Milk

The child is a separate clinical subject for the purpose of safety assessment even though exposure originated from maternal treatment.

The pharmacovigilance record should therefore retain sufficient information about the child, the exposure pathway and the outcome to support appropriate evaluation.

Longer-term outcomes may be relevant when a plausible concern cannot be adequately assessed at birth or shortly after exposure.

18. Women of Childbearing Potential

Pregnancy may be unplanned, and exposure can occur before pregnancy is recognised.

Pharmacovigilance and risk-management planning should therefore consider women of childbearing potential rather than treating pregnancy as an exposure that can always be anticipated and prevented.

Where a medicine has specific pregnancy-related risks or precautions, the effectiveness of those measures should be considered within the applicable risk-management framework.

19. Pre-Conception Exposure

Exposure before conception may require separate consideration depending on the medicine and its persistence or biological effects.

The relevant period may extend beyond the day treatment is stopped when the medicine or its effects persist. The assessment should therefore use pharmacological and clinical knowledge to define the relevant exposure window rather than relying on an arbitrary calendar boundary.

20. Follow-Up of Pregnancy Reports

Follow-up can materially improve the value of pregnancy exposure information.

The objective is not to collect every conceivable detail. It is to obtain information that can clarify exposure, timing, pregnancy outcome, maternal and child factors and clinically relevant alternative explanations.

Follow-up should therefore be proportionate, respectful and clinically appropriate, particularly when pregnancy outcomes are sensitive or distressing.

21. Loss to Follow-Up

Pregnancy cases can be difficult to follow because the outcome may occur months after the original report and the reporter may not remain accessible.

The organisation should use a controlled approach to follow-up while recognising practical and ethical limits. Failure to obtain an outcome should not be interpreted as evidence of a favourable outcome.

The limitation should remain visible in the safety record.

22. Registries and Other Structured Data Sources

Pregnancy registries and other structured data sources can provide valuable information because they may allow systematic collection of exposure and outcome information.

Their design, inclusion criteria, comparator selection, follow-up and completeness determine the evidential value of the resulting data.

Registry findings should therefore be interpreted according to the design and question being addressed rather than simply because the information was collected prospectively.

23. Studies in Pregnancy

Clinical and observational studies may provide comparative evidence that cannot be obtained from spontaneous reports alone.

Important considerations include the study population, exposure definition, comparator, outcome ascertainment, confounding and follow-up duration.

A study designed to estimate pregnancy outcomes may answer a different question from one designed to investigate a specific congenital anomaly or developmental outcome.

24. Signal Management in Pregnancy

Pregnancy-related information can generate signals through individual cases, aggregate review, studies, registries or other sources.

The general signal-management principles continue to apply, but the clinical context may require specialised assessment of timing, background risk and competing explanations.

A pregnancy signal should therefore enter the general signal-management system while retaining the population-specific information necessary for meaningful assessment.

25. Benefit-Risk Assessment During Breastfeeding

Breastfeeding requires consideration of benefits and risks for both mother and child.

The medicine may provide substantial benefit to the mother, while breastfeeding may itself provide important benefits to the infant. The assessment should therefore consider the alternatives: continuing treatment and breastfeeding, modifying treatment, temporarily interrupting breastfeeding or using another treatment where clinically appropriate.

The pharmacovigilance role is to provide evidence that supports this clinical decision rather than prescribe a universal response.

26. Communication of Pregnancy and Breastfeeding Risks

Safety communication should present pregnancy and breastfeeding evidence in a way that supports informed therapeutic decisions.

Where evidence is limited, the communication should make the uncertainty visible. Where a risk has been characterised, the information should distinguish the established risk from the background risk and explain relevant clinical implications.

Communication should avoid creating unnecessary fear that could lead patients to stop beneficial treatment without appropriate medical advice.

27. Product Information and Risk Management

Evidence concerning pregnancy or breastfeeding may lead to consideration of product-information changes or risk-management measures.

The scientific assessment should establish the evidence and uncertainty. Regulatory assessment then determines the appropriate product-information or risk-management outcome.

A pregnancy exposure signal should not automatically result in a contraindication or recommendation to avoid treatment. The regulatory response depends on the totality of evidence and the clinical benefit-risk balance.

28. Governance and QPPV Oversight

Significant pregnancy or breastfeeding safety concerns should be visible within the pharmacovigilance governance system.

The QPPV does not need to perform every pregnancy-case assessment, but the QPPV should have appropriate oversight of significant risks, signal assessments, risk-management interfaces and regulatory consequences.

The organisation should be able to demonstrate that specialist expertise is available and that significant uncertainties are escalated appropriately.

29. Scientific Interpretation of Pregnancy Evidence

Pregnancy safety assessment requires integration of several evidence streams rather than reliance on a single source.

Individual reports provide clinical detail, but may be affected by reporting and selection biases. Population studies can provide comparative estimates but may be affected by confounding. Registries can improve follow-up but depend on their design and completeness. Literature and external surveillance may identify concerns that require further evaluation.

The assessment should therefore explain what each source contributes and where uncertainty remains.

30. Confounding by Indication

Confounding by indication is particularly important in pregnancy pharmacovigilance because the underlying condition may itself influence pregnancy outcomes.

For example, a medicine used for a serious maternal disease may appear associated with an adverse pregnancy outcome because the disease, rather than the treatment, contributes to the outcome.

Comparative study designs and appropriate adjustment can help address this problem, but residual confounding may remain. The limitation should be considered when interpreting both positive and negative findings.

31. Concomitant Exposures

Pregnant women may receive several medicines during the same period, making attribution difficult.

The assessment should consider the timing and duration of each exposure and whether another medicine provides a more plausible explanation for the observed outcome.

Concomitant exposure does not automatically exclude a medicinal product from consideration. It changes the causal question and may require more detailed analysis.

32. Maternal Disease as an Alternative Explanation

The underlying maternal disease may affect fetal or neonatal outcomes directly or through changes in maternal physiology.

A credible assessment should therefore describe the disease severity, duration, complications and relevant clinical characteristics where available.

The aim is not to attribute every outcome to the disease. It is to prevent medicine exposure from becoming the default explanation simply because it occurred in temporal proximity to the outcome.

33. Teratogenicity and Developmental Toxicity

Potential developmental effects require consideration of exposure timing, developmental biology, dose and the phenotype observed.

A suspected teratogenic effect should not be inferred merely from an individual congenital anomaly after exposure. The assessment requires consideration of the specific phenotype and whether the timing and biological mechanism are compatible with the proposed effect.

Human evidence should also be interpreted alongside relevant non-clinical information where appropriate.

34. Neonatal Effects

Exposure late in pregnancy may be relevant to neonatal outcomes even when fetal structural development is not the principal concern.

Possible issues include neonatal adaptation, withdrawal, pharmacological effects or other consequences of exposure near delivery.

The timing of the final maternal dose, gestational age at delivery and neonatal clinical course can therefore be important to interpretation.

35. Long-Term Child Outcomes

Some potential effects may not be apparent at birth.

Where a plausible concern involves development, neurodevelopment or another outcome that requires longer observation, the evidence-generation strategy may need longer follow-up.

The absence of short-term neonatal findings should not automatically be interpreted as evidence that a potential longer-term effect has been excluded.

36. Breast Milk Exposure and Pharmacology

For breastfeeding, the clinical relevance of medicine transfer into milk depends on the concentration in milk, timing of exposure, infant intake, oral bioavailability and the infant's capacity to metabolise and eliminate the medicine.

These factors mean that the mere detection of a medicine in breast milk does not establish a clinically significant risk to the infant.

Conversely, a low measured concentration may not be sufficient to exclude risk for a particularly vulnerable infant or a medicine with a narrow therapeutic margin.

37. Infant Age and Maturity

The significance of exposure through breast milk can vary with infant age and maturity.

Premature infants, neonates and older infants may differ in organ maturity, feeding volume and pharmacokinetic handling. The clinical assessment should therefore consider the characteristics of the exposed infant rather than treating all breastfeeding exposure as equivalent.

38. Exposure Through Breast Milk Versus Direct Infant Treatment

A medicine detected in an infant receiving breast milk should not automatically be interpreted as equivalent to therapeutic administration to the infant.

The dose received through milk may be substantially different, and the clinical context is different. The assessment should distinguish indirect exposure from direct paediatric use while considering whether pharmacological effects are plausible.

39. Lactation Failure and Maternal Effects

A medicine may affect the ability to initiate or maintain breastfeeding even when there is no direct adverse effect on the infant.

The assessment should therefore consider maternal lactation outcomes where clinically relevant. Such effects may alter infant nutrition and consequently become part of the wider safety assessment.

40. Pregnancy and Breastfeeding in the RMP

Pregnancy or breastfeeding may constitute an important area of missing information or a specific safety concern within a risk-management system.

The RMP should reflect the relevant evidence gaps and planned pharmacovigilance activities where applicable. The activities should have a defined purpose: for example, characterising pregnancy outcomes, improving exposure ascertainment or addressing a specific uncertainty.

The existence of an RMP activity does not itself establish that a safety risk has been confirmed.

41. Additional Pharmacovigilance Activities

Additional activities may include pregnancy registries, enhanced follow-up, observational studies, database studies, literature monitoring or other suitable approaches.

The choice should follow from the information gap. A registry may be useful where prospective exposure and outcome collection is needed, while a comparative database study may be more appropriate when the central question concerns relative risk.

Evidence generation should therefore be hypothesis-driven and proportionate.

42. Data Quality

Pregnancy and breastfeeding data can be incomplete in ways that materially affect interpretation.

Missing gestational age, uncertain exposure dates, incomplete outcome descriptions, unknown concomitant medicines or inadequate follow-up may prevent meaningful analysis.

Data-quality controls should therefore focus on information that changes the scientific interpretation rather than collecting large quantities of detail without a defined purpose.

43. Coding and Clinical Terminology

The choice of coding and outcome classification can influence apparent patterns in pregnancy safety data.

Clinically related but distinct outcomes should not be merged simply to increase case numbers, while overly narrow coding can fragment a clinically meaningful signal.

The organisation should use clinically appropriate definitions and maintain consistency across analyses so that apparent changes are not artefacts of classification.

44. Denominators and Exposure Data

Pregnancy safety assessment often requires some understanding of the number and characteristics of exposed pregnancies.

Report counts alone cannot establish incidence without an appropriate denominator. Exposure estimates may nevertheless be difficult because pregnancy may be unplanned and medicine use may change after pregnancy recognition.

The limitations of exposure estimates should be explicit when interpreting rates or comparative risks.

45. Negative Findings

A study that does not identify an increased risk can provide important evidence, but the strength of that evidence depends on study size, outcome ascertainment, exposure classification and other methodological factors.

A negative study with adequate power for the specific outcome may substantially reduce uncertainty. A small study with incomplete follow-up may provide little reassurance.

The assessment should therefore distinguish evidence against an association from an absence of informative evidence.

46. Signals From Individual Cases

Individual pregnancy reports can be highly informative when the phenotype and exposure timing are well characterised.

A cluster of clinically similar cases may warrant further investigation even when the total number remains small. Conversely, heterogeneous cases with poor timing information may provide limited evidence of a common causal mechanism.

Case quality should therefore influence the weight assigned to the observation.

47. Clusters of Pregnancy Outcomes

A cluster may attract attention because several similar outcomes occur after exposure within a relatively short period.

Clustering can reflect a genuine safety issue, chance, changes in exposure, stimulated reporting or shared background factors. The organisation should investigate the cluster while considering the appropriate expected frequency and population characteristics.

A cluster is therefore a reason for structured assessment, not a conclusion of causality.

48. Reporting Bias

Pregnancy outcomes can be subject to selective reporting because unusual or adverse outcomes are more likely to be reported than uncomplicated exposures.

This can make crude reporting patterns difficult to interpret. Prospective data sources and appropriate denominators can help address some of these limitations.

The presence of reporting bias does not make spontaneous reports unusable; it changes the questions they can answer reliably.

49. Surveillance During Public Health Events

Pregnancy and breastfeeding pharmacovigilance may become particularly important during infectious-disease outbreaks, vaccination campaigns or other public-health events.

Large numbers of exposures can generate substantial reporting volume, while the underlying infection or public-health condition may itself affect pregnancy outcomes.

The assessment should therefore distinguish medicine-related risk from risks associated with the underlying public-health event and changes in healthcare utilisation.

50. Communication During Pregnancy

Pregnancy-related safety communication requires particular attention to the consequences of misunderstanding.

Patients may discontinue treatment abruptly because of concern for the fetus, even when untreated disease presents a significant risk. Conversely, inadequate communication may lead to exposure without appropriate consideration of a characterised risk.

The message should therefore support informed discussion between patient and healthcare professional and communicate uncertainty honestly.

51. Communicating Absolute and Relative Risk

Where quantitative estimates are available, communication should use measures that help the audience understand the practical magnitude of risk.

Relative increases can appear substantial when baseline risk is low. Absolute risks may therefore be more informative for clinical decision-making, provided the underlying estimate is sufficiently reliable.

The choice of presentation should remain faithful to the evidence and avoid implying greater precision than the data support.

52. Avoiding Unnecessary Treatment Avoidance

The purpose of pregnancy pharmacovigilance is not to minimise medicine use at any cost.

If a medicine effectively controls a serious maternal disease, avoiding treatment may itself cause harm. Safety communication should therefore present the known risk alongside the consequences of untreated disease and the available alternatives where relevant.

This is particularly important when evidence is incomplete and patients are making time-sensitive treatment decisions.

53. Regulatory Assessment

When pregnancy or breastfeeding evidence raises a significant concern, the regulatory assessment may consider product information, risk-management measures, additional studies or other appropriate actions.

The pharmacovigilance assessment provides the evidence base, but the regulatory outcome is a separate decision governed by the applicable regulatory framework.

The organisation should preserve the reasoning that connects the safety evidence to the proposed regulatory response.

54. Inspection Perspective

An inspector evaluating pregnancy pharmacovigilance may examine whether the organisation can identify relevant exposures, obtain meaningful follow-up, evaluate outcomes appropriately and integrate the information into signal and risk-management processes.

The inspection question is not simply whether pregnancy reports were collected. It is whether the system can convert those reports into useful safety knowledge and appropriate action.

55. Illustrative Inspection Scenario: Exposure Recorded Without Outcome Follow-Up

The organisation has a substantial collection of pregnancy exposure cases but does not have a controlled process for seeking pregnancy outcomes where follow-up is appropriate.

The potential weakness is that the system is collecting exposure information without converting it into the outcome data needed for safety characterisation.

56. Illustrative Inspection Scenario: Pregnancy Outcomes Attributed Without Background Context

An internal analysis counts congenital anomalies following medicine exposure and describes the number as evidence of increased risk without considering the expected background frequency or the characteristics of the exposed population.

The potential weakness is an interpretation problem rather than a data-collection problem. Counts require appropriate context before they can support a causal conclusion.

57. Illustrative Inspection Scenario: Breastfeeding Exposure Treated as Direct Infant Dosing

An assessment assumes that detection of a medicine in breast milk means the infant received a clinically equivalent dose.

The potential weakness is failure to distinguish indirect exposure through milk from direct administration. Infant dose, bioavailability, age, pharmacology and clinical outcome require separate consideration.

58. Illustrative Inspection Scenario: Risk Communication Ignores Untreated Disease

A communication advises women to avoid treatment during pregnancy without explaining that untreated disease may also present risks.

The potential weakness is an incomplete benefit-risk message. Pregnancy pharmacovigilance requires consideration of both treatment-related risks and the consequences of untreated disease.

59. A Structured Assessment Model

Pregnancy and breastfeeding assessment can be organised around a sequence that keeps exposure, outcome and clinical context connected:

Identify exposure
      ↓
Characterise timing and dose
      ↓
Characterise mother and child
      ↓
Define the clinical outcome
      ↓
Assess background and disease-related risk
      ↓
Consider concomitant exposures
      ↓
Evaluate biological and temporal plausibility
      ↓
Integrate comparative and other evidence
      ↓
Characterise uncertainty
      ↓
Scientific conclusion
      ↓
Risk-management / regulatory decision
      ↓
Communication and follow-up

The sequence is not intended to make every case complex. It provides a framework for ensuring that important clinical distinctions are not lost during routine processing.

60. The Difference Between Case Processing and Population Assessment

Individual cases are processed according to the general pharmacovigilance requirements, but population-level pregnancy assessment asks additional questions.

The organisation may need to understand how many pregnancies were exposed, what outcomes occurred, whether patterns are changing and whether the observed findings differ from appropriate background expectations.

Case processing therefore supplies essential observations, while aggregate and epidemiological assessment determine what those observations mean at population level.

61. Linking Pregnancy Data to Signal Management

Pregnancy information should be capable of entering the signal-management process when a potential new or changed safety concern emerges.

A signal may concern a specific congenital anomaly, pregnancy complication, neonatal effect, breastfeeding outcome or broader pattern of exposure.

The signal record should preserve the population-specific evidence and explain whether the finding represents a new risk, a new aspect of a known risk or insufficient evidence for a conclusion.

62. Linking Pregnancy Data to Aggregate Reporting

Pregnancy and breastfeeding information may also contribute to PSURs and other aggregate evaluations.

Aggregate reporting provides an opportunity to review accumulated exposure and outcomes over time and to reconsider previous conclusions as evidence grows.

The organisation should avoid treating pregnancy cases as a separate dataset that never feeds into the broader safety profile.

63. Integration With Medical Information

Medical-information contacts can provide early evidence about pregnancy exposure, treatment decisions and patient concerns.

Where relevant safety information is received, it should enter the appropriate pharmacovigilance pathway. At the same time, medical-information responses should remain consistent with the current approved safety position.

This interface is particularly important when patients contact organisations because of concerns about treatment during pregnancy or breastfeeding.

64. Integration With Clinical Development

Pre-authorisation clinical development generally provides limited pregnancy exposure because pregnant women are often excluded or underrepresented in trials.

Post-authorisation pharmacovigilance therefore becomes particularly important when exposure occurs in routine clinical practice. Information from clinical development, however, can still provide relevant baseline safety information and mechanistic context.

The two evidence streams should be integrated rather than treated as competing sources.

65. Non-Clinical Evidence

Animal reproductive and developmental toxicity data can provide mechanistic context for human pregnancy safety assessment.

Such evidence may support biological plausibility or identify areas requiring particular attention, but findings in animals cannot automatically be translated into the magnitude or clinical relevance of human risk.

Human clinical and epidemiological evidence remains essential to characterising risk in humans.

66. Exposure During Unplanned Pregnancy

Unplanned pregnancies can result in medicine exposure before the woman is aware of the pregnancy.

This makes early exposure difficult to prevent and reinforces the importance of accurate exposure information and appropriate counselling where relevant.

The pharmacovigilance system should be able to distinguish exposure before pregnancy recognition from deliberate continued treatment and assess both within the appropriate clinical context.

67. Exposure Near Conception

Exposure around conception can raise questions about whether the relevant biological window precedes implantation or extends into early embryonic development.

The assessment should use the pharmacology and reproductive biology of the medicine rather than applying a generic assumption about when exposure becomes relevant.

This is another example of why precise exposure dates can materially improve scientific interpretation.

68. Postpartum Considerations

The period immediately after delivery may involve changes in maternal treatment, breastfeeding status and infant exposure.

A medicine administered during pregnancy may continue to be relevant postpartum, while a new medicine started after delivery may create a different breastfeeding exposure scenario.

The pharmacovigilance record should therefore preserve the transition between pregnancy and breastfeeding where it affects exposure and outcome assessment.

69. Ethical and Privacy Considerations

Pregnancy and breastfeeding information can contain highly sensitive personal and clinical details.

Follow-up should therefore be clinically justified, proportionate and handled according to applicable privacy and data-protection requirements. The desire for complete data should not lead to unnecessary collection of sensitive information.

Good pharmacovigilance requires both scientific adequacy and appropriate handling of the people whose experiences provide the evidence.

70. What Makes a Pregnancy Safety Assessment Defensible?

A defensible assessment should make clear:

• what medicine was taken;
• by whom;
• at what dose and for how long;
• when exposure occurred relative to pregnancy or breastfeeding;
• what outcome occurred;
• what alternative explanations were considered;
• what background risk is relevant;
• what other evidence exists;
• what remains uncertain;
• and what action follows from the assessment.

The level of detail should be proportionate to the question and the potential significance of the finding.

71. What Makes the System Effective?

An effective pregnancy and breastfeeding pharmacovigilance system does more than accumulate cases.

It identifies exposures, obtains meaningful outcomes, characterises the relevant clinical context, integrates evidence across sources, recognises important signals and converts conclusions into appropriate risk-management and communication decisions.

Effectiveness therefore depends on the interfaces between case processing, medical assessment, epidemiology, signal management, RMP, regulatory affairs and communication.

72. Final Principle

Pregnancy and breastfeeding pharmacovigilance is fundamentally an exercise in balanced clinical reasoning.

The medicine may create risk, but untreated disease may also create risk. The observed outcome may follow exposure, but temporal sequence does not establish causality. A study may find no association, but the strength of that reassurance depends on its ability to detect the outcome. A risk may be real, but its clinical significance depends on the magnitude of risk and the benefits of treatment.

The purpose of the pharmacovigilance system is to make these distinctions visible and usable in clinical and regulatory decision-making.

Key Takeaways

Pregnancy and breastfeeding require population-specific pharmacovigilance because medicine exposure can affect both the treated woman and a developing or breastfed child, while untreated maternal disease and loss of beneficial treatment can also create risks.

The scientific assessment depends heavily on exposure timing, clinical phenotype, background risk, maternal disease, concomitant exposures and appropriate comparative evidence. Pregnancy exposure should not be equated with an adverse outcome, and an outcome following exposure should not be equated with causality.

For breastfeeding, the assessment additionally considers exposure through milk, infant age and maturity, pharmacology, infant outcomes and effects on lactation itself. The benefits of breastfeeding and maternal treatment form part of the wider benefit-risk assessment.

The mature system links individual case processing with population-level assessment, signal management, aggregate reporting, risk management, communication and regulatory decision-making. Its purpose is not to eliminate uncertainty, but to characterise it sufficiently to support proportionate and scientifically defensible decisions.

References

1. European Medicines Agency. Good Pharmacovigilance Practices (GVP), Product- or Population-Specific Considerations III: Pregnant and breastfeeding women and their children exposed in utero or via breastmilk, EMA/653036/2019 Corr.
2. European Medicines Agency. Good Pharmacovigilance Practices (GVP), Module VI — Collection, management and submission of reports of suspected adverse reactions to medicinal products.
3. European Medicines Agency. Good Pharmacovigilance Practices (GVP), Module IX — Signal Management.
4. European Medicines Agency. Good Pharmacovigilance Practices (GVP), Module V — Risk management systems.
5. European Medicines Agency. Good Pharmacovigilance Practices (GVP), Module VII — Periodic safety update report.
6. Regulation (EC) No 726/2004, as amended.
7. Directive 2001/83/EC, as amended.
8. Commission Implementing Regulation (EU) No 520/2012, as amended.

Regulatory Note

EMA's current GVP index identifies Product- or Population-Specific Considerations III as the final guidance on pregnant and breastfeeding women and their children exposed in utero or via breast milk, with reference EMA/653036/2019 Corr. and legal effective date 9 February 2026. The guidance was first published in final form on 6 February 2026. citeturn0search0

This article distinguishes regulatory requirements from scientific interpretation and recommended operational practice. Current legislation, GVP guidance and applicable EMA or national competent-authority procedures should be verified when applying the framework to a specific product or safety issue.

Inspection scenarios are illustrative and are not presented as documented regulatory findings.