Thalidomide made teratogenic-risk evidence and drug approval a pharmaceutical-accountability test

Summary

• The history has to be separated by market. Thalidomide was developed in West Germany and sold there from 1957, then licensed, supplied or distributed through different arrangements in numerous other countries. Dates, brand names, prescription status, indications and withdrawal mechanisms varied. The United States is a decisive counterexample: an application was reviewed, samples reached some people through investigational distribution, but the Food and Drug Administration did not approve thalidomide for the early sedative or pregnancy-sickness use. The FDA's institutional history of reviewer Frances Kelsey records why requests for stronger safety evidence mattered. It is inaccurate to turn international marketing into a claim of formal early US approval.

• Population-level teratogenicity and individual causation are different propositions. The characteristic pattern, timing of exposure, clinical case series, comparative observations and later experimental knowledge support thalidomide's teratogenic risk. An individual determination still asks whether a particular pregnancy involved thalidomide, when exposure occurred, whether the phenotype is compatible, what records remain and what competing explanations exist. A compensation scheme may deliberately use a different evidentiary threshold from a court or clinical diagnosis.

• No exact global total exists. Frequently repeated figures for affected births, infant deaths and living survivors are estimates assembled from different countries, definitions and periods. Some count recognized embryopathy; some count claims or compensated people; some include fetal loss or early mortality; some do not. The proper unit is a dated national register, study cohort or scheme population—not an apparently precise worldwide sum. The Australian National Museum's official historical account is useful for Australian chronology and commonly reported scale, but its international estimates should not be converted into a census.

• Withdrawal did not itself answer responsibility. Clinicians' 1961 warnings and the pattern they identified led to withdrawal in multiple markets, but not on one synchronized date. Withdrawal reduced future exposure; it did not decide what testing had been required, who knew which signal, what warnings were adequate, or what legal standard applied in a specific jurisdiction.

• Criminal process, civil settlement and compensation must remain separate. The West German criminal proceeding ended without a merits judgment after the court terminated the case; that outcome was neither a conviction nor an acquittal resolving every disputed fact. Civil settlements funded remedies without constituting comprehensive adjudications. Public schemes later adopted eligibility and payment rules designed for support. None can be substituted for the others.

• Compensation is jurisdiction-specific. Germany's foundation rests on a statutory structure; the United Kingdom combined a historic settlement with later health-grant support; Australia and Canada created their own support programs. Eligibility, proof, covered costs, periodic payment, capital payment and appeal mechanisms differ. A person recognized in one program does not establish the legal position of every person elsewhere.

• Later medical use does not rehabilitate the original safety system. Thalidomide later demonstrated benefit in restricted conditions, including complications of leprosy and multiple myeloma. Modern authorization concerns a defined indication, dose, evidence package and risk-management system. The FDA approval record for application 020785 begins decades after the early tragedy. It does not retroactively approve the earlier sedative use or erase teratogenicity.

• The durable control is an inspectable evidence chain. A defensible system connects nonclinical reproductive testing, protocol and entity records, exposure timing, adverse-event coding, clinician escalation, cross-border signal exchange, regulator authority, labeling, dispensing restrictions, pregnancy testing, contraception, pharmacy certification and outcome audit. The standard is not whether a rule exists, but whether it prevents fetal exposure, detects exceptions and supports people already harmed.

One medicine produced different national histories

The name “thalidomide disaster” can imply one product moving through one regulatory system. The historical reality was a network. Chemie Grünenthal developed the compound in West Germany. Products containing it were sold under Contergan and other names, while licensees and distributors used additional brands in other countries. Some markets permitted non-prescription sale for periods; others required prescription; some combined thalidomide with other ingredients. Promotional claims, package warnings, professional communications and adverse-event channels differed. So did the legal identity of the seller facing a claimant.

West Germany was the originating and largest documented market. Contergan entered sale in 1957. Early safety concern included peripheral neuropathy in adults; congenital-malformation recognition emerged later because the relevant harm appeared in children after exposure in pregnancy, not necessarily in the person taking the medicine. That latency divided information between neurologists, obstetricians, pediatricians, general practitioners, pharmacies, manufacturers and health authorities. A reporting system focused on a patient's immediate reaction was poorly configured to see an outcome recorded months later in another medical file.

The United Kingdom had a separate route through Distillers and products sold under names including Distaval. British regulatory arrangements before the modern licensing system relied more heavily on manufacturer responsibility and limited administrative review than the post-1968 framework. That context is described in the House of Commons Library briefing on thalidomide, which also traces settlement and state-support questions. It would be wrong, however, to import every finding about Grünenthal's German conduct into the legal position of a British licensee, or to treat a UK settlement as a judgment about every German decision.

Australia likewise involved local distribution, medical use and a national record of affected families, followed many decades later by a dedicated parliamentary examination. The Australian chronology should not be compressed into the German sales date, and recognition of an Australian survivor under a current support program is not proof of precisely which corporate document existed in Aachen in 1960. Japan, Canada, Ireland, New Zealand, countries in continental Europe and other markets each require the same discipline, even when a short comparative article cannot narrate every one.

The United States differs in kind, not just degree. Richardson-Merrell submitted an application for thalidomide, and Frances Kelsey, joined by FDA colleagues, repeatedly sought adequate evidence, including on neuropathy and safe use. Approval did not issue. Investigational samples had nonetheless been distributed to physicians, so “not approved” does not mean “no US exposure at all.” Conversely, exposure under the investigational practices of the period cannot be relabeled as marketing approval. Those two propositions can coexist and are essential to an accurate account.

Canada had authorized distribution and later built a federal support program for survivors. Australia's official record likewise recognizes exposure and a national survivor population. Those histories demonstrate why a global map needs at least four fields: authorization status, distribution mechanism, exposure evidence and withdrawal date. A colored map labeled merely “approved” can hide investigational supply, named-patient access, import, combination products or weak premarket review.

This distinction also protects causal analysis. If a person was born in a country where a licensed product was sold, that fact establishes possible availability, not maternal ingestion. If a product was never formally approved, that fact reduces one route of exposure but does not exclude samples or cross-border acquisition. The correct question is not whether thalidomide was “in the country” in an abstract sense. It is what product was available, under what legal route, to whom, during the biologically relevant pregnancy window.

Teratogenic evidence had to cross clinical silos

The strongest historical signal was not a laboratory dashboard. It was an unusual clinical pattern. In 1961, clinicians including Widukind Lenz in Germany and William McBride in Australia independently connected a marked increase in severe limb-reduction and other congenital anomalies with maternal thalidomide use. Their observations mattered because the phenotype was uncommon, the timing clustered, and exposure histories created a coherent hypothesis. The signal strengthened through additional cases and comparison rather than appearing as one perfect experiment.

“Thalidomide causes birth defects” is scientifically sound at the population and mechanism level, but accountability requires more resolution. Teratogenic effect depends strongly on developmental timing. Exposure during particular days after conception is associated with different patterns involving limbs, ears, eyes, face, heart and internal organs. Dose, duration and individual susceptibility may matter. Exposure outside a susceptible window is not analytically identical to exposure inside it. A record saying only “during pregnancy” can therefore be important yet limited public evidence for phenotype-specific reconstruction.

The premarket evidence problem was partly conceptual. Conventional toxicity work could focus on adult animals and visible maternal illness. An embryo can be harmed at a dose that produces little obvious toxicity in the pregnant adult. Species also differ: a negative result in one animal model does not establish reproductive safety in humans. A competent system needs study design suited to organ formation, appropriate species, exposure windows, fetal examination and preserved raw data.

It must also treat use by people who can become pregnant as a foreseeable exposure scenario, especially when a medicine is promoted for nausea, insomnia or anxiety associated with pregnancy.

Neuropathy was a separate early safety signal. It did not prove teratogenicity, but it challenged broad claims of harmlessness and should have raised the evidentiary burden. A medicine can have more than one adverse-effect pathway. The discovery of adult nerve injury did not require a regulator to predict the precise embryopathy; it required the manufacturer and authorities to reexamine dose, duration, contraindications, reporting and promotion rather than continuing to rely on generalized safety language.

Clinical observation also required a route to authority. A pediatrician seeing several infants could identify a pattern but not know the mother's product history. An obstetrician could know a prescription but not see the child's later specialist assessment. A distributor might receive returns or complaints without standardized congenital-anomaly coding. A national authority might receive reports divided among brand names. Cross-border licensing multiplied the fragmentation. The institutional task was to join mother, product, timing and outcome while protecting records and moving quickly enough to prevent new exposure.

Modern pharmacovigilance terminology can make this sound easy. It was not. Reporting systems were immature, and background rates of congenital anomaly were not captured uniformly. Yet uncertainty cuts both ways. A rare, severe and patterned outcome following a widely used medicine justifies urgent investigation and temporary risk reduction before a final incidence estimate exists. Waiting for a precise denominator means continuing an exposure whose consequences cannot be reversed.

The evidence standard for withdrawal is therefore not the same as the standard for damages in an individual lawsuit. A regulator can act on a credible population signal because the purpose is prevention. A civil court may ask whether, on its governing standard, this claimant's condition was caused by this product and whether a defendant breached a duty. A compensation program may accept characteristic features and a plausible exposure history because the state has chosen a remedial rather than adversarial design.

Treating those standards as interchangeable either delays safety action or imposes a false legal certainty on public-health data.

Individual causation requires more than resemblance

Thalidomide embryopathy is associated with recognizable patterns, but no responsible assessment begins and ends with visual resemblance. Limb differences can arise from genetic conditions, vascular disruption, other exposures or causes that remain unknown. Some thalidomide-associated effects involve internal organs or cranial nerves and may not fit a popular image centered on phocomelia. An overly narrow stereotype can exclude people with compatible but less publicly familiar effects; an overly broad one can attribute every congenital condition to the medicine.

An individual evidence file ideally contains maternal prescription or dispensing information, the brand and formulation, dates and dose, pregnancy dating, contemporaneous obstetric records, birth and pediatric records, imaging, family history and specialist phenotype assessment. In practice, many records were never created, were held by different institutions or were destroyed under ordinary retention rules before compensation existed. The absence may reflect the passage of time rather than deception. It still places an evidentiary cost on survivors whose exposure occurred decades earlier.

Temporal compatibility is central. Development is sequential, and the organ systems affected can help test whether a reported exposure date is plausible. But pregnancy dates themselves may be estimates, recollection can shift, and tablets may have been shared or provided as samples without a conventional prescription. A scheme that demands an intact pharmacy record from 1960 may exclude genuine cases by design. A scheme that ignores chronology and alternative explanations may make inconsistent awards.

Fairness requires published criteria, multidisciplinary review, a route to submit contextual evidence and an appeal or reconsideration mechanism.

The name of the program matters less than the decision rule. A clinical diagnosis may guide care. An administrative recognition may unlock support. A tort judgment allocates legal responsibility between parties under a defined law. A criminal proceeding tests charged offenses with still higher protections and burden. One person's inclusion in a compensation scheme does not prove a criminal offense; one criminal case ending without judgment does not disprove the person's medical condition.

This separation also applies to records of maternal ingestion. A doctor's note may establish that thalidomide was recommended but not prove it was taken. A recollection may be credible even where no label survives. A country-level sales figure can show availability but not the tablet used in one pregnancy. Conversely, the loss of a sales ledger does not undo a well-documented individual prescription. Evidence needs to be evaluated at the level of the proposition it actually supports.

Modern genetic testing can sometimes identify an alternative diagnosis, but a negative genetic test does not automatically establish thalidomide exposure. Nor should new diagnostic technology be used to reset a settled support relationship without a clear lawful basis. The responsible approach separates clinical inquiry, eligibility review and adversarial liability, tells the person what question is being asked, and explains how uncertainty was treated.

The scale cannot be reduced to a global body count

Public accounts often cite about 10,000 affected children worldwide and substantial death before or shortly after birth. Those figures communicate magnitude, but they are estimates, not the output of a complete international registry. Product names differed, distribution extended across borders, fetal loss was incompletely recorded, some congenital anomalies were never linked to exposure, and recognition rules varied. Early deaths also mean that survivor organizations and compensation registers cannot serve as birth cohorts.

A national figure can have several denominators. One source may count children believed born with thalidomide embryopathy. Another may count living people known to an association. A foundation may count approved beneficiaries. A health grant may count current recipients. A lawsuit may count filed claims. These are not redundant measures. A deceased person may belong to the birth estimate but not the current grant population; a living person may have compatible impairment but no qualifying evidence under a particular scheme.

Germany's foundation population is an administrative cohort created under German law. The Contergan Foundation Act defines the statutory architecture, while the official Contergan information portal describes benefits and administration. Neither database is a global epidemiological register. It reflects applications, decisions, survival, legal definitions and the scheme's territorial and product connections.

United Kingdom statistics similarly require labels. Numbers attached to the historic settlement, the Thalidomide Trust and later state health grants may describe overlapping but non-identical points in time. They should not be added. Australia’s current support-program entities are survivors recognized under Australian criteria, not every Australian pregnancy exposed or every child born with a congenital difference during the period. Canada's program population is likewise a federal eligibility cohort.

Even the phrase “thalidomide survivor” can be used differently. It may mean a living person with formally recognized embryopathy, a person who self-identifies based on history and phenotype, or any person affected by thalidomide-related harm, including family. Administrative writing must state which meaning applies. The survivor population also changes with deaths, new recognitions and appeals. A number accurate on one date can be misleading when detached from that date.

Mortality is the least suitable place for false precision. Miscarriages and stillbirths may not have been investigated; severely affected infants may have died without exposure being recognized; cause-of-death certification may reflect immediate organ failure rather than embryopathy. A percentage repeated across decades can acquire undeserved authority. The honest conclusion is that mortality was grave and incompletely measured. National records can illuminate portions of it, but their totals cannot be combined into one exact worldwide number.

Counting is still necessary. Governments need denominators for lifetime care, accessibility, equipment, housing support and fiscal planning. Researchers need cohorts to study late effects. Survivors need visibility. The safeguard is a metadata discipline: every number should state geography, inclusion rule, observation date, whether it is observed or estimated, whether deceased people are included and whether overlap exists with another count. That discipline is part of remedy, not a statistical footnote.

Withdrawal was a distributed control action

Once the congenital-malformation association was communicated in 1961, withdrawal began in major markets. But “withdrawn in 1961” compresses different dates, actors and mechanisms. A manufacturer could stop supply; a distributor could communicate with doctors and pharmacies; a health authority could suspend authorization; hospitals could remove stock; clinicians could stop prescribing; patients needed direct warning. Each step had its own lag.

The last mile mattered because tablets already dispensed remained in homes and clinics. A national announcement did not guarantee that every prescriber received it, that every pharmacy quarantined stock or that a person who had been told the medicine was safe understood the urgency. Combination products and multiple brand names made communication harder. Withdrawal completeness therefore required reconciliation of inventory, documented notices, returned stock and monitoring for continued prescriptions.

The signal also had to cross borders faster than commercial correspondence ordinarily moved. A case cluster in Hamburg or Sydney could be relevant wherever the same active ingredient was used. A robust regulator should know which local brands contain the compound, which companies hold authorizations, which wholesalers supplied them and which indications targeted pregnancy. Without that product dictionary, an alert using one brand name could miss another.

Withdrawal is preventive and prospective. It cannot restore a pregnancy already exposed. That asymmetry should favor rapid precaution when the suspected outcome is severe. Temporary suspension while evidence is examined is not a final accusation. It is a way to prevent irreversible harm during uncertainty. The accountability question is whether decision-makers understood that asymmetry and had legal and operational authority to act.

Post-withdrawal work was equally important. Authorities needed to preserve product files and adverse-event reports, identify affected families, support diagnosis, study outcomes and review approval law. Manufacturers and distributors needed to retain batch, sales and complaint records. Courts needed evidence. Compensation bodies created later depended on documents that could easily have been discarded under schedules designed for routine commercial records.

Regulation changed, but reform claims need jurisdictional care

Thalidomide became a catalyst for stronger drug law, but it was not the only cause of every reform attributed to it. In the United States, the 1962 Kefauver-Harris amendments strengthened requirements concerning effectiveness, safety, informed consent in investigations and regulatory oversight. The enacted Drug Amendments of 1962 are primary legal evidence. They do not establish that thalidomide had been approved; rather, the non-approval episode helped demonstrate the value of demanding adequate evidence before marketing.

In the United Kingdom, the post-disaster path included voluntary safety arrangements and ultimately the Medicines Act 1968, which created a comprehensive statutory framework for medicinal products. It would be anachronistic to apply every later licensing duty word-for-word to a 1950s decision. The appropriate historical test asks what law, professional standard and feasible scientific practice existed at the relevant time, while the reform test asks whether later law closed the exposed gaps.

European integration later created a different authorization environment. The current European Medicines Agency record for Thalidomide BMS concerns a centrally assessed product for a defined later use with explicit risk restrictions. It is not a continuation of Contergan's 1957 marketing permission. The active ingredient is the same; the evidence package, indication, labeling, distribution controls and legal authority are not.

Regulatory learning has at least six layers. First, preclinical studies must examine reproductive and developmental toxicity using methods capable of detecting embryo-fetal harm. Second, clinical development must control pregnancy exposure and follow outcomes. Third, an application must contain adequate manufacturing, safety and efficacy evidence. Fourth, authorities need power to request data, inspect, restrict, suspend and withdraw. Fifth, postmarketing systems must connect adverse outcomes across specialties and borders. Sixth, warnings and restrictions must be verified at prescribing and dispensing points.

Rules alone do not prove performance. An agency can possess suspension power but receive a weak signal late. A label can contain a boxed warning but fail to reach a shared household tablet. A registry can enroll patients but miss pregnancies after treatment stops. A risk-management plan can count signed forms rather than prevented exposures. Durable reform requires outcome measures: detected pregnancies, timing of tests, dispensing exceptions, prescriber and pharmacy compliance, fetal outcomes, audit findings and corrective action.

Data sovereignty complicates cross-border pharmacovigilance. Patient records are protected for good reason, yet privacy cannot become an excuse for failing to share a de-identified severe signal. A national regulator needs local case detail and international pattern recognition. The system should define the minimum data elements, lawful transfer route, common product identifiers, response deadlines and authority responsible for escalation. Safety intelligence must travel without turning sensitive pregnancy records into uncontrolled commercial data.

Legal outcomes did not create one comprehensive verdict

In West Germany, prosecutors brought a criminal case involving Grünenthal personnel. The lengthy proceeding ended in 1970 when the court terminated it, including conclusions about the prospects and public interest in continuing under the applicable procedure. The case did not end in a conviction after a complete merits judgment. It also did not end in an acquittal establishing that all development, warning and distribution conduct was proper. Describing the termination as either comprehensive guilt or comprehensive exoneration exceeds the procedural outcome.

Civil resolution followed a different logic. Settlement can provide money sooner, reduce litigation uncertainty and create a vehicle for long-term payments. It can also leave contested evidence untested and avoid a reasoned judgment on duty, breach and causation. The resulting German foundation therefore should be evaluated as a remedy institution with statutory rules, not treated as a criminal verdict by another name.

In the United Kingdom, negotiations and litigation over compensation unfolded alongside intense public reporting. Contempt restrictions imposed on The Sunday Times became the subject of the European Court of Human Rights judgment in Sunday Times v United Kingdom. That judgment addressed freedom of expression and the legal restriction on publication; it was not a tort judgment deciding every claimant's thalidomide causation or every corporate allegation. It is evidence about the accountability environment, not a substitute for the underlying product case.

Settlements also need exact language. A company payment can reflect compromise, litigation risk, moral response or a negotiated allocation without admitting each allegation. Government contributions may recognize social responsibility without accepting legal liability. A trust's eligibility decision may use agreed criteria. Reporting should state what the instrument says rather than converting payment into a universal admission or converting a denial clause into proof that no wrong occurred.

Limitation periods, corporate structure, evidentiary loss and jurisdiction complicated civil claims. A parent company, licensee, distributor, prescribing clinician and state authority occupied different roles. The fact that the active ingredient was common does not make every defendant interchangeable. Actor-specific analysis needs the contract, product authorization, promotional control, knowledge, warning duty, sale, exposure, limitation law and causal standard relevant to that forum.

The same care applies to later public apologies. A statement of regret or apology may be ethically important and may acknowledge a company's part in a tragedy. Its legal effect depends on text and law. It cannot supply absent details about one prescription, and it should not be minimized merely because it is not a judgment. Accountability reporting can respect its meaning while maintaining the evidentiary boundary.

Compensation systems answer different questions

Germany's response developed into a statutory foundation providing pensions and other benefits to recognized people. Its legitimacy depends on more than the existence of a fund. Eligibility must reflect the medical evidence without demanding impossible records; impairment assessment must capture internal and late-emerging effects; payments must respond to lifetime cost; decisions must be explainable; and review must be accessible. Because survivors age, a schedule designed around childhood impairment can become inadequate for pain, joint degeneration, reduced mobility and loss of informal care.

The United Kingdom's arrangement combined negotiated compensation and trust administration with later government health-grant support. The distinction matters. A settlement fund represents a civil compromise involving corporate parties. A government grant is a public policy instrument. Neither should be described as if it were the other, and current recipients should not be turned into an estimate of all historical UK births.

Australia created a dedicated federal support architecture after the Senate Community Affairs References Committee investigated survivor needs. The committee's official report documented evidence and made recommendations; it was a parliamentary inquiry, not a civil trial. The Australian Government response records which recommendations the executive accepted and how it proposed to act. Acceptance is policy evidence, not proof that every need was met.

The resulting Australian Thalidomide Survivors Support Program provides an official source for current program design, including payments and extraordinary-healthcare assistance. Program enrollment is governed by Australian criteria. An approved claim establishes entitlement under that scheme; it does not adjudicate a German criminal charge or set another country's eligibility rule.

Canada's Thalidomide Survivors Contribution Program is another distinct federal remedy. Its continuing payments and exceptional-medical-assistance features should be assessed against its own eligibility, evidence and administration. A Canadian recognition count is an administrative snapshot. It cannot be added to German, UK and Australian beneficiary totals and labeled the number of living survivors worldwide.

Good compensation design separates at least five questions. Who qualifies? What evidence is accepted? What heads of need are covered? How are payments adjusted over time? What review exists? A sixth question concerns people who died before a scheme or whose families supplied decades of unpaid care. Different jurisdictions answer these differently. Comparative evaluation should expose the choices, not erase them beneath the word “compensation.”

The evidentiary burden is a core accountability issue. When institutions failed to preserve prescribing or sales records, a strict documentary-only rule makes the injured person pay twice: first through harm, then through exclusion caused by missing system evidence. Alternative evidence can include consistent maternal history, characteristic phenotype, location and timing, clinician records, product availability and expert assessment. That flexibility still needs safeguards against arbitrary inconsistency.

Payment adequacy also changes with age. Many survivors adapted through extraordinary physical techniques that placed stress on joints and spine. Parents and partners supplied care that is no longer available. Housing, vehicles, personal assistance, dental work, hearing support and bespoke equipment can cost more as function changes. A one-time settlement measured decades earlier cannot automatically demonstrate present sufficiency. Programs need periodic, participatory review based on actual cost and health evidence.

Survivorship is a continuing health-system obligation

The public image of thalidomide frequently freezes survivors as infants in black-and-white photographs. That framing is ethically and clinically inadequate. Survivors are adults with careers, relationships, expertise and agency. Many also experience secondary musculoskeletal problems, chronic pain, nerve symptoms, hearing or vision impairment, internal-organ consequences, dental needs, fatigue and barriers created by inaccessible services. Aging can reduce compensatory strength and expose needs that were not apparent in childhood.

Routine health care can itself be difficult. Blood-pressure cuffs, imaging equipment, dental chairs, examination tables and standard positioning may not accommodate a person's anatomy. Clinicians unfamiliar with thalidomide embryopathy may attribute every new condition to it or, at the other extreme, fail to understand a relevant anatomical difference. A safe pathway needs accessible facilities, sufficient appointment time, continuity, specialist advice and a record controlled with the survivor rather than assumptions based on appearance.

Mental health and social effects should not be treated as secondary decoration. Public staring, school exclusion, employment discrimination, dependence created by inaccessible environments, bereavement and repeated eligibility examinations can cause lasting distress. Parents may have lived with blame for taking a medicine presented as safe. Remedy should avoid reproducing that blame or requiring humiliating demonstrations of impairment.

Survivor knowledge is evidence. People know which movements create injury, which equipment works and where medical systems fail. Involving representative organizations in program design is not a courtesy; it is a quality control. Consultation still needs transparency about who participated, what changed and how dissent was handled. One organization cannot be presumed to speak for every country or every phenotype.

Long-term research needs consent and governance. Registries can clarify late effects and improve services, but enrollment in compensation should not silently compel research participation. Data reuse, international linkage and withdrawal rights must be explained. Results should return to entities in accessible form. The historical failure involved information moving around people without adequate protection or disclosure; a modern survivor registry should not repeat that structure in digital form.

Later restricted use is a separate benefit-risk decision

Thalidomide did not disappear from medicine. It was later used for erythema nodosum leprosum and found to have anti-inflammatory, immunomodulatory and anti-angiogenic effects. It also became a treatment component in multiple myeloma. These benefits are real. They do not conflict with the conclusion that embryo-fetal exposure can cause severe harm. Benefit and hazard can coexist, which is why indication and controls matter.

The US approval chronology makes the boundary concrete. FDA approval in 1998 concerned erythema nodosum leprosum under exceptionally restrictive distribution conditions; a later approval included newly diagnosed multiple myeloma in combination therapy. The current FDA prescribing information carries an embryo-fetal-toxicity warning and detailed requirements. The National Cancer Institute drug page summarizes the oncology indications. Neither source supports use for ordinary insomnia, anxiety or nausea in pregnancy.

A pregnancy-prevention program must control more than the person prescribed the drug. It needs prescriber certification, patient counseling, pregnancy testing at specified intervals, contraception or documented abstinence under defined rules, pharmacy certification, limited dispensing windows, no sharing, secure storage, and instructions for missed periods or suspected pregnancy. Because thalidomide can be present in semen, risk communication also addresses patients who can impregnate a partner. Blood and semen donation restrictions protect additional pathways.

The controls must be usable. A checklist written in inaccessible language can fail. Testing that is geographically unavailable can delay cancer treatment. A portal outage can encourage workarounds. Overly burdensome administration can move use outside monitored channels, while weak controls permit preventable exposure. The governance task is to preserve the hard safety boundary while providing equitable, timely access for patients who may benefit.

Outcome audit is essential. Counting completed forms does not show that pregnancies were prevented. Programs should examine dispensing outside authorization windows, missing tests, pregnancies during or after treatment, pharmacy overrides, prescriber noncompliance and product sharing. Each exception needs investigation and corrective action. Aggregate results should be public without identifying patients.

Later authorization also requires ordinary pharmacovigilance beyond teratogenicity. Peripheral neuropathy, thrombosis, sedation and other known risks can affect adult patients. Focusing only on pregnancy can obscure them. Benefit-risk review should use indication-specific alternatives, dose, duration, comorbidity and monitoring. A drug with catastrophic reproductive risk is not necessarily without therapeutic value; it is a medicine whose controls must match the consequence of failure.

Control rights show where accountability belongs

The manufacturer controlled compound development, study design, raw-data preservation, safety interpretation, labeling proposals, promotional claims and communications to licensees. Those rights carried duties to test foreseeable use, investigate neuropathy and pregnancy signals, avoid unsupported safety claims, notify authorities and commercial partners, and act rapidly when a severe pattern emerged. A licensee or distributor with local promotion and complaint channels held its own control rights rather than serving as a passive mailbox.

National authorities controlled authorization within their legal systems, requests for further evidence, restrictions, suspension and market communication. Their powers differed in the 1950s and 1960s. Accountability should not invent authority an agency did not legally possess, but weak law is itself a governance finding for legislators and governments. Where an authority could request data or warn clinicians, its actual use of that power remains examinable.

Clinicians controlled prescribing, indication, dose and communication with patients. Their knowledge depended on available evidence, company material, professional publications and warnings. No fair account assumes every doctor knew the association before it was reported. Once a credible warning circulated, however, stopping use and contacting exposed patients became part of clinical safety. Samples and informal supply required the same care as a conventional prescription.

Pharmacists and wholesalers controlled stock. Their records could trace distribution and their quarantine action could make withdrawal real. Hospitals controlled formulary, maternity-service communications and retention of maternal and pediatric records. Researchers and journals controlled rapid but responsible communication of a signal. Governments controlled whether affected people later faced a hostile litigation-only route or an accessible remedy.

Patients did not control the hidden evidence package. A pregnant person could make a choice only if told that pregnancy exposure was relevant, what uncertainty existed and what alternatives were available. Consent cannot transfer an undisclosed development risk from institution to patient. Nor can a patient's use of a medicine for a promoted purpose be reframed as personal fault when the material hazard was not known to them.

These control rights form a chain. Development evidence must be intelligible to the regulator; authorization conditions must reach prescribers; prescriptions must reach the dispensing register; adverse pregnancy outcomes must connect back to exposure; cross-border signals must reach each market; withdrawal must reach every stock location; preserved records must support diagnosis and remedy. A break anywhere can defeat an otherwise strong control.

What remains uncertain

The exact global number of affected pregnancies, births, early deaths and living survivors remains unknown. International estimates use incomplete case recognition and inconsistent definitions. National scheme numbers are more concrete but describe eligibility and administration, not the entire exposed population. This article therefore does not add them or present a single current survivor total.

The exposure history of every country is not resolved in one common dataset. Formal authorization, licensed distribution, investigational supply, import and tablet sharing are distinct. Withdrawal dates also differed. The article's comparative examples do not imply that an unmentioned market followed the German, British, Australian, Canadian or US path.

Individual causation may remain uncertain where maternal records, product identity or timing are missing. A characteristic phenotype can strongly support recognition, while alternative diagnoses may require assessment. A compensation decision, clinical diagnosis and court judgment can reasonably reach different results because they answer different questions under different standards.

Actor-specific knowledge is date-sensitive. Evidence available to a clinician who reported cases was not necessarily available to every local prescriber. A manufacturer's central files were not identical to a distributor's complaint record. A regulator's legal power differed by jurisdiction. Broad institutional responsibility should not be turned into an allegation about a named person without that person's record and the competent forum.

The German criminal proceeding's termination did not produce a complete merits adjudication. Civil settlements did not necessarily admit every allegation or decide each claimant's causation. The European human-rights judgment concerned publication restrictions. Compensation statutes and programs create support rights. None is a universal factual verdict.

The long-term adequacy of survivor programs also remains open. A program can exist while payment levels, access, review time or healthcare coverage remain contested. Later risk-management systems can be stringent while still experiencing noncompliance or pregnancy exposure. Institutional learning must be measured in outcomes, not inferred from the existence of law, a fund or a warning label.

The accountability test

Thalidomide's history is not a simple story in which one late discovery produced one worldwide ban and one settlement. It is a distributed pharmaceutical failure: premarket methods did not protect embryonic development; adult neuropathy weakened assurances of safety; congenital-anomaly observations were divided across specialties; products moved through different commercial and legal routes; withdrawal traveled at different speeds; and evidence needed for later recognition was often incomplete.

The first discipline is jurisdiction. Germany was the development and principal early market; Britain and Australia used local licensees and systems; Canada had its own approval and remedy history; the United States withheld marketing approval while still experiencing limited investigational exposure. Those facts cannot be merged into “approved worldwide,” and US non-approval cannot be inflated into zero exposure.

The second discipline is evidentiary. Population evidence establishes a powerful teratogenic association and biologically meaningful timing. Individual decisions still require exposure and phenotype assessment. Global scale is estimated, not counted exactly. A national foundation or support-program register is a remedial population, not a worldwide epidemiological denominator.

The third discipline is legal. Withdrawal prevented prospective exposure but did not adjudicate blame. A terminated criminal case was not a conviction or a comprehensive acquittal. A settlement delivered compromise, not a complete findings judgment. A compensation award recognizes eligibility under a scheme, not criminal guilt. Keeping those tracks separate protects both survivors and procedural truth.

The fourth discipline is continuity. Survivors need accessible care and support across a lifetime, while later patients may legitimately receive thalidomide for narrow indications under strict controls. The system must preserve both obligations without treating therapeutic rediscovery as historical absolution or historical catastrophe as a reason to deny an evidence-based benefit.

The future test is inspectable: can a regulator identify every local product containing the same substance; can a severe pregnancy signal cross specialties and borders within days; can an authority suspend exposure before certainty; can records connect product, timing and outcome decades later; can remedy accept fair alternative evidence; can survivors shape aging services; and can restricted-use programs show prevented exposure rather than completed paperwork? Those are the controls by which pharmaceutical accountability should be judged.

Source notes

This article gives controlling weight to official statutes, regulator records, parliamentary and government reports, compensation-program pages and a formal court judgment. Historical institutional pages are used for bounded chronology rather than as complete adjudications. Current labels and authorization records describe later restricted uses only. National administrative cohorts are not combined into a global total, and no settlement, terminated proceeding, scheme decision or apology is treated as a comprehensive finding of civil or criminal liability. The companion source ledger records the intended use and boundary for every URL.