Psychology by BogdanovaThe Encyclopedia of Psychology for the Artificial Era

RESEARCH DOSSIER

Developmental, Reproductive, and Later-Life Psychopharmacology

Abstract. Developmental, reproductive, and later-life psychopharmacology examines how medication efficacy, exposure, adverse effects, and outcomes change across childhood, adolescence, pregnancy, lactation, and older age. This fundamental research monograph synthesizes randomized pediatric and geriatric treatment studies; reproductive pharmacoepidemiology; maternal, neonatal, and offspring outcomes; medication transfer in breastfeeding; clinical pharmacology; professional guidelines; and regulatory labeling. It critically distinguishes causal experimental findings from observational associations, evaluates confounding by indication and generalizability, and integrates the harm of untreated disease into the interpretation of treatment. The synthesis identifies life-stage-sensitive outcome measurement, drug-specific rather than class-wide inference, and longitudinal clinical follow-up as the organizing scientific principles, while specifying unresolved research problems in developmental safety, reproductive causality, frailty, polypharmacy, and artificial medication-information systems.

Keywords: developmental psychopharmacology; pediatric psychopharmacology; reproductive psychiatry; psychotropic medications in pregnancy; prenatal exposure; fetal development; antidepressants; lithium; valproate; neonatal adaptation; lactation pharmacology; breastfeeding; geriatric psychopharmacology; frailty; geriatric depression; polypharmacy; Beers Criteria; causal inference

Scientific Overview and Research Questions

Developmental, reproductive, and later-life psychopharmacology investigates how the effectiveness, safety, exposure, interpretation, and long-term consequences of medications acting on the central nervous system vary across development, reproductive transitions, and aging. The research object is a population-sensitive benefit–harm relationship. It encompasses the patient receiving treatment, the developing fetus or nursing infant when relevant, and the changing physiological and social setting in which pharmacotherapy is delivered. Developmental state influences pharmacokinetic pathways, brain maturation, psychopathology, behavioral reporting, and the time horizon for outcomes. Pregnancy adds a second developing organism and dynamic maternal physiology. Older age introduces heterogeneous organ reserve, frailty, multimorbidity, pharmacodynamic vulnerability, and treatment burden.

Four questions organize the research. Which treatment effects have been established in the population and indication actually under consideration? How does developmental or reproductive physiology change drug exposure and monitoring? What can experiments and observational studies establish about immediate and delayed harms? How should evidence accommodate both disease-related harms and treatment-related harms? These questions cannot be answered by extrapolating a trial conducted in healthy middle-aged adults to all other groups. The methodological starting point is direct measurement in defined populations, explicit comparators, and outcomes that have clinical meaning to participants and families. Population-specific recommendations are evaluated in American College of Obstetricians and Gynecologists (2023), the American Geriatrics Society 2023 Beers Criteria Update Expert Panel (2023), and the foundational pediatric trials examined below.

This is a critical narrative scientific synthesis based on identifiable original trials, population cohorts, professional guidance, prescribing criteria, and regulatory drug-information resources. It is not a prospectively registered systematic review, a new clinical trial, or a pooled meta-analysis. The work evaluates competing explanations, measurement limitations, transportability, and research priorities. All medication examples are population- and indication-specific scientific cases; contemporary treatment decisions require current product labeling, individual clinical evaluation, and relevant professional guidance.

The Historical Formation of Population-Specific Psychopharmacology

The early history of modern psychopharmacology established pharmacological intervention as an experimental approach to psychiatric symptoms, while the expansion of treatment into pediatric, reproductive, and geriatric populations exposed gaps in the evidence generated by general adult studies. The resulting research programs combined developmental psychiatry, reproductive epidemiology, geriatrics, clinical pharmacology, and regulatory science. They gradually replaced broad assumptions about medication classes with questions about age-specific indications, susceptible developmental windows, organ-function change, and longitudinal outcomes.

Pediatric drug evidence became increasingly dependent on diagnosis-specific randomized comparisons, structured adverse-event monitoring, and study of combined psychological and pharmacological treatment. The MTA Cooperative Group (1999), the Treatment for Adolescents With Depression Study (TADS) Team (2004), and the Walkup (2008) are historical milestones because they examined concrete pediatric treatment strategies rather than treating child dosage as a mere modification of adult dosage. In reproductive research, administrative cohorts, pregnancy registries, and within-family comparisons provided alternatives to randomized fetal exposure. For later life, pragmatic interventions and geriatric medication review extended outcomes beyond disorder scales to falls, cognition, daily function, treatment persistence, and quality of life.

The regulatory history also matters. The Food and Drug Administration (2014), effective in 2015, replaced simplified pregnancy letter categories for covered products with structured narrative sections intended to support benefit–risk interpretation. Pediatric antidepressant suicidality warnings made monitoring a central issue. Geriatric prescribing criteria consolidated evidence about context-sensitive adverse effects without abolishing individualized clinical judgment. These are distinct historical developments, linked by the demand for population-valid evidence.

A Life-Course Framework for Medication Effects

A life-course approach treats biological age, chronological age, developmental stage, reproductive status, and functional reserve as separate explanatory dimensions. Two adolescents of the same age can differ in pubertal development, metabolic maturation, growth trajectory, and capacity to describe internal symptoms. Two adults of the same advanced age can differ dramatically in renal function, mobility, cognition, nutritional status, and total medication burden. Pregnancy changes substantially across trimesters and around delivery; lactation adds infant age, gestational maturity at birth, and milk-transfer characteristics. A study limited to a single age threshold or an undifferentiated pregnancy exposure category may therefore aggregate clinically different states.

The basic causal sequence is treatment indication and severity; medication decision and dose; absorption, distribution, metabolism, and elimination; receptor-level and network-level effects; therapeutic and adverse outcomes; and subsequent treatment changes. Every stage is modified by social context, access to care, medication adherence, co-prescriptions, developmental physiology, and time. A meaningful research design specifies the time origin, compares clinically plausible treatment strategies, and follows outcomes long enough to detect delayed effects. This logic also explains why pregnancy safety, pediatric long-term development, and late-life fall prevention belong to a shared dossier while retaining their separate empirical literatures.

Biological and psychological outcomes must be measured together. Symptom response may coexist with weight change, sedation, sleep disruption, or altered family function. Medication avoidance may spare some adverse reactions yet expose the patient to persistent severe illness. The correct research question is the expected outcome under each feasible care strategy for a specified population, rather than the abstract danger of drug exposure alone. The framework is consistent with the comparator-sensitive analyses of Park (2018) and Sujan (2017).

Pharmacokinetics, Pharmacodynamics, and Developmental State

Pharmacokinetics describes the time course of drug exposure through absorption, distribution, metabolism, and excretion; pharmacodynamics concerns the relationship between exposure, target engagement, physiological response, and adverse effects. Both vary by physiological state. Body composition and protein binding change distribution. Hepatic enzymes and transporters vary with maturation, pregnancy, genetics, interacting drugs, inflammation, and organ disease. Renal elimination can be unusually vulnerable during aging or fluid imbalance. Identical prescribed doses therefore need not yield identical exposure, and identical measured concentrations need not produce identical clinical responses.

Developmental pharmacology must distinguish neonates, young children, adolescents, and adults. A neonatal pharmacokinetic observation cannot be generalized to a teenager, and a pediatric efficacy trial cannot by itself demonstrate absence of effects on growth, maturation, or outcomes appearing decades later. Pregnancy changes plasma volume, renal filtration, and the activity of multiple metabolic processes, sometimes altering drug concentration over time even when a maintenance dose is unchanged. In later life, reduced physiological reserve and increased sensitivity to central nervous system effects can produce sedation, impaired balance, orthostasis, or confusion despite exposure within a general adult reference interval. National Library of Medicine (2026) illustrates how maternal concentrations, milk concentrations, and infant metabolic maturity are different measures.

The appropriate methodological response is to combine population-specific clinical outcomes with exposure data when a validated concentration–response relationship exists. Therapeutic drug monitoring has particular importance for medicines with narrow therapeutic margins, but it is not a substitute for symptom, renal, thyroid, metabolic, or adverse-event monitoring as indicated by the medicine. A measured concentration answers an exposure question. It does not establish treatment success, neurodevelopmental safety, or acceptable risk by itself.

Research Designs and the Hierarchy of Inference

Randomized controlled trials can estimate causal effects of treatment assignment within their enrollment criteria. Their internal validity depends on allocation, masking where possible, follow-up, outcome definition, and handling of missing data; their generalizability depends on who was enrolled. Trials in pregnancy are restricted by practical and ethical considerations, so medication safety often relies on observational cohorts, registries, pharmacovigilance, and mechanistic information. These designs can identify signals, estimate associations, and sometimes support causal inference when assumptions are examined transparently. Clinical guidelines integrate these heterogeneous forms of evidence while adding normative judgments about acceptable outcomes.

A registry comparison of exposed and unexposed pregnancies is especially vulnerable to confounding by indication. The illness prompting treatment, illness severity, other substances, nutrition, health-care contact, genetic liability, and family environment can influence outcomes independently of medication. Active comparators, propensity-score methods, sibling comparisons, negative controls, and pre-pregnancy exposure comparisons address different biases; none automatically removes all residual confounding. Huybrechts (2014), Sujan (2017), and a large Suarez (2022) demonstrate why conclusions change after adjustment.

Time-varying treatment complicates inference. Initiation, discontinuation, dose changes, adherence, hospitalization, and relapse may affect future exposure and future outcomes. Live-birth cohorts omit some pregnancy losses by design and can produce selection effects. Observational prescription filling is an exposure proxy, not proof that tablets were consumed. A sound synthesis distinguishes randomized efficacy, real-world effectiveness, association, risk prediction, and causal attribution, and presents confidence intervals and absolute risks alongside relative measures whenever the original record permits.

Psychopharmacology of Childhood and Adolescence

Childhood and adolescence are periods of rapid changes in cognition, affect regulation, social demands, sleep, growth, and bodily development. Psychiatric presentations are shaped by developmental context: restlessness may have different meanings across ages; depressed mood may coexist with irritability and disrupted school participation; anxiety may appear through avoidance, physical complaints, and family accommodation. Pharmacological research must therefore use developmentally appropriate diagnoses, symptom measures, informants, and functionally relevant outcomes. Parents, teachers, clinicians, and young people may supply complementary and discrepant observations.

A pediatric medication trial establishes evidence for a particular disorder, population, regimen, and follow-up interval. It does not validate every use of the same molecule in children. Regulatory indications differ by product and jurisdiction, and an age-based approval for one condition does not generalize to another. Clinical research must also examine the developmental opportunity costs of untreated illness: missed education, persistent distress, family disruption, accidents, and the long-term accumulation of disability. These outcomes are difficult to estimate over the short duration of many randomized trials.

The research program integrates efficacy, adverse events, family and school functioning, longitudinal monitoring, and psychotherapy. It must examine exposure during growth, changing adherence as autonomy develops, and transition between pediatric and adult care. A guiding comparison is not simply medication against no medication, but medication management, evidence-based psychological interventions, combined care, and usual care under realistic service conditions. The MTA Cooperative Group (1999), Treatment for Adolescents With Depression Study (TADS) Team (2004), and Walkup (2008) studies supply different designs for this comparison.

Attention-Deficit/Hyperactivity Disorder: The MTA Research Program

The Multimodal Treatment Study of Children with ADHD enrolled 579 children aged seven to just under ten years in a fourteen-month randomized comparison of carefully managed medication, intensive behavioral treatment, their combination, and community care. Medication management and combined treatment generally improved core ADHD symptoms more than behavioral treatment alone or community care. Combined care did not significantly exceed medication management on the direct core-symptom comparison, although some noncore and positive-functioning outcomes favored the combined approach. The trial thus separates a specific symptom-domain result from the broader question of total functional benefit. MTA Cooperative Group (1999).

The distinction between carefully managed medication and routine community prescribing is methodologically important. Scheduled assessment, titration, monitoring, and response-based decisions are part of an intervention strategy; comparing medication labels without accounting for care organization can miss an essential causal component. Medication effectiveness is mediated by delivery quality, and classroom or family outcomes can improve through pathways not captured by a single symptom score. Interpretation also requires attention to treatment availability, family burden, adherence, and the unequal feasibility of intensive behavioral programs.

Long-term ADHD research must investigate growth trajectories, appetite and sleep effects, cardiovascular symptoms, social development, academic achievement, substance-use outcomes, and persistence of benefit across changing developmental tasks. Short-term symptom superiority cannot alone support definitive claims about adult achievement or prevention of later disorders. Pharmacologic and behavioral approaches should be evaluated as potentially complementary strategies with endpoints defined before comparison.

Adolescent Depression, Antidepressants, and Suicidality

The Treatment for Adolescents With Depression Study randomized 439 adolescents aged twelve to seventeen to fluoxetine, cognitive behavioral therapy (CBT), their combination, or placebo during an initial twelve-week trial. In the published first-phase results, response was reported in 71.0% with combined treatment, 60.6% with fluoxetine, 43.2% with CBT, and 34.8% with placebo. These numbers refer to a specific trial definition of response and follow-up period; they should not be interpreted as universal probabilities for all adolescent depression. The combined arm had the most favorable assessed benefit–risk tradeoff in that phase. Treatment for Adolescents With Depression Study (TADS) Team (2004).

The antidepressant suicidality warning is a central evidence-safety case. The FDA pooled pediatric trials reported increased suicidal thinking or behavior during initial treatment compared with placebo, and emphasized close observation for clinical worsening and unusual behavior changes. Historical FDA labeling reported an average rate of 4% among antidepressant-treated pediatric participants versus 2% on placebo in the pooled short-term analyses; there were no suicides in those trials. These figures summarize heterogeneous trial conditions rather than an individualized prediction. Food and Drug Administration (2005). Current product labeling and any applicable later revisions control contemporary regulatory wording.

The research must treat the underlying disorder and treatment-emergent events as simultaneous safety concerns. Depression itself is associated with suicidal ideation and other functional risks; careful monitoring should assess emergent agitation, mood changes, activation, worsening symptoms, and suicidal behavior independently of whether treatment has begun. Trials with relatively short follow-up provide limited information about rare events, developmental trajectories, and long-term comparative effectiveness. Observational monitoring and clinical follow-up add safety information but require their own causal cautions.

Pediatric Anxiety and Combined Treatment: CAMS

The Child–Adolescent Anxiety Multimodal Study randomized 488 young people aged seven to seventeen with separation anxiety disorder, generalized anxiety disorder, or social anxiety to CBT, sertraline, their combination, or placebo. The twelve-week trial demonstrated improvement with both active monotherapies and superior response with combined treatment. Its relevance extends beyond the particular drug. It shows how a design can compare pharmacological change, learned behavioral skills, and their conjunction with a shared assessment structure. Walkup (2008).

Interpretation requires attention to what the treatments actually deliver. CBT involves repeated behavioral practice and exposure-based learning, which may generalize beyond the acute intervention; medication provides another mechanism for symptom reduction during treatment. Combined care also involves greater treatment contact, family coordination, and the possibility that reduced anxiety facilitates behavioral engagement. These mechanisms are plausible interpretations and require designs that isolate mediators before causal certainty can be assigned.

The study cannot automatically be extended to obsessive-compulsive disorder, trauma-related disorders, or every child with anxiety symptoms. Comorbidity, developmental level, therapy availability, family resources, and treatment preference can modify comparative value. Moreover, a response at twelve weeks is distinct from sustained remission, functional participation, relapse after discontinuation, and adverse effects over years. The strongest research agenda combines long-term functional follow-up with careful measurement of both therapy and pharmacological exposure.

Pediatric Bipolar Disorder, Psychosis, and Complex Presentations

The evidence problem becomes more difficult for severe pediatric mood and psychotic presentations. Diagnostic boundaries, developmental symptoms, co-occurring neurodevelopmental conditions, and high clinical acuity complicate trial entry and external validity. Antipsychotic or mood-stabilizing medications can have substantial acute benefits for specific approved indications, while metabolic, endocrine, neurological, and movement-related adverse effects require explicit study. Weight and metabolic monitoring, sedation, extrapyramidal symptoms, and persistent involuntary movements are not peripheral variables in a developing patient.

Comparative medication studies should distinguish short-term crisis stabilization from maintenance efficacy and relapse prevention. A treatment might suppress acute agitation or psychosis yet have a different benefit–harm balance during years of maintenance. Cross-drug comparisons must account for indication, severity, baseline metabolic risk, dosing, switching, and concurrent medication. Evidence from adult schizophrenia or bipolar studies cannot by itself establish pediatric efficacy and safety, and off-label use is an evidence classification rather than a guarantee of harm or benefit.

Long-term developmental follow-up needs measurements of education, cognition, growth, peer relationships, patient-defined functioning, and discontinuation effects. Research should document differences among participants receiving the same diagnosis, including developmental disability and medically complex children. Medication effects are assessed within a multimodal clinical strategy, and regulatory status should be verified product by product using current prescribing information rather than inferred from class membership.

Growth, Sleep, Metabolic Health, and Developmental Outcomes

Pediatric safety research must address outcomes that are not fully represented in an adult adverse-event scale. Appetite suppression, growth velocity, body-mass changes, sleep disruption, social participation, academic engagement, and physical activity can alter a child's developmental environment. For some psychotropics, metabolic change can be clinically significant even when symptom improvement is strong. These consequences affect adherence and family decisions, so they must be considered alongside direct neurological adverse reactions.

Measurement requires age- and sex-relevant growth references, baseline trends, repeated observations, and documentation of puberty and nutrition. An apparent change in a single height or weight observation can arise from measurement variation or baseline trajectory. Valid inference requires longitudinal comparison and attention to disease itself, eating patterns, other medications, and family characteristics. For sleep, studies must separate baseline insomnia, medication-associated activation or sedation, and shifts produced by changes in school schedule or psychiatric symptoms.

The appropriate end point is a multivariate developmental trajectory: symptoms, physiological growth, sleep, social function, and adverse effects jointly observed. Trials powered for symptom improvement may be underpowered for a rare cardiovascular or neurological event. Product-specific postmarketing surveillance and population registries therefore complement controlled efficacy studies, but safety signals require evaluation for denominator, reporting bias, and competing explanations. A clinically meaningful pediatric evidence system integrates short-term trials with long-term, continuously updated safety assessment.

Psychopharmacology in Pregnancy: The Dual-Outcome Problem

Pregnancy changes the unit of analysis. Medication treatment is directed to a pregnant patient's clinical condition while fetal and neonatal development introduce additional outcomes, time windows, and ethical considerations. The presence of two outcome sets does not imply that the optimal strategy can be established by maximizing fetal exposure avoidance. Persistent or recurrent severe maternal psychiatric illness may itself affect health, functioning, prenatal care, safety, and the postpartum period. A defensible research framework compares plausible treatment strategies while describing maternal and child outcomes separately. American College of Obstetricians and Gynecologists (2023).

Exposures must be temporally specified. Periconception exposure, early organogenesis, later gestation, delivery, and breastfeeding are physiologically and clinically distinct. Structural congenital anomalies, preterm delivery, transient neonatal adaptation, obstetric complications, and childhood neurodevelopment are separate outcome domains. A finding for one cannot stand in for another. Studies of medication dispensation should consider actual use, discontinuation, switching, dosage, and concurrent drugs, as well as repeated periods of maternal symptoms.

Regulatory pregnancy labeling gives a structured account of human and animal data, clinical considerations, and data limitations. The FDA's Food and Drug Administration (2014) provides a framework for reading such information, while ACOG's guideline assesses psychiatric conditions and their pharmacotherapy in the perinatal period. These sources require ongoing version checks. The scientifically meaningful objective is a traceable comparison of benefits and harms for each particular medication and clinical context.

Untreated Illness, Continuation, and Relapse Risk

The consequence of stopping an effective psychiatric treatment is an exposure in its own right. In a prospective observational investigation of pregnant women with a history of major depression, Cohen (2006) studied relapse in those maintaining or discontinuing antidepressant treatment. Its clinical importance lies in documenting illness recurrence as a competing outcome in reproductive decisions. Its nonrandomized design also means that continuation and discontinuation groups may differ in illness history, patient choice, severity, and other factors that affect relapse.

The relevant clinical comparison should begin before medication changes occur and document previous episodes, current stability, treatment history, psychotherapy access, and planned follow-up. A cohort of people who discontinue because of milder illness is not automatically comparable to people who continue after recurrent severe episodes. Conversely, extrapolation from a high-risk referral population to every patient with previous depression would overstate risk. Studies should stratify by recurrence history and use measures of sustained remission and functional recovery.

Research on bipolar disorder and prior postpartum psychosis demonstrates especially consequential time windows for relapse prevention. Bergink (2012) observed different courses among women with psychosis restricted to previous postpartum episodes and those with established bipolar disorder. Their findings support diagnostic and history-sensitive investigation; the study is observational and small, so its treatment implications should be integrated with current specialist guidance rather than elevated into a universally applicable regimen.

Antidepressants and Structural Birth Outcomes

The association between antidepressant exposure and congenital anomalies has been tested in large pregnancy cohorts. Huybrechts (2014) examined antidepressant use during pregnancy and cardiac defects in a population-based analysis with progressively adjusted comparisons. The central lesson is methodological: crude or incompletely adjusted comparisons can yield different interpretations from analyses accounting more carefully for psychiatric illness and other covariates. Drug-specific questions require statistical power sufficient for uncommon malformations and rigorous consideration of multiple testing.

Structural outcomes should be separated by organ system and developmental timing. A composite of all congenital anomalies may mask an effect limited to a particular structure, while subdivision into many rare outcomes increases uncertainty and false-positive risk. Medication class is often a crude exposure definition because agents differ in pharmacology, metabolism, and treatment selection. The denominator of pregnancies also matters; studies restricted to live births may incompletely represent fetal losses and terminations, creating selection complications when outcomes and exposure affect inclusion.

Scientific interpretation should provide absolute frequencies and adjusted uncertainty alongside relative measures, distinguish replication from one-time associations, and avoid treating failure to find a statistically significant increase as proof of complete safety. Likewise, a detected association is not automatically a pharmacological causal effect. The appropriate reading of the literature is indication-specific and comparator-sensitive, with current guideline and regulatory information supplying the clinical context rather than replacing primary epidemiological evidence.

Lithium Exposure and Congenital Cardiac Risk

Lithium is a distinctive reproductive psychopharmacology case because its established role in bipolar disorder must be considered alongside narrow therapeutic monitoring and possible developmental risk. In a Medicaid cohort covering 1,325,563 pregnancies, Patorno (2017) identified cardiac malformations among 16 of 663 lithium-exposed infants (2.41%) and 15,251 of 1,322,955 unexposed infants (1.15%). Their adjusted risk ratio was 1.65, with a 95% confidence interval of 1.02–2.68. The analysis also considered lamotrigine exposure and dose categories. These results suggest a smaller relative risk than some historical fears while leaving clinically important uncertainty.

The cohort's strengths include size, detailed treatment records, and measured adjustment; its limits include residual confounding, relatively few exposed malformation events, and the use of pregnancies ending in live birth. The absolute-risk expression is essential. A statistically elevated adjusted ratio does not tell a family how probable a particular outcome is without a baseline risk, and the drug's maternal relapse-prevention role cannot be inferred from the fetal malformation study alone.

Pregnancy-related changes in lithium handling and the transition around delivery make monitoring a separate research domain. Lithium also raises questions about renal and thyroid physiology, interacting medications, hydration, and neonatal observation. The present dossier describes these scientific domains without supplying an individual dose, target blood level, or treatment plan. Current specialist guidance and product-specific labeling govern contemporary medical application.

Valproate and the Evidence of Developmental Harm

Valproate illustrates the strongest end of the reproductive risk spectrum for a psychotropic and antiseizure medication. Official product labeling describes a dose-related association with major congenital malformations, including neural tube defects, and adverse neurodevelopmental outcomes after in utero exposure. The same molecule can be used in several different clinical indications; a regulatory contraindication for one indication is not identical to the risk–benefit framework for another. This is why the precise product, jurisdiction, diagnosis, and present label must be identified before any clinical assertion. The Food and Drug Administration (2024) details congenital and developmental risks.

This evidence raises a crucial timing issue. Structural anomalies may arise during organogenesis, whereas neurodevelopmental effects can involve exposure across different stages and may only become measurable in early childhood or school years. Register-based malformation estimates, prospective developmental testing, and later diagnoses measure related but distinct endpoints. The quality of these measures, missing outcomes, concomitant antiseizure therapy, underlying maternal illness, and developmental surveillance all affect interpretation.

Regulatory hazard information has to be communicated with precision. A medication can carry major reproductive risks and still present particular dangers if abruptly withdrawn in someone whose underlying condition has severe destabilization or seizure consequences. The evidence-based subject is comparative decision quality under clinical supervision. The specific reproductive risk of valproate cannot be generalized mechanically to all mood stabilizers or antipsychotic drugs.

Antipsychotics, Metabolism, and Obstetric Outcomes

The maternal metabolic effects of some antipsychotics motivate research into gestational diabetes, weight change, fetal growth, and obstetric events. In a Medicaid cohort, Park (2018) compared continuation with discontinuation of individual second-generation antipsychotics. After adjustment, continuation of quetiapine and olanzapine was associated with a higher gestational-diabetes risk in the studied population; results varied by medicine. The central methodological advantage is comparing people with recent use of the same medication, although continuation is still a nonrandomized decision.

A Swedish register study also reported elevated gestational diabetes and large-for-gestational-age outcomes associated with a selected higher-metabolic-risk group, including olanzapine, clozapine, and quetiapine. Heinonen (2022). A separate UK and Hong Kong analysis, published online in 2020 and in print in 2021, found less consistent associations for continuation as a broader treatment category, underscoring how grouping, population, comparator selection, and adjustment can influence results. Wang (2021).

The scientifically defensible synthesis is medicine-specific and outcome-specific. A fetal growth association can arise through metabolic mechanisms, maternal factors, or clinical selection. General claims about all antipsychotics in pregnancy obscure the heterogeneity of these mechanisms and study designs. Research needs repeated metabolic measures, severity-adjusted comparators, exposure timing, and simultaneous assessment of psychiatric stability. The aim is to improve evidence quality and monitoring, rather than present a uniform decision for every patient taking an antipsychotic.

Neonatal Adaptation and Exposure Near Delivery

Late-gestation psychotropic exposure raises questions that differ from first-trimester structural teratogenicity. Neonatal adaptation outcomes include respiratory support, feeding difficulty, altered tone, irritability, tremor, and other clinically observed features, but definitions vary between studies. An outcome such as admission for respiratory assistance is clinically meaningful yet affected by hospital protocols and newborn characteristics. Research must measure the time interval from exposure to delivery and distinguish transient adaptation from persistent injury.

A cohort of 280,090 term infants in Northern California examined SSRI dispensing after twenty weeks' gestation and a composite delayed neonatal adaptation outcome. The study reported the composite in 11.2% of SSRI-exposed infants and 4.4% of unexposed infants, with adjusted odds ratio 2.14 (95% CI 1.96–2.32). The authors also explored type and dose differences. The result is evidence of an association supported by covariate adjustment; residual severity-related and delivery-related confounding remain important, and the composite does not represent a single disease mechanism. Cornet (2024).

The practical research question is how to balance maternal symptom control, any treatment adjustments near delivery, and neonatal observation using a prospective plan. A neonatal adaptation signal cannot alone establish that sudden discontinuation near birth improves the combined maternal–infant outcome. Randomized data about such strategies are limited. Studies must include psychiatric relapse, delivery outcomes, neonatal morbidity, and longer follow-up under realistic care conditions.

Child Neurodevelopment After Prenatal Antidepressant Exposure

Long-term neurodevelopment is a particularly difficult causal domain because parental psychiatric conditions, inherited characteristics, home environment, and health service use affect both prenatal treatment and child outcomes. In a Swedish study using sibling comparisons, Sujan (2017) found that apparent associations with autism and ADHD in population comparisons substantially attenuated under within-family analysis. This change is evidence that familial confounding can explain part of an unadjusted or conventional adjusted association. It does not imply that sibling designs are infallible, since exposure differences within families may themselves reflect time-varying factors.

A later cohort assembling more than three million pregnancies from US insurance data evaluated later-pregnancy antidepressant exposure and childhood neurodevelopmental disorders. The most fully adjusted and sibling analyses did not show a corresponding increase, despite much larger crude associations. The study used several approaches to confounding, illustrating the value of triangulation rather than trusting one model. Suarez (2022).

Evidence continues to develop. A Chan (2026) reported modest pooled associations with some neurodevelopmental outcomes that attenuated or became nonsignificant after confounder adjustment, while grading certainty as low to very low. A McIlvride (2026) observed associations in educational outcomes while identifying maternal depression severity as a substantial confounding concern. These findings have different exposure definitions, endpoints, and adjustment strategies; no single pooled association proves a medication-specific developmental mechanism. Research must keep causal uncertainty visible.

Lactation Pharmacology and Infant Exposure

Breastfeeding creates a pharmacological transfer pathway from maternal circulation to milk, followed by infant ingestion, absorption, metabolism, and elimination. Milk concentration alone is not an infant outcome, and the relative infant dose is a screening measure of estimated exposure rather than a safety certificate. Interpretation depends on milk production, timing of sampling, infant age, prematurity, hepatic and renal maturation, and whether other drugs were used before or after delivery. A healthy term infant and a premature infant may have very different capacities to clear the same substance.

The National Library of Medicine's National Library of Medicine (2026), describes generally low milk levels and often undetectable infant serum sertraline, while noting that some premature infants may accumulate exposure. Its appraisal draws on multiple observational studies and reviews and distinguishes measured transfer from reported developmental follow-up. These observations support population-specific interpretation for sertraline; they cannot be applied unchanged to every antidepressant or mood stabilizer.

A complete lactation analysis asks whether the medicine is essential to maternal stability, how much infant exposure has been measured, what infant harms have been documented, what follow-up duration is available, and which monitoring strategies are supported. Decisions also involve the benefits of breastfeeding, family preferences, availability of alternatives, and clinical supervision. Drug-specific current resources are critical because new case reports and pharmacokinetic studies may alter the evidence after a monograph's scientific review date.

Lithium During Lactation: Monitoring and Evidence Limits

Lithium poses a distinctive lactation challenge because infant exposure can be pharmacologically consequential and clearance capacity varies substantially in early life. A Swedish retrospective study followed thirty breastfed infants exposed to maternal lithium and recorded infant serum concentrations and clinical information. The authors reported mostly low concentrations after the early neonatal period, with unexpectedly high concentrations in two infants during the first month. The conclusion was conditional on strict follow-up rather than a universal declaration of safety. Heinonen (2022).

The study's sample size is small, lacks a randomized comparison, and principally describes a selected care setting with monitoring. Healthy full-term infants receiving specialized surveillance do not establish safety for premature infants, medically fragile infants, or families unable to obtain comparable follow-up. In addition, postpartum maternal dosing, hydration, sleep deprivation, and renal function introduce changes that cannot be inferred from prenatal exposure data.

The correct synthesis separates measured infant lithium concentrations, observed clinical effects, and the strength of evidence for longer-term developmental safety. Short-term absence of a reported adverse event is reassuring in a specific cohort but cannot exclude rare outcomes. The maternal benefit of relapse prevention, the quality of newborn monitoring, and the different infant risk strata all belong in the same research account. No generalized breastfeeding clearance or infant-monitoring prescription can be derived from these data alone.

The Postpartum Transition and Severe Mental Illness

The postpartum period brings rapid biological, sleep, caregiving, and social changes. For some people with bipolar disorder or a history of postpartum psychosis, the risk of severe episodes is concentrated around this transition. Treatment studies must separate depression, bipolar relapse, mania, and postpartum psychosis because their recurrence patterns, acute hazards, and pharmacological strategies differ. Obstetric and psychiatric coordination, continuity of medication information, and crisis recognition are system-level components of effective care.

Bergink (2012) followed seventy pregnant women at elevated risk of postpartum psychosis and documented differences between women with psychosis confined to earlier postpartum periods and those with bipolar disorder. The results made the timing of prophylaxis a central research question. However, the absence of randomization, clinical selection, and small subgroup sizes restrict causal interpretation. Contemporary guidelines therefore provide a broader evidence framework for specialist management. American College of Obstetricians and Gynecologists (2023).

The measurement problem includes maternal sleep, breastfeeding intentions, relapse signs, medication continuity, infant-care capacity, partner support, and service response time. Research should treat maternal mental health and infant health as jointly relevant outcomes while preserving the patient's rights and preferences. A medication decision cannot be evaluated through the fetal-exposure endpoint alone if maternal psychiatric instability presents significant competing risks.

Aging, Frailty, and the Heterogeneity of Older Adults

Later-life psychopharmacology studies medication effectiveness and harms in the context of aging, disease burden, frailty, cognitive change, and life circumstances. Chronological age is an enrollment marker, but clinical vulnerability is distributed unevenly. An active community-dwelling person aged eighty may tolerate a regimen differently from a younger individual with dementia, recurrent falls, malnutrition, and advanced renal disease. The design of a geriatric study must specify cognitive status, living situation, disability, multimorbidity, and functional reserve rather than relying only on an age cutoff.

Reduced renal clearance, interactions affecting hepatic metabolism, altered body composition, changes in receptor responsiveness, and impaired homeostatic compensation can modify central nervous system effects. Drugs associated with sedation, orthostatic hypotension, anticholinergic burden, and impaired psychomotor function can affect fall risk and independence. Polypharmacy makes attribution difficult: an apparent new psychiatric symptom might be the result of a nonpsychiatric drug, an interaction, a withdrawal phenomenon, or an underlying disorder. Canio (2022) reviews polypharmacy as a central geriatric problem.

Frailty also introduces competing outcomes and goals. Symptom relief, preservation of cognition, prevention of falls, avoidance of hospitalization, and quality of life may have different priorities for different patients. Evidence obtained in robust trial volunteers may underestimate adverse events in institutionalized or medically complex populations. Trial inclusion and exclusion criteria are therefore integral scientific facts, not footnotes to a treatment conclusion.

Geriatric Depression and Collaborative Care

Late-life depression intersects with physical illness, loneliness, pain, disability, sleep disturbance, and changes in role or independence. The evidence for medication must be considered within a care system capable of diagnosing depression, measuring response, identifying adverse effects, and providing psychotherapy or psychosocial support. A simple comparison between antidepressants and no treatment overlooks problems of access, follow-up, and adherence that often determine outcomes.

The IMPACT randomized trial enrolled 1,801 adults aged sixty or older across eighteen primary-care clinics and compared collaborative depression care with usual care. Its significance was demonstrating that a structured model with care management and treatment coordination can produce better depression outcomes in a real-world primary care population. The intervention was a service strategy that could include medication and psychotherapy, rather than a blinded trial of a single antidepressant compound. Unützer (2002).

This distinction matters for causal interpretation. Superior collaborative care does not prove that every participant should receive a specific drug, and the effectiveness of a drug in an efficacy trial does not prove that an unmonitored prescription will reproduce the same outcome in routine care. Research should therefore compare medication efficacy, treatment implementation, measurement-based adjustment, psychotherapy, and longitudinal functional outcomes as related but distinguishable components.

Treatment-Resistant Depression in Later Life

The evidence base for treatment-resistant depression in older adults has moved beyond extrapolation from general adult studies. A randomized placebo-controlled trial examined aripiprazole augmentation after inadequate response to venlafaxine in adults aged sixty or above. Among 181 randomized nonremitters, remission occurred in 44% assigned aripiprazole versus 29% assigned placebo; akathisia and parkinsonism were more common with aripiprazole. The result illustrates simultaneous efficacy and adverse-event tradeoffs under defined inclusion criteria. Lenze (2015).

The later OPTIMUM trial compared augmentation and switching approaches in older adults with treatment-resistant depression. Initial augmentation with aripiprazole improved psychological well-being more than switching to bupropion over the study window, while adverse-event patterns included differences in falls between strategies. Later-step comparisons of lithium augmentation and a switch to nortriptyline showed greater uncertainty. Lenze (2023). These estimates belong to trial-specific patients and protocols, not to an undifferentiated ranking of all antidepressant strategies.

Clinical interpretation requires baseline gait and fall risk, cognition, metabolic health, interacting medications, and prior treatment history. In geriatric trials, remission and quality of life should be analyzed alongside motor side effects, akathisia, treatment discontinuation, and hospital outcomes. A medication augmentation benefit remains clinically incomplete unless the patient population and adverse-effect profile are explicitly stated.

Benzodiazepines, Hypnotics, Falls, and Cognitive Safety

Benzodiazepines and related hypnotics provide a prominent example of pharmacodynamic vulnerability in older adults. The American Geriatrics Society 2023 Beers Criteria Update Expert Panel (2023) identifies increased risks of cognitive impairment, delirium, falls, fractures, and driving-related harm with benzodiazepines in many older populations. It also identifies important exceptions and indication-specific contexts; these criteria support a medication review rather than a uniform prohibition applied without clinical assessment. Shorter half-life by itself does not eliminate the class-related concern.

The research mechanisms include sedation, psychomotor slowing, reduced postural compensation, sleep-related changes, and interactions with other central nervous system depressants. Observational studies can detect fall associations but must consider that anxiety, insomnia, frailty, and acute illness may independently elevate risk. Trials may capture symptom reduction over weeks while missing less frequent injuries or cumulative dependence. Medication review must therefore examine efficacy, functional consequences, and withdrawal burden over time.

Discontinuation is itself an intervention requiring evidence. Long-term exposure can produce physiological dependence, and abrupt changes can be dangerous in certain contexts. A scientifically responsible deprescribing analysis compares structured, supported strategies, outcome measurement, relapse, withdrawal symptoms, and patient preference. It does not equate medication risk with a recommendation that individual readers suddenly stop treatment. The Beers and STOPP/START frameworks are tools for clinician-mediated assessment of appropriateness in context.

Antipsychotics in Dementia and Delirium

Antipsychotic use in dementia-related behavioral symptoms is a high-stakes benefit–risk problem. Behavioral and psychological symptoms can be distressing and at times dangerous, yet treatment may increase risks of mortality, cerebrovascular events, sedation, falls, and cognitive worsening in susceptible patients. The American Geriatrics Society 2023 Beers Criteria Update Expert Panel (2023) advises avoidance for behavioral problems of dementia or delirium except under specific conditions involving failure or infeasibility of nonpharmacologic measures and substantial harm risk. The evidence must be interpreted in the context of indication and severity.

Dementia-related agitation is heterogeneous. Pain, infection, medication side effects, sensory deprivation, changes in environment, or unmet needs can contribute to behavioral episodes. Studies that classify all behavioral disturbance as a primary psychiatric drug target risk overlooking reversible causes. Symptom scales, caregiver outcomes, emergency events, stroke, mobility, and mortality should be measured jointly, with sufficient follow-up and adverse-event reporting.

These considerations also clarify distinctions among dementia, delirium, and primary psychotic disorders. Different syndromes carry different mechanisms and drug indications. An antipsychotic may be appropriate for certain severe psychiatric conditions while carrying an unfavorable balance for routine sedation in dementia. Research should avoid collapsing separate diagnostic populations into a generic older-adult category and should analyze alternative nonpharmacologic and supportive care strategies with equal rigor.

Anticholinergic Burden, Polypharmacy, and Drug–Disease Interactions

The burden of medications with anticholinergic activity is especially relevant to cognition, constipation, urinary retention, dry mouth, falls, and delirium risk in older adults. Several psychiatric and nonpsychiatric agents contribute to this burden, creating combinations that are difficult to recognize when medicines are reviewed one by one. A pharmacologically diverse regimen can have an additive clinical effect even when each component was initially prescribed for a distinct reason.

Polypharmacy also generates prescribing cascades: an adverse reaction is interpreted as a new diagnosis and treated with another medication. Research examining this process needs medication timelines, clinical indication, symptom onset, withdrawal or dose-change information, and objective physiological measures. Drug–drug interactions are influenced by enzyme pathways and combined pharmacodynamic action, while drug–disease interactions arise when a medication aggravates an existing condition. Canio (2022) and the American Geriatrics Society 2023 Beers Criteria Update Expert Panel (2023) provide systematic approaches to these issues.

Reducing medication count is not a sufficient outcome by itself. A smaller regimen can represent improved appropriateness or undertreatment depending on which therapies were stopped and why. The O'Mahony (2023) explicitly address both potentially inappropriate prescriptions and clinically important treatment omissions. Research must therefore evaluate medication appropriateness, indication, patient-valued functioning, and health outcomes rather than treating numerical pill reduction as a universal success.

Confounding, Bias, and the Problem of Generalization

Every life-stage psychopharmacology study faces the possibility that treatment selection reflects disease severity rather than merely prescribing preference. Patients who continue medication may have recurrent illness; pregnant people who stop treatment may differ in resources and symptom history; older adults receiving sedatives may have sleep disturbance, acute pain, or frailty that already predicts injury. Investigators must define the causal question and compare treatments among populations for whom the alternatives are clinically plausible.

Confounding control can improve interpretation without guaranteeing causal truth. Propensity-score adjustment requires measured covariates; sibling comparisons address stable family characteristics while allowing changes across pregnancies; negative controls can reveal common causes; randomized trials can still fail to transport to excluded populations. Sujan (2017), Patorno (2017), and Park (2018) illustrate different comparator problems.

Transportability requires a documented target population. A drug may be effective in a school-age trial but remain uncertain in preschool children, in pregnancy complicated by multiple illnesses, or in frail older adults with dementia. Trial averages conceal individual variation, and excluded groups may experience different exposure and harm. Scientific synthesis should therefore separate the directly studied population, reasonable mechanistic extrapolation, and unverified clinical assumptions. This distinction is a foundation of evidence integrity rather than a mere qualification.

Regulatory Labeling, Guidelines, and Jurisdiction

Regulatory authorization, professional recommendations, and scientific evidence perform different functions. A product label summarizes approved uses, restrictions, contraindications, and safety information under a particular jurisdiction and date. A professional guideline integrates the available literature for practice, including judgments about disease severity and treatment alternatives. An original trial supplies empirical findings under its design. A monograph should refer to all three while distinguishing the authority of each.

The FDA's Food and Drug Administration (2014), American College of Obstetricians and Gynecologists (2023), National Library of Medicine (2026), American Geriatrics Society 2023 Beers Criteria Update Expert Panel (2023), and O'Mahony (2023) represent complementary evidence infrastructures. They are not interchangeable. In particular, a country-specific approval label should be identified as such, and a geriatric prescribing criterion is a decision-support resource rather than an exhaustive account of every patient.

Updated medication advice requires exact product, indication, formulation, population, regulatory jurisdiction, and date. Product labels and safety communications can change after the research date, and unverified class-level generalizations may be misleading. This monograph treats medication safety at the level of scientific principles and selected verified studies; it does not function as a current prescribing manual for an individual patient.

Equity, Access, Culture, and Treatment Burden

Access to specialized pediatric care, perinatal psychiatric consultation, and geriatric pharmacology varies between health systems and regions. These differences affect who receives medication, who receives evidence-based psychotherapy, how frequently adverse events are detected, and whether follow-up is possible. Observational data derived from insured populations or selected academic clinics may therefore underrepresent economically marginalized, rural, displaced, or medically complex groups.

Care responsibilities also alter the psychological consequences of treatment. A teenager may experience medication effects in relation to school performance and family decisions; a pregnant patient may navigate conflicting medical advice, stigma, and reproductive autonomy; an older adult may prioritize mobility or cognitive clarity over improvement on a clinician-rated scale. A complete scientific account requires shared decision research, caregiver burden assessment, communication accessibility, and transparent statements about whose preferences determine measured benefit.

The international character of this field demands that findings specify the health system, diagnostic procedures, population, and regulatory environment in which they were generated. Cross-country transfer may be valid for biological mechanisms but weaker for service delivery, adherence, or detection of outcomes. A rigorous monograph reports contextual limits while preserving the comparability of core psychopharmacological measures.

Digital Evidence, Artificial Systems, and Scientific Integrity

Artificial systems can assist in searching medication labels, identifying relevant randomized trials, extracting study design information, and organizing life-stage-specific drug-safety comparisons. Their scientific value depends on traceability. A generated statement about pregnancy safety must link to the exact original publication or current official label; a citation to an unrelated drug, an outdated regulatory warning, or a nonexistent numerical estimate is a material evidence failure.

The Artificial Era introduces new information pathways between clinicians, patients, families, and biomedical literature. Such systems may improve access to structured evidence while also accelerating the spread of plausible but unsupported treatment claims. Evaluation must measure reference correctness, retrieval completeness, version awareness, uncertainty calibration, subgroup accuracy, and whether a system conflates experimental observations with clinical advice. In high-risk reproductive and geriatric contexts, error should be audited at the level of clinical consequences rather than surface fluency.

The present dossier connects these research questions to the wider scientific architecture of Psychology of Artificial and Human–Artificial Relations. The philosophical horizon of Artificial Era · From Homo to Artificial supplies a context for examining artificial participation in evidence interpretation; the claims concerning drug efficacy, fetal safety, and clinical outcomes remain answerable to independent biomedical studies. No original artificial-cognition experiment or clinical validation study is claimed in this monograph.

Critical Scientific Synthesis

The evidence across childhood, pregnancy, lactation, and advanced age supports a common explanatory principle: drug effects and treatment consequences are conditional on physiological state, illness severity, clinical indication, care organization, and time. The strongest pediatric trials establish efficacy for specific indications and tested strategies; they also show that symptom, functioning, and safety outcomes require separate interpretation. Pregnancy cohorts expose the importance of confounding by indication and precise exposure windows, while drug-specific findings for lithium and valproate demonstrate why neither blanket reassurance nor blanket risk statements are scientifically defensible.

Lactation research emphasizes a pharmacokinetic transfer chain extending beyond maternal concentration to infant dose, clearance, and follow-up. Geriatric trials provide direct evidence that late-life depression can respond to structured interventions and particular augmentation strategies, while prescribing criteria expose increased vulnerability to falls, delirium, cognitive effects, and cumulative interaction burdens. These literatures are connected through pharmacology, causal inference, and clinical monitoring, but they do not collapse into a single average life-course effect.

An integrated evidence standard therefore requires explicit population, indication, comparator, duration, exposure ascertainment, outcome definition, and uncertainty. It also requires a separate account of the harm associated with untreated or undertreated illness. The strongest scientific contribution of life-stage psychopharmacology is the disciplined specification of these conditions, permitting research evidence to guide decisions without disguising missing knowledge.

Research Gaps and Future Directions

Long-term pediatric follow-up remains a major priority. Trials should combine symptom and functional measurement with growth, metabolic, social, and educational outcomes, and investigate whether changes during developmental transitions persist into adulthood. Independent replication, registries linked to carefully measured clinical severity, and better inclusion of medically complex children can strengthen generalizability. Research on comparative care packages should disentangle pharmacological action, psychotherapy, monitoring intensity, and family support.

Perinatal research needs prospective assessment before conception where feasible, more precise drug exposure and adherence measurement, carefully defined active comparators, and studies that distinguish pregnancy losses, delivery outcomes, neonatal adaptation, and later child development. The contradictory appearance of some adjusted and unadjusted neurodevelopmental findings makes causal triangulation especially important. Drug-specific pregnancy and lactation evidence must be continuously updated; rare outcomes require collaborative surveillance across populations while protecting privacy and reproductive autonomy.

Geriatric priorities include trials enrolling participants with frailty, cognitive impairment, multimorbidity, and polypharmacy, as well as pragmatic evaluations of deprescribing, treatment switching, and collaborative care. Future studies should estimate joint benefit–harm profiles instead of testing symptom endpoints in isolation. Across all age groups, patient-centered outcomes and transparent uncertainty reporting will be decisive. Research on artificial clinical-information systems must independently test source fidelity, subgroup accuracy, and the harms of mistaken guidance.

Conclusion

Developmental, reproductive, and later-life psychopharmacology is a distinct scientific field organized around variation in medication effects and consequences across biological and social stages of life. Its foundations combine clinical psychopharmacology, developmental science, reproductive epidemiology, geriatrics, pharmacokinetics, randomized evidence, and longitudinal causal inference. The empirical record demonstrates significant treatment benefits for defined indications and populations, identifies drug-specific reproductive and geriatric hazards, and establishes the need to measure physiological, psychological, developmental, and functional outcomes together.

The principal scientific conclusion is that medication benefit and harm are functions of a defined clinical context observed over time. Robust inference depends on specifying illness severity, competing care strategies, developmental or reproductive state, drug exposure, outcome timing, and the quality of available comparisons. Pediatric symptom efficacy, prenatal structural risk, neonatal adaptation, lactational transfer, geriatric remission, and falls are different causal questions. Their integration constitutes a complete research program because each contributes to understanding the consequences of psychotropic exposure throughout the life course.

Future progress depends on longitudinal and pragmatic evidence, transparent drug-specific regulation, high-quality comparisons, and more representative participants. In the Artificial Era, the reliability of systems that communicate such evidence becomes another measurable research problem. The enduring standard is that claims about treatment and safety must be traceable to the population, experiment, observation, and document that support them.

References

  • American College of Obstetricians and Gynecologists. (2023). Treatment and management of mental health conditions during pregnancy and postpartum: ACOG Clinical Practice Guideline No. 5. Obstetrics & Gynecology, 141(6), 1262–1288. https://doi.org/10.1097/AOG.0000000000005202
  • American Geriatrics Society 2023 Beers Criteria Update Expert Panel. (2023). American Geriatrics Society 2023 updated AGS Beers Criteria for potentially inappropriate medication use in older adults. Journal of the American Geriatrics Society, 71(7), 2052–2081. https://doi.org/10.1111/jgs.18372
  • Bergink, V., Bouvy, P. F., Vervoort, J. S. P., Koorengevel, K. M., Steegers, E. A. P., & Kushner, S. A. (2012). Prevention of postpartum psychosis and mania in women at high risk. American Journal of Psychiatry, 169(6), 609–615. https://doi.org/10.1176/appi.ajp.2012.11071047
  • Canio, W. C. (2022). Polypharmacy in older adults. Clinics in Geriatric Medicine, 38(4), 621–625. https://doi.org/10.1016/j.cger.2022.05.004
  • Chan, J. K. N., Zhong, A. H. F., Lam, J. Y. H., Wong, C. S. M., Solmi, M., Correll, C. U., & Chang, W. C. (2026). Maternal and paternal antidepressant use before and during pregnancy and offspring risk of neurodevelopmental disorders: A systematic review and meta-analysis. The Lancet Psychiatry, 13(6), 472–484. https://doi.org/10.1016/S2215-0366(26)00089-1
  • Cohen, L. S., Altshuler, L. L., Harlow, B. L., et al. (2006). Relapse of major depression during pregnancy in women who maintain or discontinue antidepressant treatment. JAMA, 295(5), 499–507. https://doi.org/10.1001/jama.295.5.499
  • Cornet, M.-C., Wu, Y. W., Forquer, H., Avalos, L. A., Sriram, A., Scheffler, A. W., Newman, T. B., & Kuzniewicz, M. W. (2024). Maternal treatment with selective serotonin reuptake inhibitors during pregnancy and delayed neonatal adaptation: A population-based cohort study. Archives of Disease in Childhood: Fetal and Neonatal Edition, 109(3), 294–300. https://doi.org/10.1136/archdischild-2023-326049
  • Food and Drug Administration. (2005). Antidepressant class suicidality warning in pediatric labeling: Pooled short-term clinical trial information [Historical labeling, not a current prescribing label]. https://www.accessdata.fda.gov/drugsatfda_docs/label/2005/020415s018%2C021208s009lbl.pdf
  • Food and Drug Administration. (2014/2015). Pregnancy and Lactation Labeling Rule: Content and format of labeling. https://www.fda.gov/vaccines-blood-biologics/biologics-rules/pregnancy-and-lactation-labeling-final-rule
  • Food and Drug Administration. (2024). Depakote (divalproex sodium): Prescribing information [Drug labeling]. https://www.accessdata.fda.gov/drugsatfda_docs/label/2024/018723s068lbl.pdf
  • Heinonen, E., Forsberg, L., Nörby, U., Wide, K., & Källén, K. (2022). Antipsychotic use during pregnancy and risk for gestational diabetes: A national register-based cohort study in Sweden. CNS Drugs, 36(5), 529–539. https://doi.org/10.1007/s40263-022-00908-2
  • Heinonen, E., Tötterman, K., Bäck, K., Sarman, I., Svedenkrans, J., & Forsberg, L. (2022). Lithium use during breastfeeding was safe in healthy full-term infants under strict monitoring. Acta Paediatrica, 111(10), 1891–1898. https://doi.org/10.1111/apa.16444
  • Huybrechts, K. F., Palmsten, K., Avorn, J., et al. (2014). Antidepressant use in pregnancy and the risk of cardiac defects. New England Journal of Medicine, 370(25), 2397–2407. https://doi.org/10.1056/NEJMoa1312828
  • Lenze, E. J., Mulsant, B. H., Blumberger, D. M., Karp, J. F., Newcomer, J. W., Anderson, S. J., Dew, M. A., Butters, M. A., Stack, J. A., Begley, A. E., & Reynolds, C. F., III. (2015). Efficacy, safety, and tolerability of augmentation pharmacotherapy with aripiprazole for treatment-resistant depression in late life: A randomised, double-blind, placebo-controlled trial. The Lancet, 386(10011), 2404–2412. https://doi.org/10.1016/S0140-6736(15)00308-6
  • Lenze, E. J., Mulsant, B. H., Roose, S. P., Lavretsky, H., Reynolds, C. F., III, Blumberger, D. M., Brown, P. J., Cristancho, P., Flint, A. J., Gebara, M. A., Gettinger, T. R., Lenard, E., Miller, J. P., Nicol, G. E., Oughli, H. A., Pham, V. T., Rollman, B. L., Yang, L., & Karp, J. F. (2023). Antidepressant augmentation versus switch in treatment-resistant geriatric depression. New England Journal of Medicine, 388(12), 1067–1079. https://doi.org/10.1056/NEJMoa2204462
  • McIlvride, S., Rao, N., Singh, S., Nelson, S. M., Pell, J. P., & Fleming, M. (2026). Prenatal exposure to maternal depression and antidepressants and neurodevelopmental outcomes: A population cohort study. PLOS Medicine, 23(9), e1004658. https://doi.org/10.1371/journal.pmed.1004658
  • MTA Cooperative Group. (1999). A 14-month randomized clinical trial of treatment strategies for attention-deficit/hyperactivity disorder. Archives of General Psychiatry, 56(12), 1073–1086. https://doi.org/10.1001/archpsyc.56.12.1073
  • National Library of Medicine. (2026, August 15). Sertraline. Drugs and Lactation Database (LactMed). National Center for Biotechnology Information. https://www.ncbi.nlm.nih.gov/books/NBK501191/
  • O'Mahony, D., Cherubini, A., Renom Guiteras, A., et al. (2023). STOPP/START criteria for potentially inappropriate prescribing in older people: Version 3. European Geriatric Medicine, 14(4), 625–632. https://doi.org/10.1007/s41999-023-00777-y
  • Park, Y., Hernandez-Diaz, S., Bateman, B. T., et al. (2018). Continuation of atypical antipsychotic medication during early pregnancy and the risk of gestational diabetes. American Journal of Psychiatry, 175(6), 564–574. https://doi.org/10.1176/appi.ajp.2018.17040393
  • Patorno, E., Huybrechts, K. F., Bateman, B. T., Cohen, J. M., Desai, R. J., Mogun, H., Cohen, L. S., & Hernández-Díaz, S. (2017). Lithium use in pregnancy and the risk of cardiac malformations. New England Journal of Medicine, 376(23), 2245–2254. https://doi.org/10.1056/NEJMoa1612222
  • Suarez, E. A., Bateman, B. T., Hernández-Díaz, S., Straub, L., Wisner, K. L., Gray, K. J., Pennell, P. B., Lester, B., McDougle, C. J., Zhu, Y., Mogun, H., & Huybrechts, K. F. (2022). Association of antidepressant use during pregnancy with risk of neurodevelopmental disorders in children. JAMA Internal Medicine, 182(11), 1149–1160. https://doi.org/10.1001/jamainternmed.2022.4268
  • Sujan, A. C., Rickert, M. E., Öberg, A. S., et al. (2017). Associations of maternal antidepressant use during the first trimester of pregnancy with preterm birth, small for gestational age, autism spectrum disorder, and attention-deficit/hyperactivity disorder in offspring. JAMA, 317(15), 1553–1562. https://doi.org/10.1001/jama.2017.3413
  • Treatment for Adolescents With Depression Study (TADS) Team. (2004). Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents With Depression Study (TADS) randomized controlled trial. JAMA, 292(7), 807–820. https://doi.org/10.1001/jama.292.7.807
  • Unützer, J., Katon, W., Callahan, C. M., et al. (2002). Collaborative care management of late-life depression in the primary care setting: A randomized controlled trial. JAMA, 288(22), 2836–2845. https://doi.org/10.1001/jama.288.22.2836
  • Walkup, J. T., Albano, A. M., Piacentini, J., et al. (2008). Cognitive behavioral therapy, sertraline, or a combination in childhood anxiety. New England Journal of Medicine, 359(26), 2753–2766. https://doi.org/10.1056/NEJMoa0804633
  • Wang, Z., Man, K. K. C., Ma, T., Howard, L. M., Wei, L., Wong, I. C. K., & Brauer, R. (2021). Association between antipsychotic use in pregnancy and the risk of gestational diabetes: Population-based cohort studies from the United Kingdom and Hong Kong and an updated meta-analysis. Schizophrenia Research, 229, 55–62. https://doi.org/10.1016/j.schres.2020.11.021