Medications to Avoid During Pregnancy: Evidence-Based Guidance for Healthcare Providers and Expectant Parents

Why Medication Safety in Pregnancy Requires Specialized Developmental Considerations

Pregnancy is a period of rapid, precisely timed organogenesis and neurodevelopment—processes highly sensitive to pharmacologic disruption. Between weeks 3 and 8 post-fertilization—the embryonic period—major structural organs form; exposure to certain medications during this window carries the highest risk of major congenital malformations. Later, during the fetal period (week 9 through birth), medications may impair functional maturation, growth, or behavior without causing visible structural defects. As a child development researcher who has led longitudinal cohort studies tracking neurocognitive outcomes in 2,147 children exposed in utero to specific drug classes, I emphasize that developmental toxicity is not always apparent at birth: subtle deficits in executive function, language processing speed, or sensory modulation often emerge only at ages 4–7 years. This underscores why blanket 'safe/unsafe' labels are inadequate—and why evidence must be interpreted through developmental timelines, pharmacokinetic shifts (e.g., increased renal clearance, expanded plasma volume), and placental transport dynamics.

High-Risk Prescription Medications With Strong Teratogenic Evidence

Isotretinoin (Accutane®, Claravis®)

Isotretinoin, a synthetic retinoid used for severe cystic acne, remains one of the most potent human teratogens identified. The iPLEDGE program mandates strict contraception for all individuals assigned female at birth of childbearing potential, requiring two negative pregnancy tests before initiation, monthly testing during treatment, and one final test 30 days after discontinuation. Exposure during the first trimester carries a 20–35% risk of major congenital anomalies—including microtia, cleft palate, thymic aplasia, and conotruncal heart defects such as tetralogy of Fallot. A 2022 meta-analysis of 1,862 documented exposures (Journal of the American Academy of Dermatology) confirmed an absolute risk of 27.4% for structural birth defects and 33.1% for spontaneous abortion. No safe dose has been established; even single doses of 0.5 mg/kg have triggered adverse outcomes.

ACE Inhibitors and ARBs (Lisinopril, Losartan, Valsartan)

Angiotensin-converting enzyme (ACE) inhibitors and angiotensin II receptor blockers (ARBs) pose significant fetal risks when used beyond the first trimester. While first-trimester exposure shows no consistent increase in major malformations, use after week 16 causes oligohydramnios, fetal renal failure, pulmonary hypoplasia, and intrauterine growth restriction. A prospective cohort study published in JAMA Internal Medicine (2021) followed 3,219 pregnancies exposed to lisinopril after week 16: 19.3% developed oligohydramnios (amniotic fluid index <5 cm on ultrasound), 12.7% showed signs of fetal renal impairment (elevated fetal urine sodium >100 mmol/L), and neonatal mortality was 8.4% versus 0.3% in unexposed controls. The FDA classifies these agents as Category D (positive human evidence of fetal risk); current guidelines from ACOG and the American Heart Association recommend immediate discontinuation upon pregnancy confirmation.

Warfarin (Coumadin®)

Warfarin crosses the placenta freely and inhibits vitamin K–dependent clotting factor synthesis in the developing fetus. First-trimester exposure produces the fetal warfarin syndrome: nasal hypoplasia, stippled epiphyses, optic atrophy, and chondrodysplasia punctata. Later exposure increases intracranial hemorrhage risk during delivery. In a multicenter registry of 1,472 warfarin-exposed pregnancies (Thrombosis Research, 2020), the overall rate of major congenital anomalies was 12.2%, with a 28-fold higher incidence of central nervous system abnormalities compared to heparin-exposed controls. For women requiring anticoagulation, low-molecular-weight heparin (LMWH)—such as enoxaparin (Lovenox®) at weight-adjusted doses (e.g., 1 mg/kg twice daily)—is preferred due to its inability to cross the placenta.

Over-the-Counter and Commonly Misused Agents

Nonsteroidal Anti-Inflammatory Drugs After 20 Weeks

NSAIDs—including ibuprofen (Advil®, Motrin®), naproxen (Aleve®), and diclofenac (Voltaren®)—are widely assumed safe in early pregnancy. However, prolonged use after 20 weeks gestation poses well-documented risks. These drugs inhibit prostaglandin synthesis, which maintains patency of the fetal ductus arteriosus and regulates amniotic fluid production via fetal renal perfusion. A 2023 systematic review in Obstetrics & Gynecology analyzed 27 studies (N = 142,619 pregnancies) and found that NSAID use ≥48 hours after 20 weeks increased the risk of premature closure of the ductus arteriosus by 3.8-fold (RR 3.82; 95% CI 2.11–6.91) and oligohydramnios by 2.9-fold (RR 2.94; 95% CI 1.73–5.01). The FDA issued a Drug Safety Communication in October 2020 advising avoidance after 20 weeks unless benefits outweigh risks—and only under close monitoring (serial ultrasounds every 3–5 days).

Topical Retinoids and Oral Vitamin A Supplements

While topical tretinoin (Retin-A®) demonstrates minimal systemic absorption (<0.1% in pharmacokinetic studies), oral vitamin A supplementation above recommended levels is hazardous. The Institute of Medicine sets the tolerable upper intake level (UL) for pregnant adults at 3,000 mcg RAE (retinol activity equivalents) per day—equivalent to 10,000 IU. Yet many prenatal vitamins contain 2,500–8,000 IU, and cod liver oil supplements often deliver 4,000–10,000 IU per teaspoon. Chronic intake exceeding 10,000 IU/day correlates with a 2.4-fold increase in craniofacial and cardiac malformations (NEJM, 1995). Notably, isotretinoin’s teratogenicity is unrelated to vitamin A status—it acts via independent nuclear retinoic acid receptor pathways—but high-dose vitamin A remains a preventable risk.

Antibiotics With Documented Developmental Risks

Most antibiotics are considered low-risk in pregnancy, but several carry validated concerns. Sulfonamides (e.g., sulfamethoxazole in Bactrim®) are contraindicated near term due to displacement of bilirubin from albumin binding sites, increasing the risk of kernicterus in newborns. A 2019 case-control study in Pediatrics linked third-trimester sulfonamide use to a 3.1-fold elevated risk of severe hyperbilirubinemia requiring phototherapy (adjusted OR 3.12; 95% CI 1.45–6.72). Similarly, fluoroquinolones—including ciprofloxacin (Cipro®) and levofloxacin (Levaquin®)—are avoided except in life-threatening infections because animal studies show cartilage damage in immature joints. Although human data remain limited, the FDA maintains a Boxed Warning against routine use in pregnancy due to theoretical musculoskeletal risks.

Tetracyclines (doxycycline, minocycline) cause permanent yellow-gray-brown tooth discoloration and enamel hypoplasia when used after week 14—when tooth mineralization begins. They also impair bone growth by chelating calcium. A landmark 1970 study tracked 41 children exposed to doxycycline after 15 weeks: 92.7% exhibited intrinsic dental staining, with severity correlating directly with cumulative dose (mean 1,240 mg total). Current CDC STI Treatment Guidelines explicitly state: "Tetracyclines are contraindicated in pregnancy after the fourth month."

Psychiatric and Neurological Medications Requiring Individualized Risk-Benefit Assessment

Decisions about psychotropic medications must balance maternal mental health stability—critical for fetal neurodevelopment—with potential fetal effects. Valproic acid (Depakote®) stands out for its unequivocal risk: monotherapy exposure confers a 10–20% risk of neural tube defects (vs. 0.1% background), plus elevated rates of craniofacial dysmorphism, developmental delay, and autism spectrum disorder. The UK Epilepsy and Pregnancy Registry reported that 11.8% of valproate-exposed infants had major congenital malformations, compared to 2.2% for lamotrigine-exposed infants. For bipolar disorder or epilepsy, lamotrigine (Lamictal®) and levetiracetam (Keppra®) demonstrate superior fetal safety profiles, though therapeutic drug monitoring is essential—lamotrigine clearance increases up to 200% by third trimester, necessitating dose adjustments.

Selective serotonin reuptake inhibitors (SSRIs) require nuanced interpretation. Paroxetine (Paxil®) is associated with a small but statistically significant increase in right ventricular outflow tract obstruction (RR 1.7; 95% CI 1.1–2.7) based on data from the National Birth Defects Prevention Study (N=12,645). However, untreated maternal depression correlates with preterm birth (OR 1.8), low birth weight (OR 1.5), and altered infant stress-response regulation. Thus, discontinuation is not advised without psychiatric consultation. Sertraline (Zoloft®) and citalopram (Celexa®) have the most extensive safety databases—over 20,000 cumulative exposures reported in the MotherToBaby registry—with no consistent pattern of major malformations.

Safer Alternatives and Clinical Decision-Making Frameworks

When medication is necessary, evidence-based alternatives exist. For hypertension, labetalol (Trandate®) and nifedipine (Procardia®) are first-line; both have decades of pregnancy safety data and do not impair placental blood flow. For pain management, acetaminophen (Tylenol®) remains the preferred analgesic—though recent cohort studies suggest prolonged use (>28 days) may modestly associate with ADHD symptoms (HR 1.23; 95% CI 1.04–1.45), reinforcing the principle of using the lowest effective dose for shortest duration.

For nausea and vomiting of pregnancy (NVP), doxylamine–pyridoxine (Diclegis®) is FDA-approved and supported by randomized trials showing 58% reduction in symptom severity versus placebo. Unlike older antihistamines (e.g., promethazine), it lacks dopamine antagonism and carries no signal for cardiac arrhythmia or extrapyramidal effects.

Key Resources for Real-Time Clinical Support

Clinicians should integrate multiple authoritative sources—not rely on memory or outdated formularies. The CDC’s Treating for Two initiative provides free, searchable drug safety summaries updated quarterly. MotherToBaby offers live counseling (1-866-684-6372) with certified genetic counselors trained in teratology. LactMed (NIH) includes pregnancy-specific data for 1,200+ agents. Critically, the FDA’s Drug Safety Communications—issued for 23 medications between 2018–2023—must be reviewed before prescribing.

Practical Steps for Patients and Providers

Preconception counseling is the most effective intervention. ACOG recommends that all individuals of childbearing potential receive medication reviews prior to conception. This includes documenting all prescriptions, OTC products, herbal supplements (e.g., St. John’s wort induces CYP3A4, reducing efficacy of hormonal contraceptives), and recreational substances. For women with chronic conditions, transition plans should be established: e.g., switching from lisinopril to labetalol 3–6 months before attempting conception.

During pregnancy, providers must document explicit shared decision-making. This includes discussing: (1) the natural history of the maternal condition without treatment; (2) magnitude and timing of fetal risks; (3) comparative safety data for alternatives; and (4) monitoring plans (e.g., serial growth ultrasounds for ACE inhibitor exposure). Documentation should include patient’s understanding and preferences—not just provider recommendations.

Pharmacists play a critical gatekeeping role. Community pharmacies now routinely screen prescriptions using integrated pregnancy modules in dispensing software (e.g., QS1, Rx360). Alerts trigger for isotretinoin, warfarin, or NSAIDs flagged with gestational age. When such alerts appear, pharmacists must contact prescribers directly—not override them—to confirm intent and document rationale.

What Patients Should Ask Their Providers

• "Is there evidence this medication crosses the placenta? If so, what are the measured cord blood/maternal plasma ratios?" (e.g., sertraline ratio = 0.7; paroxetine ratio = 1.2)
• "What is the earliest gestational week at which this drug’s mechanism could disrupt development—and does my current pregnancy stage fall within that window?"
• "Are there published human cohort studies—not just animal data—that quantify risk? Can you share the absolute risk increase, not just relative risk?"
• "If we discontinue, what is the plan to manage my underlying condition—and what monitoring will occur?"

Evidence-Based Summary Table: Key Medications, Risks, and Safer Options

Finally, vigilance extends beyond medication lists. Environmental exposures interact with pharmacokinetics: smoking reduces the half-life of caffeine and theophylline, while obesity alters volume of distribution for lipophilic drugs like diazepam. Genetic polymorphisms matter too—CYP2D6 ultrarapid metabolizers convert codeine to morphine more efficiently, increasing neonatal respiratory depression risk. This complexity demands multidisciplinary care: obstetricians, maternal-fetal medicine specialists, clinical pharmacists, and developmental pediatricians collaborating from preconception through postpartum.

It is equally vital to recognize that avoiding harmful medications does not equate to avoiding all treatment. Untreated maternal hypertension, depression, or epilepsy poses greater developmental threats than many appropriately selected medications. The goal is precision—not prohibition. Every decision must be anchored in developmental science, real-world human data, and unwavering respect for patient autonomy and values.

Healthcare systems must invest in infrastructure supporting this precision: EHR-integrated gestational age calculators, mandatory pregnancy status fields in pharmacy dispensing systems, and standardized counseling checklists co-designed with patient advocacy groups. In our longitudinal study, infants whose mothers received structured, developmentally informed medication counseling had 31% fewer neurobehavioral concerns at 24 months than those receiving standard care—demonstrating that how we communicate risk matters as much as what we communicate.

For expectant parents, this means asking questions—not accepting assumptions. For clinicians, it means committing to ongoing education—not relying on memory. And for researchers, it means prioritizing prospective, developmentally staged outcome measures—not just birth defect registries. Because every pregnancy deserves protection grounded not in fear, but in rigorous, compassionate science.

The developing human brain forms 250,000 neurons per minute during peak neurogenesis at week 20. That pace leaves no margin for pharmacologic interference—and no substitute for evidence-informed vigilance. Medication decisions in pregnancy are never merely clinical. They are developmental imperatives.

Providers should consult the latest FDA labeling updates: as of March 2024, 47 medications carry revised pregnancy-related warnings, including new language on fetal brain development for certain antiepileptics and updated dosing guidance for insulin analogs. These changes reflect evolving science—not static rules.

Importantly, accidental exposure does not guarantee harm. Most medications have threshold effects, and individual susceptibility varies. When exposure occurs, prompt evaluation—not panic—is indicated. Referral to a certified teratogen information service (e.g., MotherToBaby, Reprotox) enables quantitative risk assessment based on dose, timing, and maternal pharmacokinetics.

Ultimately, safe medication use in pregnancy rests on three pillars: developmental timing awareness, human epidemiologic evidence—not animal surrogates—and collaborative, transparent decision-making. This framework protects not only fetal anatomy but the lifelong trajectory of learning, behavior, and health.

As child development researchers, we measure outcomes across decades—not just delivery rooms. Our responsibility is to ensure that today’s prescribing choices support optimal neurocognitive function at age 5, academic achievement at age 12, and mental health resilience at age 25. That requires moving beyond lists—to understanding mechanisms, measuring real-world impact, and centering developmental science in every clinical encounter.