Medicines in Pregnancy and Older Age
TL;DR. Two populations where the usual rules change, and where the commonest error runs in opposite directions. In pregnancy, the error is stopping necessary treatment: untreated epilepsy, asthma, and depression carry real risks to both mother and baby, and a great deal of reflexive discontinuation causes more harm than the drug would. In older age, the error is continuing treatment: drugs accumulate, prescriptions pile up, and each additional medicine raises the risk of interactions and falls. Deprescribing is as much a clinical skill as prescribing.
Pregnancy and breastfeeding
1. What changes physiologically
| Change | Effect on drugs |
|---|---|
| Blood volume up 40 to 50 percent | Dilution; some drugs need higher doses |
| Kidney filtration up 50 percent | Renally cleared drugs eliminated faster (lithium, some antibiotics, levetiracetam) |
| Liver enzyme activity shifts | CYP3A4 and CYP2D6 induced, CYP1A2 reduced |
| Reduced plasma albumin | More free (active) drug for highly protein-bound drugs |
| Slower gastric emptying | Slower absorption |
| Increased body fat | Larger reservoir for fat-soluble drugs |
The practical consequence: several drugs need dose increases during pregnancy, notably levothyroxine (usually by 25 to 50 percent, as soon as pregnancy is confirmed) and several antiepileptics, which is why levels are monitored.
2. The timing of risk
| Stage | Risk |
|---|---|
| Weeks 0 to 2 (pre-implantation) | "All or nothing": either the embryo is lost or develops normally |
| Weeks 3 to 8 (organogenesis) | The period of maximum structural malformation risk. Often before pregnancy is known |
| Weeks 9 to birth | Growth and functional development; risks are to growth, brain development, and specific organs |
| Near term | Effects on labour, on the neonate, and on withdrawal |
That third row is why preconception planning matters: the highest-risk window is largely over by the time many people know they are pregnant. Anyone taking long-term medication who might become pregnant should have that conversation in advance rather than in a crisis.
3. The known teratogens
| Drug | Effect |
|---|---|
| Sodium valproate | Major malformations in ~10 percent; neurodevelopmental disorders in 30 to 40 percent. Contraindicated in anyone who could become pregnant unless under a strict pregnancy prevention programme. One of the most serious medicines scandals of recent decades |
| Isotretinoin | Severe malformations. Strict pregnancy prevention programme required |
| Thalidomide | Limb malformations; the case that created modern drug regulation (Chapter 63) |
| ACE inhibitors, ARBs | Fetal kidney damage, oligohydramnios, death (2nd/3rd trimester) |
| Warfarin | Fetal warfarin syndrome; heparins used instead |
| Methotrexate, mycophenolate | Miscarriage, malformations |
| Tetracyclines | Tooth and bone effects after 15 weeks |
| Live vaccines | Avoided in principle |
| NSAIDs | Avoided generally; contraindicated from 20 weeks (oligohydramnios) and absolutely from 30 weeks (premature ductus closure) (Chapter 66) |
| Alcohol | Fetal alcohol spectrum disorder. No safe level (Chapter 61) |
| Retinol (high-dose vitamin A), liver | Malformations (Chapter 15) |
4. What is generally considered safe, and the bigger error
Reasonably safe and commonly used: paracetamol (lowest dose, shortest time), most penicillins and cephalosporins, erythromycin, most inhaled asthma treatments, levothyroxine, insulin, many antihypertensives (labetalol, nifedipine, methyldopa), several antidepressants including sertraline, and folic acid, iron, and vitamin D.
The commonest and most consequential error in pregnancy prescribing is stopping necessary medication.
- Epilepsy: uncontrolled seizures risk maternal death (SUDEP) and fetal hypoxia. Stopping antiepileptics abruptly is dangerous.
- Asthma: poorly controlled asthma is associated with pre-eclampsia, prematurity, and low birth weight. Inhaled steroids should be continued.
- Depression: untreated perinatal depression is associated with poor outcomes for both, and with postnatal depression and suicide, which is a leading cause of maternal death in the year after birth.
- Thyroid disease, diabetes, hypertension, autoimmune disease: all carry risks from undertreatment.
Do not stop a prescribed medicine because you have discovered you are pregnant. Contact your prescriber the same week.
Supplements in pregnancy: folic acid 400 µg (5 mg for higher-risk groups) from preconception to 12 weeks, vitamin D 10 µg throughout, iodine adequacy, iron if deficient, and avoid vitamin A supplements and liver.
Pregnancy sickness: ginger and vitamin B6 first; then doxylamine plus B6, cyclizine, or promethazine; ondansetron is used, with a small first-trimester caution. Hyperemesis gravidarum is a different and serious condition requiring treatment, not endurance.
5. Breastfeeding
Most drugs pass into milk in small amounts, and the practically relevant question is the relative infant dose: under 10 percent of the maternal weight-adjusted dose is generally considered acceptable.
Usually compatible: paracetamol, ibuprofen, most penicillins, most SSRIs (sertraline and paroxetine have the lowest milk transfer), inhaled asthma drugs, levothyroxine, insulin, most antihypertensives, and most vaccines.
Avoid or use with caution: codeine (CYP2D6 ultra-rapid metaboliser risk; Chapter 68), aspirin at analgesic doses, lithium, amiodarone, cytotoxics, radioactive isotopes, and high-dose or prolonged benzodiazepines.
Practical tips: take the dose immediately after a feed to maximise the interval before the next one; prefer drugs with short half-lives; and consult a specialist medicines-in-lactation service (such as the UK Drugs in Lactation Advisory Service or LactMed) rather than relying on the package leaflet, which is usually written defensively and often says "avoid" where the evidence supports use.
Stopping breastfeeding unnecessarily because of a medicine is itself a harm, and it is common.
Older age
6. What changes physiologically
| Change | Consequence |
|---|---|
| Kidney function declines (roughly 1 percent a year after 40) | Renally cleared drugs accumulate. eGFR-based dosing is essential |
| Liver mass and blood flow fall | Reduced first-pass metabolism; higher levels of some oral drugs |
| Body fat increases, water decreases | Fat-soluble drugs (diazepam) have longer effects; water-soluble drugs are less diluted |
| Reduced albumin | More free drug |
| Increased blood-brain barrier permeability, more sensitive receptors | Greater sensitivity to sedatives, opioids, and anticholinergics |
| Impaired baroreceptor reflex | Postural hypotension and falls with blood pressure drugs |
| Reduced thirst and total body water | Faster dehydration, worsening drug accumulation |
Serum creatinine is misleading in older people, because reduced muscle mass produces a normal creatinine despite substantially reduced kidney function. eGFR is the number to use.
7. Polypharmacy
Taking five or more regular medicines is the usual definition, and it is common: in many countries, a majority of people over 75 meet it.
Some polypharmacy is entirely appropriate. Heart failure alone requires four drug classes (Chapter 75). The problem is inappropriate polypharmacy.
The risks compound:
- Interaction risk rises sharply with each additional drug.
- Adverse drug reactions cause an estimated 5 to 10 percent of hospital admissions in older people, and a large share are considered preventable.
- Adherence falls as regimens get more complex.
The prescribing cascade is the mechanism that generates much of it, and it is worth being able to recognise:
A drug causes a side effect. The side effect is interpreted as a new condition. A second drug is prescribed to treat it. That drug causes a further side effect, and so on.
Classic examples: a calcium channel blocker causes ankle swelling, which is treated with a diuretic, which causes incontinence, which is treated with an anticholinergic, which causes confusion and falls. Or metoclopramide causes parkinsonian symptoms, which are treated with levodopa.
The question to ask about any new symptom in someone on multiple medicines is: could this be the drugs?
8. Drugs that cause the most trouble in older people
The Beers Criteria (US) and STOPP/START (Europe) are formal lists of potentially inappropriate medicines in older adults. The recurring offenders:
| Drug class | Problem |
|---|---|
| Anticholinergics (first-generation antihistamines, oxybutynin, tricyclics, some antipsychotics) | Confusion, falls, urinary retention, constipation, dry mouth. Cumulative anticholinergic burden is associated with dementia risk |
| Benzodiazepines and Z-drugs | Falls, hip fractures, confusion, road accidents |
| Opioids | Falls, confusion, constipation |
| NSAIDs | GI bleeding, kidney injury, heart failure, hypertension |
| Antipsychotics in dementia | Increased mortality and stroke |
| Sulfonylureas | Prolonged hypoglycaemia |
| Digoxin at higher doses | Toxicity with declining renal function |
| PPIs long-term | Reviewed in Chapter 71 |
Falls are the outcome that matters most, because a hip fracture in an older person carries substantial mortality and loss of independence. "Falls risk-increasing drugs" are a recognised category, and medication review is a standard part of falls assessment.
9. Deprescribing
A deliberate, supervised process of reducing or stopping medicines where harms outweigh benefits. It is a clinical skill, not neglect.
Questions worth asking at a medication review:
- Is there still an indication? Some drugs were started for a temporary problem years ago.
- Is it working? Can we tell?
- Do the benefits still exceed the risks at this age and this kidney function?
- What is the number needed to treat, over what time? A statin's benefit accrues over years; a person with limited life expectancy may not reach it.
- Is any drug treating another drug's side effect?
- Could the dose be lower?
- What matters to the patient? Preventing a stroke in ten years and being able to get to the toilet without falling are different goals, and the second is often more important.
Some drugs must be tapered, not stopped: beta blockers, steroids, benzodiazepines, antidepressants, opioids, PPIs, and antiepileptics.
A structured medication review with a pharmacist or GP is a genuine intervention that improves outcomes, and many health systems offer them annually to people over 65 or on multiple medicines. It is worth asking for by name.
10. Practical points
- Bring everything to every appointment, in a bag: prescriptions, over-the-counter medicines, supplements, eye drops, creams, and inhalers. What is on the record and what is in the cupboard routinely differ.
- Use one pharmacy so interaction checking is complete.
- Ask for a dosette box or blister pack if the regimen is complex, and ask whether any doses can be consolidated.
- Simplify: once-daily formulations, combination tablets, and dropping anything unnecessary.
- Sick day rules: know which drugs to hold during a dehydrating illness (metformin, ACE inhibitors, ARBs, diuretics, NSAIDs, SGLT2 inhibitors) and which need increasing (steroids). Ask for a written list.
- Report new symptoms as possible drug effects, especially confusion, dizziness, falls, and incontinence.
- Do not stop anything alone. Ask for a review.
The bottom line
- In pregnancy, the commonest harmful error is stopping treatment. Untreated epilepsy, asthma, depression, thyroid disease, and diabetes carry real risks. Contact your prescriber; do not stop unilaterally.
- The highest-risk window for malformation is weeks 3 to 8, often before pregnancy is known, which is why preconception planning matters more than reaction.
- Valproate and isotretinoin are the drugs with the most serious teratogenic profiles and the strictest prevention programmes. NSAIDs are out from 20 weeks and absolutely from 30.
- Most drugs are compatible with breastfeeding, and unnecessary weaning is itself a harm. Use a specialist lactation medicines service rather than the leaflet.
- In older age, the commonest harmful error is continuing treatment. Kidney function falls, sensitivity rises, and each added drug raises the risk of falls, confusion, and admission.
- Learn to recognise the prescribing cascade, and ask of any new symptom: could this be the drugs?
- Ask for a structured medication review. Deprescribing is a skill, not neglect.
Sources and notes
Doses, cautions, and interactions follow the British National Formulary and the electronic Medicines Compendium; US figures follow FDA labelling. Physiological changes in pregnancy and their pharmacokinetic consequences follow Anderson's reviews and the BNF. Levothyroxine dose increase in pregnancy follows American Thyroid Association guidance. Teratogen listings follow UKTIS and the FDA's pregnancy labelling. Valproate's 10 percent malformation and 30 to 40 percent neurodevelopmental risk figures, and the Pregnancy Prevention Programme, follow the MHRA's 2018 and 2023 regulatory actions and Bromley et al.'s Cochrane review. NSAID restriction from 20 weeks follows the FDA's 2020 safety communication. Isotretinoin follows the MHRA pregnancy prevention programme. The harm of stopping necessary treatment follows MBRRACE-UK maternal mortality reports, which repeatedly identify untreated epilepsy and mental illness among leading causes. Breastfeeding compatibility and relative infant dose follow LactMed and the UK Drugs in Lactation Advisory Service. Codeine in breastfeeding follows the Health Canada and FDA restrictions after the 2006 infant death report. Age-related pharmacokinetic change and the unreliability of creatinine in older people follow Mangoni and Jackson, British Journal of Clinical Pharmacology, 2004. Polypharmacy, adverse drug reaction admissions, and preventability follow Pirmohamed et al., BMJ, 2004. Potentially inappropriate medicines follow the AGS Beers Criteria and STOPP/START version 3. The prescribing cascade concept follows Rochon and Gurwitz, BMJ, 1997. Anticholinergic burden and dementia follows Coupland et al., JAMA Internal Medicine, 2019. Falls risk-increasing drugs follow NICE falls guidance. Deprescribing frameworks follow Scott et al., JAMA Internal Medicine, 2015.
Open questions. Whether deprescribing improves hard outcomes, as opposed to reducing drug burden, has limited trial evidence. Pregnant and breastfeeding women remain systematically excluded from trials, so most of this evidence is observational by necessity.
👉 Next: drug interactions.