By Dr. Marcus Judge

Next Lesson - Safe Prescribing

Pharmacology

Contents

  1. Introduction
  2. The Thalidomide Tragedy
  3. Why Pharmacovigilance Matters
  4. The Yellow Card Scheme
  5. Pharmacogenetics
  6. Stratified and Personalised Medicine
  7. Summary
  8. Quiz

Abstract

  • Pharmacovigilance is the science of identifying, assessing and preventing adverse drug reactions (ADRs) once a medicine is in use. The UK system is run by the MHRA (Medicines and Healthcare products Regulatory Agency) through the Yellow Card Scheme.
  • Pre-marketing trials may miss rare or delayed adverse effects because of small numbers, short duration, narrow patient selection, and exclusion of people who have several medical conditions at once (comorbidities). The thalidomide tragedy of the 1950s and 1960s is the historical case that drove modern pharmacovigilance regulation.
  • Pharmacogenetics studies how individual genetic variation influences drug handling and response. Common examples include CYP polymorphisms (common variations in the genes for the enzymes that break down drugs; codeine, warfarin, clopidogrel), HLA-linked hypersensitivity (severe immune reactions linked to variants in the human leucocyte antigen genes; abacavir, carbamazepine), and TPMT deficiency (low activity of the enzyme thiopurine methyltransferase; azathioprine).
  • Recognising the genetic and population-level basis of drug variability is the foundation of stratified medicine (matching treatment to groups of patients who share a feature that predicts benefit or harm), the modern alternative to "one dose fits all" prescribing.

Core

Introduction

Once a drug has been licensed, the work of evaluating it has only just begun. Clinical trials study a relatively small number of selected patients for a short time. Some adverse drug reactions, especially rare, serious or delayed ones, only emerge once a drug is used by a larger and more varied group of people over a longer period. Pharmacovigilance is the system by which those reactions are identified, assessed and acted upon.

The World Health Organization (WHO) defines pharmacovigilance as "the science and activities relating to the detection, assessment, understanding and prevention of adverse effects or any other medicine/vaccine related problem". In the UK this is the responsibility of the Medicines and Healthcare products Regulatory Agency (MHRA) and, in the European Union, of the European Medicines Agency (EMA).

The Thalidomide Tragedy

Modern pharmacovigilance regulation was largely prompted by a single drug. Thalidomide was marketed from 1957 as a sedative and anti-emetic (a medicine that reduces nausea and vomiting), including for the treatment of morning sickness in pregnancy. It was sold in more than 40 countries under different names. In West Germany it was sold over the counter as Contergan. Between 1957 and 1962, an estimated 10,000 children were born worldwide with severe limb-reduction defects (phocomelia: from the Greek for "seal limb"; arms or legs absent or reduced to flipper-like stumps). Cardiac (heart), gastrointestinal (gut), ocular (eye) and auditory (ear) malformations also occurred.

The link between thalidomide and these defects was identified by Dr Widukind Lenz and Dr William McBride and reported in 1961. Thalidomide was withdrawn from most markets by the end of 1962. The episode prompted three lasting changes in regulation:

  • Mandatory teratogenicity testing (tests for whether a drug can cause birth defects) in animals before human trials.
  • Phased clinical trial requirements: Phase I to III trials before licensing (first in a small number of people, then in larger groups of people who are ill). For some medicines these are followed by Phase IV studies of safety and effectiveness once the medicine is licensed and in use.
  • The creation of spontaneous reporting systems, in which health professionals and patients voluntarily report suspected adverse events: the UK Yellow Card Scheme was established in 1964 in direct response to the thalidomide disaster.

Why Pharmacovigilance Matters

A study published in 2004 followed 18,820 patients (aged over 16) who were admitted to two Merseyside hospitals over six months. Adverse drug reactions were related to 6.5% of the admissions (1,225 admissions). In 80% of those 1,225 admissions, the reaction directly caused the admission. This is a historical two-hospital finding, not a current estimate for all UK admissions. The study illustrates the burden that ongoing medicine-safety surveillance addresses.

Limitations of Pre-Marketing Trials

Pre-marketing clinical trials, however well-conducted, cannot detect every adverse effect. Their structural limitations include:

  • Small numbers. Most Phase III trials enrol a few thousand patients at most. An ADR with an incidence of 1 in 10,000 may be missed in a trial involving a few thousand exposed patients.
  • Selected populations. Trial populations may under-represent groups such as older people with multiple conditions or people taking several medicines, limiting how well the findings apply to them.
  • Short duration. Phase III trials often last a year or more. That is shorter than the years over which chronic and delayed reactions can emerge.
  • Specialist settings. Trials are run in research centres with intensive follow-up, not in the busy clinics where the drug will eventually be used.
  • Concomitant medications excluded (other medicines taken at the same time). Drug-drug interactions can be missed if polypharmacy (taking several medicines together) is excluded from the protocol.

For all these reasons, the safety profile of a new medicine is genuinely incomplete at the moment it is licensed. Pharmacovigilance is the means by which that profile is built up over time.

Mechanisms of Adverse Drug Reactions

The classification of ADRs into types A, B, C, D, E and F (Augmented, which is dose-related; Bizarre, which is not dose-related; Chronic; Delayed; End of use, meaning withdrawal; and Failure of therapy) is covered in detail in Pharmacodynamics. Mechanistically, many ADRs can be explained by one of four mechanisms (this is a different scheme from types A to F):

  • Exaggerated pharmacological response at the intended target. Bleeding from warfarin; hypoglycaemia (low blood glucose) from insulin; bradycardia (slow heart rate) from a beta-blocker (a drug that slows the heart by blocking beta receptors).
  • Desired pharmacological action at an unintended site. Dry mouth from atropine; headache from glyceryl trinitrate; bronchospasm (narrowing of the airways) from a non-selective beta-blocker; cough from angiotensin-converting enzyme (ACE) inhibitors. ACE normally breaks down bradykinin (a signalling peptide that can provoke cough). The cough is thought to occur because blocking ACE lets bradykinin build up in the airways.
  • Off-target pharmacological effects. QT-interval prolongation (a delay in the heart's electrical recovery after each beat, seen on the ECG) from many drugs (e.g. macrolide antibiotics such as erythromycin, and antipsychotics, the medicines used to treat psychosis).
  • Immune-mediated reactions. Anaphylaxis to penicillin; Stevens-Johnson syndrome (a severe blistering reaction of the skin and mucous membranes) from sulphonamides (a group of antibacterial drugs) or carbamazepine.

The Yellow Card Scheme

The Yellow Card Scheme is the UK system for spontaneous reporting of suspected adverse drug reactions, introduced in 1964 under the Committee on the Safety of Drugs and now run by the MHRA. Reports can be submitted online, by post, by phone or via the Yellow Card mobile app. Reporter and patient details are kept confidential (not anonymous: identifiers are needed so the MHRA can follow up where required).

Who reports? Until 1997, only doctors and dentists could report. Today, reports are accepted from health professionals, patients, parents and carers, and patients have been reporting since 2005. The more people who can report, the more reactions are captured.

What gets reported?

  • All suspected ADRs to drugs marked with the inverted black triangle (▼) in the BNF (British National Formulary, the UK's standard prescribing reference); new medicines and vaccines under additional monitoring. The symbol does not mean that a medicine is unsafe. The MHRA keeps the black-triangle status until it is satisfied that the safety profile is well-established (usually 5 years, but this can be extended if required).
  • Serious suspected ADRs to any other drug, regardless of how long it has been licensed. "Serious" means resulting in death, hospital admission, prolonged hospitalisation, congenital abnormality, life-threatening illness, or significant disability.
  • Suspected ADRs to vaccines and herbal or homeopathic medicines, and problems with medical devices, defective or fake medicines and e-cigarettes. These are all reported through the Yellow Card Scheme, although medical devices have a separate reporting route in Scotland.

Strengths of the Yellow Card Scheme:

  • Simple, free, and applies to all licensed drugs throughout their life cycle.
  • Accessible to all healthcare professionals and to patients.
  • Generates the early signals (information suggesting that a medicine may cause a harmful effect and needs further investigation) that lead to formal regulatory investigation.

Weaknesses:

  • Under-reporting: the single biggest limitation. A 2006 systematic review (a structured pooling of earlier studies) of 37 studies found a median under-reporting rate of 94% for ADRs. For specific serious or severe ADR-drug combinations, the median was lower but still high, at 85%.
  • Reports are unquantifiable in isolation: the denominator (how many people were exposed) is not known.
  • Reports may be duplicated, biased, or based on weak association.
  • The system is vulnerable to publicity: reports surge after media coverage of a suspected reaction, distorting the apparent signal. The 1995 "pill scare" over third-generation combined oral contraceptives (pills containing two hormones, an oestrogen and a progestogen; third-generation pills contained newer progestogens) is the classic example.

For all these reasons the Yellow Card Scheme is best understood as a signal-generating system rather than a definitive one. Once a signal emerges, the MHRA can commission formal pharmacoepidemiological studies (studies of how medicines affect large groups of people), request data from manufacturers, and ultimately restrict, label or withdraw a drug.

Pharmacogenetics

Pharmacogenetics is the study of how individual genetic variation influences drug response. The closely related term pharmacogenomics covers the same territory at a genome-wide level. The clinical importance is that prescribing decisions made at the population level may be wrong for a particular individual. In a 2001 systematic review of 27 drugs often cited in ADR studies, 59% were metabolised by at least one enzyme with a variant known to cause poor metabolism, compared with 7% to 22% of randomly selected drugs. This suggests that genetic information could help to prevent some adverse reactions.

Response to common drugs varies widely between patients. Variants in genes for drug-metabolising enzymes, transporters, receptors and other drug targets have been linked to differences in the effectiveness and toxicity of many medicines. Genes are only one cause. For statins (cholesterol-lowering drugs), for example, differences in response have several causes (they are multifactorial), including whether patients take their tablets.

Pharmacokinetic Polymorphisms

A polymorphism is a common variation in the DNA sequence of a gene. Pharmacokinetic polymorphisms change what the body does to a drug (how it is absorbed, distributed, metabolised and excreted). The most clinically relevant variation of this kind affects drug-metabolising enzymes: principally the cytochrome P450 (CYP) family. The CYP system is covered in Pharmacokinetics; pharmacogenetic variants produce several phenotypes (the four main ones are listed below), and a patient's phenotype can mean a different dose or a different drug is needed:

  • Poor metabolisers (PMs): reduced or absent enzyme activity. Higher plasma levels and increased toxicity for parent drugs (drugs that are active as given); reduced effect for pro-drugs (inactive drugs that the enzyme must convert into an active form).
  • Intermediate metabolisers (IMs): partly reduced enzyme activity.
  • Normal metabolisers (NMs), also called extensive metabolisers (EMs) in older sources: normal enzyme activity.
  • Ultra-rapid metabolisers (UMs): gene duplication, increased enzyme activity. Reduced effect for parent drugs; toxic levels for pro-drugs.

Three CYP-mediated examples are routinely taught at pre-clinical level:

  • CYP2D6 and codeine. Codeine is a pro-drug that requires CYP2D6 to be converted to morphine. Ultra-rapid metabolisers can develop morphine toxicity from standard doses. Codeine is contraindicated in people known to be CYP2D6 ultra-rapid metabolisers at any age, and in children under 12. In children aged 12 to 18, it should only be used for acute moderate pain that cannot be relieved by other painkillers used alone, such as paracetamol or ibuprofen. It is also contraindicated in anyone under 18 having a tonsillectomy or adenoidectomy (removal of the tonsils or adenoids, lymphoid tissue at the back of the throat and nose) for obstructive sleep apnoea (repeated blocking of the airway during sleep). This is because of reports of serious or fatal respiratory depression (dangerously slow or shallow breathing). The BNF advises that codeine should not be used in breastfeeding mothers, because it can pass to the baby through breast milk.
  • CYP2C9 and warfarin. Variation in CYP2C9 and in VKORC1 (vitamin K epoxide reductase complex subunit 1, the gene for the enzyme complex that warfarin blocks) can affect warfarin dose requirements. Dosing still relies on INR monitoring (the international normalised ratio, a blood test of clotting).
  • CYP2C19 and clopidogrel. Clopidogrel is a pro-drug. Poor CYP2C19 metabolisers form less of the active form of the drug (its active metabolite) at recommended doses and have a smaller effect on platelet function. NICE (2024) recommends CYP2C19 genotype testing to assess whether clopidogrel is a suitable antiplatelet drug (a drug that stops platelets sticking together) for people who have just had an ischaemic stroke or a transient ischaemic attack (TIA, a brief episode of stroke symptoms). In other settings, treatment selection should follow the clinical indication and relevant guidance.

Beyond CYP enzymes, thiopurine methyltransferase (TPMT) is an important example. Azathioprine (a medicine that suppresses the immune system) is converted to mercaptopurine, which enters competing metabolic pathways. In one pathway, TPMT adds a methyl group to the drug (methylation). This limits the formation of active thioguanine nucleotides (the active forms of the drug, which act on dividing cells). Reduced TPMT activity increases the concentration of these nucleotides. It also raises the risk of severe myelosuppression (reduced production of blood cells by the bone marrow). The azathioprine SmPC (summary of product characteristics, the official prescribing information) describes about 0.3% (1 in 300) of patients as having two non-functional TPMT alleles, with little or no enzyme activity. About 10% have one non-functional allele and low or intermediate activity. The BNF says that the manufacturer advises considering measuring TPMT activity before starting azathioprine. The SmPC also says that testing for NUDT15 variants (variants in another gene linked to severe azathioprine toxicity) may be considered. Blood counts still need close monitoring, because TPMT tests have not been shown to identify all patients at risk of severe toxicity.

Likewise, dihydropyrimidine dehydrogenase (DPD) deficiency is a partial or complete lack of the enzyme that breaks down fluorouracil, most often caused by inherited gene variants. It increases the risk of severe, sometimes fatal, toxicity with intravenous 5-fluorouracil, capecitabine and tegafur (anti-cancer medicines known as fluoropyrimidines). The MHRA recommends testing all patients for DPD deficiency before starting any of these three medicines. Routine pretreatment DPD testing is not required for topical fluorouracil, because very little of the drug is absorbed into the bloodstream.

Pharmacodynamic Polymorphisms

Polymorphisms in drug targets (the proteins a drug acts on), rather than metabolising enzymes, can also alter response. These are called pharmacodynamic polymorphisms: they change what the drug does to the body. VKORC1 variants alter warfarin sensitivity at the drug's target. The target is an enzyme complex, not a receptor. Variants in the β2-adrenergic receptor (the receptor that salbutamol acts on in the airways) have been studied for effects on the bronchodilator response (widening of the airways) to salbutamol. A meta-analysis (a pooled analysis of several studies) found only weak evidence for an association. Variants in the 5-HT (serotonin) transporter have been associated with response to selective serotonin reuptake inhibitors (SSRIs, a group of antidepressants).

HLA-Linked Hypersensitivity

Several severe reactions that are immune-mediated (caused by the immune system) and not dose-related (Type B) are strongly linked to specific human leucocyte antigen (HLA) alleles. HLA genes code for cell-surface proteins that help the immune system recognise what is foreign. Genotyping (testing which alleles a person carries) before prescribing is advised for some drugs and in some at-risk groups, but the strength of the advice differs between the drugs below:

  • HLA-B*5701 and abacavir (an HIV medicine): in the PREDICT-1 trial, a hypersensitivity reaction was diagnosed clinically in 7.8% of patients treated without screening and in 3.4% of those screened first. Reactions confirmed immunologically (by a skin patch test) occurred in 2.7% without screening and in none with screening. No patient with a negative test had a confirmed reaction, but only 47.9% of those with a positive test did. The BNF advises testing for HLA-B*5701 before starting abacavir, or before restarting it if the status is not known.
  • HLA-B*1502 and carbamazepine (an anti-seizure medicine): strongly associated with Stevens-Johnson syndrome and toxic epidermal necrolysis (a severe skin reaction with peeling over large areas of the body), particularly in patients of Han Chinese, Thai or other South-East Asian ancestry, where the allele is far more prevalent. The BNF advises testing for HLA-B*1502 in people of Han Chinese or Thai origin, and considering testing in other at-risk Asian populations.
  • HLA-B*5801 and allopurinol (a medicine used to prevent gout, a painful joint condition caused by uric acid crystals): severe cutaneous (skin) reactions, again with strong ethnic variation in allele frequency. NICE Clinical Knowledge Summaries advise not prescribing allopurinol to people known to have the HLA-B*5801 allele. They also advise that screening should be considered in ethnic groups where the allele is known to be common (for example Han Chinese, Thai and Korean populations).

Pharmacogenetics in Oncology

Cancer treatment is one of the fields where genetic testing already guides drug choice in routine practice. NICE recommends some cancer drugs only for people whose tumour has (or lacks) a particular genetic change. A few standard examples:

  • HER2 status (HER2 is a growth-factor receptor found in large amounts on some breast cancer cells) in breast cancer determines whether a patient benefits from trastuzumab (a monoclonal antibody, which is a laboratory-made antibody, that blocks HER2).
  • BRCA1 / BRCA2 mutations (changes in two genes that normally help to repair damaged DNA) identify patients who benefit from PARP inhibitors (drugs that block PARP, poly ADP-ribose polymerase, another DNA-repair enzyme), such as olaparib.
  • EGFR mutations (changes in the gene for the epidermal growth factor receptor) in non-small-cell lung cancer (the commonest group of lung cancers) predict response to EGFR tyrosine kinase inhibitors (drugs that block an overactive enzyme driving cell growth: gefitinib, erlotinib, osimertinib).
  • BCR-ABL translocation (a swap of DNA between chromosomes 9 and 22) in chronic myeloid leukaemia creates the BCR-ABL fusion gene. This makes an overactive tyrosine kinase (an enzyme that drives cell growth), which is the target of imatinib.
  • RAS wild-type status (the normal, unmutated form of the RAS genes, which include KRAS) is required for benefit from anti-EGFR antibodies (antibodies that block the epidermal growth factor receptor: cetuximab, panitumumab) in colorectal cancer. NICE recommends cetuximab and panitumumab, in combination with chemotherapy (anti-cancer drugs), as options for previously untreated RAS wild-type metastatic colorectal cancer (cancer that has spread to other parts of the body).

Most of these markers are not germline pharmacogenetics (germline variants are inherited and present in every cell). They reflect somatic changes, which are acquired in the cancer cells only. BRCA1 and BRCA2 are the exception: changes in these genes can be inherited (germline) or found only in the tumour (somatic). In every case the underlying principle is the same: matching the drug to a molecular target.

Population-Level Differences in Drug Response

Allele frequencies vary between populations, and so do average drug responses. Two examples are worth knowing for pre-clinical exams:

  • ACE inhibitors tend to be less effective in people of Black African or African-Caribbean family origin, who tend to have low-renin hypertension (high blood pressure with low blood levels of renin, a kidney enzyme that raises blood pressure). For adults of this family origin who do not have type 2 diabetes, NICE guideline NG136 advises offering a calcium channel blocker (a drug that relaxes blood vessel walls) as the first drug. NICE also advises considering an angiotensin II receptor blocker (ARB) in preference to an ACE inhibitor for this group. The mechanism is not strictly genetic but is a population-level pharmacological difference covered in Hypertension and Heart Failure.
  • Aldehyde dehydrogenase 2 (ALDH2) deficiency affects roughly a third of East Asian people (about 36% in one 2009 review of Japanese, Chinese and Korean populations). Affected individuals accumulate acetaldehyde (a toxic breakdown product of alcohol) after drinking alcohol, producing facial flushing, tachycardia (a fast heart rate) and nausea. This is called the alcohol flush reaction. It is sometimes called "Asian flush". The same enzyme is targeted by disulfiram (a drug that causes unpleasant effects if alcohol is drunk) in alcohol-use disorder.

Group averages describe populations. They do not tell you how one patient will respond. An individual's genotype (their own gene variants) matters more than the average for their ethnic background. Assuming that a population average applies to one patient is itself a form of bias.

Stratified and Personalised Medicine

Stratified medicine matches treatment to groups sharing characteristics that predict benefit or harm. Pharmacogenetic variants are one possible basis for grouping patients. Personalised medicine uses relevant information about an individual to tailor care, including targeted genetic or protein tests where appropriate. Whole-genome sequencing (reading a person's entire DNA sequence) is not required for every patient or prescription.

Four developments are driving the move away from "one dose fits all" prescribing:

  • Developments in genetic testing, including targeted tests chosen for the clinical question.
  • Growth in evidence linking specific variants to specific drug effects.
  • Regulatory acceptance: the MHRA, EMA and the US Food and Drug Administration (FDA) increasingly include pharmacogenetic information in product labelling. For a few medicines, such as abacavir, the labelling advises testing before prescribing.
  • The NHS Genomic Medicine Service in England, launched in 2018, offers whole-genome sequencing within the NHS for selected indications. The National Genomic Test Directory sets out which genomic tests are available to which patients.

Several barriers remain: the evidence for many proposed pharmacogenetic tests is incomplete, turn-around times can be slower than the clinical decision, and the ethical questions about genetic data are unresolved. The pre-clinical student is not expected to know these barriers in detail; only to recognise that, for some medicines, "one dose fits all" prescribing is giving way to genetically informed prescribing.

Summary

  • Pharmacovigilance is the science of detecting and managing ADRs once a drug is in clinical use.
  • Pre-marketing trials can miss rare, delayed, or population-specific reactions because of small numbers, short duration and selection bias.
  • The Yellow Card Scheme is the UK signal-generating system, run by the MHRA. It is open to all healthcare professionals and patients.
  • Thalidomide is the historical event that led to the modern regulatory framework. Black triangle medicines are under additional monitoring, usually for 5 years after licensing, and all suspected ADRs to them should be reported.
  • Pharmacogenetics explains some of the variability in drug response, alongside other causes. Examples include CYP2D6 (codeine), CYP2C9 (warfarin), CYP2C19 (clopidogrel), TPMT (azathioprine), DPD (5-fluorouracil), and HLA-linked hypersensitivity (abacavir, carbamazepine).
  • The translation of these findings into stratified medicine is a growing part of modern UK prescribing.

Reviewed by: Dr. Marcus Judge

Quiz

Preview the Pharmacovigilance and Pharmacogenetics quiz

233