Adverse Drug Reaction

An adverse drug reaction (ADR) is an unintended and harmful response to a medication that occurs at doses normally used for prophylaxis, diagnosis, or treatment. It is distinct from adverse events caused by medication errors, intentional overdose, or drug abuse, and implies at least a reasonable possibility that the drug caused it.^[2]

This normal dose definition is the traditional one, set out by the World Health Organization in 1972, and it remains common in clinical pharmacology. Regulatory pharmacovigilance has since broadened it in some jurisdictions. In the European Union, for example, legislation applying since 2012 defines an adverse reaction as any noxious and unintended response to a medicinal product, which includes suspected reactions arising from medication errors, overdose, misuse, abuse, and off label use, and EU marketing authorization holders must report them to the EudraVigilance database: suspected serious reactions (in the EU or elsewhere) within 15 days of gaining knowledge, and non-serious reactions occurring in the EU within 90 days (Directive 2001/83/EC, Article 107).^[7] Studies and safety reports should therefore state which definition they use.

Adverse drug reactions are a major cause of hospitalization, morbidity, and mortality worldwide. A systematic review of 25 prospective studies found that a median of 5.3% of hospital admissions (interquartile range 2.7% to 9.0%) were associated with ADRs, with higher rates in older adults.^[6] Pharmacovigilance, the science of monitoring and evaluating drug safety, depends on accurate ADR identification and reporting. For clinical researchers, distinguishing ADRs from other adverse events is essential for correctly characterizing a medication's risk profile. Regulatory agencies use ADR data to make decisions about drug labeling, prescribing restrictions, and market withdrawal. For patients and clinicians, understanding ADRs informs treatment selection and risk management.

ADRs are classified using several frameworks. The most widely used is the Rawlins and Thompson classification, which divides reactions into Type A and Type B. Type A reactions are dose dependent, predictable based on the drug's known pharmacology, and generally common. Examples include bleeding with anticoagulants, hypoglycemia with insulin, and sedation with benzodiazepines. Type B reactions are dose independent, idiosyncratic, and unpredictable. Examples include anaphylaxis to penicillin, Stevens Johnson syndrome from certain anticonvulsants, and drug induced liver injury from agents with no known hepatotoxic mechanism. Additional categories (Types C through F) cover dose and time related reactions, delayed effects, withdrawal reactions, and treatment failure.^[3]

Causality assessment determines whether a specific drug caused the observed reaction. Several standardized tools exist for this purpose, including the Naranjo algorithm, which scores ten questions (such as temporal relationship, dechallenge and rechallenge responses, alternative explanations, and prior reports in the literature) to give a total from negative 4 to 13 that maps to doubtful, possible, probable or definite causality.^[4] The WHO Uppsala Monitoring Centre system classifies causality as certain, probable, possible, unlikely, conditional, or unassessable.

ADR detection occurs through multiple channels. Pre marketing clinical trials identify common ADRs but are often too small and too short to detect rare or delayed reactions.^[1][2] Post marketing surveillance captures ADRs through spontaneous reporting systems (such as the FDA's MedWatch and the EU's EudraVigilance), active surveillance programs, electronic health record mining, and pharmacoepidemiologic studies using large claims databases.^[1] Signal detection algorithms scan these data sources to identify unexpected patterns of ADR reporting.

ADR assessment is relevant whenever a patient experiences an unexpected harmful outcome during drug therapy and the clinician or researcher needs to determine whether the drug is responsible. It is central to pharmacovigilance, clinical trial safety monitoring, and drug regulatory processes. Researchers assessing individual case reports of suspected reactions should use a standardized causality assessment tool, and safety reports should record ADR data using established terminologies such as MedDRA (the Medical Dictionary for Regulatory Activities).

A common error is conflating adverse drug reactions with adverse events. An adverse event is any untoward medical occurrence during drug therapy, regardless of causation, whereas an ADR specifically implies a causal relationship with the drug.^[2][3] Mislabeling all adverse events as ADRs inflates the apparent toxicity of a medication, while failing to identify true ADRs understates its risks.

Another frequent mistake is underreporting. A systematic review of 37 studies found a median under-reporting rate of 94% (interquartile range 82% to 98%) of ADRs to spontaneous reporting systems.^[5] This underreporting distorts the apparent safety profile of medications and delays signal detection. Researchers and clinicians should report suspected ADRs even when causality is uncertain. Finally, neglecting to consider drug interactions as a cause of ADRs can lead to misattribution, especially in patients taking multiple medications.

A pharmacovigilance team at a regulatory agency notices a cluster of 14 reports of acute liver injury in patients taking a newly approved antifungal medication over a three month period. Using the Naranjo algorithm, they classify 9 cases as "probable" and 5 as "possible" ADRs. The signal triggers a formal safety review, which ultimately leads to a label update requiring baseline and periodic liver function monitoring for all patients prescribed the drug.