How an Adverse Event Is Defined and Reported During a Trial

Answer first: the widest term is the one that assumes nothing

The definitions live at 21 CFR 312.32(a), and they are built in layers. An adverse event means any untoward medical occurrence associated with the use of a drug in humans, whether or not considered drug related.

That final clause is the load-bearing part. The base category deliberately does not require anyone to think the drug caused anything. It is a record that something happened to a person who was taking the drug.

The next layer up adds a judgment. A suspected adverse reaction means any adverse event for which there is a reasonable possibility that the drug caused the adverse event, and for safety-reporting purposes reasonable possibility means there is evidence to suggest a causal relationship. The regulation is explicit that this is weaker than certainty: suspected adverse reaction implies a lesser degree of certainty about causality than adverse reaction, which means any adverse event caused by a drug.

So there are three rungs, not two. Something that happened. Something that may have been caused. Something that was caused. Most of what a trial collects sits on the first rung, and most public discussion of trial safety data treats it as if it sat on the third.

Serious is a defined list, not a description of how bad it felt

Under the same section, an adverse event or suspected adverse reaction is considered serious if, in the view of either the investigator or the sponsor, it results in any of the following outcomes: death, a life-threatening adverse event, inpatient hospitalization or prolongation of existing hospitalization, a persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions, or a congenital anomaly or birth defect.

That is an outcome list. It is not a measure of severity or discomfort. Something can be miserable without meeting it, and something can meet it without being distressing at the moment it occurred.

The regulation then adds a route back in for judgment. Important medical events that may not result in death, be life-threatening, or require hospitalization may be considered serious when, based on appropriate medical judgment, they may jeopardize the patient and may require medical or surgical intervention to prevent one of the listed outcomes. Its examples include allergic bronchospasm requiring intensive treatment, blood dyscrasias or convulsions that do not result in hospitalization, and the development of drug dependency or drug abuse.

Life-threatening has its own tight definition, and the tightness matters. An event is life-threatening if, in the view of either the investigator or sponsor, its occurrence places the patient at immediate risk of death. It does not include an event that, had it occurred in a more severe form, might have caused death.

Unexpected means not already in the paperwork

The fourth word is the one most often misread. Unexpected does not mean surprising in a general sense. Under 312.32(a), an event or suspected adverse reaction is unexpected if it is not listed in the investigator brochure, or is not listed at the specificity or severity that has been observed — or, where no brochure is required or available, is not consistent with the risk information described in the general investigational plan or elsewhere in the current application.

The regulation gives two worked examples of how specificity and severity work. Hepatic necrosis would be unexpected, by virtue of greater severity, if the brochure referred only to elevated hepatic enzymes or hepatitis. Cerebral thromboembolism and cerebral vasculitis would be unexpected, by virtue of greater specificity, if the brochure listed only cerebral vascular accidents.

There is a third example that catches a common assumption. Unexpected also refers to events mentioned in the brochure as occurring with a class of drugs, or as anticipated from the pharmacological properties of the drug, but not specifically mentioned as occurring with the particular drug under investigation.

In short: expected means already documented, for this drug, at this level of detail. Class-level knowledge does not make an event expected for a specific molecule.

The clocks: seven days, fifteen days, and everything else

Two deadlines run during an investigation, and they are short.

The sponsor must notify FDA and all participating investigators of potential serious risks, from clinical trials or any other source, as soon as possible but in no case later than 15 calendar days after the sponsor determines the information qualifies. Four things qualify. A suspected adverse reaction that is both serious and unexpected. Findings from other studies — epidemiological, pooled analyses, or clinical studies whether or not conducted under a federal application and whether or not conducted by the sponsor — that suggest a significant risk in humans. Findings from animal or in vitro testing that suggest a significant risk in humans, such as mutagenicity, teratogenicity, carcinogenicity, or significant organ toxicity at or near expected human exposure. And any clinically important increase in the rate of a serious suspected adverse reaction over the rate listed in the protocol or investigator brochure.

The shorter clock applies to the worst cases. The sponsor must notify FDA of any unexpected fatal or life-threatening suspected adverse reaction as soon as possible but in no case later than 7 calendar days after the sponsor's initial receipt of the information.

Each report must not stand alone. In each safety report the sponsor must identify all previous reports concerning a similar suspected adverse reaction and analyze the significance of the reaction in light of those previous reports or other relevant information. And the regulation names one kind of evidence that qualifies as suggesting causation: an aggregate analysis of specific events observed in a trial indicating those events occur more frequently in the drug treatment group than in a concurrent or historical control group.

Everything that does not meet those triggers still gets reported, more slowly. Under 21 CFR 312.33 the sponsor's annual report must include a narrative or tabular summary showing the most frequent and most serious adverse experiences by body system, a summary of all safety reports submitted during the year, a list of subjects who died during participation with the cause of death for each, and a list of subjects who dropped out in association with any adverse experience, whether or not thought to be drug related.

Who reports to whom, and one sentence about what a report is not

The chain starts at the site. Under 21 CFR 312.64, an investigator must immediately report to the sponsor any serious adverse event, whether or not considered drug related, including those already listed in the protocol or investigator brochure, and must include an assessment of whether there is a reasonable possibility that the drug caused the event. Nonserious events are recorded and reported to the sponsor according to the timetable specified in the protocol.

There is a special rule for study endpoints that is easy to misread as a loophole and is not one. Study endpoints such as mortality or major morbidity are reported to FDA as described in the protocol and ordinarily would not be reported as safety reports — because in a trial designed to count deaths, every death is the measurement. But if a serious and unexpected adverse event occurs for which there is evidence suggesting a causal relationship — the regulation's example is death from anaphylaxis — it must be reported as a serious and unexpected suspected adverse reaction even if it is a component of the study endpoint.

The sponsor's duties do not stop at forwarding. Under 21 CFR 312.56 the sponsor must monitor the progress of all investigations, review and evaluate the safety and effectiveness evidence as it comes in, and — if it determines the drug presents an unreasonable and significant risk to subjects — discontinue the investigations presenting that risk, notify FDA, all review boards and all investigators who have at any time participated, and do so as soon as possible and in no event later than five working days after making the determination.

Finally, the regulation includes a disclaimer that belongs in any honest reading of safety data. A safety report or other information submitted by a sponsor, and any release by FDA of that report, does not necessarily reflect a conclusion by the sponsor or FDA that the report constitutes an admission that the drug caused or contributed to an adverse event. A sponsor need not admit, and may deny, that it does.

That sentence is why a count of reports is not a count of harms. It is a count of things that were required to be written down.

One further point about how the numbers are assembled. FDA's draft guidance on weight-reduction drugs notes that because the number of safety events in any one trial may be limited, safety data are commonly integrated across studies, and recommends that integrated analyses generally be stratified by trial to prevent confounding when pooling trials with different randomization ratios and populations with different risks. It also lists differences that matter across studies, including different treatment durations and different methods of adverse event ascertainment — questionnaire versus general inquiry, different frequencies of querying, or substantial differences in reporting patterns. How a number was collected shapes the number.