Definition and types of AEFI
The World Health Organization (WHO) defines an AEFI as a medical incident that takes place after an immunisation, causes concern, and believed to be caused by immunisation. There are three key aspects to this definition.
Firstly, it is deliberately loose to encourage reporting of events. This is because it does not restrict the type of event (other than being a health consequence) nor limit the time window after immunisation, nor attempt to determine whether the immunisation may have been responsible, i.e. it is events, not reactions that are reported.
Secondly, it describes a belief about causality that requires investigation. The belief that immunisation was responsible may turn out to be correct, incorrect, or impossible to assess.
Thirdly, describing an event as an AEFI, does not and must not imply causality. (There is only a belief of causality). Only after investigation, it may be possible to assign causality.
Within this framework an AEFI can be either:
- Caused by the vaccine or immunisation process (causally associated) or
- a coincidental event that, by chance, happened after immunisation (temporally associated).
There are three types of AEFIs caused by immunisation, these are caused by:
- The inherent properties of the vaccine (vaccine reaction)
- An error in the immunisation process (programme error)
- Injection-related reactions arising from anxiety about or pain of the injection
Sometimes it is not possible to state with any degree of certainty what is the cause of a particular AEFI. This is particularly true at the level of the individual if there are no studies that have helped assess causality. From the above, there are five types of AEFI, described in table below.
|Vaccine reaction||Event caused or precipitated by the vaccine when given correctly, caused by the inherent properties of the vaccine.|
|Programme error||Event caused by an error in vaccine preparation, handling, or administration.|
|Injection reaction||Event from anxiety about, or pain from, the injection itself rather than the vaccine.|
|Coincidental||Event that happens after immunisation but not caused by the vaccine - a chance association.|
|Unknown||The event's cause cannot be determined.|
Source: Adapted from “Immunisation Safety Surveillance: guidelines for managers of immunisation programmes on reporting and investigating adverse events following immunisation. Manila: World Health Organization, 1999. ”
These reactions are caused by a constituent of the vaccine. In some cases this will be the vaccine antigen (the substance that generates immunity), and is thus a side effect of the immunological process of generating immunity. In other cases it will be caused by other vaccine constituents (e.g. preservatives, stabilisers, antibiotics, or residual substance from the manufacturing process) or the adjuvant that is added to boost the vaccine’s immunogenicity. Vaccine reactions can be categorised into two types:
- Common, usually minor and self-limiting
- Rare and more serious
The reason that there are only these two types of vaccine reaction is that any vaccine that caused a serious reaction that was even relatively common would either not be licensed for use or withdrawn when the frequency of the serious adverse event is identified. An example of this was the withdrawal of the first rotavirus vaccine because it was found to cause intussusception in about 1 in 10,000 vaccinees. Similarly, oral polio vaccine (OPV) is no longer used in most industrialised countries because it carries a risk of about 1 in 2 - 3 million of causing polio. However, the current use of OPV in developing countries and its previous use in other countries, reflects the fact that the vaccine risks need to be balanced against the risk of the disease and the cost of the alternative (inactivated vaccine in the case of OPV).
Common vaccine reactions (reactogenicity)
The common reactions are caused by the body’s response to the vaccine antigen(s) and the other vaccine constituents. In general they occur within 24-48 hours of vaccination and are self-limiting. However reactions following live vaccines (e.g. MMR) may be delayed and resemble a mild version of the disease.
|Symptom or sign||Maximum difference in rate*||Peak frequency (days after vaccination)|
|Local erythema (>2 cm)||0.8||0.1-1.4||2|
|Other local reaction||0.4||0-1.4||2|
|Mild fever (<38.5°C rectal)||2.7||0-6.1||10|
|Moderate fever (38.6-39.5°C)||2.9||1.6-4.3||9|
|High fever (>39.5°C)||1.4||0.7-2.1||10|
|Willingness to stay in bed||1.4||0.5-2.3||11|
|Cough and/or coryza||-1.5¹||-4.6-1.6||9|
|Nausea and/or vomiting||-0.8¹||-1.6-0||7-8|
*between MMR group and placebo group
¹more in placebo-injected children.
Adapted from: Symptoms and signs caused by MMR vaccination and day of peak occurrence. Reference: Peltola H, Heinonen OP. Frequency of true adverse reactions to measles-mumps-rubella vaccine. A double-blind placebo-controlled trial in twins. Lancet 1986;1(8487):939-42
Common reactions are usually a local reaction at the site of injection and/or systemic symptoms. For the local reaction there is pain, swelling, lump and/or redness at the injection site, which can vary in severity and usually last a day or so but sometimes longer. The systemic symptoms can include fever and a range of non-specific symptoms (e.g. irritability, malaise, ‘off-colour’, loss of appetite). For live virus vaccines, the systemic symptoms can be a mild form of the disease (e.g. fever, rash for measles vaccine).
The common reactions usually take place within a day or so of immunisation - except for live attenuated vaccines such as measles/MMR, where reactions occur 6 to 12 days after immunisation (the time taken for the vaccine virus to replicate). Common reactions usually only last a few hours to a few days. The frequency and severity of these common reactions vary by vaccine and by person. (See specific vaccine preventable diseases for details).
Rare vaccine reactions
While there is some similarity between the common reactions caused by all vaccines, the rare more serious reactions tend to be more specific to each vaccine.
Most of the rare and more serious vaccine reactions (e.g. seizures, thrombocytopaenia, hypotonic hyporesponsive episodes, persistent inconsolable screaming) do not lead to long-term problems.
Anaphylaxis, while potentially fatal, is treatable without leaving any long-term effects. Although encephalopathy is included as a possible rare reaction to measles or DTP vaccine, it is not certain that these vaccines in fact cause encephalopathy (brain inflammation) as there is no difference in incidence between vaccinated and unvaccinated people.
Programme errors result from errors and accidents in vaccine preparation, handling, or administration. They are preventable with good practice, appropriate facilities and equipment. Vaccinator technique in choosing the site for the injection, the depth of the injection, and the speed of injection may affect the rate and extent of local reactions.
A programme error may lead to a cluster of events associated with a particular provider, health facility, or even a single vial of vaccine that has been inappropriately prepared or contaminated. Programme errors can also affect many vials (e.g. by freezing vaccine during transport, leading to an increase in local reactions and loss of effect).
Globally, the most common programme errors are non-sterile injections leading to bacterial or viral infections. This is extremely rare in New Zealand, particularly as multi-dose vials are not usually used.
Injection technique can also lead to increased local reactions for example giving intramuscular injections too superficially.
Individuals can react in anticipation and as a result of the pain of any injection. This reaction is unrelated to the content of the vaccine. Some individuals may be ‘needle-phobic,’ aggravating such reactions.
The injection reaction can lead to a group event. The event is essentially a psychological reaction that spreads between individuals of the group (who are usually primed for this by high levels of anxiety), when a member of the group suffers a reaction such as a faint or other vaccine reaction.
Fainting (vasovagal reactions) is a relatively common injection reaction, but usually only affects children aged over five years. Fainting does not require any management beyond placing the patient in a recumbent position, preferably the recovery position.
The likelihood of faints can be anticipated when immunising older children, and reduced by minimising stress in those awaiting injection through short waiting times, comfortable room temperatures, preparation of vaccine out of recipient’s view, and privacy during the procedure. Avoiding injury from the fall is important, and those at particular risk should be immunised while lying down; the faint however can occur many minutes after the immunisation.
Hyperventilation as a result of anxiety about the immunisation leads to specific symptoms of light-headedness, dizziness, tingling around the mouth and in the hands, and sometimes chest pain.
IMAC 2010 Guidelines for a pre or post vaccination hyperventilation event - click here
Younger children tend to react in a different way, with vomiting a common anxiety symptom. Breath-holding may occur, which can end in a brief period of unconsciousness during which breathing resumes. They may also scream to prevent the injection or run away.
Clear explanations about the immunisation and calm, confident delivery will decrease the level of anxiety about the injections and thus reduce the likelihood of an injection reaction.
It is important to be able to distinguish the loss of consciousness resulting from a fainting (vasovagal) injection reaction to that of anaphylaxis.
|Onset||Usually at the time or soon after the injection||Usually some delay between 5-30 minutes after injection|
|Skin||Pale, sweaty, cold and clammy||Red, raised and itchy rash;swollen eyes, face; generalised rash|
|Respiratory||Normal to deep breaths||Noisy breathing from airways obstruction (wheeze or stridor)|
|Neurological||Transient LOC, good response once prone||LOC, little response once prone|
Distinguishing anaphylaxis from a faint (vasovagal reaction)Source: “Immunisation Safety Surveillance: guidelines for managers of immunisation programmes on reporting and investigating adverse events following immunisation. Manila: World Health Organization, 1999.”
This is probably the most important type of AEFI, in terms of impact on immunisation programmes. It is the main cause for the continuing controversy about immunisation safety.
The distinction between the event happening after immunisation and being caused by immunisation is not easy to make, and many people find it hard to understand. The automatic assumption is that when B follows A, it is that A caused B.
Because large numbers of immunisations are given every day, it is inevitable that in the days following immunisation many people will suffer major health events. Most vaccines are given early in life, when infections and other illnesses are common, including manifestations of an underlying congenital or neurological condition. The health event is likely to be blamed on the immunisation, simply because of the temporal association between the events that, in fact, reflect chance and coincidence and not a real causal relationship.
Examples of false allegations caused by coincidental events
One good example is the association between immunisation and Sudden Infant Death Syndrome (SIDS or cot death). The incidence of SIDS peaks around the age when infant immunisations are delivered. So, many SIDS cases will occur in children who have been recently immunised. Inevitably, when a previously healthy child dies a cause is sought. In some cases there will, purely by chance, have been an immunisation, shortly before the death. The only way to clarify if such an event is in fact due to a chance association is through careful studies. In the case of SIDS, controlled studies have shown that the association of SIDS and immunisation is purely coincidental and not causal.
More recently, there has been much concern around MMR vaccine causing autism that started from a report of 12 cases with a variable temporal association. Despite many controlled studies that have failed to find a link, many parents remain convinced that their child’s autism was due to the vaccine. This association has arisen because MMR is usually given early in the second year of life – at a time when autism is recognised.
Anticipating coincidental events
It is possible to predict how often one would expect to see a coincidental association between a health event and immunisation. Using the expected incidence (number of cases per population to be vaccinated per time period) of the health event.
For example, assume that one million children aged 1-15 years are immunised in a mass campaign and the background mortality rate for this population is 3 per 1000 per year. Then, 250 deaths can be expected in the month after immunisation and 8 deaths on the day of the immunisation, simply by coincidence. These deaths will be temporally associated with, even though entirely unrelated to, immunisation.
A similar calculation is shown in the Table below for infant (aged under one-year) deaths in selected Western Pacific countries for the number of deaths temporally associated with routine DTP immunisation (calculated in 1999).
|Infant mortality (per 1,000 live births)||Number of births per year||Number of infant deaths during one year in the:|
|month after immunisation||week after immunisation||day after immunisation|
NOTE: Assumes uniform distribution of deaths and children who are near to death will still be immunised.
Infant mortality and births from 1998 Western Pacific Region Health Data Bank.
IMR= Infant mortality rate per live birth (can substitute for rate of any event).
N = Number in population (births used as proxy for numbers aged under one year).
nv = number of immunisation doses: assumed here to be three visits.
ppv = proportion of population vaccinated: assumed here to be 90% for each dose.
Note: in calculation of deaths the first line of equation shows number of total deaths in period, second line adjusts for exposure to vaccine within that period, multiplied by the number of periods in the year.