Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Epinephrine self-injection for anaphylaxis in children

Three Part Question

In [children with anaphylaxis] does [self-injection of epinephrine] lead to [reduced mortality and morbidity]?

Clinical Scenario

A five year-old boy is admitted with severe anaphylactic shock having inadvertently ingested peanuts at a birthday party. He had a previous reaction two years ago and was given an epinephrine auto-injector for use at home. His mother had used this when the reaction first started but to no avail. You administer intramuscular epinephrine while wondering whether there is any evidence for the effectiveness of epinephrine self-injection.

Search Strategy

Ovid Medline 1950 - 2007 August Week 5
Ovid Embase 1980 - 2007 Week 35
The Cochrane Library 2007 Issue 3
Medline, Embase:
[exp Anaphylaxis/ OR anaphyla$.mp. OR exp Hypersensitivity/ OR allerg$.mp.] AND [ OR OR "epi-EZ pen".mp. OR "epi ez pen".mp. OR OR exp epinephrine/ OR] AND [(exp Home Care Services/ OR exp Outpatients/ OR OR out-patient$.mp. OR self-administr$.mp. OR OR exp First Aid/ OR first OR (exp Infant Mortality/ OR exp Child Mortality/ OR OR exp Hospital Mortality/ OR exp Mortality/ OR exp Death/ OR exp "Cause of Death"/ OR exp Death, Sudden/ OR] AND limit to human and English language and All Child 0-18 years (Ovid filter)
([MeSH heading Anaphylaxis] OR anaphyla*) AND ([MeSH heading Epinephrine] OR adrenaline OR epinephrine OR epipen)

Search Outcome

Altogether 94 papers were identified using Medline, 585 using Embase and 26 using Cochrane. Three papers provided evidence that was relevant to the three-part question.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Gold et al
68 children with a history of anaphylaxis who were prescribed EpiPens by the paediatric allergy service. Data collected by telephone interview with parents.Retrospective telephone surveyParental recall of method of administration with EpipenOnly 16 (24%) of parents were able to recall all 4 steps required for the correct use of Epipens; 5% could not recall any steps.Suboptimal study design to answer this question. A prospective observational cohort would be more informative (RCT would be ideal but probably unethical). Possible selection bias - the parents of only 80% of the patients identified were interviewed. 19% could not be contacted - their outcome is unknown. No standard criteria for initial prescription of Epipen. Time since index reaction not standardised. In-hospital adrenaline use apparently assessed by parental interview. Severity of episodes not objectively measured.
Epipen use in anaphylaxisEpipen given in 13 (29%) of 45 anaphylactic reactions. Of those not given Epipen, 15 (45%) were later given epinephrine in hospital. Of those given Epipen, 2 of 13 (15%) later received epinephrine in hospital (P<0.05)
Hospital admissionEpipen not given: 15 of 32 (47%) admitted; Epipen given: 2 of 13 (15%) admitted, P<0.05
Colver et al
United Kingdom
229 cases of children (aged under 16 years) admitted to hospital with food-allergic reactions between 1998 and 2000.Prospective surveyDeath3 children died. One received epinephrine via auto-injector at homePatients only included after reported hospital admission. Emergency Department attendances without admission not included.
Near-fatal anaphylaxisOccurred in 6 children. In 3 children excess epinephrine administration (in hospital) was implicated in clinical deterioration.
Potential reduction in severity by use of epinephrine auto-injectors6 severe cases and 7 non-severe cases received epinephrine via auto-injector before arriving at hospital. Auto-injectors could therefore have reduced the severity of a maximum of 13/229 reactions.
Potential benefits if epinephrine auto-injectors had been more widely availableOf the 58 severe cases, a maximum of 6 could have benefited had auto-injectors been available to them (the remainder either did not have epinephrine at any point, already had an auto-injector but did not use it, had epinephrine administered by primary care or ambulance staff within 10 minutes, had not had a previous allergic reaction to food or were over 12 years old and their only previous reaction had been to allergens as babies).
Pumphrey et al
United Kingdom
164 cases of fatality with anaphylaxis recorded as the cause of death between 1992 and 1998 (148 had records available for further analysis).Analysis of registry dataMedian time to respiratory or cardiac arrest30 minutes for foods (range 6-360), 15 minutes for venom (range 4-120) and 5 minutes for iatrogenic reactions (range 1-80).Registry data, not specifically designed to answer this question. Only fatal cases included. Cases where epinephrine self-treatment prevented death therefore not included. Adults and children included.
Details of patients who had been given adrenaline self-treatment kits5 did not use the kit (not with patient in 2 cases, out of date in 1 case, may have collapsed too quickly in 1 case, found dead holding unused kit in 1 case (could not assemble it?)


Epinephrine is the drug of choice for initial treatment of severe anaphylaxis as it blocks mediator release and reverses systemic effects. Many children are now prescribed epinephrine auto-injectors for emergency use in the community. The survey identified suggests that epinephrine self-injection may reduce subsequent hospital admission and need for epinephrine in hospital. However, it is notable that 76% of parents were not familiar with the correct procedure for using the devices, despite prior instruction (Gold et al, 2000). There is evidence that prompt use of epinephrine improves prognosis in severe anaphylactic reactions. Sampson et al (1992) studied six children and adolescents who died of anaphylactic reactions to foods and seven others who nearly died and required intubation. The six who died had symptoms within 3 to 30 minutes of allergen ingestion and only two received epinephrine within the first hour. All patients who survived had symptoms within 5 minutes of allergen ingestion and all but one received epinephrine within 30 minutes. Bock et al (2001) found that only 4 of 32 fatal anaphylactic reactions identified from a registry had received timely epinephrine. As 53% of children with reported allergic reactions in the United Kingdom and Ireland between 1998 and 2000 (MacDougall et al, 2002) have had a previous allergic reaction (9% requiring hospitalisation), there is certainly potential to identify children who may be eligible for epinephrine self-treatment in the future. However, the number of patients who may benefit from widespread availability of epinephrine auto-injectors is likely to be small. A significant proportion of children will not have their auto-injector available at the time of the reaction. Others may be unfamiliar with the technique for self-injection, a small number may die despite therapy and many patients have had no previous reaction (therefore being ineligible for prior epinephrine prescription). It is known that epinephrine may lead to cardiac arrhythmia. Colver et al (2005) reported three cases where excess epinephrine administration in hospital was implicated in clinical deterioration. MacDougall et al (2002) reported one case in which death was attributed to excess intravenous epinephrine administration. However, the reported search strategy did not reveal any evidence of death or cardiac arrhythmia following self-injection of intramuscular epinephrine. Although there is no evidence of high quality available to answer the three-part question, there is a sound physiological basis for the early use of epinephrine in severe anaphylaxis. At present it would be unethical to perform a randomised controlled trial. As early epinephrine may be life-saving and in the absence of evidence of harm following self-injection, measures to increase its availability should be encouraged even in the absence of high level evidence.

Clinical Bottom Line

Where available, epinephrine auto-injectors should be utilised in the community at an early stage for severe anaphylactic reactions, although high-quality evidence is lacking. To maximise any potential benefit it is important to provide education for patients, parents and carers regarding injection technique and ensuring auto-injector availability at all times.

Level of Evidence

Level 3 - Small numbers of small studies or great heterogeneity or very different population.


  1. Gold MS; Sainsbury R First aid anaphylaxis management in children who were prescribed an epinephrine autoinjector device (EpiPen) Journal of Allergy and Clinical Immunology 2000; 106(1 I): 171-176
  2. Colver AF; Nevantaus H; MacDougall CF; Cant AJ Severe food-allergic reactions in children across the UK and Ireland, 1998-2000 Acta Paediatrica 2005; 94: 689-695
  3. Pumphrey RSH Lessons for management of anaphylaxis from a study of fatal reactions Clinical & Experimental Allergy 2000; 30: 1144-1150
  4. MacDougall CF; Cant AJ; Colver AF How dangerous is food allergy in childhood? The incidence of severe and fatal allergic reactions across the UK and Ireland Archives of Disease in Childhood 2002; 86: 236-239
  5. Sampson HA; Mendelson L; Rosen JP Fatal and near-fatal anaphylactic reactions to food in children and adolescents New England Journal of Medicine 1992; 327(6): 380-384
  6. Bock SA; Munoz-Furlong A; Sampson HA Fatalities due to anaphylactic reactions to foods Journal of Allergy and Clinical Immunology 2001; 107: 191-193