Meta-analysiscited by the Tayebat archivefull text checked★ key study A-LEG-14 in-depth review
How often food allergy kills
Incidence of fatal food anaphylaxis in people with food allergy: a systematic review and meta-analysis.
Overview
Umasunthar T et al.. Clin Exp Allergy 2013 · PMID 24118190 · DOI 10.1111/cea.12211
Among food-allergic people, fatal reactions occur about 1.8 times per million person-years, less often than accidental deaths.
Does the study support the specific claim? No · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Estimate fatal anaphylaxis rate
- How was the study done?
Meta-analysis
- How many people, animals or samples were studied?
13 studies
- What was actually found?
Very rare
- How large was the observed effect?
1.81/million person-years
- What do the results mean in everyday life?
Death risk is low even for allergic people.
- What are the study's limitations?
Heterogeneity
- What does it say about Elawady's specific claim?
Limits the alarm
- Do the results apply to Germany and Europe?
High
- What can explicitly not be concluded from this study?
Cannot justify avoidance by non-allergic people.
Scientific in-depth review
1. Full reference
- Authors
- Umasunthar T, Leonardi-Bee J, Hodes M, Turner PJ, Gore C, Habibi P, Warner JO, Boyle RJ
- Journal
- Clin Exp Allergy
- Year
- 2013
- Identifier
- PMID 24118190 DOI 10.1111/cea.12211 · PMC4165304
- Verification
- verified – PubMed E-utilities efetch, 2026-10-25
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Supported by an NIHR Biomedical Research Centre; first author salary support from Lincoln Medical (adrenaline autoinjector supplier); one author is a trustee of the Anaphylaxis Campaign; authors received UK Food Standards Agency research funding and a Meda Pharmaceuticals conference grant. (Earlier entry 'Medical Research Council funding' is not supported by the full text.)
2. Research question
Incidence of fatal food anaphylaxis among food-allergic people.
3. Study design and population
- Study design
- Meta-analysis – Systematic review and random-effects meta-analysis of incidence rates (generic inverse variance on log rates) from registries, databases and cohorts; MOOSE-reported; no registered protocol
- Population
Food-allergic people (denominator estimated by applying assumed food-allergy prevalence to defined general populations)
- Country
USA, UK, Canada, Sweden, Italy, Australia, Israel, Brazil
- Period
Searches Jan 1946 - 5 Sep 2012; included data from 1986-2011
- Sample size
13 studies, 240 fatal food anaphylaxis episodes, ~165 million food-allergic person-years (overall estimate from 10 studies; 0-19 y from 10; peanut from 7)
- Inclusion and exclusion criteria
Registries/databases/cohorts with >=1 fatal food anaphylaxis case in a defined population and period; excluded fatalities not probably/definitely anaphylaxis, undefined period or unquantifiable population; 14 studies with zero deaths (~4 million person-years) could not be included
4. Methods and endpoints
- Exposure / intervention
Food allergy (assumed prevalence 3% overall, 3.9% age 0-19, 1% peanut)
- Comparator
General-population incidence of other events (accidental death, murder, fire, lightning, ED visits) from CDC Wonder and Eurostat, not age adjusted
- Endpoints
Incidence of fatal food anaphylaxis per million person-years (micromorts)
- Statistical methods
Random-effects pooling of log incidence rates (SE 1/sqrt(events)); I2 heterogeneity; funnel plots; sensitivity analyses by assumed prevalence and excluding poor-quality studies (Newcastle-Ottawa <2)
5. Actual results
- Results
Food-allergic people: 1.81 per million person-years (95% CI 0.94-3.45; range 0.63-6.68). Age 0-19: 3.25 (1.73-6.10; range 0.94-15.75). Peanut-allergic: 2.13 (1.09-4.16; range 1.03-8.77). Excluding poor-quality studies: 1.57 (0.75-3.27); 0-19: 2.70 (1.31-5.54); peanut 2.04 (1.01-4.13). Prevalence sensitivity: overall 1.35 (0.71-2.59) to 2.71 (1.42-5.17); 0-19: 1.18 (0.63-2.22) to 6.13 (3.25-11.56); peanut 0.73 (0.38-1.43) to 4.25 (2.17-8.31). High heterogeneity in all analyses. Rate >=100 times lower than accidental death in the general population (>=10 times lower at age 0-19).
- Absolute risks
1.81 fatal food anaphylaxis deaths per million food-allergic person-years (3.25 at age 0-19; 2.13 for peanut allergy); below 1 per 100,000 person-years under all prevalence assumptions
- Authors' conclusion
Fatal food anaphylaxis for a food-allergic person is rarer than accidental death in the general population.
6. Strengths
- Comprehensive multi-database search without language restriction
- Duplicate independent screening and extraction
- Absolute rates with extensive sensitivity analyses for denominator assumptions
- Explicit comparison with everyday risks
7. Weaknesses
- Denominators rely on assumed food-allergy prevalence, not measured allergic cohorts
- High heterogeneity; variable case capture (registries, media, coroners, charities)
- No registered protocol
- Quality of included studies mixed (Newcastle-Ottawa 0-3/4)
- Possible under-ascertainment (e.g. deaths coded as asthma), acknowledged
- Zero-event studies excluded, which may inflate pooled rates
- Does not report the share of deaths caused by peanut
8. Risk of bias
- Orienting tool
- AMSTAR 2
- Type of assessment
- structured assessment based on the full text by a single reviewer (no formal double review)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Single-reviewer structured AMSTAR 2 appraisal from full text; not formal (needs two independent reviewers). One critical weakness (no registered protocol) would rate 'low' confidence under strict AMSTAR 2 rules; for an incidence review the main uncertainty is the assumed prevalence denominator, which the sensitivity analyses bracket.
| Domain | Judgement | Rationale |
|---|---|---|
| Item 1 PICO components | low | Population, exposure, outcome and comparator risks defined. |
| Item 2 Protocol registered a priori (critical) | high | No registered protocol; methods stated to be planned a priori. |
| Item 3 Explanation of included designs | low | Registries/cohorts needed for incidence estimation justified. |
| Item 4 Comprehensive search (critical) | low | Seven databases, conference abstracts, reference lists, no language limits. |
| Item 5 Duplicate selection | low | Two authors independently screened. |
| Item 6 Duplicate extraction | low | Two authors independently extracted. |
| Item 7 List of excluded studies with reasons (critical) | some concerns | Zero-fatality studies listed in supplement; full list of excluded full texts with reasons not evident in main text. |
| Item 8 Description of included studies | low | Table 1 describes setting, case identification and denominator. |
| Item 9 RoB assessment of included studies (critical) | some concerns | Modified Newcastle-Ottawa scale used; limited for registry case ascertainment. |
| Item 10 Funding of included studies | high | Not reported. |
| Item 11 Appropriate meta-analytic methods (critical) | some concerns | Random-effects pooling of rates appropriate, but pooled despite high heterogeneity; SE approximation and exclusion of zero-event studies. |
| Item 12 Impact of RoB on results | low | Sensitivity analysis excluding poor-quality studies. |
| Item 13 RoB considered in interpretation (critical) | low | Limitations and possible systematic bias discussed. |
| Item 14 Heterogeneity explained | some concerns | Discussed (data capture, prevalence) but not resolved. |
| Item 15 Publication bias (critical) | low | Funnel plots; small-study effect noted. |
| Item 16 Conflicts of interest reported | low | Funding and COI declared. |
Full-text review (25 October 2026): Added peanut-specific rate 2.13 (1.09-4.16), quality-sensitivity estimates, full prevalence sensitivity table, comparator methods, AMSTAR 2 item-level appraisal; corrected funding/COI (NIHR BRC, Lincoln Medical, Anaphylaxis Campaign trustee, FSA, Meda - not MRC).
9. Assessment of the statistical interpretation
Absolute incidence estimates with wide CIs and high heterogeneity; the conclusion of very low absolute risk is robust to all tested prevalence assumptions (always <1 per 100,000 person-years).
10. Relevance to the specific claim
Limits C-LEG-11.
11. Independent contrary findings
Itself.
12. Transferability to Germany and Europe
High.
13. Overall methodological judgement
The archive's own source limits its claim.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | taken out of context |
14. Open questions
Risk stratification tools to identify the small subgroups at higher risk.
15. Linked fact checks
- Peanut leads all food-anaphylaxis fatalities; peanut and tree nuts caused more than 90% of food-anaphylaxis deaths in the US registry. – Studies contradicting or limiting the claim
Use by the Tayebat archive
Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.
- What the archive says about the study
Bibliography only (cited in support of peanut danger).
- A. What was actually studied?
Fatal anaphylaxis incidence.
- B. What did the authors find?
Very low absolute incidence: 1.81 per million food-allergic person-years overall, 2.13 for peanut allergy; far below accidental-death rates.
- C. What conclusions do the authors themselves draw?
Rarer than accidental death.
- D. What does the archive derive from it?
Implied support for danger narrative.
- E. Is this derivation scientifically justified?
Not justified: the study shows the absolute risk is very low.
- Adequacy of the interpretation (by the archive)
- taken out of context