Cohort studyindependent cross-checkfull text checked★ key study I-ABX-03 in-depth review
Antibiotics and asthma: sibling comparison (Sweden)
Antibiotics in fetal and early life and subsequent childhood asthma: nationwide population based study with sibling analysis.
Overview
Örtqvist AK et al.. BMJ 2014 · PMID 25432937 · DOI 10.1136/bmj.g6979
When siblings were compared, most of the antibiotic-asthma link disappeared.
Does the study support the specific claim? No · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Test confounding.
- How was the study done?
Sibling analysis.
- How many people, animals or samples were studied?
493,785 children.
- What was actually found?
Fetal antibiotic exposure: HR 1.28 in the whole cohort, 0.99 when comparing siblings. Antibiotics for urinary/skin infections: 1.54 vs 0.85 in siblings. Antibiotics for respiratory infections remained linked (2.36 in siblings), likely because of the infections themselves.
- How large was the observed effect?
Fetal HR 0.99 in siblings.
- What do the results mean in everyday life?
Antibiotics probably not the main cause.
- What are the study's limitations?
Reason for prescriptions not recorded; short follow-up (to about age 6); fewer informative sibling pairs for some comparisons.
- What does it say about Elawady's specific claim?
Contradicts causal claim.
- Do the results apply to Germany and Europe?
High.
- What can explicitly not be concluded from this study?
That antibiotics cause asthma; a small effect cannot be fully excluded.
Scientific in-depth review
1. Full reference
- Authors
- Örtqvist AK et al.
- Journal
- BMJ
- Year
- 2014
- Identifier
- PMID 25432937 DOI 10.1136/bmj.g6979
- Verification
- verified – PubMed efetch 2026-10-25
- Retracted
- no
- Corrections
ErratumIn: BMJ. 2014;349:g7395
- Extent checked
- full text
- Conflicts of interest and funding
Funding: Swedish Research Council (2011-3060) via SIMSAM (80748301, 340-2013-5867), Stockholm County Council (ALF), Karolinska Institutet epidemiology programme, Swedish Heart Lung Foundation; funders had no role. Competing interests: HK received institutional fees from RTI Health Solutions and several pharmaceutical companies for post-approval safety studies outside the submitted work.
2. Research question
Is the antibiotic-asthma association confounded?
3. Study design and population
- Study design
- Cohort study – Nationwide prospective register-based birth cohort with sibling-control analysis (stratified Cox model conditional on sibling group)
- Population
Children born in Sweden 2006-2010 to pregnancies starting from 1 July 2005, followed to end of 2011 (school age at most)
- Country
Sweden
- Period
Pregnancies July 2005-December 2010; follow-up to 31 December 2011
- Sample size
493,785 children in cohort analyses; 180,894 in sibling analyses (93% two-child families); 6% developed asthma; 20% exposed in fetal life, 62% had >=1 childhood antibiotic prescription
- Inclusion and exclusion criteria
Medical Birth Register linked to Multi-Generation, Prescribed Drug and Patient Registers; sibling analyses restricted to full siblings (9,896 excluded with unknown father)
4. Methods and endpoints
- Exposure / intervention
Filled prescriptions of systemic antibiotics (ATC J01) in fetal life (by trimester) and childhood (by age band), grouped as any, airway (respiratory-infection) and urinary tract/skin antibiotics; number of prescriptions
- Comparator
Unexposed children (cohort); unexposed siblings (sibling analysis)
- Endpoints
Asthma: ICD-10 J45 diagnosis plus dispensed asthma drugs (inhaled corticosteroids, LTRA or combinations)
- Statistical methods
Cox regression with attained age as time scale and sandwich SEs; stratified sibling Cox models; adjustment for maternal smoking, parity, family situation, maternal country of birth, parental education, maternal asthma (cohort only) and perinatal factors; sensitivity analyses (asthma from age 2 after first-year exposure, birth order, dose-response, Swedish-born mothers)
5. Actual results
- Results
Fetal exposure: cohort adjusted HR 1.28 (95% CI 1.25-1.32) vs sibling HR 0.99 (0.92-1.07). Childhood any antibiotics, youngest age band: cohort HR 3.71 (3.41-4.03) vs sibling 2.11 (1.61-2.76). Airway antibiotics: cohort HR 4.12 (3.78-4.50) vs sibling 2.36 (1.78-3.13); urinary tract/skin antibiotics: cohort 1.54 (1.24-1.92) vs sibling 0.85 (0.47-1.55). First-year exposure and asthma from age 2: associations in cohort, none in sibling analyses for any antibiotic group. Hazard ratio for any/airway antibiotics fell from about 4 to 2 within 6 months after exposure. Dose-response persisted in sibling analyses for any and airway antibiotics but not for urinary tract/skin antibiotics.
- Absolute risks
Overall asthma prevalence 6% in both cohorts; no absolute risk differences by exposure reported
- Authors' conclusion
The association between antibiotics and asthma is likely caused by confounding factors shared within families, in addition to confounding by indication or reverse causation from respiratory infections; antibiotics might not cause asthma, but careful consideration is needed before treating respiratory symptoms with them.
6. Strengths
- Nationwide population-based registers; prospective, objective exposure data (no recall bias)
- Sibling design controls shared genetic and family factors
- Antibiotic-type comparison (airway vs urinary/skin) as indication probe
- Lag analysis (first-year exposure, asthma from age 2)
- Validated asthma definition requiring diagnosis plus medication
7. Weaknesses
- Indication (respiratory infection) not recorded; residual confounding by indication and reverse causation for airway antibiotics
- Sibling analyses rely on exposure-discordant pairs; wider CIs (e.g. urinary/skin 0.47-1.55); sensitive to non-shared confounding and measurement error
- Short follow-up (to age <=6); early childhood asthma diagnoses include wheezing phenotypes
- Adherence and antibiotics prescribed abroad not captured
- Cannot address allergic diseases other than asthma
8. Risk of bias
- Orienting tool
- ROBINS-E
- 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 ROBINS-E-oriented judgement based on the full text; not formal (needs two independent reviewers and a protocol). Changed from low to some_concerns because residual confounding by indication and short follow-up remain, although the design is strong for detecting familial confounding.
| Domain | Judgement | Rationale |
|---|---|---|
| Confounding | some concerns | Shared family confounding handled by sibling design; non-shared confounding by indication (respiratory infections) cannot be adjusted, as authors state. |
| Measurement of exposure | low | Dispensed prescriptions from national register; adherence unknown. |
| Selection of participants | low | Complete national birth cohort; sibling subset defined by family structure. |
| Post-exposure interventions | not applicable | Not relevant. |
| Missing data | low | Register-based with few losses (4,922 emigrated, 1,448 died); complete-case for covariates. |
| Measurement of outcome | some concerns | Register diagnosis plus medication is specific, but asthma vs infection-related wheeze in young children is hard to distinguish. |
| Selection of reported result | low | Comprehensive reporting of cohort and sibling estimates, crude and adjusted, with sensitivity analyses. |
Full-text review (25 October 2026): Added follow-up to 2011, exposure prevalences, youngest-age HRs (3.71 vs 2.11), lag analysis showing no sibling association for first-year exposure and asthma from age 2, dose-response findings, funding/COI; bias overall changed from low to some_concerns (unadjustable confounding by indication, short follow-up).
9. Assessment of the statistical interpretation
Attenuation of fetal-exposure HR from 1.28 to 0.99 within siblings is strong evidence of familial confounding; the remaining airway-antibiotic association (HR 2.36) is most plausibly explained by indication/reverse causation, as it disappears with a lag (asthma from age 2) and is absent for urinary/skin antibiotics.
10. Relevance to the specific claim
C-ABX-06 (contradicts causation).
11. Independent contrary findings
12. Transferability to Germany and Europe
High.
13. Overall methodological judgement
Key study showing that much of the antibiotic-asthma association is explained by family factors and by the infections being treated; it does not support a causal effect of antibiotics on asthma, while not fully excluding a small effect.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | some concerns |
| Precision of results | high |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Longer follow-up into later childhood; other allergic outcomes; settings with different prescribing.
15. Linked fact checks
- Antibiotic exposure in early life causes asthma and allergic disease. – Studies contradicting or limiting the claim