Reduced risk of hospital admission for childhood asthma among Scottish twins: record linkage studyBMJ 2000; 321 doi: https://doi.org/10.1136/bmj.321.7263.732 (Published 23 September 2000) Cite this as: BMJ 2000;321:732
- David P Strachan, professor of epidemiology ()a,
- Sarah E Moran, statisticiana,
- Kevin McInneny, programmerb,
- Mary Smalls, head of statisticsb
- a Department of Public Health Sciences, St George's Hospital Medical School, London SW17 0RE
- b Information and Statistics Division, National Health Service in Scotland, Trinity Park House, Edinburgh EH5 3SQ
- Correspondence to: D P Strachan
- Accepted 22 May 2000
A recent study of Swedish army conscripts found a reduced prevalence of asthma and allergic rhinitis among twins.1 We analysed routine data on hospital admissions in Scotland to compare risks of asthma and other respiratory complaints among twins and singletons.
Subjects, methods, and results
We identified all twins born in Scotland during 1981–4 from computerised maternity records. Subsequent admissions of twins to Scottish hospitals during 1981–94 were ascertained by probability matching on the basis of date of birth, sex, and surname. This matching is considered 99% accurate for singletons, but for twins it is reliable only at the level of the pair: which twin is admitted cannot be identified with certainty.
We identified hospital admissions for respiratory disease (ICD-9 (international classification of diseases, 9th revision) codes 464, 466, 480-486, and 490-496) for all Scottish children born during 1981-4. Rates of hospital admission among singletons and twins up to 10 years of age were compared by cause and sex, assuming Poisson errors in the numerators.
Twins were significantly less likely than singletons to be admitted for respiratory diseases (table). This was attributable to a reduced risk of admission for asthma among twins (code 493) by more than half throughout the age range 0–10 years. In contrast, twins were at significantly increased risk of admission for acute bronchitis and bronchiolitis (code 466). Admissions for other respiratory diseases were divided more equally between twins and singletons (table).
No significant differences were found between twins of the same or different sex in admission rates for any cause or all respiratory diseases combined. The relative difference in rates of admission with asthma for twins of the same sex compared with singletons was greater for males than for females, although this sex interaction was not significant.
Our record linkage study confirms that twins are at reduced risk of asthma but not of other respiratory diseases than are singletons. The twofold difference in rates for admissions with asthma between Scottish twins and singletons is greater than the difference in asthma prevalence (4.9% versus 5.9%) reported among Swedish army conscripts.1 Cases admitted to hospital possibly represent more severely affected patients among whom the “twin effect” is more influential.
The difference in rates for admission with asthma may have been exaggerated if the admission of one twin to hospital led to recognition and prophylactic treatment of asthma in the cotwin. Alternatively, hospital staff may prefer to label episodes of wheezing in twins as “acute bronchitis” rather than “asthma.” Diagnostic transfer cannot be the whole explanation, however, because a reduced risk of admission among twins was still evident when all respiratory diagnoses were combined.
Although we were unable to distinguish monozygotic and dizygotic pairs, twin registry studies suggest a similar prevalence of asthma in monozygotic and dizygotic twins.2-4 Both types of twins tend to weigh less at birth than singletons, but low birthweight babies are generally at increased risk of asthma. Adjusting for birth weight would therefore accentuate rather than explain the twin effect, as was evident in the Swedish study.1
A more coherent explanation would be that twins have an extra sibling of the same age, and their reduced risk of asthma is a special case of the protective effect of large families on allergic disease.5 Sibship size is, however, more consistently related to hay fever and eczema than it is to asthma and wheezing.5 Swedish twins are also at a decreased risk of allergic rhinitis, but the effect of multiple gestation on asthma prevalence was not confined to conscripts with allergic rhinitis.1
We conclude that the effect of multiple gestation on risk of asthma is not readily explained by recognised risk factors. This deserves investigation as a clue to early developmental influences on asthma.
Contributors: DPS, MS, and KMcI designed the study. KMcI and SEM analysed the data. DPS took the lead in writing the paper and acts as guarantor.
Competing interests None declared.