Intended for healthcare professionals


Home environment and severe asthma in adolescence: a population based case-control study

BMJ 1995; 311 doi: (Published 21 October 1995) Cite this as: BMJ 1995;311:1053
  1. David P Strachan, senior lecturer in epidemiologya,
  2. Iain M Carey, statisticiana
  1. aDepartment of Public Health Sciences, St George's Hospital Medical School, London SW17 0RE
  1. Correspondence to: Dr Strachan.
  • Accepted 27 July 1995


Objective: To investigate the effects of the home environment on the risk of severe asthma during adolescence.

Design: A questionnaire based case-control study drawn from a cross sectional survey of allergic diseases among secondary school pupils in Sheffield in 1991.

Subjects: 763 children whose parents had reported that over the previous 12 months they had suffered either 12 or more wheezing attacks or a speech limiting attack of wheeze. A further 763 children were frequency matched for age and school class to act as controls. Analysis was restricted to 486 affected children and 475 others born between 1975 and 1980 who had lived at their present address for more than three years.

Results: Independent associations with severe wheeze were seen for non-feather bedding, especially foam pillows (odds ratio 2.78; 95% confidence interval 1.89 to 4.17), and the ownership of furry pets now (1.51; 1.04 to 2.20) and at birth (1.70; 1.20 to 2.40). These estimates were derived from subjects whose parents denied making changes in the bedroom or avoiding having a pet because of allergy. Parental smoking, use of gas for cooking, age of mattress, and mould growth in the child's bedroom were not significantly associated with wheezing.

Conclusions: Either our study questionnaire failed to detect the avoidance or removal of feather bedding by allergic families or there is some undetermined hazard related to foam pillows. Synthetic bedding and furry pets were both widespreadin this population and may represent remediable causes of childhood asthma.

Key messages

  • Key messages

  • The bedroom arrangements of one quarter of teenagers with troublesome asthma in Sheffield had been altered because of the child's allergy or chest problem. One quarter of these families had avoided pets or removed them from the home because of allergy

  • Nevertheless, about two thirds of asthmatic children were exposed to furry pets in their home, an exposure which almost doubled their risk of troublesome symptoms

  • Alterations to the bedroom commonly entailed use of non-feather bedding, used by 95% of severely wheezy children. Exposure to synthetic pillows was associated with a two to threefold increase in risk of severe wheezing, even after allowance for selective avoidance of allergens by the families of allergic or asthmatic children

  • Avoidance of feather bedding is unlikely to benefit children with asthma, and there may be a hitherto unidentified hazard associated with use of synthetic pillows


Recent concern about environmental influences on childhood asthma has focused on the possible hazards of outdoor pollutants, particularly those derived from vehicle exhausts, such as nitrogen dioxide and ozone.1 Less attention has been directed towards the indoor environment, although many people spend upwards of 90% of their time indoors.2 Certain outdoor pollutants, such as particulates and nitrogen dioxide, may be present at higher concentrations indoors, and domestic allergens such as house dust mites and pet dander (dandruff) are a major source of allergic sensitisation which relate specifically to asthma in children.3

Although childhood asthma is common, many cases are relatively mild.4 Few epidemiological studies have focused on the more severe forms of the disease, which pose a substantial burden on the health services. We present the relation between aspects of the home environment and troublesome asthma, paying particular attention to possible biases arising from avoidance of allergens by allergic families.5

Subjects and methods

A two page questionnaire was circulated at school to the parents of all children in the first to fifth forms (11-16 years old) attending state or private secondary schools in Sheffield in November 1991. Of the 35 schools, 34 participated, and replies were received for 18203 (79%) of the 23054 questionnaires issued.

Cases and controls were selected on the basis of responses to previously published questions relating to asthma and wheezing.6 Children whose parents reported that they suffered either 12 or more wheezing attacks in the past 12 months or an attack of wheeze over the same period that limited speech to only one or two words at a time between breaths were selected as cases. One child with no history of asthma or wheezing at any age and frequency matched for age and school was selected as a control for each case.

A second questionnaire was posted to parents of both groups of children in June 1993. Replies were received for 571/763 children with wheezing (75%) and 568/763 without (74%). Information was collected on household pets, methods of cooking, the child's bedding, dampness or mould, or both, in the child's bedroom, and parental smoking in the house. Respondents were asked two questions specifically about changes to the home environment as a result of asthma or allergy. These asked if they had ever got rid of a pet or decided not to have one because one of the family might be allergic to it, and also if they made any alterations or special arrangements in the child's bedroom because of allergy, asthma, or other chest problems.

Results are presented for 486 cases and 475 controls. The children were born 1975-80 inclusive and had lived at their 1993 address for at least three years. The data were analysed as an unmatched study because cases and controls were frequency matched rather than individually matched. Unadjusted odds ratios were derived from tabulations of the characteristics of the cases and controls: firstly, for all subjects; secondly, restricted to subjects whose parents denied pet avoidance; and, thirdly, restricted to children whose parents denied alterations to the child's bedroom because of asthma or allergy.

Mutually adjusted odds ratios for each indoor environmental risk factor were derived from multiple logistic regression models fitted in the generalised linear interactive modelling system (GLIM).7 These models included year of birth and sex in addition to environmental characteristics. Interaction terms were fitted to exclude children for whom measures of avoidance of allergens were reported from the assessment of risks associated with pets and bedding while retaining these subjects in the model for assessment of other risk factors.


Among the 18203 valid responses to the 1991 questionnaires about one child in four had ever wheezed (26.4%; 95% confidence interval 25.8% to 27.1%). The prevalence of wheeze in the past 12 months was 15.7% (15.1% to 16.3%). Severe wheezing was much less common, with a similar annual prevalence of speech limiting wheeze (2.9%; 2.6% to 3.1%) and frequent attacks (12 or more in the past year) (2.3%; 2.1% to 2.5%), comparable with the results of a recent nationwide survey.4 The lifetime prevalence of asthma diagnosed by a doctor was 17.7% for the whole study sample, 62.7% for children who had ever wheezed, 92.4% for those with speech limiting wheeze, and 91.7% for the group with 12 or more attacks.

The 486 cases comprised three groups: frequent wheeze only (n=148), speech limiting wheeze only (n=225), and those with both (n=113). These three groups were compared, in turn, with the 475 controls, producing unadjusted odds ratios for indoor environmental factors shown in table I.


Prevalence of exposure and unadjusted odds ratios for three mutually exclusive case groups in relation to indoor environmental risk factors. Figures are numbers (percentage); odds

View this table:

Ownership of furry pets and mould in the bedroom were significant risk factors for speech limiting wheeze (odds ratio 1.64 (P<0.05) for ever owning a pet; 2.36 (P<0.05) for damp and mould in bedroom). The only significant relation with parental smoking was for the category of mothers smoking more than 10 cigarettes a day and frequent and speech limiting wheeze in the child (2.28; P<0.05). Similar sized effects were seen for maternal smoking around the time of the child's birth (not shown). In contrast, the most consistent result across the three groups was a reduction in risk associated with feather pillows, which increased in strength with increasing severity of wheezing (odds ratio 0.31 (P<0.001) for frequent attacks; 0.22 (P<0.001) for speech limiting wheeze; and 0.10 (P<0.001) for both). There were similar, though less clear, reductions in the use of feather quilts by wheezy children.

A possible explanation for the lower proportion of wheezy children using feather bedding would be the previous removal of feather items in response to the child's symptoms. Allergic families might also tend to avoid or remove pets and thereby dilute a positive association of wheezing with pet ownership.

As expected, changes in the bedroom because of asthma, allergy, or chest problems were more commonly reported for cases (24.9% (121)) than for controls (2.5% (12)). When these subjects were excluded from the analysis the decreased risk associated with feather bedding diminished only slightly (odds ratio rising towards unity), and remained significant for feather pillows and for feather quilts (table II).


Unadjusted odds ratios for risk factors associated with bedroom before and after restriction to families denying alterations to bedroom. Figures are numbers (percentage); odds ratio (95% confidence interval)

View this table:

The families of 144 (29.6%) cases and 46 (9.7%) controls had avoided or removed household pets because of allergy. Exclusion of these subjects from the analyses relating to pet ownership increased the odds ratio for current pet ownership (1.49) but decreased that for past pet ownership (0.92). The effects of owning cats or dogs were broadly similar (table III).


Unadjusted odds ratios for pet ownership before and after restriction to families denying avoidance or removal of pet. Figures are numbers (percentage); odds ratio (95% confidence interval)

View this table:

table IV shows the independent effects of each risk factor as evaluated by multiple logistic regression. Adjustment for other factors altered the protective effect of feather pillows little (odds ratio 0.36 in table IV compared with 0.33 for “feather” and “both” combined in table II), and it remained highly significant (P<0.001). The increased risk associated with ownership of a furry pet was independent of other factors, with significant effects of approximately equal magnitude for pets now (P<0.05) and pets at birth (P<0.01).


Mutually adjusted odds ratios for indoor environmental

View this table:

After adjustment for other factors in table IV the odds ratios for pet ownership at birth, now, and both times compared with no reported pet ownership were, respectively: 1.30 (95% confidence interval 0.62 to 0.71); 1.35 (0.86 to 2.11), and 2.50 (1.60 to 3.94).

Restriction of the multiple logistic analysis to 102 cases and 142 controls with no parental history of wheeze, asthma, eczema, or hay fever slightly reduced the protective effect of feather pillows (odds ratio 0.41; 0.20 to 0.82), but it remained highly significant. Among these children the increased risk associated with furry pets persisted, albeit not significant at the 5% level (1.94; 0.99 to 3.80).


The main strength of our case-control study was the size of the population survey from which it was derived. This permitted systematic ascertainment of a large case group of children with more severe asthma, and selection bias was minimised by selecting a control group from survey respondents.

Limitations of the study are the incomplete response rates (79% in 1991; 77% in 1993) and reliance on a retrospective questionnaire assessment of exposures in the home. On the other hand, special care was taken to deal with the possible biases related to avoidance of allergens.

The most powerful risk factors identified were pet ownership and non-feather bedding. In contrast, effects of parental smoking, gas cooking, and mould growth were weak and non-significant.


Ownership of furry pets has been suggested as a risk factor for wheeze in many clinical studies,8 9 10 11 but epidemiological studies have generally reported no association or an inverse relation to furry pets,2 12 13 14 although a positive association was found in a recent study in Singapore.15

These inconsistencies could be due to dilution or reversal of positive associations by the tendency to remove pets from the home after the child (or other family members) have developed allergic complaints.

Brunekreef et al reported that 12% of families of Dutch children had removed pets from the home, and 2% had avoided household pets because of allergy.5 (This compares with 10% of our control families who avoided or removed pets.) The lowest prevalence of pet allergy was found in children who currently (but not previously) had a pet in the home, whereas the highest prevalence was found in families without pets at present who had previously owned some in the past. Kuehr et al reported similar findings for skin prick tests, the prevalence of sensitisation to cat dander being significantly raised only in children whose families had owned cats in the past but not among current cat owners.16

Our study suggests that exposure to furred pets is an independent risk factor for the more severe forms of wheeze in adolescence. Early pet exposure was no more influential than current exposure after allowance for pet avoidance. Of the control families, 62% reported current ownership of a furry pet, and even among families who said they had avoided or removed a pet because of allergy 14% owned a furry pet both now and at birth. These figures suggest widespread exposure to pet allergens and reluctance even among allergic families to remove all pets from the home. Using the adjusted relative risks for pet ownership at birth, now, and both times (cited in results) and the prevalence of these exposures in the control group (table I) we estimate that current and past ownership of a furry pet accounts for 40% of cases of severe wheeze in our population. Allergy to pet dander provides a biologically plausible mechanism for this relation. Furry household pets are likely to be a widespread but potentially removable cause of the more troublesome forms of childhood asthma.


The idea that feather pillows increase the risk of allergic asthma is widely accepted,17 but remarkably there is no previous epidemiological literature on the subject. Our study suggests a substantially lower risk of troublesome asthma among children using feather bedding relative to those using non-feather materials. At first sight, the most likely explanation for this observation was avoidance or removal of feather bedding by families of an asthmatic child or by allergic parents. The inverse relation of severe asthma to feather pillows, however, remained strong and highly significant even after restriction to non-allergic families who denied making changes in the bedroom. This raises the intriguing possibility that non-feather substitutes may pose a greater risk of asthma than any allergens associated with feather bedding.

The single “bedroom alterations” question (on the 1993 questionnaire) may have been insufficient to detect all conscious action by families to ensure that their children do not have feather pillows. Nevertheless, to explain the inverse association of asthma with feather pillows purely on the basis of undetected avoidance it would be necessary to assume that more than half of the families of cases who denied bedroom alterations and who would otherwise have used feather pillows avoided doing so because of their child's chest trouble (table II). This seems unlikely. Further epidemiological studies including more specific inquiries about allergen avoidance are required to confirm the apparent risk associated with non-feather pillows. If this association is causal, however, we estimate that it accounts for 53% of the severe asthma in our population.

Although firm conclusions cannot be drawn from a single study, our results suggest that avoidance of feather bedding is unlikely to benefit children with asthma and raise the possibility that there is a hitherto unidentified hazard associated with synthetic pillows. Volatile organic compounds released in low concentrations close to the breathing zone might increase mucosal permeability to inhaled allergens and thereby offer a speculative explanation for a possible causal link between foam pillows and childhood asthma. This possibility should be investigated by further observational and experimental studies.


We thank all the school staff and parents who took part in the 1991 and 1993 surveys; Dr Elizabeth Taylor and Ms Jayne Thompson for assistance with the 1991 survey; and Dr Jane Scarlett and Ms Joanne Griffiths for help with the 1993 mailing.


  • Funding The National Asthma Campaign

  • Conflict of interest None.