The bedding environment, sleep position, and frequent wheeze in childhood

The bedding environment, sleep position, and frequent wheeze in childhood

Anne-Louise Ponsonby

Bedding can trigger child wheeze, particularly wheeze caused by house dust mite (HDM)related airway obstruction. Bedding HDM (such as Dermatophagoides pteronyssinus [Der p]) levels predict respiratory symptoms and lung function. (1,2) Mattresses and underbedding are significant HDM allergen reservoirs. (3,4) It is less recognized that upperbedding such as quilts also can be a source of HDM exposure during sleep. A 10-fold reduction of airborne HDM allergen near the face during sleep was obtained by replacing used overlying quilts with new quilts. (5) Recently, synthetic quilts have been shown to have a markedly higher Der p content than feather quilts (6); the ratio (synthetic/feather) of the geometric mean for Der p ([micro]g/[m.sup.2]) was 15.45 (95% confidence interval [CI]: 4.28-55.8). (6) Higher HDM allergen levels have been found also in synthetic as compared with feather pillows. (7,8) The higher allergen loading near a child’s airway has been proposed as one possible explanation (7,8) for the numerous cross-sectional (9-11) and prospective associations (12,13) between synthetic bedding and childhood wheeze. However, the association may reflect selection bias associated with the preferential choice of synthetic bedding for children at risk of later allergic asthma. (11) Although randomized, controlled trials are constrained by the ethics of applying potentially deleterious synthetic bedding to children with severe wheeze, additional confirmatory evidence is needed from observational studies.

One novel approach to assess the role of bedding in asthma could be to look at the interaction between bedding and different sleep positions with regard to child wheeze. Specifically, if it could be demonstrated that synthetic quilt use, for example, conferred a different risk if a child slept in a different sleep position, then the case for synthetic quilt use being causally related to asthma would be strengthened. If the adverse effect was mediated through some mechanism that involved the bedding being in close proximity to the airway, one would postulate that the association between overlying synthetic quilt use and adverse respiratory outcomes would be observed specifically among children sleeping supine, during which the face would be closer to overlying bedding.

In the Tasmanian Asthma Survey, synthetic bedding was strongly associated with frequent wheeze. (13) The aim of this report is to examine whether the association between synthetic quilt use and frequent wheeze is significantly stronger among children who slept supine compared with children who slept nonsupine.


The 1995 Tasmanian Asthma Survey

We conducted a cross-sectional survey on all children who turned 7 years old in Tasmania, Australia, identified through primary schools and home learning and distance-education organizations. By December 31, 1995, questionnaires had been completed for 92% (6378 of 6911) of eligible children. (13) The questionnaire included questions from the International Study of Asthma and Allergies in Childhood. (14) Previous validation work has shown that the report of wheeze over the past 12 months has a sensitivity of 0.81 and a specificity of 0.85 for the physician diagnosis of asthma in childhood. (15)

Sleep position was classified as supine if the parents reported that the child usually slept on the back in response to the question: “What position does your child usually sleep in?” Children usually sleeping on their sides or stomachs were termed nonsupine sleepers. Synthetic pillows were those reported as foam/sponge/ tontine/polyester/Dacron. Quilts, doonas, or duvets of Dacron, polyester, or other synthetic composition were classed as synthetic quilts. Overlying bedding was categorized into 3 categories: any synthetic quilt, feather quilt, and other bedding (ie, neither feather nor synthetic quilt). Sheepskin was any wool fleece underbedding with either hide or material backing.

In addition, we conducted full exercise-challenge lung-function testing on 414 children at 23 randomly selected schools in Southern Tasmania; the method is described in detail elsewhere. (16) It included baseline and postexercise spirometric measures of forced expiratory volume in 1 second (FE[V.sub.1]), the latter after 6 minutes of free running. (16) These studies were approved by the University of Tasmania Ethics Committee (Human Experimentation), and parents provided informed, written consent.

Outcome Measurement

Disease misclassification within the broad spectrum of asthma has been a large problem in asthma epidemiology. For this bedding study, we were interested in identifying children who would be more likely to have HDM-triggered airway disease rather than those with asthma caused by other mechanisms. We chose to compare children with frequent wheeze (>12 episodes in the past year compared with children with no wheeze) for the following reasons: 1) children with asthma who are also sensitized to HDM are more likely to have severe or frequent asthma (17); 2) feather pillow and quilt use has been inversely associated with severe wheeze (18); and 3) in our setting, we have reported previously that HDM-sensitized children are much more likely (P < .0001) to have frequent wheeze titan nonsensitized children in our setting (prevalence ratio 19.6 [95% CI: 6.9, 55.6]). (19) In addition, we examined postexercise FE[V.sub.1], a lung measure previously observed to be lower in children with recent wheeze or asthma. (16)

Statistical Methods

First, univariate odds ratios (ORs) were calculated by using logistic regression. (20) The reference group chosen for upperbedding was feather quilts, because this bedding type has 1) lower HDM levels (7,8) and 2) has been associated previously with reduced wheeze, compared with synthetic quilts. (10) We examined the interaction between quilt type and sleep position by comparing synthetic and feather quilts. Most children (88.1%) slept under these 2 overbedding types (Table 1). For sleep position, children who slept supine were compared with nonsupine children, and children with no usual sleep position were excluded. The Breslow-Day test was used to examine differences in the association between synthetic versus feather quilt use and severe wheeze by sleep-position strata. (21) We then conducted a nested case-control study among the children who were reported to have a usual sleep position. Each child with frequent wheeze matched to control children with no wheeze over the past year. Each index child was matched to the nearest controls by date of birth who had the same status as the index child with regard to 4 characteristics: gender, foam mattress or not, electric blanket or not, and sheepskin or not. Thus, in the matched analysis, each matched set was standardized with regard to underbedding except for pillow type. Feather pillows and allergen-occlusive mattress covers were too uncommon to allow adequate matching and thus were considered as additional confounders in the matched analyses. In total, 117 children with frequent wheeze were matched to 1162 controls, with up to 10 controls per case. More than 80% of controls were born within a month of the index case. Conditional logistic regression models then were used to examine the relation between quilt use, sleep position, and reported respiratory outcomes, with control for confounders. (20) The difference in the quilt wheeze associations by sleep position was examined by the log likelihood ratio test for the reduction in deviance associated with the addition of an interaction term into the logistic model. (22) Thus, the interaction is based on a multiplicative model Any interactions observed would be present also in an additive model, but negative or absent findings cannot be taken to mean that additive interactions are not present. (23) The lung function sample was relatively small, and thus synthetic quilt use was examined as a dichotomous exposure to maximize the sample size for analysis. Multiple linear regression models were used to examine FE[V.sub.1] as a continuous outcome, with adjustment for relevant confounders. (24)


General Features

The mean age of the participants was 6.9 years (SD: 0.3), and 50.5% were male. Synthetic pillow or quilt use was common (Table 1). Feather quilt use was also common, but feather pillow use was not. Most parents were able to report a usual child sleep position, with 86% of children with frequent wheeze (117 of 136) or no wheeze (4048 of 4707) reported to have a usual position. More than one third of the children slept supine (Table 1). Among nonsupine children, the side position (n = 2564) was much more common than the prone sleep position (n = 484). Although many children had recent wheeze (23%), only 2% (n = 136) of the children were reported to have had frequent wheeze with >12 episodes in the last year (Table 1).

Sleeping Environment and Other Characteristics Associated With Frequent Wheeze or Sleep Position

Frequent wheeze was positively associated with the use of a synthetic quilt, synthetic pillow, or electric blanket and sleeping in a bottom brink (Table 2). A supine sleep position was associated with electric blanket or foam mattress use but not with frequent wheeze, rhinitis, or eczema (Table 2). Asthma medication type and use over the past year did not relate to sleep position. In Australia in 1997, bedding advice to children with asthma included the use of an allergen-occlusive mattress cover and sheepskin avoidance. (25) Consistent with this, children with frequent wheeze were more likely (OR: 18.39; P < .0001) to sleep on allergen-occlusive covers but less likely (OR: 0.54; P = .02) to sleep on a sheepskin.

The Relation Between Overbedding Composition and Frequent Wheeze by Usual Sleep Position

First, we examined the association between quilt use and severe wheeze after stratification by sleep position (Table 2). Among supine sleepers, synthetic (versus feather) quilt use was associated with frequent wheeze (OR: 3.75 [95% CI: 1.78, 8.63]). However, this apparent adverse effect was not evident among nonsupine children (OR: 1.12 [95% CI: 0.65, 1.93]). The difference in synthetic quilt effect by sleep position was significant (test for interaction: P – .007). There was no difference in the association between allergen-occlusive cover use or sheepskin use and recent wheeze by sleep position.

Second, we examined the data, taking possible confounding factors into account. We examined the nested case-control sample in which each child with frequent wheeze was matched to controls without wheeze with regard to age, gender, foam mattress use, electric blanket use, and sheepskin use. Again, synthetic compared with feather quilt use was associated (P = .02) with severe wheeze among supine sleepers (OR: 2.54 [95% CI: 1.18, 5.48]), but this association was absent for children sleeping on their sides or prone (OR: 1.09 [95% CI: 0.62, 1.92]; test for interaction: P = .005). Additional adjustment for synthetic pillow use provided matched ORs of 2.37 (95% CI: 1.08, 5.23) and 1.06 (95% CI: 0.60, 1.88), respectively (test for interaction: P = .005). Among nonsupine children, the effect of synthetic versus feather quilt use did not vary between side- and prone-sleeping children (P = .86). The significant difference in quilt effect by sleep position persisted after adjustment for premature (<37 weeks' gestation) birth, breastfeeding history, sibling number, allergen-occlusive mattress cover use, overlying sheet use, home gas use, child exposure to tobacco smoke, or exposure to fire other factors listed in Table 2. The interaction effect also persisted after adjustment for child hay fever, eczema, and family history of asthma (test for interaction: P = .01). Third, we examined 6 different sleep-position-overbedding combinations with regard to frequent wheeze (Table 3). The baseline group for each pairwise comparison was sleeping supine under a feather quilt. All other sleep combinations were associated with a higher risk of frequent wheeze, and the difference in risk for synthetic quilt use was significant regardless of sleep position used under a synthetic quilt. Fourth, we attempted to examine the effect of pillow composition on frequent wheeze by sleep position. However, this was not possible because of low numbers: only 3 children who slept supine on a nonsynthetic pillow were reported to have frequent wheeze.

Lung Function

The median FE[V.sub.1] in this sample was 1.44 L (inter-quartile range: 1.30 1.59 L). The results in Table 4 have been adjusted for child age, gender, height, distance run, ambient humidity during exercise, family history of asthma, child exposure to active smoking, and whether asthma medications were taken within 6 hours of testing. Increasing wheeze frequency over the past year was clearly associated with lower postexercise FE[V.sub.1] measures. Children with any wheeze episodes over the past year had a lower FE[V.sub.1] postexercise than children with no wheeze. Among children who slept nonsupine, synthetic quilt use was not associated with a significant FE[V.sub.1] reduction postexercise, compared with children not using this item. However, among children who did sleep supine, synthetic quilt use was associated (P = .05) with a lower postexercise FE[V.sub.1] compared with supine sleepers under feather or other bedding (Table 4). Thus, an adverse effect for synthetic quilt use, particularly among supine sleepers, was observed for this measure of lung function, similar to the findings reported above for frequent wheeze.


An interaction between quilt use and sleep position on respiratory function was observed. There was an adverse effect associated with synthetic quilt use on frequent wheeze and postexercise lung function among children who slept supine but not among children who did not sleep supine. It is of relevance to study the interaction between bedding and different sleep positions, an approach we have used previously in sudden infant death research. (26) For the issue of overlying synthetic quilts and child wheeze, if the adverse effect was mediated through some mechanism that involved the bedding being in close proximity to the airway (such as, but not restricted to, HDM allergen transfer), one would postulate that the association between overlying synthetic quilt use and adverse respiratory outcomes would be observed specifically among children sleeping supine, during which the face would be closer to overlying bedding, as was observed in the present study. In contrast, the adverse effect of an overlying synthetic quilt would be less evident for side- or prone-sleeping children, because other bedding items such as the mattress would be closer to the airway and thus relatively more important than any overlying items such as quilts. A diagrammatic representation of this is shown in Fig 1.


The strengths of this study are a high participation rate, the availability of lung-function and symptom-report data, and the use of frequent wheeze as a study outcome. Frequent wheeze is a better marker of HDM allergen-related wheeze than the global term “asthma,” because HDM atopy is strongly related to frequent wheeze. (19) The use of a nested case-control analysis with matching on underbedding reduced the likelihood that underbedding variability contributed to the observed interaction. The cross-sectional design study does not limit the study materially, because the postulated interaction is based on short-term disease processes. Synthetic bedding has been postulated to induce adverse effects through HDM allergen loading, (7) the release of volatile organic compounds, (18) or altered endotoxin exposure. (13) Bronchial provocation with aeroallergen can produce either an immediate or late response with wheeze and FE[V.SUB.1] reduction. (27,28) Volatile organic compounds can induce a decline in FE[V.SUB.1] in asthmatic subjects within 1.5 hours of exposure. (29) Inhaled endotoxin can produce a dose-related bronchoconstriction within 6 hours. (30) Thus, bedding-induced bronchoconstriction could occur over a short time period. A potential weakness of the study is that frequent wheeze, bedding, and sleep position were based on parental report only. The report of wheeze frequency showed good agreement with measures of postexercise lung function. Additionally, the findings for synthetic quilts and lung function were similar to those for frequent wheeze. The lung-function decrement associated with synthetic quilt use among supine sleepers was small, but it was two thirds the magnitude of the decrement associated with recent wheeze and thus may be of clinical relevance. Previously, the parent report of infant bedding has been shown to agree well with interviewer observations at home interview. The validity of a parental report of usual sleep position in children of this age group has not been established. Child sleep position is likely to change during the night, (31) which could lead to misclassification of usual sleep position. This misclassification would be likely to be nondifferential and thus reduce the study’s ability to detect sleep-position effects.

However, the important issue is that the interaction between synthetic bedding and sleep position cannot be explained by such misclassification. Sleep position was classified as a binary factor (supine/ nonsupine), and both differential and nondifferential misclassification of a binary environmental factor will, if it has any effect, bias a multiplicative interaction effect toward the null value when the following conditions are met. (32) First, the 2 binary factors should be independent. The lack of association between synthetic quilt use and supine sleeping in Table 2 satisfies the criteria that the 2 binary factors should be independent. (32) A second condition is that the measures must classify the true exposures better than random. (32) Past validation work shows a high level of agreement between parental report and nurse observation of infant bedding. (13) Thus, parental report is likely to classify bedding correctly, certainly better than random. To our knowledge, the parental report of child usual sleep position at age 7 has not been validated. However, findings in this study related to sleep position and electric blanket use indicate that parental classification of usual sleep position is not random. Children sleeping on electric blankets were classified by parents as significantly more likely to sleep supine, consistent with past thermal work on sleeping infants. (33) That is, the sleep-position data do provide a pattern of results consistent with past work on thermal balance, (33) which would predict that children with electric blankets would avoid the prone position, which provides a reduced capacity to lose body heat. Thus, parental report, although it may not classify all sleep positions correctly, is likely to reflect the actual predominant sleeping position.

There is growing evidence that synthetic bedding materials are associated with increased childhood wheeze. (9-11) It is now a high priority to establish whether the association is causal in nature. Previous studies have shown that the association between synthetic upperbedding and severe wheeze has a high strength of association, (13,18) a dose-response relationship, (13) biological plausibility, (7) and ecological coherence. (9-11) Two recent birth cohorts have reported that a prospective relationship is evident. (12,13) However, experimental evidence is not available and may be difficult to obtain because of the ethics of applying a potentially deleterious exposure (synthetic bedding) to children with asthma. One of the counterarguments against a causal role for synthetic bedding with childhood wheeze was that the association may merely reflect selection bias even before wheeze development (ie, synthetic bedding was preferentially selected for children at risk of subsequent wheeze). However, the finding that the adverse effect of synthetic quilts was restricted to supine-sleeping children only is evidence against this, because it is unlikely that this selection bias differs by sleep position. This is supported by the findings of no interaction between sleep position and the 2 bedding items (sheepskin and allergen-occlusive mattress cover) that do seem to reflect selection according to the asthma recommendations existing at the time of the study. In 1997, Australian asthma recommendations were to avoid sheepskins and use impermeable mattress covers. (25) The interaction reported here provides additional support for a causal interpretation, (34) because the adverse effect of synthetic quilts was most evident in children who would be more likely to be sleeping face-up near the quilt. In contrast, as predicted, the adverse effect of an overlying synthetic quilt was less evident for side- or supine-sleeping children, because other bedding items such as the pillow or mattress would be closer to the face and thus relatively more important than overlying quilts. The exact mechanisms that underlie this interaction are not understood yet. An increasing focus on the bedding environment immediately adjacent to the nose and mouth of children during sleep is required.

TABLE 1. Distribution of Sleeping Environment Characteristics and

Respiratory Wheeze in the 1995 Cross-Sectional Study

% With

Characteristic Total, n n Variable

Usual sleep position 6355

Supine (on back) sleep position 2431 38.3

Nonsupine (on side or tummy) 3048 48.0

No usual sleep position or other 876 13.8

Overbedding 6345

Any synthetic quilt use 2649 41.7

Feather quilt use 2941 46.4

Other overbedding 755 11.9

Pillow 6340

Any synthetic pillow 5709 90.0

Feather pillow 231 3.6

Other pillow 400 6.3

Sheepskin underbedding 6307 1331 21.1

Electric blanket 6307 1347 21.4

Allergen-occlusive mattress cover 6307 54 0.9

Foam mattress 6363 2038 32.0

Bed type 6354

Bottom bunk bed 686 10.8

Wheeze frequency over the past year 6107

None (0 episodes) 4707 77.1

Moderate (1-12 episodes) 1264 20.7


TABLE 2. The Association Between Sleeping Environment, Child

Characteristics, and Frequent Wheeze or Supine Sleep Position in

the Tasmanian Asthma Survey

Characteristic (Reference Group) OR (95% CI)

for Frequent


Sleeping environment

Usual supine position (usual side or

prone) 0.90 (0.61, 1.33)

Synthetic quilt (feather quilt) 1.74 (1.20, 2.51)

Other overbedding (feather quilt) 1.37 (0.78, 2.39)

Synthetic pillow (feather pillow) 3.24 (1.03, 10.24)

Other pillow (feather pillow) 2.59 (0.57, 11.73)

Sheepskin underbedding (no sheepskin) 0.54 (0.33, 0.90)

Electric blanket (no electric

blanket) 1.56 (1.07, 2.29)

Foam mattress (no foam mattress) 1.35 (0.95, 1.92)

Bottom bunk bed (other type of bed) 1.76 (1.08, 2.88)

Overbedding by sleep position

Among supine sleepers, synthetic

(feather) quilt 3.75 (l.78, 8.63) *

Among nonsupine sleepers, synthetic

(feather) quilt 1.12 (0.65, 1.93) *

Among supine sleepers, other over-

bedding (feather) quilt 2.76 (0.89, 7.90) ([dagger])

Among nonsupne sleepers, other over-

bedding (feather) quilt 0.99 (0.39, 2.21) ([dagger])

Child characteristics

Male sex 1.31 (0.93, 1.85)


Breastfed in infancy 0.61 (0.41, 0.92)

No siblings 1.31 (0.66, 2.62)

Family history of asthma 5.43 (3.50, 8.41)

Eczema 3.69 (2.60, 5.23)

Hay fever 9.05 (6.37, 12.86)

Characteristic (Reference Group) OR (95% CI)

for Supine

Sleep Position

Sleeping environment

Usual supine position (usual side or

prone) …

Synthetic quilt (feather quilt) 1.03 (0.92, 1.16)

Other overbedding (feather quilt) 0.92 (0.77, 1.10)

Synthetic pillow (feather pillow) 0.98 (0.78, 1.23)

Other pillow (feather pillow) 1.26 (0.88, 1.81)

Sheepskin underbedding (no sheepskin) 1.06 (0.93, 1.20)

Electric blanket (no electric

blanket) 1.17 (1.02, 1.33)

Foam mattress (no foam mattress) 1.21 (1.08, 1.36)

Bottom bunk bed (other type of bed) 0.96 (0.81, 1.15)

Overbedding by sleep position

Among supine sleepers, synthetic

(feather) quilt —

Among nonsupine sleepers, synthetic

(feather) quilt —

Among supine sleepers, other over-

bedding (feather) quilt —

Among nonsupne sleepers, other over-

bedding (feather) quilt —

Child characteristics

Male sex 1.07 (0.96, 1.20)


Breastfed in infancy 0.93 (0.81, 1.06)

No siblings 0.95 (0.75, 1.21)

Family history of asthma 1.03 (0.92, 1.15)

Eczema 1.12 (0.98, 1.28)

Hay fever 1.11 (0.97, 1.27)

* Difference in effect by sleep position; P = .007.

([dagger]) Difference in effect by sleep position; P = .10.

TABLE 3. The Risk of Frequent Wheeze for Various Combined Sleep

Arrangements Compared With Sleeping Supine Under a Feather


Combined Sleeping Arrangement Cases, % Matched

(n = 117) Controls, %

Sleep Position Overbedding (n = 1144)

Supine Feather quilt 8.5 20.6

Supine Other overbedding 6.0 4.9

Supine Synthetic quilt 26.5 17.7

Nonsupine Feather quilt 27.4 26.0

Nonsupine Other overbedding 6.8 6.5

Nonsupine Synthetic quilt 24.8 24.4

Combined Sleeping Arrangement Matched OR * P Value

(95% CI)

Sleep Position Overbedding

Supine Feather quilt 1.00 (reference) —

Supine Other overbedding 2.88 (0.83, 10.03) .10

Supine Synthetic quilt 2.67 (1.23, 5.80) .01

Nonsupine Feather quilt 1.87 (0.86, 4.08) .12

Nonsupine Other overbedding 6.84 (0.82, 56.75) .08

Nonsupine Synthetic quilt 2.38 (1.09, 5.18) .03

* Cases matched to controls with regard to age, gender, foam mattress

use, electric blanket use, and sheepskin use.

TABLE 4. Change in FEV, Postexercise by Wheeze Frequency and Sleep


Factor No. of Children

With Lung-Function

Tests and Data Available

Number of wheeze episodes in the past year

None (reference) 320

1-3 48

4-12 16


Supine sleep position 349

Synthetic quilt use 406

For children with nonsupine sleep position,

synthetic quilt use 196

For children with supine sleep position,

synthetic quilt use 151

Factor FE[V.sub.1] Change in mL


(95% CI)

Number of wheeze episodes in the past year

None (reference) 0% (reference)

1-3 -5.5 (-8.80 to -2.2)

4-12 -5.5 (-11.1 to 0.05)


Supine sleep position -0.1 (-2.6 to 2.4)

Synthetic quilt use -1.3 (-3.6 to 1.00)

For children with nonsupine sleep position,

synthetic quilt use -1.0 (-4.7 to 2.8)

For children with supine sleep position,

synthetic quilt use -3.6 (-7.0 to 0.03)

Factor P Value

Number of wheeze episodes in the past year

None (reference) —

1-3 .001

4-12 .05


Supine sleep position .94

Synthetic quilt use .27

For children with nonsupine sleep position,

synthetic quilt use .61

For children with supine sleep position,

synthetic quilt use .05

Lung-function measures were adjusted for child height, age, gender,

distance run, ambient humidity during exercise, family history of

asthma, child exposure to active smoking, and whether asthma

medications were taken within 6 hours of testing. Children with no

usual sleep position were excluded. Children using G-agonists on the

day of testing were excluded (n = 6).


The Public Health Research and Development Committee of the National Health and Medical Research Council (Australia) funded the 1995 Tasmanian study. The analysis of this project was funded partly by a grant from Blundstone’s Pty Ltd and a Coles Supermarket grant to the Canberra Region Medical Foundation. Part of the analysis of this project was funded by a grant from the National Priority Areas Initiative (Asthma), Department of Health and Aged Care (Australia).

We thank the parents, families, infants, and children who participated in these studies; the research staff for data collection and collation; and the hospitals participating in the infant cohort study. We thank the participating schools, the Department of Education, the Department of Cultural and Community Development, and the Catholic Education Office. We thank the Asthma Foundation of Tasmania for the equipment loan.


(1.) Jalaludin B, Xuan W, Mahmic Peat J, Tovey E, Leeder S. Association between Der p 1 concentration and peak expiratory flow rate in children with wheeze: a longitudinal analysis. J Allergy Clin Immunol. 1998;102: 382-386

(2.) Chan-Yeung M, Manfreda J, Dimich-Ward H, et al. Mite and cat allergen levels in homes and severity of asthma. Am J Respir Crit Care Med. 1995;152:1805-1811

(3.) Vanlaar CH, Downs SH, Mitakakis TZ, et al. Predictors of house-dust-mite allergen concentrations in dry regions in Australia. Allergy. 2001; 56:1211-1215

(4.) Abbott J, Cameron J, Taylor B. t louse dust mite counts in different types of mattresses, sheepskins, and carpets, and a comparison of brushing and vacuuming collection methods. Clin Exp Allergy. 1981;11:589-595

(5.) Sakaguchi M, Inouye S, Yasueda H, Shida T. Concentration of airborne mite allergens (Der 1 and Der II) during sleep. Allergy. 1992;47:55-57

(6.) Mills S, Siebers R, Wickens K, Crane J, Purdie G, Fitzharris P. House dust mite allergen levels in individual bedding components in New Zealand. N Z Med J. 2002;115:151-153

(7.) Kemp TJ, Siebers RW, Fishwick D, O’Grady GB, Fitzharris P, Crane J. House dust mite allergen in pillows. BMJ. 1996;313:916

(8.) Hallam C, Custovic A, Simpson B, Houghton N, Simpson A, Woodcock A. Mite allergens in feather and synthetic pillows. Alh, rgy. 1999;54: 407-408

(9.) Fitzharris P, Siebers R, Crane J. Pillow talk: have we made the wrong beds for our patients to lie in? Clin Exp Allergy. 1999;29:429-432

(10.) Custovic A, Woodcock A. Feather or synthetic? That is the question. Clin Exp Allergy. 1999;29:144-147

(11.) Siebers R. Feather bedding and allergic disease in children: a cover story? Clin Exp Allergy. 2002;32:1119-1123

(12.) Nafstad P, Nystad W, Jaakkola JJ. The use of feather quilt, childhood asthma and allergic rhinitis: a prospective cohort study. Clin Exp Allergy. 2002;32:1150-1154

(13.) Ponsonby AL, Dwyer T, Kemp A, Cochrane J, Couper D, Carmichael A. Synthetic bedding and wheeze in childhood. Epidemiology. 2003;14: 37-44

(14.) Worldwide variation in prevalence of symptoms of asthma, allergic rhinoconjunctivitis, and atopic eczema: ISAAC. The international Study of Asthma and Allergies in Childhood (ISAAC) Steering Committee. Lancet. 1998;351:1225-1232

(15.) Jenkins MA, Clarke JR, Carlin JB, et al. Validation of questionnaire and bronchial hyperresponsiveness against respiratory physician assessment in the diagnosis of asthma. Int J Epidemiol 1996;25:609-616

(16.) Ponsonby A-L, Couper D, Dwyer T, Carmichael A, Wood-Baker R. Exercise-induced bronchial hyperresponsiveness and parental ISAAC questionnaire responses. Eur Respir J. 1996;9:1356-1362

(17.) Peat JK, Tovey E, Gray EJ, Mellis CM, Woolcock AJ. Asthma severity and morbidity in a population sample of Sydney schoolchildren: Part II–Importance of house dust mite allergens. Aust N Z J Med. 1994;24: 270-276

(18.) Strachan DP, Carey IM. Home environment and severe asthma in adolescence: a population based case-control study. BMJ. 1995; 311 (7012):1053-1056

(19.) Ponsonby A-L, Kemp A, Dwyer T, Carmichael A, Couper D, Cochrane J. Feather bedding and house dust mite sensitization and airway disease in childhood. J Clin Epidemiol. 2002;55:556-562

(20.) Hosmer DW, Lemeshow S. Applied Logistic Regression. Second ed. New York, NY: John Wiley & Sons Inc; 2000

(21.) Breslow NE, Day NE. Statistical methods in cancer research. Volume I–The analysis of case-control studies. IARC Sci Publ. 1980;(32):5-338

(22.) Kupper LL, Hogan MD Interaction in epidemiologic studies. Am J Epidemiol. 1978;108:447-453

(23.) Koopman JS. Interaction between discrete causes, Am J Epidemiol. 1981; 113:716-724

(24.) Johnston ID, Strachan DP, Anderson HR. Effect of pneumonia and whooping cough in childhood on adult lung function. N Engl J Med. 1998;338:581-587

(25.) NSW Health Department. Asthma and the Environment: Perspectives on the Prevention of Asthma Sydney, Australia: NSW Health Department; 1997

(26.) Ponsonby AL, Dwyer T, Gibbons LE, Cochrane JA, Wang YG. Factors potentiating the risk of sudden infant death syndrome associated with the prone position. N Engl J Med 1993;329:377-382

(27.) Calhoun WJ, Dick EC, Schwartz LB, Busse WW. A common cold virus, rhinovirus 16, potentiates airway inflammation after segmental antigen bronchoprovocation in allergic subjects. J Clin Invest. 1994;94:2200-2208

(28.) Cockcroft DW, Ruffin RE, Dolovich J, Hargreave FE. Allergen-induced increase in non-allergic bronchial reactivity. Clin Allergy. 1977;7:503-513

(29.) Harving H, Dahl R, Molhave L. Lung function and bronchial reactivity in asthmatics during exposure to volatile organic compounds. Am Rev Respir Dis. 1991;143:751-754

(30.) Rylander R, Bake B, Fischer JJ, Helander IM. Pulmonary function and symptoms after inhalation of endotoxin. Am Rev Respir Dis. 1989;140: 981-986

(31.) De Koninck J, Lorrain D, Gagnon P. Sleep positions and position shifts in five age groups: an ontogenetic picture. Sleep. 1992;15:143-149

(32.) Garcia-Closas M, Rothman N, Lubin J. Misclassification in case-control studies of gene-environment interactions: assessment of bias and sample size. Cancer Epidemiol Biomarkers Prev. 1999;8:1043-1050

(33.) Sawczenko A, Fleming PJ. Thermal stress, sleeping position, and the sudden infant death syndrome. Sleep. 1996;19(10 suppl):S267-S270

(34.) Weiss NS. Can the “specificity” of an association be rehabilitated as a basis for supporting a causal hypothesis? Epidemiology. 2002;13:6-8

Anne-Louise Ponsonby, PhD, FAFPHM * ([double dagger]); Terence Dwyer, PhD, FAFPHM ([double dagger]); Leigh Trevillian, MPH *; Andrew Kemp, PhD, FRACP ([section]); Jennifer Cochrane, BA ([double dagger]) David Couper, PhD ([parallel]); and Allan Carmichael, MD, FRACP ([paragraph])

>From the * National Centre for Epidemiology and Population Health, Australian National University, Canberra ACT, Australia; ([double dagger]) Menzies Centre for Population Health and ([paragraph]) Department of Paediatrics and Child Health, University of Tasmania, Tasmania, Australia; ([parallel]) Department of Biostatistics, University of North Carolina, Chapel Hill, North Carolina; and ([section]) Department of Immunology, Royal Children’s Hospital, University of Melbourne, Melbourne, Australia.

Received for publication Jan 9, 2003; accepted Jul 14, 2003.

Address correspondence to Anne-Louise Ponsonby, PhD, FAFPHM, National Centre for Epidemiology and Population Health, Mills Rd, Australian National University, Canberra ACT 0200, Australia. E-mail:

COPYRIGHT 2004 American Academy of Pediatrics

COPYRIGHT 2004 Gale Group