Association between use of a quilt and sudden infant death syndrome: case-control studyBMJ 1998; 316 doi: http://dx.doi.org/10.1136/bmj.316.7126.195 (Published 17 January 1998) Cite this as: BMJ 1998;316:195
- Anne-Louise Ponsonby, deputy director of research ()a,
- Terence Dwyer, directora,
- David Couper, biostatisticiana,
- Jennifer Cochrane, research officera
- aMenzies Centre for Population Health Research, University of Tasmania, GPO Box 252-23, Hobart, Australia 7001
- Correspondence to: Dr Ponsonby
- Accepted 11 June 1997
The relation between an infant's sleeping environment and development of the sudden infant death syndrome depends on the infant's sleep position.1 We report how the association between the use of a quilt and the syndrome depends on sleep position.
Subjects, methods, and results
Between 1 October 1988 and 31 December 1995 in Tasmania 107 infants <1 year old died of the sudden infant death syndrome. Of the families affected, 100 (93%) participated in this study and were compared directly with 196 age matched controls. Methods are described elsewhere.1 A quilt is a coverlet made by stitching two thicknesses of fabric together with a filling (usually synthetic) enclosed between the layers. Conditional multiple logistic regression2 was used to evaluate interaction effects; egret 0.26.6 software (Cytel Software, Cambridge, MA, USA) provided matched odds ratios with logit-based 95% confidence intervals.
An adverse effect of quilt use was evident in infants who did not sleep prone but not in infants who slept prone (1). This interaction was not altered by adjustment for sleeping on sheepskin; the interaction between sleeping on sheepskin and sleeping prone; mattress liner; mattress type; use of a quilt under the infant; use of a pillow; infant illness; heating in the bedroom; amount of thermal insulation over the infant; maternal age; or family history of asthma. Adjustment for maternal smoking, breast feeding, or swaddling increased the size of the interaction.
A model was constructed to test whether the interaction was a marker for one of the previously identified modifiers of the prone position,1 with indicators for maternal smoking, use of a quilt, swaddling during last sleep, use of a natural fibre mattress, infant illness, and use of bedroom heating during last sleep; the interaction between the last five variables and the prone position was also evaluated. The interaction effect remained (P=0.01). An adverse effect of the use of a quilt was found among infants who did not sleep prone (adjusted odds ratio 6.16; 2.01 to 18.87) but not among those who slept prone. Thus the interaction could not be explained by these factors. The interaction between sleep position and use of a quilt remained for infants and controls who slept in their own cot during the index sleep.
The adverse effect of quilt use among infants who did not sleep prone was greater in infants 12 weeks old (odds ratio 6.77; 1.14 to 40.21) than in younger infants (odds ratio 1.62; 0.27 to 9.59). Adjustment for being found prone reduced the effect of quilt use (adjusted odds ratio 3.16; 1.16 to 8.60) suggesting that part of the adverse effect of quilts may occur by allowing infants to roll into the prone position. Among infants who did not sleep prone adjustment for swaddling, maternal smoking postnatally, breast feeding, and type of mattress gave an odds ratio of 5.63 (1.69 to 18.72) for quilt use. Review of the data at the infant's death scene showed that infants found supine with facial obstruction were reported by at least one witness to have had an overlying quilt. The risk of the sudden infant death syndrome for quilt use varied by usual sleep position (common odds ratio test P=0.017) and was highest for infants found supine. Unconditional analysis showed that quilt use attenuated the protective effect of sleeping supine compared with sleeping on the side (common odds ratio test, P=0.03).
Our data show that the use of a quilt increases the risk of the sudden infant death syndrome particularly among older infants who sleep supine or on their side; this is related to obstruction of the face by the quilt for some infants. At 16 weeks infants are able to pull clothing over their face.3 Infants >3 months old who sleep on their back or their side may be particularly likely to pull loose bedding over their faces; infants sleeping prone are less likely to be able to do this. The risk of the sudden infant death syndrome when a quilt is used (odds ratio 2.82; 1.95 to 4.08) is reduced after adjustment for factors that include the intermediate state of infants being found with their head under the covers.4 Duvets and quilts often do not tuck in because of their design.5 The recommendation that quilts should not be used in babies <1 year old is important for those who sleep supine or on their side.3
We thank the parents and infants who participated in this study; the professional officers of the Tasmanian Ambulance Service for collecting data at the scene of death; the Departments of Hospital Pathology, Police, and Justice for their cooperation; Ego Pharmaceuticals for providing presents for control families after the home interview.
Funding: Tasmanian State Government, Australian Rotary Health Research Fund, National Health and Medical Research Council, National SIDS Council of Australia and its constituent organisations including the Sudden Infant Death Research Foundation, and Tasmanian Sanatoria After-Care Association. A-LP holds a National Health and Medical Research Council public health research and development fellowship.
Conflict of interest: None.
Contributors: A-L P initiated the study, designed the protocol, coordinated the collection of data, and undertook the statistical analysis, and participated in writing the paper. TD participated in the design of the study, was involved in the analysis and interpretation of data, and participated in writing the paper. DC participated in the statistical analysis and interpretation of data, and contributed to the writing of the paper. JC coordinated the fieldwork, data documentation, and quality control; participated in interpreting the data; and contributed to the writing of the paper. Allan Carmichael provided paediatric advice. John Williams provided computing advice. Helen Bain, Shirley Cameron, and Dianne McCracken were the principal research interviewers for collection of data. Janelle Booth prepared the manuscript for publication and participated in the literature search. A-L P is guarantor for the paper.