Risk and preventive factors for cot death in The ... - CiteSeerX

Eur J Pediatr (1998) 157: 681 ± 688

Ó Springer-Verlag 1998

PREVENTIVE PEDIATRICS AND EPIDETIOLOGY

M. P. L'Hoir á A. C. Engelberts á G. Th. J. van Well á S. McClelland á P. Westers T. Dandachli á G. J. Mellenbergh á W. H. G. Wolters á J. Huber

Risk and preventive factors for cot death in The Netherlands, a low-incidence country

Received: 16 July 1997 / Accepted 12 revised form: 8 December 1997

Abstract In the Netherlands an 18 months case control study into cot death was undertaken as part of the European Concerted Action (ECAS) on sudden infant death syndrome to determine the relative risk of prone sleeping and other sleep practices. Physicians in the Netherlands were asked to report to the study centre all sudden and unexpected deaths of children between 1 week and 2 years of age. Non cot death cases were deleted from further analysis after a consensus was reached by three pathologists, not primarily involved in the post mortem diagnosis. A positive response of families was achieved in 91% of cases registered in the Central Bureau of Statistics. The study comprised 73 cot deaths and 146 controls, two for each case and matched for date of birth. All families were visited at home for completion of a questionnaire. The cot death rate has dropped considerably over the past 10 years after the recommendations on supine sleeping to a low of 0.26 per 1000 live born infants. In addition to the ECAS objective, we wanted to establish whether previously found risk factors are still valid in the present situation or that new factors might have emerged, some of them possibly protective. Conclusion Placing an infant prone or on side on last occasion, secondary prone position (not placed prone but turned to prone), inexperienced prone sleeping and use of a duvet, leading to head and body being covered, were shown to be risk factors. Preventive factors were using a cotton sleeping-sack and a dummy. Even in a low incidence country, such as the Netherlands, there are indications that further prevention is possible. Key words Cot death á Sudden infant death syndrome á Prone position Abbreviations ECAS European Concerted Action on SIDS á CBS Dutch Central Bureau of Statistics á FSD face straight down á SID sudden infant death

M. P. L'Hoir (&) á G. Th. J. van Well á S. McClelland T. Dandachli W. H. G. Wolters á J. Huber Wilhelmina Children's Hospital, PO Box 18009, 3501 CA Utrecht, The Netherlands, Fax: 31-30-2320707 A.C. Engelberts University Hospital Leiden, Department of Paediatrics, Leiden, The Netherlands

P. Westers Centre for Biostatistics, University of Utrecht, Utrecht, The Netherlands G. J. Mellenbergh Department of Methodology, Faculty of Psychology, University of Amsterdam, Amsterdam, The Netherlands

682

Introduction

Methods

Cot death is the sudden unexplained death of an infant or young child. When a full post mortem examination fails to demonstrate an adequate cause of death, the term sudden infant death (SID) is used. We leave out the ax ``syndrome'' since there are no signs and symptoms and the clinical death is, on the contrary, characterized by the absence of such. Much research into cot death is focused on epidemiological risk factors which when avoided might diminish the cot death rate. One important risk factor is the prone sleeping position. Although suggested previously, it was not until the late 1980s that it came to the fore [1±4]. In the Netherlands, the traditional sleeping position used to be the alternate side position. In 1972 advice to place infants prone was followed widely after publications in the medical and lay press [5]. At the end of October 1987 G. A. de Jonge initiated a nationwide campaign not to place infants prone to sleep. No preference was given as to side or supine position. Although cot death rates had begun to decline from a high in 1984, rates in 1987 further decreased from 0.91 to an alltime low of 0.26/1.000 in 1995 [6]. In the same period no signi®cant diagnostic shifts in statistical registration occurred [7]. In April 1992 the advice was modi®ed from ``not prone'' to sleep exclusively ``supine''. From 1994 onwards additional recommendations were given, namely discouraging the use of duvets, pillows and `cot-bu€ers'. In the Netherlands duvets had become popular in the early part of the 1970s. In 1988 60%, in 1994 77% and in 1996 45% of the infants slept under a duvet [8]. Prevalence studies, carried out every other year, showed that putting infants to sleep in the prone position declined from 60% before the campaign to 27% in 1988, to 7% in 1996. However, many babies are presently found in prone position after having been put to bed on the side or back. In two earlier cot death studies in the Netherlands in 1984±1987 and 1987±1992 respectively, 11% and 27% of infants who were found in the prone position had turned to that position, thus showing the secondary prone position to be of increased importance in cot death as opposed to the primary prone position [9, 10]. The cot death incidence in the Netherlands is presently one of the lowest in the Western industrialized world. However, international cot death ®gures are dicult to compare because of problems with de®nition and ascertainment [11]. With such a low cot death incidence the question arises which of the previously known factors associated with cot death are still valid and whether they are amenable to preventive strategies. The present study was conducted as part of the `European Concerted Action on SIDS' (ECAS) [12]. The aim of this article is to examine relative risks of placing infants prone and other practices associated with putting infants to sleep, including bedding and heating.

Study design A case control design was chosen. All general practitioners, (community) paediatricians, pathologists and midwives in the Netherlands were informed by letter about the project. They were requested to inform the research centre as soon as possible after an infant died suddenly and unexpectedly. Announcements appeared in more than 15 professional journals and a national press report was released. Two controls for every case were randomly selected, matched for date of birth. Case ascertainment We intended to include all cot death cases in the Netherlands from March 1995 to September 1996. Infants older than 7 days and younger than 2 years who died suddenly and unexpectedly and whose death was unexplained when ®rst found, were included. Not included were premature infants who had never left hospital and babies explicitly taken home to die. The de®nitely non-SID cases were excluded after three pathologists ®rst independently and later in consensus, had made their diagnosis. In the Netherlands all deaths are classi®ed anonymously at the Dutch Central Bureau of Statistics (CBS). Date of birth and date of death of cot death infants were compared with CBS data to assess completeness of ascertainment. Control ascertainment Two live controls were obtained for each case, born within 1 week before or after the case. They were selected by the municipal authority in whose district the death had occurred and two from the list of births in the nearest large urban area (Amsterdam, Rotterdam, Utrecht or The Hague). Ideally we would not have liked to match by geographical area or place of residence; however, in the Netherlands, for practical purposes, this was unavoidable. An earlier Dutch study has shown that there was no di€erence in the incidence of cot death and sleeping position in respect of a di€erent degree of urbanisation [9]. Selecting controls only from municipality of birth could possibly have introduced selection bias: in the rural areas there are fewer ethnic minorities. In case of more than two replies, two were randomly selected. Questionnaire The ECAS questionnaire consisted of 228 questions and was separated into subjects concerning: site of death, factors related to the health of the baby, maternal factors and factors related to socioeconomic status. The Dutch version included several additional questions among which: movement in bed, position of the face when found prone and use of a sleeping-sack (Fig. 1). All questionnaires taken from parents of cases and controls were completed at their home by six interviewers (two researchers and four medical students) who received repeated special training. The questions for the control-parents all referred to the day/night before the interview. Reference sleep is de®ned as the sleeping period of the control baby that corresponds to the period during which the index baby died as well as the sleeping period on last occasion of the cases. Many questions refer to infant care practices occurring `usually' which was de®ned as the usual routine for the time of the reference sleep.

683 Statistical analysis Data were entered with EPI-info and analysed with the Statistical Package for Social Sciences. Binomial tests were used for comparison with national data of the CBS and chi-square tests and correlation analyses for associations between the factors for cases and controls separately. Chi-square tests and t-tests were used to compare case and control groups. Logistic regression analyses were used to measure the risk of factors adjusted for confounders. Well known risk factors, established from earlier Dutch retrospective studies [7±10], literature review and ongoing international research [12, 13], were considered as confounders: infant's age, parity, mother's age at ®rst live birth, birth weight, sex, socioeconomic status, and maternal smoking during pregnancy. Birth weight and gestational age were closely related and e€ect modi®cation between these was investigated. After a factor analysis was performed, socioeconomic status was de®ned by several markers including schooling and housing. Due to the time interval between death of the case and the control interview, conditional regression analysis was dicult to apply. We decided to use unconditional logistic regression analysis after comparison of conditional and unconditional logistic regression analyses with and without the factor `age', which did not result in di€erences in the standard error. It has been demonstrated in cot death studies that conditional and unconditional logistic regression analyses give similar results [14]. In the logistic regression analyses age at reference sleep was adjusted for. A stepwise logistic regression analysis was used to determine the independent e€ect of the factors concerning bedding and heating in relation to sleeping position. We built the model with the confounders and the determinants for cot death found in univariate analyses. In the tabulations `not applicable' and `missing' data are excluded from analyses. All P-values are two-tailed, unless stated otherwise. Relative risks are expressed in odds ratios (OR) with con®dence intervals of 95% (CI 95%).

Results

Response cases Of 105 infants noti®ed at the research centre, 6 were excluded according to the criteria and 11 families refused participation. Of these 11 families the age distribution did not di€er from the response group, there were 8 boys and 1 girl (2 unknown), 3 belonged to an ethnic minority group and 2 had a strict religious background. Two nonresponse families were known to be drug users. In the

other 88 cases a home-visit was made and the questionnaire was completed. Postmortem was obtained in 63 of the 88 cases. After the consensus meeting of the three pathologists, 14 cases were excluded because of major pathology. The remaining 49 cases, together with 25 cases without postmortem were included in the study; altogether 74 cases. The mean age was 7 months (SD ˆ 5.10) and for cases