Letters

Babies sleeping with parents and sudden infant death syndrome

Invoking sudden infant death syndrome in cosleeping may be misleading

Down with smoking and babies sleeping in separate rooms

Smoking may be residual confounder in bed sharing

Authors' reply

Invoking sudden infant death syndrome in cosleeping may be misleading

EDITORBlair et al investigated the factors influencing the risk of the sudden infant death syndrome.¹ In an increasing number of infant deaths that we have investigated over recent years we have found several recurring themes: infant under 3 months; shared sleeping arrangements, particularly sharing a sofa; young carer; consumption of alcohol the night before the infant was found dead. Typically, the infant is found between the adult and the back of the sofa, often covered by a duvet. Sharing a sofa seems to be particularly common in single parent households with poor socioeconomic support, in which the mother sleeps on the sofa, often with more than one child, simply because it is the warmest place in the house. Some of the risk factors highlighted by Blair et al have long been recognised in law in the Children and Young Persons Act 1933.

We are concerned with the use of the term sudden infant death syndrome in this paper because it is potentially misleading in deaths associated with cosleeping. In such cases the infant may have been accidentally overlaid, smothered by the bedclothes, or squashed between the adult and back of the sofa. We accept that this cannot, at present, be proved beyond reasonable doubt, but we found in a recent study that intra-alveolar haemorrhage was increased in infants who died in the context of bed sharing and that this may be a marker of accidental asphyxia.² This often manifests as bloodstained fluid issuing from the nose and mouth with consequent bloodstaining of the bedding and infant's clothing. Examination of the clothing and bedding is therefore essential.

It is inappropriate to accuse a carer unjustly in a cosleeping death. Nevertheless, it strikes us as being more honest to raise the suspicion that accidental upper airway obstruction may be a factor in the death and to give the cause of death as unascertained while giving appropriate support to parents.

N Carter, senior lecturer in forensic pathology. 
G N Rutty, senior lecturer in forensic pathology. 
Medicolegal Centre, Sheffield S3 7ES

Down with smoking and babies sleeping in separate rooms

EDITORThe finding of Blair et al that infants who sleep in a separate room are at a significantly increased risk of the sudden infant death syndrome is of critical importance but was not adequately emphasised.¹ This increased risk was about the same as that for infants sharing a bed, even including smokers. Considering that in the industrialised world sleeping in a separate room is much more common than sharing a bed, many lives could be saved by discouraging this separation.

Parents commonly let their babies "cry it out" in their cribs, often in a separate room. Eventually, the infants are conditioned not to cry at all. Many paediatricians promote this suppression of the babies' inborn instinct to cry and the mothers' natural instinct to respond by providing comfort, although there is no scientific evidence to show that this practice is safe or without long term effects.

The long bouts of crying evoke many physiological responses such as increased heart rate, body temperature, blood pressure, respiratory rate, and production of stress hormones. Overheating is especially worrying as this may be a factor in the sudden infant death syndrome.²

Psychologically, infants may feel abandoned by being left alone for extended periods, a practice not shared by any other mammal or most human cultures, which have been practising cosleeping behaviour throughout their evolutionary history.³ Is it really fair, then, to ask our babies to simply shed their natural instincts, learnt over the course of millions of years, as if they were just a bad habit?

The evidence on the adverse effects of smoking during pregnancy and nursing is extensive, making it ludicrous to simply recommend mothers who smoke not to practise bedsharing, as Mitchell states in his commentary.¹ It is time for the medical community to accept some of the responsibility for the many mothers who smoke and to take the lead in curbing it. How many doctors spend time counselling their patients about these dangers?

Edwards showed that exposure to chemicals in air fresheners may have adverse effects on babies.⁴ Perfumes, deodorants, and other products may contain similar chemicals, so they might also have negative impacts, especially in infants who share a bed and are snuggled up close to their mother throughout the night.

Instead of scaring parents, as some current guidelines do,⁵ the medical community should establish recommendations for parents to safely share a bed with their infants.

Cory A Mermer, medical researcher and writer. 
PO Box 3004, Westfield, NJ 07091, USA

Smoking may be residual confounder in bed sharing

EDITORBlair et al provided some valuable evidence in their case-control study on the factors influencing the risk of the sudden infant death syndrome, particularly the dangers of an infant sharing a bed when the parents are smokers.¹ Although they adjusted for parental smoking, we are concerned that it may still act as a residual confounder and hence be an alternative explanation for the apparent increased risk of the syndrome when infants share the parental bed.

In the full multivariate analysis maternal smoking during pregnancy and parental smoking are simply categorised as yes or no variables (table 3). Usual daily postnatal exposure to smoking is crudely estimated as either 0 or 1 or more hours a day. Clearly, exposure can vary hugely in these broad categories. Smoking could still be acting as a confounding variable if parents sharing a bed with their infant were more likely to smoke heavily or for longer periods in the home. Indeed, the authors note: "Among index mothers who smoked, more of those whose infants shared the bed smoked more than 20 cigarettes per day (23.2% v 1.5% control) compared with those who did not bed share (16.6% v 5.9%)."

Similarly in the analysis of the risk of the sudden infant death syndrome by parental smoking and bed sharing (table 4) the infant's tobacco exposure is inadequately described. Parental smoking is adjusted for, but Blair et al do not specify how or whether infant exposure to smoke is included in the analysis.

We are also disturbed by the apparent lack of adjustment for smoking as a confounder in the simplified model (table 5). In particular, this table details high alcohol consumption as a variable. Since high maternal alcohol consumption is often accompanied by a similarly high prevalence of smoking, smoking may be responsible for some or all of the observed association.

We suggest that this paper may have been improved by using a greater number of categories to classify the extent of parental smoking prenatally and postnatally. In addition, an assessment of exposure to tobacco smoke immediately before the infant's last or reference sleep may have been beneficial, rather than relying on the usual daily exposure. We therefore believe that the increased risk of the sudden infant death syndrome in infants aged <14 weeks who share a bed and whose parents smoke is as yet unproved.

Amara Ezeonyeji, stage 3 medical student. 
Steph Jewitt, stage 3 medical student. 
Leigh Poyser, stage 3 medical student. 
Tom Stadward, stage 3 medical student. 
Department of Epidemiology and Public Health, University of Newcastle upon Tyne, Newcastle upon Tyne

Authors' reply

EDITORCause of death was established by a multidisciplinary committee after a full necropsy to a standard protocol including both a clinical history and a detailed description of the precise circumstances of death. Of the sofa sharing infants classified as sudden infant death syndrome, only 20% (4/20) had blood stained fluid from the nose and mouth. Carter and Rutty suggest this is a marker of accidental asphyxia, but it is not uncommon in deaths from the syndrome in which accidental asphyxia was not suspected. Blood stained fluid was also found among 13% (11/85) of infants with the syndrome who were found sleeping alone in a cot in a non-prone position without evidence of head covering (Fisher's exact test, P=0.48). Without further evidence from the necropsy or circumstances of death, the sudden infant death syndrome is the appropriate classification for infants found cosleeping on a sofa unless we wish to travel full circle and again attribute cause of death solely by the proximity of the parent.

The increased risk associated with infants sleeping in a separate room requires further investigation; 63% of the control infants shared the parental room, which should be viewed as the norm as in most societies worldwide. Recommendations for infants to sleep in a separate room from parents owe more to archaic views of child development than to scientific observation. We primarily addressed the issue of bed sharing, so where the infant slept was categorised in terms of the different forms of cosleeping. Room sharing needs to be looked at separately in terms of the time of day the sleep occurred, the type of sleeping place, and who else was present. Our preliminary findings suggest that the risks associated with infants sleeping on an adult bed alone are greater than when a parent is sharing the bed, especially among comparatively older infants who can move under the adult covers.

Infant exposure to tobacco smoke is a strong risk factor for the syndrome but does not in itself explain the disproportionate number of deaths occurring in adult beds. We collected data on the number of cigarettes smoked by the parents and the number of hours of daily infant exposure. In the large multivariate model, adjusting for the different doses of cigarettes smoked or hours of exposure had the same minimal effect on bed sharing as using dichotomous variables. In the more restricted model (controlling for adverse bed sharing conditions) we did not measure recent exposure to tobacco smoke. Adjusting for smoking using a proxy measure of infants usually exposed to high levels of tobacco smoke had little effect on the results. The risk associated with bed sharing cannot be explained by the residual confounding associated with tobacco exposure, but the reduced number of deaths occurring in the adult beds of non-smokers suggests it can only be generalised to infants of parents who smoke.

Peter Blair, medical statistician. 
Peter Fleming, professor of infant health and developmental physiology. 
Institute of Child Health, Royal Hospital for Children, Bristol BS2 8BJ

Martin Ward Platt, consultant paediatrician and senior lecturer in child health. 
Newcastle Neonatal Service, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP