Is co-sleeping putting your baby at risk?
NICE (National Institute for Health Care and Clinical Excellence) is currently updating its guidance on post-natal care for use in the NHS on reducing the risk of sudden infant death syndrome (SIDS). The recommendations below are not clear about the risk involved. There ‘may be an increase in the number of cases of SIDS, and the term ‘association’ has been used to make it clear that there is no cause-effect relationship at stake. But what will parents hear?
Have you co-slept with your baby? have you travelled to places where co-sleeping is the norm?
If you are expecting a baby or if you teach prenatal or baby classes, what do other parents to be and parents think about co-sleeping around you?
La Leche League and Baby Milk Action have offered responses to the NICE recommendation, with important arguments that fit with shared Birthlight views and values. In the early 1990s John McKenna was in Cambridge and he came to our community group to present the results obtained in his sleep-lab in the USA, where mums were observed through the night while co-sleeping with their babies. This research does not seem to have been taken into consideration in the NICE review. The people I lived with in the Peruvian Amazon would not think of sleeping any other way than with their babies, in fact when I travelled alone they often volunteered a small child to sleep next to me!
NICE addendum related to Infant Death Syndrome (July 2014)
The cause of sudden infant death syndrome (SIDS) is not known. It may be that there are many factors contributing to SIDS.
There are some factors which are known to make SIDS more likely such as placing a baby on their front or side.
We need clear evidence to say that a factor directly causes SIDS. This update reviewed evidence relating to co sleeping (parents or carers sharing a bed or sofa or chair with an infant in the first year of an infant’s life
Some of the reviewed evidence showed that there is a statistical relationship between co-sleeping and SIDS. This means that, where co-sleeping occurs there may be an increase in the number of cases of SIDS. However, the evidence does not allow us to say that co-sleeping causes SIDS. Therefore the term ‘association’ has been used in the recommendations in this update to describe the relationship between co-sleeping and SIDS.
Baby Milk Action – DRAFT response to NICE consultation on the addendum to Clinical guideline 37, Postnatal Care (reducing the risk of SIDS)
The draft guidelines appear to have been compiled with care and attention to the limits of the available evidence, in this complex area. We welcome this. Guidelines also mention person-centred care, which we understand means health professionals should take into account the context and circumstances of parents/carers and their infants, their needs and preferences, and personal choice – which we also welcome.
We believe that care of infants and the manner in which they are fed (breast, bottle, or a combination; solids also, once complementary feeding has begin) are inextricably linked. Both infant and carer behaviour are impacted upon by the manner in which a baby is fed; breastfed babies may wake more frequently at night but breastfeeding mothers may cope with this by sleeping close to their babies, thus benefiting from the fact that they do not need to get up to prepare a bottle. Breastfeeding can also help mothers get back to sleep more quickly, due to the effect of oxytocin. Co-sleeping, if breastfeeding, appears to be less risky than if bottle-feeding (due to the protective effect of breastfeeding against SIDS), more beneficial in terms of maternal rest, and possibly harder to avoid. We understand that the guidance sought to address co-sleeping specifically, but find it somewhat problematic that breastfeeding is not mentioned (while parental smoking, for instance, is).
Implementation – and likely implementation of the guidance
In the absence of clear understanding of SIDS aetiology, it seems prudent to inform parents of the various factors associated with SIDS, and not to single-out co-sleeping. We are concerned that by restricting the addendum to co-sleeping, despite the careful wording of the guidance, media and public perceptions may be that co-sleeping is the main risk factor for SIDS - which is not the case.
The guidance groups together co-sleeping and sleeping on a sofa or chair; and also, groups together accidental co-sleeping with intentional co-sleeping. This is perhaps the easiest way of analysing data from studies which used varied and sometimes unclear definitions (a major limitation in the available evidence). However, we suggest it limits the practical applicability of advice parents/carers would be given based on the recommendations.
Despite the careful wording of the draft guidance, we are concerned that in practice, a simple 'no co-sleeping' message may be delivered by health professionals; or health professionals' advice may be interpreted this way by parents/carers, and also the media.
Furthermore we are concerned that implementation of the recommendations may have unintended consequences:
(1) The draft guidelines implicitly assume that information given to parents will impact on their intentions, and that they will be able to put these intentions into practice. As early months with a new baby are often very tiring, falling asleep with the baby may occur unintentionally. We are concerned that parents, in an effort to abide by advice from their health professionals, may get out of bed at night and sit on a sofa/chair to feed their baby and get them back to sleep, and risk falling asleep unintentionally with their baby in this apparently more dangerous setting. Co-sleeping in bed could be a safer option if parents were prepared for this possibility, and could take adequate precautions to make the sleeping environment as safe as possible.
We suggest that it is inappropriate simply to tell parents not to co-sleep, and that this should be clearly stated in the guidance.
(2) Particularly in relation to breastfeeding – which has numerous, lifelong health benefits to babies and their mothers – we are concerned that the guidance does not inform parents about the protective effect of breastfeeding in relation to SIDS. That previous information parents may have received has mentioned that breastfeeding is protective, and this guidance does not, implies that the balance of evidence no longer favours breastfeeding over formula-feeding, yet this is not the case. That other factors related to SIDS, such as parental smoking, are mentioned, seems inconsistent.
In addition, there is a common perception that formula-fed babies sleep for longer, while breastfed infants wake more often (this may be the case as formula is easier to digest than breastmilk). Where mothers are exhausted, we believe the balance of evidence suggests that ceasing breastfeeding and switching to formula-feeding in the hope of gaining a better night's sleep may be more harmful to overall infant health, than continuing exclusive breastfeeding for the first 6 months and co-sleeping with adequate precautions taken (see below: we believe breastfeeding, as a factor protective against SIDS, has been inadequately addressed in Carpenter et al.'s study).
We request that the protective effect of breastfeeding should be mentioned.
We suggest if parents wish to co-sleep, they should be able to make this choice fully informed by the evidence, in order to minimise risk of SIDS (as points 3,4,5 of the guidelines, and also mentioning breastfeeding).
We suggest that all parents should be informed about how to make the sleeping environment as safe as it can be, should parents choose or feel the need to co-sleep. This would be in line with 'person-centred' care but currently is not addressed in the guidelines.
The term 'bed-sharing' requires clarification, as to whether this is sharing a bed to sleep, or (e.g.) breastfeeding in bed while awake. It appears that NICE means 'co-sleeping in bed' but this is not clear. (Ditto, 'sofa-sharing').
'Acknowledge that co-sleeping occurs' – it is unclear what this means: whether this is something health professionals are meant to say to parents, or simply be aware of, themselves? Might it better be rephrased as 'Tell parents/carers that co-sleeping sometimes occurs unintentionally – for instance falling asleep by accident; or deciding on the spur of the moment to bring the baby into bed with them to sleep on a night when they are very tired. Inform parents/carers of steps they can take to minimise risks in these situations. Inform parents that if they choose to co-sleep in bed regularly, they can also take steps to minimise risks.'
We have not reviewed the evidence in detail but are concerned about over-reliance on Carpenter et al.'s analysis. We leave aside from our methodological criticisms the highly offensive analogy drawn between co-sleeping and being a 'sow' at risk of overlaying her 'piglets' – trusting that those creating NICE guidance are free from such prejudices (and, of course, aware that the setting and population is quite different from that to which NICE guideline will apply! We note, though, that mothers who breastfeed are subject to jokes and bullying that compare them to livestock, and that this has the potential to impact on their choices to breastfeed in public or at all.) We also understand SIDS to be distinct from deaths caused by suffocation or overlaying, so we question the relevance of the analogy.
Carpenter et al.'s analysis combined data from studies which used different and sometimes unclear definitions. Use of imputation of missing variables (including known associated factors such as drug and alcohol use) is highly questionable given likely variation in these factors between the constituent study settings, and limits the certainty of the findings (see response letter by Blair et al. to Carpenter et al.'s article).
Between constituent study settings the prevalence of exclusive breastfeeding, mixed-feeding and formula feeding likely vary, but Carpenter et al's definition of 'breastfeeding' groups mixed-feeding (breast and bottle) together with exclusive breastfeeding. It would be surprising if the protective effect of breastfeeding applied uniformly irrespective of whether the infant was exclusively breastfed, or 'mixed-fed' and receiving perhaps just one breastfeed per day. (It is also unclear whether expressed milk given in a bottle would 'count' as bottle-feeding). We believe that the known association of SIDS with formula-feeding has been inadequately addressed Carpenter et al.'s study design; and also that the protective association with breastfeeding has been 'diluted' by grouping mixed-feeding with exclusive breastfeeding. However we acknowledge that it difficult to establish how infant feeding should be controlled for in studies of SIDS and co-sleeping, since the mechanisms by which formula or bottle feeding contribute to an increased risk of SIDS (or conversely, breastfeeding protects against SIDS) are not well understood.
Interpretation of a p-value of 0.062 as statistically-significant is not considered best practice (Carpenter et al., Table 3). Many would not even consider this p-value as reaching 'borderline' statistical significance. An alternative interpretation of this particular result is that where the baby is aged over 3 months and the mother smokes but does not drink, there is no statistically-significant evidence of an increase in risk with bed-sharing vs. room-sharing.
Kirkwood and Stern are referenced as stating that risk somehow infers causality to imply the term cannot be used in the context of bed-sharing but they also state ‘The purpose of statistical analysis is to quantify the magnitude of the association between one or more exposure variables and the outcome variable. (Kirkwood & Sterne 2003:13)’
Quantifying risk matters as words are not sufficient to convey variations in risk (Shaw and Dear, 1990). Communicating risk requires the risk to be put in perspective (BMA 2012:28). We think this means using numbers for each exposure variable for parents to understand and interpret the risk for themselves and decide how much effort they should invest in keeping their baby safe, particularly when having alcohol, changing sleep arrangements or stopping breastfeeding.
Good luck to you all for your responses.
BMA Board of Science (2012). Risk: what's your perspective? : a guide for healthcare professionals. London: British Medical Association.
Kirkwood, B. R., Sterne, J. A. C. & Kirkwood, B. R. (2003). Essential medical statistics, Oxford, Blackwell Science. Second edition.
Shaw, N. & Dear, P. (1990). How do parents of babies interpret qualitative expressions of probability? Arch Dis Child, 65, 520-523.
Responses from La Leche League GB to the NICE consultation on Sudden Infant Death Syndrome
La Leche League GB would like to offer input into the National Institute for Health and Care consultation on Sudden Infant Death Syndrome (SIDS).
1. The NICE Addendum states “Due to the various methodological and study design issues of the included studies in this update mentioned above, the evidence base could only illustrate associations and a cause-effect relationship could not be inferred”
• LLLGB says: Research studies on the subject of bed-sharing often use different and confusing definitions. The 2013 article by Carpenter et al, Bed-Sharing when parents do not smoke: is there a risk of SIDS?, combined five different studies with varying definitions, including what was meant by bed-sharing.
• In the studies on SIDS there was no consistency in the data collected so there are missing variables which can make a significant difference to the results. One question which was sometimes omitted was how the baby was fed, and this makes a big difference when talking about breastfeeding and bed-sharing as opposed to formula feeding and bed-sharing.
• In some research the definitions of co-sleeping have been unclear. For instance “solitary sleep” might mean a baby alone in another room, or in the same room but not the same bed. “Co-sleeping” might mean a baby in the same room as another person, or a baby in a bed with someone, or a baby asleep with someone on a sofa or other risky shared sleep surface. “Bed-sharing” might be with the baby’s mother, or with someone else, or more than one person, or sharing a sofa or recliner with his mother. The person could be awake, asleep, drunk, sober, an adult not the mother, or even another child.
• For example in 2010 there was a campaign in Milwaukee County against bedsharing. When the statistics were investigated it was found that all the “bedsharing” deaths in the past year had involved adult alcohol use, a baby on a pillow, difficult living conditions and/or smoking. Every death involved a formula fed baby.
• A 2009 study by public health officials in Alaska which aimed to tease apart known risk factors and look at them separately found that 99% of the babies who died had at least one of these risks: Face down position, sleeping with someone other than the mother, maternal tobacco use, impaired bed partner, sofa or waterbed. It was not the bed-sharing but the condition of the adult and safety of the surface1.
• Gathering information at a death scene is complicated and research conclusions are only as good as the data collected. In some places a doctor with specialised training investigates an infant death; some places don’t require any special training. Some places have a death scene report form, but it isn’t used everywhere and bed-sharing is not always clearly defined.
• An often-cited 2005 study did not collect data on alcohol consumption as previous experience had demonstrated the difficulty of obtaining accurate information2 thus ignoring a key variable for both SIDS and suffocation. Some examiners record how the baby was fed, others don’t. Some ask abut smoking, and some don’t. Sometimes information given by the adults involved isn’t accurate.
• This huge range of approaches, definitions, levels of accuracy, detail and training, provide the data for the “never bed-share” studies. When committees and support organisations include parents who have lost a baby to SIDS it is understandable that strong emotions will be involved.
• Breastfeeding mothers and babies sharing sleep is a biologically normal behaviour, while formula feeding and separate sleep are departures from the norm. It is these behaviours that need to be shown to be effective and safe, not the other way round. In some reports on SIDS and other sleep related deaths, breastfeeding mothers who do not smoke and have not consumed alcohol or arousal-altering medications, are considered a sub-group to be discounted from the analysis – whereas they should be the normal, starting point. A first step in looking at infant deaths in adult beds would be to look at what was wrong with the beds, not what was wrong with the mothers.
2. The NICE addendum states that their definition of co-sleeping includes sharing a bed or any other sleep surface such as a sofa or chair.
• LLLGB says that sofas and chairs represent a much greater risk that safe bed sharing and that to issue a blanket recommendation based on such variable circumstances is misleading. Babies have a biological need to feed during the night, and mothers are hardwired to respond. Breastfeeding releases hormones which aid rest and relaxation in the nursing dyad and make both sleepy. If warnings about bed-sharing means mothers stop taking their babies into bed with them they still need to feed their babies somewhere. This will probably be in a chair or on a sofa, which are riskier places to fall asleep with a baby than in a bed when appropriate measures have been taken.
3. The NICE Addendum states “it would be inappropriate to use the term risk when considering SIDS and co-seeping as the causes of SIDS are likely to be multi-factional and a possible causality link with co-sleeping is not clearly defined”. It went on to say “It cannot be definitely stated that co-sleeping is a risk for SIDS. Some of the reviewed evidence showed that there is a statistical relationship between co-sleeping and SIDS. This means that, where co-sleeping occurs there may be an increase in the number of cases of SIDS. However, the evidence does not allow us to say that co-sleeping causes SIDS."
• LLLGB says that no one has proposed a physiological mechanism what would cause a baby to die of SIDS just because he is next to his non-smoking mother - and since mothers and babies are hardwired to be together it wouldn’t make any sense.
• LLLGB would like to draw attention to the differentiation between SIDS and Accidental Suffocation and Strangulation in Bed (ASSB). The four biggest risk factors associated with SIDS are smoking3, laying a baby tummy down for sleep4, leaving a baby unattended5 and formula feeding6. It happens in a small group of vulnerable babies who have very specific but undiagnosed health issues. Babies may become distressed if their need to be close to their mothers at night is denied and a vulnerable baby may also be more sensitive to stress. SIDS is different to the risk of Accidental Suffocation and Strangulation in Bed (ASSB) which happens when a baby is in an unsafe sleep environment with an impaired carer.
• The lowest SIDS rates in the world are in countries where bed-sharing is traditional, for instance parts of Asia and South Asia7. It doesn’t seem to be a matter of geography as when people from a low-risk cultures move to other countries if they bring their traditions with them they also tend to bring along a low rate of SIDS8. The US has a higher rate of SIDS than just about anywhere else and the four big risk factors for this are mentioned above. Bed-sharing is not one of them.
• Putting two different risks together is misleading and does not allow parents to make informed decisions. While SIDS is something which affects a small group of vulnerable babies, parents need to be aware of ASSB (breathing hazard) risks. Making sure a bed is as safe as possible, free of suffocation and injury risks, having a smoke free home, not using arousal- altering medications or alcohol, and breastfeeding are all ways to ensure a healthy baby and his family get a good night’s sleep. Even if a mother does not intend to bed-share, ensuring her sleeping area is safe will help on those nights when she finds herself nursing in bed and falling asleep.
• Four researchers who have used very clear definitions in their studies are Drs. Helen Ball (UK), Nils Bergman (South Africa), Kathleen Kendall-Tackett and James McKenna (USA). Their conclusions support the safety and normalcy of bed sharing.
• Dr McKenna says that the amount of COÇ the mother expires in her breath acts to stimulate infant breathing. Expelled COÇ appears to act as a potential back-up should the baby's own internal drive to breathe falter or slow, since the baby's nasal regions can both detect and respond to the presence of this gas by breathing faster.
• He also says that babies who sleep with their mothers and breastfeed spend less time in the deepest stages of sleep (three and four), from which arousal is more difficult should the baby need to awaken quickly to terminate a dangerous apnoea. Instead, co-sleeping babies spend more time in lighter stages of sleep (one and two) which is thought to be physiologically more appropriate for young infants, and more natural and conducive to safe sleep for babies. The shorter durations of deeper stage sleep promoted by co-sleeping can potentially protect those infants born with arousal deficiencies (suspected to be involved in SIDS).
• LLLGB believes that suggestions have to be “doable”. Losing a baby is such an incredible tragedy that it is understandable to look for ways to prevent this. However most breastfeeding mothers sleep with their babies at least some of the time and it is unrealistic to try to stop this. Creating one rule for a specific high-risk situation and applying it to everyone will not work. Tired mothers have to feed their babies somewhere and making everyone feel they are at risk can cause damage in other ways. Mothers may turn to formula or cereal in the hope of getting their baby to sleep longer; they may fall asleep in unsafe places or undertake a sleep-training programme which has associated risks.
• LLLGB believes it would be better to offer information to parents to make shared sleep as safe as possible. Whilst no sleeping environment can be entirely risk free, studies at Durham Parent-Infant Sleep Lab have found that mothers who sleep with their breastfed babies in bed adopt a protective position that makes overlaying difficult, and smothering by bedding unlikely. It has been observed that babies “demonstrably do not overheat” and that they breastfeed more successfully and for longer, which has significant health benefits for mother and child.
• Safe Sleep campaigns which tell mothers their baby should not sleep in their bed, a couch or a chair, give that mother no information about where she can feed her baby at night when she is likely to fall asleep. Mothers need to have a clear explanation of SIDS and ASSB risks, with different recommendations based on their different lives. They need full information about their options, not alarmist threat messages.
• Policies designed to protect the health of a few babies should not be applied to all if harm can result. The blanket recommendation of this analysis that parents “simply avoid bed-sharing” may well scare women into making decisions which are not right for them, their family or their baby, and could lead to babies being breastfed in places where it would be riskier for women to fall asleep. Ultimately it could also lead to women stopping breastfeeding because they are too worried about where to feed.
• If this recommendation closes down the opportunity for discussions about safe bedsharing it denies parents whose babies are at low risk for SIDS (healthy term births, breastfed, non-smoking, non alcohol-consuming parents) the opportunity to make an informed decision and may lead to parents feeling they need to lie about their choices.
• It is vitally important that parents are aware of the need for safe sleeping arrangements and of circumstances which might cause risk. They need to know abut the risks and benefits of co-sleeping and unsafe co-sleeping practices so they can make their own informed decisions. Trying to instigate a ban is not a reasonable or effective response to such an instinctive and natural human behavior.
------------------ Notes --------------------------
1 Blabey M.H., Gessner D.G., Infant bed-sharing practices and associated risk factors among babies and
infant deaths in Alaska. Public Heath Reports 124, No 4 (2009): 527-534
2 Tappin D., Ecob R., Brooke H., Bed-sharing, room sharing and sudden infant death syndrome in
Scotland; a case control study Journal of Pediatrics 147, No1 (2009): 32
3 Zhang K, Wang X, Maternal Smoking and increased risk of sudden infant death syndrome: a meta
analysis. Legal Medicine (Tokyo) 2013. Fleming P, Blair P.S. Sudden infant death syndrome and parental
smoking. Early Human Development.2007
4 Mitchell E.A., Taylor B.J., Ford R.P.K. et al, Four modifiable and other major risk factors for cot death,
the New Zealand Study Journal of Paediartrics and Child Health, 1992
5 Carpenter R.G., Irgens L.M., Blair P.S. et al, Sudden unexplained infant death in 20 regions in Europe,
case control study Lancet 2004.
6 Hauck F.R., Thompson J.M., Tanabe K.O. et el, Breastfeeding and reduced risk of sudden infant
syndrome: a meta analysis Pediatrics 2011
7 Liamputtong P., Chil-rearing and Infant Care Issues: A cross-cultural Perspective Hauppauge, NY: Nova
8 Ball H.L., Moya E., Fairley L. et al Infant Care Practices Related to sudden infant death syndrome in
South Asian and white British families in the UK. Paediatric and Perinatal Epidemiology 26, no 1 (2012) 3-