Webchat with Professor George Haycock about co-sleeping

This is an edited transcript of a live webchat with Professor George Haycock on 30 June 2009. Professor Haycock is scientific adviser to the Foundation for the Study of Infant Deaths and emeritus professor of paediatrics at Guy's and St Thomas' Hospitals, London

Co-sleeping | Risk factorsBreastfeeding/formula feeding | Dummies | Ill children | Daytime naps | Smoking | Background information/references



Q. gingerninja: I would like to know why death in a parental bed through eg suffocation is catagorised as SIDS when there clearly is an obvious cause and why figures for death through lone sleeping are also not covered. This gives a very distorted picture.

A. GeorgeHaycock:  Death in a parental bed (or anywhere else) due to suffocation or overlying is not classified as SIDS. Figures for lone sleeping are extensively covered in the papers I have cited (see box), and indeed in many others.

Q. VeniVidiVickiQV: I have several questions:

  1. How is the data, statistics and evidence gathered when researching SIDS - I'm assuming it's all done retrospectively? Are socio-economic factors included etc? I think it would be helpful for those of us who aren't experienced at interpreting studies and research to see what you look at, how you obtain data etc in order to help validate our own decisions.
  2. What is your definition of co-sleeping exactly?
  3. Are the protective effects of breastfeeding and the more common act of co-sleeping whilst breastfeeding factored into the co-sleeping advice?

A. GeorgeHaycock: (I like the classical allusion!) The epidemiological data on SIDS are indeed gathered retrospectively, since one cannot in general prospectively randomise families to different infant care practices, especially if there is good reason already to believe that one or more of the options confers added risk.

There are one or two exceptions. For example, there are four prospective, randomised studies of dummy use and its relation to breastfeeding exclusivity and duration (reviewed in O'Connor NR et al. Archives of Pediatrics and Adolescent Medicine (2009) 163: 378-382), which found no evidence of an adverse effect of dummy use on either index of breastfeeding success.

"I avoid the use of the term co-sleeping because it includes too many different and diverse practices. I prefer to talk about bed sharing, sofa sharing, room sharing etc, to avoid confusion (which has unfortunately affected many published studies in the field)."
Professor Haycock

Yes, socioeconomic factors are very definitely included in all the serious studies.

I avoid the use of the term co-sleeping because it includes too many different and diverse practices. I prefer to talk about bed sharing, sofa sharing, room sharing etc, to avoid confusion (which has unfortunately affected many published studies in the field).

Yes, the protective effects of breastfeeding are factored into the research into bed sharing, in some but not all studies. Where it has been looked at, the combination of breastfeeding and non-bed sharing is the safest practice and the combination of bed sharing and not breastfeeding is the least safe.

Q. wuglet: Is there a difference between regular co-sleepers and occasional ones? My point being that if parents do not routinely co-sleep there may have been some reason they chose to do so on that particular night ie baby more unsettled then usual (which may itself be connected to some as yet unknown underlying cause).

A. GeorgeHaycock: This is covered by my answer to flamingobingo. Your second point is interesting, but a study by the German SIDS Study (a different one from the one referred to above, Venneman M M et al. Archives of Disease in Childhood (2005) 90: 520-522) found that symptoms of minor illness were no longer a predictor of SIDS in a population where the great majority of infants now sleep supine. Their results were obviously obtained from the German population but the general description fits the UK population as well.

Q. hunkermunker: Of the 'co-sleeping' incidents of SIDS, how many are sofa-sharing, how many are bed-sharing? And of the bed-sharing ones, how many are following the current safest guidelines to the letter? And how does that compare to cot-sleeping incidents of SIDS? Thank you - and so sorry to those of you with personal experience of this particular nightmare - unbearable.

A. GeorgeHaycock: More deaths occur in the parental bed than on sofas, but a higher proportion of sofa sharers than bed sharers die, because bed sharing is a much commoner practice that sofa sharing. In an important study from Avon (Blair P S et al. Lancet (2006) 367: 314-319) comparing the statistics from 1984 to 2003, in the most recent cohort (1999-2003) 50% of deaths occurred in infants sleeping alone in cots, 39% in the parental bed and 11% were sofa sharing.

Recent studies from Germany and the Netherlands suggest that following the guidelines 'to the letter' reduces but does not abolish the risk of bed sharing, especially in the youngest infants.

Q. LupusinaLlamasuit: I'm sure MNHQ can point you in the direction of the recent thread on MN in which your recent poster campaign was discussed. Many of us on MN are used to interpreting scientific data.

I looked at the poster and your press releases attached to it. I used my academic position to explore where the research that the poster and press releases were based upon were published. I couldn't find them published anywhere. Given that the studies mentioned were based on a very small sample of pathology reports it seemed somewhat surprising that you would base a major campaign, claiming 'expert advice based on new scientific studies'. If you do have the references, could you please post them here? And if it is not published in peer-reviewed journals, do you think it is wise to act using your considerable credibility and influence to send out a message that many parents might misinterpret?

What made many of us cross on the other thread was the rather scaremongering tone of the poster in response to what was a very inconclusive piece of small scale research. There was no relation of the data to the actual population, and no separation of the independent risk factors (crucially sleeping on a bed and sofa etc were included together). This is poor data interpretation and reportage. To base a frightening poster campaign, directly attacking safe bedsharing, I felt, was completely irresponsible.

Given that we know that public health messages DO have an impact on parental behaviour (pace the large decrease in SIDS when the back to sleep campaing was implemented), will the FSID be funding research to explore whether its own campaign may have increased infant deaths by encouraging more unsafe co-sleeping? We had messages on the days after our discussion from parents who said 'oh I fall asleep on the sofa with her because I'm scared to have her in the bed...'

Why you allowed this to be released staggers me. The ALSPAC study in Bristol will surely provide much more robust evidence.

A. GeorgeHaycock:  (I hope your screen name refers to the wolf and not the disease!) I have given a list of relevant references to Geraldine, which I hope will answer your first point (see box).

The FSID recommendations were actually based on published papers based on 2,616 SIDS cases and 9,296 controls – not such a small sample really. All of these were published in peer-reviewed journals. FSID funds research projects for which funding is sought by independent researchers. If you or anyone you know has a coherent proposition for research of the kind you suggest, feel free to apply for support.

Q. hedgiemum: My HV told me that a link between co-sleeping and SIDS wasn't proven for "exclusively breastfeeding women who apply common sense", the link only being proven for falling asleep with a baby (on sofa or bed) in general, with formula-feeding mothers lumped in with breastfeeding mothers in the research. Is this true ?

She said this in the context of encouraging me to try co-sleeping rather than give up breastfeeding at 8 weeks as I couldn't function, I was so exhausted. I ended up co-sleeping with each of my three children until they stopped breastfeeding at 12-18 months, at which point they migrated back into their cots.

A. GeorgeHaycock: I'm afraid the information given to you by your HV is not correct (see previous answers and my accompanying references). Your personal predicament is a common one. It is worth remembering that both the risk of SIDS in general, and the risk of bed sharing, applies mainly to the youngest infants, i.e. those less than 2-3 months old.

Q. hunkermunker: If co-sleeping is not safe 'full-stop' how are there children alive in most of the world? Sorry, that does sound slightly facetious, but still, a vast number of babies sleep with their mothers/fathers/in a heap like mice and not in tastefully beige /primary-coloured bedding by themselves. And they survive - nay, thrive.

A. GeorgeHaycock:  Even at its worst in this country, before the 'back to sleep' campaign in the early 1990s, the cot death rate was about 2 per 1,000 live births (about 1,300 cases a year in England and Wales) which has now fallen to about 0.4 cases per 1,000 live births (about 300 a year). So, even when front sleeping, maternal smoking pregnancy and other avoidable risk factors were either not understood or not avoided, the great majority of babies survived. As I have tried to point out elsewhere, a risk factor is not a single cause. To use an analogy, nobody seriously doubts nowadays that cigarette smoking is a major risk factor for lung cancer, but most lifelong heavy smokers (90% in fact) don't get lung cancer. The recommendations for avoiding SIDS are not designed to save the 700,000 or so babies born every year in this country but the few hundred who are still dying of this still incompletely understood condition.

The FSID guidelines do include the recommendation to breastfeed your baby. To go back to your previous point, the huge majority of formula fed babies survive (and thrive!). The best evidence, incidentally, suggests that breastfeeding modestly reduces the risk of SIDS compared with bottle feeding with an odds ratio of about 0.64.

This figure comes from the USA Agency Agency for Healthcare Research and Quality report no. 153, available online at http://www.ahrq.gov/clinic/tp/brfouttp.htm. Beware, this report is about 400 pages long: the section on breastfeeding and SIDS begins on page 97.

Q. GreenMonkies: For me, these figures speak for themselves, if less that 40% of babies who succumb to SIDS are in the parental bed, then this means that the remaining 60% were not in the parental bed. Which means more non-bed-sharing babies die from SIDS, which means a baby is more likely to die from SIDS if it is sleeping alone in a cot or on a sofa etc.

How can anyone interpret these figures to say that "co-sleeping"/bed-sharing increases the risk of SIDS?

In parts of the world where babies sleep with their mothers (in slings etc during the day and in the bed during the night) and are breastfed through the night, and are not put in cots or alone in rooms, or subjected to "sleep training" to make them sleep unnaturally long, deep sleeps, and not fed formula, SIDS is essentially unheard of. It is only in the "West" where babies are fed formula, put into cots/rooms alone to sleep, and "encouraged" to sleep longer and deeper by being in automatic rocking swings and being left to "whinge" and "settle themselves" to sleep etc that we have a problem with SIDS. Surely this isn't a coincidence?

The human race has been bed-sharing for millions of years, and in most of the world still does. It is not shared sleep that is dangerous, it is separation and artificial feeding that increases the risks. Safe bed-sharing is exactly that, SAFE, a side-car cot, a king size bed, baby next to mum not between mum and dad, breastfeeding on demand throughout the night, these things are not risk factors, they are common sense and natural, and safe.

And one study stating that dummy use doesn't interfere with breastfeeding doesn't overturn the several that say it does. Dummies cause all kinds of problems, poor latch, reduced feeding, malocclusion, narrowed nasal airways and increase the risk of ear infections and obstructive sleep apnea. WHy would anyone recommend the use of a dummy to help a baby stay in a light enough sleep when a breastfed, bed-sharing baby gets this and more.

Why don't you consult with and take advice from Helen Ball instead of MAM?

A. GeorgeHaycock: Several questions here. First, many more babies sleep in cots than in their parents' beds, so not surprisingly there will be more deaths in that setting. However, the proportion of bed sharing babies that die is significantly greater than the proportion of non bed sharing babies that die.

It is true that in many parts of the world infants do sleep with their mothers (interestingly, in some of these populations the father is excluded from the maternal bed as long as this is continuing). It is also true that most of those places have infant mortality figures that would be considered horrendous by European standards, so the safety is a matter of speculation, Also, in many of these populations the baby sleeps with the mother on a firm, hard mattress on the floor, not on a modern soft mattress with many coverings, so the findings are not necessarily capable of extrapolation to UK circumstances.

It is four studies, not one, that found no evidence of any adverse effect of dummy use on breastfeeding. They are reviewed in a single study (O'Connor et al 2009), which is unique among all the published studies on this subject in that it included only those research papers that used the best methodology (prospective, randomised controlled studies) and are therefore free of bias, whether conscious or unconscious.

Again, the evidence for a protective effect of dummy use againt SIDS is based on at least 8 or 9 good quality studies and appears to apply to both breast and bottle fed babies.

Perhaps I could add that my generation (babies born during the second world war) probably had the lowest incidence of breastfeeding of any generation of modern times, but we have so far had the greatest life expectancy of any generation in history. This is not an argument against breastfeeding but perhaps a caution not to overstate its benefits on survival.

I know Helen Ball well. As it happens, I am a member of the steering committee of her latest piece or research. I am unfortunately bound by contract not to discuss that project until it is finished and published. For what it is worth, I last met her at a meeting convened by the University of Leicester on June 18 and 19, 2009. 

Background information

This short contribution addresses some of the issues raised by various Mumsnetters, sometimes repeatedly, and I hope will be seen as helpful as a background to further discussion. Some of your correspondents seem to think that the advice on bed sharing rests only on the recent figures from the north of England, referred to in the FSID press release. This is not the case.

There are currently at least 8 case control studies, (ie all of them compare infants who died of SIDS with a group of control infants who did not die, usually two, three or four control infants for each index case).

They also all control for other factors including parental smoking, socio-economic indices and in some cases breastfeeding as well.

These studies come from Ireland, Scotland, England, New Zealand, the USA, the Netherlands and Germany, as well as one which is a multinational European study from 20 regions in Europe. (1)

All of these are published in mainstream, peer reviewed journals and between them they include 2,616 SIDS cases and 9,296 control infants.

The references for these publications are given below and should be easily obtained by anyone with access to a University or Medical School library. (1-8)

In summary, all 8 of these found that bed sharing (as opposed to sofa sharing, which is a much higher risk) increased the risk of SIDS.

The difference was statistically significant in 6 of these studies; (1-3, 5, 7, 8) in the other two (4, 6) the risk did not reach conventional statistical significance (P <0.05) for non-smoking mothers but the trend was the same: odds ratio for bed sharing vs non-bed sharing 1.3 and 1.35 respectively.

The risk was inversely related to age: most studies found that the main risk was to infants aged less than 3-4 months.

On dummy use, a meta-analysis (9) of 7 research papers (1, 10-15) was set up by the American Academy of Pediatrics. Again, these were all published in good, peer-reviewed scientific journals and selected from a much larger number of studies in the literature because of their good scientific methodology.

Combined, these studies included 2,215 SIDS cases and 6,816 control infants. They found that the summary odds ratio for dummy use during the last sleep compared with the reference sleep for the control infants was 0.39 with 95% confidence intervals of 0.30-0.50.

This means that babies offered a dummy for the last sleep were less than half as likely to die than those not offered one.

Incidentally, a very recent publication from the USA (16) has reviewed the total published literature on dummies and breastfeeding and found that (to quote the authors directly): "The highest level of evidence does not support an adverse relationship between pacifier use and breastfeeding duration or exclusivity." This paper is particularly strong because it only uses the most rigorous form of investigation, namely prospective, randomised controlled trials, which are statistically much stronger than other types of study in inferring causation.

George Haycock, Scientific Advisor to the Foundation for the Study of Infant Deaths
Emeritus Professor of Paediatrics, Guy’s and St Thomas’ Hospitals School of Medicine


1 Carpenter RG, Irgens LM, Blair PS, England PD, Fleming P, Huber J, et al. Sudden unexplained infant death in 20 regions in Europe: case control study. Lancet 2004; 363: 185-91
2 McGarvey C, McDonnell M, Hamilton K, O'Regan M, Matthews T. An 8 year study of risk factors for SIDS: bed-sharing versus non-bed-sharing. Arch Dis Child 2006; 91: 318-23
3 Tappin D, Ecob R, Brooke H. Bedsharing, roomsharing, and sudden infant death syndrome in Scotland: a case-control study. J Pediatr 2005; 147: 32-7
4 Blair PS, Fleming PJ, Smith IJ, Platt MW, Young J, Nadin P, et al. Babies sleeping with parents: case-control study of factors influencing the risk of the sudden infant death syndrome. CESDI SUDI research group. British Medical Journal 1999; 319: 1457-61
5 Mitchell EA, Taylor BJ, Ford RP, Stewart AW, Becroft DM, Thompson JM, et al. Four modifiable and other major risk factors for cot death: the New Zealand study. Journal of Paediatrics & Child Health 1992; 28:S3-8
6 Hauck FR, Moore CM, Herman SM, Donovan M, Kalelkar M, Christoffel KK, et al. The contribution of prone sleeping position to the racial disparity in sudden infant death syndrome: the Chicago Infant Mortality Study. Pediatrics 2002; 110: 772-80
7 Ruys JH, de Jonge GA, Brand R, Engelberts AC, Semmekrot BA. Bed-sharing in the first four months of life: a risk factor for sudden infant death. Acta Paediatr 2007; 96: 1399-403
8 Vennemann MM, Bajanowski T, Brinkmann B, Jorch G, Sauerland C, Mitchell EA. Sleep environment risk factors for sudden infant death syndrome: the German Sudden Infant Death Syndrome Study. Pediatrics 2009; 123: 1162-70
9 Hauck FR, Omojokun OO, Siadaty MS. Do pacifiers reduce the risk of sudden infant death syndrome? A meta-analysis. Pediatrics 2005; 116: e716-23
10 Fleming PJ, Blair PS, Pollard K, Platt MW, Leach C, Smith I, et al. Pacifier use and sudden infant death syndrome: results from the CESDI/SUDI case control study. CESDI SUDI Research Team. Archives of Disease in Childhood 1999; 81: 112-6
11 Hauck FR, Herman SM, Donovan M, Iyasu S, Merrick Moore C, Donoghue E, et al. Sleep environment and the risk of sudden infant death syndrome in an urban population: the Chicago Infant Mortality Study. Pediatrics 2003; 111: 1207-14
12 McGarvey C, McDonnell M, Chong A, O'Regan M, Matthews T. Factors relating to the infant's last sleep environment in sudden infant death syndrome in the Republic of Ireland. Arch Dis Child 2003; 88: 1058-64
13 Mitchell EA, Taylor BJ, Ford RP, Stewart AW, Becroft DM, Thompson JM, et al. Dummies and the sudden infant death syndrome. Arch Dis Child 1993; 68: 501-4
14 L'Hoir MP, Engelberts AC, van Well GT, Damste PH, Idema NK, Westers P, et al. Dummy use, thumb sucking, mouth breathing and cot death. Eur J Pediatr 1999; 158: 896-901
15 Tappin D, Brooke H, Ecob R, Gibson A. Used infant mattresses and sudden infant death syndrome in Scotland: case-control study. Bmj 2002; 325: 1007-1012
16 O'Connor NR, Tanabe KO, Siadaty MS, Hauck FR. Pacifiers and breastfeeding: a systematic review. Arch Pediatr Adolesc Med 2009; 163:  378-82

Q. GreenMonkies: First, many more babies sleep in cots than in their parents' beds, so not surprisingly there will be more deaths in that setting. However, the proportion of bed sharing babies that die is significantly greater than the proportion of non bed sharing babies that die.

According to who? Most co-sleepers either lie about it, or don't realise that falling asleep with the baby in the bed is co-sleeping. It is true that in many parts of the world infants do sleep with their mothers (interestingly, in some of these populations the father is excluded from the maternal bed as long as this is continuing). Yes, this is called safe co-sleeping, baby is not placed between mum and dad, but dad is removed from the area where baby is sleeping (like using a side-car cot).

It is also true that most of those places have infant mortality figures that would be considered horrendous by European standards, but not from SIDS. Infections etc perhaps, but not from SIDS.

So the safety is a matter of speculation, Also, in many of these populations the baby sleeps with the mother on a firm, hard mattress on the floor, not on a modern soft mattress with many coverings, so the findings are not necessarily capable of extrapolation to UK circumstances.

It is four studies, not one, that found no evidence of any adverse effect of dummy use on breastfeeding. They are reviewed in a single study (O'Connor et al 2009), which is unique among all the published studies on this subject in that it included only those research papers that used the best methodology (prospective, randomised controlled studies) and are therefore free of bias, whether conscious or unconscious. I think I take the word of internationally respected breastfeeding support networks over all in this case.

Again, the evidence for a protective effect of dummy use againt SIDS is based on at least 8 or 9 good quality studies and appears to apply to both breast and bottle fed babies.

Perhaps I could add that my generation (babies born during the second world war) probably had the lowest incidence of breastfeeding of any generation of modern times, but we have so far had the greatest life expectancy of any generation in history. This is not an argument against breastfeeding but perhaps a caution not to overstate its benefits on survival. What? There is no lack of surveys showing that from a high rate of breastfeeding in the industrializing world of the early 20th century, after World WarII the incidence declined to a nadir around 1960.

A. GeorgeHaycock: We simply don't know that infants in these developing countries don't die of SIDS. The infant mortality from other causes is so high that a few cases of SIDS simply wouldn't be noticed (for example, if the SIDS rate were two per thousand births, bearing in mind that SIDS can only be properly diagnosed by a skilled investigation including an expert postmortem examination, and the overall infant mortality was 10% or even higher, as it is in some places).

So we simply don't know how many babies in these societies die of SIDS. In many cases we have no idea why they die at all since many of them never reach a doctor or other health care professional, let alone a hospital. And yes, I have worked in a developing country.

Risk factors

Q. gardeningmum05: I lost my daughter to sudden infant death, and she was asleep in our bed. There were no pillows, no duvet near her, we had not been drinking, all advice and safety issues thought of, but there is not a day goes by that I dont cry with guilt that we co-slept with her that night. I have had 2 children since she died and they have never ever slept in our bed. Trust me, is it worth the risk? No new mother should be encouraged to co-sleep

A. GeorgeHaycock: I am very sorry to hear of your experience. It is important to recognise that risk factors are not causes in a simple sense. Whatever the epidemiological evidence may show about population risks, it is never possible to identify a definite cause for an individual case of SIDS (if a cause were identified, it would not be classified as SIDS).

Q. flamingobingo: What is riskier when it comes to SIDS? Is it not breastfeeding, or is it co-sleeping? Because it is proven that planned co-sleeping increases breastfeeding duration, which protects against SIDS. But there is no research that I know of that shows that planned co-sleeping increases the SIDS risk. Therefore, wouldn't it be better to encourage safe co-sleeping than to demonise it completely?

I feel so sad that babies die in cots too, but no one ever regrets putting their baby in a cot! My DD3 stopped breathing in the night when she was four months old. She would be dead now if we hadn't been co-sleeping because I knew immediately and picked her up and she took a gasp of air and was fine. I would have not been anything like as aware or quick to respond had she been in a cot.

A. GeorgeHaycock: It is not proven that planned co-sleeping increases breastfeeding duration. Furthermore, there are no statistics that I know that differentiate between planned bed sharing (I avoid the term co-sleeping) and unplanned bed sharing.

Some of the studies have looked at both 'usual' bed sharing and bed sharing during the last sleep. Both appear to confer risk. In the recently published (April 2009) study by the German SIDS Study (Vennemann M M et al. Pediatrics (2009) 123:1162-1170), it was specifically those infants who were accustomed to bed sharing and who shared during the last night that were at increased risk on multivariate (adjucted) analysis: those who slept in the parental bed on the last night but were not accustomed to doing so were not at increased risk in the multivariate analysis.

Yes, babies die in cots too, but in the studies conducted and published by the FSID research group in Bristol, the incidence of infants dying in cots has fallen six fold since the back to sleep campaign was introduced but has only fallen by half in those bed sharing.

Q. GH: I imagine someone else will have asked this question already as it is always being discussed in breastfeeding circles so if anyone has put it more eloquently than me then please ignore this one!

We are always told that the risk of SIDS is higher when parents sleep with their babies either in bed or on a settee. I would like to see the breakdown between these two sets of figures. The ones I read about in the local paper seem to always involve a woman who has consumed a lot of alcohol falling asleep on the sofa and suffocating her baby. I have not read about a case where two parents choose to sleep with their baby (or one parent) in their bed every night (obviously breastfeeding parents as research indicates that af mothers don't have the same responses as bf mothers) and yet managed to suffocate their baby.

I don't have exact figures to hand but it was something like 40% die in cots, 60% die on settees or in the parents' bed. I would like 3 figures: the % who die in cots; the % who die on sofas or somewhere other than a cot or the parents' bed and the % who die in bed with their parents/mother. Do these figures exist? If not, why not? It seems a very obvious distinction to make to me. 

A. GeorgeHaycock: The papers I have referred to include many cases of babies who have died during regular sleeping in their parents' bed. All the studies that have differentiated bed sharing from sofa or settee sharing have shown that the risk is far higher in the sofa situation than in an adult bed.

It is not possible to give overall figures for the three categories because (a) they differ somewhat between the various studies, and (b) the risk varies substantially with the age of the infant.

Q. AnarchyAunt: I'd like to know the risk of SIDS when intentionally co-sleeping and following guidelines for doing so (ie avoiding smoking/alcohol/pillows), as distinct from falling asleep on the sofa with the baby. The risk factors involved are not sufficiently differentiated for a true comparison to be made IMO.

I have heard that occasional/accidental co-sleeping is more dangerous than consciously doing so every night. Is this true and what is the research to back this up?

A. GeorgeHaycock: The German study just referred to conveys the opposite message, as far as bed sharing is concerned. Occasional or accidental co-sleeping on, for example, as sofa is very much more dangerous and I think everybody is agreed on that.

Q. ReallyTired: Is cot death more or less common among firstborn babies?

A. GeorgeHaycock: Not as simple a question as it might appear. It is more common overall, but subsequent babies born in a family where one such death has occurred are at greater risk than those born in a famlly without this history.

Q. Leningrad: "50% of deaths occurred in infants sleeping alone in cots, 39% in the parental bed and 11% were sofa sharing." If I understand this correctly, it is 'safer' to have the baby in bed isn't it?

A. GeorgeHaycock: No, it isn't safer to have the baby in the bed. The explanation is that more babies sleep in cots than bed share, and the risk as a total of all babies is higher for bed sharers than for babies sleeping in a separate cot in the parents' bedroom (the last part is important: more than one study has shown that the risk for babies in the first few months who sleep in a separate room is double that for those in a cot in the parents' room).

Leningrad: It would be interesting to see a study which showed what the stats are for babies breastfed on demand who sleep in the parental bed (or just with the mother really) with no pillows/duvets and no smoking, drinking or medication from the mother. Essentially, with all those risk factors stripped out, would this be the 'safest' option or is there something about our mattresses that might be found to be a problem?

Q. gardeningmum: Are their any statistics or records of SIDS victims siblings going on to have children and their children dying of SIDS?

A. GeorgeHaycock: Good question. The short answer is no, but there is a subgroup of babies who were initially diagnosed as SIDS but who turned out on subsequent investigation to have rare genetic conditions (usually very rare). This subgroup my be at increased risk of familial occurrence of infant death but they would no longer be classified as SIDS.

Q. littleminsky: I would like to know the vaccine status of SIDS cases - co-sleeping or not.

A. GeorgeHaycock:  Careful studies by the German SIDS study group (lead investigator Mechtild Vennemann) have shown that immunised infants are less likely than non-immunised ones to die of SIDS.

littleminsky: That is really interesting. Do you have the figures of vaccinated babies dying of SIDS and non vaccinated babies dying of SIDS or where I could find it. I have been reading a book called "DPT- Ashot in the Dark" which may suggest a link between the Pertussis vaccine and SIDS. I think there was a peak in deaths 2 days after getting the vaccine so I didn't know if any further studies have been carried out about it.

Q. peppapighastakenoverm...: Have there been any analyses looking at which of the risk factors are the biggest predictors of cot death?

A. GeorgeHaycock: There are many epidemiological studies of the relative importance of various risk factors for SIDS. One of the most important of these is actually a book, rather than a paper, called the CESDI report into SUDI studies which was published by HM Stationery Office in 2000. This is obtainable from HMSO (cost £25 when I bought mine) which is easily accessed online.

Other important publications that bear on this include studies from New Zealand, Australia, Germany, Holland, USA and Canada and the Nordic countries. What is interesting is how similar the results are from these different countries, suggesting I think that they are describing real risk factors and not just local phenomena (although all these are, of course, developed - i.e. rich - countries. 

Q. NoHotAshes: Prof Haycock, I understand that this would be speculation, but do you have any personal hypotheses as to why bedsharing is a risk factor for SIDS? Especially given that presumably separate sleeping is a relatively new phenomenon for our species. And I'd like to add my thanks for the detailed responses.

A. GeorgeHaycock: I wish I could answer this but I can't. There are many theories, some of which have a reasonable amount of evidence in support of them. One is that babies who are at increased risk of SIDS have impaired arousal responses to lack of oxygen (hypoxia) or accumulation of carbon dioxide (hypercapnia), and that in the prone position they are more prone to rebreathe expired air and therefore asphyxiate.

Incidentally, this fits well with smoking during pregnancy as a risk factor. Studies both in animals and post mortem studies of SIDS babies (very recently) appear to show that antenatal nicotine exposure impairs the development of the part of the brain that controls arousal (the serotoninergic system in the brain stem, to be technical).

Another theory (that does not necessarily invalidate the first) is called the common bacterial toxin hypothesis. This holds that abnormally exaggerated inflammatory responses to minor infection (largely due to our old friend, or enemy, Staphylococcus aureus) may be triggered by the bacteria being incubated at higher temperatures (about 37 degrees Celsius), and that face down sleeping, and also head covering and overheating, reinforce this. The key element in this argument is that it is known that toxic strains of these bacteria only produce a certain dangerous toxin when incubated at these higher temperatures.

Breastfeeding/formula feeding

Q. Tambajam: I work as a volunteer breastfeeding counsellor and I often find that mothers choose to co-sleep as they feel it is safer than the alternative especially in the early days. Last week at a home visit a mother was tearfully describing how she found herself in her nursing chair slumped over her baby fast asleep after a night feed and on another occasion woke to find her baby trapped against the arm of the chair. She had then familiarised herself with the UNICEF guidelines using their bed-sharing leaflet and had set up a space in her adult bed which she felt had significantly less risk. This seemed sensible to me.

As the infant feeding survey shows 61% of breastfeeding mothers co-sleep at some point and other surveys show figures even higher than this, isn't it a bit unrealistic to expect co-sleeping to end? And by making it less acceptable don't we simply drive it underground and mean mothers won't receive the important safety information they require?

We could play tit-for-tat with research (eg look at Professor Helen Ball's work at Durham University) but the bottom line is, surely, co-sleeping IS going to happen just as it has throughout human history. So rather than try and demonize it, why don't we all get on the same page and ensure mothers are educated on how to do it as safely as possible?

A. GeorgeHaycock: The fact that some mothers feel it is safer to bed share in the early days unfortunately doesn't make it so. In fact, if you look at the papers I have highlighted, it is precisely in the first few weeks that the risk of bed sharing appears to be the greatest, with odds ratios ranging from about 5 to about 19 in different studies (that is the risk compared with non-bed sharing infants).

I prefer to avoid the term co-sleeping since it means different things to different people: some mean bed sharing, some mean room sharing and some mean any sleeping arrangement whereby a sleep surface is shared by a baby and another person. So, except where a specific question refers to another form of co-sleeping, I will confine my comments to bed sharing.

I know Helen Ball well (and, as it happens, am a member of the steering committee of her latest research project. Unfortunately, because the study is not completed or published I am unable to comment in any way on the progress of that). She and others, such as Jim McKenna of South Bend, Indiana, have shown an association between bed sharing and breastfeeding, but not causation. It may well be that the cause of the association is that when mothers stop breastfeeding they tend to stop bed sharing, rather than the other way around. Unfortunately, it is a fact of epidemiological research that what are called observational studies, which include case-control studies, can only demonstrate association, not causation in either direction.

FSID does give advice on how to make bed sharing as safe as possible and has being doing this for a number of years. On the other hand, we would hardly be doing the public a service by failing to draw attention to research that is already in the public domain. 

Q. Poface: Professor Haycock, the problem I find with this research is that often it seems to regard breastfeeding as no more than an incidental factor, and as actually breastfeeding has a huge effect on infant mortality, this should not be the case. Also, breastfeeding mothers have to make decisions on the safety of their babies based on personal experience.

I was terrified of co sleeping with my first child, and dragged myself out of bed 3 or 4 times a night to breastfeed him on the sofa, exhausted and sleep deprived. I ate snacks, watched TV, to keep myself awake. I stopped doing this when I momentarily fell asleep on the sofa and dropped him aged 6 weeks old. From that point on we co-slept, safely, (no smoking or alcohol or medication, he was not under duvet, bed guard). I am not an isolated case and you must consider this issue when offering advice. Why is there no research focussing solely on breastfed babies and co-sleeping safety?

A. GeorgeHaycock: Although I am as strong a supporter of breastfeeding as anyone, it is actually an exaggeration to say that breastfeeding has 'a huge effect on infant mortality' in developed countries. It has an effect but nowhere near as great as that in developing countries where infection, especially gastroenteritis, is the leading cause of infant mortality (not long ago the WHO estimated that 5 million infants die each year from this cause in the Indian subcontinent alone.

I am not aware that anyone has stated that bed sharing is always the worst option in all circumstances, and I would certainly never criticise anyone who found that it was the only way to cope, as you seem to have done. As I and others have repeatedly said, for low risk parents (especially non-smokers) the risk of bed sharing is very low, but I cannot state in view of the published evidence that it is zero.

There is research focussing (although not solely) on bed sharing safety for breast-fed babies. The German study previously referred to found no interaction between breastfeeding and bedsharing: the risk was increased but from a vanishingly small level to a low level.

Q. hanaflower: I advised a mum against co-sleeping the other day (people ask me about it, because my co-sleeping is well known in the baby groups I go to). The reason I gave her was that she was formula feeding, so her sleep pattern was different. This research does not seem to be well known. The mum was also on the large side so not safe for that reason, but presumably stats covering suffocation are not included in the SIDS data? Or is it impossible to separate them?

A. GeorgeHaycock: Stats covering suffocation are not included in SIDS data, by definition, but there is one fairly recent paper from the USA covering this specifically (Scheers N J et al. Pediatrics (2003) 112: 883-889).


Q. peppapighastakenoverm: With regard to the evidence suggesting using a dummy decreases the odds ratio - did you consider breastfeeding as a covariate or examine the ratio for different groups based on breastfeeding duration?

I would have thought that for an infant who is truly breastfed on demand - kept close to the mother, frequent feeding, no bottles or nipple substitutes - the mothers nipple acts in a number of ways like a dummy. The infant does not sleep deeply and for long periods of time as they wake to feed. The infant would also fall asleep often on the nipple, sucking similarly to a dummy. Often, if co-sleeping, the infant may be latched on for extended periods, not actually feeding, but comfort sucking (for want of a better term).

Is it possible to give an odds ratio for mothers and infants in that situation please?

A. GeorgeHaycock: The careful meta-analysis conducted by Fern Hauck and her associates (O'Connor et al, reference somewhere above!) showed no evidence at all of a negative effect of dummy use on breastfeeding exclusivity or duration in any of the four controlled trials in the published literature. The claim that there is such a (negative) association is based solely on observational studies, which by their nature cannot demonstrate causation and which inevitably (albeit unconsciously) subject to bias.

The protective effect of dummy use is therefore applicable equally to breast and bottle fed infants. It is not possible to give an odds ratio for the specific situation you describe because no studies have been done which addressed that question.

Q. JimmyMcNulty: I have a question about dummies. Your research talks about babies who were 'offered' a dummy and that this cut their risk of SIDS. What does that mean exactly? If the dummy falls out (as my ds's always used to at some point in the night, sometimes in the first five mins after falling asleep) what happens to the risk then?

A. GeorgeHaycock: An excellent question! Research has shown that most dummies do indeed fall out within the first half hour of a sleep period but, strangely or not, the protective effect seems to remain. This reflects, I think, the fact that we really have no idea of how or why dummies are protective. Incidentally, the same is true of the most important intervention yet devised, that of sleeping babies on their backs. We really don't know why this is effective but the evidence is overwhelming that it is.

Ill children

Q. Upwind: I have two questions:

  1. Many parents only co-sleep when their child is particularly unsettled or ill, which may itself be a warning that something is wrong. How do you factor into your research that both, the co-sleeping and the sudden death, may be prompted by an underlying problem?
  2. I have been told that distinguishing between SIDS and eG suffocation at post mortem is often subjective. Does this mean that some of the success of the Back to Sleep campaign may actually reflect changes in pathologists' practice?

A. GeorgeHaycock:  Your first question is covered in a previous answer.

The second point is difficult. I am not sure that the word 'subjective' is really the right one, but there may be differences of opinion on the cause of death in a particular case between different pathologists and, more importantly, different coroners since it is coroners who issue the death certificate in these cases. Most pathologists that I know, who actually do post mortem examinations in SIDS cases, believe they can tell the difference.

Q. treedelivery: I'm interested in the issue of 'ill' babies and the SIDS risk. What happens to the risk, and when is an 'under the weather' baby at increased risk of SIDS?

A. GeorgeHaycock: There are two studies addressing this question in the 'post-back sleeping' era. One is by the German SIDS group (I seem to be quoting them a lot!), which did not find any increased risk for babies with symptoms of minor illness, and another from the University of Leicester (lead investigator Michael Wailoo) which found a small risk.

Why this should be so is speculative, but there is a considerable body of evidence that infection and inflammation may play a role in the causation of at least some cases of SIDS. To find these papers Google 'Pubmed', and put either 'vennemann mm' or 'wailoo m' into the entry box and the studied should come up. In most cases you will be able to access the abstracts directly via Pubmed, but not the full text. If I understand this correctly, it is 'safer' to have the baby in bed isn't it?

Q. tiktok: Sometimes, babies are bed sharing because they are unsettled/ill/miserable, and this may reflect something 'wrong' and it's the something 'wrong' that's increased the risk of SIDS. The bed sharing is incidental. I am not sure that all studies control for this. And as far as I can tell, there are no studies that really control for safe co-sleeping, but I might be wrong there. Co-sleeping with pillows and a drugged mother in hospital seems to me to be poor care.

A. GeorgeHaycock: I have partly dealt with this in earlier postings. Recent evidence is not unanimous as to whether symptoms of minor infection (restlessness, irritability, low grade fever etc.) are risk factors for SIDS in the 'back to sleep' era. However, the recent German study that I have quoted already suggests that it is infants who habitually bed share who are at greatest risk, rather than those who are taken into the parents' bed only occasionally.

tiktok: Thanks, ProfGeorge, but babies can be habitually unsettled and miserable (if not actually habitually 'ill') and habitually bedshare because this unsettledness is dealt with more effectively that way....but there could be a reason for habitual unsettledness which is linked with subsequent SIDS. This then skews the stats. Are their studies which might throw light on this?

Daytime naps 

Q. priyag: Can you give more details about the the latest advice that babies should not be put down to sleep in a seperate room for daytime naps ?

A. GeorgeHaycock: There are I think two recent papers, one from England and one from New Zealand, suggesting that infants should not be put down to sleep alone (that is to say with nobody else in the room) for daytime naps as well as for the night sleep.

As I am in Mumsnet Towers I do not have my database with me so I can't give you the references. However, if you enter Pubmed and put in 'blair ps AND fleming pf' and 'mitchell ea' you should find them (this is the correct syntax, use lower case except for operators such as 'AND'.


Q. Upwind: I had a quick look at the abstracts of the first four studies cited. Study 1 found that: "If the mother smoked, significant risks were associated with bed-sharing, especially during the first weeks of life (at 2 weeks 27·0 [13·3-54·9]). This OR was partly attributable to mother's consumption of alcohol."

The abstracts of 2, 3 make no mention of alcohol consumption. The results and conclusion of 4 do not seem to support FSID advice: "Results: In the multivariate analysis infants who shared their parents' bed and were then put back in their own cot had no increased risk (odds ratio 0.67; 95% confidence interval 0.22 to 2.00). There was an increased risk for infants who shared the bed for the whole sleep or were taken to and found in the parental bed (9.78; 4.02 to 23.83), infants who slept in a separate room from their parents (10.49; 4.26 to 25.81), and infants who shared a sofa (48.99; 5.04 to 475.60). The risk associated with being found in the parental bed was not significant for older infants (>14 weeks) or for infants of parents who did not smoke and became non-significant after adjustment for recent maternal alcohol consumption (>2 units), use of duvets (>4 togs), parental tiredness (infant slept =<4 hours for longest sleep in previous 24 hours), and overcrowded housing conditions (>2 people per room of the house). Conclusions: There are certain circumstances when bed sharing should be avoided, particularly for infants under four months old. Parents sleeping on a sofa with infants should always be avoided. There is no evidence that bed sharing is hazardous for infants of parents who do not smoke."

A. GeorgeHaycock: Study 1 did find risk for non-smoking mothers for infants below the age of 7-8 weeks. This is shown in Figure 1 from that paper, which shows that the lower 95% confidence interval for non-smokers is above the unity risk line below that age. This is not referred to in the abstract but the principal author, Professor Bob Carpenter, has re-analysed the data and now says that it is significant up to 16 weeks. I do not have the paper in front of me, but if you get hold of it you will find it is so. Bob's re-analysis is published in a less well-known journal, which I did not include in my paper because it is not easily available. Incidentally, he no longer supports the conclusion with which you conclude your comment (which is unfortunate, I agree).

As I wrote in my summary, references 4 and 6 found a non-significant trend towards risk for non-smoking mothers (OR 1.3 and 1,35 respectively) but all the others were significant, especially for the younger infants. As far as I know nobody has yet conducted a formal meta-analyis of these studies but it would inevitably come out with a significant summary odds ratio.

Upwind: Thank you for your response. I've just had a quick look at USA Agency Agency for Healthcare Research and Quality report no. 153, which you say provides the best evidence that breastfeeding modestly reduces the risk of SIDS (OR of about 0.64).

They conducted their own meta-analysis, using data on "ever versus never breastfeeding". Given that a large proportion of mothers in the developed world give up breastfeeding in the first few weeks, do you feel that potentially underestimates the association between formula feeding and SIDS? Many of the mothers who have ever breastfed, will have ceased to do so by the time their babies are at an age where they are at increased risk of SIDS. 

Last updated: 9 months ago