Take the 2-minute tour ×
Skeptics Stack Exchange is a question and answer site for scientific skepticism. It's 100% free, no registration required.

Among some circles, co-sleeping is highly advocated for newborns/infants due to ease of breastfeeding and potential developmental bonding to the mother (for more examples, read about proposed advantages here). The immediate and obvious red flag that came to mind was rolling over one's baby. A defense I've heard is that you... just won't:

[In response to inquiry about the safety of co-sleeping] You won't squash him. You couldn't even comfortably roll onto a teddy bear in your sleep, let alone your own baby that your subconscious is ALWAYS aware of. (source)

you wont roll onto your baby. there is no way you forget they are there even when you're dead to the world. (source)

Physically and psychologically it is HIGHLY unlikely that you will roll over onto your baby when co-sleeping. It’s an evolutionary, parental instinct sorta thing (don’t remember the exact name for this but definitely learned about it in my psych classes). Essentially as a parent your conscious and unconscious self is aware that protecting your baby is a priority and you won’t smush or smother them. (source)

However, according to the Consumer Product Safety Commission, via Kid's Health:

According to the CPSC, at least 515 deaths were linked to infants and toddlers under 2 years of age sleeping in adult beds from January 1990 to December 1997:

  • 121 of the deaths were attributed to a parent, caregiver, or sibling rolling on top of or against a baby while sleeping
  • more than 75% of the deaths involved infants younger than 3 months old

Here's another pro-co-sleeping advocate (emphasis mine):

At the University of Notre Dame's Mother-Baby Behavioral Sleep Laboratory, our studies of breastfeeding mothers who sleep with their 2- to 4-month-olds reveal that both mothers and their babies are extremely sensitive throughout the night to each other's shifting position in the bed.

During my many years of studying sleep-sharing, I've never heard of a single instance in which, under safe conditions, it was proven that a mother suffocated her child. Notice that I said safe conditions: Babies can and do accidentally suffocate when one or both parents doesn't know a baby is in the bed, is drunk or desensitized by drugs, or is indifferent to the baby's presence.

So, to follow in the vein of the last bit:

Is co-sleeping safe when parents are aware of the baby's presence, are not under the influence of substances, and are not indifferent to the baby's presence?1

In layman's terms: is co-sleeping safe under normal conditions?

1 Is that criterion last just a catch-all no true Scotsman variant? If the first two conditions aren't true then the parents must have been indifferent at some level?

share|improve this question
I think any answer should be in terms of relative risks, against the alternatives. Are the tragic deaths of 515 infants and toddlers comparatively fewer than the deaths of children in separate beds? –  Oddthinking Aug 30 '12 at 2:10
Think about the numbers, too. ~Five hundred deaths over seven years is about 0.2 deaths/day. With how many people cosleep, that really doesn't seem like a very high rate. –  William Grobman Aug 30 '12 at 2:36
Purely anecodtal: We were advised by, oh, pretty much every medical professional we came in contact with, that co-sleeping should be avoided at all costs due to the massive increase in SIDS risks. Additionally, and not relevantly, co-sleeping is a pain in the royal arse. Ever tried sharing a bed with a 2-year-old? It's a nightmare and then they want to come back every single night. Never again! –  Mark Henderson Aug 30 '12 at 3:31
@jozzas: I am not sure I understand your objection. It is not SIDS versus rolling over. It is all mortality (non-sharing) versus all mortality (sharing), plus compensation for as many confounding variables as can be come up with. –  Oddthinking Aug 30 '12 at 4:42
@Bill: Of course, your family has my sympathy, and 515 deaths is 515 individual tragedies too many. But we shouldn't make important decisions based only on anecdote: I do not consider 515 deaths to be low (or high), until we have some number to compare it against. If that same population of children, sleeping separately, would lead to 1000 more deaths (a totally invented number), then 515 deadth would actually be low, and we should encourage co-sleeping. If it would lead to 5 deaths, then 515 is terribly high, and we should discourage co-sleeping. –  Oddthinking Sep 1 '12 at 2:11

2 Answers 2


Co-sleeping is unsafe, particularly when compared to placing a child into a suitable cot / crib. Most of this risk comes from the bed and bedding not being suitable for infants, but suffocation by overlying is also a significant risk.

The only safe place for an infant to sleep is on its back in a crib/bed that meets relevant standards.

A retrospective review of death-scene and medical reports for SIDS and related deaths showed that children sleeping in adult beds increases their risk of death by at least a factor of 20:

Using cribs as the reference group and adjusting for potential confounders, the multivariate ORs showed that ... the risk of suffocation was approximately 40 times higher for infants in adult beds compared with those in cribs. The increase in risk remained high even when overlying deaths were discounted (32 times higher) or the estimate of rates of bedsharing among living infants doubled (20 times higher).

Instances of overlying (suffocation by a person sharing the bed) were rare, and the data is not as conclusive:

The diagnosis by medical examiners and coroners that overlying of an infant while sharing an adult bed was the “cause of death” remains controversial. More overlying deaths were reported by medical examiners and coroners in the 1990s (70 deaths) than in the 1980s (7 deaths). In approximately 40.3% of the cases (31 of 77), the narratives reported that a third party found the infant covered by an adult or a child, there were compression marks on the infant, or other findings suggesting the likelihood of overlying (eg, infant sleeping in twin bed with 2 adults). In both decades, overlying deaths were associated with very young infants, with an average age of 1.9 months. Only 1 overlying death occurred after 6 months of age, a report of a 10-month-old found with another child over him.

That said, this is not the major risk with co-sleeping, it's suffocation by other means - bedding, soft mattresses, a child getting wedged between the headboard/footboard/bedframe and the mattress.

Another population-based death-scene study of SIDS and related deaths found that for their selected records (all sleep-based deaths of children under two in a particular geographical area), nearly half involved bed sharing:

Deaths Occurring While Sharing a Bed or Other Sleep Surface

Nearly one half (56) of the infants (47.1%) died while sharing a sleep surface with one or more bedmates (1.4 ± .7; range: 1–4 bedmates; Table 3). For the majority, deaths while bedsharing were diagnosed as SIDS, but for 13 the diagnoses were suffocation or undetermined (23.2% of bedsharing deaths). All deaths occurred on sleep surfaces that were not designed specifically for infant sleep. In 13 cases (23.2%), the scene investigation showed evidence for entrapment of the infant, either by a bedmate or by the sleep surface. In 18 cases (33.0%), the bedsharing infant was found dead on a pillow or comforter, items specifically identified in earlier studies as bedding that increases risk for sudden death when used by infants.18,36,38 The pillows and comforters were on the shared sleep surface and the infant had been placed on top of them.

Controlling for smoking is also an important factor when assessing the risks of bedsharing, but it appears there is a definite effect when this is taken into account:

The impact of bedsharing on risk for sudden infant death remains controversial. Three case–control studies suggest that bedsharing increases risk for sudden death,24,26,45 but the risk is lessened when the high rate of maternal smoking in these studies is considered. In England, in particular, the rate of smoking among mothers whose infants died while bedsharing is so high that the risk for nonsmoking mothers cannot be calculated from the data.24 In the United States, a case–control study27 from Washington, DC showed increased risk especially when black infants bedshare. Finally, preliminary results from the Chicago Infant Mortality Study, a large, recent case–control study, strongly indicate an effect of bedsharing that is independent of smoking. 46 There are no recent published results addressing risk for infants sleeping alone outside of cribs, but data from the US Consumer Products Safety Commission suggest that the risk may be high...

With specific regard to alcohol consumption and bed sharing, a nationwide case-control study in New Zealand found that maternal alcohol consumption did not increase the risk of death while bed sharing, though it should be noted that there doesn't appear to be any conclusive research on this matter - there are studies whose conclusions are directly contradictory.

Neither maternal alcohol consumption nor the thermal resistance of the infant's clothing and bedding interacted with bed sharing to increase the risk of sudden infant death, and alcohol was not a risk factor by itself.

CONCLUSION--Infant bed sharing is associated with a significantly raised risk of the sudden infant death syndrome, particularly among infants of mothers who smoke.

share|improve this answer
nearly half involved bed sharing would actually mean its safer to share the bed ... –  Stefan Aug 30 '12 at 13:24
@Stefan - I do not know if that is true but that statistic definately does not stand on its own merit in this answer. –  Chad Aug 30 '12 at 13:54
This is a bit of a strawman though. No cosleeping proponent I've met says to use bedding (other than perhaps a pillow) until the baby is old enough to turn over by themsleves. I think it's obvious that putting a kid in bed with things that can smother them is dangerous. Your overlying stats didn't mention if it controlled for intoxication either, which was specifically what was asked. –  William Grobman Aug 30 '12 at 15:19
@Stefan most people (I'd say the vast majority) don't do bed sharing, so I don't think it would mean that it's safer. Bed sharing deaths are overrepresented in the data. I'll try to dig up a stat that covers percentage of time children bed share / don't bed share, but I don't like my odds of finding anything. –  jozzas Aug 30 '12 at 22:02
@WilliamGrobman Many of the deaths come from children being wedged between a pillow / the wall / a headboard / the mattress / the bedframe, etc. There are some pretty gruesome pictures in the studies I've linked, that occurred in what would look like fairly safe sleeping situations to most people. Soft bedding like blankets is absolutely not the only risk, though the fact that all deaths in the second study occurred on surfaces / situations not designed for infants to sleep in is somewhat telling. –  jozzas Aug 30 '12 at 22:05

I think it's worth providing an alternative viewpoint to this question.

Outside of the Western, developed world co-sleeping is the standard way for parents to sleep with young children, not the exception.

This article's fonts makes me want to burn my eyes out, but the final paragraph has a very interesting graphic created from NIMH data:

enter image description here

In China, where I live, co-sleeping is so common as to be the assumed situation. To do otherwise is strange and discussion-worthy.

This NYT article from 2007 says data is inconclusive.

Another article mentions author Margot Sunderland quoting:

“In the UK, 500 children a year die of Sids,” Sunderland writes. “In China, where it [co-sleeping] is taken for granted, Sids is so rare it does not have a name.”

She seems to have decent credentials.

share|improve this answer
The reference to SIDs is a red herring. There is no medical or scientific evidence linking SIDs with co-sleeping or vice versa. The few groups that have discussed it all all say that co-sleeping would make SIDs more likely, not less. Never mind the total lack of data by Sunderland et al. Comments like "so rare it doesn't have a name" are useless without any real data and could easily be based on half a dozen different factors such as simply not having good reporting systems in China. –  shtimss1970 Sep 3 '12 at 14:12
Also the lack of SIDS in China is probably more due to lack of reliable statistical information gathering, rather than a lack of SIDS deaths. It's typical that when China is an outiler in data, it's usually because insufficient data is being collected. That said, your original premise may still be valid. –  Mark Rogers Sep 3 '12 at 15:27
@Mark Rogers - I think your eagerness to dismiss the China & HK data is unfounded. As immediately visible in the link, the study the chart data comes from was run by the Chinese University of Hong Kong. If anyone can get sufficient data it's probably them. I offered this alternate answer because I see no evidence in the No answer to support the strong assertion being made, and because there are probably more infants co-sleeping as I write this than there are people in the entire U.S., the fact that all studies in the answer above are focused on the Western world is telling, I think. –  Andrew Heath Sep 3 '12 at 21:51
The question cites 5 US and 1 UK site promoting the practice for mothers in their countries. Citing data from a country with completely different sleeping practices wouldn't be addressing the claim adequately, IMO. Would adding "bedsharing is okay if you use a rolled up towel instead of a pillow, sleep on a thin mat on the floor, don't use bedding, don't smoke and aren't overweight" (a 'typical' Japanese arrangement) be helpful to many people in the US/UK? –  jozzas Sep 3 '12 at 22:56
Smoking is common, yet it's unsafe. Eating fast food is common, yet it's unsafe. Etc. Not all that is common is safe! The rest of the answer is equally based on poor references. –  Sklivvz Sep 4 '12 at 6:38

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.