US guidelines warn against bed-sharing. Here is the evidence they rest on.

American guidelines advise parents to sleep in the same room as their baby but not in the same bed, and the recommendation rests on case-control studies linking bed-sharing to a higher risk of sudden infant death. That risk is real and rises steeply when a parent smokes or has been drinking, but for a breastfed infant of non-smoking, non-drinking parents the absolute risk remains very low in both arrangements — which is exactly why the topic is contested rather than settled [s1] [s2].

This is a subject where families and guidelines often diverge, and where the honest job is to lay out what the evidence shows and how it is weighed, not to tell any family what to do. A sleeping arrangement for an infant is a decision for parents and their clinician.

What the guideline says

The American Academy of Pediatrics' 2022 policy statement opens with the scale of the problem: each year in the United States, about 3,500 infants die of sleep-related deaths, a category that includes sudden infant death syndrome (SIDS), ill-defined deaths, and accidental suffocation and strangulation in bed [s1]. After a sharp decline in the 1990s, the overall rate has remained stagnant since 2000 [s1].

The AAP's core recommendations for a safe sleep environment are: placing the baby on its back to sleep; using a firm, non-inclined sleep surface; room-sharing without bed-sharing; and avoiding soft bedding and overheating [s1]. Alongside these it lists factors associated with lower risk, including human-milk feeding, routine immunisation, pacifier use, and avoiding exposure to nicotine, alcohol, marijuana, opioids, and illicit drugs [s1]. The 2022 update explicitly revised its guidance on bed-sharing and, notably, added information intended to help parents and clinicians assess the risk of specific bed-sharing situations rather than treating them as uniform [s1].

The conceptual model behind the advice is the "triple-risk" hypothesis: SIDS is thought to occur when an intrinsically vulnerable infant meets an external trigger, such as an unsafe sleep environment, during a critical developmental window [s1]. The sleep-environment recommendations target the one element of that trio a family can change.

What the bed-sharing evidence actually shows

The strongest quantitative evidence on bed-sharing comes from an individual-level pooled analysis of five major case-control studies, combining data on 1,472 SIDS cases and 4,679 control infants across 19 studies in the UK, Europe, and Australasia [s2]. In the combined dataset, 22.2% of the babies who died had been bed-sharing during their last sleep, versus 9.6% of the living controls, giving an adjusted odds ratio across all ages of 2.7 (95% CI, 1.4 to 5.3) [s2].

But the analysis's central contribution was to separate out the highest-risk circumstances. When neither parent smoked, the baby was under three months, breastfed, and had no other risk factors, the adjusted odds ratio for bed-sharing versus room-sharing was 5.1 (95% CI, 2.3 to 11.4) [s2]. Smoking and alcohol use, the authors reported, greatly increased the risk of bed-sharing [s2]. This is the finding most often lost in the debate: the relative risk of bed-sharing is highest, not lowest, among the very young, and it is amplified by smoking and drinking.

Relative risk versus absolute risk

The same study is also the source most often cited by those who think the guidance overstates the danger — because it reported absolute risks, and they are small. For that low-risk group (breastfed infants under three months of non-smoking, non-drinking parents with no other risk factors), the estimated absolute risk was 0.08 per 1,000 live births when room-sharing (95% CI, 0.05 to 0.14) and 0.23 per 1,000 when bed-sharing (95% CI, 0.11 to 0.43) [s2].

Both numbers are low. A near-tripling of a very small risk is still a very small risk in absolute terms — and it is also a real increase across a population of millions of infants. Which framing matters more is a genuine value judgment, not a fact the data resolve: public-health guidance tends to weigh the population-level increase, while an individual family may weigh the low absolute number, alongside benefits they perceive from bed-sharing such as breastfeeding and sleep. The evidence supplies both numbers; it does not adjudicate between the two ways of reading them.

Where guidance and practice diverge

The gap between the recommendation and what families do is well recognised, and the 2022 AAP update responded to it not by softening the core advice against bed-sharing but by adding detail on how specific situations differ in risk — an acknowledgement that many parents will bed-share at least sometimes and are better served by graded information than by a single prohibition [s1]. The circumstances the evidence flags as most dangerous are consistent across sources: bed-sharing with a parent who smokes, who has used alcohol or sedating drugs, on a sofa or armchair, or with a very young infant [s1] [s2].

What the evidence establishes is that bed-sharing raises SIDS risk in relative terms, most steeply for the youngest infants and when smoking or alcohol are involved, while the absolute risk for low-risk breastfed infants of non-smokers is small [s1] [s2]. What it does not do is make the decision for any family.

This article is informational and is not medical advice. Questions about a safe sleep arrangement for a specific infant, particularly a premature or very young baby, are best discussed with a clinician who knows the family's circumstances.