What the evidence supports for reducing the risk of sudden infant death

The single most consistent finding in the science of sudden infant death is that where and how a baby is placed to sleep changes the risk. The American Academy of Pediatrics' 2022 guidance distils decades of study into a short list — put the baby on its back, on a firm, flat, non-inclined surface, in the parents' room but not the parents' bed, and keep soft bedding out of it [s1]. About 3,500 infants still die of sleep-related causes in the United States each year, a toll that fell sharply in the 1990s and has not moved since 2000 [s1] [s2].

That plateau is the reason the guidance keeps being restated rather than retired. This article reports what the AAP recommends and the reasoning behind it; it is not medical advice, and clinical questions about an individual infant belong with a paediatrician.

What "sleep-related infant death" covers

The 3,500 figure is not sudden infant death syndrome (SIDS) alone. It bundles three categories that share a sleep setting: SIDS proper, coded R95; ill-defined deaths, coded R99; and accidental suffocation and strangulation in bed, coded W75 [s1]. Grouping them matters, because a death once recorded as SIDS may now be classified as suffocation, so the labels shift even when the underlying problem — an infant dying in an unsafe sleep environment — does not.

The AAP frames the mechanism as a "triple risk model": SIDS tends to occur when an intrinsically vulnerable infant, often one with impaired arousal or cardiorespiratory and autonomic responses, meets an external trigger such as an unsafe sleep surface during a narrow developmental window [s1]. The model explains why no single factor is decisive and why the recommendations target the one link a caregiver can actually change — the environment.

The core recommendations

Four elements sit at the centre of the 2022 statement. Infants should be placed for sleep on the back — supine — for every sleep [s1]. They should sleep on a firm, non-inclined surface; the guidance specifically warns against inclined products, a change reflecting deaths tied to sitting and inclined sleepers [s1]. They should room-share without bed-sharing, meaning the infant sleeps on a separate surface in the caregiver's room [s1]. And soft bedding — pillows, quilts, bumpers, soft toys — and overheating should be avoided [s1].

The bed-sharing line is where the guidance is most often contested, and it is worth being precise about what the AAP does and does not say. Its position is that room-sharing lowers risk while bed-sharing raises it, particularly alongside other exposures. Health Newspapers has examined the underlying bed-sharing literature and the debate over how absolute that advice should be in a separate piece on co-sleeping and the evidence; the 2022 statement itself adds new detail on short-term emergency sleep locations and on the circumstances that make bed-sharing especially hazardous [s1].

The secondary list, and its uneven evidence

Beyond the sleep surface, the AAP lists several measures it says further reduce SIDS risk: human-milk feeding; avoidance of nicotine, alcohol, cannabis, opioids and illicit drugs before and after birth; routine immunisation; and pacifier use at sleep [s1]. The strength of evidence behind these is not uniform, which the accompanying evidence-base report is candid about — the recommendations rest on a mix of trial and observational data of varying quality [s2].

Two points deserve emphasis for a reader trying to weigh them. First, the associations for measures such as pacifier use and breastfeeding come largely from observational studies, where families who adopt one protective behaviour tend to adopt others, so the isolated effect of any single measure is hard to pin down — the same confounding problem that clouds the wider breastfeeding outcomes literature. Second, the smoking and substance-exposure link is among the more robust, and it operates both prenatally and in the household after birth [s1].

Why the message hasn't closed the gap

If the guidance is sound, why has the death rate been flat since 2000? The evidence-base report points to two things: adherence is imperfect, and disparities persist, with markedly higher rates in some populations than others [s2]. A recommendation only works if it is followed, and the AAP notes that unsafe surfaces, soft bedding and unplanned bed-sharing remain common in the settings where deaths occur [s1].

The honest summary is that the science supports a small number of environmental changes with reasonable confidence, supports a second tier of measures more weakly, and cannot promise that any of them eliminates risk for a given infant. What the plateau shows is not that the advice is wrong but that knowing it and living it are different problems.