Sleep Position and SIDS: Reading the Probability, Not the Headline

7 min read日本語版あり
Audience
Parents of children 0–1 year
Target length
~1,600 words
Status
Draft v2 (translated from Japanese v1)
Original
../22_sids_probability.md

Lead

Most parents have done it: put the baby down, left the room, then gone back a few minutes later to check that the chest is still rising and falling.

The thought hovering in the background is almost always "" — sudden infant death syndrome. Search for it and you get a cascade of declarative statements: prone sleeping is dangerous, bed-sharing is dangerous, breastfeeding protects. None of that is false. But each figure comes with conditions — absolute risk versus relative risk, what an odds ratio actually means, which populations the numbers apply to — and when those conditions get stripped out, parents are left with fear and no workable framework.

This article works through the public health evidence on SIDS: what the back-sleeping recommendation accomplished, how to read the risk factors without confusing relative and absolute numbers, and where the current science genuinely leaves room for household-level judgment.

How the back-sleeping finding changed history

One of the foundational studies establishing the link between sleep position and SIDS is the New Zealand Cot Death Study. Conducted between 1987 and 1990 as a case-control study (485 cases, 1,800 controls), Mitchell and colleagues demonstrated that prone sleeping was a significant independent risk factor for SIDS [1].

The adjusted odds ratio was approximately 4.6 (95% CI 3.4–6.3) for infants usually placed prone, and rose dramatically — to approximately 19 (95% CI 8.2–44.8) — for infants whose caregivers normally placed them supine but placed them prone on the index night [1]. The most dangerous scenario, counterintuitively, was an unfamiliar position tried for the first time.

Around the same period, back-to-sleep campaigns launched in the United Kingdom, United States, Australia, and Japan. After the U.S. Back to Sleep campaign (later renamed Safe to Sleep) began in 1994, SIDS mortality roughly halved over the following decade, with reductions of up to 80% reported in some regions [2]. The back-sleeping recommendation ranks among the most successful public health interventions of the late twentieth century.

In Japan, the Cabinet Office's Children and Families Agency (Kodomo Katei-cho) runs an ongoing SIDS Prevention Awareness Month each November, promoting back sleeping, breastfeeding, and avoidance of passive smoke exposure as the three primary preventive measures [3]. The annual number of SIDS deaths in Japan has in recent years been approximately 50–70 [3].

Separating relative risk from absolute risk

A detour into probability is warranted here.

When someone says "prone sleeping quadruples SIDS risk," the intuitive reading is that prone sleeping is overwhelmingly dangerous relative to back sleeping. For the question of sleep position — back, always — the practical conclusion is indeed correct and not in dispute.

But "fourfold" is a (or odds ratio). The absolute risk is a separate number. SIDS mortality in Japan occurs at approximately one in 6,000 to 7,000 births [3]. A fourfold relative increase applied to that baseline still leaves an absolute risk that, while not zero and therefore worth preventing, is far smaller than the relative figure alone implies.

This distinction matters most when layering multiple risk factors. Carpenter and colleagues conducted an individual-level meta-analysis of five major case-control studies, published in BMJ Open in 2013, examining bed-sharing and SIDS [4]. The adjusted odds ratio for bed-sharing was 2.7 across all ages — but when restricted to the condition where neither parent smokes, the infant is breastfed, and the infant is under three months of age, the estimated absolute risk rose from 0.08 per 1,000 live births (room-sharing, same bed) to 0.23 per 1,000 live births (bed-sharing) [4].

A relative risk exceeding fivefold, yet an absolute risk of 0.23 per thousand. "How much does the risk multiply?" and "How often does the event occur in absolute terms?" are different questions. Carpenter's data answer both simultaneously, which is why the paper remains central to the bed-sharing debate more than a decade later.

Major risk factors and approximate odds ratios

The following summarizes the principal risk factors, drawing primarily from the 2022 AAP evidence review [5], the Hauck and colleagues breastfeeding meta-analysis [6], and Carpenter and colleagues 2013 [4]:

The 2022 AAP policy statement consolidates these into a core set of recommendations [5]: supine sleep position, firm and flat sleep surface, room-sharing on a separate sleep surface, elimination of soft bedding and loose objects, avoidance of overheating, avoidance of tobacco and alcohol and drugs, promotion of breastfeeding, and pacifier use.

A point worth underscoring: the risk factors act independently. No household perfectly follows every recommendation every night. The practical implication is that adhering to most factors on most nights meaningfully lowers overall risk — one imperfect night does not void the benefit of consistent practice elsewhere.

Bed-sharing: where guidelines still diverge

Bed-sharing remains the most contested point across international guidelines.

The AAP has consistently recommended against bed-sharing through its 2022 update [5]. Some UK guidelines and breastfeeding-support bodies have taken a more conditional position, noting that when neither parent smokes, neither is under the influence of alcohol or sedating drugs, and neither is sleeping on a sofa or armchair, the risk associated with feeding-related bed-sharing may be smaller than initially estimated.

The Carpenter 2013 data sit at the center of both arguments [4]. "The risk does rise, but in low-risk conditions the absolute magnitude remains small" — that finding is cited by both sides.

What matters for the individual household is which conditions apply. Both parents smoke, the household is prone to sofa sleep-overs, and the infant is in the newborn period: Carpenter's tables show clearly elevated absolute risk. Non-smoking parents, infant over six months, a firm mattress, deliberate same-bed arrangements: the absolute risk is limited. In neither case does the baseline recommendation change — back sleeping, firm surface, soft-bedding elimination. What changes is how conservatively to interpret the remaining margin.

This is not unusual. The public health optimum and the household-level optimum sometimes diverge; that divergence does not make either position dishonest.

Two things you can do tomorrow

Two practical steps, rather than a longer list.

First: write down the household sleep setup in two lines and share it with anyone who puts the baby to sleep. Something like "back position, crib only, room temperature 20–22°C, sleep sack instead of blanket." Risk tends to rise most at handoffs — when the caregiver changes and the sleep position or bedding changes silently with it.

Second: log near-miss observations along with the emotional response. "Found the baby rolled slightly sideways." "The blanket had shifted toward the face." These near-misses do not directly prevent SIDS, but they are operational data. Recorded in an app like Memori with a date and note, they reveal patterns in how the household's own practices drift over weeks — and they give a concrete agenda for conversation with a pediatrician or public health nurse.

On nights when anxiety runs high, the more productive choice is usually to put a call to the pediatrician or local health center on tomorrow's list, rather than recalculating SIDS probabilities online at midnight. Carrying worry alone costs more than sharing it.

Summary

SIDS risk is probabilistic, shaped by multiple independent factors — sleep position, bedding, smoking, breastfeeding, bed-sharing, and infant age — each contributing separately [4,5,6]. The back-sleeping recommendation is among the most effective public health interventions of the late twentieth century [2].

When reading risk numbers, separate relative from absolute. "X times higher" is alarming by design; what governs household decisions is the absolute magnitude [4]. And because the factors are independent, consistent adherence to most of them on most nights produces meaningful risk reduction even when the picture is imperfect.

The goal is not perfection. Covering seven out of ten risk factors reliably is, in practice, substantially protective.

References

  1. Mitchell EA, Scragg R, Stewart AW, et al. Results from the first year of the New Zealand cot death study. N Z Med J. 1991;104(906):71–76. PMID: 2020450.
  2. Moon RY, Carlin RF, Hand I; Task Force on Sudden Infant Death Syndrome and the Committee on Fetus and Newborn. Sleep-Related Infant Deaths: Updated 2022 Recommendations for Reducing Infant Deaths in the Sleep Environment. Pediatrics. 2022;150(1):e2022057990. doi:10.1542/peds.2022-057990.
  3. Children and Families Agency, Cabinet Office, Japan (Kodomo Katei-cho). Sudden Infant Death Syndrome (SIDS). https://www.cfa.go.jp/policies/boshihoken/kenkou/sids
  4. Carpenter R, McGarvey C, Mitchell EA, et al. Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case-control studies. BMJ Open. 2013;3(5):e002299. doi:10.1136/bmjopen-2012-002299. PMID: 23793691.
  5. Moon RY, Carlin RF, Hand I; Task Force on Sudden Infant Death Syndrome and the Committee on Fetus and Newborn. Evidence Base for 2022 Updated Recommendations for a Safe Infant Sleeping Environment to Reduce the Risk of Sleep-Related Infant Deaths. Pediatrics. 2022;150(1):e2022057991. doi:10.1542/peds.2022-057991. PMID: 35921639.
  6. Hauck FR, Thompson JM, Tanabe KO, Moon RY, Vennemann MM. Breastfeeding and reduced risk of sudden infant death syndrome: a meta-analysis. Pediatrics. 2011;128(1):103–110. doi:10.1542/peds.2010-3000. PMID: 21669892.
  7. Mitchell EA, Hutchison L, Stewart AW. The continuing decline in SIDS mortality. Arch Dis Child. 2007;92(7):625–626. doi:10.1136/adc.2007.116194. PMID: 17405855.

Related