Blair et al. 2009. Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. BMJ. 339:b3666
The Southwest Infant Sleep Scene (SWISS) study
A case control study (80 cases; 82 high risk controls; 87 random controls) conducted in the South-West region of England, investigating all unexpected deaths of infants under 2 years between 2003-2006.
- 43 of 79 (54%) SIDS infants were found co-sleeping with an adult; of these 13 coslept on a sofa, and 24 slept with an adult who had consumed drugs or more than 2 units of alcohol.
- “The combination of recent maternal alcohol consumption and cosleeping with an infant on a bed or sofa were the strongest predictors of SIDS in the multi-variable model, regardless of which control was used as the comparison.” (2009: 6)
- “A decreased but significant risk was associated with cosleeping in the absence of alcohol or drug use, although this included infants who coslept on a sofa. The proportion of SIDS infants found cosleeping in a bed with parents who had drunk two units or less of alcohol and taken no drugs was no different from that of the random control infants (18% v 16%)”. (2009: 6-7)
- Smoking was a significant factor, with 60% of mothers of SIDS infants smoking in pregnancy compared to 14% of random controls. Use of pillows, swaddling, preterm birth and being in fair or poor health were also significant risk factors regardless of which control group was used for comparison.
- Comparison of SIDS infants with random and high risk control groups indicated that the factors identified were specific circumstances related to risk, rather than simply surrogate markers for socio-economic deprivation.
A ‘high risk’ control group selected on the basis of several epidemiological risk factors was included, to investigate whether some hypothesised risk factors might in fact be markers of socio-economic deprivation.
For statistical reasons (due to some small sub-sample sizes) some analyses of ‘co-sleeping’ included both bed-sharing and sofa-sharing; this led to considerable misinterpretation of the results of this study by the media.
The association of SIDS with alcohol consumption, sofa-sharing and smoking was confirmed by the results of this study. There was no evidence to support the suggestion that bed-sharing in the absence of these was an independent risk factor. The authors suggested that advising against all bed-sharing might have harmful consequences, including parents choosing to sleep on a sofa instead.
Vennemann et al. 2009. Sleep Environment Risk Factors for Sudden Infant Death Syndrome: The German Sudden Infant Death Syndrome Study. Pediatrics. 123(4):1162-1170
Further analysis of data from the German Sudden Infant Death Syndrome (GeSID) study.
- Did NOT find lone sleeping to be associated with increased risk (aOR: 1.72, CI 0.97-3.04)
Factors associated with increased risk:
- prone position (aOR: 7.08, CI 3.69-13.60)
- bed sharing (aOR: 2.73, CI 1.34-5.55) (especially < 13 weeks = aOR: 19.86, CI 2.33-169.50)
- duvets (aOR: 2.20, CI 1.21- 4.00] [thick] and 1.92, CI 1.07-3.45] [thin] vs. a sleeping bag or thin cotton blankets)
- prone on sheepskin (aOR: 27.92, CI 6.45-120.91)
- sleeping in the house of a friend or a relative (aOR: 4.39, CI 1.11-17.38)
- sleeping in the living room (aOR: 2.41,CI 1.06-5.51) increased the risk for SIDS.
Factor associated with decreased risk:
- Dummy use associated with reduced SIDS risk (aOR: 0.39, CI 0.25- 0.59)
NOTE: It is not clear whether ‘living room’ = sofa; alone or in company; or includes other sleep locations such as pram, bouncy chair, bean bag, cot, floor.
High OR obtained for infants bed-sharing under 13 weeks, but no information given re: bed-partners; correction for postnatal smoking, alcohol or drug use, breastfeeding status, thermal environment (clothing), parental tiredness.
Definition of breastfed = breastfed for 2wks+, even if not breastfed at time of death.
Unusual amongst large case-control studies as lone sleeping not found to be associated with increased risk of SIDS.
Some problematic logic: “A limitation of the study was that we did not collect data on parental alcohol consumption, as this might impair the parent’s ability to respond to the infant. Many bed sharing infants are breastfed. It was reassuring that toxicology screening for alcohol and central acting drugs in the cases were negative in all cases (unpublished data), which suggest the infants were not influenced or impaired by these drugs, and that parental alcohol consumption was unlikely to be an important factor, at least among breastfeed infants.” It is clear that this logic is flawed: a) alcohol consumption by a breastfed infant’s father will not appear in infant toxicology tests; b) alcohol consumed by a mother after breastfeeding will not appear in infant toxicology tests; c) alcohol consumed by a mother whose infant is fed expressed human milk or formula on the night of death will not appear in infant toxicology tests; d) alcohol consumed by a mother whose infant WAS breastfed for 2 weeks but is not currently breastfed (the definition of breastfed used here) will not appear in infant toxicology tests.
Additionally no data on the clothes worn by infants were collected. Both thermal environment and alcohol consumption are considered to be important factors related to SIDS risk.
Ball et al. 2006a. Randomised trial of infant sleep location on the post-natal ward. Archives of Disease in Childhood. 91:1005-10
A randomised controlled trial (RCT) in a tertiary-level UK hospital.
Objective: To examine how mother-infant sleep contact might contribute to the establishment and continuation of breastfeeding Summary: Overnight videos were made of mother-baby dyads randomised to 3 sleep locations for their postnatal ward stay: (1) baby in the standard cot at mother’s bedside; (2) baby in a side-car crib attached to mother’s bed; (3) baby in mother’s bed with rail attached to bedside- known as the cot, crib, and bed conditions, respectively.
- Babies in the bed or crib exhibited significantly more frequent attempted and successful feeds than those infants in the cot
- There were no significant differences found in feeding frequency measures between the bed and crib conditions
The use of the stand-alone cot impeded breastfeeding by introducing a barrier between mother and baby preventing contact; inhibited the baby’s ability to root and initiate suckling; obscured the baby’s cues from the mother; and by its height prevented mothers from retrieving their babies without either assistance or the need to get out of bed, thereby substantially hampering the ease and speed of maternal response.
Blair et al. 2006. Sudden Infant Death Syndrome and the time of death: factors associated with night-time and day-time deaths. International Journal of Epidemiology. 35:1563-1569
The Confidential Enquiry into Stillbirths and Deaths in Infancy (CESDI) Study
Objective – to investigate how SIDS deaths in the CESDI sample (325 SIDS, 1300 controls) were distributed day vs night, and their association with known risk factors
- Most deaths occurred during night-time sleep (83%).
- Most risks associated with SIDS were significant for both day and night-time sleep.
- Position: The risk associated with side sleeping was more marked for day-time deaths. Nearly half of day-time SIDS infants who were placed on their side to sleep were found prone.
- Smoking: The risk associated with fathers’ smoking was more marked for night-time deaths (mOR = 3.25, CI 1.88-5.62).
- Lone sleep: Lone sleeping was found to be an important risk factor for day as well as night-time deaths (75% of day-time deaths occurring in this environment – mOR=10.57, CI 1.47-75.96), and more infants sleeping alone were found with their heads covered compared to infants sleeping accompanied (24.8% vs 11.3%).
Analysis reinforces the importance of room-sharing and supine position for day-time, as well as night-time sleep.
Main message is that it is not the cot in parents’ room environment that is protective, but the presence of a parent or caregiver in that room. Adult supervision may prevent head covering or rolling prone.
Carpenter, RG. 2006. The hazards of bed sharing. Paediatr Child Health. 11 (Suppl A): 24A-28A
Further analysis of data from The European Concerted Action on SIDS ‘ECAS’ Study (for study details see here)
- Age at which adjOR associated with bed-sharing for non-smokers becomes significant revised to <7weeks (down from <8 weeks in Lancet publication) (at 2 weeks adjOR 2.3, CI 1.2-4.5).
- Inclusion of confounders not recorded by all sites in the study resulted in non-smoking adjOR for bed-sharing increasing to 5.1, 1.9-13.6 at 10 weeks.
- Bed-sharing was associated with increased duration of breastfeeding
- Full breastfeeding halved risk of SIDS
Good definitions of bed-sharing with relation to variability in behaviour, and excluding sofa-sharing.
Smoking refers to smoking in pregnancy only, with no variable for smoking postnatally
Numbers start getting very small for non-smokers; only 16 non-smoking cases aged under 10 weeks (17.8%)
Fails to consider that the significant OR associated with bed-sharing for non-smokers <7weeks may be affected by confounders NOT in the data-set (such as bed-partners, tiredness, drug/medication use, postnatal smoking).
Modelling of cumulative mortality based on bed/not-bed with breast/bottle feeding leads the author to conclude that “it may be safer to bottle-feed and not bed share than to breastfeed and bed share, although the difference is not significant” (p27A). Although a numerical value of this non-significant difference is not provided, examination of the figure indicates that the increase in death rate in this scenario is to 1/10,000 from 0.5/10,000.
McGarvey et al. 2006. An 8 year study of risk factors for SIDS: bed-sharing versus non-bed-sharing. Archives of Disease in Childhood. 91:318-323
A case-control study conducted in Ireland between 1994 and 2001. 287 cases and 831 controls.
Objective – to evaluate the effect of bed-sharing on risk factors for SIDS.
- Increased risk of SIDS with bed-sharing (mOR 3.53, CI 1.4-8.93)
- In analysis by infant’s age, including adjustment for factors including maternal smoking and drinking, this increased risk remained significant for infants <10 weeks (mOR 8.02, CI 1.97-32.52) but not for older infants.
- The OR associated with bed-sharing was greater for infants found between two adults (mOR 4.68, CI 1.09-19.99) than next to one adult (mOR 3.29, CI 1.05-10.26)
The risk associated with bed-sharing was significantly greater where:
- Tog of clothing/bedding ≥10 (aOR 6.14, CI 1.1-34.42)
And was significantly reduced with:
- Greater weight for gestation at birth (by z scores) (aOR 0.37, CI 0.14-0.97)
- Maternal smoking was a risk factor for infants who did not bed or sofa-share (OR 3.74, CI 2.31-5.99)
- 87% (109/126) of bed-sharing SIDS cases had mothers who smoked during pregnancy, compared to 17% (17/101) of controls
- The risk associated with maternal smoking for bed-sharing infants was greatly increased (OR 24.78, CI 15.79-38.86).
Detailed analysis of factors and interactions with bed-sharing. Separation of bed-sharers from combined bed/sofa-sharers.
Very small number of non-smoking bed-sharers in the SIDS group makes it difficult to precisely assess the risk associated with bed-sharing for non-smokers.
Tappin et al. 2005. Bedsharing, Roomsharing, and Sudden Infant Death Syndrome in Scotland: A case-control study. The Journal of Pediatrics. 147:32-37
A case-control study conducted in Scotland between 1996 and 2000. 123 cases and 263 controls.
Objective – to examine the hypothesis that bed-sharing is associated with an increased risk of SIDS.
- Sharing a sleep surface was associated with an increased risk of SIDS (OR 2.89, CI 1.40-5.97)
- The largest risk was associated with sofa-sharing (OR 66.9, CI 2.8-1597)
- Bed-sharing was associated with a greater risk of SIDS for younger infants <11 weeks (OR10.2, CI 2.99-34.8) but not for older infants >11 weeks (OR 1.07, CI 0.32-3.56)
- The risk associated with bed-sharing remained significant for non-smokers (OR 8.01, CI 1.2-53.3) and breastfeeders (OR13.1, CI 1.29-133)
- In contrast to the CESDI study, sleeping alone in a room was not found to be associated with an increased risk of SIDS in the sample as a whole (OR 1.32, CI 0.67-2.60). The risk of lone sleeping was significant only for infants of smoking parents (OR 12.2 CI 2.25-66.4)
Analysed data based on feeding method.
Infants who bed-shared for part of the night but subsequently died whilst in a cot (1/4 of the ‘bed-sharing’ cases) were counted as bed-sharing deaths. No data collected on smoking in pregnancy, alcohol consumption or bed-covers used.
Vennemann et al. 2005. Modifiable risk factors for SIDS in Germany: Results of GeSID. Acta Pædiatrica. 94:655-660
The German Sudden Infant Death Syndrome (GeSID) Study
A case control study conducted in Germany between 1998-2001. 333 SIDS cases and 998 controls.
- Usual sociocultural and biological factors associated with increased risk (lower maternal age, single marital ‘family status’, smoking in pregnancy, greater n previous live births and lower socio-economic status, lower gestational age and birthweight).
- Prone sleep associated with increased risk (mOR 6.08, CI 3.33-11.08) but not side sleep (mOR 0.82, CI 0.52-1.28).
- Cosleeping (bed-sharing) = increased risk (mOR 2.71, CI 1.44-5.10)
- Association of risk of cosleeping with smoking not statistically significant but authors indicate it is suggested by the figures (smoking=no=OR 2.20, CI 0.99-4.91; smoking yes= OR 6.44, CI 2.62-15.81).
- Not breastfeeding for >2wks increased risk (mOR 1.71, CI 1.06-2.77).
- Dummy use (mOR 0.39, CI 0.25-0.59) associated with reduced risk.
Large, single-country study.
Definition of ‘breastfeeding’ = for more than 2 weeks (not necessarily at time of death); No definition of bed-sharing given; No indication as to whether presence of partners in bed, alcohol use etc were considered in bed-sharing risk.