Introduction

Questions regarding infant sleep location continue to vex parents and health professionals alike, and research evidence on the topic can seem contradictory. Some contradictions simply reflect the way studies are reported in the media; controversy sells newspapers and alarmist headlines should be critically examined. Other contradictions relate to the perspectives of researchers, the cultural context in which they are working, the questions they are attempting to address, and the populations they are examining. It is often hard to unpick these issues without accessing the original publication. A flurry of studies over the past three to four years have progressed understanding of parentinfant sleep contact and the relationship of bed -sharing and breastfeeding, but some questions require further work. This review examines the latest UK research and summarises findings from relevant international studies. We also announce a new resource on infant sleep issues for both health professionals and parents to be launched this year.

Safeguarding and wellbeing

Two broad clinical agendas address infant sleep location: prevention of infant death and safety awareness (safeguarding); and promotion of breastfeeding, bonding and infant mental development (wellbeing). Both agendas seek the best outcomes for infants - but the outcomes with which they are concerned clearly differ. Success for the safeguarding agenda is measured in terms of reduced rates of fatalities and injuries; for the wellbeing agenda it is measured in terms of breastfeeding rates, reduced infant morbidity, appropriate growth and development, and secure attachment relationships.

These agendas clearly intersect; however, those at the 'coalface' of health promotion can feel they are in competition as researchers addressing the outcomes for one agenda may ignore their impact on the other. However, after almost two decades we are beginning to see the two agendas are converging in their perspectives, at least in part.

Breastfeeding and bed sharing

The first important point is confirmation of the relationship between bed sharing and breastfeeding. We began reporting on this relationship over a decade ago (Ball et al, 1999; Bau, 2002, 2003; Blair and Ball, 2004) when other research evidence was sparse. However, numerous researchers have now found that women who breastfeed are more likely to bed share; and that bed sharing is associated with a greater duration of breastfeeding (Lahr et al, 2007; Hauck et al, 2008; Santos et al, 2009; Tan et al, 2009; Anuntaseree et al, 2010; Blair et al, 2010; Kendall-Tackett, 2010).

Researchers in Brazil (Santos et al, 2009) investigated breastfeeding outcomes at 12 months by interviewing mothers of 4231 infants at birth, three and 12 months about their breastfeeding and bed-sharing characteristics. Bed sharing was defined as habitual sharing of a bed between mother and child for all or part of the night. Breastfeeding prevalence at 1 2 months was 59% for those who bed-shared at three months and 44% for those who did not. Among infants exclusively breastfed at three months, 75% of bed-sharers were still breastfed at 12 months, versus 52% of non-sharers. The authors accepted these results as evidence of bed sharing being protective against early weaning; of course they are not - the existence of an association cannot determine the direction of causality. As was the case with our previous work (Ball, 2003), evidence that bed-sharers breastfeed for longer is not evidence bed sharing protects against early breastfeeding termination. The relationship may simply be that mothers who are inclined to breastfeed for longer may also be more inclined to bed-share.

North-East Cot Trial

The NECOT trial was designed to unpick one aspect of the relationship between breastfeeding and sleep-sharing (Ball et al, 2011). We hypothesised that as sleep contact between mother and baby increases breastfeeding frequency (Ball et al, 2006; Gettler and McKenner, 2010), and because frequent breastfeeding promotes effective lactation (De Carvalho et al, 1983), mothers and babies in 24-hour close contact following delivery may breastfeed for longer than mothers who room-in. A total of 1 204 pregnant women with an intention to breastfeed were recruited at the Royal Victoria Infirmary in Newcastleupon-Tyne, and randomised to one of two sleeping arrangements on the postnatal ward. Half the women were allocated to normal rooming-in (stand-alone cot at bedside); the other half were allocated to close-contact (side-car crib clamped to the mother's bed-frame).

Following hospital discharge mothers reported weekly on their breastfeeding status and infant at-home sleep location; 870 mothers provided data for six months. Intention to treat analysis using Cox regression, adjusting for maternal age, education, delivery type and previous breastfeeding, found no significant difference between the groups for duration of any or exclusive breastfeeding. No adverse events were reported with either cot-type. Bed sharing at home occurred for 67% of side-car and 64% of rooming-in infants (no significant difference) (Ball et al, 2011).

Therefore, the trial indicated that postnatal sleep proximity did not influence long-term breastfeeding outcomes in a busy tertiary level hospital where many women were discharged within 24 hours. We did, however, replicate our previous findings and those of Santos et al (2009) in that mothers and infants who bed shared in the first 13 weeks were twice as likely than non-sharers to breastfeed to six months (Ball et al, unpublished data). The short duration of current postnatal hospitalisation and the relationship of bed sharing and breastfeeding at home means the question about directionality of the association now needs examining in the home - however, randomising mothers and infants to different sleep locations (and ensuring compliance) in a domestic setting would be challenging.

Bed-sharers change over time

The longitudinal patterns of bed-sharing and breastfeeding were also examined using data on infant sleep location at five time points between birth and 45 months of age for 14000 families from the ALSPAC (Avon Longitudinal Study of Parents and Children) cohort study (Blair et al, 2010). Sleep location and infant feeding data were collected prospectively by postal questionnaire. Bed sharing was defined as an infant or child spending some of their nocturnal sleep in the same bed as an adult. Breastfeeding duration was categorised at 15 months.

Latent class analysis identified four groups of families that the researchers described as non-sharers (66%), early bed-sharers (only in infancy, 13%), late bed-sharers (commencing after the 1st year, 15%), and constant bed-sharers (throughout the four- year period, 6%). Non-white ethnicity was significantly associated with all groups of bed-sharers, and particularly with those who bed-shared constantly. Early bed sharing was associated with greater maternal education and fewer indicators of deprivation; late bed-sharing was associated with less maternal education and higher deprivation.

The prevalence of breastfeeding in each of the 15 months following birth was significantly higher among the groups that bed shared constantly or early, than among the late or non-sharers. Multivariate logistic regression, using breastfeeding at 12 months as an outcome, found that although 11% of infants overall were stul breastfed at 12 months, this differed according to the latent classes: 9% of non-sharers, 14% of late bed-sharers, 19% of early bed-sharers, and 34% of constant bed sharers were breastfeeding at 12 months. The relationship with breastfeeding remained significant after controlling for confounders. The authors conclude that the characteristics of families that are most likely to bed-share with their infant in the months following birth place them at very low risk of SIDS (Sudden Infant Death Syndrome) - therefore, any benefit from preventing bed sharing in this group will be very small, and by following such advice breastfeeding may suffer. The authors recommend that risk reduction messages to prevent SIDS be targeted specifically to unsafe infant care practices; in this way safeguarding would avoid undermining breastfeeding outcomes for those infants already at low risk of unexpected death.

Alcohol and sofas

Infant sleep safety in the UK has been examined directly in the South West Infant Sleep Study (SWISS) (Blair et al, 2009); a four-year population based case-control study of all unexpected infant deaths (157/184,800 births) in the southwest region. Eighty of 157 infant deaths remained unexplained following a multi-agency investigation and so met the criteria for SIDS. These were compared to two age-matched control groups of 87 randomly selected infants and 82 infants at high risk of SIDS (young, socially deprived, multiparous mothers who smoked). In this study 'cosleeping' was defined to include any aduit and baby sleeping together on a bed or a sofa.

Among the SIDS infants, 54% died while cosleeping compared with 20% who coslept during the reference sleep for both control groups. Excess deaths were explained by a significant interaction between cosleeping and recent parental use of alcohol or drugs (31% v 3% random controls) and the greater proportion of SIDS infants cosleeping on a sofa (17% v 1%).

The authors concluded that many of the SIDS infants had slept with an adult in a hazardous environment. The major influences on risk, regardless of markers for socioeconomic deprivation, were the use of alcohol or drugs before sharing a bed, and sofa sharing. Although data on whether or not mothers 'attempted to breastfeed' were compared for cases and controls, no association was found with SIDS; however, more specific data on infant feeding type at time of death or reference sleep were not reported.

The SWISS study results reflect those of BIabey and Gessner (2009) who examined 13 years of data on Alaskan infant deaths while bed sharing to assess the contributions of known risk factors. In 99% of cases at least one additional risk factor was present including maternal tobacco use (75%) and sleeping with an impaired person (43%). Frequent bed sharing was reported for 38% of Alaskan infants, with most bed-sharers reporting no risk factors.

The authors conclude their data indicate 'infant bed sharing in the absence of other risk factors is not inherently dangerous and thus do not support the American Academy of Paediatrics (AAP) recommendation against all bed sharing'. These studies, therefore, allow us to now begin to identify those situations where bed sharing or other forms of sleep contact is particularly risky, and those where it is not.

Breastfeeding and SIDS

Although neither of the above studies illuminated the relationship between infant feeding type and SIDS, a meta-analysis of the data from 18 case-control studies recently did so (Hauck et al, 2011) producing a multivariable summary odds ratio (SOR) for SIDS of 0.55 for infants receiving any breast milk compared with those receiving none.

The univariable SOR was 0.4 for any breastfeeding; 0,38 for any breastfeeding at two months of age, and 0.27 for any period of exclusive breastfeeding (insufficient data precluded calculation of multivariable SORS for these latter variables). Breastfeeding is, therefore, protective against SIDS and this effect seems stronger when breastfeeding is exclusive. The only recent study to directly examine the combined effect of breastfeeding and bed-sharing on SIDS-rate (Ruys et al, 2007) was excluded from the above meta-analysis as 30% of the infant deaths attributed as SIDS were not autopsied which casts doubt upon the findings.

UK ethnic differences in infant sleep location

We investigated the relationship between bed sharing, breastfeeding and ethnicity in the recently completed Bradford Infant Care Study (BradICS) (Ball et al, 2011). This telephone survey of 2560 families enrolled in the Born in Bradford cohort study compared the night-time infant care practices of 968 white British and 1212 Pakistani families.

We aimed to explore ethnic differences in infant care that might explain the observed discrepancy in SIDS rates between infants of South Asian and white British origin (the latter having a SIDS-rate four times greater than the former).

In this study, bed sharing and sofa sharing were defined as an infant sleeping with a parent while the parent was asleep in the specified location. Both groups had a very low and similar incidence of prone infant sleep; however, compared to white British infants, Pakistani infants were significantly more likely to sleep in an adult bed, ever bed share and regularly bed share, but were significantly less likely to ever sofa share. Pakistani infants were also more likely to have ever breastfed, and to have breastfed for more than eight weeks. White British infants were more likely to sleep in a room alone, sofa share, be exposed to maternal smoking and have parents who consume alcohol. South Asian infant care practices were more likely to protect infants from the most important SIDS -risks such as maternal smoking, alcohol consumption, sofa sharing and solitary sleep. In both groups breastfed infants were more likely to bed share and bed sharing was particularly prevalent among white British mothers who breastfed for eight or more weeks.

Conclusion

The majority of infant sleep safety research has been conducted in the USA, capturing dangerous sleep practices among minority populations, and leading to stringent anti-bed sharing recommendations (AAP, 2005). In the past few years, UK research has unpicked the characteristics of safe and unsafe bed sharing, and identified the characteristics of bed sharers at high and low SIDS-risk. Multiple international studies have confirmed the relationship between bed sharing and breastfeeding: UK 'Reduce the Risk' advice for SIDS now emphasises the importance ofbreastfeeding, and specifies those circumstances under which bed-sharing should be avoided (DH, 2010). Meanwhile, we have been funded by a UK research council to produce an online Infant Sleep Information Source (ISIS) for parents and health professionals which aims to provide updates on research findings, interpretations, and implications in this complex area. Working with a range of UK parenting support organisations we hope to have a beta-site available at the beginning of 2012 at www.isisonline.org.uk and look forward to hearing your feedback.

References

American Academy of Pediatrics. (AAP) (2005) The Changing Concept of Sudden Infant Death Syndrome: diagnostic coding shifts, controversies regarding the sleeping environment, and new variables to consider in reducing risk. Pediatrics 116: 1245-55.

Animtaseree W, Mo-suwan L, Vasiknanonte P et al. (2008) Night waking in Thai infants at 3 months of age: Association between parental practices and infant sleep. Sleep Mea 9(5): 564-71.

Ball HL, Hooker E, Kelly PJ. (1999) Where will the baby sleep? Attitudes and practices of new and experienced parents regarding cosleeping with, their new-born infants. Am Anthrop 101(1): 143-51.

Ball HL, Moya E, Fairley L et al. (2011) Infant care practices related to Sudden Infant Death (SIDS) in South Asian and White British families in the UK: the Bradford Infant Care Study (BradICS). Paethatr Pertnat Epidemial 26:3-12.

Ball HL, Moya E, Fairley L et al. (in press) Bed and sofa-sharing practices in a UK bi-ethnic population. Pediatrics.

BaU HL, Ward-Platt MP, Heslop E et al. (2006) Randomised trial of mother-infant sleep proximity on the post-natal ward: implications for breastfeeding initiation and infant safety. Arch Dis Child 91(12): 1005-10.

Ball HL, Ward-Platt MP, Howel D et al. (201 1) Randomised trial of sidecar crib use on breastfeeding duration (NECOT). Arch Dis Child 96(7): 630-1.

Ball HL. (2002) Reasons to bed-share: why parents sleep with their infants. J Repro Inf Psych 20(4): 207-22.

Ball HL. (2003) Breastfeeding, bed-sharing and infant sleep. Birth 30(3): 181-8.

Blabey MH, Gessner BD. (2009) Infant bed-sharing practices and associated risk factors among births and infant deaths in Alaska. Pub Health Reports 124: 527-34.

Blair PS, Ball HL. (2004) The prevalence and characteristics associated with parent-infant bed-sharing in England. Arch Dis Child 89: 1106-10.

Blair PS, Heron J, Fleming Pf. (2010) Relationship between bed sharing and breastfeeding: longitudinal, population-based analysis. Pediatrics 126(5): 1119-126.

Blair PS, Sidebotham P, Evason-Coombe C et al. (2009) Hazardous cosleeping environments and risk factors amenable to change: casecontrol study of SIDS in south west England BMJ 339:doi:10.1136/bmj. b3666

Department of Health (DH). (2009) Reduce the Risk of Cot Death, www.dh.gov.uk/en/ Publicationsandstaîstics/Publîcations/ PublicationsPolicyAndGuidance/DH_4123625

Gettler LT, McKenna II. Evolutionary perspectives on mother-infant sleep proximity and breastfeeding in a laboratory setting. Am J Phys Anthropoi 144(3 ):454-62.

Hauck FR, Signore C, Fein SB et al. (2008) Infant Sleeping Arrangements and Practices During the First Year of Life. Pediatrics 122: sl!3-20.

Hauck FR, Thompson IM, Tanabe KO et al. (2011) Breastfeeding and Reduced Risk of Sudden Infant Death Syndrome: A Meta-analysis. Pediatrics 128(1): 103-10.

Kendall-Tackett KA, Cong Z, Hale TW. (2010) Mother-infant sleep locations and nighttime feeding behaviors: US data from the Survey of Mothers' Sleep and Fatigue. Clitt Laci I: 27-36.

Lahr M, Rosenberg K, Lapidus JA. (2007) Maternal-Infant Bedsharing: Risk factors for bedsharing in a population-based survey of new mothers and implications for SIDS risk reduction. Matern Child Health J 1 1(3): 277-86.

McCoy RC, Hunt CE, Lesko SM. (2004) Frequency of bed sharing and its relationship to breastfeeding. J Dev and Behav Peds 25(3): 141.

Ruys JH, de Jonge GA, Brand R, Engelberts AC, Semmekrot BA. (2007) Bed-sharing in the first four months of life: a risk factor for sudden infant death. Acta Paediatr 96(10): 1399-403.

Santos IS, Mota DM, Matijasevich A et al. (2009). Bed-sharing at 3 months and breastfeeding at 1 year in Southern Brazil. J Pediatrics 155: 505-09.

Tan KL, Ghani SN, Moy FM. (2009) The prevalence and characteristics associated with mother-infant bed-sharing in Klang District, Malaysia. Med J Malaysia 64(4): 311-15.

Author affiliation:

Helen Ball PhD, MA, BScfHons)

Professor of Anthropology

Director, Parent-Infant Sleep Lab

Department of Anthropology

Durham University

Correspondence: h.l.ball@durham.ac.uk