We commend Carpenter et al (2013) for examining risks associated with incidence of SIDS but question their conclusions as unsubstantiated because of faulty and missing data, as well as confounding of criteria used to define bedsharing and risks—a challenge in any meta-analysis.
Their study examines some of the most salient risk factors for SIDS events—infant sleep position, parental use of cigarette smoking, infant birth weight and age. These risks have been well-documented as increasing risk of SIDS events. Thus, it is not surprising or informative to note that these factors remain risks in a re-evaluation of these findings.
While the risks examined do contribute significantly to increasing possibility of SIDS (see Chart 1 below), so do other factors such as bedding and temperature (see Box below for lists of risks not considered). Without consideration of these risks, it is not possible to determine that one variable, such as bedsharing itself is inherently responsible for risk remaining in this study. Nor is it possible to say that one of the variables within the nighttime care routine, such as breastfeeding, is not protective.
Chart 1. Adjusted Odds Ratios from Carpenter et al. (2013)
Verbal description of chart 1:
Side position: 1.4% risk; Prone position: 10.5% risk; Mother smoke: 1.5% risk; Partner smoke: 1% risk; Both smoke: 7.3% risk; drug use: 11.5; alcohol use: 4.8
In addition to these major limitations in making broad, sweeping statements about risk based on this meta-analysis, there are two additional issues that are of significant concern in the paper as a whole. We address these herein.
The first is the treatment of breastfeeding. Buried deep in the last section of the paper is the recommendation that breastfeeding be supported as a mechanism for protecting infant health, the construction of the hypotheses explored here lead to a very different framework. In attempting to examine whether breastfeeding is protective against risk of SIDS when parents bedshare seems to jumble the role of breastfeeding in a manner that undermines one of the stated objectives of the authors… to address health costs associated with early infant care by reducing SIDS events. Further the authors seem to overlook the AOR for bottlefeeding and SIDS risk (see Chart 2).
Chart 2. Adjusted Odds Ratios (percent risk) including bottlefeeding from Carpenter et al. (2013)
NOTE: BW = Birthweight
Verbal description of Chart 2:
Bottlefeeding: 1.4% risk; Side position: 1.4% risk; Prone position: 10.5% risk; Mother smoke: 1/5% risk; Partner smoke: 1% risk; Both smoke: 7.3% risk; birthweight 2500-3499 grams:1/5% risk; birthweight 2000-2499: 4% risk; birthweight under 2000 grams: 9.4% risk
In examining the role of breastfeeding, the authors seem to overlook one essential aspect of infant development—breastfeeding contributes positively to both immediate and later infant health outcomes, not just a reduction in SIDS, though it serves as a protective factor there as well (Alm et al., 2002; Ford et al., 1993; Horne et al., 2004; McVea et al., 2000; Mitchell et al., 1992; Mosko et al., 1997; Scragg et al., 1993). Thus, important in consideration from any perspective is to encourage mothers’ breastfeeding through the infant’s first year of life. However, the authors seem to couch this protective factor in the arena of risk, thus confusing the message for practitioners and parents.
Instead of looking at how each of the variables in the dataset can contribute to risk of infants’ breathing or compromise arousal—the authors focus on whether the act of breastfeeding protects against all risk of SIDS. Clearly that is a standard that cannot be reached. We can, however, easily answer whether breastfeeding protects against SIDS regardless of parental behavior without the necessity of meta-analyses, the imputing of data from 5 of 12 variables, the compromising operational definitions of nighttime care contexts. The answer is simple, though not informative. Yes, there is still a risk. Why? Because there are multiple risk factors that compromise infants’ capacity to breathe and infants’ ability to arouse. Breastfeeding does not vaccinate against all risks (e.g., pillow in the face).
The authors give lip service to breastfeeding, but suggest that any claim that bedsharing helps breastfeeding is ill-advised. The use of the Netherlands as a key example of lowering bedsharing but increasing breastfeeding rates fails to make their point given the relatively low rates and low increases over the 10 year period discussed (a rise of 7% and 8% of any breastfeeding at 3 and 6 months respectively). It is unclear if the strong anti-bedsharing campaign inhibited greater growth in breastfeeding, something that should be of concern when examining the costs associated with infant health. In the US alone, a cost-analysis found that if we could get 80% mothers to breastfeed exclusively for six months (the WHO recommendations), the US would save $10.5 billion a year in health-related costs (Bartick & Reinhold, 2009).
Furthermore, it is misguided and dangerous to argue if bedsharing were recognized as a means of supporting breastfeeding, then we would see more SIDS events. Even more dangerous is to abandon support of breastfeeding in favour of supporting bottlefeeding if it were to detour bedsharing. Although the AOR in the current meta-analysis suggests that bottle-feeding is a lower risk factor than bed-sharing (the validity of which will be discussed below), it only concerns itself with SIDS events, not the more general protective benefits of breastfeeding on infant health. As previously mentioned, breastfeeding confers many health benefits, both immediate and long-term, to children (Horta et al., 2007; Ip et al., 2007; Martin et al., 2005; Owen et al., 2002) and to only consider SIDS events ignores the effects of lower breastfeeding rates on myriad other diseases.
The second issue pertains to the risk factors included and not included in the analysis. The authors have thankfully confirmed some of the major risk factors associated with SIDS, both independently and when interacting with sleep location, such as sleep position, parent smoking, alcohol use, drug use, birth weight, and infant age. The authors solidified many risks, as they were stated individually in the reports associated with each large data set. With this, researchers, practitioners, and parents now have a clear documentation of these specific risks. Very clearly presented is confirmation of known risks and quantification of those risks. For example, maternal smoking remains to be one of the most salient risks associated with SIDS—with paternal smoking contributing to risk as well. Similarly, infant sleep position (i.e., prone and side sleep), contribute significant risk of SIDS events.
Missing from the analysis are other known risk factors that are recognized as part of the triple risk model [i.e., infant age, infant vulnerability, and exogenous, environmental stressors]. For example, in the analysis there is not consideration of exogenous risks such as bedding or vulnerability risks such as infant prematurity. Additionally, the authors fail to include data sets that do examine these risk factors and that come to very different conclusions about the inherent risk of bed sharing on SIDS events (e.g., Blabey & Gessner, 2009). The authors argue that bed sharing is causally related to SIDS events via theories about infant breathing and arousability. Specifically, the authors state, “The proposition that bed sharing is causally related to SIDS is coherent with theories that respiratory obstruction, re-breathing expired gases, and thermal stress (or overheating),which may also give rise to the release of lethal toxins, are all mechanisms leading to SIDS, in the absence of smoking, alcohol or drugs. Infants placed prone are exposed to similar hazards.”
Is the implication in the press release for this article verifiable? Are breastfed, bedsharing babies at inherent risk of SIDS events? The answer is equally as simple, but much more informative. No.
Again what places an infant at risk of SIDS events is what places infants’ breathing and arousability at risk. The elements of the sleep context that place infants’ breathing and arousability at risk are well defined:
These authors seem to be arguing that parenting behavior that can be associated with risk, even if the source of risk in not the behavior, should be stopped (i.e., bedsharing). This is problematic given that bedsharing is a universal, evolved practice, and is often preferred by parents. In fact, the absence of bedsharing does not eliminate risk of SIDS events. The diminishing of bedsharing however is associated with decreases in other behaviors shown to provide protection against SIDS events, such as breastfeeding.
Certainly, without question, a nighttime care context that includes bedsharing and breastfeeding can include elements that compromise infants’ breathing and ability to arouse. Importantly, we know that breastfeeding not only does not contribute to the risk, but serves to help reduce these risks. See Table 1 whereby bottlefed infants are at a greater risk of a SIDS even regardless of sleep location.
What of bedsharing per se? The authors would have us believe bedsharing per se increases the risk of compromising infant breathing and arousability. However, they fail to acknowledge or discuss the fact that there are other factors that influence breather and arousability such as bedding, temperature, and premature status (which is correlated with birth weight, but carries with it unique risk factors that must be considered). Data from Alaska between 1993 and 2004 examined the same question of bed sharing risk, only they also included other known risk factors such as sleep surface (not just sofa, but the type of bed) and sleeping with a non-caregiver and compared the data not just to controls (Blabey & Gessner, 2009). Additionally, the comparison group was taken from a state-wide monitoring system which does not focus on answering one day of bedsharing habits, but rather asks parents about usual bedsharing habits. As such, they most likely had more accurate information on bedsharing than the studies included in the current review. What was found in Alaska? Of the SIDS events that took place while bedsharing, 99% included at least one risk factor, and thus the authors conclude that “infant bed sharing in the absence of other risk factors is not inherently dangerous”.
So, let’s stop going around in circles talking about secondary issues and focus on discussion on primary issue: decreasing the risk of SIDS events. If we want to decrease risk of SIDS events, then we must assure infants are in the best possible situation to support breathing and arousability.
How to do that?
Address Maternal and Infant Health that reduces risk:
Address Nighttime Care practices to ASSURE Breathing and Arousability
Continued monitoring of sleep space
Despite a long history of efforts to reduce bedsharing, this nighttime care practice remains to be the preferred practice for many, is increasing in some areas, and provides many protective or health-benefiting outcomes for mothers and infants. Assuring infants’ safety at night is compromised when discussions shift from the criteria above to admonitions to sleep separately. A focus on protection, a discussion of what underlies risk will be much more successful in reducing risk of SIDS—as well as improving the health context postnatally.
Ten Important Variables Not Considered in The Carpenter et al. (2013) Paper
1. The researchers importantly did not consider whether the bedsharing was planned. Previous research from Venneman (2009) showed no increased risk in planned bedsharing (versus unplanned). This is an incredibly important omission.
2. The paper did not consider the effects of the mother smoking during pregnancy, only smoking post birth. This is a missing risk factor.
3. Breastfeeding information is too limited to draw conclusions. No difference has been drawn between frequency and percentage of breastfeeds versus formula feeds for those ‘partially feeding’.
4. The paper only considered ‘illegal drug use’. Many postnatal mothers (0-12weeks after the birth) are prescribed analgesic medication for related birth induced injuries including but not limited to Caesarean healing, known to have a sedative effect. This was not considered at all. This is a missing risk factor.
5. Prematurity was not considered at all. This is a missing risk factor.
6. Parental exhaustion was not considered at all. Some experts suggest this is considered to be less than 4-5 hours sleep in the past 24 hour period, other experts advise parents to use their instinct. Parental exhaustion naturally impacts on responsive to infant cues. This is a missing risk factor.
7. The researchers did not examine the effect of maternal (and paternal) obesity. This is a missing risk factor.
8. No differentiation was made between having one or both parents in the bed and importantly the location of the baby. It is advisable that the mother sleeps in between the father and infant. Equally it was not noted if older siblings were also present in the bed. This is a missing risk factor.
9. The researchers did not consider fully the impact of alcohol consumption by the father when bedsharing. This is a missing risk factor.
10. No mention was made of whether parents were aware of the risks of bedsharing and how to minimize these before sharing a bed with their infant.
Posts in Sleep Series:
Also, check out: Dangers of "Crying it Out"
Sarah Ockwell-Smith, BabyCalming.com
Professor Wendy Middlemiss, University of North Texas
Tracy Cassels, University of British Columbia, EvolutionaryParenting.com
Helen Stevens, Safe Sleep Space
Professor Darcia Narvaez, University of Notre Dame
Professor Kathy Kendall-Tackett, Texas Tech and University of New Hampshire.
Blabey, M.H., & Gessner, B.D. (2009). Infant bed-sharing practices and associated risk factors among births and infant deaths in Alaska. Public Health Reports, 124,527-534.
Carpenter, R., McGarvey, C., Mitchell, E.A., Tappin, D.M., Vennemann, M.M., Smuk, M., Carpenter, J.R. (2013). Bedsharing when parents do not smoke: Is there a risk of SIDS? An individual level analysis of five major case-control studies. British Medical Journal Open, BMJ Open 2013;3:e002299. doi:10.1136/bmjopen-2012-002299
Ford RPK, Taylor BJ, Mitchell EA, et al. Breastfeeding and the risk of sudden infant death syndrome. Int J Epidemiol. 1993;22:885- 890
Horne RS, Parslow PM, Ferens D, Watts AM, Adamson TM. Comparison of evoked arousability in breast and formula fed infants. Arch Dis Child. 2004;89(1):22-25
Horta BL, Bahl R, Martinés JC, et al. Evidence onthe long-term effects of breastfeeding: systematicreview and meta-analyses. Geneva: World Health Organization; 2007:1-57.
Ip S, Chung M, Raman G, et al. Breastfeeding and maternal and infant health outcomes in developed countries. Evid Rep Technol Assess (FullRep). 2007;153:1-186.
Martin RM, Gunnell D, Smith GD. Breastfeeding in infancy and blood pressure in later life: systematic review and meta-analysis. Am JEpidemiol. 2005;161:15-26.
McVea KL, Turner PD, Peppler DK. The role of breastfeeding in sudden infant death syndrome. J Hum Lact. 2000;16:13-20
Mitchell EA, Taylor BJ, Ford RPK, et al. Four modifiable and other major risk factors for cot death: the New Zealand study. J Paediatr Child Health. 1992;28(suppl 1):S3-S8
Mosko S, Richard C, McKenna J. Infant arousals during mother-infant bed sharing: implications for infant sleep and sudden infant death syndrome research. Pediatrics. 1997;100:841- 849
Owen CG, Whincup PH, Gilg JA, et al. Effect of breast feeding in infancy on blood pressure in later life: systematic review and meta-analysis.BMJ. 2003;327:1189-1195.
Owen CG, Whincup PH, Odoki K, Gilg JA, Cook DG. Infant feeding and blood cholesterol: a study in adolescents and a systematic review. Pediatrics. 2002;110:597- 608
Scragg LK, Mitchell EA, Tonkin SL, Hassall IB. Evaluation of the cot death prevention programme in South Auckland. N Z Med J. 1993;106: 8 -10