Copyright: Kirill Federspiel

Copyright: Kirill Federspiel

Sleeping with your baby is as old as human history.  It is a biological norm for our babies to want to be close to us (and for most of us, for us to be close to them).  It is a practice that has been condemned in our current society, with fears of suffocation and SIDS leading the way, and most parents have been told flat-out to avoid it.  However, the practice is increasing as more and more parents realize the benefits of sleeping close to baby and even some health authorities are started to provide more nuanced messages about safety and bedsharing

[1].

Great, right?

Yes, for a certain group of parents.  For breastfeeding mothers in particular, the message now being shared is that with certain careful considerations, they can safely bedshare which will help maximize the breastfeeding relationship.  In fact, in a recent talk, Dr. James McKenna of the University of Notre Dame (who has spent his career examining infant-mother sleep) called the practice of bedsharing by breastfeeding mothers “breastsleeping”.  This highlights how the bedsharing debate and discussion has been focused on the link to breastfeeding and not the other benefits, such as closeness, ease of responsiveness, physiological synchrony, and increased touch.

The question we need to be asking is: Is it fair?

For parents who are not breastfeeding for any reason, should we be telling them not to share that precious sleep time as well?  Is it an all or none situation or is bedsharing and the closeness it can provide be something families can enjoy if done safely?  Is feeding method really a risk factor for it?  Herein, I want to look at what we do know in hopes of helping families make their own educated decisions about what will work for them when it comes to infant sleep.

Formula Feeding and SIDS

The first issue that needs to be addressed is the link between formula use and SIDS.  Given the main concern with bedsharing is SIDS (or suffocation, and we’ll get to that in a moment), the known increased risk of SIDS when using formula has to be discussed.  There are numerous studies that have found an increased risk of SIDS when a mother is using formula over breastfeeding (e.g., [2]; for a review, see [3]).  Notably, the risk is dose-dependent, with a greater association between formula feeding and SIDS when it is exclusive and ongoing for the peak periods.  For example, [2] found that the risk of SIDS was halved when breastfeeding was exclusive at 1 month, but the risk was not statistically different from formula use when infants were only partially breastfed at 1 month.  This suggests the mechanism for the increased risk for formula use is greater when formula use is exclusive.

 

Similarly, in a meta-analysis examining 18 case-control studies (which means the individuals are matched in key demographic information pertaining to the infant but often not other factors), any breastfeeding for any duration reduced the risk of SIDS by 64% in a univariate analysis and by 45% when other variables were considered (this drops to 51% and 32%, respectively, when studies that were excluded in a sensitivity analysis – i.e., didn’t meet 1 or more quality criteria – were included; this suggests that some articles that do not find an increase are of lower quality).  Any breastfeeding at age 2 months was associated with a reduction in risk of 62% (no multivariate analysis possible), and exclusive breastfeeding of any duration was associated with a reduction in risk of 73% (no multivariate analysis possible).  Again, one can see a greater risk of formula use on SIDS when it is exclusive and for a longer duration, but also that it has been lower-quality studies that seem to show a lower risk of formula.  (Note the research speaks of reduction of risk for breastfeeding, but as that is the biological norm, the risk is in formula use, hence my terminology.)

At this stage, I believe it is fair to state a greater risk of SIDS with formula use (even though the absolute risk of SIDS is still low, with 2000-2500 deaths per year in the United States).  The question, however, is about the interaction with bedsharing.  Should parents who cannot or choose not to breastfeed be told they should not engage in what many parents find to be a wonderful time with their babies and children?

Formula Feeding, Bedsharing, and SIDS Risk

What’s first notable here is that there is a lot of discussion about the issue of formula feeding and bedsharing, not much of it positive, which would naturally give many women pause when considering bedsharing.  The reason, though, is that the bedsharing and breastfeeding relationship is the biological norm of our species (and many others) and thus the change from the biological norm in one area naturally raises questions about the safety or validity of the remaining part of the equation[4][5][6].  It’s a logical conclusion, but oddly one for which we don’t have much research, except to know that the two are highly intertwined and seem to facilitate each other (for a review, see [7]).  Earlier studies looking at the risk of bedsharing often omitted any inclusion of feeding status in their analyses (e.g., [8]) or simply included it as a control factor to examine bedsharing without looking at the actual interaction with bedsharing (e.g., [9]).

Recently, a meta-analysis which is highly flawed in many areas did try to examine the effect of feeding method on bedsharing and SIDS[10].  Before I provide more information, I have to be clear that this analysis was flawed because the original articles were flawed.  The assessment of breastfeeding was never complete and did include “ever initiated” as a variable, making assessment problematic.  One study did not even report breastfeeding data in their original analysis.  This particular meta-analysis, however, did not find an interaction of feeding and bedsharing, though (i.e., they found an increased risk for bedsharing, which I disagree with given their data and other studies, but that’s a different issue, but the increased risk for bedsharing was of the same magnitude for both breastfeeding and formula feeding dyads and given the increased risk of SIDS in formula feeding infants, the absolute risk was higher in formula feeding dyads, but only relative to feeding method, not sleeping method).  In a later section in which they predict risk based on very flawed analyses, they do find that bottle-feeding would interact with smoking and alcohol use to astronomically increase the risk of SIDS, but again, even these data are questionable given what was done herein.

So what can we say?  The studies that have included feeding status in their analyses have found that the risk of bedsharing does decrease once that factor is controlled for (e.g., [2][9]).  The problem is that that is never the only factor controlled for; often there are many factors that are included as controls, including birthweight, SES, maternal age, birth problems, etc.  This means we cannot determine the exact influence of formula feeding on SIDS rates or how formula and bedsharing interact.  A further problem is that many of the studies include “breastfeeding initiation” as their feeding variable (e.g., [11]) instead of more nuanced definitions of breastfeeding, including duration and exclusivity, which we know to be related to SIDS rates (see above).  The meta-analysis is so flawed it’s difficult to take home much of any message, especially as it suffers the same definitional issues, but adds to it problems with statistics, missing data, and interpretation.  This means even the one piece of evidence we have that suggests no interaction between feeding method and bedsharing is also something we can’t really take a firm message from.

Possible Mechanisms and the Research Behind Them

Source: Unknown

Source: Unknown

If we don’t have much in the way of actual numbers, what do we have?  Well, we can look at the hypothesized explanations for concern and see what research there is to support them.  This is possibly the best recourse at the moment, for if people know the mechanisms, they may be able to avoid problems altogether.

The main issues concerning formula feeding and bedsharing tend to fall into one of two main concerns: Positioning of the baby and arousability (of either parent or baby).

Positioning of the baby

One concern with formula feeding and bedsharing is how the baby is positioned in bed.  With breastfeeding dyads, it is noted that baby often spends the vast majority of the night at breast level[12][13][14], likely to facilitate breastfeeding.  One known study has also looked at positioning in formula feeding mothers and found that these mothers were more likely to position their babies closer to the pillow, as if they were sleeping next to another adult[12].  The problem with this is that it can lead to greater airway coverings from pillows, blankets, and human hair, raising the risk of suffocation.  Indeed, one of the findings from breastfeeding-bedsharing dyads is that the mother kept blankets at waist level whereas fathers or non-breastfeeding-bedsharing dyads kept the blankets at the more typical chin level.

 

This one study hasn’t been replicated and there are case studies in which breastfeeding mothers show sleep arrangements more like that of formula feeding moms in [12] and where bottle-feeding mothers are sleeping more like breastfeeding mothers[15].  These case studies suggest that if positioning is an issue, it is one that may be modifiable by presenting parents with the information needed to make sure they position their infant at breast level, regardless of feeding type.  I say “may” because the idea that we can teach people to bedshare like a breastfeeding dyad is an assumption that has not been tested and it’s one we also need to be aware may not work out.  We just simply don’t know.

It is also worth mentioning that airway coverings do happen in breastfeeding and bedsharing infants as well.  In one study, there were 110 episodes of head covering from 22 infants examined for one night’s sleep and they were all breastfed infants[14].  Most were inadvertent and most were cleared promptly by either the infant or the parent and resulted in no distressing situations for the infant.  In another study that examined oxygen saturation and heart rate during airway coverings in infants while bedsharing, it was found that despite at times prolonged bouts of airway covering, there was no consistent effect on either oxygen saturation or heart rate[16].  This means that perhaps the fear of face covering that is often brought up in the formula-bedsharing discussion isn’t limited to this particular dyadic situation, but rather reflects something that is common in breastfeeding-bedsharing dyate risk of SIDS is still low, with 2000-2500 deaths per year in the United States).  The question, however, is about the interaction with bedsharing.  Should parents who cannot or choose not to breastfeed be told they should not engage in what many parents find to be a wonderful time with their babies and children?

Formula Feeding, Bedsharing, and SIDS Risk

What’s first notable here is that there is a lot of discussion about the issue of formula feeding and bedsharing, not much of it positive, which would naturally give many women pause when considering bedsharing.  The reason, though, is that the bedsharing and breastfeeding relationship is the biological norm of our species (and many others) and thus the change from the biological norm in one area naturally raises questions about the safety or validity of the remaining part of the equation[4][5][6].  It’s a logical conclusion, but oddly one for which we don’t have much research, except to know that the two are highly intertwined and seem to facilitate each other (for a review, see [7]).  Earlier studies looking at the risk of bedsharing often omitted any inclusion of feeding status in their analyses (e.g., [8]) or simply included it as a control factor to examine bedsharing without looking at the actual interaction with bedsharing (e.g., [9]).

Recently, a meta-analysis which is highly flawed in many areas did try to examine the effect of feeding method on bedsharing and SIDS[10].  Before I provide more information, I have to be clear that this analysis was flawed because the original articles were flawed.  The assessment of breastfeeding was never complete and did include “ever initiated” as a variable, making assessment problematic.  One study did not even report breastfeeding data in their original analysis.  This particular meta-analysis, however, did not find an interaction of feeding and bedsharing, though (i.e., they found an increased risk for bedsharing, which I disagree with given their data and other studies, but that’s a different issue, but the increased risk for bedsharing was of the same magnitude for both breastfeeding and formula feeding dyads and given the increased risk of SIDS in formula feeding infants, the absolute risk was higher in formula feeding dyads, but only relative to feeding method, not sleeping method).  In a later section in which they predict risk based on very flawed analyses, they do find that bottle-feeding would interact with smoking and alcohol use to astronomically increase the risk of SIDS, but again, even these data are questionable given what was done herein.

So what can we say?  The studies that have included feeding status in their analyses have found that the risk of bedsharing does decrease once that factor is controlled for (e.g., [2][9]).  The problem is that that is never the only factor controlled for; often there are many factors that are included as controls, including birthweight, SES, maternal age, birth problems, etc.  This means we cannot determine the exact influence of formula feeding on SIDS rates or how formula and bedsharing interact.  A further problem is that many of the studies include “breastfeeding initiation” as their feeding variable (e.g., [11]) instead of more nuanced definitions of breastfeeding, including duration and exclusivity, which we know to be related to SIDS rates (see above).  The meta-analysis is so flawed it’s difficult to take home much of any message, especially as it suffers the same definitional issues, but adds to it problems with statistics, missing data, and interpretation.  This means even the one piece of evidence we have that suggests no interaction between feeding method and bedsharing is also something we can’t really take a firm message from.

Possible Mechanisms and the Research Behind Them

Source: Unknown

Source: Unknown

If we don’t have much in the way of actual numbers, what do we have?  Well, we can look at the hypothesized explanations for concern and see what research there is to support them.  This is possibly the best recourse at the moment, for if people know the mechanisms, they may be able to avoid problems altogether.

The main issues concerning formula feeding and bedsharing tend to fall into one of two main concerns: Positioning of the baby and arousability (of either parent or baby).

Positioning of the baby

One concern with formula feeding and bedsharing is how the baby is positioned in bed.  With breastfeeding dyads, it is noted that baby often spends the vast majority of the night at breast level[12][13][14], likely to facilitate breastfeeding.  One known study has also looked at positioning in formula feeding mothers and found that these mothers were more likely to position their babies closer to the pillow, as if they were sleeping next to another adult[12].  The problem with this is that it can lead to greater airway coverings from pillows, blankets, and human hair, raising the risk of suffocation.  Indeed, one of the findings from breastfeeding-bedsharing dyads is that the mother kept blankets at waist level whereas fathers or non-breastfeeding-bedsharing dyads kept the blankets at the more typical chin level.

This one study hasn’t been replicated and there are case studies in which breastfeeding mothers show sleep arrangements more like that of formula feeding moms in [12] and where bottle-feeding mothers are sleeping more like breastfeeding mothers[15].  These case studies suggest that if positioning is an issue, it is one that may be modifiable by presenting parents with the information needed to make sure they position their infant at breast level, regardless of feeding type.  I say “may” because the idea that we can teach people to bedshare like a breastfeeding dyad is an assumption that has not been tested and it’s one we also need to be aware may not work out.  We just simply don’t know.

It is also worth mentioning that airway coverings do happen in breastfeeding and bedsharing infants as well.  In one study, there were 110 episodes of head covering from 22 infants examined for one night’s sleep and they were all breastfed infants[14].  Most were inadvertent and most were cleared promptly by either the infant or the parent and resulted in no distressing situations for the infant.  In another study that examined oxygen saturation and heart rate during airway coverings in infants while bedsharing, it was found that despite at times prolonged bouts of airway covering, there was no consistent effect on either oxygen saturation or heart rate[16].  This means that perhaps the fear of face covering that is often brought up in the formula-bedsharing discussion isn’t limited to this particular dyadic situation, but rather reflects something that is common in breastfeeding-bedsharing dyads as well.

Arousability

The second concern regarding formula feeding and bedsharing is arousability – specifically both that of mom and baby.  Breastfeeding mothers are known to regularly wake briefly, check on the baby, and return to sleep; indeed, one study found the average “checking time” during the night was 10.7 minutes[14], which is actually quite a lot given how quickly a mother can check.  Dr. James McKenna has long hypothesized that for certain subgroups of infants, this arousability that is linked to breastfeeding and bedsharing is protective of SIDS, though so far no research has supported the hypothesis (nor has any debunked it given the focus is on those infants at-risk for SIDS and that is difficult to ascertain ahead of time) (e.g., [5]).  Problems with arousability is one hypothesis regarding SIDS, and has been given some support by research into brain stem abnormalities in infants who have died of SIDS[17].  How does it related to formula feeding and/or bedsharing?

Research on infant arousability by feeding status has found that infants who are formula fed are less arousable in active sleep, though there are no differences in quiet sleep[18].  Most notable, this difference was only noticeable at 2-3 months, the peak age for SIDS (the other time periods tested were 2-4 weeks and 5-6 months).  It is worth noting is that the arousability was assessed as evoked arousability, which is to say that it was not natural arousals but stimulated ones (in this case, using a jet of air to the nostrils).  In one sense, the artificial nature of the experiment may make the results less valid, but in another sense, they may be even more damning as one would expect natural arousability to be less than evoked arousability.  Thus, if the infants are not responding to external stimuli to the same degree as breastfed infants, then it is likely they are less likely to arouse on their own as well.  However, this study has not been replicated to my knowledge which means, like the study on infant positioning in sleep for formula fed infants, it is unclear if the findings will hold.

In the aforemented study on bedsharing and formula feeding, [12], mother and infant arousals were also examined.  In addition to the different positioning, formula feeding mother-infant dyads showed significantly fewer maternal, infant, and mutual arousals than breastfeeding dyads.  This lends some support to the concern that if mothers are not waking as regularly, there may be greater concern for infant safety with respect to some of the hazards of the modern bed (e.g., airway covering, smothering, etc.).  Again, we are at the whim of research that desperately needs to be replicated.

Conclusions

The issue of formula feeding and bedsharing is a very tricky one.  There is the established link between formula use and SIDS independent of bedsharing, but the question of an interaction remains.  Preliminary evidence could suggest a greater risk based on infant positioning and decreased arousals for both mother and infant when bedsharing and using formula.  However, we must note two very important things: (1) these results have not been replicated, and (2) these results on arousability in bedsharing have not been compared to arousability in solitary sleep.

This second point I mention because the issue at hand isn’t about the risk of SIDS when formula feeding (that link is pretty firmly established), but rather the interaction with bedsharing and whether bedsharing while formula feeding increases the risk of SIDS even more.  The results on arousability simply can’t speak to whether or not this decreased arousability relative to breastfeeding moms is any greater when bedsharing than during solitary sleep.  Indeed, it is possible that arousals are greater during bedsharing and may even help prevent some SIDS cases of formula-fed infants.  There just isn’t the research to say one way or the other.

What are you to do if you want to bedshare and you are using formula?  Well, it’s up to you.  Currently, the recommendations are to not bedshare at all, but as you now know, the data isn’t quite as conclusive.  It’s up to each family to decide the risks they are willing to take and the benefits of bedsharing for some families may be great enough to say the risks are worth taking, especially given the absolute low risk of SIDS.  However, other families will feel differently and opt to avoid bedsharing altogether.  A middle-ground option is to avoid bedsharing for the first 3-4 months, opting for room-sharing instead, and then transition to bedsharing after the peak age for SIDS has passed.

All in all, far more research is needed.  As someone who firmly believes in historical and evolutionary child caregiving practices, it is difficult to see that one change can lead to the unavailability of other practices that promote closeness, synchrony, and touch.  I acknowledge that this may very well be the case and future research may tell us that formula feeding and bedsharing is an interaction that increases risk to an unacceptable level; however we are not there right now.  As such it is up to families to decide what will work best for them with the knowledge of what the research is telling us as well as its limitations.

For more information on bedsharing, please check out