Source: Unknown

Source: Unknown

Recently an article has made the rounds on various news sources stating that co-sleeping has negative effects for mom.  The research is not a peer-reviewed journal, but a conference presentation at the American Public Health Association’s annual meeting, which is legitimate, but means we have nothing to go on except mass media, which can be notoriously wrong.  However, taking what we can, it seems that the researchers – headed up by Clarissa Simon from Northwestern University – studied cortisol levels over one day for mothers six months after giving birth.  They found that women who breastfed, but did not co-sleep showed their peak in the morning with a sharp, steady decline throughout the day with the lowest cortisol levels at night.  Women who either did not breastfeed or who co-slept (in this case it seems to refer to bedsharing) did not show the steepest declines (though by the researcher’s comments, it seems they still showed a steady decline during the day).  Based on this, the researchers concluded that co-sleeping was suboptimal for mother’s health and recommendations in the mass media have proliferated telling mothers that co-sleeping is bad for them, but is it?

What does this ‘steep decline pattern’ really mean?

As stated in the article, this steep decline has previously been associated with optimal health in women.  Researchers have long known that hypercortisolism (a consistently high pattern of cortisol) and hypocortisolism (a consistently low pattern of cortisol) are associated with negative mental health and functioning

[1][2].  But there is also evidence that the failure to show a decline from morning to afternoon and evening cortisol levels (i.e., showing a flat diurnal pattern, even if moderate) was associated with an increased risk of mortality among women with breast cancer[3], patients with post-traumatic stress disorder[4], and women with high stress and low support environments[5].  However, these studies only refer to a decline versus a flat pattern, as opposed to the actual degree of decline.  The only study for which the actual steepness of the decline showed a relationship with anything was with respect to marital satisfaction[6].


Given what I can interpret from the interviews, it seems that women who breastfeed, but do no co-sleep have the steepest decline, suggesting other mothers actually show a decline as well, just not as steep.  Previous research would suggest that so long as there is a decline, we are in the realm of optimal health.  It is also important to remember that this typical diurnal pattern is not as universal as many individuals would have us believe.  In fact, depending on the study, it seems that only 50-60% of individuals actually show this pattern, with the remaining 40% showing either a flat or up-and-down pattern, and this is irrespective of any psychological or physical distress[7][8].  In short, this pattern may mean something, but not necessarily – the authors are making some rather strong leaps without having presented evidence that these mothers are actually doing any better than their non-breastfeeding or co-sleeping counterparts (or it is being incorrectly reported, something I must consider).

But why the difference at all?

The fact that there is a difference does suggest that something may be going on, right?  What could that be?  The researchers hypothesize that co-sleeping negatively affects mom’s sleep.  If that is the case, and you co-sleep and get better sleep (as I did and I know many others did), then this argument doesn’t hold water.  And given that sleep ratings were not reported as being obtained, it seems this is mere speculation on behalf of the researchers.  What else do we have?

Well, my own theory stems from research on the synchrony that is typically found in mother-infant dyads.  Not only is there behavioural synchrony in responsive mother-infant relationships, but also hormonal (or physiological) synchrony begins early in the infant’s life[9][10].  Importantly, this synchrony is key to the development of secure attachment[11] and the degree to which a mother is responsive to her child has been found to affect this physiological synchrony [9][10].  In one study, Wendy Middlemiss and colleagues found that extinction methods (crying-it-out) led to asynchrony in mother-infant dyads after just three days of a sleep training program[12].  Importantly, according to this research, after sleep training had been initiated there was still synchrony prior to bedtime after a full day playing and engaging in activities together (though the synchrony was weaker than prior to even beginning sleep training), but that after sleep, that synchrony had disappeared (due to the continued raised levels of cortisol in infants, but the decreased levels in mothers).  This suggests that levels of physiological attunement between mother and infant are highly sensitive to the exact times they are together and responsive to each other.

Co-sleeping is one mechanism by which this synchrony may be influenced and in the case of the current research, it may look like a negative effect though it may simply be attunement to the infant.  How would this happen?  One possibility involves the fact that the development of a cortisol diurnal pattern in infancy is subject to incredible individual differences [13], which means that for many of these six-month old children, their diurnal pattern is not set.  If the mother-infant dyad is synchronous at bedtime (which, let’s face it, is more likely to be stronger if the infant is in bed with mom given the physical proximity), the likelihood that mom has her lowest cortisol at bedtime is less if her infant does not show a lowered cortisol level at the same time.  A second possibility is to do with the cortisol awakening response (CAR).  This is the peak of our cortisol that, as adults, appears right after waking.  The pattern the researchers referred to in their study involving the steep decline depends upon this peak being high upon waking.  However, very recent research from Melissa Bright and colleagues has shown that the CAR does not exist in infancy, with children up to 17 months being included in the analysis[14].  Although Ms. Bright and her colleagues also found synchrony between mother-infant dyads (and presumably not all were co-sleeping), it would not be unreasonable to assume greater synchrony upon wakening between co-sleeping dyads than non-co-sleeping dyads.  Given the lack of CAR in infancy, a blunted cortisol response upon awakening in mothers – simply due to being in tune with her infant – would not be far-fetched.  Finally, research on diurnal patterns in children who are still napping (including young infants) shows that this “ideal” diurnal pattern is not normal for young children and in fact isn’t seen in many children even at 3 years of age[15].  In fact, they show the pattern that is described for moms in this study.  Hmmm… need I connect the dots for you?

Take-home message?

Contrary to what the popular media would have us believe, there is really no evidence that has been presented that suggests co-sleeping is bad for mom.  First, the researchers did not address social or emotional functioning – the primary correlates of the “optimal” cortisol diurnal pattern – and thus we have no idea if these women even present with any mental health (or even physical health) problems.  Second, previous research has only found consistent differences with respect to a decline pattern versus a flat pattern, yet the researchers suggest that co-sleeping is only associated with a less steep decline in cortisol throughout the day.  Based on previous research, we have no reason to assume that these women are categorically different than their non-co-sleeping counterparts.  And finally, even if there is a different pattern for co-sleeping moms, this may be explained through maternal-infant synchrony.  As infants don’t have either the same diurnal pattern as most adults (though not all) or the circadian awakening response, mothers who are in synch with their infants AND co-sleeping should show a different rhythm given that they are more likely to be influenced by their infants’ physiology during the nighttime hours (including right before bed and upon wakening).  So if the recent coverage of the cortisol and co-sleeping research had you questioning whether you were doing the right thing for yourself and your baby, please relax.  Even if it turns out that co-sleeping means a ‘less than optimal’ cortisol pattern for a period of time, it if works best for you and baby, that’s all that should matter.

Recommended Reading after the References

[1] Fries E, Hesse J, Hellhammer J, Hellhammer DH.  A new view on hypocortisolism.  Psychoneuroendocrinology 2005; 30: 1010-1016.

[2] Burke HM, Davis MC, Otte C, Mohr DC.  Depression and cortisol responses to psychological stress: a meta-analysis.  Psychoneuroendocrinology 2005; 30: 846-856.

[3] Sephton S, Sapolsky R, Kraemer H, Spiegel D.  Diurnal cortisol rhythm as a predictor of breast cancer survival.  Journal of National Cancer Institute 2000; 92: 994-1000.

[4] Lauc G, Zvonar K, Vuksic-Mihaljevic Z, Flogel M.  Post-awakening changes in salivary cortisol in veterans with and without PTSD.  Stress and Health 2004; 20: 99-102.

[5] Ambercrombie HC, Grese-Davis J, Sephton S, Epel S, Turner-Cobb JM, Speigel D.  Flattened cortisol rhythms in metastatic breast cancer patients.  Psychoneuroendocrinology 2004; 29: 1082-1092.

[6] Vedhara K, Tunistra JT, Miles JN, Sanderman R, Ranchor AV.  Psychosocial factors associated with indices of cortisol production in women with breast cancer and controls.  Psychoneuroendocrinology 2006; 31: 299-311.

[7] Smyth JM, Ockenfels MC, Gorin AA, Catley D, Porter LS, Kirschbaum C, et al. Individual differences in the diurnal cycle of cortisol.  Psychoneuroendocrinology 1997; 22: 89-105.

[8] Ice GH, Katz-Stein A, Himes J, Kane RL.  Diurnal cycles of salivary cortisol in older adults.  Psychoneuroendocrinology 2004; 29: 355-370.

[9] Sethre-Hofstad L, Stansburty K, Rice MA.  Attunement of maternal and child adrenocorticol response to a child challenge.  Psychoneuroendocrinology 2002; 27: 731-747.

[10] Thompson LA, Trevarthan WR.  Cortisol reactivity, maternal sensitivity, and learning in 3-month-old infants.  Infant Behavioral Development 2008; 31: 92-106.

[11] Isabella RA, Belsky J.  Interactional synchrony and the origins of mother-infant attachment: a replication study.  Child Development 1991; 62: 373-384.

[12] Middlemiss W, Granger DA, Goldberg WA, Nathans L.  Asynchrony of mother-infant hypothalamic-pituitary-adrenal axis activity following extinction of infant crying responses induced during the transition to sleep.  Early Human Development 2011; 88: 227-232.

[13] de Weerth C, Zijl RH, Buitelaar JK.  Development of cortisol circadian rhythm in infancy.  Early Human Development 2003; 73: 39-52.

[14] Bright MA, Granger DA, Frick JE.  Do infants show a cortisol awakening response?  Developmental Psychobiology 2012; 54: 736-743.

[15] Watamura SE, Donzella B, Kertes DA, Gunnar MR.  Developmental changes in baseline cortisol activity in early childhood: relations with napping and effortful control.  Developmental Psychobiology 2004; 45: 125-33.