Babies sleeping with parents: case-control study of factors influencing the risk of the sudden infant death syndromeCommentary: Cot death—the story so far
BMJ 1999; 319 doi: https://doi.org/10.1136/bmj.319.7223.1457 (Published 04 December 1999) Cite this as: BMJ 1999;319:1457
All rapid responses
This excellent article should prove helpful to those of us struggling
to offer the best advice to parents. However, I am concerned that the
deaths involving "sofa sleeping" are incorrectly ascertained as SIDS
cases.
In residency, I had the unfortunate luck to have two infant deaths on
Emergency Room call, both occurring while the infant slept on the parent's
chest while on a sofa. The children apparently slipped betweent the back
of the sofa, and the parent. The medical examiner (Dr. Pat Lantz, noted
expert on child accidental death and abuse) ruled the cause as
"restrictional" or "pressure" asphyxia. Can similar cases in the study
truly be called "SIDS?"
In addition, I have always abided by the definition of SIDS that
requires a scene investigation by a medical examiner. This does not seem
to have been performed in each case, thus I am left wondering if the case
numbers are reliable in the study. In the UK, does the assignment of SIDS
as a cause of death require a similar scene investigation?
Thank you,
Dudley Bell, M.D., F.A.A.P.
Competing interests: No competing interests
Dear Editor,
Blair et al. have provided some valuable evidence in their case
control study concerning the factors influencing the risk of sudden infant
death syndrome. In particular this study highlights the dangers of infant
bed-sharing where parents are smokers.
Although the authors have adjusted for parental smoking in their
analysis, we have concerns that it may still act as a residual confounder
and hence be an alternative explanation for the apparent increased risk of
SIDS where infants share their parents bed.
In the full multivariate analysis (table 3) maternal smoking during
pregnancy and parental smoking are simply categorized as ‘yes’ or ‘no'
variables. Usual daily postnatal exposure to smoking is crudely estimated
as either ‘0’ or ‘1 or more’ hours per day. Clearly, levels of exposure
can vary hugely in these broad categories. Smoking could still be acting
as a confounding variable if parents of bed sharers were more likely to
smoke heavily, or for longer periods in the home. Indeed the authors note
that:
"Among index mothers who smoked, more of those whose infants shared
the bed smoked more than 20 cigarettes per day (23.2% vs. 1.5% control)
compared with those who did not bed share (16.6% vs. 5.9%)."
Similarly in the analysis of risk of SIDS by parental smoking and bed
sharing status (table 4) the infant’s level of tobacco exposure is
inadequately described.
Adjustment is made for parental smoking but it is neither specified how,
nor whether, infant exposure to smoke is included in the analysis.
We are also disturbed by the apparent lack of adjustment for smoking
as a confounder in the simplified model (table 5). In particular this
table details high alcohol consumption as a variable. Since high maternal
alcohol consumption is often accompanied by a similarly high prevalence of
smoking, it is possible that it is smoking which may be responsible for
some or all of the observed association.
We suggest that this paper may have been improved by using a greater
number of categories to classify the extent of parental smoking during
both the pre and postnatal periods. In addition, an assessment of exposure
to tobacco smoke in the period immediately before the infant’s last or
reference sleep may have been beneficial, rather than relying on the usual
daily exposure.
Given these concerns we believe that the increase risk of SIDS in bed
sharing infants aged <_14 weeks="weeks" where="where" parents="parents" smoke="smoke" is="is" as="as" yet="yet" unproven.="unproven." p="p"/> Yours sincerely,
Amara Ezeonyeji, Steph Jewitt, Leigh Poyser, Tom Stadward
Stage 3 medical students.
Department of Epidemiology and Public Health,
University of Newcastle upon Tyne
Competing interests: No competing interests
The study by Blair, et al, brings to light many important facts
concerning bed-sharing (co-sleeping) and Cot Death (SIDS) (1). Their
finding concerning the dangers of sofa-sharing is particularly important
and could potentially save many lives. It is also just as important that
this study may finally lay to rest the myth that co-sleeping is, in and of
itself, a dangerous practice.
However, a critical finding in the study, although mentioned in the
text, is not given it’s proper prominence in either the ‘Conclusions’ or
‘Key Messages’ sections or in the accompanying Commentary by Mitchell, and
therefore was most likely missed in the media reports. It was also not
mentioned in the BMJ press release or ‘Editor’s Choice’ section.
I am referring to the fact that infants, who were put to sleep in a
cot (crib) in a separate room, were found to be at a significantly
increased risk of Cot Death (SIDS). This risk was about the same as that
for those sharing a bed (co-sleeping), even when including infants whose
parents smoked, as well as those with other increased risk factors.
This, coupled with the fact that infants sleeping in a separate room
is a much more common situation than is co-sleeping, means that many more
lives could be saved by discouraging this separation, than could be by
discouraging smoking parents from co-sleeping.
What are some of the possible mechanisms of the association between
separate sleeping and Cot Death (SIDS)? I would like to propose several
possibilities.
It is common practice for parents to let their babies “cry it out” in
their cribs. Eventually, the infants are “trained” or conditioned not to
cry at all, which pleases most parents who are exhausted and are
understandably grateful to be able to enjoy some peaceful and restful
sleep.
Unfortunately, this unnatural suppression of both the babies’ inborn
instinct to cry and the mothers’ natural instinct to respond by providing
comfort is promoted by many Pediatricians, and therefore parents feel it
must be the right thing to do. However, there is absolutely no scientific
evidence to show that this practice is safe or without long-term effects,
either physical or psychological.
The long bouts of unconsoled crying evoke many physiological
responses from the baby. For example, heart rate, body temperature, blood
pressure, and respiratory rate all rise. The rise in temperature is
especially troubling in regards to cot death, in light of the fact that
overheating may be a contributing factor (2). The body is also forced to
produce increased levels of stress hormones such Adrenaline and Cortisol,
which can exert other effects on the baby’s immature system.
Psychologically these infants, although their brains are not fully
developed, may still feel a sense of “abandonment” or simply undue “fear”
by being left alone for extended periods, a practice not shared by any
other mammal on earth or even most human cultures.
Nor is it a practice that was likely shared by our ancestors, prior
to quite recently. As a matter of fact, human beings have been practicing
co-sleeping behavior throughout their evolutionary development. Is it
really fair then, to ask our babies to simply shed their genetically
implanted instincts as if they were just a bad habit?
Luckily, not all doctors agree with this practice, as evidenced in a
recently published letter in which it was mentioned that newborn infants
have spent the last 9 months in as close contact as you can get with their
mothers, only to be then placed all alone without all of the sensory
stimulation they were used to (3). Additionally, studies have shown that
co-sleeping infants and their mothers develop a physical synchronicity (4)
and that infants experience significantly more arousals (5) during co-
sleeping.
Another possible reason for the increased risk with separate-room
sleeping arrangements, is the fact that signs of distress, exhibited by
the infant, may be missed by parents, who would be able to respond to such
signs when room-sharing.
An even greater opportunity to save lives comes from the easily
modifiable risk factor of smoking. As a matter of fact, considering all
the well-known ill-health effects of smoking I would even go further than
the Commentary by Mitchell, who states “It is time to recommend that
mothers who smoke should not share a bed with their babies.” (1). A more
obvious and ethical recommendation would be to say that “It is time to
recommend that women who smoke should not become pregnant, and if they
unexpectedly do, it is imperative that they stop immediately.” Anything
less than this is unethical and borders on criminal, since we are talking
about saving babies lives and regardless of the sleeping arrangements
smoking has been shown to cause large increases in the risk of Cot Death.
It is time for the medical community to accept a good portion of the
blame for the large number of mothers who continue this harmful practice,
and to take the lead in curbing it. How many doctors (and midwives) take
the time to counsel their patients about these dangers and how
emphatically and vociferously do they do so?
The scientific evidence, which now exists concerning the adverse
effects of smoking during pregnancy/nursing, is quite extensive.
Therefore, in my mind, the proposal, by Mitchell, to recommend that
smoking mothers not sleep with their babies seems about as ludicrous as
asking your adolescent child to make sure that they only use clean needles
when they use heroin. Of course, this seems ridiculous because any
competent and caring parent would tell them not to use the drugs in the
first place.
I would just like to present one last hypothesis as to another
mechanism that might contribute to the increased risk of Cot Death in co-
sleepers. One of the authors of the study by Blair et al, Jean Golding,
was cited in another source as having shown that exposure to chemicals in
air fresheners may have adverse effects on babies, as well as mothers (6).
Since perfumes and other fragranced cosmetics contain similar chemicals,
is it not possible that these products may also have negative impacts?
Exposure would be especially high in co-sleeping infants, who are snuggled
-up close, throughout the night, to a mother wearing perfume.
Theoretically, even fragranced detergent and fabric softeners used on
bedding and deodorant (under-arm) could have adverse effects.
To conclude, in response to the finding by Blair et al. (1)of a lack
of increased risk of Cot Death for co-sleeping in non-smoking families, it
is time for the medical community to stop scaring parents away from this
practice, as some current guidelines do (7). Instead, they should
establish a set of recommended guidelines for the safe practice of co-
sleeping.
The various risk factors need to be discussed and developed further
and most importantly, presented to parents. Among other things, the items
which should be addressed are: use of sofa, alcohol, drugs (recreational
or medicinal), obesity, post-partum depression, use of heavy blankets,
position of infant in relation to parents, over-crowded conditions, and
use of perfumes and other frangranced products. If and when these
guidelines are adopted and widely accepted, it may turn out that the non-
significant risk of co sleeping found by Blair et al, becomes a
significantly decreased risk.
Also, sleeping in separate rooms and on sofas should be discouraged,
unless and until other studies can prove that these practices are actually
not dangerous. Lastly, a better job needs to be done to discourage
smoking during pregnancy and lactation.
This subject seems to be getting a lot of long-overdue attention
lately, and hopefully, the study by Blair, et al, will help to keep the
momentum going on this issue. They should be commended for their work.
Cory A. Mermer
Westfield, New Jersey, USA
References:
1. Blair PS, Fleming PJ, Smith IJ, Platt MW, Young J, Nadin P, Berry
PJ, Golding, J, CESDI SUDI research group, Mitchell E. Babies sleeping
with parents: case-control study of factors influencing the risk of sudden
infant
death syndrome; Commentary: Cot death - the story so far. BMJ 1999;
319:1457-1462.
2. Nelson EA, Taylor BJ, Wetherall IL. Sleeping position and infant
bedding may predispose to hyperthermia and the sudden infant death
syndrome. Lancet 1989; i: 199-201.
3. Moy JG. Putting babies "Back to Sleep" [letter]. JAMA 1999;
281: 983.
4. McKenna J, Mosko S. Sleep and arousal, synchronicity and
independence, among mothers and infants sleeping apart and together (same
bed): an experiment in evolutionary medicine. Acta Paediatr Suppl 1994;
397: 94-102.
5. 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; 100:5 199-201.
6. Edwards R. Far from fragrant. New Scientist 4 Sept 1999.
(Comments by Jean Golding at conference on indoor air pollution in
Edinburgh in August 1999)
7. American Academy of Pediatrics Task Force on Infant Positioning
and SIDS. Does bed-sharing affect the risk of SIDS? Pediatrics 1997;
100:2 Pt 1 272.
Competing interests: No competing interests
The different sleeping environments of the surviving control infants
prior to
interview suggest sofa-sharing is relatively uncommon, for one cot death
to
occur in such an environment over 200 deaths would need to be observed.
However
more than one in twenty infants in this study died whilst asleep with a
parent
on a sofa, a ten-fold difference suggesting a highly significant
association. We took a detailed narrative account from the parents soon
after the death but did not have adequate resources in this study to
conduct a death-scene investigation. We therefore do not have
information for instance on the size or type of sofa or the exact
sleeping position of the parent in relation to the infant. The
narrative account suggests at least 4 of the 20 infants were ‘wedged’
between the parent and the back of the sofa and 8 of the cosleeping
parents had recently consumed alcohol. Most of these infants were less
than 4 months old.
We can hypothesize possible causative mechanisms such as entrapment
and
the inability of the infant to lose heat but further detailed study is
required before we can fully understand why sofa-sharing infants are at
increased risk.
Epidemiological studies can quantify the degree of risk associated
with
particular factors but not necessarily elucidate the causative
mechanism. For instance, prone sleeping is strongly associated with cot
death but we are not certain the role this position plays in the chain
of causality. Regardless of our limited knowledge, advice to place
infants on their back to sleep in both the UK and several other
countries has led to a large reduction in the number of cot deaths.
Given the findings of this study we feel it prudent to advise parents
not sleep with their infant on a sofa with the hope that this may lead
to further reductions.
Dr Peter Blair
p.s.blair@bristol.ac.uk
FSID Unit, Dept Child Health,
Royal Hosp for Children,
St Michael's Hill, Bristol, BS2 8BJ.
Competing interests: No competing interests
I have read the paper "Babies sleeping with parents: case-control
study of factors influencing the risk of the sudden infant death syndrome"
with great interest. I have a 4 month old daughter who was a premature
infant born at 30 weeks. This makes her at a higher risk of cot death, a
fact that, naturally, causes me some concern. She likes to sleep with me
at night and, as this often leads to longer sleep for both of us, I am
relieved to learn this will not greatly increase her risk of cot death. I
am interested to learn that sharing a sofa to sleep will. Both my husband
and I often fall asleep on the sofa with Ellie. I am interested to learn
why this practise can increase the risk of cot death.
Competing interests: No competing interests
Epidemiology or associated symptoms?
I applaud the authors of this article for giving us a study worth
reading and for presenting their data in a straightforward and scientific
manner. All too often researchers present their findings as cause and
effect, when all that has been proven is an association. Epidemiologists
need to remember that proving something is contributing to the etiology of
SIDS is quite different than showing something is associated with SIDS,
and forgetting this will lead to huge problems in our understanding of the
disease.
What I'm discussing may be easier to understand using an example.
Diabetics experience increased urine output, thirst, and increased fluid
intake, and the association between these symptoms of hyperglycemia and
diabetes would be statistically significant if compared to a control group
without diabetes. But this statistically significant association says
nothing about the etiology of diabetes, and one should not conclude that
these symptoms cause the disease or that decreasing fluid intake or
attempting to avoid urinating will prevent it. These are associated
symptoms of hyperglycemia; the etiology is a problem with insulin
regulation. The same situation applies to SIDS. If children who sleep
under certain conditions have a higher incidence of SIDS, this may be a
symptom or a contributing factor in why children die of this disease.
So little clinical information exists about SIDs that I am unable to
form an opinion about this. Do these children have sleep problems which
result in co-sleeping and sofa sleeping, (in my experience, parents of
children who sleep poorly usually end up sleeping with their children and
once the habit is formed, continue it even if the child is "settling") or
do they have parents who choose to sleep in bed or on a sofa with their
child, and thus increase the likelihood of SIDs occurring? No one ever
asks the parents, so I can't find the answer. The authors show that the
control group - not the SIDS group - was more likely to have settling
problems the day of the SIDS event, but they fail to discuss sleep history
for these children or ask why the parents chose to co-sleep or sofa sleep.
They also do not point out that children dying of SIDS are frequently
lethargic in the day prior to death and would not be expected to have a
settling problem that day. If SIDS is related to sleep apnea, it may be
possible these children sleep "better" than the control group because of
sleep deprivation and decreased sleep latency, but without a better sleep
history, not much can be said. It is interesting that in the control
group, the number of children who were bedsharing remained almost constant
throughout all the age groups, but in those dying of SIDS it was roughly
double the control group in the youngest infants, with this number
decreasing over time, but always above the control group. I would
hypothesize that within the control and SIDs groups a certain percentage
of parents choose to co-sleep, and this number remains constant in the
control group because it represents those parents. The number is
increased in the SIDS group because this number is comprised of those who
voluntarily bedshared, and those who were forced into bed-sharing by a
child who wouldn't settle alone. As the child became more sleep deprived
due to their sleep problem, they developed decreased sleep latency,
allowing the parents to move the child back to their own bed where they
had wanted them all along. But again, this is a hypothesis and we don't
have a clinical history to draw on which would answer whether this
possible interpretation is correct. Why did these parents choose not to
bedshare any more? Did they ever want to bedshare or did they do so just
to get some sleep? It would seem these are important questions, and it is
not clear why a good clinical history isn't included since these are
answerable questions that would make interpreting the data more easy.
In looking into these questions, I found that no detailed clinical
histories exist in the available medical literature about SIDs - none.
Those that appear to have been published are so old as to be unattainable,
and any clinical signs and symptoms are typically reported as isolated
features of the disease. This lack of clinical information doesn't appear
to be due to SIDs having no clinical evidence of its presence. Children
are reported to exhibit abnormal cries, lethargy, weight loss, poor fluid
intake, congestion, and other signs of illness in the days prior to death.
I would like to hear more about these clinical findings and less about
risk factors since they would allow us to spot and possibly treat the
child who is ill, not just prevent random children from dying. Certainly
it is important to counsel parents on good sleep habits and any practices
that lower the incidence of SIDs, but what is it we should be looking for
to clinically detect these children and stop their deaths? Concentrating
on risk factors and forgetting clinical information is counterproductive
if we lose sight of this. There are no diseases without clinical evidence
of their presence, just doctors in need of more education as to what the
signs and symptoms of disease might be. We cannot see what we don't know,
and unless we ask parents the right questions and more closely observe
these infants, we never will know how to diagnose and treat SIDs.
Competing interests: No competing interests