You are here
- Home
- Archive
- Volume 3, Issue 5
- Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case–control studies
Responses
PDF
Other responses
Jump to comment:
- David T MagePublished on: 23 June 2015
- Elizabeth BernsteinPublished on: 31 January 2014
- Elizabeth S. BernsteinPublished on: 19 December 2013
- Edward J. O'HaganPublished on: 7 October 2013
- Katharina SchuellerPublished on: 6 August 2013
- Robert CarpenterPublished on: 1 August 2013
- Edward J. O'HaganPublished on: 2 July 2013
- Elizabeth BernsteinPublished on: 21 June 2013
- Roger W. ByardPublished on: 14 June 2013
- Catherine M. FetherstonPublished on: 12 June 2013
- Herbert Renz-PolsterPublished on: 10 June 2013
- Paul N GoldwaterPublished on: 23 May 2013
- Charlotte K. RussellPublished on: 23 May 2013
- Sara L MacbayPublished on: 22 May 2013
- Peter S. BlairPublished on: 22 May 2013
- Published on: 23 June 2015A Possible Cause of SIDS is Found in SIDS Data of Carpenter et al., BMJO 2013, Figure 1 and Table 1
- David T Mage, Health Scientist
- Other Contributors:
- E. Maria Donner
Show MoreSIDS is characterized and diagnosed by exclusion, especially when expert pediatric pathologists, trained forensic investigators, and learned epidemiologists are unable to find a sufficient cause of infant death at autopsy, at the scene, or in the data, respectively. Therefore the cause of SIDS may be invisible or immeasurable. Furthermore, SIDS is characterized by the fact that in every such case that the parents had no...
Show LessSIDS is characterized and diagnosed by exclusion, especially when expert pediatric pathologists, trained forensic investigators, and learned epidemiologists are unable to find a sufficient cause of infant death at autopsy, at the scene, or in the data, respectively. Therefore the cause of SIDS may be invisible or immeasurable. Furthermore, SIDS is characterized by the fact that in every such case that the parents had no premonition that their infant was at imminent risk of death and "there was a complete absence of prodromal symptoms or any departure from the normal sufficient to justify the parent in seeking medical advice (1)." Therefore, most authors now look at risk factors that are not causes of death (prone sleep position, maternal smoking, etc.) However, they overlook three possible causal factors that alone might be insufficient to cause SIDS, but may act simultaneously in order to do so, that appear in this paper, as follows:
1) An unknown X-linked recessive allele for SIDS susceptibility with frequency q = 2/3, in Hardy-Weinberg-Equilibrium (HWE), that predicts the 50% male excess of SIDS for equal numbers of males and females at risk. The XY male is at risk with q = 2/3 and the XX female is at risk with q x q = 4/9, a 50% male excess. Given a nominal 5% male excess birth rate the expected male fraction is 0.612 for 60 x 1.05 males per 40 females (2). Table 1 reports 898 male and 578 female SIDS, for male fraction = 0.613;
2) Physiological anemia (3) causing the lognormal-type age distribution of SIDS in their Figure 1 that is its most unique characteristic. Unfortunately, "cases and controls over 1 year of age were excluded", which is like 'throwing the baby out with the bathwater.' The upper 1-year limit for SIDS is an artifact of the perception that SIDS over 1 year may include more false positives, so they should not be included in research studies.(4) Physiological anemia occurs when fetal hemoglobin (HbF) decays rapidly after birth while adult hemoglobin (HbA) slowly begins to replace the HbF. The term infant has its lifetime maximal total Hb at birth and it rapidly falls to their lifetime minimal Hb at or about 2-months of life, and then it follows a slow increase until a new equilibrium Hb concentration is established [we neglect here the different binding strengths of oxygen molecules to HbF and HbA, and pre-term Hb relations]. We estimated 19 as the missing number over 1-year that the authors excluded, by fitting an exponential decay curve to the numbers of SIDS from 8 weeks to 52 weeks in Figure 1 and extrapolating it to 178 weeks as determined previously as the upper limit for SIDS.(5) We then fit these weekly data (w) by the Johnson SB transformation that y = Log[(w + 1.343)/(178 - w)] where y = mu + sigma z, and mu = -1.025, sigma = 0.3, w is age in weeks, and z is a standard normal deviate. We interpret this Figure 1 as the number of the most anemic genetically susceptible infants whose total mix of Hb = HbF + HbA falls below some critical threshold, and therefore cannot survive a respiratory infection (see below) that reduces blood oxygenation in the lungs sufficiently to create an acute anoxic encephalopathy resulting in neuronal death. Hemoglobin concentration cannot be measured accurately at autopsy because of the gravitational settling of red blood cells during hemostasis leading to lividity so it is missing in virtually all SIDS autopsy studies (6); and
3. A prodromal respiratory infection (PRI) contracted from a family member that begins to fulminate shortly before or after the infant is placed for its final sleep. Farber (7) reported such a case where a baby was cared for by a nurse because the mother had a severe sore throat at the time. After a 2 p.m. uneventful feeding by the nurse, the baby was found dead 1.5 hours later, "lying on his back. There was no evidence of bedclothes, pillow or any other object over his face." We propose that the only thing that could have changed from the immediately previous sleep condition, sufficiently to cause death and be unnoticed, had to be a PRI that began to fulminate and cause fatal anoxic encephalopathy "before an important or appreciable amount of lung parenchyma has been involved." The author's Table 1 can provide no data on a potentially fatal respiratory infection presence in the SIDS cases, because, if such visible evidence existed at autopsy, the diagnosis by definition would not be SIDS. However, these data in Table 1 cast a visible shadow of a fulminating PRI. The authors list the SIDS and Control Live-Birth-Order (LBO) as shown below. We assume that Cohabiting Family Members (CFM) = 2 adults + (LBO - 1) Siblings = LBO + 1. We then assume that the probability of not carrying a communicable respiratory infection (CRI) (symptomatic or asymptomatic) is P so the probability of at least one CFM carrying a CRI is 1 - P^(CFM). Setting P = 0.92 by least squares gives the Model values shown. We previously fit U.S. LBO and SIDS data with P = 0.90 (8). Note that a plot of SIDS Fraction or SIDS Model vs CFM goes to the origin (0, 0) that implies that there are no SIDS cases here that are independent of CFM.
LBO = 1, SIDS = 407, Control = 1836, SIDS Fraction = 0.181, Model = 0.152;
LBO = 2, SIDS = 491, Control = 1566, SIDS Fraction = 0.239, Model = 0.219;
LBO = 3, SIDS = 280, Control = 748, SIDS Fraction = 0.272, Model = 0.280;
LBO = 4, SIDS = 149, Control = 304, SIDS Fraction = 0.329, Model = 0.337;
LBO = 5+, SIDS = 122, Control = 200,SIDS Fraction = 0.379, Model = 0.390.
In summary, SIDS may occur from these three causal factors (genetic susceptibility, physiological anemia, PRI) acting simultaneously that are a function of chance, age and CFM, respectively. There is no explanation we can find in the medical literature , other than an X-linkage in HWE, for a constant male fraction that is different from that of the male birth fraction. There is no explanation other than the maximum hemoglobin at birth for the virtual absence of SIDS immediately after live birth that is the most risky time for all other causes of infant death, the peak SIDS rate at or about the total Hb nadir, followed by an exponential decrease in SIDS rate as the infant's total HbA increases. We know of no other invisible cause of a sudden and unexpected infant death than a fulminating PRI that leaves no visible trace for the pathologist at autopsy or under the microscope in a genetically-susceptible physiologically-anemic infant.
REFERENCES
1. Davison WH. Accidental Infant Suffocation. Br Med J. 1945;2(4416):251-2.
2. Mage DT, Donner EM. A genetic basis for the sudden infant death syndrome sex ratio. Med Hypotheses 1997;48:137-42.
3. O'Brien RT, Pearson HA. Physiologic anemia of the newborn infant. J Pediat 1971;79:132-8.
4. Willinger M, James LS, Catz C. Defining the sudden infant death syndrome (SIDS): deliberations of an expert panel convened by the National Institute of Child Health and Human Development. Pediatr Pathol. 1991;11(5):677-84.
5. Mage DT. A probability model for the age distribution of SIDS. J Sudden Infant Death Syndrome and Infant Mortality 1996;1(1):13-31.
6. Poets CF, Samuels MP, Wardrop CA, et al. Reduced haemoglobin levels in infants presenting with apparent life-threatening events--a retrospective investigation. Acta Paediatr. 1992;81(4):319-21.
7. Farber S. Fulminating streptococcus infections in infancy as a cause of sudden death. NEJM 1934;211:154-9.
8.Mage DT, Donner M, A unifying theory for SIDS. Int J Pediat 2009; 2009:368270
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 31 January 2014Studies cited by authors in their reply do not in fact show routine bed-sharing to be a risk
- Elizabeth Bernstein, Author
Show MoreIn their August 1, 2013 reply to the various responsive letters, the authors characterize as a misunderstanding a previous statement by Vennemann et al. (1) that only non-routine bed sharing is a risk. They say that the two studies cited for that proposition show bed sharing on the last night to be a risk whether or not it is routine. This is incorrect as to both of the studies cited: the 1993 Scragg et al. paper from New...
Show LessIn their August 1, 2013 reply to the various responsive letters, the authors characterize as a misunderstanding a previous statement by Vennemann et al. (1) that only non-routine bed sharing is a risk. They say that the two studies cited for that proposition show bed sharing on the last night to be a risk whether or not it is routine. This is incorrect as to both of the studies cited: the 1993 Scragg et al. paper from New Zealand (2), and the 2009 Vennemann et al. paper from Germany (3).
The results of the 1993 New Zealand study were heavily linked to maternal smoking. With respect to non-smoking mothers the authors found that "the measure of bed sharing in the last sleep had no increase in risk."
As for the 2009 German study, it can only be understood by examining the coding used by its authors with respect to "usual" bed sharing. Among the parents of 333 SIDS cases asked what bed the child usually used in the last 4 weeks, 26 families gave the response "parental bed." This group, compared to controls, showed no increased risk of SIDS. (Table 2.) There were 146 families, however, who were coded as taking their child in the parental bed "sometimes every night" as opposed to "never" - apparently the only two options allowed. It was these families who gave rise to the results about increased risk for "bed sharing usual and bed shared during last sleep." (Table 3.)
In other words, the infants whom the authors of the 2009 study categorized as "usual" bed-sharers were mostly infants who primarily slept alone but who were taken into the parental bed on occasion. (For both cases and controls, roughly three-quarters of the "usual" bed-sharers did not have the infant in the parental bed during the last or reference sleep, indicating this was far from the primary sleeping arrangement among the "bed sharing usual" group.) Among parents who flex their sleeping arrangements this way, certainly one reason for taking the infant into the parental bed would be for comforting or settling when something seems "wrong." As a statistical matter, one might then see elevated instances of SIDS occurring in the parental bed not because the sleep location was a causative factor, but because that something "wrong" elevated both the probability of the infant being taken into the parental bed and the risk of SIDS.
It is a terrible disservice to lead non-smoking parents to believe that they endanger their infants by establishing a "family bed" from birth, when the evidence simply does not support that.
Elizabeth Bernstein
Bisbee, Arizona, United States
REFERENCES
1. Vennemann MM, Hense H-W, Bajanowski T, et al. Bed sharing and the risk of sudden infant death syndrome: Can we resolve the debate? J Pediatr 2012;160: 44-8.
2. Scragg R, Mitchell EA, Taylor BJ, Stewart AW, Ford RP, Thompson JM, et al. Bed sharing, smoking, and alcohol in the sudden infant death syndrome. New Zealand Cot Death Study Group. BMJ 1993;307:1312-8.
3. Vennemann MM, Bajanowski T, Brinkmann B, Jorch G, Sauerland C, Mitchell EA. Sleep environment risk factors for sudden infant death syndrome: the German Sudden Infant Death Syndrome Study. Pediatrics 2009;123:1162-70.
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 19 December 2013The authors' reply leaves unresolved the critical question of primary bed-sharing
- Elizabeth S. Bernstein, Author
Show MoreIn their August 1, 2013 reply letter, the authors characterize as a misunderstanding the conclusion of a 2012 meta-analysis (1) -- using many of the same databases and with one of the same co-authors as the instant study -- that only non-routine bed-sharing is associated with an increased risk of SIDS. In this way they dismiss a critical limitation of their own study, one which calls fundamentally into question the soundn...
Show LessIn their August 1, 2013 reply letter, the authors characterize as a misunderstanding the conclusion of a 2012 meta-analysis (1) -- using many of the same databases and with one of the same co-authors as the instant study -- that only non-routine bed-sharing is associated with an increased risk of SIDS. In this way they dismiss a critical limitation of their own study, one which calls fundamentally into question the soundness of the suggestion that the medical profession "should take a more definite stand against bed sharing." In fact no evidence has been produced against the possibility that primary bed-sharing, which is to say an arrangement in which the baby routinely spends most of every night in the parental bed, might be the safest sleeping arrangement of all. In the absence of such evidence, it is a terrible disservice to warn non-smoking parents against establishing the bed-sharing routine from birth.
The 2012 meta-analysis cited two papers -- a 1993 article based on New Zealand data (2) and a 2009 article based on the German 1998-2001 SIDS Study (3) -- for the proposition that "routine bed-sharing did not increase the risk of SIDS." The authors now say that those two studies in fact prove the opposite, that bed-sharing on the last night is a risk factor whether or not it is routine.
The New Zealand study found with respect to non-smoking mothers that "the measure of bed sharing in the last sleep had no increase in risk," so it hardly supports the present claims. As for the 2009 German paper, a closer examination reveals acute methodological limitations with respect to the question of routine vs. non-routine bedsharing. Those limitations explain how two subsequent interpretations could reach opposite conclusions on this essential point. They further illustrate dramatically the fact that warnings against primary bed-sharing to the present day lack evidentiary foundation.
The problem arises from the classification of families where parents occasionally allow the infant into the parental bed as "usual" bed- sharers. Among parents who flex their sleeping arrangement this way, certainly one reason for taking the infant into the parental bed would be for comforting or settling when something seems "wrong." One might then see elevated instances of SIDS occurring in the parental bed not because the sleep location was a causative factor, but because that something "wrong" elevated both the probability of the infant being taken into the parental bed and the risk of SIDS.
The only way to avoid this methodological problem would be to break out families which engage in "primary" bed-sharing, which is to say families in which the infant is in the parental bed for most of every night.
According to the 2009 paper on the German SIDS Study data set, of 333 cases only 26 families responded "parental bed" when asked what bed the child usually used in the last 4 weeks. But the researchers created an additional division of families into two groups, those who stated that their infants were never in the parental bed, and those who stated that they were in the parental bed either "sometimes" or "every night." There were 146 families of SIDS cases coded as being in the latter category. (Among the "sometimes/every night" control families, less than a quarter had the infant in the parental bed at the time of the reference sleep, emphasizing the fact that this grouping was in no way composed of primary bed-sharers.)
It was this larger group of 146 families, the group whose choice to take the infant into the parental bed on the last night might have been influenced by the perception of something wrong, which somewhat exceeded the control group in the percentage of infants sharing a bed during their last sleep, thereby giving rise to the finding about increased risk for "bed sharing usual and bed shared during last sleep." Among the smaller group of 26 families who reported the parental bed as the child's primary place of sleep, there was no increase seen in the risk of SIDS.
This latter result seems the more relevant one on the question of "routine" bed-sharing. At the very least it appears we have come a long way down the path of warning parents against the family bed without having resolved one of the most fundamental issues. The current study continues this unfortunate pattern.
Elizabeth S.Bernstein
Bisbee, Arizona, United States
REFERENCES
1. Vennemann MM, Hense H-W, Bajanowski T, et al. Bed sharing and the risk of sudden infant death syndrome: Can we resolve the debate? J Pediatr 2012;160: 44-8.
2. Scragg R, Mitchell EA, Taylor BJ, Stewart AW, Ford RP, Thompson JM, et al. Bed sharing, smoking, and alcohol in the sudden infant death syndrome. New Zealand Cot Death Study Group. BMJ 1993;307:1312-8.
3. Vennemann MM, Bajanowski T, Brinkmann B, Jorch G, Sauerland C, Mitchell EA. Sleep environment risk factors for sudden infant death syndrome: the German Sudden Infant Death Syndrome Study. Pediatrics 2009;123:1162-70.
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 7 October 2013Re:Authors' Reply to responses on BMJ Open to Bed sharing when parents do not smoke: Is there a risk of SIDS? An individual level analysis of five major cases-control studies.
- Edward J. O'Hagan, Retired
Show MoreLanguage in and of itself has no meaning . Only people have meaning. For example:
Through the Looking Glass, by Lewis Carroll
'I don't know what you mean by "glory",' Alice said.
Humpty Dumpty smiled contemptuously. 'Of course you don't --till I tell you. I meant "there's a nice knock-down argument for you!"'
'But "glory" doesn't mean "a nice knock-down argument",' Alice objected....
Show LessLanguage in and of itself has no meaning . Only people have meaning. For example:
Through the Looking Glass, by Lewis Carroll
'I don't know what you mean by "glory",' Alice said.
Humpty Dumpty smiled contemptuously. 'Of course you don't --till I tell you. I meant "there's a nice knock-down argument for you!"'
'But "glory" doesn't mean "a nice knock-down argument",' Alice objected.
'When I use a word,' Humpty Dumpty said, in rather a scornful tone, it means just what I choose it to mean -neither more nor less.'
'The question is,' said Alice, 'whether you can make words mean so many different things.'
'The question is,' said Humpty Dumpty, 'which is to be master -- that's all.'
******************************
With that in mind , and amending 'Dr.' to 'Mr.', the opening paragraph of the authors' response reads as follows :
Dr. O'Hagan's suggestion that SIDS may be due to maternal vitamin C deficiency seems unlikely because baby formula milk includes vitamin C supplement. We would therefore expect on his hypothesis that bottle fed babies would, if anything, be at lower risk than breast fed infants, which is not the case(1,2).
Comment: The above statement advisedly merits re- examination by its author(s) with a view to either its deletion or by their acknowledgement that what it concludes is totally in the realm of the non sequitur... and that's putting it mildly !
The Vennemann et al study (1) was structured to compare the two types of infant feeding and the totals of the diagnoses of SIDS respectively. However, neither it nor Dr. Carpenter's report (2), can be cited to support such an entirely erroneously based and thoroughly misleading conclusion.
Note that the implied objective of the response was to refute the " ...suggestion that SIDS may be due to maternal vitamin C deficiency ...". Accordingly, and inasmuch as the terms 'breast fed infants' and 'bottle fed babies' have been employed absent further qualification, they ambiguusly ... and need it be said , most conveniently... provide complete disregard of the the requirement of distinguishing between their employment in (1) and (2), and similarly nourished infants/ babies, who from the moment of their births are deficient in vitamin C.
In other words, Carpenter et al would have been well-advised to 'measure twice prior to cutting once'. Paradoxically, the direct opposite is what they should 'expect' if they were to give a ittle more thought to it ; namely, that vitamin C deficient "... bottle fed babies would, if anything, be at lower risk than breast fed infants ".
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 6 August 2013The validity of Carpenter's arguments from a statistical point of view
- Katharina Schueller, Statistician
- Other Contributors:
- Claudia Stuckart
Imputation of missing dataShow More
The REALCOM-IMPUTE imputation method, mentioned in the appendix of the article by Carpenter et al., is based on the assumption that the value of the variable of interest is missing completely at random (MCAR) or that the missing value can be explained by other variables and does not depend on the (unknown) value of the variable of interest itself (MAR, missing at random). Initial an...Imputation of missing dataShow Less
The REALCOM-IMPUTE imputation method, mentioned in the appendix of the article by Carpenter et al., is based on the assumption that the value of the variable of interest is missing completely at random (MCAR) or that the missing value can be explained by other variables and does not depend on the (unknown) value of the variable of interest itself (MAR, missing at random). Initial analyses and the study design allow the conjecture, according to carpenter et al., that the missing values of all variables are subject to MAR. There is no discussion in detail which methods were used and which assumptions were to take in order to carry out those initial analyses. Notably with respect to the survey questions on the parental use of alcohol or drugs we reasonably assume that the missing value depends on the value itself (NMAR, not missing at random). In that case the imputation method in use would be invalid.
The proportion of missing values in those questions on alcohol and drug use adds up to more than 60%. We consider an imputation based on only 40% of the data to be examined critically. After imputation, the odds ratio for alcohol consumption (odds ratio) increases from 5.1 to 6.4, i.e. by approximately 30%. On imputing drug consumption, OR increases from 19.2 to 30.7, i.e. by 60%. In general, we expect an increase of the OR when we deal with questions like those, as we assume that parents consuming alcohol and/or illegal drugs conceal their consumption depending on the extent of their drug use itself or depending on a combination of their drug use and other factors. The considerable difference in the increase of both odds ratios is not necessarily consistent with our expectation. If both variables are equally well-known risk factors, we would not expect that large differences in increase. Since imputing missing data produces a considerably larger increase in the OR for drug use, compared to the increase in the OR for alcohol use, we could explain the differences with the respondents being worried about possible legal effects provoked by true answers. Social desirability and the fear of criminal or civil prosecution might affect the respondents' willingness to answer so that, as a result, missing data could only be more or less arbitrarily imputed. This clarifies that unlike in the survey design at hand, such critical variables should be surveyed by the use of indirect measures so that a reliable data set can be acquired.
Generality and recommendations for parents
Overall statements on cause and effect which allow us to deduct recommendations require w.l.o.g., at least a quasi-experimental study design (selection of cohorts or randomized groups where the values of the potentially influencing variables - here: socio-demographic characteristics, use of legal/illegal drugs etc., and the exposition factor "bed sharing" - are controlled before the event of interest - here: child dies of SIDS - occurs). The selection of studies and/or the construction of the case and control groups do not comply with such a design. If correlations are found, we strongly advise to speak of coincidence or association, not of causation. Therefore the study cannot make any statements on causal relations that would allow us to draw conclusions on recommendations for today, the year 2013. Regardless of the fact that for ethical reasons, a study design other than a case control design is (hardly) practicable, our criticism on the generality of the statements remains unaffected.
Definition of variables
Furthermore we would like to question whether the variable definition chosen in the study is suitable to investigate the authors' hypotheses. The variable "bed sharing" is defined as the sleeping behaviour in the night before the interview (controls) or the night of the child's death (cases). That point in time will not necessarily reflect the regular parental behavior. We can imagine a variety of circumstances where parents take their child to their own bed as an exception only, for instance in response to their child behaving abnormally, which itself might indicate a hidden variable (sudden disease, stress factors, etc.) causing, ultimately, the SIDS. Therefore such a self-reported variable might not represent the parental behavior properly, resulting in biased outcomes. At least there should have been control questions asking for the frequency of and reasons behind the parental behavior.
Statistical regression model
The mixed logit model, generally applicable to data with a repeated measures structure, was used to model the regressions. However we could not find any details on whether only random intercepts or random slopes as well were included, and for which reasons. Additionally it is not clear to us which software and algorithms were used to estimate the model parameters. It is well-known that different algorithms may lead to different results. We would not consider that a major issue, however we would appreciate those technical details to be clarified.
Parameters
The authors fitted models with main effects and interaction terms. It is though not clearly comprehensible why only interactions between age, smoking, and alcohol consumption were includes or, at least, published. While "bed sharing" is the variable of primary interest, it is not included in any of the interaction terms. There are various reasons why that variable should be part of several interaction terms. For example, the mother's alcohol consumption might be her motive for "bed sharing", so that the latter should be a mediator variable within the statistical model. Bivariate analyses reveal as well that a very low birth weight appears to be a risk factor for SIDS. The multivariate model however lacks of an interaction term between alcohol consumption, smoking, and birth weight. The OR for SIDS increases dramatically when both the mother is consuming alcohol and the child is very young. The younger the child, normally the less is its weight. Simultaneously the birth weight is affected by the mother's alcohol consumption during pregnancy. There are obviously various interactions. It still remains to be explored how the combination of all those factors affects the risk for SIDS and whether the variable "bed sharing" has the character of an endogeneous rather than that of an independent variable. Last but not least further potential risk factors besides smoking, alcohol and (illegal) drug use, such as the mother's therapeutic drug use or any comorbidityies of parents and child, have so far not been considered.
Conclusion
The imputation of missing data is based on questionable assumptions, the risk factors to be considered are not exhaustive, plausible mediator affects are not part of the model, and the study design does not allow statements on causality. Therefore we judge most critically the authors' reasoning on the reduction of SIDS cases seemingly achievable by means of avoiding "bed sharing".Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 1 August 2013Authors' Reply to responses on BMJ Open to Bed sharing when parents do not smoke: Is there a risk of SIDS? An individual level analysis of five major cases-control studies.
- Robert Carpenter, Hon.Professor
- Other Contributors:
- McGarvey C, Mitchell EA, Tappin D, Vennemann MM, Smuk M, Carpenter JR.
Show MoreDr. O'Hagan's suggestion that SIDS may be due to maternal vitamin C deficiency seems unlikely because baby formula milk includes vitamin C supplement. We would therefore expect on his hypothesis that bottle fed babies would, if anything, be at lower risk than breast fed infants, which is not the case(1,2).
Professor Byard and Dr Hunsaker are concerned that the possibility of accidental suffocation was not addres...
Show LessDr. O'Hagan's suggestion that SIDS may be due to maternal vitamin C deficiency seems unlikely because baby formula milk includes vitamin C supplement. We would therefore expect on his hypothesis that bottle fed babies would, if anything, be at lower risk than breast fed infants, which is not the case(1,2).
Professor Byard and Dr Hunsaker are concerned that the possibility of accidental suffocation was not addressed. However, there are striking variations in certification of sudden unexpected death in infancy both within and between countries, and over time. Such cases might be certified as SIDS (R95), Unascertained (R99) or Accidental suffocation in bed with another person (W75). A strength of this analysis is that the studies were done at a time when all such deaths were labelled as SIDS, and thus cases were not excluded. To take an extreme example, if all bed sharing deaths were labelled as asphyxia (W75), then bed sharing would be protective for SIDS. Unlike Professor Byard (3), we find no evidence of an interaction between sex and bed sharing, p > 0.5. The percentage of males in all non bed sharing cases is 63.1 (95% CI 59.8 to 66.5) and for bed sharing cases 58.4 (95% CI 52.5 to 64.2). When the parents were non-smokers, the percentage of male cases is 63.6% that died when bed sharing and 62.8% for those that died elsewhere.
Fetherstone, Young and Kriske cite the percentages of case and control infants of non-smokers who were bed sharing, reported by Blair et al (4). as grounds for thinking that bed sharing is safe for this groups. However, Blair et al. do not give an adjusted odds ratio with a confidence limit for this group. It is essential that position the baby is put down to sleep in the cot, is taken into account when assessing the risk associated with bed sharing. To fail to do so, as was done in Blair et al's 1999 paper - see our response to the referees' comments on the web site - can be seriously misleading.
All studies either included direct questions on how the baby was fed at the time of death/interview, or include questions on the duration of breast feeding, from which the information can be obtained. There have been several studies which include feeding at the time of death(1,2,5). Belittling of Dr. L'Hoir's data on bed sharing fails to acknowledge their achievement of an exceptionally low SIDS rate.
The suggestions that the demonstration of the risks associated with bed sharing for non-smokers will result in more mothers falling asleep in armchairs or sofas, or that we have not emphasised the risks associated with bed sharing when combined with smoking, alcohol and drugs is irrelevant. Both these risks are all well established and the figures in the tables amply demonstrate the latter. To criticise the paper because we do not take account of putative benefits associated with bed sharing in relation to breast feeding similarly misses the point. Our paper is not a public health message, but provides a basis for the formulation of public health messages in relation to bed sharing. We do not say that 'bed sharing should never be undertaken in any circumstances'. But our finding that 88% of the bed sharing deaths in our data set are attributable to bed sharing and the implication to the current situation, set out in our penultimate paragraph, lead to the unavoidable conclusion that a further substantial reduction of SIDS rates could be achieved if parents avoided bed sharing. The observation by Elizabeth Bernstein that for some groups, the risks are extremely low is why we have set out how the absolute risk may be computed in almost any situation. Thus, our paper provides essential data for evidence based advice in relation to bed sharing.
Drs. Renz-Ploster and De Bock raise a number of criticisms. They argue that the epidemiology has changed over time, as has infant care practices, such as the use of sidecar cribs. We would challenge this. The magnitude of the odds ratios for SIDS risk factors is remarkably consistent between studies. What differs is the prevalence of these risk factors, for example few infants now sleep prone.
The use of sidecar cribs is infrequent, and few high risk families can afford such devices. The production of devices to minimise the risks associated with co-sleeping goes back a long way, as for example of the Arcutio presented to the Royal Society in 1731(6). This was a cage in which the baby was placed in the adult bed, so that the baby could be fed, but the bedclothes would not smother the baby. So far the safety of these devices is not evidence based. There is an urgent need for continuous ongoing case-control studies such as could readily be set up in the UK by standardising and integrating data from the Child Death Overview Panel (CDOP) enquiries, which are compulsory for every child death. These would provide an ongoing monitor of the safety of new devices and the ever- changing patterns of infant care.
Selection of controls, from the same maternity ward, is unlikely to eliminate social class bias due to the wide catchment areas most cover. The suggestion that we should have adjusted for socioeconomic status by household income would be problematic as mean incomes differ in the different countries studied and would also have varied over time. Maternal age and parity together are powerful predictors of socioeconomic status. Adjusting for these two variables has been found to be a better predictor of SIDS in Sheffield than the small a area index of Economic Deprivation Affecting Children (R Coombs and R Carpenter, personal communication).
With regard to criticisms regarding missing data, also repeated by other correspondents, we would draw attention to the 3rd paragraph of the discussion, which describes the care that we took to ensure that the data were complete for the key subgroup. Also see the response by Smuk and Carpenter. This shows that the significant risk associated with bed sharing when the parents do not smoke cannot be due to the imputation of missing data relating to maternal alcohol consumption in the last 24 hours.
The contention that the partners are probably responsible for the bed sharing deaths because they were under the influence of drink or drugs is inherently unlikely because they were also non-smokers. See also paragraph 2 of the discussion.
We think the there may be miss understanding in Vennemann et al's assertion that only non-routine bed sharing is a risk factor for SIDS (7). Both papers cited (5,8) show that after adjusting for other relevant factors, bed sharing on the last night is a risk factor, whether or not it was routine.
The contention that infants have evolved as co-sleepers, and consequently bed sharing is natural and safe, ignores the fact that until about three hundred years ago up to 50% of children died by the age of five. Thus, risks associated with co-sleeping would be minimal when compared with other risks when human infants evolved.
Unlike some of the other commentators Drs Goldwater and Bettelheim state that our findings are not surprising. They state we did not discuss the role of infection in the aetiology of such deaths. Bacterial colonization and toxin induction is a possible mechanism. Indeed, this mechanism may explain the strong interaction in risk of SIDS between parental smoking and bed sharing, because parental smoking is likely to greatly increase the bacterial colonization of the nasopharynx of their infants.
We were pleased to see that Drs Russell and Ball recognise that our paper allows parents to make an informed decision about any potential risk associated with their personal and cultural infant care beliefs and behaviours. Not only do we agree, but would argue that parents have the right to know the risk and health professionals have the duty to educate parents.
Dr. Helen Ball is correct that the comparative prevalence of bed sharing in controls would have been underestimated had we done as stated. However, our text should have read "Controls were bed sharing if the baby was in bed with them during the reference sleep." Where the reference sleep is defined as the period in the past 24 hours during which the corresponding case died. The care that was taken to ensure that control data was as closely matched as possible to the case data in respect of bed sharing is well illustrated in the detailed account of how the New Zealand control data was to be collected, given in the first section of our Appendix. We apologise for this error.
The risks from "hazardous" bed-sharing are well established and no- one we hope would support or encourage bed sharing if the parents smoked, or took alcohol or drugs, or slept on a sofa with baby. Our paper highlighted these risks. What the study also showed was that the risk was very low (0.8/10,000) if the socioeconomic and other demographic factors were in the low risk group and the parents followed the current SIDS prevention advice. The literature was unclear as to whether or not there was a risk from bed sharing in this low risk group. We showed clearly that the risk was increased, although the absolute risk is still small. However, when other risk factors are present the risk multiply and may produce substantially elevated odds ratios.
Dr. Blair and Professor Fleming's statement that we have no data on paternal alcohol consumption is not correct, see the 2nd paragraph of the discussion. They also suggest that the AOR for bed sharing for non- smokers must be lower than 2.7 (95% CI 1.4 to 5.3), because this is an overall figure, and therefore includes the risks of bed sharing for smokers. This is not correct, because all the models include the interactions of bed sharing with sleep position, smoking, maternal alcohol use and drug use. The AORs for bed sharing for the three parental smoking groups and when parental smoking is combined with maternal alcohol are given in Table 4. The AOR for bed sharing of 2.7 applies to all infants of non-smokers when bed sharing are compared with similar room sharing infants placed on their backs, for whatever combination of feeding, sex, birth weight, maternal age, parity, marital status, or race, because none of these factors interacted with bed sharing. Because the non-white racial group in this study is heterogeneous, we conclude that our findings primarily apply in the 'white European context' (3 lines from the end of page 8).
Dr Macbay is appalled that this paper managed to get past peer review and reach publication. As BMJ Open publishes online the reviews and the responses she will see that there was extensive peer review and all the criticisms were answered to the editor's satisfaction. She suggests that because we have grouped fully and partially breast fed infants together that our findings are irrelevant. Fully breast fed and partially breast fed infants are at lower risk of SIDS than bottle fed infants (1,2), but we found no interaction between type of feeding and risks associated with bed sharing. Consequently, our analysis suggests that the risks of bed sharing apply equally to fully, partially breast fed and bottle fed infants. The risk of bed sharing for fully breast fed infants when the mother did not smoke was demonstrated by Carpenter in 2006 (2). The other criticisms have been addressed above.
References.
1. Vennemann MM, Bajanowski T, Brinkmann B, et al. Does Breastfeeding Reduce the risk of Sudden Infant Death Syndrome? Pediatrics 2009; e406. doi:10.1542/peds.2008-2145
2. Carpenter RG. The hazards of bed sharing. The Canadian J Paediatrics and Child Health 2006; 11: Suppl A 24A - 28A.
3. Byard RW, Elliot J, Vink R. Infant Gender, cosleeping and sudden death . J Paediatr Child Health. 2012; 160: 44-8.
4. Blair PS, Sidebotham P, Edmonds M, et al. Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. BMJ 2009;339:b3666.
5. Scragg R, Mitchell EA, Taylor BJ, Stewart AW, Ford RP, Thompson JM, et al. Bed sharing, smoking, and alcohol in the sudden infant death syndrome. New Zealand Cot Death Study Group. BMJ 1993;307:1312-8.
6. St John O, Phil Trans 1731:422. Reported in Gentelmans and Historical Magazine 1746; 16:7 http://books.google.co.uk/books?id=Z3FIAAAAYAAJ&pg=PA8&dq=Gentlemans+magazine+Jan+1746&hl=en&sa=X&ei=ak3
7. Vennemann MM, Hense H-W, Bajanowski T, et al. Bed sharing and the risk of sudden infant death syndrome: Can we resolve the debate? J Pediatr 2012;160: 44-8.
8. Vennemann MM, Bajanowski T, Brinkmann B, Jorch G, Sauerland C, Mitchell EA. Sleep environment risk factors for sudden infant death syndrome: the German Sudden Infant Death Syndrome Study. Pediatrics 2009;123:1162-70.
Conflict of Interest:
As declared at the end of the paper
Conflict of Interest:
None declared. - Published on: 2 July 2013What medical science cannot explain, it can always conveniently choose to disregard
- Edward J. O'Hagan, Retired
(1) The stated objective of the study was: To resolve uncertainty as to the risk of Sudden Infant Death Syndrome (SIDS) associated with sleeping in bed with your baby if neither parent smokes and the baby is breastfed. Comment: How about parental poverty and malnutrition? Hypovitaminosis and more specifically maternal vitamin C deficiency have to be considered. Those aspects motivate this response, not alone by drawing attention...Show More(1) The stated objective of the study was: To resolve uncertainty as to the risk of Sudden Infant Death Syndrome (SIDS) associated with sleeping in bed with your baby if neither parent smokes and the baby is breastfed. Comment: How about parental poverty and malnutrition? Hypovitaminosis and more specifically maternal vitamin C deficiency have to be considered. Those aspects motivate this response, not alone by drawing attention to what follows re SIDS, but also by having become aware of the successful and wide-ranging employment of the ascorbate in treating everything from the poisonous bites of snakes and scorpions to cancer. http://www.ra-infection-connection.com/KalokerinosAA-vaccinesRefs.htm provides an abundance of information in that regard. (2)The conclusions: Bed sharing for sleep when the parents do not smoke or take alcohol or drugs increases the risk of SIDS. Risks associated with bed sharing are greatly increased when combined with parental smoking, maternal alcohol consumption and/or drug use. A substantial reduction of SIDS rates could be achieved if parents avoided bed sharing. Comment: Substance abusers are even more deficient in vitamin C. How about checking for such things as errors in concomitant variation and/or shaping things entirely to one's own preferences ... statistical interpretation disorder syndrome (SIDS) no less, which not infrequently can result in non sequitur conclusions. For example, if bed sharing is a major contributor to the 'SIDS' (?) fatality statistics which is clearly being claimed above, then it would follow that the observance of the-back-to-sleep (BTS) and all the other ' safe sleeping ' protocols, which includes ensuring that babies are not bed sharing but being (quote) 'placed supine on a cot in the parents' room', would automatically result in the prognosticated 'substantial reduction in SIDS rates' being evidenced not alone in parental abodes, but also and more especially when - and- where the infants are being supervised by trained care-providers in licensed daycare environments However, and paradoxically, that is very far being the case in the United States where annually approximately 20% all SIDS fatalities occur in daycare environments. 500 infants are dying every year in those locations; of which number 250 die during the first week, and 80-90 never make it past the very first day, and not infrequently in well-under 6 hours in the new environment ... sleeping supine and alone, in uncluttered cots, etc., and just as the 'experts' would approvingly have advocated. However, it seems that they prefer not to want to discuss these paradoxical occurrences which according to statisticians should be around 7% and not an alarmingly upwardly skewed 20%. Where we are to-day in SIDS goes all the way back to the following seminal study: Lancet. 1984 Sep 8;2(8402):535-7. Sudden infant death syndrome: increased carotid-body dopamine and noradrenaline content. http://www.ncbi.nlm.nih.gov/pubmed/6147601 The authors reported finding 10X the amount of dopamine in the carotid bodies of victims of SIDS as compared to the carotids of infants who had died from known causes. They pointed out that the neural discharge from the carotid bodies results in a decrease of ventilation, frequency of respiration and tidal volume, and appears to inhibit the carotids response to hypoxia. Additionally, and despite the fact that they concluded their report by noting that Denoroy et al had earlier reported decreased activity of dopamine beta-hydroxylase (DBH) in brainstem nuclei of SIDS infants, they did not draw attention to the fact that elevation of substrate dopamine to norepinephrine in the adrenal chromaffin cells, is dependent on the presence of vitamin C, copper and the DPH . In other words, vitamin C deficiency would have been an appropriate consideration ab initio in addressing the question as to why the elevated levels of dopamine were presenting in the first place. However, the authors' reference to the decrease in DBH activity, would lead one to infer that they were not unaware that any reference to subclinical scorbutus in infants being equated to SIDS, was a consideration which had already been rejected by NIH. N.B.For more on that aspect click on: http://www.bmj.com/content/341/bmj.c6654?tab=responses Later in 1992, Filiano and Kinney in the report of their study, Arcuate nucleus hypoplasia in the sudden infant death syndrome, http://www.ncbi.nlm.nih.gov/pubmed/1619439 , postulated that "... arcuate hypoplasia may lead to death by dyssynergy between cerebellar coordination of ventilation and autonomic/chemosensory/arousal integration, especially during sleep, hypercarbia, and in a critical developmental period '. Thereafter in 2006 , Kinney et al in a continuing series of related studies succeeded in identifying Multiple serotonergic brainstem abnormalities in sudden infant death syndrome. http://www.ncbi.nlm.nih.gov/pubmed/17077377 In 2010 it was followed by the study Brainstem serotonergic deficiency in sudden infant death syndrome... http://www.ncbi.nlm.nih.gov/pubmed/20124538 ... which concluded that compared to controls, SIDS was associated with lower 5-HT(serotonin) and TPH2 (central tryptophan hydroxylase) levels, consistent with a disorder of medullary 5-HT deficiency. (N.B. TPH2 is the key biosynthetic enzyme which is involved in elevating l-tryptophan to 5-HTP (5-hydroxytryptophan), the rate-limiting step in the synthesis of serotonin. The TPH2 requires vitamin C as a critically necessary cofactor. Absent the ascorbate, the upgrading process to serotonin can be further inhibited by deficits of folate and iron. In January 2011 Brainstem deficiency of the 14-3-3 regulator of serotonin synthesis: a proteomics analysis in the sudden infant death syndrome was published. http://www.ncbi.nlm.nih.gov/pubmed/21976671 Once again the deficiencies of TPH2 and 5-HT levels were noted; and based on the retrieved data, these abnormalities together with those relating to 5-HT(1A) receptor binding, were viewed as being related to 14-3-3 proteomics isoform deficits which had been found in the same SIDS cases. It was suggested that a potential molecular defect in SIDS related to TPH2 regulation may be occurring, inasmuch as 14-3-3 proteomics plays a critical role in that regard. Perhaps before proceeding further to studying the proteomics related aspects, SIDS researchers might consider reviewing a study conducted back in 2003: Cytochrome P450-mediated oxidative damage of nuclear membrane proteins and its prevention by vitamin C. (http://www.ncbi.nlm.nih.gov/pubmed/22900324 ), The authors noted that ascorbic acid exclusively prevents the oxidative modification and degradation of the membrane proteins, and that the observation assumes significance, particularly in subclinical ascorbic acid deficiency, where oxidative damage of the nuclear membrane would occur.Show LessConflict of Interest:
Corrected version herewith I have no competeing interests
Conflict of Interest:
None declared. - Published on: 21 June 2013The data in the appendix provide a different perspective
- Elizabeth Bernstein, Author
Show MoreTo the editor,
I am not a health care professional, but a mother who brought my babies into my bed a couple of decades ago when it appeared that that was the best way for me to get enough rest to be able to care for them. I have read with interest the recent meta-analysis by Carpenter et al. concluding that bed-sharing increases the risk of SIDS even in the case of breast- feeding, non-smoking mothers. [1]
...Show LessTo the editor,
I am not a health care professional, but a mother who brought my babies into my bed a couple of decades ago when it appeared that that was the best way for me to get enough rest to be able to care for them. I have read with interest the recent meta-analysis by Carpenter et al. concluding that bed-sharing increases the risk of SIDS even in the case of breast- feeding, non-smoking mothers. [1]
I thank Drs. Renz-Polster and De Bock for pointing out in their June 10 letter the importance of distinguishing between routine, planned bed- sharing and bed-sharing which may occur as a spur-of-the-moment deviation from the routine. For parents who want to know if they endanger their babies by establishing the bed-sharing habit from birth, a study which does not make this distinction is of limited practical use.
I have a further concern about the present analysis which has not been mentioned by the other letter-writers. In the section of the article intended to give an "overview of the absolute risks and increases in risk associated with bed sharing," the authors chose to focus on a "typical cohabiting white mother aged 26-30 having a second normal weight baby with birth weight between 2.5 and 3.5 kg," which they characterize as "the most common situation of a mother completing her family." They utilized, in other words, a weight categorization which includes the lower end of normal birth weights, where SIDS rates are higher, but excludes the entire upper end of the normal range, from 3.5-4.0 kg, where SIDS rates are extremely low. Of course all groupings of this sort are essentially arbitrary, but if a one-kilogram span was desired, then the range of 3.0- 4.0 kg would include far more babies than the one that was used. Among the control groups in this meta-analysis - which ought to better represent the general populations than the cases of SIDS - fully 50.1% of the babies had birth weights over 3.5 kg. It is possible that this percentage drops to less than half under the specific conditions of race, age, living situation and birth order cited above, but even so it is clear that the 3.0-4.0 kg category would be more representative than the 2.5-3.5 kg highlighted as "most common."
The many parents - myself among them - whose babies were born weighing over 3.5 kg would have to look at the online supplement to the article to find predicted rates of SIDS in room-sharing versus bed-sharing conditions. It turns out that for "cohabiting white women age 30+ with 1st baby birth weight > 3500g," the expected rate of SIDS for breast-fed infants sleeping supine and sharing a room with parents who are not smoking or intoxicated is 11 out of a million. For bed-sharing infants the expected rate is 31 out of a million.[2] In other words, my husband and I were theoretically running an additional risk of 1 in 50,000 by bed- sharing. That happens to be, according to the BMJ, the same as the risk of dying while playing soccer - something we also allowed our children to do. [3]
Elizabeth Bernstein
Bisbee, Arizona, United States
REFERENCES
1. Carpenter R, McGarvey C, Mitchell EA et al. Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case-control studies. BMJ Open 2013;3:e002299.doi:10.1136/bmjopen-2012-002299.
2. http://bmjopen.bmj.com/content/suppl/2013/05/10/bmjopen-2012- 002299.DC1/bmjopen-2012-002299supp_appendix.pdf
3. http://www.bmj.com/content/suppl/2003/09/25/327.7417.694.DC1
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 14 June 2013ARE ALL INFANT DEATHS IN PARENTAL BEDS DUE TO SIDS?Show More
Roger Byard, Professor of Pathology, The University of Adelaide, SA, Australia; John Hunsaker, Forensic Pathologist, University of Kentucky, Frankfort, Ky, United States.
Dear Sir,
We read with interest the recent analysis of five large case-control studies involving sudden infant death syndrome (SIDS) with respect to bed sharing [1]. Our concern is that the possibility of accidental suffocation was n...
Show LessRoger Byard, Professor of Pathology, The University of Adelaide, SA, Australia; John Hunsaker, Forensic Pathologist, University of Kentucky, Frankfort, Ky, United States.
Dear Sir,
We read with interest the recent analysis of five large case-control studies involving sudden infant death syndrome (SIDS) with respect to bed sharing [1]. Our concern is that the possibility of accidental suffocation was not addressed. While we would certainly agree with the authors that "certifying such deaths under headings other than SIDS does nothing to minimize the tragedy", it may help us to recognise and better understand different lethal mechanisms. For example, the different sex ratio in infants found dead alone (2:1) compared to a shared sleeping situation (1:1) suggests that these groups may not be the same [2]. The current analysis is not alone in "lumping" all shared sleeping deaths under the rubric of SIDS as a meta-analysis by Venneman et al in 2012 did exactly the same [3]. In previous years any infant who died in a cot was labelled a "cot death". While we came to understand the error of this approach, it now seems that any infant who dies in a parental bed is being labelled as SIDS. How does this approach help us recognise the possibility that some of these deaths may be due to accidental asphyxia [4]? Surely the increased risk of infant death if a parent has consumed alcohol has more to do with diminished parental arousal with overlaying, than having an effect on an infant's intrinsic homeostatic mechanisms?
As in many investigations of infant fatalities, the analysis of sudden unexpected deaths in infancy associated with bed sharing entails a broad, complex differential matrix prior to reaching a final decision on causality. As an example specifically relevant to this discussion, in a case of an infant found deceased in a standard adult bed with one or more adults, there may well be other compelling information, investigative omission, or physical evidence that concerns the forensic pathologist to such a degree that he/she is suspicious of a type of death that cannot be determined by either the information available or the autopsy findings. That is, the contribution of bed sharing is uncertain, irrespective of the role of parental smoking or substance abuse.
One of the great success stories in SIDS in recent years has been the realisation that this is a heterogeneous group of infant deaths composed of a number of distinct (although possibly inter-related) subsets. By identifying subsets we have been able to show that some of these deaths have been due to a range of conditions including accidental asphyxia, inflicted injury, metabolic disturbances, infections and cardiac channelopathies [5]. Unfortunately, John Emery's concern that SIDS may be used as a "convenient diagnostic dustbin" [6] if all deaths are grouped together is perhaps again being realised.
Roger Byard MD, John Hunsaker MD
Conflict of Interest: None
References
1. Carpenter R, McGarvey C, Mitchell EA, et al. Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case-control studies. BMJ Open 2013;3:e002299.
2. Byard RW, Elliott J, Vink R. Infant gender, cosleeping and sudden death. J Paediatr Child Health 2012;48:517-9.
3. Vennemann MM, Hense H-W, Bajanowski T, Blair PS, Complojer C, Moon RY, Kiechl-Kohlendorfer U. Bed sharing and the risk of sudden infant death syndrome: can we resolve the debate? J Pediatr 2012;160:44-8.
4. Byard RW. The triple risk model for shared sleeping. J Paediatr Child Health 2012;48:947-8.
5. Byard RW. Sudden Death in the Young. 3rd ed. Cambridge, UK, Cambridge University Press, 2010.
6. Emery JL. Is sudden infant death syndrome a diagnosis? Or is it just a diagnostic dustbin? Brit Med J 1989;299:1240.
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 12 June 2013Concerning conclusions that may negatively impact on safe sleeping, successful breastfeeding, parent wellbeing and open communication between health professionals and parents
- Catherine M. Fetherston, Associate Professor, School of Health Professions, Murdoch University
- Other Contributors:
- Jeanine Young, Professor of Nursing, School of Nursing and Midwifery, University of the Sunshine Coast and Chair, SIDS and Kids National Scientific Advisory Group. Sue Kruske, President, Australian College of Midwives and Director, Queensland Centre fo
Show MoreDear Editor,
The paper published in your journal by Carpenter et al [1], claims to resolve uncertainties associated with bedsharing and SIDS. We believe however, it does little other than fuel further confusion in the minds of parents and the health professionals charged with providing them with information and support.
As outlined in other letters the authors chose to use a selection of older studies...
Show LessDear Editor,
The paper published in your journal by Carpenter et al [1], claims to resolve uncertainties associated with bedsharing and SIDS. We believe however, it does little other than fuel further confusion in the minds of parents and the health professionals charged with providing them with information and support.
As outlined in other letters the authors chose to use a selection of older studies (data collected between 1987 and 2003) in their analysis that contained incomplete data on key variables such as alcohol, substance and medication use. Whereas a 2009 British study by Blair et al. [2] did collect data on smoking, alcohol, prescribed and illegal drugs in the 24 hours prior to death. These researchers did not find an increased risk for bedsharing and SIDS when these variables and other hazardous cosleeping arrangements, such as sofa and chair cosleeping were excluded. Instead they found that the number of babies bedsharing in the SIDS group (6%) was actually lower than in the control group where there were no deaths (10%).
It is also problematic that the older studies analysed by Carpenter et al. also used variable definitions for SIDS and for bedsharing and did not report breastfeeding data except in a few studies where only initiation data was collected. The lack of detail on how the breastfeeding data in Carpenter et al's analysis was obtained and what operational definitions were used for exclusive, partial and initiation of breastfeeding were not made available. Due to the known impact breastfeeding has on the prevention of SIDS [3] the lack of detail in this area places significant limitations on the study's findings.
Additionally the author's explanation of the saving of health costs associated with the reduction in bedsharing in relation to its effects on successful breastfeeding completely overlooks the very broad range of morbidities and mortalities that breastfeeding protects against across the lifespan. Likewise there is no acknowledgement when quoting unpublished data from the Netherlands, where there are active campaigns against bedsharing, that there is a strong likelihood that parents would not disclose their bedsharing activities due to fear of censure, resulting in significant potential for under reporting of bedsharing.
Furthermore there is an unhelpful emphasis on the comparably small risk for SIDS they have calculated from their analysis in relation to bedsharing. This potentially leads the novice reader, media and health consumer to conclude these results, and the author's subsequent finding that bedsharing should never be undertaken in any circumstance, are irrefutable. In reality, the significant limitations imposed by the study design make such a conclusion most inappropriate. Even the attempt to soften this conclusion by the authors saying they do not suggest babies shouldn't breastfeed in bed, only that they shouldn't fall asleep in bed, shows significant lack of insight into the realities and pressures of night time parenting.
Removing the option of bedsharing as a commonly used night-time parenting strategy is likely to have significant effects on the wellbeing of parents. It is also likely to impact negatively on successful breastfeeding, a factor associated with reduced infant mortality and morbidity, and may create a riskier sleep environment for some babies, when parents fall asleep on sofas or armchairs while trying to follow the authors' advice of 'never bedsharing'. The lack of acknowledgement of these important issues in the authors' discussion, overlooks important health promotion opportunities.
Consequently, we are most concerned by the message this paper sends to the general community. Mothers have already indicated in previous research [4] that advice against any form of bedsharing doesn't ring true for them and those who do decide to bedshare would not now disclose their intentions to health professionals. Such a response to the blanket advice against any bedsharing shuts down communication channels with parents and removes the opportunity for health professionals to have open discussions regarding safe sleeping options. Health professionals have also reported policies requiring them to advise against any bedsharing did not allow them to accept and respond to the realities of people's lives and negotiate low risk alternatives that were appropriate to their individual situations (4). As health professionals were then unable to meet parent's needs in this respect this created feelings of personal conflict and moral tension in having to comply with such policies.
Evidence based practice has been defined as the integration of best research evidence, clinical expertise and patient values [5]. The authors have proposed a risk elimination approach to the complex issue of bedsharing in this paper, which is not evidence based and will not be effective in reducing infant deaths for three key reasons: 1) current evidence is not reflected for low risk families; 2) bedsharing is a valued cultural practice in many communities including those of low and high risk; and 3) the practicalities of where mothers feed their babies at night needs serious consideration so that we do not introduce recommendations that actually increase risk for babies.
A consistent national risk minimisation approach to co-sleeping, in which the values and circumstances of families is integrated into the informed decision making process [6], is the only responsible and ethical approach to address this issue [7], and is required if we are going to succeed in further reducing sudden infant deaths associated with shared sleeping environments.
We are extremely disappointed with the conclusions reached by the authors of this study and the potentially negative impact it may have on safe sleeping, successful breastfeeding, parent wellbeing and open communication between health professionals and parents.
1. Carpenter R, McGarvey C, Mitchell EA, Tappin DM, Vennemann MM, Smuk M, Carpenter JR. 2013, Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case- control studies. BMJ Open 2013;3:5 e002299 doi:10.1136/bmjopen-2012- 002299.
2. Blair PS, Sidebotham P, Evason-Coombe C, Edmonds M, Heckstall- Smith, EMA, Fleming, P 2009: Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. BMJ 339(Oct13 1): b3666-b3666. doi: 10.1136/bmj.b3666
3. Hauck F, Tanabe K 2008, International trends in sudden infant death syndrome: stabilization of rates requires further action. Pediatrics 122(3): 660-666.
4. Dodd J 2012, Evaluation of the Department of Health Western Australia Operational Directive Statewide Co-Sleeping/Bed-Sharing Policy for WA Health Hospitals and Health Services Collaboration for Applied Research and Evaluation: Telethon Institute for Child Health Research under contract with the Department of Health, Western Australia.
5. Sackett, D. (2000) Evidence-based medicine: how to practice and teach EBM, Volume 2. London: Churchill Livingstone.
6. Young, J., Watson, K., Ellis, L., Raven, L. 2012, Responding to the evidence: Breastfeed baby if you can - the sixth public health recommendation to reduce the risk of sudden unexpected death in infancy. Breastfeeding Review 20(1): 7-15.
7. Fetherston, C., Leach, J. 2012. Analysis of the ethical issues in the breastfeeding and bedsharing debate. Breastfeeding review 20(3): 7- 17.
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 10 June 2013More questions than answers
- Herbert Renz-Polster
- Other Contributors:
- Freia De Bock
Show More
Dear Editor,
This publication [1] is confusing to parents and physicians because it draws far reaching conclusions from unconvincing data.
METHODOLOGICAL PROBLEMS
In our opinion, the validity of this analysis is threatened by selection bias, imputation bias and presentation bias. We think that the conclusions drawn from this study are not generalizable and that many current issues pert...Show Less
Dear Editor,
This publication [1] is confusing to parents and physicians because it draws far reaching conclusions from unconvincing data.
METHODOLOGICAL PROBLEMS
In our opinion, the validity of this analysis is threatened by selection bias, imputation bias and presentation bias. We think that the conclusions drawn from this study are not generalizable and that many current issues pertaining to SIDS epidemiology have not been adequately addressed.
Old data sets
The authors perform a secondary analysis of 5 previously published datasets from 1987 through 2003 from different parts of the world. It remains unclear by which criteria these studies have been selected and why the authors have chosen to include exactly these 5 studies and leave out other datasets from the same time period (a priori inclusion criteria are neither described in the paper nor in the supplementary data).
Sleep environments and habits have considerably changed since the time when most cases re-analyzed in this study have been recorded.[2] Public campaigns have resulted in a safer sleeping environment and the use of babybays (sidecar cribs) has changed the practice of bed-sharing profoundly. It is therefore questionable if the results of this study can be generalized to today's situation and questions of young families.
Problematic selection of control groups and adjustment
The weaknesses of some of the older studies now re-analyzed by Carpenter et al have not been done away with by lumping the datasets together. In 3 of the 5 studies now combined the controls born at the same maternity ward have not been selected randomly or by a predefined protocol, which may explain the wide discrepancies between cases and controls for important characteristics like marital status or maternal age in Carpenter's current publication (table 1). These discrepancies indicate that the control sample has not been drawn from the same source population as the cases - which in a case-control design of a condition like SIDS with its much higher prevalence in lower socioeconomic strata may introduce substantial bias.[3] Unfortunately, direct measures of socioeconomic status like household income or educational background do not seem to have been part of the original data sources of Carpenter's study. So, as a matter of fact, this study of a condition that is clearly more prevalent in disadvantaged social settings remains unadjusted for any direct measures of socioeconomic status. This may explain some of the surprising findings in table 1, according to which for instance - after multivariate adjustment - a mother in her late twenties is at nearly double the risk of losing her baby to SIDS than a mother aged over 30 years.
Problematic imputation
Alcohol and drug consumption are among the most important risk factors associated with SIDS in bed-sharing infants and are also more prevalent in the social settings in which SIDS most frequently occurs. In an analysis of all SIDS cases in Alaska between 1992 and 2004 for example, 32% of the bed-sharing victims had slept with an adult person impaired by alcohol or drugs during the night of their death [4] - a percentage that underscores that accurate information on this exposure has to be considered crucial for any valid statement on risk attribution to bed-sharing versus room-sharing.
In the source datasets of the study by Carpenter et al data on alcohol and drug consumption are missing for over 60% of the cases (table 1). To fill the gaps the authors resort to imputing the missing data. However, given the large amount of missing data, this strategy may introduce substantial bias.[5] [6] The imputation model as described by the authors in the supplementary online material is based on the assumption that values are missing at random between their datasets. This is almost certainly not the case for alcohol or drug consumption, the patterns of which differ along secular trends and between cultures and socioeconomic strata.[7] [8] In a study that draws cases from countries with widely disparant cultural backgrounds and drinking habits like Ireland and New Zealand it is, in our opinion, not valid to assume equal distribution and determinants for alcohol and drug consumption.
Important exposures missing
In spite of its claim to analyze risk factors "in depth as never before" (p 8) this study does not include some important risk factors for SIDS like paternal aclohol and drug use. This is not a trivial shortcoming as alcohol and drug use are one of the few plausible explanations why SIDS cases occur significantly more often on weekends than on week days[9], with an exceptional peak on New Year.[10] The authors do not see this as a threat to the validity of their result and claim in the discussion that paternal alcohol and drug use "does not add further to the risk of SIDS" (above and beyond maternal alcohol and drug use). They base this statement on an analysis from their ECAS study. [11] We remain unconvinced. For one, maternal alcohol and drug use is far from reliably ascertained in this study (see above). Also, it is plainly not conceivable to us why a drugged or drunken father sleeping next to his infant should not pose an additional and independent risk to baby's welfare.
Data on other important risk factors for SIDS, like the quality of the sleep environment or sleeping with a non-caregiver (see below), are also absent.
It could be argued that the less-than-perfect ascertainment or imputation of environmental influences is negligible as "it all dials out somehow between groups". This, however is not true and a serious misconception. Many of the exposures relevant in SIDS "act" primarily through the proximity to the baby and therefore affect bed-sharing infants and room-sharing infants differentially. In SIDS research, an equal influence of missing or unreliable data on subgroups therefore can not be assumed. Any mind altering influence for example (like alcohol or drug use) will have a much greater impact on the risk associated with bed-sharing than on the risk associated with room-sharing. Therefore, the less a study is able to capture exposures which act more effectively on bed-sharing infants than on room-sharing infants, the more risk will be falsely attributed to the bed-sharing cases. We wish the authors had commented on this very relevant issue in their discussion.
No information on intended versus non-intended bed-sharing
One of the co-authors, Mechthild Vennemann, has, in her recent meta-analysis of SIDS-cases[12] as well as in the report of her German study (which is also part of the dataset of the current study) [13] brought up an important issue. It is well known that bed-sharing occurs under two fundamentally different circumstances: in some families sharing the bed with the baby is part of the normal, everyday sleep routine (these cases have been called primary, routine or "usual" bed-sharers). Other mothers or families use bed-sharing only on special, exceptional occasions. It has to be assumed that some of the latter cases are associated with a higher a priori or subsequent risk. The baby may have been taken into the family bed because it has been unusually fussy during daytime (which may indicate the onset of an infection), or the mother may not have felt up to getting up at night, therefore taking her baby to bed with her. Some of these maternal choices may represent the influence of mind altering influences like alcohol, drug or medication use.
The differentiation of routine versus non-routine bed-sharing has proven important. In Vennemann's meta-analysis SIDS was twice as likely for non-routine bed-sharers than for primary bedsharers. As a matter of fact, in her dataset primary bed-sharing was not associated with a higher risk of SIDS at all ("Routine bed sharing did not increase the risk of SIDS. However, bed sharing during the last night when bed sharing was not routine increased the risk two-fold."). Given the importance of this ongoing discussion, brought up by one of the study authors herself, it certainly comes as a surprise that this issue is not even mentioned as a limitation of this study.
POTENTIAL MISCLASSIFICATION BIAS
Problematic definition of breastfed cases
The authors report breastfed versus bottle-fed cases. However, in their definition of "breastfeeding" they also include cases that are being partially bottle-fed. This is problematic as some of the cases labeled as "breastfed" may indeed not have been breastfed during sleep time but instead may have been in the process of being weaned during the period of their death.
Weaning is not a trivial issue in SIDS research as the peak-incidence of SIDS (with most cases occurring between 2 and 4 months of life) parallels the incidence of weaning especially in the social groups where SIDS is most common. [14] [15] This association asks for explanation and may indeed be causally related if one considers findings from sleep research: sleep regulation is entirely different in bed-sharing babies being breastfed during the night than in babies being bottle-fed at nighttime, with the former babies spending less time in deep sleep and showing higher arousability than the latter. [16] [17] A baby that is being actively weaned and for example not nursing at nighttime may therefore indeed be ridded of the protective dyadic sleep regulation typical for the fully breastfed infant.
The issue of a protective effect of a lighter, more active sleep is being widely discussed in the context of pacifier use, which has been shown to be one of the few protective influences against SIDS [18 ] [19] and which presumably also acts through its influence on sleep regulation - frequent suckling providing for an overall lighter, more active sleep with more arousals. This explanation is also in keeping with the epidemiological evidence that clearly shows a link between the intensity of breastfeeding and SIDS. [20]
Grouping fully breastfed and partially weaned babies together in a study of SIDS therefore is problematic to say the least. Indeed this lumping may be equivalent to lumping together protective factors and risk factors in one exposure - with all the confusing implications when it comes to interpreting the results. And again, as in the case of missing data, this putative misclassification may lead to an exaggerated risk attribution to the bed-sharing cases.
Possible lack of identification of non-caregivers
We are also concerned about another potential misclassification. Studies suggest that sleeping with a "non-caregiver" may pose a higher risk of SIDS to the infant than sleeping with a "usual caregiver" (who is usually a parent). In the study by Margaret Blabey of SIDS and bed-sharing, 29% of cases had slept with a non-caregiver (like an aunt, sibling or friend) during the night of their death. [21] It is not clear to us if the cases in the Carpenter study have really been classified along the line of parents versus non-parents, at least there is no mention about any exclusion of cases based on their status as non-parents or non-caregivers.
How easily misclassification or missing data can lead to false risk attribution to bed sharing may be illustrated by the following example: A mother may breastfeed her baby at daytime but bottle-feed at nighttime. As bottle feeding can be delegated, she may, to catch a good night's sleep, ask her partner to feed the baby at night while she sleeps in a seperate bed. Had the baby died during that night while sleeping at the side of the mom's partner, the case would have been classified and analyzed in this study as a breastfed baby dying of SIDS while bed-sharing. Had the partner chosen to drink alcohol in the evening, it would not have been identified as an additional risk.
INADEQUATE DISCUSSION OF ISSUES RELATED TO RISK INTERPRETATION OF BED SHARING
Breastfeeding has a protective effect on SIDS, cutting its risk by about 50%.[22] It also has a protective effect on maternal and child health, including risk reduction for infections, autoimmune disorders, and metabolic disorders. [23] It is well known that bed-sharing facilitates breastfeeding [24] [25] and is associated with higher rates of breastfeeding (as well as a longer duration). [26] [27] Therefore the question if general advice against bed-sharing - as put forward by Carpenter et al - may inadvertently jeopardize the practice and promotion of breastfeeding, is an important one as it may introduce the risk of not being breastfed in the name of avoiding possible risks associated with bed-sharing. [28] [29] The authors deal with this complex issue by referring to secular data from the Netherlands, in which bed-sharing incidence, according to a personal communication to the author, apparently dropped while breastfeeding rates went up. This resorting to ecological associations that have not been examined for causal relationship is unsatisfartory in this debate, and clearly other studies show opposite associations (i.e. decreasing incidence of SIDS in the face of increasing prevalence of bed-sharing) [30] [31], none of which are mentioned or discussed.
Bed-sharing as a lifestyle choice
The authors describe bed sharing as easily avoidable. This misses two important points. First, babies do sleep better in proximity with their attachment person, they fuss less, cry less and need less soothing. [32] For many families life with a small baby is a stressful period, as evidenced by the high prevalence of maternal fatigue, depression (peaking at 4 - 12 weeks post partum) and emotional dysregulation on the baby's part (e.g. excessive crying resp. "colics", peaking at 4 - 8 weeks post partum). [33] [34] So, for many parents, it is not easy to relinquish soothing care practices like bed-sharing. [35] [36] Second, it has been suggested that parents anxious about possible hazards of bed-sharing may choose alternative settings for breastfeeding instead, like sofas or armchairs, with the obvious risk of falling asleep in a clearly hazardous sleep environment. [37]
SUMMARY: A MISSED OPPORTUNITY
SIDS and bed sharing have been an interesting topic for scientists from many fields and specialties. Anthropologists point out that human infants have evolved as co-sleepers, i.e. they have, for reasons of thermoregulation and protection from predators, among others, slept with their (nursing) mothers during most of human history. According to this line of scientific enquiry, a baby sleeping solitarily would have been a dead baby until very recently. In fact, the overwhelming majority of human infants who have ever slept on earth have done so in close proximity to a nurturing, protecting and warming mother. [38] [39]
Modern sleep research supports this assumption of an evolved dyadic sleep arrangement by pointing out that the sleep physiology of an infant is clearly geared toward reciprocal regulation: sleep states are co-regulated in a sense that brain activity and sleep stages are synchronized. [40] [41] As part of intuitive interactions during light sleep the co-sleeping baby is frequently being repositioned by the mother and placed face to face in a position facilitating subsequent nursing, i.e. on its side or on its back. [42] [43] [44] Indeed, the observations of widespread supine sleeping in cultures where co-sleeping is part of the traditional cultural practice has been the starting point for questioning the prone position as the safest option for baby sleep in the 1980ies and has given decisive support to the back-to-sleep campaign so successful in preventing SIDS in industrial nations worldwide since the early 1990ies. [45] [46]
So from an ethological perspective it is ever more important to approach the core question in SIDS research with appropriate scientific tools - namely, if the well described risks associated with bed-sharing are a feature of bed-sharing in and of itself or if they stem from potentially avoidable environmental exposures that in some cases go along with bed-sharing.
We do not think that the study by Carpenter et al adds to answering this question. From a public health standpoint we are instead left with more confusion, as the answer given by Carpenter et al is based on an analysis with several shortcomings, including inadequate imputation, missing or unreliable ascertainment of confounders (like paternal and maternal drug use), missing data on routine versus non-routine bed-sharing as well as misclassification of some exposures (like partial bottle-feeding) - introducing a clear und possibly unidirectional bias in risk-attribution to the bed-sharing cases.
Given these constraints we would have hoped for a more cautionary interpretation. SIDS epidemiology has changed considerably since the times during which the data of this study have been collected, with the SIDS rate in countries like Germany now 89% lower than just 20 years back (down from a high of 1285 cases in 1991 to 147 cases in 2011). [47] In fact two recent analyses of SIDS cases in the post back-to-sleep era show that 99% of cases are now associated with at least 1 risk factor and that the vast majority of cases are now occurring in high risk environments. [48] So instead of making far reaching and statistically tenous claims about the dangers of bed-sharing even for those who do avoid the known risk factors, this study would have served parents and policymakers better by fitting their claims to their modest methodology.
The authors have shown that the vast majority of SIDS cases are related to avoidable environmental exposures: 78% of their SIDS cases were exposed to tobacco smoke, 65 % were not breastfed (and many more may have only partially been breastfed), 43 % were placed on their tummy for sleep and 43% died while sleeping in a separate room from their parents. These data are real and well ascertained. But the recommendations of the authors are directed against any and all bed-sharing in all kinds of circumstances during the first 3 months - which for us, given the source and limitations of their data, represents an overgeneralization. The authors, in our opinion, are advancing an issue to their core message, in which their data are the weakest.
No study so far has convincingly shown that there is a risk of dying for fully breastfed babies sleeping with their non-smoking or non-obtunded mother in a safe environment. We should therefore be more careful with the messages we bring out to the community, as many parents presume that whatever is being published by scientist represents rigorous research. Of course we should inform the public about the risks that can be associated with bed-sharing and that bed sharing needs to be practiced in a safe way. We should warn about the dangers of using mind altering substances or tobacco while co-sleeping with a baby. We should warn about unsafe sleep environments like sofas, armchairs or waterbeds. We should inform families that co-sleeping with a baby may not be fully safe if practiced by a non-caregiver or even a non-breastfeeding mother or the father.
However, we should be careful not to overextend our warnings. We should bear in mind that the decision to share the bed with baby also represents a lifestyle and relational choice. [49] Burdening this choice with warnings derived from less than rigorous data does not serve families. It also does not serve the debate when Carpenter et al, in an attempt to infer a causal relationship between SIDS and bed-sharing per se, refer to the behaviour of piglets, implying that the dangers of bed-sharing may indeed be a mammalian heritage (see their application of Bradford Hill's criteria on page 9): "Sows are normally separated by a bar from piglets to prevent them from being crushed when she turns over, but allowing her piglets to feed." This, to us, is biologistic folklore.
What we find especially disconcerting is the mode and style in which the discussion and the key messages are being put forward to the reader. Most parts of the discussion of this paper read like a discussion of a seminal randomized controlled trial of immense power and homogeneity, while in fact this is a secondary analysis of a convenience sample of mostly old case-control studies from different sociocultural regions of the world with limited adjustment, missing and probably mislabeled data and daring imputation. This is not all the fault of the authors, some of these constraints are inherent to the methodology and data sets that SIDS research, by its very nature, is restricted to. But these restrictions should be ever more of a reason to stay away from inflated claims and generalizations.
REFERENCES
1. Robert Carpenter, Cliona McGarvey, Edwin A Mitchell, David M Tappin, Mechtild M Vennemann, M Smuk, JR Carpenter. Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case-control studies. BMJ Open 2013;3:e002299
2. Trachtenberg FL, Haas EA, Kinney HC, Stanley C, Krous HF. Risk factor changes for Sudden Infant Death Syndrome after initiation of Back-to-Sleep campaign. Pediatrics 2012;129;630-8
3. Case-Control Studies. Encyclopedia of Quantitative Risk Analysis and Assessment
4. Blabey MH, Gessner BD. Infant bed-sharing practices and associated risk factors among births and infant deaths in Alaska. Public Health Rep 2009;124:527-34
5. Jin Hyuk Lee & John Huber Jr., 2011. "Multiple imputation with large proportions of missing data: How much is too much?," United Kingdom Stata Users' Group Meetings 2011 23, Stata Users Group.
6. Inngunn Myrtveit, Erik Stemsrud, Ulf H.Olsson: Analyzing data sets with Missing Data: An Empirical evaluation of imputation methods and likelihood-based Methods. IEEE Transactions on software Engeneering, Vol 27, Nr, 11, November 2011, p999-1013
7. Shield KD, Rylett M, Gmel G, Gmel G, Kehoe-Chan TA, Rehm J. Global alcohol exposure estimates by country, territory and region for 2005--a contribution to the Comparative Risk Assessment for the 2010 Global Burden of Disease Study. Addiction. 2013 May;108(5):912-22.
8. Day NL, Cottreau CM, Richardson GA. The epidemiology of alcohol, marijuana, and cocaine use among women of childbearing age and pregnant women. Clin Obstet Gynecol. 1993 Jun;36(2):232-45.
9. M?llborg P, Alm B. Sudden infant death syndrome during low incidence in Sweden 1997-2005. Acta Paediatr. 2010 Jan;99(1):94-8.
10. Phillips DP, Brewer KM, Wadensweiler P. Alcohol as a risk factor for sudden infant death syndrome (SIDS). Addiction. 2011 Mar;106(3):516-25.
11. Carpenter RG, Irgens LM, Blair P, et al. Sudden unexplained infant death in Europe: findings of the European Concerted Action on SIDS, ECAS. Lancet 2004;363:185 - 91.
12. Vennemann MM, Hense HW, Bajanowski T, Blair PS, Complojer C, Moon RY, Kiechl-Kohlendorfer U. Bed sharing and the risk of sudden infant death syndrome: can we resolve the debate? J Pediatr. 2012 Jan;160(1):44-8.e2
13. Vennemann MM, Bajanowski T, Brinkmann B, Jorch G, Sauerland C, Mitchell EA; GeSID Study Group. Sleep environment risk factors for sudden infant death syndrome: the German Sudden Infant Death Syndrome Study. Pediatrics. 2009 Apr;123(4):1162-70.
14. Wright CM, Parkinson KN, Drewett RF. Why are babies weaned early? Data from a prospective population based cohort study. Arch Dis Child. 2004 Sep;89(9):813-6.
15. Li R, Fein SB, Chen J, Grummer-Strawn LM. Why mothers stop breastfeeding: mothers' self-reported reasons for stopping during the first year. Pediatrics. 2008 Oct;122 Suppl 2:S69-76.
16. Volpe LE, Ball HL, McKenna JJ. Nighttime parenting strategies and sleep-related risks to infants. Soc Sci Med. 2013 Feb;79:92-100.
17. McKenna JJ, Mosko SS, Richard CA. Bed sharing promotes breastfeeding. Pediatrics 1997; 100: 214-219.
18. Hauck FR, Omojokun OO, Siadaty MS. Do pacifiers reduce the risk of sudden infant death syndrome? A meta-analysis. Pediatrics. 2005 Nov;116(5):e716-23.
19. Mitchell EA, Blair PS, L'Hoir MP. Should pacifiers be recommended to prevent sudden infant death syndrome? Pediatrics. 2006 May;117(5):1755-8. Review.
20. Fredrickson D, Sorenson J, Biddle A, Kotelchuck M. Relationship of sudden infant death syndrome to breastfeeding duration and intensity. American Journal of Diseases of Children 1993; 147: 460.
21. Blabey MH, Gessner BD. Infant bed-sharing practices and associated risk factors among births and infant deaths in Alaska. Public Health Rep 2009;124:527-34
22. Hauck FR, Thompson JM, Tanabe KO, Moon RY, Vennemann MM. Breastfeeding and reduced risk of sudden infant death syndrome: a meta-analysis. Pediatrics 2011;128:103-10
23. Dieterich CM, Felice JP, O'Sullivan E, Rasmussen KM. Breastfeeding and health outcomes for the mother-infant dyad. Pediatr Clin North Am. 2013 Feb;60(1):31-48.
24. Ball HL, Ward-Platt MP, Heslop E, Leech SJ, Brown KA. Randomised trial of infant sleep location on the postnatal ward. Arch Dis Child 2006;91:1005-10
25. Gettler LT, McKenna JJ. Evolutionary perspectives on mother-infant sleep proximity and breastfeeding in a laboratory setting. American Journal of Physical Anthropology 2011;144:454-62
26. Blair PS, Heron J, Fleming PJ. Relationship between bed sharing and breastfeeding: longitudinal, population-based analysis. Pediatrics 2010;126:e1119-e1126
27. Santos IS, Mota DM, Matijasevich A, Barros AJ, Barros FC. Bed-sharing at 3 months and breast-feeding at 1 year in southern Brazil. J Pediatr 2009;155:505-9
28. McCoy RC, Hunt CE, Lesko SM, Vezina R, Corwin MJ, Willinger M, Hoffman HJ, Mitchell AA. Frequency of bed sharing and its relationship to breastfeeding. J Dev Behav Pediatr. 2004 Jun;25(3):141-9.
29. Ball HL. Breastfeeding, bed-sharing, and infant sleep. Birth. 2003 Sep;30(3):181-8.
30. Trachtenberg FL, Haas EA, Kinney HC, Stanley C, Krous HF. Risk factor changes for Sudden Infant Death Syndrome after initiation of Back-to-Sleep campaign. Pediatrics 2012;129;630-8
31. Nelson E, Taylor B, Jenik A et al. International child care practices study: infant sleeping environment. Early Human Development 2001; 62: 43-55.
32. Morgan BE, Horn AR, Bergman NJ. Should neonates sleep alone? Biol Psychiatry 2011;70:817-25
33. Bayer JK, Hiscock H, Hampton A, Wake M. Sleep problems in young infants and maternal mental and physical health. J Paed Child Health. 2007;43:66-73.
34. Wake M, Morton-Allen E, Poulakis Z, Hiscock H, Gallagher S, Oberklaid F. Prevalence, stability, and outcomes of cry-fuss and sleep problems in the first 2 years of life: prospective community-based study. Pediatrics. 2006 Mar;117(3):836-42.
35. Kennedy HP, Gardiner A, Gay C, Lee KA. Negotiating sleep: a qualitative study of new mothers. J Perinat Neonatal Nurs. 2007 Apr-Jun;21(2):114-22.
36. Hooker E, Ball HL, Kelly PJ. Sleeping like a baby: attitudes and experiences of bedsharing in northeast England. Med Anthropol. 2001;19(3):203-22.
37. Blair PS, Sidebotham P, Evason-Coombe C, Edmonds M, Heckstall-Smith EM, Fleming P. Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. BMJ. 2009 Oct 13;339:b3666.
38. Gettler LT, McKenna JJ. Evolutionary perspectives on mother-infant sleep proximity and breastfeeding in a laboratory setting. American Journal of Physical Anthropology 2011;144:454-62
39. Konner M. 2005. Hunter-gatherer infancy and childhood: The !Kung and others. In: Hunter-gatherer childhoods: Evolutionary, developmental and cultural perpectives. BS Hewlett and ME Lamb (eds). New Brunswick: Transaction Publishers.
40. McKenna J, Mosko S, Richard C, Drummond S, Hunt L, Cetel MB, Arpaia J: Experimental studies of infant-parent co-sleeping: mutual physiological and behavioral influences and their relevance to SIDS (sudden infant death syndrome). Early Hum Dev. 1994 Sep 15;38(3):187-201.
41. 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 Nov;100(5):841-9
42. Baddock SA, Galland BC, Bolton DP, Williams SM, Taylor BJ. Differences in infant and parent behaviors during routine bed sharing compared with cot sleeping in the home setting. Pediatrics. 2006 May;117(5):1599-607.
43. Richard C, Mosko S, McKenna J. Sleeping position, orientation, and proximity in bedsharing infants and mothers. Sleep 1998; 19: 667-684.
44. 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 Nov;100(5):841-9
45. Davies DP. Cot death in Hong Kong: a rare problem? Lancet. 1985 Dec 14;2(8468):1346-9.
46. Nelson E, Taylor B, Jenik A et al. International child care practices study: infant sleeping environment. Early Human Development 2001; 62: 43-55.
47. http://www.gbe-bund.de, last accessed June 3rd, 2013
48. Trachtenberg FL, Haas EA, Kinney HC, Stanley C, Krous HF. Risk factor changes for Sudden Infant Death Syndrome after initiation of Back-to-Sleep campaign. Pediatrics 2012;129;630-8
49. Ball HL, Volpe LE. Sudden Infant Death Syndrome (SIDS) risk reduction and infant sleep location - moving the discussion forward. Soc Sci Med. 2013 Feb;79:84-91.
Authors:
Herbert Renz-Polster, MD, Paediatrician, Associated Scientist, Mannheim Institute of Public Health
Freia De Bock, MD, MPH, Paedatrician, Senior Scientist and Project Leader "Child & Health", Mannheim Institute of Public Health
Mannheim Institute of Public Health (MIPH),
Medical Faculty Mannheim, Heidelberg University Social and Preventive Medicine
Ludolf-Krehl-Strasse 7-11
D-68167 Mannheim
www.miph.uni-hd.deConflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 23 May 2013Re:Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case-control studies
- Paul N Goldwater, Senior Consultant
- Other Contributors:
- Karl A Bettelheim
Show MoreIn their paper1 Carpenter et al. conclude that bed-sharing of infants under 3 months with their parents even when the latter did not smoke and had no other risk factors the adjusted odds ratio for SIDS was 5.1 (2.3 to 11.4). This risk was greatly increased when the parents smoked, took alcohol or drugs. These findings are not surprising and confirm that co- sleeping is particularly dangerous for babies under three months of...
Show LessIn their paper1 Carpenter et al. conclude that bed-sharing of infants under 3 months with their parents even when the latter did not smoke and had no other risk factors the adjusted odds ratio for SIDS was 5.1 (2.3 to 11.4). This risk was greatly increased when the parents smoked, took alcohol or drugs. These findings are not surprising and confirm that co- sleeping is particularly dangerous for babies under three months of age. The main message is that co-sleeping (with or without other risk factors such as alcohol, other drugs or smoking) should never occur, but it is OK to breast feed in the parental bed and then to put the baby to sleep "face up" in its own cot (with clean linen) in the same room as the parent(s). The study pre-supposes that accidental smothering is the reason for these deaths while providing no autopsy evidence to that effect. The paper does not discuss the role of bacterial infection and only briefly alludes to it peripherally in terms of thermal stress (overheating) and the release of lethal toxins. It has been our view, that the theory of respiratory compromise due to smothering may account for only a very small minority of cases. The parental bed or sofa represents a dangerous sleeping surface, as both of these are heavily contaminated with bacteria such as Escherichia coli and Staphylococcus aureus which are equipped with a variety of lethal toxins. A baby sleeping in such a contaminated environment could inhale or ingest shed skin cells covered in these bacteria. A genetically susceptible baby is at risk in the co-sleeping situation. This risk is further increased with every additional risk factor.
Infection has long been recognised as a risk factor for SIDS as shown by symptoms and signs of respiratory and/or gastrointestinal infection in the days preceding SIDS death.2 Infection of normally sterile sites with major bacterial pathogens in SIDS has been independently described by two research groups.3,4 This finding could represent the 'footprint' of a bacteraemic episode prior to death.5,6 Bloodstream infection is a profound inducer of hypoxaemia; the mechanism of which is not fully understood, however, the cytokine storm almost certainly plays a major role7 and the hypoxaemia the storm produces seems to precede the final demise in monitored SIDS cases. In addition, it is plausible that the findings of Vargas et al.8 in relation to Pneumocystis lung colonization/infection, with increased mucus production, could contribute to hypoxaemia.
Not only is infection (via sepsis or toxaemia) an inducer of hypoxaemia, it is also a key thermogenesis inducer and is usually overlooked by SIDS researchers married to linking prone sleep to respiratory physiology. On the other hand, supportive evidence for plausible links between prone sleep position and infection have been proposed based on bacterial colonization and toxin induction with raised nasopharyngeal temperatures prone,9,10 and bacterial contamination of the sleeping surface promoting colonization of the infant's nasopharynx and gut.9-11 From studies in a neonatal ward conducted many years ago, it has long been known that organisms, such as E. coli are shed into the air and onto surrounding furniture by infants and their care-givers and that these bacteria can then colonize other infants and sometimes cause serious infection.12 Furthermore, as skin scales decorated with Staphylococcus aureus, and coliform (e.g. Escherichia coli) bacteria are shed in "dangerous sleeping environments" (parental bed, sofa, etc.) it is not unreasonable to conjecture that prone babies will inhale or ingest these bacteria and thereby provide the first step in a pathway leading to SIDS via the so-called "Common Bacterial Hypothesis of SIDS"13 or via toxigenic effects of these bacteria,5,6,14-24 Hypotheses of pathogenesis fail to adequately link all the known risk factors for SIDS into a holistic, all- encompassing hypothesis. Respiratory and brainstem and cardiac hypotheses compete to explain hypoxemia in the progression to SIDS and therefore do not satisfy a complete pathogenetic mechanism. A microbiological basis needs to be considered as the key for the epidemiological risk factors and the physiological findings provide a plausible explanation for SIDS.
Conflict of Interest: The authors have no conflicts of interest to disclose. Word count: 804 (excluding abstract and references)
Contributors' Statement: Paul N. Goldwater: Professor Goldwater drafted the initial manuscript and approved the manuscript as submitted. Karl A. Bettelheim: Dr Bettelheim conceptualized the article and reviewed and revised the manuscript and approved the manuscript as submitted.
REFERENCES 1. Carpenter R, McGarvey C, Mitchell EA, Tappin DM, Vennemann MM, Smuk M, Carpenter JR Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case-control studies. BMJ Open Access Research, May 21, 2013.
2. Hoffman HJ, Damus K, Hillman L, Krongrad E: Risk factors for SIDS. Results of the National Institute of Child Health and Human Development SIDS Cooperative Epidemiological Study. Ann N Y Acad Sci. 1988; 533: 13- 30.
3. Weber MA, Klein NJ, Hartley JC, Lock PE, Malone M, Sebire NJ: Infection and sudden unexpected death in infancy: a systematic retrospective case review. Lancet 2008; 371: 1848-1853.
4. Goldwater PN: Sterile site infection at autopsy in sudden unexpected deaths in infancy. Arch Dis Child. 2009;94(4):303-7. Epub 2008 Sep 15.
5. Goldwater PN. A perspective on SIDS pathogenesis. The hypotheses: plausibility and evidence. BMC Medicine 2011;9:64. http://www.biomedcentral.com/1741-7015/9/64.
6. Highet AR, Berry AM, Goldwater PN. Novel hypothesis for unexplained sudden unexpected death in infancy (SUDI) Arch Dis Child 2009; 94:841-843.
7. Hotchkiss RS, Karl IE. The pathophysiology and treatment of sepsis. New Engl J Med. 2003; 348(2): 138-150.
8. Vargas SL, Ponce CA, Gallo F, Felipe Astorga J, Bustamante R, Chab? M, Durand-Joly I, Iturra P, Miller RF, Moukthar Aliouat EI, Dei-Cas E. Near universal prevalence of Pneumocystis and associated increase in mucus in the lungs of infants with sudden unexpected death. Clin Infect Dis 2013; 56: 171179.
9. Blackwell CC, Gordon AE, James VS, MacKenzie DA, Mogensen-Buchanan M, El Ahmer OR, Al Madani OM, T?r? K, Csuk?s Z, S?tonyi P, Weir DM, Busuttil A. The role of bacterial toxins in sudden infant death syndrome (SIDS). Int J Med Microbiol 2002; 291:561-570.
10. Molony N, Blackwell CC, Busuttil A. The effect of prone posture on nasal temperature in children in relation to induction of staphylococcal toxins implicated in Sudden Infant Death Syndrome FEMS Immunol. Med. Microbiol. 1999; 25: 109-114.
11. Goldwater PN. Sudden infant death syndrome: A critical review of approaches to research Arch Dis Child 2003, 88:1095-1100
12. O'Farrell SM, Lennox-King SMJ, Bettelheim KA, Shaw EJ, Shooter, RA. Escherichia coli in a maternity ward. Infection 1976, 4:146-152.
13. Morris JA. The common bacterial toxins hypothesis of sudden infant death syndrome. FEMS Immunol Med Microbiol 1999; 25:11-17.
14. Bettelheim KA, Dwyer BW, Smith DL, Goldwater PN, Bourne, AJ. Toxigenic Escherichia coli associated with sudden infant death syndrome. Med J Aust. 1989; 151:538.
15. Bettelheim, KA, Goldwater, PN, Dwyer, BW, Bourne, AJ & Smith, DL. Toxigenic Escherichia coli associated with Sudden Infant Death Syndrome. Scand J Infect Dis.1990; 22:467-476.
16. Bettelheim, KA, Smith, H, Goldwater, PN, Morris, JA, Murrell, TGC, Sweet, C, Weaver, SA. Sleeping position and cot deaths Lancet 1991; 338:192.
17. Bettelheim KA, Chang BJ, Elliott SJ, Gunzburg ST, Pearce, JL. Virulence factors associated with strains of Escherichia coli from cases of sudden infant death syndrome (SIDS). Comp Immunol Microbiol Infect Dis. 1995; 18:179-188.
18. Bettelheim KA, Luke RKJ, Johnston N, Pearce JL, Goldwater, PN. A Possible Murine Model for Investigation of Pathogenesis of Sudden Infant Death Syndrome Curr Microbiol 2012, 64:276-282.
19. Goldwater, PN, Bettelheim, KA. Curliated Escherichia coli, soluble curlin and the sudden infant death syndrome (SIDS): J Med Microbiol 2002; 51:1009-1012.
20. Highet, AR, Goldwater PN. Staphylococcal enterotoxin genes are common in Staphylococcus aureus intestinal flora in Sudden Infant Death Syndrome (SIDS) and live comparison infants. FEMS Immunol Med Microbiol. 2009; 57:151-155.
21. Pearce JL, Luke RKJ, Bettelheim KA. Sudden infant death syndrome: What questions should we ask? FEMS Immunol Med Microbiol 1999; 25:7-10.
22. Pearce JL, Luke RKJ, Bettelheim KA. Extraintestinal Escherichia coli isolations from SIDS cases and other cases of sudden death in Victoria, Australia FEMS Immunol Med Microbiol 1999; 25:137-144.
23. Pearce JL, Luke RKJ, Bettelheim KA. Infection and food: a factor in sudden infant death syndrome? FEMS Immunol Med Microbiol 2004; 42:66- 75.
24. Pearce JL, Bettelheim KA, Luke RKJ, Goldwater, PN. Serotypes of Escherichia coli in Sudden Infant Death Syndrome. J Appl Microbiol 2010; 108:731-735.
Conflict of Interest:
None declared
Conflict of Interest:
None declared. - Published on: 23 May 2013Re: Bed sharing when parents do not smoke: is there a risk of SIDS?
- Charlotte K. Russell, Research Associate
- Other Contributors:
- Helen L. Ball
Show MoreCharlotte K Russell - Research Associate, Department of Anthropology, Durham University
Helen L Ball - Professor of Anthropology, Department of Anthropology, Durham University; Director of the Parent-Infant Sleep Lab
Dear Editor,
This publication analyses SIDS-risks associated with bed-sharing under different circumstances using data from five historical SIDS studies. Unlike previous analyse...
Show LessCharlotte K Russell - Research Associate, Department of Anthropology, Durham University
Helen L Ball - Professor of Anthropology, Department of Anthropology, Durham University; Director of the Parent-Infant Sleep Lab
Dear Editor,
This publication analyses SIDS-risks associated with bed-sharing under different circumstances using data from five historical SIDS studies. Unlike previous analyses of these data, this analysis includes data on feeding type. It promises, at last, to enable those of us working with parents and the staff who support them to be able to answer complicated but commonly asked questions about SIDS, and allow parents to make informed decisions about any potential risk associated with their personal and cultural infant care beliefs and behaviours.
Because this analysis is based on old data, however, it can only inform us as to the risks as they existed 15-26 years ago: knowledge both of bed-sharing behaviour and safety issues not considered here have advanced considerably in the intervening period. The authors present analysis of a dataset in which 'safer' bed-sharing is considered as distinct from bed-sharing in conjunction with known hazards (that is, breastfeeding infants bed-sharing in the absence of smoking, alcohol, drugs and sofa-sharing). Babies were categorized as bed-sharing if they were found dead in the parental bed, or woke up in the parental bed in the morning following the 'index night' (deviating from the original analyses which used many different definitions for 'bed-sharing'). This definition under-estimates the number of 'control' babies who bed-shared as some infants bed-share for a portion of the night and are placed in a cot following the last feed, so would not wake up in the parents' bed (and would be classed here as 'room-sharers'), however babies who die would not be returned to the cot. 'Accidental' (and consequently unplanned) bed- sharing is therefore also likely to be overrepresented in the SIDS group (Ball 2012). Both factors would inflate the risk associated with bed- sharing.
The authors predict a SIDS-rate of approximately 1 per 10,000 babies for room-sharers, and 2 per 10,000 babies for bed-sharers. The current rate of SIDS in UK is 1/3000, or 3.4/10,000, which means that both sleep locations for breastfed infants of non-smoking parents in the absence of alcohol experience very few SIDS deaths. It is curious, therefore, that the authors focus attention only to this small difference in predicted SIDS rates for breastfed babies of non-smoking parents who bed-share compared to room-sharing-while ignoring the hugely inflated risks associated with hazardous bed-sharing environments. It appears as though the authors choose to target breastfeeding mothers in this way as they are a sub-group with strong opinions about the benefits of bed-sharing, even though the infants of these mothers contribute negligibly to UK SIDS rates.
The recommendations of the authors that parents are advised to 'simply avoid bedsharing' indicates a worrying lack of cultural awareness or sensitivity to childrearing beliefs of different groups of parents on the part of these SIDS researchers. Such a recommendation does not allow parents, especially those whose infants are at low risk for SIDS (healthy term births, breastfed, not exposed to parental smoking or alcohol consumption), to make an informed choice to bed-share or not. While single message recommendations may have been appropriate and effective in previous campaigns targeting simple infant care practices such as supine vs. prone sleep position, they are inappropriate and ineffective for addressing infant care issues involving relational behaviours and cultural beliefs (Ball and Volpe 2012). Closing down all discussion of the reasons why parents might bed-share with their infant by issuing a dogmatic recommendation inhibits health professionals from raising the topic, causes parents to lie about their behavior, and stifles the provision of information about hazardous sleeping environments and the degrees of risk involved (Fetherston and Leach, 2013).
We argue that SIDS is not the only issue that must be taken into consideration when considering parent-infant bed-sharing, and that risk minimization, involving parent education and facilitating informed choice, is a more logical and ethical approach to the bed-sharing issue, than one focusing on risk elimination.
References:
Ball, H. L. (2012). Sleeping with the baby. IBFAN Breastfeeding Briefs, No. 53, September 2012.
Ball, H. L., & Volpe, L. E. (2013). Sudden Infant Death Syndrome (SIDS) risk reduction and infant sleep location - moving the discussion forward. Social Science & Medicine, 79, 84-91.
Fetherston C.M., & Leach, J.S. (2012) Analysis of the ethical issues in the breastfeeding and bedsharing debate. Breastfeeding Review, 20(3): 7-17.
Conflict of Interest:
Ball and Russell run the Infant Sleep Information Source (ISIS), an online source of infant sleep research information for parents and health professionals (www.isisonline.org.uk). They do so as part of their academic outreach activities and do not receive any income from the project.
Conflict of Interest:
None declared. - Published on: 22 May 2013Inadequate data, too many confounders and a ridiculous analogy!
- Sara L Macbay, GP
Show MoreI am appalled that this paper managed to get past peer review and reach publication. The authors define 'breastfed' as 'partially or completely breastfeeding at time of death or interview'. This implies there were babies entered into the breastfed group who were receiving formula in quantities that are unknown. What exactly is partially breastfed? Is this one, two, three bottles of formula per day? One breastfeed per da...
Show LessI am appalled that this paper managed to get past peer review and reach publication. The authors define 'breastfed' as 'partially or completely breastfeeding at time of death or interview'. This implies there were babies entered into the breastfed group who were receiving formula in quantities that are unknown. What exactly is partially breastfed? Is this one, two, three bottles of formula per day? One breastfeed per day? I cannot seem to find any more data about the feeding. They seem to have grouped partial and fully breastfeeding together, so one cannot extrapolate this data to exclusively breastfed infants.
The authors do not have data on maternal alcohol/ drug consumption in the previous 24 hours in three out of five of the studies. They have instead used imputed data to make up for this. How one goes about imputing values from studies conducted across the world with different cultural habits is beyond me!
I am struggling to understand why they looked at 'mothers who took 2 units or more of alcohol in the previous 24 hours'. This implies that there were mothers in the control group who had alcohol on board who co- slept with their babies which is clearly against advice regarding sleeping safely with your baby. Recalling units is also extremely difficult, given that most people (in my professional experience as a GP) cannot accurately estimate how many units they are drinking.
We also have no information on the partners' alcohol and drug consumption. Surely this must be an important variable to include in a study looking at all the hazards of bed-sharing?
If only the authors asked the right question (i.e. what is the risk of SIDS in babies who are exclusively breastfed who co-sleep with parents who do not drink alcohol, smoke or take drugs?), then we may actually have some useful answers.
Oh yes, I also object to being compared to a sow! The authors use the analogy of a sow crushing her piglets to meet the Bradford-Hill criteria. This is really scraping the barrel. If we [breastfeeding mothers] are now to be compared with animals, at least have the decency to compare us to animals of the same order who live in their natural environment and are not intensively farmed!
This study has now been splashed around the media as if it is gospel, despite its many flaws. I am deeply concerned about the message this study will send to exclusively breastfeeding mothers who will now be petrified of bed-sharing so will fall asleep on sofas/ breastfeeding chairs instead. In the meantime, mothers will co-sleep anyway without access to information on safe bed-sharing.
The BMJ Open has a duty to filter out the studies which draw conclusions from inadequate data which is subject to all kinds of bias, where the confounding issues have not been appropriately accounted for. I am disappointed BMJ Open - the consequences of publishing such a study could be disastrous!
Conflict of Interest:
Salaried GP, breastfeeding mother
Conflict of Interest:
None declared. - Published on: 22 May 2013A lack of data and questionable emphasis
- Peter S. Blair, Senior Research Fellow, University of Bristol
- Other Contributors:
- Peter J Fleming (Professor of Infant Health and Developmental Physiology, University of Bristol) and Sue Ashmore (Unicef UK Baby Friendly Initiative)
Show MoreDear Editor,
The report by Carpenter et al [1] on whether there is a risk of SIDS associated with bed sharing when parents do not smoke fails to meet its stated objective because the data needed to resolve any uncertainty is not available from the studies presented. The over-arching argument is whether bed-sharing in itself poses a risk to infants or whether the risk is within the hazardous circumstances in whi...
Show LessDear Editor,
The report by Carpenter et al [1] on whether there is a risk of SIDS associated with bed sharing when parents do not smoke fails to meet its stated objective because the data needed to resolve any uncertainty is not available from the studies presented. The over-arching argument is whether bed-sharing in itself poses a risk to infants or whether the risk is within the hazardous circumstances in which we bed-share. Parental smoking, recent alcohol consumption, use of illegal drugs and the use of sofas have all been identified in a detailed sleep scene investigation from a more recent SIDS study in the UK. [2] None of these older studies (data collected between 1987 and 2003) have data on recent parental drug consumption prior to the last sleep, none have data on recent paternal alcohol consumption and only two of the five studies have data on recent maternal alcohol consumption. It is not clear whether the parent who drank alcohol was the parent sleeping next to the infant for the last sleep and trying to impute values of maternal alcohol consumption from different cultures in different countries at different time periods when the question was not even asked makes unreasonable equivalence assumptions about maternal drinking.
The authors do have data on smoking status but conspicuous by its absence is the overall risk associated with bed-sharing amongst smokers and non-smokers. The reported adjusted risk for bed-sharing is 2.7 [95% CI: 1.4 to 5.3] and the age specific breakdown in Figure 2 suggests this risk is much higher amongst smokers (almost ten-fold). Conversely the overall adjusted OR for bed-sharing amongst non-smokers must be considerably lower than 2.7, probably just above unity and possibly not significant. The impression however given from the abstract is a five- fold increase in risk for bed-sharing when neither parent smoked but this is in reference to a group at very low risk of SIDS. The baseline group used in Table 3 is limited to breastfed girls placed on their back to sleep by the side of the parental bed with no other risk factors present serving to inflate subsequent adjusted risks. Thus the risk of bed-sharing when parents smoke is 21.6 and the risk when they also consume alcohol is 151. The lack of emphasis on such findings is quite astounding.
The strategy that parents should avoid bed-sharing as advocated by the authors ignores the fact that parents often fall asleep with their baby whether they intend to or not, particularly when they are feeding or caring for their baby during the night. We also know that some parents choose to bring their baby into bed for cultural reasons. Any guidance that recommends the total avoidance of bed-sharing can end up forcing parents into much riskier practices such as feeding on sofas and armchairs where they can easily fall asleep. It is surprising in this paper that this risk factor has not been explored as 3 of the 5 studies have previously reported on sofa-sharing with a combined univariable risk of almost 30, 10-fold higher than the risk associated with the parental bed.
UNICEF's leaflet and accompanying health professional guidelines on 'Caring for your baby at night' [3] do not actively promote bed-sharing but acknowledge that some parents will choose to sleep with their baby and fall asleep whether they intend to or not. Carpenter et al are not suggesting the baby is never brought into the bed so some parents, especially in the first months after the baby is born, will inevitably fall asleep. It is therefore important to emphasise risks that are a magnitude greater than other infant care practices and alcohol consumption, drug consumption, smoking and sofas fall into this category.
The risk of bed-sharing when the parents do not smoke appears quite negligible (when using the whole dataset) in comparison with the almost inestimably large risk of bed-sharing under hazardous circumstances.
References 1. Carpenter R, McGarvey C, Mitchell EA, Tappin DM, Vennemann MM, Smuk M, Carpenter JR. Bed sharing when parents do not smoke: is there a risk of SIDS? An individual level analysis of five major case-control studies. BMJ Open 2013;3:e002299. doi:10.1136/bmjopen-2012-002299 2. Blair PS, Sidebotham P, Evason-Coombe C, Edmonds M, Heckstall-Smith EM, Fleming P. Hazardous cosleeping environments and risk factors amenable to change: case-control study of SIDS in south west England. BMJ 2009;339:b3666. doi: 10.1136/bmj.b3666. 3. http://www.unicef.org.uk/BabyFriendly/Resources/Resources-for- parents/Caring-for-your-baby-at-night/
Dr Peter S Blair, Prof Peter J Fleming & Sue Ashmore
Conflict of Interest:
None
Conflict of Interest:
None declared.