Short-Term Effects of Breastfeeding
Short-Term Effects of Breastfeeding
Short-Term Effects of Breastfeeding
of breastfeeding
A SYSTEMATIC REVIEW
ON THE BENEFITS OF BREASTFEEDING
ON DIARRHOEA AND
PNEUMONIA MORTALITY
Contents
Summary
Chapter 1. Introduction
Chapter 5. Diarrhea
12
30
Chapter 7. Conclusions
43
References
45
iii
Summary
Background
Several studies suggest that breastfeeding has clear short-term benefits, particularly reducing morbidity and mortality due to infectious diseases in childhood. These benefits have been reported in
low and middle income and in high-income countries.
Objective
This systematic review and meta-analysis was aimed at assessing the effect of breastfeeding on respiratory infections and diarrheal disease in childhood.
Search strategy
Two independent literature searches were carried out, comprising the MEDLINE (1966 to December
2011) and Scientific Citation Index databases.
Selection criteria
We selected observational and randomized studies, published in English, French, Spanish or Portuguese that evaluated the associations between breastfeeding and diarrhea or respiratory infections
outcomes in children younger than 5 years of age. Studies that did not use an internal comparison
group were excluded from the meta-analyses. The type of categorization of breastfeeding varied by
study, but in all of them it was possible to compare a group with more intense breastfeeding practices with another with less intense breastfeeding. (e.g., ever versus never breastfed; breastfed for x
months versus breastfed for less than x months, exclusively versus partially or not breastfed, etc.).
Data extraction and analysis
Two reviewers using a standardized protocol independently evaluated the manuscripts; any disagreements were solved by consensus. Heterogeneity among studies was assessed with the Q-test
and I-square. Because heterogeneity was evident for all outcomes, random-effects models were used
throughout.
Effect on diarrhea
We identified 15 studies that provided 18 estimates on the effect of breastfeeding on diarrhea morbidity among children < 5 years. More intense breastfeeding practices were associated with a pooled
relative risk of diarrhea incidence of 0.69 (95% confidence interval: 0.58; 0.82) compared to less intense breastfeeding. Among infants aged 6 months, we obtained 49 estimates from 23 studies,
with a corresponding pooled relative risk 0.37 (95% confidence interval: 0.27; 0.50). We also identified
1
11 studies that evaluated children aged > 6 months, among whom the pooled relative risk was 0.46
(95% confidence interval: 0.28; 0.78). Breastfeeding also decreased the risk of hospitalization from
diarrhea [pooled relative risk: 0.28 (95% confidence interval: 0.16; 0.50) and diarrhea mortality [pooled
relative risk: 0.23 (95% confidence interval: 0.13; 0.42)]. Furthermore, we identified three randomized
trials of breastfeeding promotion; diarrhea morbidity was lower in the group receiving the intervention [pooled relative risk: 0.69 (95% confidence interval: 0.49; 0.96)].
Effect on respiratory infection
We identified 18 studies that provided 22 estimates on the effect of breastfeeding on any respiratory
infection outcome for any subgroup of under-five children, and 16 studies that restricted the analysis
to infants aged 6 months. Breastfeeding reduced the risk of hospitalization for respiratory infection by 57% [pooled relative risk: 0.43 (95% confidence interval: 0.33; 0.55)], and this protective effect
did not change with age. Studies that compared breastfed with non-breastfed children reported
the highest protective effect [pooled relative risk: 0.33 (95% confidence interval: 0.24; 0.46)] against
hospitalization for respiratory infection. Mortality from lower respiratory tract infections was also
reduced among breastfed children [pooled relative risk: 0.30 (95% confidence interval: 0.16; 0.56)].
Furthermore, breastfeeding also reduced the prevalence or incidence of lower respiratory tract infection [pooled relative risk: 0.68 (95% confidence interval: 0.60; 0.77)].
Limitations
Because nearly all studies included in the analyses are observational, we were not able to completely
rule out the possibility that the beneficial effect of breastfeeding was due to self-selection of breastfeeding mothers or residual confounding. Nevertheless, we identified three randomized trials in
which breastfeeding promotion reduced the risk of diarrhea.
Reviewers conclusion
The available evidence suggests that breastfeeding reduces the risk of diarrhea and respiratory infection. All effects were statistically significant, and for most outcomes the magnitude of the effects
were large. Protection was observed both in low income and high income countries.
CHAPTER 1
Introduction
Breastfeeding has well-established short-term benefits, particularly the reduction of morbidity and
mortality due to infectious diseases in childhood. A pooled analysis of studies carried out in middle/
low income countries showed that breastfeeding substantially lowers the risk of death from infectious diseases in the first two years of life (1). These benefits have also been reported in high-income
countries. Based on data from the United Kingdom Millennium Cohort, Quigley et al (2) estimated
that optimal breastfeeding practices could prevent a substantial proportion of hospital admissions
due to diarrhea and lower respiratory tract infection.
A systematic review by Kramer et al (3) confirmed that exclusive breastfeeding in the first 6 months
of life decreases morbidity from gastrointestinal and allergic diseases, without any negative effects
on growth. Given such evidence, it has been recommended that in the first six months of life, every
child should be exclusively breastfed, with partial breastfeeding continued until two years of age (4).
This systematic review and meta-analysis was aimed at assessing the effect of breastfeeding on respiratory infection and diarrhea disease in childhood.
CHAPTER 2
Methodological issues
Randomized controlled trials often provide the best evidence on the association between an exposure such as breastfeeding and a health outcome. Randomization results in a high likelihood that
the study will not be affected by confounding or self-selection (5). Furthermore, existing guidelines
propose standards for conducting, analyzing and reporting clinical trials, which help increase the
validity of the evidence (6).
On the other hand, the short-term benefits of breastfeeding evaluated in the present meta-analyses
are an ethical challenge to the design of randomized trials on the consequences of breastfeeding. It
is currently unethical to randomly allocate subjects to receive breastmilk. But, it is ethically sound to
allocate mothers to receive or not to receive breastfeeding counseling. In Belarus, the Promotion
of Breastfeeding Trial (7) randomly assigned maternity hospitals and their affiliated polyclinics to the
Baby-Friendly Hospital Initiative. The proportion of infants exclusively breastfed at 3 and 6 months
was substantially higher among infants from the intervention group. This trial is ethically sound because mothers were randomly assigned to receive intense breastfeeding promotion, compared to
usual care in the hospitals. On the other hand, compliance to the intervention was far from universal,
only 43.3% of the infants in the intervention group were exclusively breastfed at 3 months compared
to 6.4% in the comparison arm. In Mexico, Morrow et al (8) randomly allocated mothers to one of
the intervention group (six or three breastfeeding-counseling home visits) or to the control group.
The proportion of exclusively breastfed infants at 3 months was higher among those whose mother
received six visits. In another trial in India, mothers were assigned to receive or not visit on promotion
of exclusive breastfeeding, at 3 months the proportion of exclusively breastfed infants was higher
among infants in the intervention group (9). In these trials, intervention and control groups represented a mixture of breastfeeding practices. Therefore, the effect of breastfeeding is underestimated,
and statistical power is reduced.
The assessment of the evidence on the health consequences of breastfeeding is mostly based on
observational studies because of the small number of randomized controlled trials. Prospective birth
cohort studies are the next-best design in terms of strength of evidence.
Below, we discuss the strengths and weaknesses of observational studies, as well as approaches that
may help overcome their main shortcomings.
Factors affecting internal validity
Losses to follow-up
If losses to follow-up are high, selection bias may be introduced. This may affect both randomized
and observational studies. In order to assess the study susceptibility to selection bias, baseline data,
such as breastfeeding duration, should be compared between those subjects who were followed
up and those who were not. If attrition rates are not related to breastfeeding duration or other base 4
line characteristics, selection bias is unlikely (10). Unfortunately, this information was not available for
some studies.
Misclassification
Inaccurate measurement of exposure or outcome leads to misclassification. And misclassification
may be differential or non-differential.
Retrospective studies are more susceptible to recall bias and direction of bias may be modified. For
example, Huttly et al (11) observed that Brazilian mothers of high socioeconomic status tended to
overestimate the breastfeeding duration, whereas among poor mothers this was not the case. This
differential recall of breastfeeding duration would overestimate the protective effect of breastfeeding
because high socioeconomic status is associated with a lower risk of infectious diseases in childhood.
On the other hand, if the measurement error is not related to exposure or outcome, non-differential
misclassification occurs. Such bias underestimates the measure of association, and, therefore, reduces the likelihood of observing a significant association. Indeed, in a meta-analysis on the relationship
between maternal smoking in pregnancy and breastfeeding duration, the odds ratio for weaning at
3 months was inversely related to the length of recall for exposure and outcome (12).
Confounding
Confounding is one of the challenges in interpreting the evidence of observational studies. Even
large studies that managed to measure the possible confounders may still be affected by residual
confounding, if the confounder variables were not properly measured or adjusted for. Some methods
have been suggested to improve causal inference. These include comparison of siblings in withinfamily analyses, which allow controlling for unmeasured maternal and family variables (socioeconomic status, maternal variables) as well as for self-selection bias, because these characteristics are
shared among siblings. Usually, sibling studies assess the effect of discordance on breastfeeding duration or complementary feeding on the outcome. A limitation of these studies is that heterogeneity
in breastfeeding duration is smaller among siblings than that observed among unrelated individuals
and the sample size for the sibling analysis are smaller, decreasing statistical power. In the present
systematic reviews and meta-analyses, we did not observe any study on the short-term effects of
breastfeeding that have used this approach.
Another strategy involves the comparison across studies with a different confounding structure. In
this approach, if an association is causal, the association should be observed in every setting, in spite
of differing confounding structures.
Reverse causality
Reverse causality occurs when breastfeeding is stopped as a consequence of hospitalization or an
episode of diarrhea or respiratory disease (13). Cross-sectional and retrospective studies are more
susceptible to this bias that tends to overestimate the protective effect of breastfeeding against infections because the prevalence of breastfeeding is underestimated among those who developed
diarrhea or respiratory infection.
This bias can be avoided by the following strategies:
"" exclusion of deaths or episodes occurring within the first 7 days of life;
"" assessment of infant feeding practices before the onset of the episode.
5
CHAPTER 3
Search methods
Selection criteria for studies
In the present meta-analyses, we searched for observational and randomized studies, published in
English, French, Spanish or Portuguese that evaluated the associations between breastfeeding and
diarrhea and respiratory infections outcomes.
Studies that did not use an internal comparison group were excluded. We did not apply any restrictions on the type of categorization of breastfeeding (never versus breastfed, breastfed for more or
less than a given number of months, exclusively breastfed for more or less than a given number of
months). Instead, as discussed in the previous section, the type of categorization of breastfeeding
was considered as possible source of heterogeneity among the studies.
Type of outcome measures
In the present systematic reviews and meta-analyses, we searched for manuscripts that have assessed
the following diarrhea and respiratory infection outcomes:
"" Mortality
"" Hospitalization
"" Incidence
"" Prevalence
For the review on respiratory infection, we excluded those studies that evaluated the relationship
between breastfeeding and upper respiratory infection.
In the review on diarrhea outcomes, studies that evaluated the effect of breastfeeding on pathogenspecific diarrhea, such as Shigella or rotavirus, were excluded from the review.
Search strategy
We tried to identify as many relevant studies as possible, minimizing the likelihood of selection bias.
Two independent literature searches were carried out, using the terms described below. Initially, we
searched Medline (1966 to December 2011) using the following terms for breastfeeding: breastfeeding; breast feeding; breastfed; breastfeed; bottle feeding; bottle fed; bottle feed; infant feeding; human milk; formula milk; formula feed; formula fed; weaning.
Every breastfeeding term was combined with each of the following terms for the outcomes:
"" Mortality: Infant mortality; pneumonia AND mortality; pneumonia and death; respiratory infection AND mortality; respiratory infection and death; lower respiratory tract infection and mortality; lower respiratory tract infection and death; diarrhea AND mortality; diarrhea AND death.
7
"" Hospitalization: hospitalization; AND infant OR childhood; AND pneumonia OR respiratory infection OR lower respiratory tract infection OR diarrhea
"" Incidence/prevalence: infant OR childhood; AND pneumonia OR respiratory infection OR lower
respiratory tract infection OR diarrhea
Initially, we scanned through the titles of studies identified in the electronic search to exclude those
that were clearly irrelevant. Thereafter, abstracts were perused to further exclude studies. Finally, the
full text of the remaining studies was retrieved and relevant articles were identified. In addition to the
electronic search, reference lists of the articles identified was searched, and we perused the Web of
Science Citation Index for manuscripts citing the identified articles. Attempts were made to contact
the authors of all studies that did not provide sufficient data to estimate the pooled effect. We also
contacted the authors to clarify any queries on the study methodology or result.
CHAPTER 4
Review methods
Assessment of study characteristics
In the present systematic review and meta-analyses, study quality was not evaluated using summary scores. On the other hand, study characteristics considered as being relevant methodological
aspects were assessed and the contribution of each one to the heterogeneity among studies was
evaluated (19). The following study characteristics were abstracted.
TABLE 4.1
Categorization
Sample size
Continuous
Follow-up rates
(if applicable)
Continuous
Type of study
Randomized trial
Birth cohort
Other
Categorization of breastfeeding
Outcome
Incidence
Prevalence
Mortality
Hospitalization
None
Age
Socioeconomic and demographic variables Socioeconomic, demographic and
birthweight
Socioeconomic, demographic, birthweight and maternal smoking
Low income
Middle income
High income
Continuous
Continuous
Yes
No
Data abstraction
Two independent reviewers extracted data on the above characteristics from each study using a
standardized protocol, and disagreements were resolved by consensus rating.
Data analysis
Pooled effect estimates
In the meta-analyses, effect measures were presented as pooled relative risks. Definition of exposure
to breastfeeding followed the classification used in each study, and a relative risk <1 indicated that
breastfed subjects presented a lower risk of the outcome.
Fixed or random-effects model
To pool the studies estimates, we used a fixed and a random-effects model. Under the fixed-effect
model, we assume that there is one true effect size and the difference among studies results is due
to random variation. In the fixed-effect model, studies are weighted by their precision (inverse of the
standard error) (20). On the other hand, under the random-effects model we assume that the true
effects also vary, and the pooled effect needs to take into consideration the additional source of variation. In the random-effect model, studies are weighted by their precision plus the estimate of the
between studies variance (heterogeneity) (21). By incorporating a second source of variability (variance between studies) in the estimate of the variance, the confidence interval in the random-effect
model is wider than that for the fixed-effect model. Because the between studies variance is the
same for every study, the random-effect model gives greater weight to smaller studies.
In the present meta-analyses, heterogeneity among studies was assessed with the Q-test and
I-square; if either method suggested that between-studies variability was higher than that expected
by chance, a random-effects model was used (21). Otherwise, a fixed-effect model is recommended.
In this series of meta-analyses, heterogeneity was evident for all outcomes, and thus random-effects
models were used throughout.
Publication bias
Studies reporting statistically significant associations are more likely to be published and to be cited
by others articles, whereas small studies with negative findings are less often published. Therefore,
studies reporting an association are more likely to be included in a systematic-review. Publication
bias is more likely to affect small studies because the great amount of resources (time and money)
spent in larger studies makes them more likely to be published, regardless of their results (20). In
meta-analysis, publication bias is a type of selection bias.
Funnel plots and Eggers test are usually employed to assess the presence of publication bias (22),
but in the present meta-analyses we did not estimate the likelihood of selection bias with the funnel plot or the Eggers test because several comparisons were done. On the other hand, the analyses
were stratified according to study size, in order to assess the impact of publication bias on the pooled
estimate.
Reverse causality
Cross-sectional and retrospective studies on the short-term consequences of breastfeeding are susceptible to reverse causality. Because breastfeeding may be stopped due to an illness or hospital
admission, the assessment of feeding status at the moment of the interview may increase the proportion of non-breastfed infants among those who developed the outcome. This bias tends to overestimate the short-term protective effect of breastfeeding. In order to avoid such bias, the study should
evaluate the feeding practices prior to the onset of the episode. In the present meta-analyses, we
evaluate whether the studies were susceptible to reverse causality.
10
Assessing heterogeneity
The last phase of the analyses relied on meta-regression to assess the contribution of study characteristics to between-study variability (23). In this approach, if the data are homogenous or if the
heterogeneity is fully explained by the covariates, the random-effects model is reduced to a fixed
effect model. This analysis was performed using the METAREG command within STATA. Each of the
items listed in table 4.1 were included as covariates in the meta-regression, one at a time, rather than
using an overall score. This approach allows the identification of aspects of study design that were
responsible for heterogeneity between studies (24).
11
CHAPTER 5
Diarrhea
In spite of recent progress, diarrhea remains as one of the leading causes of death among children
<5 years. In 2010, diarrhea was estimated to have caused about 800,000 child deaths globally (25).
In 1984, a comprehensive review indicated that promotion of breastfeeding was one of the most important interventions for controlling diarrhea among children (26). In the 2003 Lancet Child Survival
series, breastfeeding promotion was again identified as one of the most cost-effective interventions
against under-five deaths in general, and against diarrhea in particular (27).
Biological plausibility
Several mechanisms for a possible protective effect of breastfeeding against gastrointestinal infections have been proposed, including the presence in breastmilk of substances with antimicrobial or
immunological properties, avoidance of contamination (as in non-human milk or baby bottles), and
the general nutritional status of breastfed infants.
Breastmilk contains several antimicrobial and anti-inflammatory factors, hormones, digestive enzymes and growth modulators that protect against infections. Below, we briefly discuss the evidence
on the protective effect of some of the components of breastmilk.
Oligosaccharides are the third largest solid component of human milk. It has been suggested that
oligosaccharides homology to cell surface carbohydrates would block the attachment of pathogens
to the infants mucosa, preventing the development of gastrointestinal infections (28).
Breastmilk also confers immunity against gastrointestinal infections by carrying antibodies (secretory
IgA) produced by mothers who have been exposed to such pathogens, protecting the infant from
developing an infection (29,30).
Lactoferrin, one of the main proteins in human milk can destroy pathogens and reduce inflammatory responses. Furthermore, lactoferrin increases the activity of the immune system because it is a
growth factor for lymphocytes (31).
A second mechanism is that non-breastfed infants are more exposed to pathogens that may cause
diarrhea than breastfed subjects. Many studies attest to the presence of pathogens in foods offered
to infants. For example, in The Gambia, Rowland et al (17) observed that weaning foods traditionally
given to children were contaminated with microorganisms that could cause gastrointestinal infections. Another study from Chile showed that most feeding bottles harbored large numbers of pathogens that could cause gastrointestinal infection (18).
Last, it has been proposed that in low-income settings optimal breastfeeding practices can prevent
undernutrition associated with repeated infections and with the use of over-diluted breastmilk substitutes (32). Good nutrition is essential for non-specific immunity that contributes to fighting infections in general.
12
CHAPTER 5. DIARRHEA
When reviewing the mechanisms through which breastfeeding may protect against diarrhea, it is important to consider the different settings where studies are carried out. Whereas the biological characteristics of breastmilk do not seem to vary markedly in high- and low-income societies, the other
two mechanisms (contamination and nutritional status of non-breastfed infants) are likely to play a
much larger role in poor societies without proper sanitation and with inadequate weaning foods.
Overview of the evidence
The protective effect of breastfeeding against mortality and morbidity from diarrhea has been widely studied. In the electronic search, we identified five systematic reviews and/or meta-analyses on this
subject.
In 1984, Feachem & Koblinsky (26) systematically reviewed the evidence on the association between
diarrhea morbidity and infant feeding. In relation to exclusively or partially breastfed infants, the median risk of morbidity from diarrhea among infants who were not breastfed was 3.0 for those younger
than 2 months of age, 2.4 for infants aged 3 to 5 months and about 1.4 for those aged 6 to 11 months.
Among children older than 1 year of age, no association between breastfeeding and diarrhea was
observed. Furthermore, the relative risk of diarrhea in infants under 6 months of age who were not
breastfed ranged from 3.5 to 4.9, in comparison to those who were exclusively breastfed. In this age
group, it was estimated that promotion of breastfeeding would reduce diarrhea morbidity by 8% to
20%, depending on different assumptions, and mortality would decrease by 24 to 27%.
Huttly et al (33) updated in 1997 the Feachem & Koblinsky (26) review of potential interventions for
the prevention of morbidity from diarrhea in childhood. Breastfeeding was again pointed as one of
the key preventive strategies for prevention of childhood diarrhea.
Also in 1997, Golding et al (34) systematically reviewed the evidence on the relationship between
breastfeeding and cause-specific diarrhea or gastroenteritis. Both in developed or in developing countries, exclusive breastfeeding protected infants under six months from diarrhea and gastroenteritis.
The protective effect was not consistent for rotavirus infection but was clearly observed for non-viral
pathogens. The authors concluded that breastfeeding protected the infant against non-viral diarrhea.
In 2004, Kramer et al (3) reviewed the evidence on the effect on child health and growth of exclusive breastfeeding for 6 months. Morbidity from gastrointestinal diseases was lower among infants
who were exclusively breastfed for 6 months, in comparison to infants exclusively breastfed for 34
months.
Recently, Lamberti et al (35) evaluated the effect of breastfeeding duration on morbidity and mortality from diarrhea. Among infants younger than 6 months, the risk of dying from diarrhea was 10.5
(95% confidence interval: 2.79; 39.6) times higher among those infants who were not breastfed in
relation to those who were exclusively breastfed. Among children in the age range 6 to 23 months,
the protective effect of breastfeeding was smaller, but still statistically significant [relative risk 2.18
(95% confidence interval: 1.44; 4.16)]. Therefore, the evidence from this recent review also indicated
that breastfeeding protects against diarrhea.
In the present systematic review and meta-analysis, the electronic search identified 41 studies that
provided 81 estimates on the relative risk of morbidity, mortality or hospitalization from diarrhea according to infant feeding.
Studies published since 2007 that were not included in any of the previous cited systematic reviews
and meta-analyses reported on research carried out in Denmark (36), Bangladesh (37), Philippines (38),
13
Zimbabwe (39) and Guinea (40), and two from the United Kingdom (2, 41), all of which confirmed the
presence of strong protection by breastfeeding. Further details on these studies are provided in Tables 5.1 to 5.3.
Update of the existing meta-analyses
A new meta-analysis was carried out including (a) studies included in the existing meta-analyses; (b)
the recently published studies described above, and (c) a small number of older studies that had not
been identified by the earlier reviews but were detected in our computerized search starting in 1966.
Studies reporting on subgroups of infants for example, preterm or low birthweight were not included in the review. Neither were studies reporting on diarrhea due to specific agents.
Due to the large number of available studies, we calculated pooled effects separately for morbidity,
hospitalizations and mortality. For each outcome we present initially analyses that included all children under five years, followed by studies restricted to children under six months of age, and to those
including children aged 6-59 months.
For children < 5 years of age, we identified 15 studies that provided 18 estimates on the effect of
breastfeeding on morbidity (Table 5.1). Figure 5.1 shows that among children < 5 years of age, breastfed children were less likely to present diarrhea. When studies reporting on incidence and prevalence
were combined, the pooled relative risk was 0.69 (95% confidence interval: 0.58; 0.82)].
The effect of breastfeeding among infants aged 6 months was also assessed. We identified 23
studies (Table 5.2) that provided 49 estimates on the effect of breastfeeding on diarrhea morbidity
or mortality. Figure 5.2 shows that morbidity due to diarrhea was lower among breastfed infants.
[pooled relative risk 0.37 (95% confidence interval: 0.27; 0.50)]. Because there was clear heterogeneity
among studies, the estimates were pooled using a random-effect model. With respect to age at assessment of morbidity, we observed that the protective effect of breastfeeding seemed to be largest
among infants younger than 4 months. On the other hand, because the confidence interval of the
estimates from younger children overlapped with those for children aged 5 and 6 months, the difference was not statistically significant. (Figure 5.3). In Figure 5.4, the association between morbidity
(incidence or prevalence) from diarrhea and infant feeding was stratified according to the categories
of feeding used for comparison. Those studies that compared exclusive breastfed infants with nonexclusive breastfeeding and those that compared partial breastfed infants with those not breastfed
were those reporting the smallest mean effect of breastfeeding on morbidity. On the other hand,
the effect of breastfeeding was highest in studies that compared exclusively breastfed with nonbreastfed infants.
In the literature search, we identified 11 studies that provided 14 estimates on the association between breastfeeding and morbidity or mortality from diarrhea among children aged > 6 months.
(Table 5.3) The risk of morbidity from diarrhea was lower among those infants who were breastfed
[pooled relative risk 0.46 (95% confidence interval: 0.28; 0.78)]. (Figure 5.5) With respect to the categorization of breastfeeding, most studies on children > 6 months of age compared children who
were breastfed with those who were not breastfed. Independent of the categorization, there were
consistent inverse associations between breastfeeding intensity and diarrhea outcomes. (Figure 5.6)
Figure 5.7 shows that breastfeeding decreased the risk of hospitalization from diarrhea [pooled relative risk: 0.28 (95% confidence interval: 0.16; 0.50)]. There was marked heterogeneity among studies,
and the protective effect of breastfeeding was higher among young infants.
14
CHAPTER 5. DIARRHEA
With respect to mortality, breastfeeding markedly decreased the risk of diarrhea mortality [pooled
relative risk: 0.23 (95% confidence interval: 0.13; 0.42)]. Similarly to the observed for hospitalization,
the effect of breastfeeding was higher among infants younger than 6 months. (Figure 5.8)
Table 5.4 shows that infants who were exclusively breastfed presented lower risk of morbidity from
diarrhea even in relation to predominantly breastfed infants. Furthermore, infants who were not
breastfed had the greatest risk of morbidity or hospitalization.
We identified three randomized trials in which breastfeeding promotion was related to diarrhea outcomes, whose results are summarized below. Because such analyses do not entail a comparison of
breastfeeding categories, they could not be incorporated in our main meta-analyses. In Mexico, Morrow et al (8) showed that breastfeeding promotion increased the duration of breastfeeding; diarrhea
episodes in the intervention group were reported for 12% of all infants, compared to 26% in the
comparison group, a protection of 52%. In a randomized trial of exclusive breastfeeding promotion
in India, the 7-day diarrhea prevalence was lower in the intervention than in the control communities
at 3 months [0.64 (95% confidence interval: 0.44; 0.95)] and 6 months [0.85 (95% confidence interval:
0.72; 0.99)] (9). In the Belarus PROBIT trial (7), maternity hospitals were randomized to receive or not
to receive promotion of exclusive breastfeeding. Children in the intervention group were less likely to
present one or more episodes of gastrointestinal infections [odds ratio 0.60 (95% confidence interval:
0.40; 0.91)]. Hospital admissions were similar in both groups [odds ratio 0.92 (95% confidence interval:
0.62; 1.37)].
Data from the Belarus trials were also analyzed to compare risks of diarrhea according to breastfeeding categories, and these are incorporated in the present meta-analyses (Table 5.1). Taken together,
the results of these trials support the presence of a causal effect of breastfeeding promotion against
diarrhea morbidity. The pooled results on diarrhea morbidity from these three trials are presented in
Figure 5.9, with a pooled relative risk of 0.69 (95% confidence interval: 0.49; 0.96). Such a significant
effect was detected even though analyses were based on intent to treat, that is, both intervention
and comparison groups included compliers and non-compliers.
Conclusion
In the assessment of the evidence on the short-term consequence of breastfeeding, confounding is
an important methodological issue that should be taken into consideration, as discussed in a previous section. In low-income countries confounding is expected to underestimate the benefit of breastfeeding on diarrhea outcomes because breastfeeding tends to be more frequent among the poor. In
high-income countries, where the rich tend to breastfeed for longer than the poor, confounding may
be expected to operate in the opposite direction. In the meta-analysis on morbidity from diarrhea
among infants 6 months of age, most studies were from low and middle-income countries. We observed that the pooled estimates were similar between studies that only reported unadjusted results
[pooled relative risk: 0.39 (95% confidence interval: 0.30; 0.49) and those that adjusted their estimates
for socioeconomic and other variables [pooled relative risk 0.35 (95% confidence interval: 0.23; 0.54)].
This suggests that the present meta-analysis was not affected by confounding. Furthermore, in lowincome settings we expected negative confounding, that is, underestimation of the true effect, and
results from most studies showed strong protection.
A second type of bias is reverse causality, that is, breastfeeding may have been interrupted or modified because of the diarrhea episode, thus leading to an association in the opposite direction than
the one being postulated (13). Only one study explicitly accounted for this possibility, by ensuring that
15
information on feeding practices was obtained for a date previous to the onset of the episode. This
study showed substantial protection associated with breastfeeding (42).
Publication bias is another methodological issue that should be taken into consideration in the assessment of evidence from meta-analyses. The selective publication of small positive studies may
overestimate the benefit of an intervention. Funnel plot is one of the strategies used to assess the
susceptibility of the meta-analysis to publication bias. In the present meta-analysis, given the large
number of different comparisons being made, we opted for not generating funnel plots. Instead, we
stratified the analyses by sample size and observed that among infants 6 months of age, the mean
effect of breastfeeding on morbidity from diarrhea was similar among studies with a sample size <
1000 subjects (pooled relative risk 0.39) and those with 1000 subjects (pooled relative risk 0.36).
This finding suggests that publication bias is not distorting the results of the review.
In light of different comparisons of breastfeeding categories in the available studies (e.g. breastfed
versus non-breastfed; predominantly breastfed versus partially breastfed; exclusively breastfed versus non-breastfed; and many other combinations), the overall pooled results must be interpreted
with due caution. In these pooled results we compared children with greater exposure to breastfeeding against those in a lower exposure category which in some cases included infants who
were not receiving any breastmilk, but in other comparisons included children who were partially, or
non-exclusively breastfed. Because many studies were available on diarrhea morbidity, it was possible to carry out separate meta-analyses for different types of comparisons (Figures 5.4 and 5.6) but
for hospital admissions and mortality this was not possible. Nevertheless, it is reassuring that results
from the different types of comparisons all point out to protective effects of breastfeeding.
Of the 81 comparisons included in this review, only four (4346) showed higher risks associated with
more intense breastfeeding, and these four had confidence intervals that included the unity.
The findings from our review suggest that breastfeeding substantially protects against morbidity/
mortality from diarrhea and that such protection is higher among infants who are exclusively breastfed in the first 6 months of life. The protection afforded by more intense breastfeeding is in the orders
of 8090% for mortality and hospital admissions, and of 50% for morbidity. These results are robust,
being observed in high and low-income settings, and across a number of different diarrhea related
outcomes. Our updated and expanded results are consistent with the conclusions of previous reviews of the literature. The protection afforded by breastfeeding against diarrhea is certainly one of
the most consistent findings in the epidemiological literature on any type of outcome, in the same
category as for example the association between smoking and lung cancer.
16
TABLE 5.1
17
012 months
Randomized trial
(observational analyses)
Case-control
Cross-sectional
Cohort
03 years
012 months
09 months
08 months
012 months
012 months
Randomized trial
(observational analyses)
Cross-sectional
012 months
04 years
012 months
018 months
012 months
012 months
012 months
012 months
012 months
Age at assessment
of the outcome
Cohort
Cohort
Cohort
Cohort
Case-control
Case-control
Cohort
Cohort
Cohort
Study design
Benner, 2011
Not available
Cross-sectional
19992001
2005
20002001
2000 2001
19961997
19961997
Not available
19931995
19811983
19871989
19841986
19841985
1979
19711975
19741976
Year of birth of
subjects
Author, Year
Prevalence
Prevalence
Prevalence
Hospitalization
Prevalence
Hospitalization
Prevalence
Mortality
Prevalence
Incidence
Hospitalization
Mortality
Mortality
Incidence
Incidence
Hospitalization
Outcome
0.67(0.47; 0.94)
Comparison
Breastfeeding and diarrhea outcomes among children < 5 years of age: studies included in the meta-analysis in ascending order of year of publication
and subjects age at which outcome was measured
CHAPTER 5. DIARRHEA
TABLE 5.2
18
19831984
19831984
19831984
19831984
19831984
Not available
19921993
19831984
19831984
19831984
Not available
Not available
19671968
Year of birth of
subjects
Grantham-McGregor (56),
1970
Author, Year
Cross-sectional
Cross-sectional
Cohort
Cohort
Cohort
Cohort
Cohort
Cohort
Case-control
Case-control
Cohort
Cross-sectional
Cohort
Study design
< 5 months
06 months
46 months
46 months
24 months
24 months
02 months
02 months
45 months
23 months
05 months
05 months
04 months
Age at assessment
of the outcome
Prevalence
Prevalence
Prevalence
Prevalence
Prevalence
Prevalence
Prevalence
Prevalence
Hospitalization
Hospitalization
Incidence
Prevalence
Incidence
Outcome
Comparison groups
Breastfeeding and diarrhea outcomes among children < 6 months of age: studies included in the meta-analysis in ascending order of year of publication
and subjects age at which outcome was measured
19
19972000
1997
19992001
20022003
2007
19951997
19951997
1994
Not available
2001
19841991
1988
19951997
Not available
1994
Year of birth of
subjects
Author, Year
Cohort
Case control
Cohort
Cohort
Cross-sectional
Cohort
Cohort
Cohort
Cross-sectional
Cross-sectional
Pooled analysis
Cohort
Cohort
Cohort
Cross-sectional
Study design
06 months
06 months
36 months
05 months
03 months
626 weeks
626 weeks
03 months
6 months
06 months
06 months
06 months
06 months
06 months
05 months
Age at assessment
of the outcome
Hospitalization
Hospitalization
Prevalence
Prevalence
Prevalence
Mortality
Hospitalization
Prevalence
Prevalence
Prevalence
Mortality
Prevalence
Incidence
Incidence
Prevalence
Outcome
Comparison groups
CHAPTER 5. DIARRHEA
TABLE 5.3
Not stated
19871989
19841986
20
19992007
Birth cohort
Pooled analysis
Cross-sectional
Cross-sectional
Cross-sectional
Cohort
Case control
Cohort
Case-control
Case-control
Cohort
Cohort
Cross-sectional
Study design
612 months
611 months
1223 months
611 months
614 months
1235 months
423 months
611 months
811 months
67 months
1236 months
1236 months
611 months
Age at assessment
of the outcome
Mortality
Prevalence
Prevalence
Outcome
Partial breastfed
Not breastfed
Not breastfed
Predominant breastfed
Partial breastfed
Predominant breastfed
Exclusive breastfed
Exposure category
Outcome
Prevalence
Mortality
Prevalence
Prevalence
Predominant breastfed
Reference category
Number of
episodes
Partial breastfed
Comparison groups
Mortality
Prevalence
Incidence
Hospitalization
Hospitalization
TABLE 5.4
19901991
19841991
19901991
19831984
19831984
19931994
19851986
Not stated
19851986
Year of birth
of subjects
Author, Year
Breastfeeding and diarrhea outcomes among children > 6 months of age: studies included in the meta-analysis in ascending order of year of publication
and subjects age at which outcome was measured
FIGURE 5.1
Relative risk and its 95% confidence interval of diarrhea morbidity (prevalence or incidence) in children < 5 years of life comparing breastfeeding
categories in different studies
CHAPTER 5. DIARRHEA
21
Relative risk and its 95% confidence interval of diarrhea morbidity (prevalence or incidence) in children 6 months of life comparing breastfeeding
categories in different studies
FIGURE 5.2
22
Relative risk and its 95% confidence interval of diarrhea morbidity (prevalence or incidence) in children 6 months of age according to age at assessment
of morbidity from diarrhea
FIGURE 5.3
CHAPTER 5. DIARRHEA
23
Relative risk and its 95% confidence interval of diarrhea morbidity (prevalence or incidence) in children 6 months of age according to the type of
comparison between categories of infant feeding
FIGURE 5.4
24
FIGURE 5.5
Relative risk and its 95% confidence interval of diarrhea morbidity (prevalence or incidence) in children > 6 months of life comparing breastfeeding
categories in different studies
CHAPTER 5. DIARRHEA
25
FIGURE 5.6
Relative risk and its 95% confidence interval of diarrhea morbidity (prevalence or incidence) in children > 6 months of age according to the different
comparisons between categories of infant feeding
26
FIGURE 5.7
Relative risk and its 95% confidence interval of hospitalization from diarrhea comparing breastfeeding categories in different studies
CHAPTER 5. DIARRHEA
27
FIGURE 5.8
Relative risk and its 95% confidence interval of mortality from diarrhea comparing breastfeeding categories in different studies
28
FIGURE 5.9
Relative risk and its 95% confidence interval of morbidity from diarrhea comparing children in the breastfeeding intervention group in relation to those
in the control group
CHAPTER 5. DIARRHEA
29
CHAPTER 6
Respiratory infection
In spite of recent progress, respiratory infections continue to be the leading cause of death among children < 5 years, worldwide. In 2010, it was estimated that 1.384 million deaths among children under
5 years were caused by pneumonia (25). Promotion of breastfeeding has been considered as one of
the most cost-effective interventions against such deaths (27). A systematic review concluded in 2009
that breastfeeding also protect infants against respiratory infections in industrialized countries (73).
Biological plausibility
Several mechanisms explaining a possible protective effect of breastfeeding against respiratory infections have been proposed, including the presence in breastmilk of substances with antimicrobial
or immunological properties and the improved general nutritional status of breastfed infants.
Breast milk contains immune cells, antibodies, immune modulators and growth modulators that protect the child against respiratory infection. For example, secretory IgA antibodies may transfer immunity from previously exposed mothers to their children (74). Furthermore, cytokines and growth
factors may be transferred via human milk and stimulate the infants immune system (75).
It has also been suggested that oligosaccharides may inhibit the attachment of pathogens to the
infants mucosa, preventing respiratory infections (75).
Furthermore, in low-income settings breastfeeding can reduce the risk of undernutrition due to repeated infections and use of improper weaning foods (32). Because adequate nutritional status is
essential for non-specific immunity that contributes to fighting infections in general, improved nutrition is a possible mechanism explaining the protective effect of breastfeeding.
Overview of the evidence
The protective effect of breastfeeding against mortality and morbidity from respiratory infections
has been widely studied. In the electronic search, we identified five reviews and/or meta-analyses on
this subject.
In 1984, Kovar et al (77) reviewed the evidence on the association between infant feeding and infant
health, reporting that many of the studies identified had observed a protective effect of breastfeeding
against respiratory infections. The authors did not reach a firm conclusion on the effect of breastfeeding for the following reasons: a) most studies did not adjust their estimates for possible confounding
variables; and b) several of the studies that did not detect statistically significant associations failed
to compare extreme categories of breastfeeding, i.e., exclusively vs. non-breastfed infants. In 1997,
Golding et al (78) reviewed the evidence on the protective effect of breastfeeding against respiratory
and other infections. Six studies on lower respiratory tract infections were identified. In three studies,
the adjusted odds ratio was not statistically significant, whereas among the other three studies the
30
crude results were statistically significant, but no adjusted estimates were provided. For this reason,
the authors concluded that there was no evidence for an association between breastfeeding and
lower respiratory tract infection. It should be noted that these early reviews were severely limited by
the poor quality of most studies available at the time.
With the advent of meta-analytic techniques, the quality of available reviews improved. A first metaanalysis by Bachrach et al (79) in 2003 assessed the relationship between breastfeeding and the risk of
hospitalization for lower respiratory disease among term infants living in high-income settings. Data
from seven cohort studies were pooled, leading to the conclusion that breastfeeding reduced the
risk of hospitalization by 72% [pooled relative risk: 0.28 (95% confidence interval: 0.14; 0.54)].
In 2009, Duijts et al (73) produced an updated systematic review of the effect of breastfeeding on
infections during infancy in industrialized countries. The studies included in the review had to fulfill
at least three of the following internal validity criteria: (a) avoidance of detection bias; (b) adjustment
for possible confounding variables, such as socioeconomic status, size of the family, maternal smoking and maternal schooling; (c) use of clear definition of infant feeding; and (d) having well-defined
outcomes. With respect to respiratory infections, 13 of the 16 included studies reported a protective
effect of breastfeeding. There was no attempt to pool the results of these studies through metaanalysis.
In 2010, McNeil et al (80) also reviewed the evidence on the effect of exclusive breastfeeding on the
risk of hospitalization for lower respiratory tract infection, and six studies were identified. All of these
reported lower risks of hospitalization among exclusively breastfed infant, but in only two studies the
confidence interval did not include the unity. The authors concluded that any formula use was associated with an increased risk of hospitalization. Pooled results were not presented.
In the present systematic review and meta-analysis, we were able to include a substantially larger
number of studies than in any of the above-described reviews and apply modern meta-analytic tools
for obtaining pooled estimates The electronic search identified 36 studies that provided 50 estimates
on the relative risk of morbidity, mortality or hospitalization from respiratory infections according to
infant feeding.
New studies, that appeared since 2007, included research carried out in the United Kingdom (2), Brazil
(81), Bangladesh (37, 46), Philippines (82), Zimbabwe (39), Guinea (40), Canada (83), Hong Kong (84),
and Netherlands (85), Seven of these ten studies reported statistically significant protective effects of
breastfeeding.
The meta-analysis was carried out including the recently published studies cited above, those included in the previous systematic reviews and meta-analyses, and a few other older studies. We identified 18 studies that provided 22 estimates on the effect of breastfeeding on any respiratory infection
outcome for any subgroup of under-five children (Table 6.1), but not restricted to children under 6
months. Table 6.2 shows that 16 studies provided 24 additional estimates on the effect of infant feeding on morbidity or mortality from respiratory diseases among children aged 6 months. Three studies (42, 86, 87) had already been included in a pooled analyses, and we used the results from the latter
instead of the individual study results (1). Table 6.3 shows that we also identified four studies that
evaluated the effect of breastfeeding on respiratory infections among children older than 6 months.
Results of the meta-analyses were separated by outcome, and are presented in Figures 6.16.5. Concerning hospitalization for respiratory infection (respiratory, lower respiratory tract infection or pneumonia), breastfeeding reduced the risk by 57% [pooled relative risk: 0.43 (95% confidence interval:
31
0.33; 0.55)]. Figure 6.1 shows that in contrast to what has been observed for diarrhea outcomes, the
protective effect of breastfeeding was not modified by the age at which children were evaluated. For
example, among the four studies (45, 82, 84, 88) that assessed infants aged < 6 months the pooled
relative risk of hospitalization among breastfed infants was 0.41 (95% confidence interval: 0.25; 0.69),
whereas among the seven studies (47, 54, 81, 8992) that evaluated children younger than 12 months
the pooled effect was 0.42 (95% confidence interval: 0.25; 0.69).
Figure 6.2 shows that those studies that compared breastfed with non-breastfed children reported
the highest protective effect of breastfeeding [pooled relative risk: 0.33 (95% confidence interval:
0.24; 0.46) against hospitalization for respiratory infection. For four other comparisons between categories of breastfeeding (breastfed vs. not breastfed; exclusive breastfed vs. not breastfed, predominant breastfed vs. not breastfed, and partial breastfed vs. not breastfed) we identified two or more
studies. Figure 6.2 shows that the confidence interval of the pooled effect of each one of these comparisons did not include the unity. Therefore, in spite of the different categories being compared,
breastfed infants were less likely to be hospitalized.
With respect to mortality from lower respiratory tract infections, we identified four studies that provided six estimates on the effect of breastfeeding. Figure 6.3 shows that three of the four estimates
were homogeneous, with relative risks ranging from 0.34 to 0.42. On the other hand, Bahl et al (45)
reported a much stronger protective effect of breastfeeding and its confidence interval did not include the estimate from the remaining studies. Breastfeeding reduced the risk of death for respiratory infection by 70% [pooled relative risk: 0.30 (95% confidence interval: 0.16; 0.56)]. Given the small
number of studies that assessed the effect of breastfeeding on mortality and the heterogeneity of
categories of breastfeeding that were compared, we did not stratify the analysis according to categories of breastfeeding.
Figure 6.4 shows the studies that assessed the effect of breastfeeding on morbidity (prevalence or
incidence) from lower respiratory infection. Breastfeeding also reduced the prevalence or incidence
of lower respiratory tract infection [pooled relative risk: 0.68 (95% confidence interval: 0.60; 0.77)].
Similarly to the observed for hospitalization and mortality from lower respiratory tract infections, the
effect of breastfeeding on morbidity was not modified by the age at assessment of morbidity.
Figure 6.5 shows that the effect of breastfeeding on incidence or prevalence of lower respiratory
tract infection does not seems to vary according the types of categories of breastfeeding that were
compared, but these results must be interpreted with caution because several types of comparisons
were adopted only by one or two studies, providing therefore imprecise estimates.
In the literature search, we identified the Belarus PROBIT trial (8), in which maternity hospitals were
randomized to receive or not to receive promotion of exclusive breastfeeding. The proportion of
children who were hospitalized for respiratory infection was similar among the groups [odds ratio:
0.85 (95% confidence interval: 0.57; 1.27)]. As mentioned in the previous section, this analysis did not
compare breastfeeding categories and therefore it could not be incorporated in the meta-analysis.
On the other hand, the Belarus trial also compared the risk of hospitalization for respiratory infection
according to breastfeeding categories, and this result is incorporated in the present meta-analyses
(Table 6.1) (54).
Conclusion
Methodological pitfalls of analyses of breastfeeding and disease were laid out many years ago, yet
few recent studies have taken these into consideration (13). In particular, self-selection of mothers
32
who breastfeed for longer periods of time can bias results of existing studies. This is particularly
problematic in high-income settings, where mothers who breastfeed tend to be more educated and
health-conscious (93). Adjustment for socioeconomic position and maternal education is essential,
but even so residual confounding remains as a possibility.
Another problem is reverse causality, that is, breastfeeding being stopped due to an illness or hospital admission. This poses a special problem in cross-sectional or retrospective studies, which can be
avoided by asking about feeding practices prior to the onset of the episode. Nevertheless, few of the
existing studies seem to have taken this into consideration, and reverse causality bias tends to overestimate the protective effect of breastfeeding against infections. Indeed, for hospitalization from
respiratory infection, the protective effect of breastfeeding was higher among those four studies (82,
89, 94, 95) that avoided the reverse causality bias [pooled relative risk 0.33 (95% confidence interval:
0.23; 0.49)], than among the 13 that did not avoid this bias [pooled relative risk: 0.46 (95% confidence
interval: 0.34; 0.61)].
As discussed in the previous section, confounding is one of the methodological issues that should
be taken into consideration when assessing the short-term consequences of breastfeeding. Similarly
to diarrhea, in low-income countries the benefit of breastfeeding on respiratory outcomes are likely
underestimated by confounding because breastfeeding tends to be more frequent among the poor,
among whom mortality is also higher. In high-income countries, where the rich tend to breastfeed
for longer than the poor, confounding may be expected to operate in the opposite direction. In the
meta-analysis on respiratory infection outcomes among infants 6 months of age, studies from developing countries reported that the protective effect of breastfeeding was similar in studies that
only reported unadjusted results [pooled relative risk: 0.60 (95% confidence interval: 0.36; 1.01) and
studies that adjusted their estimates for socioeconomic and other variables [pooled relative risk 0.50
(95% confidence interval: 0.33; 0.75)], suggesting that confounding is not a likely explanation for the
findings.
Publication bias is another methodological issue that should be taken into consideration in the assessment of evidence from meta-analyses. Funnel plot is one of the strategies used to assess the
susceptibility of the meta-analysis to publication bias. As in the previous meta-analysis for diarrhea
outcomes, we did not generate funnel plots because several different comparisons were done. On the
other hand, we stratified the analyses by sample size and observed that among infants 6 months of
age, the mean effect of breastfeeding on respiratory infection outcomes was similar among studies
with a sample size < 1000 subjects (pooled relative risk 0.59) and those with 1000 subjects (pooled
relative risk 0.56). This finding suggests that publication bias is not distorting the results of the review.
The only randomized trial on hospital admissions due to respiratory infections compared a group of
children born in hospitals with breastfeeding promotion programs, to children born in similar hospitals without such a program. It showed a non-significant reduction of 15% 7. Given that compliance
with breastfeeding promotion in this trial was only partial, these results are not inconsistent with the
levels of protection documented in the present meta-analysis.
Our review suggests that breastfeeding protects against respiratory infection outcomes. Levels of
protection were around 30% for morbidity, about 50% for hospital admissions and about 60% for
mortality, suggesting that breastfeeding affects not only the incidence but also the severity of these
infections. These results are robust, being observed in high and low-income settings, across different
respiratory infections related outcomes, and evident in studies using different definitions of breastfeeding categories.
33
TABLE 6.1
19601977
34
1993
198992
19931995
19961997
19921993
19881990
19831986
19801984
1981
19841985
Cohort
Not
available
Cohort
19741976
Randomized trial
(Observational
analysis)
Cohort
Cohort
Case-control
Cohort
Case-control
Cohort
Cohort
Cohort
Case-control
Cohort
Cohort
19711975
Study design
Year of birth
of subjects
Author, Year
012 months
012 months
012 months
012 months
012 months
024 months
012 months
012 months
018 months
012 months
012 months
024 months
012 months
012 months
Age
Incidence of pneumonia
Comparison
Outcome
Breastfeeding and respiratory infections among children < 5 years of age: studies included in the meta-analysis in ascending order of year of publication
and subjects age at which outcome was measured. Studies restricted to children under 6 months of age are not included
Author, Year
20002002
2005
19961998
20002001
Year of birth
of subjects
Case-control
Cross-sectional
Casecontrol
Cohort
Study design
024 months
09 months
012 months
08 months
Age
Comparison
Outcome
35
TABLE 6.2
36
19972000
19992001
20022003
19951997
19951997
19941996
19841991
1988
19881990
19821983
Not available
19791982
19921993
19581960
Year of birth
of subjects
Author, Year
Cohort
Cohort
Cross-sectional
Cohort
Cohort
Cohort
Pooled analysis
Cohort
Cohort
Cohort
Cross-sectional
Case-control
Casecontrol
Cross-sectional
Study design
36 months
05 months
03 months
626 weeks
626 weeks
05 months
06 months
06 months
06 months
06 months
05 months
06 months
03 months
03 months
Age
Prevalence of respiratory
infection
Prevalence of respiratory
infection
Incidence of pneumonia
Comparison group
Prevalence of pneumonia
Outcome
Breastfeeding and respiratory infections among children < 6 months of age: studies included in the meta-analysis in ascending order of year of
publication and subjects age at which outcome was measured. Studies with an upper limit of age above 6 months are not included
1997
Cohort
Case control
Cohort
Study design
06 months
06 months
6 months
Age
Comparison group
Outcome
Year of birth
of subjects
19861994
19841991
20022006
19992007
Author, Year
37
Cohort
Cohort
Pooled analysis
Cross-sectional
Study design
612 months
712 months
611 months
624 months
Age
Comparison
Prevalence of pneumonia
Outcome
Breastfeeding and respiratory infections among children > 6 months of age: studies included in the meta-analysis in ascending order of subjects age at
which outcome was measured
TABLE 6.3
2007
20022006
Year of birth
of subjects
Author, Year
FIGURE 6.1
Relative risk and its 95% confidence interval of hospitalization due to respiratory infection, lower respiratory tract infection or pneumonia
in children < 5 years of age
38
FIGURE 6.2
Relative risk and its 95% confidence interval of hospitalization for respiratory infection, lower respiratory tract infection or pneumonia
in children < 5 years of age, according to breastfeeding categories
39
FIGURE 6.3
Relative risk and its 95% confidence interval for mortality for respiratory infection or lower respiratory tract infection in children < 5 years of age
40
FIGURE 6.4
Relative risk and its 95% confidence interval of lower respiratory tract infection (prevalence or incidence) comparing breastfeeding
categories in different studies
41
Relative risk and its 95% confidence interval of lower respiratory tract infection (prevalence or incidence), according to breastfeeding categories
FIGURE 6.5
42
CHAPTER 7
Conclusions
The available evidence suggests that breastfeeding protects against diarrhea and respiratory infection in childhood. Because the meta-analyses are almost exclusively based on observational studies
the possibility of self-selection and residual confounding must be taken into consideration.
With respect to confounding, in the present review, we believe that the findings were not susceptible
to residual confounding because a strong protective effect of breastfeeding was observed in lowincome countries. In these countries duration of breastfeeding is higher among the poor. Therefore,
confounding by socioeconomic status should underestimate the protection afforded by breastfeeding.
Reverse causality should also be taken into consideration in the assessment of the evidence on the
short-term effects of breastfeeding. As previously discussed, breastfeeding may have been interrupted or modified by an episode of infectious diseases, thus leading to an association in the opposite direction than the one being postulated (13). For the review on breastfeeding and diarrhea, only
one study explicitly accounted for this bias, by ensuring that information on feeding practices was
obtained for a date previous to the onset of the episode. This study showed substantial protection
associated with breastfeeding (42). For hospitalization from respiratory infection, we identified four
studies that avoided the reverse causality bias and the protective effect of breastfeeding was higher
among these studies (38,89,94,95).
Publication bias is another methodological issue that should be taken into consideration; the selective publication of small positive studies may overestimate the benefit of breastfeeding. In the
present review, we stratified the analyses by sample size and observed that the protective effect of
breastfeeding was not modified by sample size. Suggesting, therefore, that publication bias is not
distorting the results.
Interpretation of results from observational studies may be aided by also taking into consideration
the findings of randomized studies. We identified three randomized trials in which breastfeeding
promotion was related to diarrhea outcomes (79). Diarrhea morbidity was lower in the group receiving the intervention [pooled relative risk: 0.69 (95% confidence interval: 0.49; 0.96)]. This protection
was observed even though the analyses were based on intent to treat, that is, both intervention and
comparison groups included compliers and non-compliers. For respiratory infection, we identified
one randomized trial that compared hospital admissions due to respiratory infections between a
group of children born in hospitals with breastfeeding promotion programs and those who were
born in similar hospitals without such a program. It showed a non-significant reduction of 15% (7).
Given that compliance with breastfeeding promotion in this trial was only partial, these results are
not inconsistent with the levels of protection documented in the present meta-analysis.
43
44
References
1. Effect of breastfeeding on infant and child mortality due to infectious diseases in less developed countries: a pooled
analysis. WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality. Lancet.
2000;355:4515.
2. Quigley MA, Kelly YJ, Sacker A. Breastfeeding and hospitalization for diarrheal and respiratory infection in the United
Kingdom Millennium Cohort Study. Pediatrics. 2007;119:e83742.
3. Kramer MS, Kakuma R. The optimal duration of exclusive breastfeeding: a systematic review. Adv Exp Med Biol.
2004;554:6377.
4. Organization WH. Global strategy for infant and young child feeding. The optimal duration of exclusive breastfeeding.
Geneva: World Health Organization; 2001.
5. Chalmers I. Unbiased, relevant, and reliable assessments in health care: important progress during the past century,
but plenty of scope for doing better. BMJ. 1998;317:11678.
6. Moher D, Hopewell S, Schulz KF, Montori V, Gotzsche PC, Devereaux PJ, et al. [CONSORT 2010 Explanation and Elaboration: updated guidelines for reporting parallel group randomised trials (Chinese version)]. Zhong Xi Yi Jie He Xue
Bao.8:70141.
7.
Kramer MS, Chalmers B, Hodnett ED, Sevkovskaya Z, Dzikovich I, Shapiro S, et al. Promotion of Breastfeeding Intervention Trial (PROBIT): a randomized trial in the Republic of Belarus. JAMA. 2001;285:41320.
8. Morrow AL, Guerrero ML, Shults J, Calva JJ, Lutter C, Bravo J, et al. Efficacy of home-based peer counselling to promote
exclusive breastfeeding: a randomised controlled trial. Lancet. 1999;353:122631.
9. Bhandari N, Bahl R, Mazumdar S, Martines J, Black RE, Bhan MK. Effect of community-based promotion of exclusive
breastfeeding on diarrhoeal illness and growth: a cluster randomised controlled trial. Lancet. 2003;361:141823.
10. Victora CG, Barros FC, Lima RC, Behague DP, Gon alves H, Horta BL, et al. The Pelotas birth cohort study, Rio Grande do
Sul, Brazil, 19822001. Cad Saude Publica. 2003;19:124156.
11. Huttly SR, Barros FC, Victora CG, Beria JU, Vaughan JP. Do mothers overestimate breast feeding duration? An example
of recall bias from a study in southern Brazil. Am J Epidemiol. 1990;132:5725.
12. Horta BL, Kramer MS, Platt RW. Maternal smoking and the risk of early weaning: a meta-analysis. Am J Public Health.
2001;91:3047.
13. Habicht JP, DaVanzo J, Butz WP. Does breastfeeding really save lives, or are apparent benefits due to biases? Am J
Epidemiol. 1986;123:27990.
14. Thompson SG. Why sources of heterogeneity in meta-analysis should be investigated. BMJ. 1994;309:13515.
15. Promislow JH, Gladen BC, Sandler DP. Maternal recall of breastfeeding duration by elderly women. Am J Epidemiol.
2005;161:28996.
16. Kuh D, Ben-Shlomo Y, Lynch J, Hallqvist J, Power C. Life course epidemiology. J Epidemiol Community Health.
2003;57:77883.
17. Rowland MG, Barrell RA, Whitehead RG. Bacterial contamination in traditional Gambian weaning foods. Lancet.
1978;1:1368.
18. Monckeberg F. Factors conditioning malnutrition in Latin America with special reference to Chile. In: Gyorgy P, Kline
O, editors. Malnutrition is a Problem of Ecology. New York: Karger AG; 1970.
19. Juni P, Witschi A, Bloch R, Egger M. The hazards of scoring the quality of clinical trials for meta-analysis. JAMA.
1999;282:105460.
20. Greenland S. Quantitative methods in the review of epidemiologic literature. Epidemiol Rev. 1987;9:130.
21. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7:17788.
45
22. Egger M, Smith GD. Bias in location and selection of studies. BMJ. 1998;316:616.
23. Berkey CS, Hoaglin DC, Mosteller F, Colditz GA. A random-effects regression model for meta-analysis. Stat Med.
1995;14:395411.
24. Greenland S. Invited commentary: a critical look at some popular meta-analytic methods. Am J Epidemiol.
1994;140:2906.
25. Liu L, Johnson HL, Cousens S, Perin J, Scott S, Lawn JE, et al. Global, regional, and national causes of child mortality in
20002010: an updated systematic analysis. Lancet. 2012;379:215161.
26. Feachem RG, Koblinsky MA. Interventions for the control of diarrhoeal diseases among young children: promotion of
breast-feeding. Bull World Health Organ. 1984;62:27191.
27. Jones G, Steketee RW, Black RE, Bhutta ZA, Morris SS. How many child deaths can we prevent this year? Lancet.
2003;362:6571.
28. Newburg DS. Do the binding properties of oligosaccharides in milk protect human infants from gastrointestinal bacteria? J Nutr. 1997;127:980S4S.
29. Nathavitharana KA, Catty D, McNeish AS. IgA antibodies in human milk: epidemiological markers of previous infections? Arch Dis Child Fetal Neonatal Ed. 1994;71:F1927.
30. Hayani KC, Guerrero ML, Morrow AL, Gomez HF, Winsor DK, Ruiz-Palacios GM, et al. Concentration of milk secretory immunoglobulin A against Shigella virulence plasmid-associated antigens as a predictor of symptom status in
Shigella-infected breast-fed infants. J Pediatr. 1992;121:8526.
31. Hashizume S, Kuroda K, Murakami H. Identification of lactoferrin as an essential growth factor for human lymphocytic
cell lines in serum-free medium. Biochim Biophys Acta. 1983;763:37782.
32. Schlaudecker EP, Steinhoff MC, Moore SR. Interactions of diarrhea, pneumonia, and malnutrition in childhood: recent
evidence from developing countries. Curr Opin Infect Dis. 24:496502.
33. Huttly SR, Morris SS, Pisani V. Prevention of diarrhoea in young children in developing countries. Bull World Health
Organ. 1997;75:16374.
34. Golding J, Emmett PM, Rogers IS. Gastroenteritis, diarrhoea and breast feeding. Early Hum Dev. 1997;49 Suppl:S83103.
35. Lamberti LM, Fischer Walker CL, Noiman A, Victora C, Black RE. Breastfeeding and the risk for diarrhea morbidity and
mortality. BMC Public Health. 11 Suppl 3:S15.
36. Ethelberg S, Olesen B, Neimann J, Schiellerup P, Helms M, Jensen C, et al. Risk factors for diarrhea among children in
an industrialized country. Epidemiology. 2006;17:2430.
37. Mihrshahi S, Ichikawa N, Shuaib M, Oddy W, Ampon R, Dibley MJ, et al. Prevalence of exclusive breastfeeding in Bangladesh and its association with diarrhoea and acute respiratory infection: results of the multiple indicator cluster
survey 2003. J Health Popul Nutr. 2007;25:195204.
38. Hengstermann S, Mantaring JB, 3rd, Sobel HL, Borja VE, Basilio J, Iellamo AD, et al. Formula feeding is associated with
increased hospital admissions due to infections among infants younger than 6 months in Manila, Philippines. J Hum
Lact.26:1925.
39. Koyanagi A, Humphrey JH, Moulton LH, Ntozini R, Mutasa K, Iliff P, et al. Effect of early exclusive breastfeeding on
morbidity among infants born to HIV-negative mothers in Zimbabwe. Am J Clin Nutr. 2009;89:137582.
40. Diallo FB, Bell L, Moutquin JM, Garant MP. The effects of exclusive versus non-exclusive breastfeeding on specific
infant morbidities in Conakry. Pan Afr Med J. 2009;2:2.
41. Fisk CM, Crozier SR, Inskip HM, Godfrey KM, Cooper C, Roberts GC, et al. Breastfeeding and reported morbidity during
infancy: findings from the Southampton Womens Survey. Matern Child Nutr.7:6170.
42. Victora CG, Smith PG, Vaughan JP, Nobre LC, Lombardi C, Teixeira AM, et al. Evidence for protection by breast-feeding
against infant deaths from infectious diseases in Brazil. Lancet. 1987;2:31922.
43. Wray JD. Direct nutrition intervention and the control of diarrheal diseases in preschool children. Am J Clin Nutr.
1978;31:207382.
44. Cushing AH, Anderson L. Diarrhea in breast-fed and non-breast-fed infants. Pediatrics. 1982;70:9215.
45. Bahl R, Frost C, Kirkwood BR, Edmond K, Martines J, Bhandari N, et al. Infant feeding patterns and risks of death and
hospitalization in the first half of infancy: multicentre cohort study. Bull World Health Organ. 2005;83:41826.
46. Mihrshahi S, Oddy WH, Peat JK, Kabir I. Association between infant feeding patterns and diarrhoeal and respiratory
illness: a cohort study in Chittagong, Bangladesh. Int Breastfeed J. 2008;3:28.
46
REFERENCES
47. Ellestad-Sayed J, Coodin FJ, Dilling LA, Haworth JC. Breast-feeding protects against infection in Indian infants. Can
Med Assoc J. 1979;120:2958.
48. Cunningham AS. Morbidity in breast-fed and artificially fed infants. II. J Pediatr. 1979;95:6859.
49. Garrido F, Borges G, Cardenas V, Bobadilla J, Ibarra J, Ruiz-Matus C. Mortalidad postneonatal por diarreas: un estudio
de casos y controles. Salud Publica Mex. 1990;32:2618.
50. Chen Y. Relationship between type of infant feeding and hospitalization for gastroenteritis in Shanghai infants. J Hum
Lact. 1994;10:1779.
51. Mondal SK, Gupta PG, Gupta DN, Ghosh S, Sikder SN, Rajendran K, et al. Occurrence of diarrhoeal diseases in relation
to infant feeding practices in a rural community in West Bengal, India. Acta Paediatr. 1996;85:115962.
52. Molbak K, Jensen H, Ingholt L, Aaby P. Risk factors for diarrheal disease incidence in early childhood: a community
cohort study from Guinea-Bissau. Am J Epidemiol. 1997;146:27382.
53. Arifeen S, Black RE, Antelman G, Baqui A, Caulfield L, Becker S. Exclusive breastfeeding reduces acute respiratory infection and diarrhea deaths among infants in Dhaka slums. Pediatrics. 2001;108:E67.
54. Kramer MS, Guo T, Platt RW, Sevkovskaya Z, Dzikovich I, Collet JP, et al. Infant growth and health outcomes associated
with 3 compared with 6 mo of exclusive breastfeeding. Am J Clin Nutr. 2003;78:2915.
55. Vieira GO, Silva LR, de OVT. [Child feeding and diarrhea morbidity]. J Pediatr (Rio J). 2003;79:44954.
56. Grantham-McGregor SM, Back EH. Breast feeding in Kingston, Jamaica. Arch Dis Child. 1970;45:4049.
57. Brown KH, Black RE, Lopez de Romana G, Creed de Kanashiro H. Infant-feeding practices and their relationship with
diarrheal and other diseases in Huascar (Lima), Peru. Pediatrics. 1989;83:3140.
58. Mahmood DA, Feachem RG, Huttly SR. Infant feeding and risk of severe diarrhoea in Basrah city, Iraq: a case-control
study. Bull World Health Organ. 1989;67:7016.
59. Popkin BM, Adair L, Akin JS, Black R, Briscoe J, Flieger W. Breast-feeding and diarrheal morbidity. Pediatrics. 1990;86:874
82.
60. Ketsela T, Asfaw M, Kebede D. Patterns of breast feeding in western Ethiopia and their relationship to acute diarrhoea
in infants. J Trop Pediatr. 1990;36:1803.
61. Brito Hernandez ML, Barreras Meneses K, Quintero Fleitas F, Lopez Callejo Hiorj-Lorenzen M. [The effect of breast
feeding and its duration on the health of the infant (December 1992February 1993)]. Rev Cubana Enferm. 1995;11:14
24.
62. al-Ali FM, Hossain MM, Pugh RN. The associations between feeding modes and diarrhoea among urban children in a
newly developed country. Public Health. 1997;111:23943.
63. Lopez-Alarcon M, Villalpando S, Fajardo A. Breast-feeding lowers the frequency and duration of acute respiratory
infection and diarrhea in infants under six months of age. J Nutr. 1997;127:43643.
64. Clemens J, Elyazeed RA, Rao M, Savarino S, Morsy BZ, Kim Y, et al. Early initiation of breastfeeding and the risk of infant
diarrhea in rural Egypt. Pediatrics. 1999;104:e3.
65. Raisler J, Alexander C, OCampo P. Breast-feeding and infant illness: a dose-response relationship? Am J Public Health.
1999;89:2530.
66. Khadivzadeh T, Parsai S. Effect of exclusive breastfeeding and complementary feeding on infant growth and morbidity. East Mediterr Health J. 2004;10:28994.
67. Macias-Carrillo C, Franco-Marina F, Long-Dunlap K, Hernandez-Gaytan SI, Martinez-Lopez Y, Lopez-Cervantes M.
[Breast feeding and the incidence of acute diarrhea during the first three months of life]. Salud Publica Mex. 2005;47:49
57.
68. Tarrant M, Kwok MK, Lam TH, Leung GM, Schooling CM. Breast-feeding and childhood hospitalizations for infections.
Epidemiology.21:84754.
69. Briend A, Wojtyniak B, Rowland MG. Breast feeding, nutritional state, and child survival in rural Bangladesh. Br Med J
(Clin Res Ed). 1988;296:87982.
70. Knight SM, Toodayan W, Caique WC, Kyi W, Barnes A, Desmarchelier P. Risk factors for the transmission of diarrhoea in
children: a case-control study in rural Malaysia. Int J Epidemiol. 1992;21:8128.
71. Molbak K, Gottschau A, Aaby P, Hojlyng N, Ingholt L, da Silva AP. Prolonged breast feeding, diarrhoeal disease, and
survival of children in Guinea-Bissau. BMJ. 1994;308:14036.
72. Mulder-Sibanda M, Sibanda-Mulder FS. Prolonged breastfeeding in Bangladesh: indicators of inadequate feeding
practices or mothers response to childrens poor health? Public Health. 1999;113:658.
47
73. Duijts L, Ramadhani MK, Moll HA. Breastfeeding protects against infectious diseases during infancy in industrialized
countries. A systematic review. Matern Child Nutr. 2009;5:199210.
74. Tamura S, Funato H, Hirabayashi Y, Suzuki Y, Nagamine T, Aizawa C, et al. Cross-protection against influenza A virus infection by passively transferred respiratory tract IgA antibodies to different hemagglutinin molecules. Eur J Immunol.
1991;21:133744.
75. Hanson LA, Korotkova M, Haversen L, Mattsby-Baltzer I, Hahn-Zoric M, Silfverdal SA, et al. Breast-feeding, a complex
support system for the offspring. Pediatr Int. 2002;44:34752.
76. Schlaudecker EP, Steinhoff MC, Moore SR. Interactions of diarrhea, pneumonia, and malnutrition in childhood: recent
evidence from developing countries. Curr Opin Infect Dis. 2011;24:496502.
77. Kovar MG, Serdula MK, Marks JS, Fraser DW. Review of the epidemiologic evidence for an association between infant
feeding and infant health. Pediatrics. 1984;74:61538.
78. Golding J, Emmett PM, Rogers IS. Does breast feeding protect against non-gastric infections? Early Hum Dev. 1997;49
Suppl:S10520.
79. Bachrach VR, Schwarz E, Bachrach LR. Breastfeeding and the risk of hospitalization for respiratory disease in infancy:
a meta-analysis. Arch Pediatr Adolesc Med. 2003;157:23743.
80. McNiel ME, Labbok MH, Abrahams SW. What are the risks associated with formula feeding? A re-analysis and review.
Birth. 2010;37:508.
81. Macedo SE, Menezes AM, Albernaz E, Post P, Knorst M. [Risk factors for acute respiratory disease hospitalization in
children under one year of age]. Rev Saude Publica. 2007;41:3518.
82. Hengstermann S, Mantaring JB, 3rd, Sobel HL, Borja VE, Basilio J, Iellamo AD, et al. Formula feeding is associated with
increased hospital admissions due to infections among infants younger than 6 months in Manila, Philippines. J Hum
Lact. 2010;26:1925.
83. Banerji A, Greenberg D, White LF, Macdonald WA, Saxton A, Thomas E, et al. Risk factors and viruses associated with
hospitalization due to lower respiratory tract infections in Canadian Inuit children : a case-control study. Pediatr Infect
Dis J. 2009;28:697701.
84. Tarrant M, Kwok MK, Lam TH, Leung GM, Schooling CM. Breast-feeding and childhood hospitalizations for infections.
Epidemiology. 2010;21:84754.
85. Duijts L, Jaddoe VW, Hofman A, Moll HA. Prolonged and exclusive breastfeeding reduces the risk of infectious diseases in infancy. Pediatrics. 2010;126:e1825.
86. Yoon PW, Black RE, Moulton LH, Becker S. Effect of not breastfeeding on the risk of diarrheal and respiratory mortality
in children under 2 years of age in Metro Cebu, The Philippines. Am J Epidemiol. 1996;143:11428.
87. Hanson LA, Ashraf R, Zaman S, Karlberg J, Lindblad BS, Jalil F. Breast feeding is a natural contraceptive and prevents
disease and death in infants, linking infant mortality and birth rates. Acta Paediatr. 1994;83:36.
88. Pisacane A, Graziano L, Zona G, Granata G, Dolezalova H, Cafiero M, et al. Breast feeding and acute lower respiratory
infection. Acta Paediatr. 1994;83:7148.
89. Cesar JA, Victora CG, Barros FC, Santos IS, Flores JA. Impact of breast feeding on admission for pneumonia during
postneonatal period in Brazil: nested case-control study. BMJ. 1999;318:131620.
90. Oddy WH, Holt PG, Sly PD, Read AW, Landau LI, Stanley FJ, et al. Association between breast feeding and asthma in 6
year old children: findings of a prospective birth cohort study. BMJ. 1999;319:8159.
91. Howie PW, Forsyth JS, Ogston SA, Clark A, Florey CD. Protective effect of breast feeding against infection. BMJ.
1990;300:116.
92. Nafstad P, Jaakkola JJ, Hagen JA, Botten G, Kongerud J. Breastfeeding, maternal smoking and lower respiratory tract
infections. Eur Respir J. 1996;9:26239.
93. Der G, Batty GD, Deary IJ. Effect of breast feeding on intelligence in children: prospective study, sibling pairs analysis,
and meta-analysis. BMJ. 2006;333:945.
94. Leventhal JM, Shapiro ED, Aten CB, Berg AT, Egerter SA. Does breast-feeding protect against infections in infants less
than 3 months of age? Pediatrics. 1986;78:896903.
95. Victora CG, Fuchs SC, Flores JA, Fonseca W, Kirkwood B. Risk factors for pneumonia among children in a Brazilian
metropolitan area. Pediatrics. 1994;93:97785.
96. Fergusson DM, Horwood LJ, Shannon FT, Taylor B. Breast-feeding, gastrointestinal and lower respiratory illness in the
first two years. Aust Paediatr J. 1981;17:1915.
48
REFERENCES
97. Forman MR, Graubard BI, Hoffman HJ, Beren R, Harley EE, Bennett P. The Pima infant feeding study: breastfeeding and
respiratory infections during the first year of life. Int J Epidemiol. 1984;13:44753.
98. Chen Y, Yu SZ, Li WX. Artificial feeding and hospitalization in the first 18 months of life. Pediatrics. 1988;81:5862.
99. Wright AL, Holberg CJ, Martinez FD, Morgan WJ, Taussig LM. Breast feeding and lower respiratory tract illness in the
first year of life. Group Health Medical Associates. BMJ. 1989;299:9469.
100. French JG. Relationship of morbidity to the feeding patterns of Navajo children from birth through twenty-four
months. Am J Clin Nutr. 1967;20:37585.
101. Beaudry M, Dufour R, Marcoux S. Relation between infant feeding and infections during the first six months of life. J
Pediatr. 1995;126:1917.
102. Cushing AH, Samet JM, Lambert WE, Skipper BJ, Hunt WC, Young SA, et al. Breastfeeding reduces risk of respiratory
illness in infants. Am J Epidemiol. 1998;147:86370.
103. Koch A, Molbak K, Homoe P, Sorensen P, Hjuler T, Olesen ME, et al. Risk factors for acute respiratory tract infections in
young Greenlandic children. Am J Epidemiol. 2003;158:37484.
104. Chantry CJ, Howard CR, Auinger P. Full breastfeeding duration and associated decrease in respiratory tract infection
in US children. Pediatrics. 2006;117:42532.
105. Fisk CM, Crozier SR, Inskip HM, Godfrey KM, Cooper C, Roberts GC, et al. Breastfeeding and reported morbidity during
infancy: findings from the Southampton Womens Survey. Matern Child Nutr. 2011;7:6170.
49