We investigated whether delivery by C/S is associated with asthma and atopy outcomes at age 7–9 years and assessed whether this relationship differs in children with and without a family history of allergic diseases. After adjusting for potential confounders, positive associations of C/S delivery with ever wheeze, ever asthma diagnosis and atopic sensitization were observed. While no effect modification by family history of atopy was observed in terms of the asthma outcomes, there was some evidence that the association of C/S with atopic sensitization was particularly pronounced in the presence of family history of allergies. In fact, among those without family history of allergies, atopic sensitization was not found to be associated with C/S.
Despite the cross sectional design of the study, the temporal sequence of the exposure and outcomes in this study is clear. In addition the self – reporting nature of the mode of delivery is unlikely to have lead to misclassification of the exposure as parents usually report accurately the mode of delivery of their child. Of course, information bias in terms of self reporting asthma outcomes as well as other covariates (e.g. family history of allergies or birth weight) cannot be excluded. Nevertheless, there is no reason to believe that any such bias should occurred differentially with regards to the mode of delivery. Even though the participation rate in the questionnaire survey was rather low, the final sample corresponds to about half of all children in this age range. While it is not unlikely that asthmatic children may have been more inclined to participate in the study and may have been overrepresented in the sample (and as a result inflate the prevalence estimates), there is no reason to believe that the participation rate could have affected the study outcomes differentially in terms of mode of delivery (i.e. there is no reason why more asthmatics born by C/S or more non-asthmatics born vaginally would have participated in the study). Another limitation of our study is the lack of available information on some potential confounding factors such as transient tachypnea of the newborn
, pregnancy complications (e.g. prematurity and intrauterine growth retardation -IUGR) and delivery complications. However, in view of the fact that many of these variables as well as birth weight are strongly correlated
, it is not uncommon for studies to avoid adjusting for all, especially if adding them to the model does not seem to affect the findings
. In this study, we used low birth weight as a surrogate parameter primarily for both prematurity and IUGR and all analyses were repeated after exclusion of children with a low birth weight. Nevertheless, we adjusted in our model for other important confounding factors such as maternal smoking during pregnancy and exclusive breastfeeding for longer than two months and parental educational level
. Finally it should be noted that we only investigated atopic sensitization using inhalant allergens and have not included food allergens in our assessment although summary of evidence so far has shown the association of C/S delivery with atopic sensitization to be stronger in the case of food allergens than inhalant allergens
 and hence this association would have probably been stronger should we have included food allergens in our panel of antigens in SPT’s.
Association between delivery by C/S and asthma
A number of epidemiological studies have investigated the risk of having asthma in individuals born by C/S. Two recent meta-analyses have reported a modest increase in the risk of asthma in children born by C/S
[20, 26]. In this study, we have also shown associations between delivery by C/S and all asthma outcomes investigated. Specifically, in the case of ever wheeze and ever asthma, the effect estimates did not attenuate after adjusting for confounders and the observed associations remained statistically significant.
Due to unavailability of data, it was not possible to investigate the association of CS with asthma outcomes differentially in children born by elective and emergency CS. Likewise, it would have been preferable to adjust the observed association for prematurity; nevertheless in our study inferences remain unaffected both when adjusting for birth weight (as a surrogate for prematurity) as well as when the analyses are restricted to those with a birth weight higher than 2.5 Kg. Furthermore, prematurity in the literature has been associated with increased risk of asthmatic symptoms but not with the development of allergies
 and thus is unlikely to confound the observed association of C/S with atopic sensitization.
In both meta analyses, the authors commented on the inconsistency of findings in the literature and suggested that to some extent the observed heterogeneity may be attributed to differences in methodology, study designs and the populations under study (adults vs children), use of different asthma outcome definitions, the effect of bias and lack of appropriate adjustment for confounders, limited statistical power due to small sample size etc. It is also likely that the association of delivery by C/S and asthma differs in atopic and non atopic children and, therefore, by not distinguishing on the basis of atopic background effect estimates may be reflecting a mixed picture.
Indeed, the study by Roduit et al.
 found that the association of C/S and asthma was more pronounced in children with one or two allergic parents, although no test for effect modification was provided. In contrast, our analysis did not reveal any difference in the estimates of the effect of C/S delivery on any of the asthma outcomes in children with and without family history of allergies indicating probably that the positive association of CS delivery with asthma outcomes found in our study is important in both children with and without family history of allergies. On the contrary, the findings of the longitudinal study by Pistiner et al. did not find any evidence of a positive association of C/S delivery and asthma disease or symptom in 432 nine year old children with parental history of atopy. Therefore, the effect of genetic predisposition on the association of asthma with C/S delivery remains unclear and further research is needed to address this potential interaction.
Association between delivery by C/S and atopic sensitization
Nine studies to date, all prospective in nature, have investigated the potential association of atopic sensitization with C/S delivery using objective methods i.e. food challenge
, IgE specific levels
[9, 23, 29] and SPT
[1, 3, 15, 16, 30] to food and/or inhalant allergens. While some of these studies have a small sample size
 or have tested for only a small number of allergens
, the majority could not replicate an association between delivery by C/S and atopic sensitization
[6, 8, 9, 23, 31, 32]. A recent review
 as well as one of the meta-analyses
 could not verify this association and concluded that, if anything, delivery by C/S might be associated with food but not with inhalant atopy.
Four of these studies examined the potentially modifying effect of family history of atopy or, alternatively, restricted their investigations to children born to atopic parents
[8, 21, 23, 28] with mixed findings. Maitra et al. found that the relationship between C/S delivery and atopy was not differentially affected by maternal history of hayfever amongst 5,916 seven year-old children in England
. However, a degree of misclassification cannot be excluded since only maternal history of hayfever was considered and neither asthma nor eczema were included in the definition used. Roduit et al. showed no significant association of C/S with atopy
. Even though, as mentioned earlier, the authors reported that the association between C/S and asthma was more pronounced amongst children of allergic parents, they found the opposite to be true for atopy with a more pronounced effect in children of non-allergic parents. Nevertheless, it is not known to what extent the particularly low rates of C/S in this population together with the differential non-participation rate observed in this study (with higher loss to follow up amongst the subjects of the atopic background group) may have possibly affected the direction of estimates.
In contrast, Pistiner et al. showed that C/S was associated with a two-fold increase in the odds of atopy (OR 2.1, CI 1.1-3.9, p=0.02) in children with parental history of atopy
. In addition, the study by Eggesbo et al. showed that the risk of having a positive food challenge to an egg was 7 times higher in those born by C/S to allergic mothers compared to those born vaginally to non allergic mothers amongst 2803 two year-old Norwegian children
. In agreement with these findings in this study, we also found evidence of an overall association between C/S delivery and atopy. In addition, we showed a modest evidence to suggest that this association might not be independent of family history of allergies as children with a family history of atopy have 2.6 times the odds of having atopic sensitization at age 8–9 years if born by C/S while no similar association was observed in children with no family history of allergies.
The differential effect of atopic background has also been studied in relation to other early life environmental exposures and asthma development
. Indeed, maternal overweight has been shown to be associated with an increased risk of asthma in children predisposed to asthma at the age of 8 years whilst no association was observed in children without predisposition. The differential effect of atopic background might be suggestive of gene-environment interaction that can influence the development of the immune system. It has been previously shown that atopic and non-atopic children develop different T-helper lymphocyte responses during the first year of life
. More recently, a study investigating the interaction of day care exposure in the first year of life with 72 genetic polymorphisms at 45 candidate loci demonstrated 22 significant interactions of 16 polymorphisms with day care attendance on cytokine profiles and atopic phenotypes
. Furthermore, there is growing evidence that gene-environment interactions are also involved in the development of childhood food allergy
. Therefore, the interaction of genetic predisposition with early life environmental exposures, such as in the case of delivery by C/S, might play an important role in the development of allergic and probably other immune related conditions; this will need to be explored further in future studies.