Increased risk of allergic rhinitis among children delivered by cesarean section: a cross-sectional study nested in a birth cohort

The prevalence of asthma and chronic rhinitis are high in Brazil [1, 2]. The average prevalence rate of symptoms of asthma in the age range 6 and 7 years is 24 % and for chronic rhinitis, 25.7 % [3]. In Feira de Santana, State of Bahia, the prevalence of symptoms of chronic rhinitis and allergic rhinitis was 33 and 17 %, respectively, in the year of the 2003, among the highest in Brazil [2, 3]. The prevalence rate of asthma physician-diagnosed among adolescents is 12.4 and 6.4 % in children 6 to 7 years old [3, 4].

The frequency of birth by cesarean section (CS) is growing fast in the Brazilian public and private health services. The rate of CS among all births was 33.8 % in 2004 and increased to 45.9 % in 2008 [5], being considerated the highest in the world. The World Health Organization recommends a rate of CS of less than 15 %. However, in the majority of countries the actual rate exceeds this recommendation [6].

Several studies have demonstrated an association between symptoms of asthma and allergic rhinitis with CS, but there are controversial results reports [7]. Some cohort studies demonstrated an association between CS and asthma in children, with odds ratio (OR) ranging from 1.09 to 1.82 [8, 9]; whereas other studies have found no association [10, 11]. In Brazil, the study of Menezes et al. evaluated the association between CS and asthma in children in two age ranges and did not find a significant association [12]. A meta-analysis has favored a positive association with an estimate increased risk of 20 % for having asthma among children born by CS (Thavagnanam et al., 2008) [13].

There are a few studies that have evaluated the association between allergic rhinitis, asthma and birth by CS in children with genetic predisposition, and in those that have demonstrated a positive association, the OR ranged between 1.37 and 1.80 [14, 15]. Nevertheless, a meta-analysis have not demonstrated a significant increase in the risk of allergic rhinitis in children born by CS (Bager et al., 2003) [16].

The aim of the present study was to investigate whether children born by CS have a higher chance of developing asthma and allergic rhinitis at 6 years of age, looking for other potential risk factors in a birth cohort and evaluating family history of asthma as a possible effect modifier.

In this study 672 mother-child dyads (98.2 %) of the 684 being currently followed-up in the cohort were included in analysis. The 12 remaining dyads were not identified because of recent change of their home addresses.

The rate of CS was 48.3 % and there was no significant differences between the two groups divided by type of birth in relation to birth weight, gender, gestational age, pneumonia ever, bronchiolitis in the 1st year of life and maternal smoking during gestation. However, different mother’s ages at the time of birth, family income, mother’s schooling up to 8 years of study, number of people who sleep in the room with the child, attendance of a nursery to the age of 2 years and institution where the birth took place (private/public) were different when comparing children’s birth by CS versus vaginal delivery (Table 1).

The overall prevalence of asthma symptoms was 13.8 %, being 14.8 % and 13 % among subjects born by CS and vaginal delivery respectively. There was no statistically significant difference in the prevalence of symptoms of active asthma or severe asthma episodes between the children born by CS or vaginal delivery.

The prevalence rate of symptoms related to chronic rhinitis was 36.9 %, and of allergic rhinitis, 23.5 %; The prevalence of symptoms of chronic and allergic rhinitis was higher in children born by CS when compared with those born by vaginal delivery (Table 2).

Active and severe asthma symptoms were not associated with delivery by CS, even after the adjustment for the following covariates: gestational age, birth weight, breastfeeding up to the 4th month, mother’s smoking during gestation, family income, mother's educational level, number of siblings, number of persons sleeping in the same room as the child, attendance to nursery before the age of 2 years and physician-diagnosed pneumonia ever. The frequency of CS section and vaginal delivery in children with active asthma was 51.6 % versus 48.4 % respectively (adjusted OR 1.16; 95%IC 0.75–1.80). Chronic and allergic rhinitis were associated with delivery by CS in the bivariate analysis as well as in the multivariate logistic regression adjusted for the same covariates as for asthma (Table 2).

Table 3 demonstrates the frequency of asthma in children delivered by CS and vaginal route with and without parental history of asthma. In children born by CS, the parental history of asthma was associated with asthma (OR = 1,97; 95 % CI 1.04–3.74) whereas in children born by vaginal delivery, the parental history of asthma was also associated with asthma in the child (OR = 2.05; 95 % CI 1.07–3.94). There was also an association between birth by CS and symptoms of chronic rhinitis and allergic rhinitis only in children with a parental history of asthma, after adjusting for all covariates, as demonstrated in Table 3.

Therefore the effect modifier of parental history of asthma was demonstrated in children born by CS, as they were significantly more likely to develop allergic rhinitis and chronic rhinitis, when exposed to the two variables combined (p value for Likelihood Ratio Test for effect modification = 0.02 and 0.10, respectively) (Table 3).

This study demonstrated a positive association between CS, chronic rhinitis and allergic rhinitis, after adjusting for potential confounders, but the association was significant only in children at higher risk of chronic rhinitis and allergic rhinitis as indicated by parental history of asthma. CS was not significantly associated with symptoms of asthma at age of 6 years in our study population. A cross-sectional study by Kolokrotoni et al. [17] found a positive association between CS, asthma and atopic sensitization in children aged 8–9 years. When children were stratified by family history of allergy, the association between CS and atopic sensitization was more pronounced but this was not the case for asthma outcomes. While more than a two-fold increase in the odds of being atopic was observed in children with a family history of allergy if born by CS, no association was observed in children without a family history of allergies.

The diagnosis criteria of active asthma considered in the present study was based on symptoms of the disease in the 12 months before the application of the ISAAC questionnaire, ruling out the cases of transient wheezing at an earlier age.

The possibility of selection bias doesn’t seem to be relevant in the present study, because children born by CS did not differ significantly from those born by vaginal delivery with regard to the main confounding variables (gestational age, birth weight, time of exclusive breastfeeding and mother smoking during gestation). Moreover, there were not differences between the proportions of exclusions due to complications during gestation or related to birth, such as differences in proportions of newborns who presented perinatal respiratory complications. The absence of association between CS and asthma may be explained because the majority of cases of symptoms of asthma related symptoms reported in urban centers in Brazil have not been attributed to atopy, as demonstrated in a study in children conducted in Salvador-Bahia, just 100 km from Feira de Santana, where the present study was conducted, by Cunha et al. [18].

Previous studies that assessed the association between CS and asthma presented different results. Roduit et al. [19], in a prospective cohort study of 2.917 children, found a significant association between delivery by CS and asthma (OR = 1.77) and the risk was increased in the presence of parental history of atopy. In a prospective cohort study of 12,367 children, Maitra et al. [20] found no association between delivery by CS and asthma (OR = 1.14; 95%CI 0.9–1.4). In two populations included in a cohort study conducted in Southern of Brazil by Menezes et al. [12], no significant association was found between the type of birth and dyspnea at 4 years of age (OR = 0.96), nor from 4 to 11 years (OR = 1.18) neither at 15 years (OR = 1.02). A meta-analysis conducted by Bager et al. [16], found a significant increase in the risk for asthma in children delivered by CS with (OR = 1.18;95%CI1.05–1.32) and risk for allergic rhinitis with (OR = 1.24;95%CI 1.08–1.43) [17]. In another meta-analysis, Thavagnanam et al. also found an increased risk for asthma in children delivered by CS (OR = 1.22; 95%CI 1.14–1.29) [13]. However, in these analyses there was high heterogeneity among the studies. They differed in many issues such as definition of asthma, age of the children when the diagnosis of asthma was made, in the study design, evaluation of environmental factors, time duration of exclusive breastfeeding and the use previous of BCG vaccination [21, 22].

The significant association between delivery by CS, chronic and allergic rhinitis observed in the present study was also identified in the studies carried out by Pistiner e et al. (OR = 1.80); Renz-Polster e et al. (OR = 1.37); Salam e et al. (OR = 1.57) [8, 14, 15]. Other studies found no association between delivery by CS and allergic rhinitis: Park e et al. (OR = 1.14); Xu e et al. (OR = 1.28); Mallen e et al. (OR = 2.82) [10, 23, 24]. Parental history of atopy is considered an important risk factor for the development of asthma and atopic diseases [25], and is a potential effect modifier in the evaluation of the association between delivery by CS, asthma, allergic rhinitis and atopic sensitization [14, 15, 17]. In our study, CS was associated allergic rhinitis and chronic rhinitis only in children with parental history of asthma, which suggests that the association between CS and allergic rhinitis and CS and chronic rhinitis could be dependent on the genetic predisposition. Our finding is consistent with the results of study of Pistiner e et al. which found increased risk of allergic rhinitis among children with a parental history of atopy but did not find any evidence of a positive association between CS and asthma in children with parental history of atopy.

The controversial reports on the association between allergic rhinitis and delivery by CS may be due to the different diagnostic criteria considered for the definition of allergic rhinitis, or the presence of confounding factors, such as maternal smoking during gestation, mother's history of asthma, and primiparity [26]. In the present study, the inclusion of the main confounding variables in the multivariate analysis of the association between delivery by CS and allergic rhinitis might have contributed to a greater discrimination of the findings.

In the present study the birth rate by CS was higher in children born in private hospital, reproducing the current situation of the type delivery that occurs in Brazil [5]. Therefore differences between children born by CS and vaginal delivery may be related to socioeconomic differences between children born in public health services versus private hospitals.

The possible biological mechanism that may explain the direct association between delivery by CS and allergic rhinitis could be related to the hygiene hypothesis: the characteristics of the intestinal microbiota in children born by CS is different from that found among children born by vaginal delivery [27]. The intestinal microflora of children born by vaginal delivery is composed of bacteria from the vaginal flora by exposure while in the birth canal, whereas the intestinal biome of children born by CS is composed of bacteria from colonization of the skin and those related to the hospital environment [28]. According to the hygiene hypothesis, there would be a predominance of the immunologic response of the Th2 type in children born by CS, due to the lack of physiological modulation of the immune system conferred by the natural biome. That condition entails the later development of atopic asthma, allergic rhinitis and other atopic diseases [29]. Children born by CS present elevated levels of IL-13 and interferon γ in the umbilical cord and lower levels of IL-10, a regulatory cytokine [30].

The limitations of our study could be related to the lack of a multilevel approach analysis given the 10 institutions where participants were selection; the study design not robust enough to infer causation, as it did not allow the collection of biomarkers of atopy (skin prick tests or specific IgE), which could have reinforced the causality association by bringing in mechanistic evidence; we had no stool cultures to examine the type of intestinal microflora present in the children born by different methods, which could also bring further evidence to the theoretical explanation; possible misclassification due to self-reported of outcomes, as to parental history of asthma related to mothers; lack of classification of the CS type (emergency and elective), considering that babies born through emergency CS might be at less risk of allergies due to the exposure of the children to vaginal flora during labor. Despite these limitations, the results were similar to previous ones carried out in different settings and with different populations, and provide further evidence to fill the gap of knowledge on the association between, chronic and allergic rhinitis in children with parental history of asthma.

Our results indicate that elective CS increases significantly the risk of developing chronic rhinitis and allergic rhinitis, chronic diseases that imply in recurring discomfort and impaired quality of life among children with a family predisposition for asthma.

In conclusion, birth by CS was associated with both chronic rhinitis and allergic rhinitis in children with parental history of asthma, in this cross sectional study nested in a birth cohort.