Analysis of maternal and perinatal determinants of allergic sensitization in childhood
Allergy, Asthma & Clinical Immunology volume 16, Article number: 71 (2020)
Non-communicable diseases, such as allergies, are influenced by both genetic and epigenetic factors. Perinatal determinants conceivably modify the epigenetic makeup of the developing fetal immune system preventing or predisposing the development of allergies. The aim of this study therefore was to identify independent perinatal factors associated with allergic sensitization in childhood.
In a single center retrospective case-cohort study electronic obstetric medical records and available skin prick testing results of children were analyzed. For the analysis 286 skin prick test positive (sensitized) children [median (IQR): 3.47 (1.70–7.34) years] were compared with data from all remaining live births in the obstetric cohort (n = 66,583).
Sensitized children more frequently had a mother born in Asia (19.1% vs. 10.2%; P < 10–6). Applying backward elimination logistic regression, seven out of 23 initially entered perinatal factors remained in the model. High maternal age (> 35 years; OR: 1.912; P < 0.001), male offspring sex (OR: 1.423; P < 0.01) and assisted conception (OR: 1.771; P < 0.05) increased the risk for allergic sensitization. In contrast, maternal smoking (OR: 0.469; P < 0.005), increasing parity (OR: 0.881; P < 0.05), maternal pre-pregnancy overweight (OR: 0.742; P < 0.005) and preterm birth (OR: 0.484; P < 0.05) decreased the risk for allergic sensitization.
In addition to supporting previous findings, this study is first to report an increased risk of allergic sensitization after assisted conception. Beyond this finding’s potential implementation in preventative strategies, exploration of this association could further pathophysiological understanding of allergic disease.
Affecting around 30–40% of the world’s population, allergies are the most common and earliest-onset non-communicable diseases . With a remarkable and rapid increase in prevalence especially in industrialized countries, their prevention and treatment have become a public health priority .
Despite increasing interest, the pathogenesis of allergic disease remains unclear. Even though the importance of a hereditary component is well established, there is considerable evidence that the effect of family history partly reflects shared environmental risk factors . Epidemiological studies emphasize the importance of early environmental factors . In utero adaption to the maternal environment by means of epigenetic modifications are considered to have evolved as effective survival strategies preparing the fetus for postnatal life. However, fetal adaptation comes with the risk of a mismatch between the predicted and the actual encountered environment and is assumed to contribute to the development of allergic disease [5, 6].
This study aims to investigate the effect of several maternal and perinatal factors and their association with allergic sensitization in a pediatric population.
Materials and methods
Study population and design
This is a retrospective study linking skin prick test (SPT) data with hospital records contained in the state-wide mandatory electronic obstetric database (Obstetrix™)  from a metropolitan teaching hospital in Western Sydney, Australia. SPTs conducted on children attending Nepean Hospital between January 2010 and December 2017 were reviewed. 286 atopic children with a positive SPT were linked with Obstetrix™ ranging from 2000 to 2017 (n = 67,268), using medical record numbers. After exclusion of stillbirths (n = 399) from Obstetrix™, maternal and perinatal characteristics of SPT positive children (n = 286) were compared with the remaining children who did not undergo allergy assessment (n = 66,583), as shown in Fig. 1. The rate of atopy in the remainder of the obstetric cohort was assumed to be either similar, or lower than the average rate of atopy in Australian children (approximately 20%) . For more information regarding data collection see Supplementary Information.
Skin prick testing
The allergens tested were selected following specialist consultation for suspected allergic disease. Standard commercial food and aeroallergen extracts were used as recommended by manufacturers (Additional file 1: Table S1). Skin prick testing and interpretation was carried out following Australasian Society of Clinical Immunology and Allergy (ASCIA) recommendations .
As birth weight, length and head circumference (HC) are naturally confounded by gestational age and sex, z-scores were used for further analysis. Multiple births and stillbirths were excluded from the data, leaving 64,204 entries. Entries beyond three standard deviations were regarded entry errors and excluded from further calculation. The LMS parameters described by Cole  were then calculated based on 32,460 weight, 30,666 length and 31,526 HC entries for boys and 30,171 weight, 27,671 length and 29,480 HC entries for girls. No z-scores were calculated for entries < 26 weeks gestation (n = 301) and > 42 weeks gestation (n = 140).
All statistical analyses were performed using SPSS statistics (version 24, IBM Corporation, Chicago, Illinois, USA). Descriptive statistics are reported as means and standard deviation (mean ± SD) for normally distributed data, non-normally distributed data are reported as median and interquartile ranges [median (IQR)]. Normality was assessed using histograms. Descriptive statistics of categorical variables were reported as frequency and percentage, or only percentage. Continuous outcomes were compared between groups using t-tests or Mann–Whitney U-tests. Categorical outcomes were compared using chi-square or Fisher exact tests. Tests were two-sided. To select predictive factors for allergic sensitization, binary logistic regression was carried out. Backward elimination, excluding variables with a significance level of P ≥ 0.1, ensured that the most parsimonious multivariable model was selected. To ensure that the binary logistic regression model’s assumption of variable independency was not violated, variance inflation factors (VIFs) for all variables in the final regression model were calculated. Significance was assumed at P < 0.05.
Characteristics of atopic children
Firstly, we compared children with positive SPT and available obstetric data (included children), with children with positive SPT for which obstetric data was not available (excluded children), Table 1. Generally, children with positive SPTs were challenged for [median (IQR)] 19 [11–27] allergens. While excluded children were sensitized to a similar number of allergens, they were significantly older and more frequently sensitized to inhaled allergens than included children. Because there was no obstetric data available prior to the year 2000, this study naturally excluded older children. The sensitization pattern for inhaled allergens has in previous studies been associated with increasing age , which might explain these differences. The sensitization frequencies to individual allergens for included children are shown in Additional file 1: Table S2.
Comparison of perinatal outcomes
Comparison of maternal and perinatal characteristics between atopic children and children born in the same time period (Table 2) showed that atopic children were more frequently male (P < 0.01). The high frequency of males in the sensitized sub-population explains why birth weight (mean ± SD; 3.395 ± 0.591 kg vs. 3.305 ± 0.685 kg; P < 0.05) was generally higher in SPT positive children, while z-scores did not differ between groups.
Mothers of atopic children had more frequently received assisted conception (P < 0.005), were less likely to smoke (P < 0.001), had fewer prior gravidities (P < 0.001) and consequently a lower parity (P < 0.005, Table 2). Interestingly, mothers of sensitized offspring also were older at delivery (P < 0.005) and more frequently born outside of Australia (P < 0.05). Generally, assisted conception and maternal age were significantly correlated (Pearson correlation r = 0.095, P < 10−98, Additional file 1: Figure S1). When stratifying our data by maternal origin of birth, it became apparent that atopic offspring more frequently had a mother born in Asia (19.1% vs. 10.2%; P < 10−6), while other children born in the same time period more frequently had mothers born in Australia (76.7% vs. 71.3%; P < 0.05; Additional file 1: Table S3) or in the rest of Oceania (7.1% vs. 3.2%; P < 0.05). Positive SPT frequency did not differ between children to European mothers (4.3% vs 3.3%; P = 0.36), North American mothers (0.4% vs. 0.4%; P = 0.84), South American mothers (1.1% vs 0.4%; P = 0.05) and mothers from Africa (0.7% vs. 1.9%; P = 0.14).
Independent risk factors for allergic sensitization
All variables in Table 2 were entered in the logistic regression model, except for gravidity, birth weight, length and HC. Gravidity was excluded because of the high correlation to parity; z-scores were used instead of birth measurements, to avoid confounding by sex and gestational age. After backward elimination, seven predictors remained in the model (Table 3; Fig. 2), namely maternal age, smoking maternal BMI, infant sex, preterm birth, parity, conception type and parity. The independence of these risk factors was ensured by calculation of VIFs for all independent variables in the final binary logistic regression model (Additional file 1: Table S4).
To ensure that these results did not only arise from the heightened sensitivity seen when a small population is compared to a bigger one, we computed binary logistic regressions on five randomly selected subpopulations, similar in size to the group with sensitized children (comprising ca 0.4% of the Obstetrix™ cohort, Additional file 1: Table S5). This analysis showed that results most likely are not a product of increased sensitivity. We also investigated whether time-dependent trends in the populations might have confounded analysis, since the birth year distribution of atopic children in our study slightly deviated from the distribution of the remaining population (Additional file 1: Figure S2). For this purpose, we investigated collinearity between birth year and factors identified in the binary logistic regression (Table 3). Overall correlation strength between SPT outcome and factors from the binary logistic regression remained unchanged when adjusted for birth year (Additional file 1: Table S6). To further ensure that results from the binary logistic regression model were not confounded, we forced birth year together with all previously identified factors (Table 3) into the regression model. This did not increase the regression models R2 and OR as well as P values of all previously identified factors remained largely unchanged, indicating that birth year distribution has not affected our results (Additional file 1: Table S7).
Prior work has established the importance of perinatal and maternal factors in the development of allergic disease. Godfrey et al. for example reported that disproportionate fetal growth was associated with raised IgE concentrations in adult life . However, these studies mostly focused on anthropometric measurements, which are potentially confounded by gestational age and infant sex. In this study, we tested the extent to which maternal factors and routinely measured perinatal factors could be associated with the development of allergic sensitization in children. Aside from strengthening the association of previously suggested risk factors of allergic sensitization, one novel perinatal factor was found to have an effect on childhood atopy, namely assisted conception.
While few previous studies have investigated the effect of assisted conception and/or maternal subfertility on allergic disease, our study is the first to report an increased risk of allergic sensitization in offspring after assisted conceptions. In accordance with our findings, some prior studies have reported an association between assisted conceptions/maternal subfertility and asthma [12,13,14]. Noteworthy, however, is that preceding studies used questionnaires, diagnosis codes, or prescription of antihistamines to identify atopic patients. While antihistamines can be used for treatment of other diseases, and diagnosis codes and questionnaires don’t guarantee underlying allergic status, the use of SPTs is a major strength of this study.
Assisted conception was found to be mildly correlated to maternal age. Though slightly correlated both variables remained in the binary logistic regression model as predictors after backward elimination, thus indicating that they have independent effects on allergic sensitization status during childhood. We further established that the assumptions of the binary logistic regression were not violated by the co-existence of these two variables in the model (Additional file 1: Table S4). Furthermore, we show that this finding is not due to time-dependent changes of the availability of assisted conception (Additional file 1: Figure S2, Tables S6, S7). Therefore, we assume that our analysis identified a true effect of assisted conception on pediatric allergic sensitization status. Though this has to be evaluated carefully in future studies, we hypothesize that such an effect might be explained by immunological shifts in the placenta during pregnancy.
Notable is that this study fails to identify differences in anthropometric measurements between sensitized children and the general population, after adjusting for gestational age and sex. This is not surprising, since a comprehensive review  pointed out, that methodological differences between previous studies have made the relationship between anthropometric measurements and allergic sensitization ambiguous. However, our study shows that both gestational age and infant sex, were significantly associated with allergic sensitization. While the association of male sex with increasing risk for allergic disease during early childhood is established , our study provides additional support to the fairly new concept that prematurity may decrease the risk of allergic sensitization in childhood.
Even though findings are heterogeneous, evidence that prematurity might be associated with a lower prevalence of allergic sensitization is accumulating. Two large Swedish register studies [17, 18] report that prematurity is associated with a lower prevalence of allergic rhinitis, in agreement with our findings. The relationship between gestational age and prevalence of allergic sensitization might be explained by hormone dependent shift in the placental immune environment. It can be hypothesized that a lower gestational age is associated with lower explosibility for progesterone which in turn leads to a less Th2 predominant environment in utero . Alternatively, it might reflect the effect of underlying risk factors of premature delivery.
In parallel with some previous research, we find that advanced maternal age (> 35 years) is a risk factor for allergic sensitization in childhood. As proposed by Dowhower Karpa et al. , increasing maternal age might expose offspring to a greater risk for atopic disease due to age-dependent changes in the gut’s microbiome, in particular the decrease of Lactobacilli and the increase in Clostridia have been linked to increased risk for atopy in the offspring.
Our results also strengthen the widely established association of increasing parity/birth order as a protective factor. Previous studies have suggested that increasing parity leads to in utero programming, which results in a more anti-inflammatory intrauterine environment, reflected by lower cord blood IgE levels at birth .
In our study, we found that mothers of sensitized children were less frequently overweight and smoked less. Therefore, our regression model suggests that maternal pre-pregnancy overweight and maternal smoking might be associated with a lower prevalence of allergic disease (Fig. 2). In regards to maternal BMI, some preceding studies suggested that the relationship of maternal BMI to allergic disease was U-shaped , implying that maternal underweight as well as maternal overweight were associated with a higher risk for atopy. Since most previous studies identifying maternal overweight as a risk factor, were conducted in the northern hemisphere, it is possible that differences in diet, or gut microbiome, could explain that maternal overweight was associated with a lower prevalence of allergic sensitization in this study. Furthermore, some previous studies  suggested that maternal smoking might be associated with a lower prevalence of allergic sensitization. Though the majority of scientific reports indicate that smoking and obesity are associated with a increased prevalence of allergic sensitization in offspring [24, 25]. A possible explanation for this studies deviating results might therefore be that, since a family history of allergic disease is a strong predictive factor for allergic sensitization , mothers of sensitized children are likely to opt for a healthier lifestyle motivated by their own allergic disease.
Interestingly, positive SPT were more common in offspring of mothers born in Asia. This supports previous findings made by the HealthNuts study group  and could suggest that maternal continent of birth affects the development of atopy.
We compared children with confirmed allergic sensitization with a population including approximately 20% sensitized children . While this aggravated identification of risk factors, it also ensured that the risk factors identified differed strongly between groups and hence should be repeatable in other populations. Furthermore, the use of z-scores is a strength of this study, as they were calculated based on a geographically and contemporaneously matched population. Another major strength of this study is the use of SPT to quantify allergic sensitization. However, the retrospective design did not allow controlling for allergic family history, maternal education or socioeconomic status, which in previous reports have been associated with allergic disease [16, 26]. Furthermore, this was a monocentric study and while the population size in this study is comparable to many other studies in this field, larger multicenter cohorts are needed to confirm these findings. It should also be mentioned again that children in the presented population underwent SPT due to clinical indications and hence were not screened with standard allergen panels. While clinical indication for SPT allows us to identify a group of children with symptomatic allergic sensitization, the absence of standard panels complicates comparison between groups. Furthermore, some previous studies , as well as the comparison between included and excluded children presented here (Table 1), indicate that sensitization pattern of children might be age dependent, though this dimension was not considered in the present study it should be considered in future studies of pediatric populations.
This study has identified associations between perinatal and maternal factors and allergic sensitization. In part our results confirm some previous reports but add to the current literature by using an objective measure of allergic sensitization and applying a multiple regression backward elimination approach. Interestingly, we show an increased risk of atopy after assisted conception and in offspring of mothers born in Asia, neither of which have been described in a more general context of allergic sensitization. These findings support the notion that early periconceptional and other environmental factors are likely to predispose towards atopy and warrant further mechanistic and epidemiological studies, aiming to develop primary preventive strategies for childhood allergies.
Availability of data and materials
The data analysed in this study was obtained from hospital records. The datasets for this manuscript are not publicly available to preserve patient integrity and confidentiality. Requests to access the datasets should be directed to the Nepean Blue Mountains Local Health District.
Body mass index
Skin prick test
Variance inflation factor
Pawanker R, Canonica GW, Holgate ST, Lockey RF. WAO White Book on Allergy 2011–12: executive summary. Milwaukee: World Allergy Organization; 2011.
Halken S. Prevention of allergic disease in childhood: clinical and epidemiological aspects of primary and secondary allergy prevention. Pediatr Allergy Immunol. 2004;15(Suppl 16):4–5, 9–32.
Brook JS, Whiteman M, Brook DW. Transmission of risk factors across three generations. Psychol Rep. 1999;85(1):227–41.
Godfrey KM, Barker DJ, Osmond C. Disproportionate fetal growth and raised IgE concentration in adult life. Clin Exp Allergy. 1994;24(7):641–8.
Gluckman PHM. Mismatch: why our world no longer fits our bodies. Human biology & health. Oxford: Oxford University Press; 2006.
Godfrey KM, Lillycrop KA, Burdge GC, Gluckman PD, Hanson MA. Epigenetic mechanisms and the mismatch concept of the developmental origins of health and disease. Pediatr Res. 2007;61(5 Pt 2):5R–10R.
ObstetrixTM. Sydney: Meridian Health Informatics; 2018.
Allergy in Australia 2014: a submission for allergic diseases to be recognised as a National Health Priority Area. Australasian Society of Clinical Immunology and Allergy Inc.; 2014. https://www.allergy.org.au/images/stories/reports/ASCIA_Allergy_in_Australia_2014_NHPA__Submission.pdf.
Skin prick testing for the diagnosis of allergic disease—a manual for practicioners. Australasian Society of Clinical Immunology and Allergy Inc.; 2016. https://www.allergy.org.au/images/stories/pospapers/ASCIA_SPT_Manual_March_2016.pdf.
Cole TJ. The LMS method for constructing normalized growth standards. Eur J Clin Nutr. 1990;44(1):45–60.
Chiu CY, Huang YL, Tsai MH, Tu YL, Hua MC, Yao TC, et al. Sensitization to food and inhalant allergens in relation to atopic diseases in early childhood: a birth cohort study. PLoS ONE. 2014;9(7):e102809.
Kallen B, Finnstrom O, Nygren KG, Otterblad OP. Asthma in Swedish children conceived by in vitro fertilisation. Arch Dis Child. 2013;98(2):92–6.
Carson C, Sacker A, Kelly Y, Redshaw M, Kurinczuk JJ, Quigley MA. Asthma in children born after infertility treatment: findings from the UK Millennium Cohort Study. Hum Reprod. 2013;28(2):471–9.
Harju M, Keski-Nisula L, Raatikainen K, Pekkanen J, Heinonen S. Maternal fecundity and asthma among offspring-is the risk programmed preconceptionally? Retrospective observational study. Fertil Steril. 2013;99(3):761–767.e1.
Wandalsen GF, Chong-Neto HJ, de Souza FS, Sole D, Bacharier LB. Early weight gain and the development of asthma and atopy in children. Curr Opin Allergy Clin Immunol. 2014;14(2):126–30.
Arshad SH, Stevens M, Hide DW. The effect of genetic and environmental factors on the prevalence of allergic disorders at the age of two years. Clin Exp Allergy. 1993;23(6):504–11.
Crump C, Sundquist K, Sundquist J, Winkleby MA. Gestational age at birth and risk of allergic rhinitis in young adulthood. J Allergy Clin Immunol. 2011;127(5):1173–9.
Braback L, Hedberg A. Perinatal risk factors for atopic disease in conscripts. Clin Exp Allergy. 1998;28(8):936–42.
Santner-Nanan B, Peek MJ, Khanam R, Richarts L, Zhu E, Fazekas de St Groth B, et al. Systemic increase in the ratio between Foxp3+ and IL-17-producing CD4+ T cells in healthy pregnancy but not in preeclampsia. J Immunol. 2009;183(11):7023–30.
Dowhower Karpa K, Paul IM, Leckie JA, Shung S, Carkaci-Salli N, Vrana KE, et al. A retrospective chart review to identify perinatal factors associated with food allergies. Nutr J. 2012;11:87.
Karmaus W, Arshad H, Mattes J. Does the sibling effect have its origin in utero? Investigating birth order, cord blood immunoglobulin E concentration, and allergic sensitization at age 4 years. Am J Epidemiol. 2001;154(10):909–15.
Forno E, Young OM, Kumar R, Simhan H, Celedon JC. Maternal obesity in pregnancy, gestational weight gain, and risk of childhood asthma. Pediatrics. 2014;134(2):e535–46.
Strachan DP, Harkins LS, Johnston ID, Anderson HR. Childhood antecedents of allergic sensitization in young British adults. J Allergy Clin Immunol. 1997;99(1 Pt 1):6–12.
Kramer U, Lemmen CH, Behrendt H, Link E, Schafer T, Gostomzyk J, et al. The effect of environmental tobacco smoke on eczema and allergic sensitization in children. Br J Dermatol. 2004;150(1):111–8.
Scholtens S, Wijga AH, Brunekreef B, Kerkhof M, Postma DS, Oldenwening M, et al. Maternal overweight before pregnancy and asthma in offspring followed for 8 years. Int J Obes. 2009;34:606.
Mielck A, Reitmeir P, Wjst M. Severity of childhood asthma by socioeconomic status. Int J Epidemiol. 1996;25(2):388–93.
Koplin JJ, Peters RL, Ponsonby AL, Gurrin LC, Hill D, Tang ML, et al. Increased risk of peanut allergy in infants of Asian-born parents compared to those of Australian-born parents. Allergy. 2014;69(12):1639–47.
This research was not supported by any specific grant.
Ethics approval and consent to participate
This study complies with the Helsinki Declaration and relevant national guidelines and was approved by the Human Research Ethics Committee of the Nepean Blue Mountains Local Health District (Ethics number: 10/16).
Consent for publication
Not applicable, as analysis was conducted retrospectively on data base.
Authors declare that they have no affiliations with or involvement in any organization or entity with any financial interest, or non-financial interest in the subject matter or materials discussed in this manuscript.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Additional methods. Table S1. Compilation of allergens used in skin prick tests. Table S2. Skin prick test based sensitization rates for the 286 included children. Only allergens for which at least one child tested positive are shown. Table S3. Continent of birth for mothers of atopic children and children from the remaining population. Table S4. Correlation coefficients between independent risk factors for allergic sensitization. Table S5. Binary logistic regression of independent risk factors from Table 3 comparing a randomly chosen subpopulation consisting of 0.4% children to the remaining cohort. Table S6. Correlation of birth year with the outcomes of the binary logistic regression model. Table S7. Modelling the effect of birth year on the outcomes of the binary logistic regression model (Table 3). Figure S1. Histogram of maternal age in the context of conception type. Figure S2. Histogram of birth year for sensitized children and the remaining population.
About this article
Cite this article
Schäfer, S., Liu, A., Campbell, D. et al. Analysis of maternal and perinatal determinants of allergic sensitization in childhood. Allergy Asthma Clin Immunol 16, 71 (2020). https://doi.org/10.1186/s13223-020-00467-5
- Assisted conception
- Skin prick test
- Fetal development