Article Text

PDF

The role of prenatal, perinatal and postnatal factors in the explanation of socioeconomic inequalities in preschool asthma symptoms: the Generation R Study
  1. Esther Hafkamp-de Groen1,2,
  2. Lenie van Rossem1,2,
  3. Johan C de Jongste3,
  4. Ashna D Mohangoo1,2,4,
  5. Henriëtte A Moll3,
  6. Vincent W V Jaddoe1,3,5,
  7. Albert Hofman5,
  8. Johan P Mackenbach2,
  9. Hein Raat2
  1. 1The Generation R Study Group, Erasmus MC, Rotterdam, The Netherlands
  2. 2Department of Public Health, Erasmus MC, Rotterdam, The Netherlands
  3. 3Department of Paediatrics, Erasmus MC-Sophia Children's Hospital, Rotterdam, The Netherlands
  4. 4TNO, Netherlands Organisation for Applied Scientific Research, Department Child Health, Leiden, The Netherlands
  5. 5Department of Epidemiology, Erasmus MC, Rotterdam, The Netherlands
  1. Correspondence to Esther Hafkamp-de Groen, Public Health/Generation R Study, Erasmus Medical Center, PO Box 2040 (Room Ae−003), Rotterdam 3000 CA, The Netherlands; e.hafkamp{at}erasmusmc.nl

Abstract

Background The authors assessed whether socioeconomic inequalities in asthma symptoms were already present in preschool children and to what extent prenatal, perinatal and postnatal risk factors for asthma symptoms mediate the effect of socioeconomic status (SES).

Methods The study included 3136 Dutch children participating in the Generation R Study, a prospective cohort study. Adjusted ORs of asthma symptoms for low and middle SES (household income and maternal education) compared to high SES were calculated after adjustment for potential confounders and also adjusted for prenatal, perinatal and postnatal mediators at preschool age.

Results At age 1 year, low-SES children had a 40% lower risk of asthma symptoms compared to high-SES children (p<0.01). However, the risk of asthma symptoms in 3- and 4-year-old low-SES children was 1.5 times higher compared to their high-SES age mates (p<0.05). The positive associations at age 1 year were particularly modified by postnatal factors (up to 38%). In toddlers, prenatal factors explained up to 58% of the negative associations between SES and asthma symptoms.

Conclusions SES indirectly affects asthma symptoms at preschool age. The inverse association between SES and asthma symptoms emerges at age 3 years. This is particularly due to a high level of adverse prenatal circumstances in low-SES toddlers. Future research should evaluate public health programs (during pregnancy) to reduce socioeconomic inequalities in childhood asthma.

  • Asthma
  • child health
  • preventive medicine
  • social inequalities
  • birth defects
  • infant mortality
  • perinatal epidemiology
  • public health
  • social epidemiology
  • health expectancy
  • asthma symptoms
  • maternal educational level
  • household income
  • preschool children
  • socioeconomic status

Statistics from Altmetric.com

Introduction

Recently, marked variations in the prevalence of asthma were shown between countries, with the highest rates in children living in countries undergoing rapid development.1 Also within a country, the prevalence of asthma showed a mixed picture and disproportionally affected various socioeconomic status (SES) groups.2

It remains unclear to what extent disparities in preschool asthma symptoms are due to socioeconomic differences. In-depth reports on socioeconomic inequalities in asthma symptoms in preschool children are scarce, and results are conflicting. While some studies report that asthma prevalence is disproportionately high among low-SES children,3–8 others found no or only a weak association between SES and asthma.9–13 Four of these studies analysed preschool children.3 ,6–8 In preschool children, an asthma-symptom-based rather than an asthma-diagnosis-based approach has been proposed because it is difficult to diagnose asthma prior to age 5.14 Our main hypothesis is that SES may indirectly affect asthma symptoms, such as wheezing and breathlessness: low-SES children are more likely to be susceptible to asthma symptoms due to a high level of common risk factors, such as tobacco smoke exposure,15 whereas protective factors such as breastfeeding16 are less common in low-SES families.17

This is the first longitudinal study in a large ethnically homogeneous population to investigate the association between SES and asthma symptoms at preschool age. We examined to what extent known risk factors for asthma in the prenatal, perinatal and postnatal period mediate the effect of SES. This study elucidates the mechanisms underlying the association between SES and asthma symptoms at preschool age and helps identify areas needing attention to promote child healthcare.

Methods

Study design

This study was embedded in the Generation R Study, a population-based prospective cohort study.18 ,19 Consent for postnatal follow-up was available for 7295 children. Since socioeconomic disparities in asthma may vary by ethnicity, the present study was restricted to an ethnically homogeneous population.20 A total of 3824 children were assigned Dutch ethnicity. In accordance with the Dutch Standard Classification, we assigned a Dutch ethnicity to a child if both parents were born in the Netherlands.21 To take into account third-generation immigrants, a child was considered Dutch if both parents were born in the Netherlands and at least one grandparent of both parents was born in the Netherlands. If children had one or both parents born abroad, and all four grandparents born in the Netherlands (n=54), these children were also considered Dutch. The study was conducted in accordance with the guidelines proposed in the Declaration of Helsinki. The Medical Ethics Committee of the Erasmus Medical Center, Rotterdam, approved the study, and written informed consent was obtained.

Socioeconomic status

Two individual indicators of SES were used in this study: maternal educational level and household income. Maternal educational level was established at enrolment and categorised as follows: low (<4 years of high school), mid-low (college), mid-high (bachelor) and high (master).22

Data on income were available at age 2 years. Parents reported their own average net monthly income. Responses were categorised into three levels: low (<€2000/month, ie, below modal income),23 middle (€2000−€3300/month) and high (>€3300/month).

Asthma symptoms

In preschool children, it is difficult to diagnose asthma because symptoms are non-specific, often transient, and no diagnostic tests are available. In preschool children, asthma has commonly been defined as the presence of parent-reported asthma symptoms.24 Parentally retrieved questionnaires were obtained at ages 1, 2, 3 and 4 years. ‘Wheezing and breathlessness during the past year (yes, no)’ were measured with validated questions taken from the International Study of Asthma and Allergies in Childhood.25

Covariates

Selection of covariates was based on reports of early determinants of childhood asthma.26 ,27 Child's gender and exact age at measurement and age of mother at enrolment were treated as confounders. The effect of SES on the risk of asthma symptoms is likely to act through mediators (see figure 1). The following covariates (in italics) were treated as mediators (categorised in prenatal, perinatal and postnatal mediators):

Figure 1

Simplified conceptual framework for the association between socioeconomic status and asthma symptoms at age 1, 2, 3 and 4 years.

Information on prenatal mediators was established using postal questionnaires during pregnancy. These included smoking during pregnancy (yes, no); maternal atopy (yes, no); maternal psychopathology during pregnancy as assessed using the Global Severity Index (GSI) of the Brief Symptom Inventory (a validated self-report measure, which consists of 53 positive and negative self-appraisal statements)28; long-lasting difficulties during the year preceding the pregnancy as evaluated with a 12-item checklist29 and (poor) family functioning as measured with the Family Assessment Device (FAD: a validated self-report 12-item questionnaire) during pregnancy.30 Respective item scores were summed to derive a total score of the GSI (range 0–2.29), checklist for long-lasting difficulties (range 0–18) and FAD (range 1–3.75), with higher scores denoting more symptoms. Total scores were divided into tertiles (cut-off points: GSI (1.25 and 1.75); checklist for long-lasting difficulties (1 and 3); FAD (0.08 and 0.19)).

Perinatal factors included birth weight (in grams) and gestational age at birth (in weeks). Both were obtained from medical records.

Postnatal factors were established using questionnaires and included: breastfeeding at age 6 months (yes, no); keeping pets (yes, no) at age 1 year; having siblings (yes, no) at ages 2 and 3 years; day-care attendance (yes, no) at ages 1, 2 and 3 years; tobacco smoke exposure (yes, no) measured at age 6 months and ages 2 and 3 years; eczema (yes, no) at age 3 years and respiratory tract infections at ages 1, 2, 3 and 4 years. Parents were asked whether their child has been to a doctor with fever and cough/runny or blocked nose/ear ache in the preceding year to define respiratory tract infections (yes, no).

Statistical analyses

The associations between SES and asthma symptoms in children at ages 1, 2, 3 and 4 years were analysed using generalised estimating equation models (using complete cases) to address the analysis using multiple observations per child. To save space, we only explained the positive association at age 1 and negative association at age 4 between SES and asthma symptoms. Because the missing values were not completely at random, complete-case analysis was likely to introduce biased results. A multiple imputation method was used to impute missing values (with a maximum percentage missing of 20%).31 Missing values in the study variables ranged from 0% (birth weight) to 29% (tobacco smoke exposure at age 6 months). Ten imputed data sets were generated using a fully conditional specified model to handle missing values. Imputations were based on the relations between all variables in the study. We computed five multivariable logistic regression models. We used the ENTER method to construct our models. This method enters all variables at the same time. The highest income level and maternal educational level were set as reference. First, we fit a model that was adjusted for confounders (Basic model). When results of the Basic model showed significant results, we added the hypothesised mediators separately (prenatal, perinatal and postnatal mediators) to show the impact on the association between SES and asthma symptoms. Finally, we adjusted for all variables simultaneously (Full model). For each adjustment, the percentage change in OR for the SES level with a decreased or increased risk of asthma symptoms was calculated by (100× (ORBasic Model−OR+mediators)/(ORBasic model−1)).

No differences in results were observed between analyses with imputed missing data or complete cases. All measures of association are presented in OR with their 95% CI. All analyses were performed using SPSS V.18.0 for Windows (Statistical Package of Social Sciences; SPSS Inc).

Non-response analysis

Families with missing data on household income (n=688) were compared with families who filled out the questions on household income (n=3136). Differences between responders and non-responders were present in covariates, except for gender, maternal atopy, siblings, tobacco smoke exposure at age 3 years, eczema and respiratory tract infections (p>0.05) (see online appendix 1).

Results

General characteristics

Complete data on household income were available in 3136 (1592 boys and 1544 girls) of the 3824 children (82%). For 3136 children, the parents had returned at least one of the questionnaires at ages 0–4 years. Maternal educational level was available in 99.7% of the 3136 children. Table 1 shows that 11% of the children were in the lowest income level and 53% were in the highest income level, 8% of the mothers were in the lowest educational level and 40% were in the highest educational level. Tobacco smoke exposure decreased from 14% in the first 2 years of life to 10% at age 3 years. Respiratory tract infections were most frequently reported at age 2 years (47%). Day-care attendance increased from 71% at age 1 year to 95% at age 3 years. Income differences were present in all outcomes and covariates, except for gender, maternal atopy, breastfeeding and respiratory tract infections at ages 2 and 4 years. Children from low-income families had a lower mean birth weight, less siblings, less day-care attendance and less respiratory tract infections at age 1 year, but more often respiratory tract infections at age 3 years, compared to children from high-income families. Mothers with highest tertile psychopathology scores, long-lasting difficulties scores and poor family functioning scores during pregnancy more often were in the lowest income group. Low-income mothers more often had a shorter gestational duration and kept pets compared to high-income mothers.

Table 1

Characteristics of the total study population, by household income level (n=3136)

Associations between SES and asthma symptoms

The prevalence of asthma symptoms decreased with increasing age. In the first year of life, wheezing and breathlessness showed a positive household income gradient and at ages 3 and 4 years a negative household income gradient (figure 2). After adjustment for potential confounders, low-income children were at lower risk of wheezing at age 1 year (adjusted OR (aOR)=0.71, 95% CI 0.53 to 0.95), at higher risk of wheezing at ages 3 and 4 years (aOR=1.57, 95% CI 1.09 to 2.26 and aOR=1.53, 95% CI 1.06 to 2.22, respectively); and low-income and middle-income children were at higher risk of breathlessness at age 3 years (aOR=1.87, 95% CI 1.31 to 2.67 and aOR=1.43, 95% CI 1.12 to 1.84, respectively) compared to high-income age mates (figure 3).

Figure 2

Prevalence of wheezing and breathlessness by socioeconomic status (household income and maternal educational level) at preschool age. Prevalences are unadjusted (n=3136).

Figure 3

Associations between socioeconomic status (household income and maternal educational level) and wheezing and breathlessness, based on generalised estimating equation models. Models were adjusted for maternal age and child's gender. Adjusted ORs (aOR) and 95% CI were given (allowing for a time trend) for each year of age separately (n=3136).

A negative maternal educational gradient in child's wheezing and breathlessness was found after the second year of life (figure 2). After adjustment for potential confounders, children of low-educated mothers were at lower risk of wheezing and breathlessness at age 1 year (aOR=0.58, 95% CI 0.41 to 0.82 and aOR=0.63, 95% CI 0.44 to 0.92, respectively), at higher risk of breathlessness at age 3 years (aOR=1.63, 95% CI 1.06 to 2.51) and at higher risk of breathlessness at age 4 years (aOR=1.62, 95% CI 1.05 to 2.50); children of mid-low educated mothers were at higher risk of wheezing and breathlessness at age 3 years (aOR=1.56, 95% CI 1.15 to 2.12 and aOR=1.69, 95% CI 1.26 to 2.27, respectively) and at higher risk of wheezing at age 4 years (aOR=1.43, 95% CI 1.06 to 1.94); and children of mid-high educated mothers were at lower risk of wheezing at age 1 year (aOR=0.81, 95% CI 0.66 to 0.99) and at higher risk of wheezing at age 3 years (aOR=1.35, 95% CI 1.01 to 1.82) compared to age mates of high-educated mothers (figure 3).

Table 2 showed that the 28% lower risk of wheezing in low-income children compared to high-income age mates was neutralised after adjustment for postnatal factors at age 1 year. In 1-year-old children of low-educated mothers, postnatal factors explained 19% ((0.58−0.66/0.58−1)×100) and 38% ((0.63−0.77/0.63−1)×100) of the decreased risk of wheezing and breathlessness, respectively. This was mainly due to the variables day-care attendance, respiratory tract infections and presence of siblings (see online appendix 2).

Table 2

Multivariable logistic regression models fitted on wheezing and breathlessness at ages 1 and 4 years (n=3136)

At age 3 years (data not shown), prenatal factors explained 74% ((1.87−1.50/1.50−1)×100) of the elevated risk of breathlessness in low-income children. This was mainly due to the variables maternal psychopathology and maternal atopy. At age 4 years, adjustment for prenatal factors reduced the aOR for the association between low income and wheezing and breathlessness to 1.24 and 1.15, respectively. Prenatal factors explained 58% ((1.62−1.26/1.62−1)×100) and postnatal factors explained 32% ((1.62−1.42/1.62−1)×100) of the elevated risk of breathlessness in children of low-educated mothers. The aOR in the full model only remained significant for the association between mid-low maternal educational level and child's wheezing at age 4 years.

Discussion

This longitudinal cohort study in an ethnically homogeneous group showed that the direction of the association between SES and asthma symptoms changed from a positive association at age 1 year into a negative association at age 3 and 4 years. The pathway between SES and asthma symptoms particularly was mediated by postnatal factors in the first year of life and by prenatal factors in toddlers.

Comparison with other studies

Mielck et al 32 reviewed 22 studies on the association between SES and childhood asthma and demonstrated conflicting results. Although findings regarding the strength and direction of the SES gradient remain mixed, most studies revealed that children from low-SES families more often have asthma symptoms or an asthma diagnosis.3–8 ,17 ,33 Comparison of our results with earlier findings is hampered due to different age groups, indicators of SES and various asthma outcomes that were applied. Several studies used dichotomised physician-diagnosed asthma outcomes.4 ,5 ,8 ,10 ,12 ,13 ,17 Some studies applied wheezing as an outcome in the association between SES and asthma.9 ,11 ,13 ,17 Only one study has investigated the association between SES and asthma symptoms in preschool children at three different time points, and they identified pathways through which income might influence childhood asthma symptoms. They found a mediating effect of some (grand) parental risk factors.11

We evaluated household income and maternal education as two separate indicators of SES in relation with asthma symptoms. This study shows that both household income and maternal education affect asthma symptoms at preschool age in a similar way. Furthermore, associations with these two indicators of SES showed a similar pattern for wheezing and breathlessness. This supports the evidence for the presence of an association between SES and asthma symptoms at preschool age. This is the first longitudinal study that showed a change in the direction of the association between SES and asthma symptoms at preschool age. The inconsistent findings on the association between SES and asthma in previous studies may (in part) be due to the use of cross-sectional data at one moment in time.

Association between SES and asthma symptoms

Most preschool children with asthma symptoms, such as wheezing and breathlessness, do not really develop asthma.34 Wheezing and breathlessness are non-specific, many times related to respiratory tract infections. Therefore, adjustment was made for (indicators of) respiratory tract infections.

Interestingly, the positive association between SES and asthma symptoms at age 1 was particularly explained by postnatal factors (including respiratory tract infections); the postnatal factors considerably attenuated the association between SES and asthma symptoms compared to prenatal and perinatal factors. Possible mechanisms by which these postnatal factors may influence asthma symptoms in the first year of life have previous been reported35: postnatal factors such as day-care attendance and the presence of siblings were associated with transient early wheeze, probably because they increase the risk of respiratory tract infections. So, at age 1 year, it is likely that wheezing and breathlessness are symptoms of infection.

In toddlers, we showed that particularly prenatal factors mediated the associations found between SES and asthma symptoms. Prenatal factors such as maternal psychopathology, long-lasting difficulties and poor family functioning during pregnancy might be indicators of prenatal stress. Previous studies showed that prenatal stress, smoking during pregnancy and maternal atopy are associated with asthma symptoms.36–39 Possible mechanisms by which these prenatal factors may influence the development of asthma symptoms have also been reported: (1) prenatal stress may contribute to asthma pathogenesis via neuroendocrine and immune pathways,36 (2) pulmonary/airway development goes ‘off track’ in utero in children born of smoking mothers40 and (3) maternal atopy could be seen as an indicator of genetic predisposition to childhood asthma.41

The concept that childhood asthma symptoms comprise several heterogeneous wheezing phenotypes may be in line with our findings. Rusconi et al 42 found different patterns of risk factors for different wheezing phenotypes. Having siblings and day-care attendance were risk factors for transient early wheezing. Maternal atopy and maternal smoking during pregnancy were more likely to be associated with persistent wheezing.41 Taken together, this may suggest that high-SES children more often have early transient wheezing and low-SES children are more susceptible to develop persistent wheezing, which is often considered a risk factor for developing asthma.43 In the future, the follow-up of our cohort will determine whether the increased prevalence of asthma symptoms in certain SES groups represents a temporary association in early childhood or predicts progression to asthma.

While SES is strongly related to perinatal factors,44 these factors hardly contributed to the explanation of the observed socioeconomic differences in asthma symptoms at preschool age, suggesting that a low SES does not influence a child's risk to asthma symptoms through its link with birth weight and gestational age.

The strongest associations were found for SES (maternal educational level) and wheezing at age 1 and associations between SES (household income or maternal educational level) and breathlessness at age 3; these associations remain statistically significant after applying a Bonferroni correction for multiple testing (p<0.003; ie, 0.05/16).

A substantial proportion of the effect of SES on asthma symptoms remained unexplained; it could be argued that genetic factors and gene-by-environment interactions among distinct socioeconomic groups might predispose infants to the development of asthma symptoms.45 It should be acknowledged that, in the present study, unmeasured variables, such as traffic air exposure or different attitudes towards the use of the healthcare, could (in part) explain the association between SES and asthma symptoms.

Methodological considerations

Strengths of this study are the design with repeated measurements of asthma symptoms and covariates. Stratification by asthma symptoms and the use of both household income and maternal educational level as indicators of SES are other original contributions of our study.

Some limitations need to be addressed. Selection bias due to non-response would be present if the associations of household income with asthma symptoms differ between those with (n=3824) and those without (n=688) data on household income. Although the general characteristics of those with versus without data on household income were different, no differences in asthma symptoms were found. Thus, selection bias due to non-response on household income seems unlikely but cannot be excluded. Another limitation was that the population studied appeared to be relatively affluent: 53% was categorised as high income and 40% had a mother with a high educational level. Therefore, our results may not be generalisable to more deprived populations. Because the highest household income category was predefined (>€3300/month), we were not able to study the effect of two or three times the modal income on asthma symptoms. We recommend that future studies focus on asthma symptoms in more detailed household income subgroups.

Asthma symptoms were parent reported in the Generation R Study. It remains debatable whether or not parents' reports on asthma symptoms are accurate or not.46 ,47 We used validated questions on the frequency of asthma symptoms, taken from the International Study of Asthma and Allergies in Childhood questionnaires as they were previously used in the Dutch PIAMA cohort.48

We recommend future studies to explore the association between SES and asthma symptoms, with the use of structural equations models in addition to a logistic regression framework, to gain more insight in the mediating pathways.

Conclusions

SES indirectly affects asthma symptoms, already at preschool age. Socioeconomic inequalities in preschool asthma symptoms have their origin early in life and come to expression as an inverse association at the third year of life. SES in early life is important since studies found that changes in later family income did not offset the effects of early-life SES in terms of children's risk of having asthma.12 Follow-up is needed to establish any effect of SES on the persistence of asthma symptoms later in life. We recommend more studies in varied populations to confirm or reject these findings and to evaluate public health programmes (during pregnancy) to reduce socioeconomic inequalities in childhood asthma.

What is already known on this subject

  • SES is a relevant determinant of children's health.

  • In-depth reports on socioeconomic inequalities in asthma symptoms in preschool children are scarce, and results are conflicting.

  • Pathways through which SES might influence the development of asthma symptoms in early childhood are still open to debate.

What this study adds

  • This study showed an inverse socioeconomic gradient in asthma symptoms at ages 3 and 4 years and showed that adverse prenatal circumstances particularly explain the inverse association between SES and asthma symptoms in toddlers.

Acknowledgments

We gratefully acknowledge the contribution of general practitioners, hospitals and midwives in Rotterdam. The Generation R Study is conducted by the Erasmus Medical Center in close collaboration with the School of Law and Faculty of Social Sciences of the Erasmus University Rotterdam, the Municipal Health Service Rotterdam area, Rotterdam, the Rotterdam Homecare Foundation, Rotterdam and the Stichting Trombosedienst & Artsenlaboratorium Rijnmond (STAR-MDC), Rotterdam. We gratefully acknowledge the contribution of children and parents, general practitioners, hospitals, midwives and pharmacies in Rotterdam. The general design of Generation R Study is made possible by financial support from the Erasmus Medical Center, Rotterdam, the Erasmus University Rotterdam, the Netherlands Organization for Health Research and Development (ZonMw), the Netherlands Organisation for Scientific Research(NWO), the Ministry of Health, Welfare and Sport and the Ministry of Youth and Families.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Funding The first phase of the Generation R Study was funded by Erasmus Medical Centre and Erasmus University Rotterdam and the Netherlands Organization for Health Research and Development (ZonMw). The present study was supported by an additional grant from ZonMw (grant number 22000128).

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval The Medical Ethics Committee of the Erasmus Medical Center, Rotterdam, approved the study.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.