Prenatal maternal psychological stress and childhood asthma and wheezing: a meta-analysis

Kim F.E. van de Loo, Marleen M.H.J. van Gelder, Jolt Roukema, Nel Roeleveld, Peter J.F.M. Merkus, Christianne M. Verhaak
European Respiratory Journal 2016 47: 133-146; DOI: 10.1183/13993003.00299-2015

Abstract

The aim of this study was to systematically review and meta-analyse observational studies on prenatal maternal psychological stress and the subsequent development of asthma and wheezing in early childhood.

All available published literature from 1960 until November 2013 was systematically searched through electronic databases (PubMed, Embase, PsycInfo and Web of Science). All observational studies assessing associations between any form of prenatal maternal psychological stress and respiratory morbidity in the child were included. Data extraction, quality assessment and meta-analyses were performed.

The overall meta-analysis included 10 studies and showed that the prevalence of wheezing, asthma and other respiratory symptoms is higher in children of mothers who were exposed to or experienced some form of psychological stress during pregnancy than in mothers who did not (pooled OR 1.56 (95% CI 1.36–1.80)). Comparable results were observed in subgroup analyses of stress exposure, perceived stress, asthma and wheezing.

This study demonstrates that prenatal maternal psychological stress is associated with respiratory morbidity, including asthma and wheezing in the child. Future studies examining the early origins of asthma and wheezing need to account for the impact of prenatal maternal stress.

Abstract

Study showing an association between prenatal psychological stress and subsequent respiratory morbidity in children http://ow.ly/USkLN

Introduction

Asthma is a common chronic disease affecting ∼300 million people worldwide [1]. The prevalence of asthma is increasing, especially in children [1, 2], and it is now the most common chronic disease in childhood [3]. A diagnosis of asthma is often preceded by wheezing in infancy [4, 5], for which prevalence estimates range 1.7– 32.2% [6].

Recent literature states that asthma originates in early life [5]. The Barker hypothesis [7], also known as “early life programming”, describes that maternal disease and lifestyle factors can affect fetal development and growth, and can be linked with adverse effects on health outcomes in the offspring later in life [8]. Early life programming also applies to lung development and functioning. Lung development is influenced by both genetic and environmental factors, as well as by interactions between these factors, which may have persistent effects on lung function and respiratory health in later life [5, 9]. Therefore, the prenatal environment and condition of the fetus are important for the development of the respiratory system and may also play a role in the development of asthma [10].

Several factors may affect developmental processes and are associated with respiratory disease in later life. Potential risk factors during pregnancy include maternal smoking, suboptimal fetal nutrition, insufficient dietary intake, vitamin D deficiency, obesity and exposure to air pollution [10–14].

Associations between maternal psychological factors and in utero developmental processes were investigated in several prospective studies, which demonstrated long-term adverse physical and mental health outcomes in the child (e.g. emotional problems, attention deficit hyperactivity disorder, impaired cognitive development and diabetes) [15–17]. However, the effects of maternal psychological stress on the prenatal development of the respiratory system and postnatal respiratory health have not been studied extensively. The mechanisms through which psychological stress during pregnancy could influence the in utero development of the respiratory system are thought to involve disruption of the balance between the neuroendocrine system, the autonomic nervous system (ANS) and the immune system [18]. More specifically, prenatal maternal stress may influence the neuroendocrine system and hence the immune system through fetal programming of the hypothalamic–pituitary–adrenal (HPA) axis [19]. Based on these theories, it is conceivable that prenatal maternal psychological stress may affect respiratory function in the child. Investigating this relationship could provide new insights into the role of prenatal maternal psychological stress in the origins of asthma.

However, an overview of the influence of psychological stress during pregnancy on lung function in the child is lacking. The purpose of this systematic review and meta-analysis is to assess whether an association exists between prenatal maternal psychological stress (stress exposure and perceived stress) and the development of respiratory morbidity in childhood, specifically asthma and wheezing, and to gain a better understanding of prenatal vulnerabilities for lung development.

Methods

For the analysis of the association between prenatal psychological stress and the development of respiratory morbidity, we used search strategies that followed the Cochrane guidelines [20]. In addition, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for adequate reporting were followed [21].

Search strategy

All available published literature from 1960 until 15 November 2013 was systematically searched using electronic databases. We searched PubMed, Embase, PsycInfo and Web of Science, including the Science Citation Index Expanded, the Social Sciences Citation Index and the Arts & Humanities Citation Index. Search terms were identified based on previous research and discussed by four of the contributing authors (K.F.E. van de Loo, J. Roukema, P.J.F.M. Merkus, and C.M. Verhaak). All search terms were individually adapted for each database, based on standardised subject terms, e.g. Medical Subject Headings (MeSH) terms or subject headings, and free-text terms. All search terms were checked by a librarian and the title, keywords and abstract of potentially eligible manuscripts were searched. To prevent bias, a broad search was performed based on separate search terms for participants, exposures and outcome measures. The participant terms were related to pregnant women (search terms were, for example, “pregnancy”, “pregnant women”, “maternal exposure” and “prenatal exposure”). The exposure variable search terms included all forms of prenatal maternal psychological stress, including all types of stress exposure (experiencing a stressful event) and perceived stress (experiencing stress symptoms) (search terms were, for example, “stress”, “psychological stress”, “mental health”, “anxiety” and “depression”). We used respiratory morbidity, including both asthma and wheezing in childhood, as outcome search terms (search terms were, for example, “respiratory sounds”, “asthma” and “bronchial hyperreactivity”). See online supplementary material S1 for a more detailed description of the search strategy. The results of all databases were merged using reference management software Endnote X5 (www.endnote.com) and duplicates were removed. No studies were excluded based on language.

Inclusion criteria

The inclusion criteria for the systematic review were: 1) data pertaining to an original study (i.e. no review articles, editorials or commentaries); 2) no strict definitions necessary for “prenatal stress” (any study reporting “prenatal stress” was eligible regardless of the measurement performed) and “asthma” or “wheezing” (any study assessing any form of respiratory morbidity was included, regardless of the measurement), but a clear statement or understanding that “prenatal stress” was the exposure of interest and “asthma” and/or “wheezing” were the outcomes of interest in each eligible study was necessary; and 3) provision of an adjusted relative risk (RR), odds ratio (OR) or hazard ratio (or adequate data in order to compute these parameters).

Study selection

Studies were first selected based on careful screening of the titles by one author (K.F.E. van de Loo). To ensure no relevant studies were missed, the titles of the first 170 articles were screened by two authors (K.F.E. van de Loo and C.M. Verhaak). The results were compared and checked using the abstracts in case of disagreement. Secondly, two authors (K.F.E. van de Loo and M.M.H.J. van Gelder) independently screened the abstracts and subsequently the remaining full-text articles. Disagreements were resolved by discussion with a third author (C.M. Verhaak) until consensus was reached. κ statistics were calculated to assess the levels of agreement [22]. A snowballing method was also used to search for additional studies. The authors of some primary studies were contacted to provide missing or additional data when needed for the inclusion or exclusion of these studies (n=3).

Data extraction

Data on study design, time period, country, study cohort, number of participants, demographic variables, age of the children, assessment measures, timing of exposure, outcome variables and covariates were extracted by two authors (K.F.E. van de Loo and C.M. Verhaak). For all studies, the adjusted estimates as provided in the articles were extracted. Furthermore, we collected the available data on the number of pregnant women who experienced some form of psychological stress and the number of women who did not experience stress. Within these groups, we were interested in the number of children with and without respiratory morbidity. We contacted all authors by email for additional information when these data were not provided in the publications. On methodological grounds, we excluded studies that used data from identical study populations, based on identical samples, to avoid bias due to duplicate results.

Quality assessment

The methodological quality was scored using an adapted version of the Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses (online supplementary material S2) [23]. Quality assessment was based on three sets of criteria. First, the selection criteria, including representativeness of the cohort and included population, ascertainment of stress exposure and measurement of stress (prospective, retrospective). Secondly, the comparability criteria were used to investigate whether cohorts were comparable based on design or analyses by inclusion of predefined confounders, such as maternal smoking. Thirdly, the outcome criteria were applied, including the assessment of the outcome measure (asthma, wheezing) and adequacy of follow-up. Two independent authors (K.F.E. van de Loo and M.M.H.J. van Gelder) individually scored the quality assessment checklist for all included articles. Subsequently, the results were compared and discussed. Based on these discussions, the scales were adapted where necessary and were again independently scored, compared and discussed. A third author (C.M. Verhaak) was involved in the final discussion for an independent third opinion, which resulted in final quality ratings. Based on these ratings, the studies were grouped into low-quality (0–3), moderate-quality (4–6) and high-quality (7–8) studies.

Data analysis

To study the overall effect of prenatal maternal psychological stress on the development of any kind of respiratory morbidity in the child, we performed separate meta-analyses based on crude and adjusted measures. First, we performed an overall meta-analysis based on crude measures, calculated using dichotomous data on the number of pregnant women (stress versus no stress) and the number of children (asthma/wheezing versus no asthma/wheezing) from all included articles. Furthermore, separate meta-analyses (also using these dichotomous data) were performed for the effects of stress exposure and perceived stress, and for the outcome measures asthma and wheezing.

Secondly, we performed a meta-analysis based on the adjusted ORs with 95% confidence intervals (CIs) as mentioned in the original articles, regardless of which factors (for example, maternal age and smoking) were included in the adjusted analyses by the authors. If the adjusted OR was not mentioned, the article was not included in the adjusted meta-analysis. To analyse the results, the generic inverse variance method was used [20].

For all meta-analyses, to avoid bias by using one study sample more than once in each analysis, only one exposure and one outcome measure were included from each study. Regarding prenatal maternal psychological stress, if available, a composite score of multiple stress measures (multiple questionnaires) was used. When this was not possible, the score of only one questionnaire was entered in the meta-analysis. For example, when both an anxiety and a depression questionnaire were applied in a study, only one of these was entered in the meta-analysis. We choose to enter the measure that was most frequently used in the included studies, to make results more comparable. In this case, the anxiety measure was included in the meta-analysis as it was more frequently used than depression. The scores of the alternative measures (for example, depression) were entered exploratively to investigate whether these would have changed the results. As we used dichotomous measures, a differentiation was made between levels of stress exposure and/or perceived stress above versus below the clinical cut-off score.

For the overall effect on respiratory health, asthma was used when a study investigated both asthma and wheezing. When multiple types of wheezing were analysed, we divided the results into two groups: wheezing (sum of all different types) versus no wheezing. Where studies presented the results of multiple age groups, we used the oldest group for analysis.

All meta-analyses were performed using a random effect model according to the DerSimonian and Laird method [24] to calculate the pooled ORs with 95% CIs. Heterogeneity between the studies was investigated using I2 [25]. If heterogeneity was >50%, it was rated as substantial. Some heterogeneity was expected beforehand, due to clinical diversity (differences in populations, participants, measurements and outcomes) between the studies. The clinical diversity regarding, for example, psychological stress (which measurements) and respiratory morbidity (which measures, how they were assessed), was investigated by exploring the study characteristics. Additionally, subgroup analyses were performed based on quality scores to investigate methodological bias in both the crude and adjusted meta-analyses. Outliers were not excluded from the analyses to avoid bias, but the results with and without outliers were described. Visual inspection of a funnel plot was used to investigate publication bias when ≥10 studies were available. All analyses were performed using RevMan version 5.0 (http://www.tech.cochrane.org/revman).

Results

Study selection

The systematic searches of the databases yielded 2572 non-duplicated results of which 2175 articles were excluded based on title only. Reasons for exclusion were, for example, asthma or wheezing in the mother instead of the child, and animal studies. Good agreement was found between the authors for screening of the first 170 titles (κ=0.70). Initial disagreement did not lead to inclusion of any extra articles. The abstracts of the 397 selected articles were independently screened by two authors with “excellent agreement” (κ=0.83). This resulted in 24 full-text articles being screened for eligibility (κ=0.72; “good agreement”) and a total of 14 studies that were included based on the inclusion criteria (figure 1). The snowball method revealed no additional studies. Three out of 14 eligible publications were excluded from the meta-analysis because they studied the same cohorts [26–28]. One study was excluded because of insufficient data and unsuccessful contact with the authors [29].

FIGURE 1
FIGURE 1

Study selection.

Study characteristics

Table 1 shows a detailed description of the study characteristics. Eight studies were cohort studies, one used a case–control design [36] and one was a cross-sectional study [34]. The quality ratings based on the NOS ranged 5–7 on a scale of 0–8. Three (30%) studies were classified as high quality and seven (70%) were classified as moderate quality (online supplementary material S3). In total, the studies included 58 different potential confounders in the multivariable models used to calculate adjusted risk estimates (between three and 25 per study). However, most studies included all potential confounders in the multivariable models, irrespective of whether these confounded the association of interest or not. The most frequently included co-variables were: preterm birth, breastfeeding, race/ethnicity, parental asthma or allergy (all in four studies), maternal age (five studies), smoking during pregnancy (six studies), maternal educational level and sex of the child (both in eight studies).

View this table:
TABLE 1

Characteristics of the studies included in the meta-analysis

The psychological stress measures used in the included studies were diverse. Studies focused on stress exposure [27, 32], perceived stress [30, 31, 35–37] or both [33, 38]. Stress exposure was assessed in terms of experience of general life events (for example, any situation of loss or uneasiness) [32–34] or difficult life circumstances (such as financial strains) [38] in some studies, and as specific events (such as the loss of a spouse or child) [27] in other studies. Perceived stress was assessed in terms of general stress symptoms (for example, “I feel unhappy”) [30, 31, 35, 37, 38], anxiety (“I feel tense”) [33, 36] or depression (“I feel hopeless”) [33, 36], as well in relation to a specific source of stress, such as pregnancy-specific anxiety or stress (“I worry about the health of the baby”) [31, 38].

Five (50%) studies assessed wheezing as an outcome measure [30, 32, 36–38], while one study assessed asthma [27] and three (30%) studies assessed both [33–35]. One study assessed respiratory health in general as an outcome measure [31]. Wheezing was assessed at different ages in childhood, ranging 0–14 years (table 1 provides a more detailed description). Only one study provided results of wheezing at multiple ages [35]. Three studies reported results for the different types of wheezing: transient, late onset or persistent [35, 37], and single wheeze versus multiple wheeze [38]. Asthma was also assessed at different ages: at 6 [35] or 7.5 years of age [33], and at a mean±sd age of 8.5±3.2 years [34]. One study assessed asthma hospitalisation at different ages from birth and also performed an additional analysis on children from 6 years of age onwards [27].

Meta-analyses

Meta-analysis based on dichotomous measures

The overall meta-analysis using the numbers of children with asthma, wheezing and other respiratory morbidity in the groups who were exposed versus unexposed to prenatal stress showed a statistically significant result (OR 1.56 (95% CI 1.36–1.80), I2=18%, 10 studies) with a low level of heterogeneity (figure 2). The subgroup analyses based on quality scores also revealed statistically significant results for the moderate-quality group (OR 1.54 (95% CI 1.30–1.84), I2=0%, seven studies) and the high-quality group (OR 1.77 (95% CI 1.18–2.67), I2=74%, three studies) separately. The level of heterogeneity in the high-quality group indicates a large degree of variability, partly due to the small number of studies included. There was no statistically significant difference between the high- and moderate-quality group (p=0.54; I2=0%), suggesting a negligible chance of methodological bias. The funnel plot revealed a slight asymmetry, indicating some evidence of indirect publication bias (figure 3).

FIGURE 2
FIGURE 2

Overall meta-analysis (dichotomous measures, random effects model) of the association between any form of prenatal maternal psychological stress and respiratory morbidity in childhood, stratified according to study quality. #: heterogeneity: Tau2=0.09, Chi-squared=7.59, df=2 (p=0.02), I2=74%; test for overall effect: Z=2.75 (p=0.006). : heterogeneity: Tau2=0.00, Chi-squared=3.15, df=6 (p=0.79), I2=0%; test for overall effect: Z=4.89 (p<0.00001). +: heterogeneity: Tau2=0.01, Chi-squared=10.94, df=9 (p=0.28), I2=18%; test for overall effect: Z=6.21 (p<0.00001); test for subgroup differences: Chi-squared: 0.37, df=1 (p=0.54), I2=0%.

FIGURE 3
FIGURE 3

Studies included in the overall meta-analysis stratified by study quality.

Stress exposure versus perceived stress

Subgroup analyses revealed statistically significant effects of stress exposure during pregnancy (OR 1.58 (95% CI 1.26–1.99), I2=34%, four studies), as well as perceived stress (OR 1.59 (95% CI 1.29–1.96), I2=18%, six studies) on respiratory morbidity in the child (online supplementary material S4).

Asthma or wheezing as an outcome measure

Subgroup analyses also showed that children of mothers who experienced some form of stress during pregnancy were more often diagnosed with asthma compared with children of mothers who did not experience stress (OR 1.45 (95% CI 1.25–1.68), I2=13%, four studies) (online supplementary material S5). In addition, wheezing complaints were more often found among children whose mothers experienced psychological stress during pregnancy compared with mothers who did not (OR 1.87 (95% CI 1.42–2.45), I2=76%, eight studies) (online supplementary material S6). A substantial level of heterogeneity was present in the latter subgroup analysis. Visual inspection of the forest plot revealed one potential outlier (OR 3.22 (95% CI=2.38–4.35)) [35]. Exclusion of this outlier decreased the overall OR to 1.57 (95% CI 1.33–1.84) and the I2 to 25%.

Adjusted meta-analysis

The meta-analysis of the adjusted ORs as reported in the articles also showed that the prevalence of asthma, wheezing and other respiratory morbidity was higher in children whose mothers experienced some form of psychological stress during pregnancy compared with mothers who did not (OR 1.59 (95% CI 1.25–2.01), I2=68%, eight studies) (figure 4). The OR was similar to the OR from the crude meta-analysis, but the level of heterogeneity was substantial. The potential confounders that were included in the individual studies differed greatly in number and type of confounder (table 1). Subgroup analysis based on quality scores, which were partly based on inclusion of the most important potential confounders (online supplementary material S2), revealed no statistically significant difference between the high- and moderate-quality groups (p=0.58). The results were statistically significant for the moderate-quality group (OR 1.51 (95% CI 1.11–2.05), I2=68%, five studies) and the high-quality group (OR 1.71 (95% CI 1.25–2.33), I2=32%, three studies), separately, despite the substantial level of heterogeneity in the moderate-quality group.

FIGURE 4
FIGURE 4

Overall adjusted meta-analysis (random effects model) using the generic inverse variance method on the association between any form of prenatal maternal psychological stress and respiratory morbidity in childhood, stratified according to study quality. #: heterogeneity: Tau2=0.03, Chi-squared=2.93, df=2 (p=0.23), I2=32%; test for overall effect: Z=3.35 (p=0.0008). : heterogeneity: Tau2=0.06; Chi-squared=12.39, df=4 (p=0.01), I2=68%; test for overall effect: Z=2.61 (p=0.009). +: heterogeneity: Tau2=0.06, Chi-squared=21.59, df=7 (p=0.003), I2=68%; test for overall effect: Z=3.83 (p=0.0001); test for subgroup differences: Chi-squared=0.31, df=1 (p=0.58), I2=0%.

Timing of exposure

The exposure to prenatal stress was measured in different time periods (different trimesters, for example) and covered different time frames (the whole pregnancy or a single trimester) (table 1). Subgroup analyses on timings of the exposure were not possible due to this heterogeneity and the small number of studies. Only one study assessed the effect in the three trimesters of pregnancy separately, and found that the RRs of asthma related to exposure to maternal bereavement were elevated in each trimester, but were not statistically significant [27]. However, due to the very small number of exposed cases in each trimester, it was impossible to tell whether there were truly no effects or whether this was because of a lack of study power. Another study assessed the second and third trimester separately, and found evidence of a dose–response relationship for maternal anxiety symptoms reported at 32 weeks of pregnancy compared to 18 weeks, with larger effect sizes for non-atopic compared with atopic asthma [33]. Some studies assessed exposure any time during pregnancy [34, 36], or in the second [35] or third trimester only [31], while one study measured exposure during the third trimester as stress during the past year [37]. One study assessed the second trimester [35]. The remaining studies assessed the results of exposure in different trimesters together: the first and third trimester [30], or the second and third trimester [32, 38].

Discussion

This is the first systematic review and meta-analysis of the association between prenatal maternal psychological stress and respiratory morbidity in the child. Both the meta-analysis based on crude data and the adjusted meta-analysis showed that prenatal psychological stress is associated with an increased risk of respiratory morbidity in the child. This association was also observed for the separate stress measures, stress exposure and perceived stress, and the individual outcome measures of asthma and wheezing, in both high- and moderate-quality studies. Literature regarding the influence of prenatal psychological stress on the development of respiratory morbidity in the child was searched from 1960 onwards, but the first article on this subject was only published in 2009. Between 2009 and the end of 2013, an additional 13 studies on this topic were published. This underscores the increasing interest in this new topic of research.

Before interpreting the results, it is important to discuss the strengths and limitations of this study. One limitation was the relatively small number of studies that could be included. By performing a meta-analysis, however, small but clinically relevant effects could be detected and a global hypothesis could be formed because of the heterogeneous groups [39]. Because the majority of studies used ORs as effect measures, we also calculated ORs in our meta-analyses. However, a sensitivity analysis in which we calculated RRs for the raw data using the numbers in figure 2, indicated that the ORs in the original studies and in this meta-analysis may overestimate the true effect sizes, due to the relatively frequently occurring outcomes, especially in some studies of moderate quality [36–38]. However, the ORs observed may underestimate the true effect sizes in some studies, because of non-differential misclassification in exposure and outcome variables. Although the RR from the overall meta-analysis (RR 1.38 (95% CI 1.23–1.54)) was somewhat lower than the overall OR (OR 1.56 (95% CI 1.36–1.80)), the result remained statistically significant and does not change the interpretation.

Because we performed separate meta-analyses, based on the crude and the adjusted-effect measures, the results were investigated with and without adjustment for confounders. Although the crude and adjusted-effect measures differed considerably within individual studies, the results of the meta-analyses were remarkably similar, with only slightly larger confidence intervals in the adjusted analyses. However, as all studies included different numbers and types of potential confounders and most did not separate out true confounders, conclusions about the influence of specific confounders could not be drawn. Inferences about indirect effects, such as for smoking or breastfeeding [40], could not be made either. The inclusion of potential confounders was also taken into account in the quality assessment, but the results of the meta-analyses showed no statistically significant differences between the moderate- and high-quality studies.

As dichotomous variables were used in the overall meta-analysis, it was not possible to obtain specific insights into the influence of the quantity and/or quality of stress experienced, or to assess a dose–response relationship. The amount of stress had to be above a cut-off score or in the 4th quartile for women to be classified as exposed, so only serious levels of stress were included in the exposure–outcome associations.

Visual inspection of the funnel plot showed some indication of publication bias. Moderate and high levels of asymmetry are common in meta-analyses with small study sizes, even in the absence of publication bias, and vice versa [41]. So there may be many reasons for the slight asymmetry observed in the funnel plot: publication bias, location bias, true heterogeneity, data irregularities, artefacts or chance [42]. It is possible that smaller studies without statistically significant effects remained unpublished. However, the quality assessment indicated that all studies included were of moderate or high quality, and the results showed no differences in the outcomes between the quality scores. Therefore, there seemed to be no indication of methodological bias. One study was excluded based on insufficient data and unsuccessful contact with the authors [29]. Given the size of the study population (n=66 203), inclusion of this study would have changed the overall results somewhat, but not drastically.

Before we performed this meta-analysis, we expected some clinical diversity among the different studies. This was confirmed by substantial levels of heterogeneity in the adjusted overall meta-analysis and in several subgroup meta-analyses. The levels of heterogeneity observed could be explained by true heterogeneity among populations regarding the prevalence of asthma and wheezing [2, 6], or may be due to heterogeneity in different types and measurements of respiratory health, timing of exposure, age of the child, confounders adjusted for, or the small numbers of studies included. Exclusion of one study from the sub-analysis on wheezing revealed low levels of heterogeneity among the remaining studies without substantial consequences for the effect measure.

Heterogeneity in definitions and measurements of psychological stress and respiratory morbidity hampered straightforward interpretation of the results. The psychological stress results showed that the studies included used different conceptualisations and different means of assessment. By exploring “stress” we made a distinction between stress exposure (for example, experiencing a stressful event, which could be major or minor) and perceived stress (experiencing stress symptoms, such as feeling tense), as exposure to a stressor does not necessarily result in perceived stress [43]. Perceived stress is dependent on individual differences in dealing with the stressor, due to individual differences in perceived social support and/or coping styles, for example [44, 45]. By making this distinction, some important critical notes need to be addressed. First, the conceptualisations of perceived stress in the included studies are rather different, ranging from more global measures of stress (perceiving more or fewer symptoms of stress) to psychological dysfunction (normal versus pathological stress). Regarding the subgroup analysis on perceived stress, for example, one study assessed stress as a general measure of experiencing distress [35], while another assessed a diagnosis of anxiety or depression disorder [36]. Comparisons between more global measures of distress and assessment of psychological dysfunction could be interesting, but were not possible in the present study due to the small sample sizes.

Secondly, different types of stress could yield different physiological responses [44]. In cases of acute stress (such as a life event), the physiological stress response could be protective and adaptive to maintain homeostasis, but could also be over-activated when stress is chronic; the latter results in allostatic overload and could have damaging effects [46, 47]. As this meta-analysis only encompassed stress exposure and perceived stress, specific conclusions about the distinction between stress exposure, perceived stress, psychological dysfunction or physiological reactions could not be drawn.

In addition to the differences in the concept of stress, a wide range of measurements are available for each definition, which complicates comparisons between studies [48]. As a result, conclusions about specific forms of stress exposure or perceived stress could not be drawn in this study. However, despite the heterogeneity in concepts and assessment methods used, the meta-analyses consistently revealed that prenatal psychological stress, and both stress exposure and perceived stress, are associated with adverse respiratory health outcomes in the child. This is in line with other studies regarding the negative impact of prenatal maternal stress on adverse health outcomes in the child [15, 16, 49–54].

Both asthma and wheezing were studied for child respiratory health. Recent literature shows that asthma is seen as a heterogeneous phenotype that originates in early life [55, 56]. A doctor's diagnosis of asthma is considered the most reliable and robust measure. In the studies included, asthma was often reported through a questionnaire completed by the child's mother [33, 35] or through a patient register [27]. One study asked a single question only, without referring to a doctor's diagnosis [34]. Wheezing has been described in different subtypes: transient early, late-onset, and persistent wheezing [57]. In the majority of children, wheezing is a transient condition in the first 3 years of life and is associated with a diminished airway function. For a small group of children, wheezing is a precursor for asthma. In the studies included, wheezing was measured in multiple ways: some made a distinction between the subtypes described above [35, 37], while others measured repeated wheeze as more than one or two [38], two [32], or three episodes [36], by one single question [34] or through maternal report of a doctor's diagnosis [30, 33]. The differences in diagnostic assessment measures were taken into account in the quality assessment ratings. As differences in methods used for diagnoses are a well-known problem, some protocols were developed to fill this gap, such as the International Study of Asthma and Allergies in Children (ISAAC) initiative [2, 58]. For wheezing, an international consensus group of the European Respiratory Society (ERS) task force classified preschool wheezing and preferred pharmacological treatment [59]. The application of standards is clearly required for future studies to reduce heterogeneity and avoid underestimation of effects.

This meta-analysis supports the hypothesis that prenatal maternal psychological stress has a negative influence on respiratory health in the offspring. However, the underlying mechanisms, such as how the psychophysiological stress responses from the mother are transferred to the child, are only partly understood. The following is hypothesised. In utero, lung development passes through several critical periods, starting around 26 days of gestation and following multiple stages [60]. It is known that prenatal stress has an influence on the programming of the immune function, which starts early in pregnancy when the stem cells are formed [61]. Prenatal maternal stress, acting through neuroendocrine, immune/inflammatory and/or vascular pathways, may affect lung maturation, growth and function [52, 53, 62]. More specifically, maternal stress could disturb the balance between the interrelated neuroendocrine system, the ANS and the immune system, thereby leading to impaired lung function and immunity in the child [18]. It is conceivable that the HPA axis and the ANS are involved in this complex mechanism [18]. Furthermore, it is hypothesised that gene–environment interactions play an important role in the development of asthma [10], suggesting that multiple factors, both genetic and non-genetic, are involved.

For future prospective studies, we recommend the specification of outcome measures, such as different types of wheezing as well as asthma [59], and the incorporation of various types of stress exposure, perceived stress, psychological dysfunction and physiological reactions. Furthermore, the role of mediating factors (for example, smoking and exposure to air pollution) needs to be explored. Future research should also focus on the effects of prenatal maternal stress at different time points in pregnancy, relating to critical developmental periods, as well as on dose–response relationships. Finally, it would be interesting to investigate the various mechanisms by which different types of prenatal maternal stress may influence in utero development of the respiratory system. Although this meta-analysis provides evidence for a negative impact of maternal stress during pregnancy on the development of the child, the clinical implications are still unclear.

In conclusion, this study shows that prenatal psychological stress plays a negative role in the development of asthma and wheezing in offspring. Based on this systematic review and meta-analysis, future studies examining the early origins of asthma need to account for the impact of prenatal maternal psychological stress.

Acknowledgements

We would like to acknowledge A.R.T. Donders (Department for Health Evidence, Radboudumc, Nijmegen, The Netherlands) for statistical advice in performing the meta-analyses.

Footnotes

  • This article has supplementary material available from erj.ersjournals.com

  • Support statement: This study was supported by the Lung Foundation Netherlands (grant no. 3.4.10.007). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

  • Conflict of interest: None declared.

  • Received February 21, 2015.
  • Accepted August 17, 2015.

References

  1. ↵
    1. Braman SS
    . The global burden of asthma. Chest 2006; 130: Suppl. 1, S4–S12.
    OpenUrlCrossRefPubMedWeb of Science
  2. ↵
    1. Asher MI,
    2. Montefort S,
    3. Bjorksten B, et al.
    Worldwide time trends in the prevalence of symptoms of asthma, allergic rhinoconjunctivitis, and eczema in childhood: ISAAC Phases One and Three repeat multicountry cross-sectional surveys. Lancet 2006; 368: 733–743.
    OpenUrlCrossRefPubMedWeb of Science
  3. ↵
    1. Masoli M,
    2. Fabian D,
    3. Holt S, et al.
    The global burden of asthma: executive summary of the GINA Dissemination Committee report. Allergy 2004; 59: 469–478.
    OpenUrlCrossRefPubMedWeb of Science
  4. ↵
    1. Henderson J,
    2. Granell R,
    3. Heron J, et al.
    Associations of wheezing phenotypes in the first 6 years of life with atopy, lung function and airway responsiveness in mid-childhood. Thorax 2008; 63: 974–980.
    OpenUrlAbstract/FREE Full Text
  5. ↵
    1. Henderson AJ,
    2. Warner JO
    . Fetal origins of asthma. Semin Fetal Neonatal Med 2012; 17: 82–91.
    OpenUrlCrossRefPubMed
  6. ↵
    1. Patel SP,
    2. Jarvelin MR,
    3. Little MP
    . Systematic review of worldwide variations of the prevalence of wheezing symptoms in children. Environ Health 2008; 7: 57.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Barker DJ
    . In utero programming of chronic disease. Clin Sci 1998; 95: 115–128.
    OpenUrlCrossRefPubMed
  8. ↵
    1. Capra L,
    2. Tezza G,
    3. Mazzei F, et al.
    The origins of health and disease: the influence of maternal diseases and lifestyle during gestation. Ital J Pediatr 2013; 39: 7.
    OpenUrlCrossRefPubMed
  9. ↵
    1. Harding R,
    2. Maritz G
    . Maternal and fetal origins of lung disease in adulthood. Semin Fetal Neonatal Med 2012; 17: 67–72.
    OpenUrlCrossRefPubMedWeb of Science
  10. ↵
    1. Duijts L
    . Fetal and infant origins of asthma. Euro J Epidemiol 2012; 27: 5–14.
    OpenUrlCrossRef
    1. Anto JM
    . Recent advances in the epidemiologic investigation of risk factors for asthma: a review of the 2011 literature. Curr Allergy Asthma Rep 2012; 12: 192–200.
    OpenUrlCrossRefPubMed
    1. Burke H,
    2. Leonardi-Bee J,
    3. Hashim A, et al.
    Prenatal and passive smoke exposure and incidence of asthma and wheeze: systematic review and meta-analysis. Pediatrics 2012; 129: 735–744.
    OpenUrlAbstract/FREE Full Text
    1. Narkowicz S,
    2. Plotka J,
    3. Polkowska Z, et al.
    Prenatal exposure to substance of abuse: a worldwide problem. Environ Int 2013; 54: 141–163.
    OpenUrlCrossRefPubMedWeb of Science
  14. ↵
    1. O'Reilly JR,
    2. Reynolds RM
    . The risk of maternal obesity to the long-term health of the offspring. Clin Endocrinol 2013; 78: 9–16.
    OpenUrlCrossRefPubMed
  15. ↵
    1. Glover V
    . Maternal depression, anxiety and stress during pregnancy and child outcome; what needs to be done. Best Pract Res Clin Obstet Gynaecol 2014; 28: 25–35.
    OpenUrlCrossRef
  16. ↵
    1. Beydoun H,
    2. Saftlas AF
    . Physical and mental health outcomes of prenatal maternal stress in human and animal studies: a review of recent evidence. Paediatr Perinat Epidemiol 2008; 22: 438–466.
    OpenUrlCrossRefPubMedWeb of Science
  17. ↵
    1. Virk J,
    2. Li J,
    3. Vestergaard M, et al.
    Prenatal exposure to bereavement and type-2 diabetes: a Danish longitudinal population based study. PLoS One 2012; 7: e43508.
    OpenUrlCrossRefPubMed
  18. ↵
    1. Wright RJ
    . Perinatal stress and early life programming of lung structure and function. Biol Psychol 2010; 84: 46–56.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Wright RJ
    . Prenatal maternal stress and early caregiving experiences: implications for childhood asthma risk. Paediatr Perinat Epidemiol 2007; 21: Suppl. 3, 8–14.
    OpenUrl
  20. ↵
    1. Higgins JPT,
    2. Green S
    . Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0; updated March 2011. www.cochrane.handbook.org Date last updated: March 2011. Date last accessed: June 10, 2015.
  21. ↵
    1. Moher D,
    2. Liberati A,
    3. Tetzlaff J, et al.
    Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009; 339: b2535.
    OpenUrlFREE Full Text
  22. ↵
    1. Orwin RG
    . Evaluating Coding Decisions. In: Cooper H, Hedges LV, eds. The Handbook of Research Synthesis. New York, Russell Sage Foundation, 1994; pp. 150–151.
  23. ↵
    1. Wells GA,
    2. Shea B,
    3. O'Connell D,
    4. Peterson J, et al.
    The Newcastle-Ottawa Scale (NOS) for assessing the quality if nonrandomized studies in meta-analyses. www.ohri.ca/programs/clinical_epidemiology/oxford.asp Date last updated: October 19, 2009. Date last accessed: June 10, 2015.
  24. ↵
    1. DerSimonian R,
    2. Laird N
    . Meta-analysis in clinical trials. Control Clin Trials 1986; 7: 177–188.
    OpenUrlCrossRefPubMedWeb of Science
  25. ↵
    1. Higgins JP,
    2. Thompson SG,
    3. Deeks JJ, et al.
    Measuring inconsistency in meta-analyses. BMJ 2003; 327: 557–560.
    OpenUrlFREE Full Text
  26. ↵
    1. Chiu YH,
    2. Coull BA,
    3. Sternthal MJ, et al.
    Effects of prenatal community violence and ambient air pollution on childhood wheeze in an urban population. J Allergy Clin Immunol 2014; 133: 713–722.
    OpenUrlCrossRef
  27. ↵
    1. Khashan AS,
    2. Wicks S,
    3. Dalman C, et al.
    Prenatal stress and risk of asthma hospitalization in the offspring: a Swedish population-based study. Psychosom Med 2012; 74: 635–641.
    OpenUrlAbstract/FREE Full Text
  28. ↵
    1. Sternthal MJ,
    2. Coull BA,
    3. Chiu YHM, et al.
    Associations among maternal childhood socioeconomic status, cord blood IgE levels, and repeated wheeze in urban children. J Allergy Clin Immunol 2011; 128: 337–345.
    OpenUrlCrossRefPubMedWeb of Science
  29. ↵
    1. Tegethoff M,
    2. Greene N,
    3. Olsen J, et al.
    Stress during pregnancy and offspring pediatric disease: a national cohort study. Environ Health Perspect 2011; 119: 1647–1652.
    OpenUrlCrossRefPubMedWeb of Science
  30. ↵
    1. Alton ME,
    2. Tough SC,
    3. Mandhane PJ, et al.
    Street drug use during pregnancy: potential programming effects on preschool wheeze. J Dev Orig Health Dis 2013; 4: 191–199.
    OpenUrlCrossRefPubMed
  31. ↵
    1. Beijers R,
    2. Jansen J,
    3. Riksen-Walraven M, et al.
    Maternal prenatal anxiety and stress predict infant illnesses and health complaints. Pediatrics 2010; 126: E401–E409.
    OpenUrlCrossRefPubMed
  32. ↵
    1. Chiu YHM,
    2. Coull BA,
    3. Cohen S, et al.
    Prenatal and postnatal maternal stress and wheeze in urban children: effect of maternal sensitization. Am J Respir Crit Care Med 2012; 186: 147–154.
    OpenUrlCrossRefPubMed
  33. ↵
    1. Cookson H,
    2. Granell R,
    3. Joinson C, et al.
    Mothers’ anxiety during pregnancy is associated with asthma in their children. J Allergy Clin Immunol 2009; 123: 847–853.
    OpenUrlCrossRefPubMedWeb of Science
  34. ↵
    1. De Marco R,
    2. Pesce G,
    3. Girardi P, et al.
    Foetal exposure to maternal stressful events increases the risk of having asthma and atopic diseases in childhood. Pediatr Allergy Immunol 2012; 23: 724–729.
    OpenUrlCrossRefPubMed
  35. ↵
    1. Guxens M,
    2. Sonnenschein-van der Voort AM,
    3. Tiemeier H, et al.
    Parental psychological distress during pregnancy and wheezing in preschool children: the Generation R Study. J Allergy Clin Immunol 2014; 133: 59–67.
    OpenUrlCrossRefWeb of Science
  36. ↵
    1. Lefevre F,
    2. Moreau D,
    3. Semon E, et al.
    Maternal depression related to infant's wheezing. Pediatr Allergy Immunol 2011; 22: 608–613.
    OpenUrlCrossRefPubMed
  37. ↵
    1. Reyes M,
    2. Perzanowski MS,
    3. Whyatt RM, et al.
    Relationship between maternal demoralization, wheeze, and immunoglobulin E among inner-city children. Ann Allergy Asthma Immunol 2011; 107: 42–49.
    OpenUrlCrossRefPubMedWeb of Science
  38. ↵
    1. Wood RA,
    2. Bloomberg GR,
    3. Kattan M, et al.
    Relationships among environmental exposures, cord blood cytokine responses, allergy, and wheeze at 1 year of age in an inner-city birth cohort (Urban Environment and Childhood Asthma study). J Allergy Clin Immunol 2011; 127: 913–919.
    OpenUrlCrossRef
  39. ↵
    1. Chiappelli F,
    2. Brant XM,
    3. Oluwadara OO, et al.
    Introduction: Research Synthesis in Evidence-Based Clinical Decision-Making. Berlin, Springer, 2010; pp. 3–16.
  40. ↵
    1. Wright RJ,
    2. Cohen S,
    3. Carey V, et al.
    Parental stress as a predictor of wheezing in infancy: a prospective birth-cohort study. Am J Respir Crit Care Med 2002; 165: 358–365.
    OpenUrlCrossRefPubMedWeb of Science
  41. ↵
    1. Terrin N,
    2. Schmid CH,
    3. Lau J
    . In an empirical evaluation of the funnel plot, researchers could not visually identify publication bias. J Clin Epidemiol 2005; 58: 894–901.
    OpenUrlCrossRefPubMedWeb of Science
  42. ↵
    1. Egger M,
    2. Davey Smith G,
    3. Schneider M, et al.
    Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 629–634.
    OpenUrlAbstract/FREE Full Text
  43. ↵
    1. Atkinson RL,
    2. Atkinson RC,
    3. Smith EE, et al.
    Hilgards’ introduction to psychology. 13th Edn. Hartcourt Inc, Fort Worth 2000.
  44. ↵
    1. Graignic-Philippe R,
    2. Dayan J,
    3. Chokron S, et al.
    Effects of prenatal stress on fetal and child development: a critical literature review. Neurosci Biobehav Rev 2014; 43: 137–162.
    OpenUrlCrossRefPubMed
  45. ↵
    1. Lazarus RSSF
    . Stress, Appraisal and the Coping Process. New York, Springer, 1984.
  46. ↵
    1. McEwen BS
    . Stressed or stressed out: what is the difference? J Psychiatry Neurosci 2005; 30: 315–318.
    OpenUrlPubMedWeb of Science
  47. ↵
    1. Wright RJ
    . Stress and atopic disorders. J Allergy Clin Immunol 2005; 116: 1301–1306.
    OpenUrlCrossRefPubMedWeb of Science
  48. ↵
    1. Nast I,
    2. Bolten M,
    3. Meinlschmidt G, et al.
    How to measure prenatal stress? A systematic review of psychometric instruments to assess psychosocial stress during pregnancy. Paediatr Perinat Epidemiol 2013; 27: 313–322.
    OpenUrlCrossRefPubMed
  49. ↵
    1. Huizink AC,
    2. De Medina PR,
    3. Mulder EJH, et al.
    Prenatal maternal stress, HPA axis activity, and postnatal infant development. Int Congr Ser 2002; 1241: 65–71.
    OpenUrlCrossRef
    1. Van den Bergh BR,
    2. Mulder EJ,
    3. Mennes M, et al.
    Antenatal maternal anxiety and stress and the neurobehavioural development of the fetus and child: links and possible mechanisms. A review. Neurosci Biobehav Rev 2005; 29: 237–258.
    OpenUrlCrossRefPubMedWeb of Science
    1. Wadhwa PD
    . Psychoneuroendocrine processes in human pregnancy influence fetal development and health. Psychoneuroendocrinology 2005; 30: 724–743.
    OpenUrlCrossRefPubMedWeb of Science
  52. ↵
    1. Wadhwa PD,
    2. Culhane JF,
    3. Rauh V, et al.
    Stress and preterm birth: neuroendocrine, immune/inflammatory, and vascular mechanisms. Matern Child Health J 2001; 5: 119–125.
    OpenUrlCrossRefPubMed
  53. ↵
    1. Wadhwa PD,
    2. Culhane JF,
    3. Rauh V, et al.
    Stress, infection and preterm birth: a biobehavioural perspective. Paediatr Perinat Epidemiol 2001; 15: Suppl. 2, 17–29.
    OpenUrlCrossRefPubMedWeb of Science
  54. ↵
    1. Weinstock M
    . The potential influence of maternal stress hormones on development and mental health of the offspring. Brain Behav Immun 2005; 19: 296–308.
    OpenUrlCrossRefPubMedWeb of Science
  55. ↵
    1. Cottini M,
    2. Asero R
    . Asthma phenotypes today. Eur Ann Allergy Clin Immunol 2013; 45: 17–24.
    OpenUrl
  56. ↵
    1. Bisgaard H,
    2. Jensen SM,
    3. Bonnelykke K
    . Interaction between asthma and lung function growth in early life. Am J Respir Crit Care Med 2012; 185: 1183–1189.
    OpenUrlCrossRefPubMedWeb of Science
  57. ↵
    1. Martinez FD,
    2. Wright AL,
    3. Taussig LM, et al.
    Asthma and wheezing in the first six years of life. The Group Health Medical Associates. N Engl J Med 1995; 332: 133–138.
    OpenUrlCrossRefPubMedWeb of Science
  58. ↵
    1. Asher MI,
    2. Keil U,
    3. Anderson HR, et al.
    International Study of Asthma and Allergies in Childhood (ISAAC): rationale and methods. Eur Respir J 1995; 8: 483–491.
    OpenUrlAbstract
  59. ↵
    1. Brand PL,
    2. Caudri D,
    3. Eber E, et al.
    Classification and pharmacological treatment of preschool wheezing: changes since 2008. Eur Respir J 2014; 43: 1172–1177.
    OpenUrlAbstract/FREE Full Text
  60. ↵
    1. Merkus PJ,
    2. ten Have-Opbroek AA,
    3. Quanjer PH
    . Human lung growth: a review. Pediatr Pulmonol 1996; 21: 383–397.
    OpenUrlCrossRefPubMedWeb of Science
  61. ↵
    1. Veru F,
    2. Laplante DP,
    3. Luheshi G, et al.
    Prenatal maternal stress exposure and immune function in the offspring. Stress 2014; 17: 133–148.
    OpenUrlCrossRefPubMed
  62. ↵
    1. Pinkerton KE,
    2. Joad JP
    . The mammalian respiratory system and critical windows of exposure for children's health. Environ Health Perspect 2000; 108: Suppl. 3, 457–462.
    OpenUrlCrossRefPubMedWeb of Science
View Abstract
Back to top
View this article with LENS
Email
Print
Citation Tools

Prenatal maternal psychological stress and childhood asthma and wheezing: a meta-analysis
Kim F.E. van de Loo, Marleen M.H.J. van Gelder, Jolt Roukema, Nel Roeleveld, Peter J.F.M. Merkus, Christianne M. Verhaak
European Respiratory Journal Jan 2016, 47 (1) 133-146; DOI: 10.1183/13993003.00299-2015
del.icio.us logo Technorati logo Twitter logo Connotea logo Facebook logo Mendeley logo
Full Text (PDF)

Jump To

Subjects