Discussion
To our knowledge, this is the first prospective, population-based,
birth-cohort study considering wheeze phenotypes during early to
mid-childhood in relation to stress assessed as life events at different
time points during perinatum. Prior studies rarely distinguished among
longitudinal phenotypes of wheeze and typically measured stress
indirectly as psychological distress. The present study documents that
early exposure to stressful events is linked to wheezing. The timing of
stress exposure matters for which wheeze phenotype likely develops. The
likelihood of wheeze with late onset increases with stress exposure
during pregnancy. Stress exposure during the first six months after
birth, in contrast, predicts persistent wheeze, which manifests by age
three years and persists until age seven years or longer. In prior
literature, persistent wheeze showed a relationship with poor lung
function and chronic asthma38, warranting better
understanding of this particular wheeze phenotype.
Specifying pathways leading from maternal stress to child wheezing
represents a contribution since these pathways were rarely examined in
prior literature. Perhaps without surprise is the finding that women who
experience more stress during pregnancy also tend to experience more
stress after delivery, which accounts in part for the link between
prenatal stress and wheezing. More intriguing is the finding that
mediation through postnatal stress is partial only, leaving an
independent effect of prenatal stress of a substantial magnitude, with
the increase in relative risk ranging 44%-69%. We examined several
additional mechanisms from perinatal stress to wheezing. LRTI diagnosis
emerged as a mediator, suggesting that children exposed to more
postnatal stress are more likely to develop LRTIs. This accounts for
some, but not for all, of the increase in wheezing among children of
mothers exposed to more postnatal life events. In other words, the
significant association between early-life stress exposure and wheezing
persists after adjusting for LRTI and covariates.
Multivariate modelling strengthens the evidence of relationships
independent of confounders, but causal interpretation is not warranted
with observational data. Nevertheless, using rigorous statistical
controls and mediation analysis, we rule out several alternative
mechanisms. This is an important contribution since mechanisms for the
stress gradient in wheezing are under-explored. Maternal emotional
dysregulation, for instance, is a candidate mechanism. It is commonly
preceded by stress exposure. Literature suggests that maternal emotional
difficulties are linked to child respiratory
problems3. Other mechanisms warranting future research
include neuro-endocrine-immune processes involved in airway functioning.
Childhood wheeze is commonly discussed in literature on allergic asthma.
Allergic asthma bronchiale is defined by allergic eosinophil
inflammation with bronchial obstruction. Wheeze is its symptom but it is
also a broader nosologic unit with heterogeneous pathogenic origins.
Factors triggering wheeze are not limited to allergic reaction; they
also include infection, irritation, and sudden change in ambient
temperature. Prior analysis of ELSPAC-CZ found that children with
persistent wheeze had higher risk of allergic asthma during school
age38. Surprising, our supplementary analyses
indicated that allergic asthma diagnosis at ages five and fifteen years
did not relate to prenatal or postnatal life events in adjusted models
(five years, high vs. low prenatal life events odds ratio [OR]=2.63,
95% CI 0.74-9.34, p=0.113; five years, high vs. low postnatal life
events OR=0.71, 95% CI 0.22-2.26, p=0.653; fifteen years, high vs. low
prenatal life events OR=1.01, 95% CI 0.59-1.73, p=0.938; fifteen years,
high vs. low postnatal life events OR=0.99, 95% CI 0.57-1.73, p=0.975).
Moreover, neither prenatal nor postnatal life events were associated
with physician-diagnosed allergy among children ages 0-18 months, as
indicated in Figure 3. We speculate that early-life stress is important
for the types of wheeze unrelated to allergic asthma. Our data did not
distinguish between allergic vs. non-allergic wheeze; mothers simply
reported whether their child experienced any wheezing episodes. Yet, the
observation that LRTI partially mediated between life events and wheeze
suggests the involvement of respiratory infections, potentially
accounting for some cases of wheeze unrelated to allergy, though this
explanation is tentative.
Strengths of this study include prospective design with a large sample
of mother-child dyads with fully completed questionnaires (n=1,849);
accounting for multiple mechanisms and confounders; using established
methodology to categorize wheeze phenotypes28, and
separately considering stress exposures during prenatal vs. postnatal
periods. Both periods are developmentally sensitive but they differ in
the types of developmental processes taking place. Assessment of life
events was quite broad with 42 different types of events, yielding a
more comprehensive picture of the overall stress levels compared to more
narrowly-focused instruments.
Methodological limitations must be noted. The cohort included mostly
young gravidas with uncomplicated pregnancies. Mean age at delivery
(25.2 years) and the caesarean section rate (8.1%) roughly correspond
to the demographic profile of the Czech population at the time of the
study39, suggesting good generalizability except for
older and complicated gravidas. Generalizability to societies outside of
the Czech Republic, however, is unknown. Child wheeze was
mother-reported. Maternal reports are routinely used in epidemiological
surveys of child health40. They are cost-effective and
capture symptoms known only to the mother, such as wheezing episodes for
which healthcare was not obtained. Potential biases include under-report
(child wheezed; mother reported no wheezing), over-report (child did not
wheeze; mother reported wheezing), and misreport (child wheezed; mother
reported another condition). Finally, biomarkers quantifying biological
mechanisms between early-life stress and wheeze were not available.