Associations between maternal mental health and early child wheezing in a South African birth cohort

Abstract Background Wheezing in early childhood is common and has been identified in high‐income countries (HIC) as associated with maternal antenatal or postnatal psychosocial risk factors. However, the association between maternal mental health and childhood wheezing has not been well studied in low and middle‐income countries (LMIC), such as South Africa. Methods A total of 1137 pregnant women over 18 year old, between 20 and 28 weeks’ gestation, and attending either of two catchment area clinics were enrolled in a South African parent study, the Drakenstein Child Health Study (DCHS). Psychosocial risk factors including maternal depression, psychological distress, early adversity, and intimate partner violence (IPV), were measured antenatally and postnatally by validated questionnaires. Two outcomes were evaluated: Presence of wheeze (at least one episode of child wheeze during the first 2 years of life); and recurrent wheeze (two or more episodes of wheezing in a 12‐month period). Logistic regression was used to investigate the association between antenatal or postnatal psychosocial risk factors and child wheeze, adjusting for clinical and socio‐demographic covariates. Results Postnatal psychological distress and IPV were associated with both presence of wheeze (adjusted OR = 2.09, 95%CI: 1.16‐3.77 and 1.63, 95%CI: 1.13‐2.34, respectively), and recurrent child wheeze (adjusted OR = 2.26, 95%CI: 1.06‐4.81 and 2.20, 95%CI: 1.35‐3.61, respectively). Conclusion Maternal postnatal psychological distress and IPV were associated with wheezing in early childhood. Thus, screening and treatment programs to address maternal psychosocial risk factors may be potential strategies to reduce the burden of childhood wheeze in LMICs.


| INTRODUCTION
Wheezing in early childhood is very common, with 50% of children from high-income countries (HIC) reported to have experienced an episode of wheezing before 6 years of age. 1 Wheezing illness comprises a spectrum of disease, ranging from transient to recurrent, a proportion of which is associated with asthma. 2 Asthma is the most common chronic illness in children, and particularly high in Africa; thus, it is important to understand the risk factors associated with wheeze onset. 2 There are many causes of wheezing in early childhood and several risk factors associated with the development or severity of wheezing. The most common risk factors include environmental tobacco smoke (ETS) exposure; genetic predisposition; early viral lower respiratory tract infections (LRTI); low socioeconomic status and poor living conditions; as well as an increased risk in males. 3 A more recent focus is on maternal psychosocial exposures and the impact these have on child wheeze development and recurrence.

| Setting
This study was a sub-study of the Drakenstein Child Health Study (DCHS), a multidisciplinary birth cohort investigating the epidemiology and etiology of childhood respiratory illness and the early life determinants of child health in a peri-urban area in Paarl, South Africa. 14 The catchment population is approximately 200 000, consisting mainly of those with low socio-economic status, who reside in informal settings or crowded conditions. 14, 15 More than 90% of the population access public healthcare services for their primary care. 14

| Participants
Participants were those enrolled in the DCHS. Inclusion criteria were women 18 years or older, who were at 20-28 weeks' gestation, attended one of two local clinics, provided written informed consent and intended to remain in the area for at least 1 year. 16 Women were followed through childbirth and mother-child pairs were followed through childhood.

| Design
The birth cohort recruited pregnant women attending one of two primary healthcare clinics: Mbekweni, which predominately served a population with African-ancestry and TC Newman which mostly served a mixed-ancestry population. 15 Child clinical and respiratory symptom questionnaires were completed at each of the study visits, which occurred at birth, 6-10 and 14 weeks and 6, 12, 18, and 24 months post-delivery at primary healthcare clinics.

| Measures
Risk factor and outcome data collection is ongoing and recorded longitudinally as part of the DCHS. The primary outcome of this study was child wheeze through 2 years of age.

| Wheeze outcomes
Child wheeze was measured through maternal report at each of the study visits, as well as episodes identified through the active surveillance for respiratory symptoms associated with lower respiratory tract infections (LRTI). Active surveillance was performed by nurses at the primary clinics and assessed in real time. 14,17 These nurses were trained in respiratory examination of children and had to attend frequent competency assessments. 17 Measurements of LRTI included ambulatory and hospitalized pneumonia cases, as defined by World Health Organization (WHO) criteria. 14,17,18 As the mothers were interviewed frequently, it was also possible to retrospectively capture respiratory events occurring at other facilities or outside the area. 17 Any information on respiratory events captured outside of the clinics was obtained by review of medical records. 17 Two binary outcome variables were considered: Whether the child experienced at least one episode of wheeze during the first 2 years of life, or whether the child experienced recurrent wheeze episodes (2 or more wheeze episodes in a 12-month period).
Wheeze was considered present if it was reported during any routine study visit or identified by study staff when examining the child at a LRTI visit in the first 2 years of life.

| Maternal psychosocial measures
Maternal psychosocial data was collected antenatally, and postnatally at 6-10 weeks and 6, 12, 18, 24 months postpartum. 15 Several validated questionnaires were used to measure psychosocial risk factors: The Edinburgh Postnatal Depression Scale (EPDS) is a widely used and reliable measure of depressive symptoms and was used to measure maternal depression. 19 Each of 10 questions were scored 0-3 and totalled. 15 A cut-off value of 13 was used to separate the participants into above-or below-threshold groups. 19,20 The Self-Reporting Questionnaire 20-item (SRQ20), 21 a widely used and validated measure, was used to determine the presence of maternal psychological distress. 22,23 Each item was scored 0-1, and a total score generated. 15 A cut-off value of 8 dichotomized participants into an above-or below-threshold group. 15,22,24 The Intimate Partner Violence (IPV) Questionnaire was used to assess maternal physical, emotional and sexual violence exposure. 25,26 Exposure to IPV was dichotomized by those recently experiencing any one of the three violent exposures versus no exposure.
Other psychosocial measures included: the Childhood Trauma Questionnaire, to assess childhood abuse and neglect, 15,27 which was dichotomized into above-or below-threshold based on any exposure versus no exposure; the Modified Post-Traumatic Stress Disorder Symptom Scale used to screen for current post-traumatic stress disorder (PTSD), 28 which was categorized into three mutually exclusive levels (no exposure, trauma exposed and suspected PTSD).

| Clinical and sociodemographic data
Covariates considered for the analyses included: child feeding practices; HIV exposure; maternal smoking, through self-report, and environmental tobacco smoke (ETS) exposure, assessed by the number of smokers in the child's household; alcohol consumption during pregnancy, measured by the Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) 29 ; maternal or family history of asthma ascertained by maternal report; birth characteristics, such as gestational age and birth weight, measured by study staff; child vaccination; socio-economic status (SES) based on a composite score considering four socio-economic variables: level of education, employment status, household income, and number of asset. 30 Standardized scores were divided into quartiles, which included "low," "low-moderate," "high-moderate," and "high" groups. A time variable, in months, was also generated to measure a child's follow-up time throughout the 24-month period.   Multivariable logistic regression was used to model the associations of maternal psychosocial risk factors, both antenatally and postnatally, with the occurrence and recurrence of child wheezing, adjusting for confounding clinical and sociodemographic covariates. As multicollinearity was present among the psychosocial risk factors, they were considered individually in a series of logistic regression models.

| Statistical analysis
The postnatal psychosocial risk factor measures were also found to be correlated over time, thus we utilized data from the 6-month scheduled visit as a proxy for postnatal exposure, as most wheezing episodes occurred within the first 6 months of life.
Diagnostic checks were generated for all multivariable models.
Based on Pearson's chi-squared and/or the Hosmer-Lemeshow test, all the models were found to correctly specify the association between perinatal maternal psychosocial risk factors and child wheezing outcomes.

| Socio-demographics and clinical factors
The median maternal age was 26 (22.3-31.1) years; 22% of the women were HIV-infected, with a significantly higher prevalence (37%) of HIV in the Mbekweni participants compared to those of TC Newman (3%), Table 1. Approximately 27% of the women smoked during pregnancy, with the majority (53%) from TC Newman. In addition, a higher number of household smokers, and antenatal maternal alcohol consumption, were reported in TC Newman relative to those attending Mbekweni,

| Birth characteristics
An even distribution of males and females were born; a small proportion (17%) of births were premature (<37 weeks' gestation).
Median birth weight was approximately 3 kg. Approximately 92% of the children had initiated breastfeeding, with a median duration of exclusive breastfeeding of 1 month during the follow-up period, Table 1. Two (0.17%) of the children were HIV positive.

| Wheezing episodes
In total, there were 924 wheeze episodes (crude incidence rate = 497 cases per 1000 person-years of follow-up time) throughout the 24-month follow-up period, most of which occurred in those children

| Postnatal psychosocial risk factors and child wheeze
The 6-month postnatal data was used to build the postnatal models, as outcomes at all the scheduled visits were highly correlated (Supplementary Table S4). In addition, a high proportion of wheezing episodes (49%) also took place within the first 6 months of life, Table 1.
The socio-demographics of those who attended and did not attend the 6-month psychosocial visit showed similar characteristics between the two groups, Supplementary Table S5. However,  there was a higher number of smokers in those that attended the visit relative to those that did not attend. In addition, a higher proportion of those attending the visit were from TC Newman.
Psychosocial risk factors measured antenatally and at the 12-month postnatal visit were also compared between those attending and not attending the 6-month psychosocial visit. There was no significant difference found in key exposures investigated (Supplementary Table S5) when comparing antenatal and postnatal psychosocial risk factors between those included and excluded from postnatal analyses.

| Postnatal psychosocial risk factors and presence of child wheeze
Maternal postnatal psychological distress and IPV were found to be significantly associated with child wheeze, when considered independently.
In the multivariable models, exposure to postnatal maternal psychological distress resulted in a twofold increased odds of developing at least one wheeze episode compared to those not

| Postnatal psychosocial risk factors and recurrent child wheeze
From Table 6, the odds of experiencing recurrent wheezing episodes increased by 96% in those whose mothers displayed postnatal depressive symptoms, compared to those whose mothers did not,  Previous studies have found that antenatal maternal psychosocial exposure impacts birth weight as well as lung development, which may lead to airway obstruction. 13,32 As this is the case, there could be an   have included impaired maternal-child relationship and a mother's inability to provide care for her child. 34 A mother that is either exposed to IPV or suffering from mental illness is also more likely to engage in harmful behaviors such as drinking or smoking. This was found to be true in the context of this study, as maternal smoking and drinking habits were found to be associated with postnatal psychosocial risk factors.
As antenatal and postnatal maternal psychosocial risk factors were closely correlated in this study, the effect of postnatal exposure may represent cumulative exposure beginning antenatally and continuing through the postnatal period. An example of this could be biological mechanisms, such as higher cortisol levels in the children due to increased stress hormones being passed from mother to child in-utero. 35 At risk children may be less able to respond to stressors, and thus more prone to wheezing or asthma diagnosis later in life. 36 This may occur, for example, if the mother-child interaction is disturbed through maternal psychosocial risk factor exposures; such alterations have been observed in children whose mothers are exposed to postnatal IPV. 36 Due to this study's findings, biological mechanisms should be investigated in these children to better understand the relationship between maternal psychosocial risk factor exposure and child wheezing in a LMIC context.