Maternal mental illness and child atopy: a UK population-based, primary care cohort study

Background The number of children exposed to maternal mental illness is rapidly increasing and little is known about the effects of maternal mental illness on childhood atopy. Aim To investigate the association between maternal mental illness and risk of atopy among offspring. Design and setting Retrospective cohort study using a UK primary care database (674 general practices). Method In total, 590 778 children (born 1 January 1993 to 30 November 2017) were followed until their 18th birthday, with 359 611 linked to their hospital records. Time-varying exposure was captured for common (depression and anxiety), serious (psychosis), addiction (alcohol and substance misuse), and other (eating and personality disorder) maternal mental illness from 6 months before pregnancy. Using Cox regression models, incidence rates of atopy were calculated and compared for the exposed and unexposed children in primary (asthma, eczema, allergic rhinitis, and food allergies) and secondary (asthma and food allergies) care, adjusted for maternal (age, atopy history, smoking, and antibiotic use), child (sex, ethnicity, and birth year/season), and area covariates (deprivation and region). Results Children exposed to common maternal mental illness were at highest risk of developing asthma (adjusted hazard ratio [aHR] 1.17, 95% confidence interval [CI] = 1.15 to 1.20) and allergic rhinitis (aHR 1.17, 95% CI = 1.13 to 1.21), as well as a hospital admission for asthma (aHR 1.29, 95% CI = 1.20 to 1.38). Children exposed to addiction disorders were 9% less likely to develop eczema (aHR 0.91, 95% CI = 0.85 to 0.97) and 35% less likely to develop food allergies (aHR 0.65, 95% CI = 0.45 to 0.93). Conclusion The finding that risk of atopy varies by type of maternal mental illness prompts important aetiological questions. The link between common mental illness and childhood atopy requires GPs and policymakers to act and support vulnerable women to access preventive (for example, smoking cessation) services earlier.


INTRODUCTION
Mothers play a key role in the development of offspring: from pregnancy to early adulthood, mothers are likely to be the primary caregivers.
Accumulating evidence suggests that children exposed to maternal mental illness have poorer physical health than unexposed children. 1 Some evidence highlights the risk of atopic diseases, such as asthma, eczema, allergic rhinitis, and food allergies, [2][3][4] which are increasingly prevalent, now occupying some of the commonest global health problems among children. 5For example, asthma affects almost 14% of children worldwide and, in the UK, one in 11 children receive asthma treatment, costing the NHS approximately £1 billion per year. 6The prevalence of eczema, allergic rhinitis, and food allergies is also increasing: 7 20% of the UK population has ≥1 atopic disease. 8urrently, almost 25% of children in the UK live with a mother experiencing mental illness, 9 which the authors' research group estimated was associated with an excess annual NHS cost of £560 million in health utilisation. 10topic disease in children is common and associated with common mental illness (CMI) among young people, 11 and maternal mental illness is linked to increased risk of offspring atopy. 1 This points to common aetiological mechanisms shared between depression, anxiety, and atopic disease. 12][15][16][17][18][19][20] Further, studies fail to account for potential confounders such as younger maternal age, or poverty and other socioeconomic factors; or examine environmental pathways potentially leading to atopy risk, including poor antenatal care, maternal smoking, (both while pregnant and as a parent), fetal growth restriction, or maternal antibiotic exposure.Finally, if environmental factors are important in the aetiology, in offspring with atopy whose mothers have serious mental illness (SMI) and higher levels of adversity they should experience higher risk of atopic disease compared with their peers whose mothers are well or have CMI only. 21his study used a large representative population-based cohort of mothers and children to compare the risk of atopic diseases in children of mothers with mental e924 British Journal of General Practice, December 2023 illness and mothers without mental illness.It was hypothesised that: • children of mothers with mental illness would have a higher risk of developing atopic diseases compared with children of mothers without mental illness; • the risk of atopic diseases would be higher among children exposed to serious maternal mental illness (affective and non-affective psychosis) compared with more common maternal mental illnesses (depression and anxiety); and • the risk of admission to hospital for asthma or food allergy (indicating a serious atopy event) would be higher among children of mothers with mental illness than mothers without mental illness.

Data source
This retrospective cohort analysis utilised the Clinical Practice Research Datalink (CPRD) GOLD with anonymised primary healthcare records for about 10% of UK general practices. 22Clinical event data are collected routinely on consultations (including diagnosis), referrals, and prescriptions.The analysis cohort was constructed using the CPRD mother-baby link, an algorithmic linkage of children and mothers based on pregnancy, delivery, and birth records, and a household identifier. 23he hospital episode statistics (HES) dataset holds anonymised electronic healthcare records of all NHS (free at the point of access) hospital visits in England; approximately 75% of English practices registered with CPRD GOLD consented for their patients' records to be linked. 22ocioeconomic data are based on Index of Multiple Deprivation (IMD) linked to GP practice postcode.The IMD is a rank score of area-level deprivation, divided into quintiles, derived using seven domains including: income, employment, education, health and disability, crime, barriers to housing and services, and the lived environment.

Cohort selection
Eligible children included those born between 1 January 1993 and 30 November 2017, whose mother was registered at a general practice for >6 months before the child's date of birth (Figure 1).
Follow up started on the latest date of the: child's birth; the practice started collecting data deemed 'up-to-standard'; child's registration; study start date (1 January 1993).Follow up ended at the earliest date of the: child or mother's death or transfer out of general practice; child's 18th birthday; end of data collection; and study end date (31 December 2017).This included 590 778 children and 428 924 mothers with an average follow up of 5.24 person-years.To investigate the risk of secondary healthcare utilisation for atopic diseases among children exposed to maternal mental illness, 359 611 of this cohort were selected who had HES linkage.

Exposure
Children's exposure to maternal mental illness was identified using primary care data on each mother's recorded diagnoses, symptoms, prescriptions, and referral to external services (see Abel et al 9

How this fits in
Prior reports of increased risk of asthma and allergic rhinitis in children of mothers with depression or anxiety has excluded important confounders and not considered potential environmental pathways.This study found an increased risk of asthma and allergic rhinitis among children exposed to maternal depression or anxiety, and explores the role that maternal smoking plays (both smoking while pregnant and/or as a parent) in childhood atopy.Future research, to test if the specific association seen here means maternal depression/anxiety causes childhood atopy, is required.Tailoring communication and preventive services to the health needs of women experiencing mental illness would benefit both women and their children.British Journal of General Practice, December 2023 e925 details).Children were defined as exposed from the first record of maternal depression, anxiety, non-affective and affective psychosis, alcohol or substance misuse, eating disorder, or personality disorders from 6 months before pregnancy until the end of follow up.
For this analysis, mental illnesses were grouped into: CMI (anxiety and depression), SMI (non-affective and affective psychosis), addiction (alcohol and substance misuse), and 'other' (eating and personality disorders).

Outcome
Atopic diseases included incident diagnosis of childhood asthma, eczema, allergic rhinitis, or food allergy.Clinical codes identifying diagnoses are available on github. 24o analyse the risk of admission to hospital for serious atopic disease all records were captured of hospital admission (day, night, planned, or unplanned) in HES where there was an International Classification of Diseases diagnosis of either asthma or food allergy. 25

Covariates
Data on maternal age at birth, child sex, maternal history of atopic disease, antibiotic use during pregnancy, and smoking status were extracted from the CPRD.Validated HES data on child ethnicity (Asian/British Asian, Black/Black British, White, Mixed, and Other) were used and supplemented with CPRD data if missing. 26

Statistical analysis
The risk of atopic disease (time to first event) for children exposed and unexposed to maternal mental illness was calculated and presented as incidence rates (events per 1000 person-years).Cox proportional hazard models were used to compare risks for maternal mental illness in the exposed and unexposed children, presented as hazard ratios.Potential confounders were adjusted for: maternal atopy history, antibiotic use during pregnancy, and age at birth; and child sex, ethnicity, delivery year, deprivation quintile, and region in model 1.The authors considered smoking on the causal pathway; therefore, a second model also included it as a covariate (model 2).
In the cohort with HES linkage, the time to an atopy hospital admission (asthma and food allergy) was compared in children in those exposed to mothers with and without maternal mental illness.
In all regression analyses, continuous variables were centred and a squared term included.Clustering by maternal sibships was accounted for by calculating the standard error using the Huber-White sandwich estimator. 27Data were analysed using Stata (SE 15.0).
This study meets the requirements of the RECORD statement. 28In this study the authors considered hazard ratios (HRs) where the 95% confidence interval (CI) did not cross 1.0 and a significance P-value <0.05 to be more conclusive or stronger evidence of an association.HRs were reported to two decimal places.

Sensitivity analysis
Although it is understood that some cases of infantile eczema can be severe and if untreated may require secondary care, the authors were concerned that any association between maternal mental illness and childhood eczema might be explained by mild, highly prevalent and time-limited infantile eczema.A sensitivity analysis was therefore conducted that started follow up from 3 years of age.

RESULTS
The analysis included 590 778 mother-child pairs (Figure 1) and a total of 3 840 135 person-years follow up (median time in analysis: 5.24 years; interquartile range 2.12-10.01)with a mean maternal age of 30  2).Both other maternal mental illness (aHR 1.04, 95% CI = 0.91 to 1.19, model 1) and CMI (aHR 1.16, 95% CI = 1.13 to 1.20, model 1) increased the risk of allergic rhinitis.After adjustment for smoking only, maternal CMI increased the risk of allergic rhinitis (aHR 1.17, 95% CI = 1.13 to 1.21, model 2).Children exposed to maternal CMI had a similar risk of asthma and allergic rhinitis.
Food allergies.Unexposed children had the highest incidence of food allergies (2.75 per 1000 person-years) (Table 2).Compared with unexposed children, children exposed to CMI had a 5% reduced risk (crude HR 0.95, 95% CI = 0.90 to 0.99), which was inconclusive after adjusting for potential confounders.Children exposed to maternal addiction had a reduced risk of allergy (crude HR 0.51, 95% CI = 0.37 to 0.69) that persisted following adjustment for confounders (aHR 0.65, 95% CI = 0.45 to 0.93, model 2).
Covariates.Overall, male children, children in the most deprived quintile, and those exposed to maternal asthma, maternal antibiotic use during pregnancy, or maternal smoking had increased risk of asthma.In contrast, developing eczema and food allergies was associated with children being less deprived, without exposure to maternal smoking (see Supplementary Tables S1  and S2).

Hospital admission for asthma and food allergy
There was a 29% increased risk of admission to hospital for asthma among children exposed to CMI (aHR 1.29, 95% CI = 1.20 to 1.38, model 1, Table 3).For the other types of maternal mental illness there was weaker evidence of an association after adjusting for potential confounders (for example, maternal addiction) (aHR 1.27, 95% CI = 0.95 to 1.71, P = 0.113, model 1).unexpectedly, with reduced risk for eczema and food allergies.
The influence of deprivation varied by outcome: children from the most deprived households had the highest risk of asthma and lowest risk of eczema.This pattern was apparent for food allergies, but less so for allergic rhinitis.Finally, asthma- related admission to hospital was greater among children exposed to maternal CMI compared with unexposed children.

Strengths and limitations
The CPRD provides a large enough cohort to estimate, with precision, links between individual maternal mental illnesses and offspring atopic disorders.However, some limitations remain.It was not possible to capture eczema severity, which is potentially vital to explain why SMIs are associated with reduced childhood eczema in primary care.The effects of maternal mental illness on risk of atopy reported here do not account for unmeasured confounders such as housing quality and paternal mental health/atopy.Important risk factors other than smoking that might be on the causal pathway, such as air pollution, were not recorded.
The study could only measure geographical region and area-level indices of deprivation.Effects of mental illness in ethnicities other than White British mothers may be underestimated because they are relatively underrepresented in this primary care cohort. 29Finally, using HES linkage to examine risk of admission to hospital excluded many mother-child pairs in England and children from Scotland, Wales, or Northern Ireland, since HES linkage is only available for NHS England and some English general practices do not consent to this linkage.

Comparison with existing literature
An increased risk of asthma and allergic rhinitis in children of mothers with CMI replicates other large cohort studies. 18,19,30 e current study has extended these findings by demonstrating the specificity of this association, suggesting shared factors increase maternal depression/anxiety and childhood atopy, 31 which may be expressed in utero. 12he current study does not replicate smaller studies reporting greater risk of eczema in children exposed to maternal CMI. 15,17,32owever, removing those with infantile eczema unmasked excess eczema in children exposed to maternal CMI; and maternal smoking and socioeconomic deprivation explained some of the association -children from the most deprived areas manifest reduced risk of eczema/food allergies.Eczema is common in infancy (0-3 years) and likely to be mild and self-limiting.It could be that mothers in higher socioeconomic groups are more health literate and spot eczema earlier than women who live in more deprived circumstances. 33s a group, atopic disorders share an immune aetiology; 34 however, the current results suggest heterogeneity in the pathways for different disorders.Similar heterogeneity is observed for parental mental illness and autoimmune disease. 35t is increasingly recognised that infantile and pregnancy exposure to diversity in the microbiota is important for development of immune responses 36 and that diet is key to maintaining those through life. 37This might explain why children exposed in utero to antibiotics develop more atopy.Also, if household crowding is more common in those with serious mental illness or addiction this is one way early immune responses could be strengthened and may explain why children in these settings develop less atopy. 12ifferences in maternal health anxiety may also clarify some of the different risk associations, for example, mothers with anxiety disorders being more vigilant and/ or reporting more symptoms.In the current study, the authors report that children exposed to maternal addiction are less likely to be seen in primary care for eczema, and cautiously report an increased likelihood of admission to hospital for eczema.Milder disorders may be undetected by ill mothers; mothers living in deprived areas, ethnic minority, and migrant mothers may be less able to attend/arrange a GP appointment because of limited knowledge, lack of English, access to transport, or childcare. 29,38plications for research and practice Public health policies and practice guidelines, including the Healthy Child Programme, 39 should be modified to tailor support to mothers with mental illness whose health literacy may be compromised.For instance, children were at high risk of developing asthma following exposure to maternal CMI; but also if they were exposed to maternal smoking during their lifetime or to antibiotics in utero; and if they lived in the most deprived areas.Therefore, expanding preventive health programmes to target vulnerable women ante-and postnatally (for example, smoking cessation), and tailoring information specifically to their needs stands out as a clear implication of the authors' findings for general practice.Such tailored information is important because, increasingly, the evidence suggests that people with mental illness underestimate their risk of poor physical health, and are less likely to adopt preventive health measures or be able to change behaviour following widespread public health campaigns. 40This means that we run the risk of increasing health inequalities if those at greatest risk are unable to access the benefits of available public health care.Training for healthcare providers in general practice who socially prescribe and signpost (for example, making every contact count), 41 might help reduce the consistently high rates of smoking in mothers with mental illness.
In conclusion, common mental illnesses in mothers increases asthma and allergic rhinitis in offspring.Aetiological mechanisms in atopic disorders remain elusive, but this and other findings make it likely that maternal mental illness plays a significant aetiological role.Future research should triangulate findings from studies of different designs to discern causality and mitigate flaws inherent in each, to better account for unmeasured confounding. 42vidence of the population effects of providing tailored support to mothers with mental illness is needed.Evidence of the impact of implementing tailored support for mothers on child health outcomes is lacking, and data are limited to a few randomised controlled trials with mixed results, primarily in low and middle-income countries.Such tailoring would allocate resources more equitably and might deliver significant health and economic benefits for all.

Table 1 . Characteristics of mother-child cohort, children born between 1 January 1993 and 31 December 2015 (N = 590 778)
a Other: eating and personality disorders.CMI = common mental illness.IMD = Index of Multiple Deprivation.SMI = serious mental illness.British Journal of General Practice, December 2023 e927

Table 3 . Unadjusted and adjusted HRs showing the association between maternal mental illness and offspring admission to hospital for asthma (N = 359 611 children) a
Owing to CPRD's small cell policy, patient counts with fewer than five observations have been suppressed.Therefore, food allergy-related admission to hospital results are not shown in this Table.b Adjusted for maternal history of atopic disease, antibiotic use during pregnancy, maternal age, child sex, child ethnicity, birth season, birth year practice IMD, and practice region.c Adjusted for all variables in model 1, plus maternal smoking.d Other: eating and personality disorders.CMI = common mental illness.CPRD = Clinical Practice Research Datalink.HR = hazard ratio.IMD = Index of Multiple Deprivation.Ref = reference.SMI = serious mental illness.
a British Journal of General Practice, December 2023 e929