The association of asthma, atopic dermatitis, and allergic rhinitis with peripartum mental disorders

Abstract Background Atopic diseases are characterized by dysregulated inflammatory response, which may incur the onset of peripartum mental disorders, but the impact remains unknown. This study examined whether and to what extent the history of atopic diseases is associated with newly onset peripartum mental disorders. Methods Using population‐based registries, we identified all primiparous women who gave birth to live singletons in Denmark during 1978–2016 (n = 937,422). The exposure was hospital contact due to the three major types of atopic diseases—asthma, atopic dermatitis, and allergic rhinitis—before conception. The primary outcome was any hospital contact for mental disorder during pregnancy and 1‐year postpartum, which was further classified into affective disorders, neurotic, stress‐related and somatoform disorders, and substance abuse. The follow‐up started from the date of conception and ended at the date of the first diagnosis of mental disorders, 1‐year postpartum, death, emigration, or December 31, 2016, whichever came first. Cox regression was used, adjusted for calendar year, age at childbirth, education, residence, and Charlson comorbidity index. Results A total of 24,016 (2.6%) women received diagnosis of at least one of the three atopic diseases before conception (asthma, 1.7%; atopic dermatitis, 0.6%; and allergic rhinitis, 0.8%). Exposure to asthma, atopic dermatitis, or allergic rhinitis was associated with a 37% increased overall risk of peripartum mental disorders (hazard ratio [HR], 1.37; 95% confidence interval [CI], 1.27–1.49). Higher risks were observed among women with more frequent hospital contacts for atopic disease (HR, 1.80; 95% CI, 1.37–2.35; ≥5 times), and with recent hospital contacts for atopic disease (HR, 1.74; 95% CI, 1.48–2.06; within 2 years before conception). Specific associations were observed between asthma and neurotic, stress‐related and somatoform disorders (HR, 1.40; 95% CI, 1.21–1.62), and between atopic dermatitis and substance abuse (HR, 1.62; 95% CI, 1.12–2.34). Conclusions History of asthma, atopic dermatitis, and allergic rhinitis before conception was associated with increased risks of peripartum mental disorders. Women who have atopic diseases before pregnancy may benefit from systematic mental health monitoring.

Conclusions: History of asthma, atopic dermatitis, and allergic rhinitis before conception was associated with increased risks of peripartum mental disorders.
Women who have atopic diseases before pregnancy may benefit from systematic mental health monitoring.

K E Y W O R D S
allergic rhinitis, asthma, atopic dermatitis, epidemiology, peripartum mental disorders S C H L Ü S S E L W € OR T E R asthma, atopische dermatitis, allergischer schnupfen, epidemiologie, peripartale psychische störungen

| INTRODUCTION
Peripartum period is a distinct and significant period of neural plasticity for females, driven by changes in hormones and immune molecules, resulting in increased susceptibility to mental disorders. 1,2 Peripartum mental disorders affect over 10% of pregnant and postpartum women, associated with both adverse pregnancy outcomes in mothers and behavioral problems in children. [3][4][5] Though the underlying mechanisms remain to be elucidated, peripartum mental disorders have been related to several immune-related disorders, such as inflammatory bowel disease, pre-eclampsia, and gestational diabetes. [6][7][8] Atopy is a personal or familial propensity to produce IgE antibodies in response to environmental triggers, contributing to atopic diseases. 9,10 Atopic dermatitis, asthma, and allergic rhinitis are the most common atopic diseases that affect approximately 20% of the population worldwide, and often develop sequentially, known as the "atopic march". 9,10 The shared pathogenesis of dysregulated inflammatory response has been reported to trigger the onset of depression and anxiety in the general population. [11][12][13][14] The findings on the association between asthma and postpartum depression have, however, been inconsistent. [15][16][17] The prevalence of atopic diseases, mostly developed before childbearing ages, is on the increase. 18 To the best of our knowledge, the association of atopic dermatitis and allergic rhinitis with peripartum mental disorders has not been studied. Moreover, while most studies investigated risk factors for postpartum-onset mental disorders, 19 the

Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition
(DSM-5) has adopted the definition of peripartum mental disorders, emphasizing that pregnancy-related mental disorders onset both during pregnancy and postpartum. 20 In this population-based cohort study using data from Danish national registries, we aimed to evaluate the association between the three atopic diseases (asthma, atopic dermatitis, and allergic rhinitis) before conception and newly onset peripartum mental disorders in pregnant women.

| Study population and follow-up
We studied all primiparous women, who gave birth to a live singleton from 1978 to 2016 in Denmark, recorded in the Danish Medical Birth Registry (n = 1,045,372), which prospectively collects data from pregnancy to the peripartum period for approximately 99% of births in Denmark. 21

| Exposure
The unique civil registration number assigned to all Danish residents permits accurate linkage of individual-level data from different national registers. 21 The diagnosis of atopic diseases was retrieved from the Danish National Patient Register, which has recorded information on all patients discharged from Danish nonpsychiatric hospitals since 1977, and on emergency and outpatient contacts since 1995. The Patient Register contains no data on visits to general practitioners. The diagnoses were coded according to from 1969 through 1993, and ICD-10 since 1994. We defined the exposure as the diagnosis of either asthma (ICD-8 code 493 and ICD-10 code J45-J46), or atopic dermatitis (ICD-8 code 691 and ICD-10 code L20), or allergic rhinitis (ICD-8 code 507 and ICD-10 code J30) before conception. 13,22,23 .
Atopic multimorbidity was defined as the diagnosis of more than one type of the three atopic diseases. 10 The frequency of atopic disease hospital contacts before conception was used as a proxy for recurrent atopic diseases (1, 2-5, or ≥5). 24 The period from the last hospital contact of atopic diseases to conception was used as a proxy for disease activity (<2, 2-10, or ≥10 years). 25 As the pathogenesis may differ between childhood-and adult-onset atopic diseases, we categorized the exposed women into three groups based on their age at the first diagnosis of atopic disease (<10, 10-20, or >20 years old). [26][27][28] Considering seasonality may influence the severity of atopic diseases, we categorized the exposed women according to the season of childbirth (winter, December-February; spring, March-May; summer, June-August; autumn, September-November). 29

| Outcome
The diagnosis of mental disorders was retrieved from the Danish Psychiatric Central Research Register, which contains every psychiatric inpatient since 1970 and outpatient and emergency room contacts since 1995. 21  xx, and 304.xx; ICD-10 codes F10 through F19). 31 Mental disorders unclassified in the former three clusters were classified as other mental disorders, mainly comprised of the following diagnostic entities: unspecified type of mental disorder (ICD-10 code, F99), personality disorder (ICD-8 code 301, ICD-10 code, F60), and mental and behavioral disorders associated with the puerperium (ICD-10 code, F53).

| Covariates
Potential confounders were selected based on directed acyclic graphs ( Figure S1), showing the best known relations between the variables in this study. Apgar score (<7, 7-9, or 10), low birth weight (<2500 or ≥2500 g), and newborn death (death within the first year of age; yes or no).
Maternal smoking during pregnancy (yes or no) was only available after 1991, thus was only adjusted for in the sensitivity analyses.

| Statistical analyses
Cox proportional hazards regression was applied to estimate the HRs and 95% CI, adjusted for the above-mentioned potential confounders. The proportional hazard assumption was assessed by a logminus-log plot ( Figure S2). The two curves showed a roughly parallel pattern, and we found no obvious evidence supporting the interaction between HR and time (p for interaction 0.51). Thus, it's reasonable to accept the proportional hazard assumption. We evaluated risks of peripartum mental disorders in relation to the timing of diagnosis of mental disorder (during pregnancy or postpartum).
Subgroup analyses were performed in the three types of atopic diseases. We used the likelihood-ratio test to test for a trend in the analyses according to the atopy recurrence and atopy activity, with weights assigned according to the median of each category. 33 A trend in atopy recurrence was assessed with weights of 1 in the group of women who had only one contact of atopic disease, 2 in the group two to four times, and 8 in the group five times or more. A trend in atopy activity was assessed with weights of 1 in the group of women whose latest contact of atopic disease before conception were within persists in multiparous settings, we identified all women giving birth to a live singleton from 1978 to 2016 in Denmark, with the same exclusion criteria described above, and randomly selected one pregnancy for multiparous women as the study population; (5) parturients with adverse pregnancy outcomes are often at higher risks of postpartum mental disorders, thus we restricted the analysis in women with favorable pregnancy outcomes (excluding offspring with 5-min Apgar score <7, low birth weight, pre-term birth, and neonatal mortality); (6) as the three atopic diseases might be diagnosed during pregnancy, the exposure was treated as a time-varying variable during pregnancy; (7) the frequency of hospital contacts for each of the three atopic diseases was counted separately; (8) the test for trend was performed with different weights assigned to each category. In addition, Henriksen et al. 29 reported an algorithm to identify atopic diseases with increased sensitivity by incorporating medication data from the Danish National Prescription Registry. A subanalysis was performed using the same algorithm with restriction on women giving birth to a live singleton between 1996 and 2016, when medication data were available (detailed methods were described in Appendix S1). All analyses were performed using Stata 15 and R version 3.6.1. (18.6% of exposed women) had atopic multimorbidity (e.g., diagnoses of more than one type of the three atopic diseases). The women with asthma, atopic dermatitis, or allergic rhinitis were more likely to be older at pregnancy, have a high level of education, live in urban areas, and suffer from somatic comorbidities and pregnancy complications (Table 1).

| RESULTS
Peripartum mental disorders affected 2.7% of the women exposed to asthma, atopic dermatitis, or allergic rhinitis, whereas only 1.2% of the unexposed were affected. After adjustment for potential confounders, we observed a 37% increased risk in exposed women (HR, 1.37; 95% CI, 1.27-1.49; Figures 1 and S2). The same pattern was observed for mental disorders diagnosed either during pregnancy or postpartum (Table 2). Subgroup analyses revealed that the risk was more eminent during pregnancy than postpartum (  Figure 2). Asthma, atopic dermatitis, and allergic rhinitis were all associated with affective disorders at a similar level (Figure 2). In particular, asthma was associated with a higher risk of neurotic, stress-related and somatoform disorders (HR 1.40; 95% CI, 1.21-1.62; Figure 2), atopic dermatitis with a higher risk of substance abuse (HR 1.62; 95% CI, 1.12-2.34; Figure 2).

| DISCUSSION
In this population-based cohort study, we observed an overall 37% increased risk of peripartum mental disorders associated with a before-conception diagnosis of either of the three atopic diseases: asthma, atopic dermatitis, and allergic rhinitis. The risk was higher among women with atopic multimorbidity, recurrent atopic disease, and active atopic disease. All the three types of atopic diseases were associated with affective disorders, while strong association was found between asthma and neurotic, stress-related and somatoform disorders, and between atopic dermatitis and substance abuse.
The association between pre-pregnancy asthma and peripartum mental disorders was reported in two population-based cohort studies, which found a 52% increased risk of depression during pregnancy and a 58% increased risk of depression postpartum, respectively. 16,17 Both studies only focused on pregnancy-related depression. Using this large population-based cohort, we observed that women with a diagnosis of asthma, atopic dermatitis, and allergic rhinitis before conception had a higher overall risk of newly onset peripartum mental disorders. Atopic diseases are often poorly controlled during peripartum period, due to the systematically altered immune status and the non-adherence to pharmacological treatment in peripartum women. 34,35 As a result, aggravated symptoms and atopy may influence mental disorders on peripartum period though a number of causal pathways. Aggravated atopic symptoms may contribute to mental stress, through sleep disorder induced by severe pruritus or dyspnea. 34,36 Also, allergic reactions may dysregulate autoimmune system, which might trigger following mental disorders. 37 The peripheral inflammation ignited from skin, nasal cavities, and respiratory tract may affect the brain, transmitted by dysregulated inflammatory cytokines, such as interleukin-6, interleukin-1β, and tumor necrosis factor. 38,39 These mechanisms may explain and support our findings of a higher risk of peripartum mental disorders in women with more frequent or recent hospital contacts for atopic diseases. In addition, shared genetic susceptibility may be part of the underlying causes given the genetic overlaps between asthma and attention deficit hyperactivity disorder, anxiety, and major depressive disorder. 40 Currently, most studies on peripartum mental disorders have focused on depression, even though anxiety is more prevalent than depression during the peripartum period. 30 Our analyses found that peripartum neurotic, stress-related and somatoform disorders, mainly comprised of anxiety, were specifically associated with asthma. This pattern has been recognized in the general population, partially due to poor asthma control. 41,42 It also suggests a potentially different mechanism underlying the association between asthma and anxiety from that for the association between atopic disease and depression. This is possibly due to the significant heterogeneity in immunopathology in asthma, compared with the generally consistent characteristic in atopic dermatitis and allergic rhinitis (immune imbalance toward a T-helper-2 response). [43][44][45] In addition, we observed a novel association between atopic dermatitis and peripartum substance abuse. Similar results were reported in the general Danish population, where the prevalence of smoking and alcohol abuse was higher among atopic dermatitis patients. 46 T A B L E 1 Baseline characteristics of primiparous women exposed and unexposed to asthma, atopic dermatitis, and allergic rhinitis before pregnancy in Denmark from 1978 to 2016  Our findings should also be interpreted in the light of the following limitations. First, the incidence of 1.2% for peripartum mental disorders in this study was lower than the reported incidence of ∼10% worldwide. 49 Also, the prevalence of the three atopic diseases of interest was lower than that was reported in many areas. 9 However, similar low incidence and prevalence were reported in other Nordic registry-based studies with a high validity of diagnosis from hospital discharges, given the fact that patients with mild symptoms might not always seek medical care. 13,50,51 We performed an additional sub-analysis utilizing medication data to define atopic diseases, which was validated to increase the diagnostic sensitivity. 52 The association persisted, while the risk estimate was a little lower.
We cannot rule out the possibility of misclassification in both exposure and outcomes, in which case the observed elevated risks would be overestimated and reflect the associations in relation to Fourth, other atopic diseases, like food allergy, could not be characterized by the ICD-code-based diagnoses. This may cause misclassification bias and led to an underestimation of the association between atopic diseases and peripartum mental disorders.
However, the studied three types represent the major atopic diseases. 9 Fifth, there was a latency period between the date of diagnosis from hospital and actual disease onset, thus the timing of outcomes (during pregnancy and postpartum) should be interpreted with caution. Sixth, as we used hospital-based data to identify and exclude women with a history of mental disorder, mild cases with no hospital contacts were still included in the study population. When diagnosed during peripartum period, these patients with mild mental disorders would be misclassified as peripartum mental disorders.
This could result in an overestimation of the incidence of peripartum mental disorders.

| CONCLUSION
History of asthma, atopic dermatitis, and allergic rhinitis were associated with increased risks of peripartum mental disorders. All the three types of atopic diseases were associated with affective disorders. Strong associations were observed between asthma and neurotic, stress-related and somatoform disorders, and between atopic dermatitis and substance abuse. Women who have atopic diseases before pregnancy may benefit from future systematic mental health monitoring.