Key Points
Question
Is maternal autoimmune disease associated with attention-deficit/hyperactivity disorder (ADHD) in children?
Findings
In this cohort study of 63 050 children, children whose mothers had any autoimmune disease, type 1 diabetes, rheumatic fever or rheumatic carditis, or psoriasis were significantly more likely to be diagnosed with ADHD. Additionally, in a meta-analysis of 5 studies, children whose mothers had any autoimmune disease, type 1 diabetes, hyperthyroidism, or psoriasis were significantly more likely to be diagnosed with ADHD.
Meaning
Maternal autoimmune diseases appear to be associated with increased ADHD among children, and more research is necessary to understand mechanisms underlying the association.
This cohort study uses data from the New South Wales Perinatal Data Collection database to assess whether the presence of maternal autoimmune disease is associated with attention-deficit/hyperactivity disorder in children.
Abstract
Importance
Maternal autoimmune disease has been associated with increased risk of neurodevelopmental disorders in offspring, but few studies have assessed the association with attention-deficit/hyperactivity disorder (ADHD).
Objective
To examine the association between maternal autoimmune disease and ADHD within a population-based cohort and combine results in a subsequent systematic review and meta-analysis.
Design, Setting, and Participants
A cohort study was conducted of singleton children born at term gestation (37-41 weeks) in New South Wales, Australia, from July 1, 2000, to December 31, 2010, and followed up until the end of 2014; and a systematic review evaluated articles from the MEDLINE, Embase, and Web of Science databases to identify all studies published before November 20, 2019. A total of 12 610 children exposed to maternal autoimmune disease were propensity score matched (1:4) to 50 440 unexposed children, for a total cohort of 63 050. A child was considered to have ADHD if they had (1) an authorization or filled prescription for stimulant treatment for ADHD or (2) a hospital diagnosis of ADHD. Children linked to a first ADHD event before 3 years of age were excluded. Data were analyzed from January 13 to April 20, 2020.
Exposures
One or more maternal autoimmune diagnoses in linked hospital admission records between July 1, 2000, and December 31, 2012. Thirty-five conditions were considered together and individually.
Main Outcomes and Measures
The main outcome was child ADHD identified from stimulant authorization or prescription data and diagnoses in linked hospital admission records. Multivariable Cox regression was used to assess the association between maternal autoimmune disease and ADHD adjusted for child sex. Pooled hazard ratios (HRs) were calculated using random-effects meta-analysis with inverse-variance weights for each exposure reported by 2 or more studies.
Results
In the population-based cohort analysis, 831 718 singleton, term infants born to 831 718 mothers (mean [SD] age, 29.8 [5.6] years) were assessed. Of 12 767 infants (1.5%) who were linked to a maternal autoimmune diagnosis, 12 610 were propensity score matched to 50 440 control infants, for a total study cohort of 63 050 infants. In this cohort, any autoimmune disease was associated with ADHD in offspring (HR, 1.30; 95% CI 1.15-1.46), as was type 1 diabetes (HR, 2.23; 95% CI, 1.66-3.00), psoriasis (HR, 1.66; 95% CI, 1.02-2.70), and rheumatic fever or rheumatic carditis (HR, 1.75; 95% CI, 1.06-2.89). Five studies (including the present study) were included in the meta-analysis. Any autoimmune disease (2 studies: HR, 1.20; 95% CI, 1.03-1.38), type 1 diabetes (4 studies: HR, 1.53; 95% CI, 1.27-1.85), hyperthyroidism (3 studies: HR, 1.15; 95% CI, 1.06-1.26), and psoriasis (2 studies: HR, 1.31; 95% CI, 1.10-1.56) were associated with ADHD.
Conclusions and Relevance
In this cohort study, maternal autoimmune diseases were associated with increased ADHD among children. These findings suggest possible shared genetic vulnerability between autoimmune disease and ADHD or a potential role for maternal immune activation in the expression of neurodevelopmental disorders in children. Future studies measuring disease activity, modifiers, and medication use are required to better understand the mechanisms underlying this association.
Introduction
Autoimmune diseases are a collection of conditions sharing a common cause: the body’s immune system inappropriately attacking its own cells.1 More than 80 individual autoimmune diseases have been identified; prior studies estimate that together they affect 3% to 9% of the world population,2,3,4,5 disproportionately affecting women of reproductive age.4,6 There is growing evidence that immune-related cells and proteins play a role in brain development and function7 and that maternal immune activation, including infection, autoimmune disease, and chronic inflammation during pregnancy, increases the risk of neurodevelopmental disorders among children.8,9,10 Maternal autoantibodies and proinflammatory cytokines are hypothesized to cross the placenta and alter fetal brain development. Potential mechanisms include epigenetic modulation of neurodevelopmental genes,11 activation of microglia (the innate immune cells of the brain), and modification of synapse formation and function.12
Neurodevelopmental disorders are caused by disruptions in brain development and include autism spectrum disorder, attention-deficit/hyperactivity disorder (ADHD), and learning disabilities.13 Previous studies have primarily focused on the association of maternal autoimmune diseases with autism spectrum disorder. The most recent 2016 systematic review and meta-analysis of 10 studies by Chen et al14 reported an overall increased odds of autism among children of women with an autoimmune disease.
In contrast, the association of maternal autoimmune diseases with ADHD has been less studied. Attention-deficit/hyperactivity disorder is characterized by persistent problems with inattention or hyperactivity-impulsivity15 and high levels of comorbidity with other neurodevelopmental disorders. It is estimated to affect 5% of children globally.16 Only 1 study, to our knowledge, has examined the association between any maternal autoimmune disease and ADHD, reporting an increased incidence rate of ADHD among children of women with any autoimmune disease.17 Additional studies have reported increased risk of ADHD among children of women with specific autoimmune conditions, including type 1 diabetes (T1D),17,18,19,20 multiple sclerosis,18 rheumatoid arthritis,18 autoimmune hepatitis,17 psoriasis,17 and ankylosing spondylitis,17 but these studies have been limited, and most have a small sample size or limited case numbers.
This study has 2 main aims: (1) to examine the association between maternal autoimmune disease overall, as well as by specific conditions, and ADHD in children and (2) to incorporate those results into a systematic review and meta-analysis of the published literature on this topic.
Methods
Population-Based Cohort Study
Study Population
All women with a singleton live birth at term (37-41 weeks’ gestation) in New South Wales (NSW), Australia, from July 1, 2000, to December 31, 2010, were identified from the NSW Perinatal Data Collection, a database of all infants born at 20 weeks gestation or later or weighing at least 400 g at public and private hospitals and at home in the state. Birth data were linked to all hospital admissions between July 1, 2000, and December 31, 2012, for both the mother and child. Hospital admission data were obtained from the NSW Admitted Patients Data Collection, a population-level data set of all public and private hospital admissions in the state. Birth data were also linked to pharmaceutical data for the child for the treatment of ADHD between July 1, 2000, and December 31, 2014. In NSW, prescription of psychostimulant medications requires prior authorization from the Ministry of Health. Pharmaceutical data were obtained from the NSW Pharmaceutical Drugs of Addiction System, which contains authorizations and prescriptions for stimulants to treat ADHD, including dexamphetamine, methylphenidate, and lisdexamfetamine. Ethics approval for the study was obtained from the NSW Population and Health Services Research Ethics Committee, which granted a waiver of consent for the study on the grounds that contacting all individuals in the cohort was not feasible. The systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline.
Attention-Deficit/Hyperactivity Disorder
A child was considered to have ADHD if they were linked to 1 or more of the following: (1) an authorization or filled prescription for stimulant treatment for ADHD or (2) a hospital diagnosis of ADHD (International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification [ICD-10-AM] codes F90, F98.8). Given the difficulty of diagnosing ADHD in young children, individuals linked to a first ADHD event before 3 years of age were excluded. The time between the child’s third birthday and the date of the first ADHD event (diagnosis, authorization, or prescription) was calculated for use in survival analysis. Children were censored at the time of death or at the end of the follow-up period, December 31, 2014.
Maternal Autoimmune Disease
Maternal autoimmune disease status was determined based on diagnosis codes from the mother’s linked hospitalization records. Each hospital admission can have up to 51 relevant diagnoses recorded using ICD-10-AM codes. A list of 35 conditions and their relevant diagnosis codes were developed based on a review of the literature and consultation with clinical stakeholders (eTable 1 in the Supplement).2,3,4,5 Children whose mother had at least 1 hospitalization with relevant diagnosis codes at any time in the study period were considered exposed to maternal autoimmune disease. An additional sensitivity analysis was conducted by restricting maternal autoimmune exposure to 1 or more diagnosis codes before or within 180 days of birth.
Propensity Score Matching
To control for potential confounding due to differences in maternal demographic details, pregnancy, and other health factors between mothers with and without autoimmune disease, propensity score matching was used. Propensity scores were calculated using logistic regression incorporating maternal age, maternal country of birth (Australia or New Zealand vs elsewhere), socioeconomic disadvantage (quintiles), place of residence (major city vs regional vs remote areas), parity, smoking during pregnancy, and maternal mental health diagnosis (ICD-10-AM F00-F99, Z86.5). Children of mothers with an autoimmune disease were optimally matched to 4 unexposed children born within 6 months, based on the logit of the propensity score (caliper size = 0.00001) to create a final study cohort.21 Absolute standardized difference measures were calculated for exposed and unexposed children in the full population and the study cohort to confirm successful balancing (difference <0.1) of potential confounders.
Statistical Analysis
To account for different follow-up periods among children in the cohort, Cox regression models were used to examine the association of maternal autoimmune disease and ADHD. Hazard ratios (HRs) with robust 95% CIs were calculated to account for the matched design of the study. Analyses were conducted for autoimmune disease overall and for individual conditions with at least 15 cases of ADHD among exposed children. Models were adjusted for child sex, and potential effect modification by child sex was examined by modeling relevant interaction terms (P < .05). To examine the effect of limited follow-up time for children born late in the study period, analyses were repeated on a subset of the cohort with at least 8 years of follow-up. The proportional hazards assumption was assessed by plotting scaled Schoenfeld residuals for all covariates, and all were within limits. All analyses were performed using SAS, version 9.4 software (SAS Institute) from January 13 to April 20, 2020. All P values were 2 sided, and P < .05 was considered significant.
Systematic Review
Search Strategy
A systematic search of the MEDLINE, Embase, and Web of Science databases was performed to identify all studies published before November 20, 2019 (eTable 2 in the Supplement). Reference lists of identified review articles and included studies were searched to identify additional eligible studies.
Study Eligibility
To be eligible for inclusion studies had to meet all of the following criteria:
Cohort or case-control design to test the association between 1 or more autoimmune diseases and ADHD. Studies without a disease-free or exposure-free comparison group were excluded.
Maternal autoimmune disease as a study exposure established by 1 of the following: (1) results from diagnostic tests or assessments or specialty clinic attendance, (2) diagnosis code documented in linked records or medical records, or (3) self-report of past diagnosis. In addition to specific autoimmune thyroid conditions (ie, Graves disease and Hashimoto thyroiditis), studies of hyperthyroidism and hypothyroidism were also included, because the majority of these cases among women of reproductive age are autoimmune in nature.22,23
ADHD diagnosis in the child established by 1 of the following: (1) diagnosis assigned by a medical professional or diagnosis code documented in linked records or (2) receipt of medication for the treatment of ADHD documented in a prescription database.
Study Screening and Data Extraction
Assessment of studies for inclusion and data extraction were performed independently by 2 reviewers (T.C.N. and S.J.L.). Any differences were resolved by discussion, with a third reviewer (N.N.) consulted to resolve disagreements. Corresponding authors were contacted in the event of missing information. Data extraction was performed using a standardized data extraction form and included study design, inclusion criteria, definition of autoimmune exposure, definition of ADHD outcome, analysis methods used, additional covariates controlled for, treatment of missing data, reported effect measure estimates, and CIs.
Study Analysis
All included studies were assessed for study quality using the Newcastle-Ottawa Quality Assessment Scale (NOS).24 The NOS is an assessment tool for observational studies that evaluates the selection of study groups, comparability of study groups, and assessment of exposure (case-control studies) or outcome (cohort studies). Reviewers assign a score for each domain (0-4 points for selection, 0-2 points for comparability, 0-3 points for exposure or outcome) and a total score of 0 to 9 points with a higher score indicating higher study quality. Both reviewers assigned NOS scores independently and then resolved any differences by discussion to assign a final score. A meta-analysis of the association of any maternal autoimmune disease as well as specific conditions with offspring ADHD diagnosis reported by 2 or more studies was conducted by calculating pooled HRs using a random-effects model and inverse-variance weighting.25,26 Heterogeneity between studies was examined by calculating I2 statistics. Meta-analysis calculations were conducted from February 10 to April 20, 2020, using Review Manager (RevMan), version 5.3 software (Cochrane Collaboration). All P values were 2 sided, and P < .05 was considered significant.
Results
Population-Based Cohort Study
A total of 916 257 children were born during the study period; 15 970 (1.7%) were excluded due to missing demographic variables, and 68 569 (7.5%) were excluded as preterm or multiple births. The remaining 831 718 singleton, term infants were born to 831 718 mothers (mean [SD] age, 29.8 [5.6] years), and 12 767 (1.5%) were linked to a maternal autoimmune diagnosis. Mothers with an autoimmune disease were more likely to also have a mental health diagnosis (2983 of 12 767 [23.4%] vs 109 272 of 818 951 [13.3%]). Propensity score matching produced a balanced study cohort of 12 610 exposed children and 50 440 matched controls for a total study cohort of 63 050 children. Maternal characteristics are presented in Table 1.
Table 1. Maternal Characteristics by Autoimmune Status in NSW, Australia (July 2000-December 2010).
| Characteristic | Full population (n = 831 718) | Propensity score–matched sample (n = 63 050)a | ||||
|---|---|---|---|---|---|---|
| Autoimmune diagnosis, No. (%) | ASD | Autoimmune diagnosis, No. (%) | ASD | |||
| Yes (n = 12 767) | No (n = 818 951) | Yes (n = 12 610) | No (n = 50 440) | |||
| Age at birth, y | ||||||
| <20 | 317 (2.5) | 30 945 (3.8) | 0.104 | 304 (2.4) | 1253 (2.5) | 0.006 |
| 20-24 | 1589 (12.5) | 116 726 (14.3) | 1569 (12.4) | 6242 (12.4) | ||
| 25-29 | 3483 (27.3) | 229 598 (28.0) | 3465 (27.5) | 13 897 (27.6) | ||
| 30-34 | 4447 (34.8) | 270 119 (33.0) | 4409 (35.0) | 17 615 (34.9) | ||
| 35-39 | 2433 (19.1) | 142 856 (17.4) | 2397 (19.0) | 9552 (18.9) | ||
| ≥40 | 498 (3.9) | 28 707 (3.5) | 466 (3.7) | 1881 (3.7) | ||
| Country of birth | ||||||
| Australia or New Zealand | 9955 (78.0) | 579 052 (70.7) | 0.167 | 9861 (78.2) | 39 437 (78.2) | <0.001 |
| Other | 2812 (22.0) | 239 899 (29.3) | 2749 (21.8) | 11 003 (21.8) | ||
| Parity, birth | ||||||
| First | 5337 (41.8) | 340 062 (41.5) | 0.061 | 5264 (41.7) | 20 996 (41.6) | 0.001 |
| Prior | 7430 (58.2) | 478 889 (58.5) | 7346 (58.3) | 29 444 (58.4) | ||
| Mental health diagnosis | ||||||
| Yes | 2983 (23.4) | 109 272 (13.3) | 0.261 | 2864 (22.7) | 11 501 (22.8) | 0.002 |
| No | 9784 (76.6) | 709 679 (86.7) | 9746 (77.3) | 38 939 (77.2) | ||
| Smoking in pregnancy | ||||||
| Yes | 1641 (12.9) | 110 712 (13.5) | 0.061 | 1557 (12.4) | 6224 (12.3) | 0.001 |
| No | 11 126 (87.2) | 708 239 (86.5) | 11 053 (87.7) | 44 216 (87.7) | ||
| Socioeconomic disadvantage, quintile | ||||||
| 1 (Most disadvantage) | 1963 (15.4) | 143 217 (17.5) | 0.079 | 1916 (15.2) | 7552 (15.0) | 0.010 |
| 2 | 1849 (14.5) | 129 245 (15.8) | 1818 (14.4) | 7341 (14.6) | ||
| 3 | 3098 (24.3) | 185 889 (22.7) | 3072 (24.4) | 12 418 (24.6) | ||
| 4 | 2454 (19.2) | 157 511 (19.2) | 2422 (19.2) | 9587 (19.0) | ||
| 5 (Least disadvantage) | 3403 (26.7) | 203 089 (24.8) | 3382 (26.8) | 13 542 (26.9) | ||
| Place of residence | ||||||
| Major city | 10 298 (80.7) | 640 416 (78.2) | 0.069 | 10 223 (81.1) | 40 875 (81.0) | 0.010 |
| Inner regional | 1922 (15.1) | 134 529 (16.4) | 1879 (14.9) | 7477 (14.8) | ||
| Outer regional | 511 (4.0) | 40 142 (4.9) | 491 (3.9) | 2000 (4.0) | ||
| Remote or very remote | 36 (0.3) | 3864 (0.5) | 17 (0.1) | 88 (0.2) | ||
Abbreviations: ASD, absolute standard difference; NSW, New South Wales.
Sample created by 1:4 optimized matching (caliper = 0.00001).
A total of 1094 of 32 294 male children (3.4%) and 332 of 30 756 female children (1.1%) in the study cohort were diagnosed with ADHD during follow-up, with 664 of 1094 (61%) and 190 of 332 (57%) first diagnosed before their eighth birthday. An ADHD diagnosis was more common among autoimmune disease–exposed than unexposed children of either sex (6.98 vs 5.48 per 1000 person-years for boys and 2.32 vs 1.70 per 1000 person-years for girls). Crude and adjusted HRs calculated for maternal autoimmune disease overall and for 12 specific conditions are presented in Table 2. Any maternal autoimmune disease (adjusted HR, 1.30; 95% CI, 1.15-1.46) was associated with an increase in the risk of ADHD diagnosis, as well as T1D (adjusted HR, 2.23; 95% CI, 1.66-3.00), psoriasis (adjusted HR, 1.66; 95% CI, 1.02-2.70), and rheumatic fever or rheumatic carditis (adjusted HR, 1.75; 95% CI, 1.06-2.89). Restricting the cohort to children with at least 8 years of follow-up did not appreciably change estimates (eTable 3 in the Supplement), and there was no evidence of effect modification by child sex (eTable 4 in the Supplement).
Table 2. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Disease Status in New South Wales, Australia (July 2000-December 2010).
| Maternal autoimmune condition | ADHD cases, No. | Incidence (per 1000 person-years) | HR (95% CI) | |||
|---|---|---|---|---|---|---|
| Exposed | Unexposed | Exposed | Unexposed | Crudea | Adjustedb | |
| Any autoimmune condition | 346 | 1080 | 4.67 | 3.63 | 1.29 (1.14-1.46) | 1.30 (1.15-1.46) |
| Graves disease | 58 | 207 | 4.38 | 3.87 | 1.14 (0.85-1.52) | 1.18 (0.88-1.58) |
| Type 1 diabetes | 66 | 125 | 7.99 | 3.65 | 2.20 (1.64-2.96) | 2.23 (1.66-3.00) |
| Crohn disease | 28 | 115 | 3.59 | 3.63 | 0.99 (0.65-1.50) | 0.99 (0.66-1.51) |
| Ulcerative colitis | 18 | 97 | 2.35 | 3.12 | 0.75 (0.46-1.25) | 0.73 (0.44-1.21) |
| Celiac disease | 32 | 99 | 5.52 | 4.17 | 1.33 (0.88-2.01) | 1.32 (0.87-1.99) |
| Systemic lupus erythematosus | 20 | 76 | 3.68 | 3.39 | 1.09 (0.68-1.75) | 1.13 (0.71-1.82) |
| Immune thrombocytopenic purpura | 15 | 64 | 3.08 | 3.20 | 0.97 (0.55-1.69) | 0.93 (0.54-1.63) |
| Autoimmune thyroiditis | 15 | 54 | 3.70 | 3.21 | 1.16 (0.65-2.08) | 1.17 (0.65-2.11) |
| Psoriasis | 23 | 60 | 7.34 | 4.55 | 1.62 (1.00-2.64) | 1.66 (1.02-2.70) |
| Rheumatic fever or carditis | 21 | 50 | 6.57 | 3.73 | 1.77 (1.07-2.94) | 1.75 (1.06-2.89) |
| Multiple sclerosis | 15 | 52 | 5.06 | 4.17 | 1.22 (0.69-2.18) | 1.27 (0.71-2.25) |
| Rheumatoid arthritis | 17 | 51 | 5.99 | 4.26 | 1.42 (0.82-2.45) | 1.38 (0.79-2.41) |
Abbreviations: ADHD, attention-deficit/hyperactivity disorder; HR, hazard ratio.
Cohort propensity score matched to control for year of birth, parity, maternal age, country of birth, maternal mental health conditions, socioeconomic disadvantage, place of residence, and smoking during pregnancy.
Regression model additionally adjusted for child sex.
For the sensitivity analysis, the cohort was restricted to exposed children whose mothers were diagnosed before or within 180 days of birth (n = 9700) and their matched controls (n = 38 800) (eTable 5 in the Supplement). Although any maternal autoimmune disease (HR, 1.28; 95% CI, 1.10-1.48) and T1D (HR, 1.95; 95% CI, 1.38-2.76) remained associated with ADHD diagnosis, the associations for psoriasis (HR, 1.67; 95% CI, 0.93-3.00) and rheumatic fever or rheumatic carditis (HR, 1.35; 95% CI, 0.73-2.49) did not persist (eTable 6 in the Supplement), and there was no evidence of effect modification by child sex (eTable 7 in the Supplement).
Systematic Review
A total of 131 unique studies were identified, of which 6 (including the present cohort study) were included in the systematic review (Figure 1).17,18,19,20,27 Included studies were conducted in the Nordic countries (4 studies),17,18,19,27 North America (1 study),20 and Australia (present study), and all were of high quality (NOS 8-9) (Table 3). Most studies focused on specific autoimmune conditions; only the Nielsen et al17 2017 Danish cohort study and the present population-based study estimated the association with maternal autoimmune disease overall.
Figure 1. PRISMA Flowchart.
Articles excluded at the screening stage did not meet study inclusion criteria based on review of the title and abstract. Articles excluded after full-text review included studies of maternal thyroid hormone levels, studies without a comparison group of children, and studies with incompatible outcomes. PRISMA indicates Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
Table 3. Studies of Maternal Autoimmune Disease and Offspring Attention-Deficit/Hyperactivity Disorder Included in Systematic Review and Meta-Analysis.
| Source | Location and study period | Study design and population | Autoimmune exposure | Outcome | Covariates controlled | Effect estimate (95% CI) | NOS score |
|---|---|---|---|---|---|---|---|
| Nielsen et al, 2020 (current study) | Australia; children born 2000-2010 | Cohort; 12 610 exposed; 63 010 total | Any AD (35 specified conditions in state-hospital records) | ADHD (ICD-10-AM F90 or F98.8 in hospital records or ≥1 stimulant authorization or rx) | Birth year, parity, mat age, country of birth, mat mental health conditions, socioeconomic disadvantage, place of residence, mat smoking, child sex | HR, 1.30 (1.15-1.46) | 9 |
| Xiang et al,20 2018 | United States; children born 1995-2012 | Cohort; 522 exposed; 333 182 total | T1D (medical records and insulin prescription) | ADHD (ICD-9 314.X or ≥2 medication refills) in electronic medical records | Birth year, mat age, parity, mat education, mat ethnicity/race, area-level median household income, mat comorbidity, mat ADHD, child sex | HR, 1.56 (1.09-2.25) | 9 |
| Ji et al,19 2018 | Sweden; exposure diagnosed 1970-2012 | Cohort; 15 615 exposed; 1 396 444 total | T1D (ICD-8 250, ICD-9 250, or ICD-10 E10 before age 20) in inpatient and outpatient register | Primary diagnosis of ADHD (ICD-9 314 or ICD-10 F90) in hospital discharge data | Child year of birth, child sex, parental ADHD, parental psych disorder, parental education, parental age, disposable income, mat UTI, mat smoking, mat hypoglycemia, small for gestational age, APGAR score | HR, 1.35 (1.18-1.55) | 9 |
| Nielsen et al,17 2017 | Denmark; children born 1990-2007 | Cohort; 983 680 total; time-varying exposure | Any AD (30 specified conditions in national inpatient and outpatient hospital register) | ADHD (ICD-10 F90.X or F98.8) in national psychiatric register and hospital register | Calendar year, child age, child sex, parental psych history | IRR, 1.12 (1.06-1.19) | 9 |
| Instanes et al,18 2017a | Norway; children born 1967-2008 | Case-control; 47 944 cases; 2 322 657 total | MS (birth record) | ADHD (medication dispensing at >3 y old during 2004-2012) | Mat age, parity, year of birth, mat marital status, mat education | OR, 1.80 (1.20-2.50) | 8 |
| RA (birth record) | OR, 1.70 (1.50-1.90) | ||||||
| T1D (birth record) | OR, 1.60 (1.30-2.00) | ||||||
| Hypothyroidism (birth record) | OR, 1.20 (1.10-1.40) | ||||||
| Hyperthyroidism (birth record) | OR, 1.20 (0.90-1.50) | ||||||
| Andersen et al,27 2014 | Denmark; children born 1991-2004 | Cohort; 30 295 exposed; 857 014 total | Hyperthyroidism (ICD-8 242.00-242.29; ICD-10 E05.0-E05.9, exc. E05.4, E05.8A, E05.9A) in national in/outpatient register and ≥1 antithyroid prescription (ATC 'H03B') in pharmacy register OR ≥2 antithyroid prescriptions | ADHD (ICD-10 F90) in hospital or central psychiatric register or ≥2 dispensed prescriptions (ATC 'N06BA') in pharmacy register after age 3 y | Child sex, year of birth, mat age, mat place of birth, mat residence, mat cohabitation, mat income, parity | HR, 1.18 (1.03-1.36) | 9 |
| Hypothyroidism (ICD-8 243.99 and 244.00-244.09, exc. 244.02; ICD-10 E03.0-E03.9 and E89.0, exc. E03.0A, E03.1B, E03.4) in national in/outpatient register, ≥1 thyroid hormone prescription (ATC 'H03A'), and no antithyroid prescriptions in pharmacy register OR ≥2 thyroid hormone and no antithyroid prescriptions | HR, 1.10 (0.98-1.25) |
Abbreviations: AD, autoimmune disease; ADHD, attention-deficit/hyperactivity disorder; APGAR, appearance, pulse, grimace, activity, and respiration scoring system; HR, hazard ratio; ICD, International Statistical Classification of Diseases; ICD-19-AM, International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification; IRR, incidence rate ratio; mat, maternal; MS, multiple sclerosis; NOS, Newcastle-Ottawa Quality Assessment Scale; OR, odds ratio; psych, psychological; RA, rheumatoid arthritis; rx, prescription; T1D, type 1 diabetes; UTI, urinary tract infection.
Instanes et al18 was not included in the meta-analysis.
Five retrospective cohort studies were included in the meta-analysis, including the present study,17,19,20,27 with 11 specific conditions reported in 2 or more studies. The case-control study by Instanes et al18 was excluded because it identified ADHD from medications dispensed at different ages, including only in adulthood. Pooled HRs were calculated for each condition and are presented with forest plots in Figure 2. Any autoimmune disease (2 studies: HR, 1.20; 95% CI, 1.03-1.38), T1D (4 studies: HR, 1.53; 95% CI, 1.27-1.85), hyperthyroidism (3 studies: HR, 1.15; 95% CI, 1.06-1.26), and psoriasis (2 studies: HR, 1.31; 95% CI, 1.10-1.56) were all associated with ADHD.
Figure 2. Pooled Hazard Ratios of ADHD by Maternal Autoimmune Disease Status Across 5 Studies.
The figure shows the published effect estimates from 5 studies and pooled hazard ratios from the meta-analysis. Estimates are presented in forest plots for any maternal autoimmune disease and all specific conditions reported in 2 or more studies. The squares represent individual study estimates, and the diamonds represent pooled estimates, both with 95% CIs. ADHD indicates attention-deficit/hyperactivity disorder; df, degrees of freedom; IV, inverse variance.
Discussion
This study found an association between maternal autoimmune disease and ADHD among children both in a population-based cohort study and a combined systematic review and meta-analysis. In the cohort study, a maternal diagnosis of any autoimmune disease, T1D, rheumatic fever or rheumatic carditis, or psoriasis during the study period was associated with an increased risk of ADHD. When combined with previous studies in the meta-analyses, hyperthyroidism, as well as any autoimmune disease, T1D, and psoriasis, were associated with an increased risk of ADHD. Restricting exposure to maternal diagnoses before or shortly after birth produced similar results, although estimates were more variable among conditions less prevalent in the cohort; this was most likely the result of smaller study numbers.
Any maternal autoimmune disease was associated with ADHD in the cohort study and in the meta-analyses of 2 studies. Another study17 included in the meta-analysis reported a similar but smaller association based on ICD-10-AM codes for 30 specific autoimmune conditions. There was evidence of considerable heterogeneity between studies, potentially owing to differences in exposure definitions. Our cohort study identified maternal autoimmune disease using inpatient hospital admissions both before and after birth, whereas Nielsen et al17 defined autoimmune status before birth based on both inpatient and outpatient records. When the results of these 2 studies were combined for individual conditions, lower heterogeneity was reported.
Maternal T1D exhibited the largest HRs in our cohort study, and it was the most commonly studied condition identified in the systematic review. Three large population-based studies based on administrative data came to similar conclusions and reported similar effect estimates that were smaller than in our cohort study,17,19,20 possibly owing to different methods of identifying T1D including insulin prescription20 and outpatient records.17,19 The observed association may also be related to nonimmune aspects of T1D, such as glycemic control, as nonautoimmune diabetes has been associated with ADHD among children.20 Similarly, autoimmune thyroid conditions have both immune and endocrine aspects. There is a larger body of literature on the association of general thyroid dysfunction during pregnancy and ADHD. A 2017 systematic review included 7 studies focused on ADHD and concluded that both overactive and underactive thyroid function during pregnancy were associated with ADHD among offspring.28 Given the observational design of our cohort study and studies included in our systematic review, we cannot draw specific conclusions related to causal mechanisms. Of the specific conditions reported by only 2 studies, only maternal psoriasis was associated with ADHD when calculating pooled HRs. Additional studies are needed for many specific conditions that are either rare or underidentified in administrative data.
Our results were consistent with previous research reporting an association between autoimmune disease and impaired mental health, particularly depression and psychosis,29,30,31 and between maternal autoimmune disease and child neurodevelopmental disorders, including autism spectrum disorder,14 obsessive compulsive disorder, and tics or Tourette syndrome.32 Results were also consistent with the general hypothesis that adverse maternal immune function during pregnancy alters fetal neurodevelopment via direct action of cytokines and autoantibodies, epigenetic modulation, or microglia activation. The observed association may also be a product of shared genetic vulnerability between autoimmune diseases and ADHD. Attention-deficit/hyperactivity disorder is highly heritable,33 and population-based cohort studies have reported associations between allergic and autoimmune diseases and ADHD within individuals.17,34 A pooled copy number variant study found an association between ADHD and immune-related pathways, although it was not robust to sensitivity analysis.35 The interaction between inflammation, genes, and environment and ADHD requires further evaluation.
Strengths and Limitations
This study has many strengths, including the use of a hybrid study design that both summarizes and adds to the published literature. The population-based study cohort was large, and restricting to singleton, term births avoided issues of confounding or mediation by gestational age and complications associated with twin pregnancy. This study also has limitations, most significantly the lack of outpatient or primary care records for identifying maternal autoimmune exposure, which is likely underascertained, particularly for disorders such as psoriasis that do not commonly result in hospitalization. We also had no data on whether women had controlled disease and what medications they received during pregnancy, both of which might influence the association with ADHD in the child. Similarly, although stimulant medications are the most common treatment for ADHD, we could not identify children with ADHD not receiving these medications, which may have led to a more severely impaired cohort or children who left NSW and received medication treatment elsewhere. We also were unable to fully assess symptom severity, domains (eg, hyperactivity and inattention subtypes), or comorbidities among ADHD cases that may complicate the association. Additionally, defining exposure based on maternal diagnoses at any time in the study period allowed for autoimmune disease to be documented after the child’s first ADHD event and may have introduced selection bias. However, our results remained robust after restricting to diagnoses before or shortly after birth. Finally, given differences in study design and definitions, the pooled HRs presented in the meta-analysis need to be treated cautiously. However, all included studies were of high quality, and random-effects models were used.25
Conclusions
In this cohort study, maternal autoimmune disease was associated with increased risk of ADHD among offspring. Our study provides justification for future studies that examine the effect of maternal autoimmune diseases, including biomarkers, condition severity, and management in pregnancy and in the periconception period, on neurodevelopmental disorders in children. It also highlights the importance of high-quality multidisciplinary care for women with autoimmune diseases and their children. Health care professionals should discuss reproductive goals with women who have autoimmune disease, and clinicians should encourage planning pregnancies when the disease is stable and well managed.36,37 Children of women with autoimmune disease may benefit from additional follow-up and support for developmental issues. The causes of neurodevelopmental disorders are complex and multifactorial; however, our study suggests maternal autoimmunity may represent one avenue for further investigation.
eTable 1. ICD-10AM Codes Used to Identify Maternal Autoimmune Disease
eTable 2. Search Strategy for Systematic Review of Studies of Maternal Autoimmune Disease and Offspring Attention-Deficit/Hyperactivity Disorder (ADHD)
eTable 3. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Diagnosis Among Children With at Least 8 Years of Follow-Up, NSW, Australia (July 2000–December 2007)
eTable 4. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Disease Status and Child Sex in NSW, Australia (July 2000–December 2010)
eTable 5. Maternal Characteristics by Autoimmune Diagnosis Prior to or Within 180 Days of Birth, NSW, Australia (July 2000–December 2010)
eTable 6. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Diagnosis Prior to or Within 180 days of Birth, NSW, Australia (July 2000–December 2010)
eTable 7. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Disease Status Prior to or Within 180 Days of Birth and Child Sex in NSW, Australia (July 2000–December 2010)
References
- 1.Davidson A, Diamond B. Autoimmune diseases. N Engl J Med. 2001;345(5):340-350. doi: 10.1056/NEJM200108023450506 [DOI] [PubMed] [Google Scholar]
- 2.Cooper GS, Bynum ML, Somers EC. Recent insights in the epidemiology of autoimmune diseases: improved prevalence estimates and understanding of clustering of diseases. J Autoimmun. 2009;33(3-4):197-207. doi: 10.1016/j.jaut.2009.09.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Eaton WW, Rose NR, Kalaydjian A, Pedersen MG, Mortensen PB. Epidemiology of autoimmune diseases in Denmark. J Autoimmun. 2007;29(1):1-9. doi: 10.1016/j.jaut.2007.05.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hayter SM, Cook MC. Updated assessment of the prevalence, spectrum and case definition of autoimmune disease. Autoimmun Rev. 2012;11(10):754-765. doi: 10.1016/j.autrev.2012.02.001 [DOI] [PubMed] [Google Scholar]
- 5.Jacobson DL, Gange SJ, Rose NR, Graham NM. Epidemiology and estimated population burden of selected autoimmune diseases in the United States. Clin Immunol Immunopathol. 1997;84(3):223-243. doi: 10.1006/clin.1997.4412 [DOI] [PubMed] [Google Scholar]
- 6.Ngo ST, Steyn FJ, McCombe PA. Gender differences in autoimmune disease. Front Neuroendocrinol. 2014;35(3):347-369. doi: 10.1016/j.yfrne.2014.04.004 [DOI] [PubMed] [Google Scholar]
- 7.Kipnis J. Multifaceted interactions between adaptive immunity and the central nervous system. Science. 2016;353(6301):766-771. doi: 10.1126/science.aag2638 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Estes ML, McAllister AK. Immune mediators in the brain and peripheral tissues in autism spectrum disorder. Nat Rev Neurosci. 2015;16(8):469-486. doi: 10.1038/nrn3978 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Knuesel I, Chicha L, Britschgi M, et al. Maternal immune activation and abnormal brain development across CNS disorders. Nat Rev Neurol. 2014;10(11):643-660. doi: 10.1038/nrneurol.2014.187 [DOI] [PubMed] [Google Scholar]
- 10.Estes ML, McAllister AK. Maternal immune activation: Implications for neuropsychiatric disorders. Science. 2016;353(6301):772-777. doi: 10.1126/science.aag3194 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Nardone S, Elliott E. The interaction between the immune system and epigenetics in the etiology of autism spectrum disorders. Front Neurosci. 2016;10:329. doi: 10.3389/fnins.2016.00329 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gumusoglu SB, Stevens HE. Maternal inflammation and neurodevelopmental programming: a review of preclinical outcomes and implications for translational psychiatry. Biol Psychiatry. 2019;85(2):107-121. doi: 10.1016/j.biopsych.2018.08.008 [DOI] [PubMed] [Google Scholar]
- 13.Thapar A, Cooper M, Rutter M. Neurodevelopmental disorders. Lancet Psychiatry. 2017;4(4):339-346. doi: 10.1016/S2215-0366(16)30376-5 [DOI] [PubMed] [Google Scholar]
- 14.Chen SW, Zhong XS, Jiang LN, et al. Maternal autoimmune diseases and the risk of autism spectrum disorders in offspring: a systematic review and meta-analysis. Behav Brain Res. 2016;296:61-69. doi: 10.1016/j.bbr.2015.08.035 [DOI] [PubMed] [Google Scholar]
- 15.American Psychiatric Association . Neurodevelopmental Disorders: DSM-5 Selections. American Psychiatric Association; 2016. [Google Scholar]
- 16.Sayal K, Prasad V, Daley D, Ford T, Coghill D. ADHD in children and young people: prevalence, care pathways, and service provision. Lancet Psychiatry. 2018;5(2):175-186. doi: 10.1016/S2215-0366(17)30167-0 [DOI] [PubMed] [Google Scholar]
- 17.Nielsen PR, Benros ME, Dalsgaard S. Associations between autoimmune diseases and attention-deficit/hyperactivity disorder: a nationwide study. J Am Acad Child Adolesc Psychiatry. 2017;56(3):234-240.e1. doi: 10.1016/j.jaac.2016.12.010 [DOI] [PubMed] [Google Scholar]
- 18.Instanes JT, Halmøy A, Engeland A, Haavik J, Furu K, Klungsøyr K. Attention-deficit/hyperactivity disorder in offspring of mothers with inflammatory and immune system diseases. Biol Psychiatry. 2017;81(5):452-459. doi: 10.1016/j.biopsych.2015.11.024 [DOI] [PubMed] [Google Scholar]
- 19.Ji J, Chen T, Sundquist J, Sundquist K. Type 1 diabetes in parents and risk of attention deficit/hyperactivity disorder in offspring: a population-based study in Sweden. Diabetes Care. 2018;41(4):770-774. doi: 10.2337/dc17-0592 [DOI] [PubMed] [Google Scholar]
- 20.Xiang AH, Wang X, Martinez MP, et al. Maternal gestational diabetes mellitus, type 1 diabetes, and type 2 diabetes during pregnancy and risk of ADHD in offspring. Diabetes Care. 2018;41(12):2502-2508. doi: 10.2337/dc18-0733 [DOI] [PubMed] [Google Scholar]
- 21.Austin PC. The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med. 2014;33(7):1242-1258. doi: 10.1002/sim.5984 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Cooper DS, Laurberg P. Hyperthyroidism in pregnancy. Lancet Diabetes Endocrinol. 2013;1(3):238-249. doi: 10.1016/S2213-8587(13)70086-X [DOI] [PubMed] [Google Scholar]
- 23.Teng W, Shan Z, Patil-Sisodia K, Cooper DS. Hypothyroidism in pregnancy. Lancet Diabetes Endocrinol. 2013;1(3):228-237. doi: 10.1016/S2213-8587(13)70109-8 [DOI] [PubMed] [Google Scholar]
- 24.Wells GA, Shea B, O'Connell D, et al. The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomised studies in meta-analyses. Accessed March 24, 2020. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp
- 25.Greenland S. Quantitative methods in the review of epidemiologic literature. Epidemiol Rev. 1987;9:1-30. doi: 10.1093/oxfordjournals.epirev.a036298 [DOI] [PubMed] [Google Scholar]
- 26.DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7(3):177-188. doi: 10.1016/0197-2456(86)90046-2 [DOI] [PubMed] [Google Scholar]
- 27.Andersen SL, Laurberg P, Wu CS, Olsen J. Attention deficit hyperactivity disorder and autism spectrum disorder in children born to mothers with thyroid dysfunction: a Danish nationwide cohort study. BJOG. 2014;121(11):1365-1374. doi: 10.1111/1471-0528.12681 [DOI] [PubMed] [Google Scholar]
- 28.Fetene DM, Betts KS, Alati R. Mechanisms in endocrinology: maternal thyroid dysfunction during pregnancy and behavioural and psychiatric disorders of children: a systematic review. Eur J Endocrinol. 2017;177(5):R261-R273. doi: 10.1530/EJE-16-0860 [DOI] [PubMed] [Google Scholar]
- 29.Derry HM, Padin AC, Kuo JL, Hughes S, Kiecolt-Glaser JK. Sex differences in depression: does inflammation play a role? Curr Psychiatry Rep. 2015;17(10):78. doi: 10.1007/s11920-015-0618-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Pollak TA, Lennox BR, Müller S, et al. Autoimmune psychosis: an international consensus on an approach to the diagnosis and management of psychosis of suspected autoimmune origin. Lancet Psychiatry. 2020;7(1):93-108. doi: 10.1016/S2215-0366(19)30290-1 [DOI] [PubMed] [Google Scholar]
- 31.Sæther SG, Rø ADB, Larsen JB, Vaaler A, Kondziella D, Reitan SK. Biomarkers of autoimmunity in acute psychiatric disorders. J Neuropsychiatry Clin Neurosci. 2019;31(3):246-253. doi: 10.1176/appi.neuropsych.18040069 [DOI] [PubMed] [Google Scholar]
- 32.Mataix-Cols D, Frans E, Pérez-Vigil A, et al. A total-population multigenerational family clustering study of autoimmune diseases in obsessive-compulsive disorder and Tourette’s/chronic tic disorders. Mol Psychiatry. 2018;23(7):1652-1658. doi: 10.1038/mp.2017.215 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Faraone SV, Larsson H. Genetics of attention deficit hyperactivity disorder. Mol Psychiatry. 2019;24(4):562-575. doi: 10.1038/s41380-018-0070-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Chen M-H, Su T-P, Chen Y-S, et al. Comorbidity of allergic and autoimmune diseases among patients with ADHD. J Atten Disord. 2017;21(3):219-227. doi: 10.1177/1087054712474686 [DOI] [PubMed] [Google Scholar]
- 35.Thapar A, Martin J, Mick E, et al. Psychiatric gene discoveries shape evidence on ADHD’s biology. Mol Psychiatry. 2016;21(9):1202-1207. doi: 10.1038/mp.2015.163 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Borchers AT, Naguwa SM, Keen CL, Gershwin ME. The implications of autoimmunity and pregnancy. J Autoimmun. 2010;34(3):J287-J299. doi: 10.1016/j.jaut.2009.11.015 [DOI] [PubMed] [Google Scholar]
- 37.Tincani A, Dall’Ara F, Lazzaroni MG, Reggia R, Andreoli L. Pregnancy in patients with autoimmune disease: A reality in 2016. Autoimmun Rev. 2016;15(10):975-977. doi: 10.1016/j.autrev.2016.07.017 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. ICD-10AM Codes Used to Identify Maternal Autoimmune Disease
eTable 2. Search Strategy for Systematic Review of Studies of Maternal Autoimmune Disease and Offspring Attention-Deficit/Hyperactivity Disorder (ADHD)
eTable 3. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Diagnosis Among Children With at Least 8 Years of Follow-Up, NSW, Australia (July 2000–December 2007)
eTable 4. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Disease Status and Child Sex in NSW, Australia (July 2000–December 2010)
eTable 5. Maternal Characteristics by Autoimmune Diagnosis Prior to or Within 180 Days of Birth, NSW, Australia (July 2000–December 2010)
eTable 6. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Diagnosis Prior to or Within 180 days of Birth, NSW, Australia (July 2000–December 2010)
eTable 7. Hazard Ratios of Attention-Deficit/Hyperactivity Disorder by Maternal Autoimmune Disease Status Prior to or Within 180 Days of Birth and Child Sex in NSW, Australia (July 2000–December 2010)