Adverse childhood experiences are associated with disease and mental health outcomes in childhood-onset SLE

Abstract

Objective

To investigate the relationship between adverse childhood experiences (ACEs) and health outcomes in childhood-onset SLE (cSLE).

Methods

We conducted a retrospective cohort study of cSLE outpatients aged 9–18 years at a Canadian tertiary centre from July 2018 to June 2020. ACEs data were collected from electronic medical records, categorised according to the US Centers for Disease Control-Kaiser framework and in relationship to time of cSLE diagnosis. Primary outcomes included disease activity, disease damage and comorbid psychiatric diagnoses (not attributable to cSLE). Secondary outcomes included neuropsychiatric SLE (NPSLE), lupus nephritis and arthritis. Regression models examined the relationship between (1) presence (yes/no) and (2) timing of ACEs (none vs before vs after cSLE diagnosis) and the outcomes. Covariates included sex, ethnicity, age at diagnosis, disease duration, nephritis/NPSLE and glucocorticoid use.

Results

Of 224 patients with cSLE, 41% had documented ACEs. The presence of ACEs was associated with NPSLE (adjusted OR=3.40, 95% CI 1.55 to 7.78, p=0.003). ACEs occurring before diagnosis were associated with disease damage (adjusted OR 3.10, 95% CI 1.21 to 8.04, p=0.018) and NPSLE (adjusted OR=2.67, 95% CI 1.02 to 6.98, p=0.043). ACEs occurring after cSLE diagnosis were associated with comorbid psychiatric diagnosis (adjusted OR=5.01, 95% CI 1.20 to 21.12, p=0.025). ACEs were not associated with disease activity, nephritis or arthritis.

Conclusion

ACEs exposure was prevalent in this cSLE cohort, associated with NPSLE diagnosis, disease damage and comorbid psychiatric diagnoses. These findings suggest that early-life stressors may contribute to cSLE morbidity, particularly neuropsychiatric features, emphasising the importance of addressing psychosocial issues in the context of medical care.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Adverse childhood experiences (ACEs) are associated with an increased risk of developing autoimmune diseases such as SLE as well as poor mental health outcomes, but little is known about the relationship between ACEs and outcomes for patients with childhood-onset SLE (cSLE).

WHAT THIS STUDY ADDS

• This study found ACEs to be prevalent in 41% of children with cSLE, and those with ACEs exposure had higher odds of having a neuropsychiatric SLE (NPSLE) diagnosis.

• ACEs exposure before cSLE diagnosis was associated with greater organ damage and neuropsychiatric SLE (NPSLE), while exposure after diagnosis was associated with the presence of a comorbid psychiatric disorder.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The study findings provide preliminary evidence to support assessment for ACEs in clinical care of patients with cSLE, and underscore the importance of considering ACEs when evaluating neuropsychiatric manifestations and risk for disease damage.

• Future prospective studies are needed to examine the effect of cumulative ACEs over time in patients with cSLE, as well as the potential biological pathways linking ACEs to immune dysregulation and neuropsychiatric outcomes in cSLE.

Introduction

SLE is a chronic autoimmune disease with heterogeneous clinical manifestations of varying degrees of severity.¹ Childhood-onset SLE (cSLE) is characterised by onset of symptoms prior to 18 years of age and accounts for approximately 20% of all SLE cases.² Children and adolescents with SLE have more active disease, increased disease damage and require more aggressive immunosuppressive treatment compared with their adult counterparts.3,6 In cSLE, arthritis affects 61–71% of patients, and major organ disease is common with lupus nephritis affecting 20–53% and neuropsychiatric SLE (NPSLE) affecting 15–20% of patients.^{7 8} Additionally, 33% of patients with cSLE experience other psychiatric disorders such as comorbid depression and anxiety⁹ that are not attributable to NPSLE and have been linked to treatment non-adherence and poor health-related quality of life. Our limited understanding of factors contributing to disease heterogeneity and comorbid psychiatric disorders in cSLE presents ongoing challenges in patient care.

There is increasing literature indicating a dose–response relationship between chronic diseases and adverse childhood experiences (ACEs), defined as potentially traumatic events that occur during the first 18 years of life. These traumatic events include exposure to violence, abuse, neglect, parental separation or divorce, caregiver substance use, mental illness and criminal behaviour.¹⁰ ACEs can increase allostatic load, reflecting the body’s wear and tear from experiencing chronic stress.¹¹ This can in turn lead to epigenetic changes within the neuroendocrine-immune system,¹² contributing to inflammation^{13 14} and the development and progression of chronic autoimmune diseases like SLE.¹⁵ ACEs also present a threat to mental health, as they occur during critical developmental stages where the brain is highly vulnerable to effects of stress and trauma.¹⁶ ACEs are also associated with increased risk of anxiety and depression in adolescence.¹⁷ Moreover, ACEs can cause overactivation of the hypothalamic–pituitary axis, resulting in dysregulation of the body’s stress response and elevated susceptibility to psychopathology in adults.¹⁸

Recent studies show that exposure to ACEs is a significant risk factor for developing adult-onset SLE and nephritis in women.^{19 20} Evidence also indicates that high levels of ACEs exposure are associated with worse self-reported disease activity, flares and depression in adults with SLE.²¹ Furthermore, research suggests that young adults with a history of at least two ACEs carry an elevated risk of hospitalisation due to rheumatic diseases, including SLE.²² Additionally, the timing of ACEs critically shapes the development of psychiatric disorders, with evidence indicating that individuals with a history of ACEs in early and middle childhood experience more severe psychiatric symptoms.²³ Yet there remains limited understanding of the effects of ACEs timing in the context of SLE diagnosis, and their relationship to SLE disease outcomes and heterogeneity. This is particularly relevant for children who may be significantly impacted by the combined deleterious effects of ACEs and SLE at their vulnerable neurodevelopmental stage.

The objective of this cross-sectional study was to determine the association of ACEs with cSLE outcomes. Specifically, we examined the relationship between presence and timing of ACEs, relative to cSLE diagnosis and disease outcomes including disease activity, disease damage and psychiatric comorbidity. Our secondary aim examined the relationship between the presence and timing of ACEs and disease manifestations of NPSLE, nephritis and arthritis.

Patients and methods

Study design, participants and setting

We conducted a retrospective cohort study of patients from the Lupus Clinic at the Hospital for Sick Children (SickKids) in Toronto, Canada. Data were obtained from retrospective and prospective cohorts comprising a total of 224 patients. The retrospective cohort included 181 patients who attended the clinic at least once between July 2018 and July 2020. Patients from the prospective cohorts included those enrolled in a mental health study (n=19) between June 2017 and May 2019, and a brain health study (n=24) between January 2020 and March 2023. All patients were between 9–18 years old during the observation periods and met the classification criteria for SLE as defined by the American College of Rheumatology (ACR) or the SLE International Collaborating Clinics (SLICC).²⁴

Patient and public involvement

Patients and the public were not directly involved in the design or conduct of the study. They will be involved in the dissemination of the results to the patient community.

Study variables

ACEs exposure variable

The exposure of interest was ACEs guided by both the original ACEs study,²⁵ which used the US Centers for Disease Control (CDC)-Kaiser ACE framework comprised of conventional household (eg, abuse, neglect, domestic violence) indicators, and the Philadelphia ACE Study which expanded on the CDC-Kaiser framework to include additional community-level (eg, neighbourhood safety, community violence) adversities relevant to more socio-economically diverse populations.²⁶ In consideration of the racially and socio-economically diverse cSLE population, we used these frameworks to identify the following 18 types of household and community-level ACEs: separated/divorced parents, witnessing domestic violence, emotional abuse, emotional neglect, parental mental illness, parental alcohol/substance abuse, incarcerated parent, death of a parent, physical abuse, physical neglect, sexual abuse, bullying, racism/discrimination, living in an unsafe neighbourhood, witnessing violence, experiencing a violent act, other environmental trauma and other interpersonal trauma. ACEs data were collected using a systematic, multilayered approach. In collaboration with a neuropsychologist, a keyword search strategy (online supplemental table 1) for all 18 ACEs was created and used with the built-in search function of the Epic electronic health record (EHR) platform to identify ACEs across all EHR encounters and documentation. Thorough reviews of EHR notes from rheumatology, psychiatry, adolescent medicine and social work were completed to further identify the types and number of ACEs and timing in relationship to cSLE diagnosis. Additionally, the 13-item post-traumatic stress disorder segment of the Kiddie Schedule for Affective Disorders and Schizophrenia (K-SADS) was used to supplement the identification of ACEs for patients (n=43) who had completed the assessment during their research study visits. Traumatic events addressed by the K-SADS included physical and sexual abuse, witnessing domestic violence, being a victim of violence, being involved in a car/fire/other accident and experiencing war zone and terrorism-related trauma. ACEs identified in the K-SADS were compared with those found in the patient’s EHR to avoid the collection of duplicated ACEs. For patients who had not completed the K-SADS, only the EHR was used to identify ACEs.

The timing of ACEs relative to cSLE diagnosis was determined by comparing the clinician-confirmed date of diagnosis and the documented occurrence of ACEs. To investigate the effect of ACEs timing on cSLE outcomes, we classified ACEs as occurring before or after a cSLE diagnosis. Only ACEs with clearly defined timing were classified, and patients with undetermined ACE timing or with multiple ACEs occurring both before and after diagnosis were omitted from the timing classification.

Outcome variables

The primary outcomes of interest were the presence of (1) disease activity, (2) disease damage and (3) comorbid psychiatric diagnosis, inclusive of mood and/or anxiety disorder. Disease activity was measured using the SLE Disease Activity Index 2000 (SLEDAI-2K).²⁷ The adjusted mean SLEDAI-2K (AMS) was determined by calculating the area under the SLEDAI-2K curve over time during the study period (July 2018 to July 2020 for the retrospective cohort, and June 2017 to May 2019 and January 2020 to March 2023 for the two prospective cohorts). Disease damage was calculated using the SLICC/ACR Damage Index (SDI).²⁷ We defined the presence of disease damage as an SDI score of >0 at the latest clinic visit in the observation period for the retrospective cohort and at the baseline study visit for the prospective cohort. Comorbid psychiatric diagnoses encapsulated clinician documented psychiatric diagnoses, inclusive of mood and anxiety disorders, that were not attributed to cSLE.

The secondary outcomes of interest were the presence of clinical SLE manifestations for (1) lupus nephritis, confirmed by renal biopsy; (2) arthritis; and (3) NPSLE, based on clinician diagnosis and guided by the ACR neuropsychiatric SLE case definitions.²⁸ These manifestations were identified by clinician diagnosis and documentation in the EHR. Patient characteristics were collected to represent the latest point during the observation period: for the retrospective cohort, either 1 July 2020, or at the discharge date if the patient transitioned from paediatric to adult care services prior to that date; for the prospective cohort, the date of study visit.

Demographic and disease variables

Patient demographic and disease characteristics were obtained through EHR and included age, sex, ethnicity, socio-economic status (SES), age at disease diagnosis, disease duration, presence of major organ disease and any-time use of glucocorticoids. Ethnicity was based on the census categories for country of family origin used by Statistics Canada,²⁹ categorised as European (includes British, French, Northern, Southern, Eastern and Western European), Asian (includes East Asian, South Asian, Southeast Asian, West Asian and Pacific Islander), Arabic, African and Caribbean, American (includes Indigenous and Latin American), Multiethnic (indicating more than one ethnic origin) and Other (self-identified other ethnic origin). SES was assessed using the 2021 Ontario Marginalisation Index dimension for material resources, which evaluates the level of community poverty.³⁰ Patient postal codes from medical charts were converted into dissemination areas using the Postal Code Conversion File.³¹ Disease duration was defined as the period from the date of diagnosis by ACR or SLICC classification criteria to the last clinic visit during the observation period. Major organ disease included nephritis and/or NPSLE, as defined above. Glucocorticoid use was determined from the date of diagnosis until the last clinic visit during the observation period.

Statistical analysis

Demographic and disease characteristics were tabulated using descriptive statistics. ACEs were modelled in two ways: (Model 1) binary ACEs exposure (present/absent), and (Model 2) categorical exposure (none vs ACEs only before cSLE diagnosis vs ACEs only after cSLE diagnosis). For Model 2, patients with non-specific timing of ACEs (either indeterminate or occurring both before and after diagnosis) were excluded from analyses. Univariate and multivariable linear regression models examined the relationship between ACEs exposure and disease activity. Univariate and multivariable logistic regression models examined associations between ACEs exposure and disease damage, comorbid psychiatric diagnoses and clinical manifestations in cSLE. Multivariable models for disease activity and disease damage outcomes were adjusted for age at diagnosis, sex, ethnicity and disease duration. The multivariable model for the comorbid psychiatric diagnosis outcome adjusted for the same covariates with the addition of any time glucocorticoid use. Patients with an NPSLE diagnosis were excluded from the analyses due to the high frequency of coexisting NPSLE diagnoses and comorbid psychiatric diagnoses deemed not definitively attributable to SLE.³² The multivariable models for the clinical manifestation outcomes adjusted for age at diagnosis, sex, ethnicity, disease duration, any time glucocorticoid use, arthritis, nephritis and NPSLE (excluding the outcome manifestation). For all regression analyses, ethnicity was grouped into six mutually exclusive categories based on the ethnic distribution of the cohort: East Asian, South Asian, European, African/Caribbean, Multiethnic and Other. Missing data (<2%) was determined to be missing completely at random, and listwise deletion was deemed appropriate for the regression analyses. All analyses were performed using R Studio V.2024.04.2 statistical software. P values <0.05 were considered statistically significant.

Results

Demographic and disease characteristics for the 224 identified patients are summarised in table 1. The median age at disease diagnosis was 13.24 years (IQR 10.70–15.35), and the median disease duration was 1.37 years (IQR 0.13–4.27). The median AMS was 2.55 (IQR 0.68–4.99). Disease damage was present in 17.1% (n=38), comprised of 57 damage items across the following domains: neuropsychiatric (n=16 items including seizures, cognitive impairment and cerebrovascular), musculoskeletal (n=15 items including avascular necrosis (7), deforming arthritis/ruptured tendon (4), osteoporosis (4)), renal (n=7 items for proteinuria), ocular (n=7 items for cataracts (6) and retinal changes (1)), endocrine (n=7 items for premature gonadal failure (5) and diabetes (2)), peripheral vascular (n=2 items for venous thrombosis) and n=1 item each for gastrointestinal (resection), pulmonary (infarction) and skin (scarring alopecia). Comorbid psychiatric disorders were observed in 20.1% (n=45) of patients, primarily comprised of anxiety disorders in n=30 and mood disorders in n=27.

Table 1. Patient characteristics in adolescents with cSLE, n=224.

Baseline demographic characteristics	Full cohort n=224	ACEs n=92	No ACEs n=132
Age in years, median (IQR)	15.86 (13.88 to 17.00)	15.84 (13.73 to 17.09)	15.90 (14.37 to 16.98)
Female, n (%)	190 (84.8)	77 (83.7)	113 (85.6)
Ethnicity, n (%)
East Asian	55 (24.6)	14 (15.2)	41 (31.1)
South Asian	46 (20.5)	16 (17.4)	30 (22.7)
European	38 (17)	21 (22.8)	17 (12.9)
African/Caribbean	26 (11.6)	10 (10.9)	16 (12.1)
Multiethnic	27 (12.1)	13 (14.1)	14 (10.6)
Latin American	15 (6.7)	11 (12)	4 (3.0)
Arabic	13 (5.8)	4 (4.4)	9 (6.8)
Indigenous	2 (0.89)	2 (2.2)	0 (0)
Other	2 (0.89)	1 (1.1)	1 (0.8)
Material resources score (range −3 to +3), median (IQR)*	−0.18 (−0.57 to 0.45)	0.03 (−0.28 to 0.91)	−0.35 (−0.68 to 0.12)
Disease characteristics
Age at disease diagnosis in years, median (IQR)	13.24 (10.70 to 15.35)	13.54 (11.41 to 15.83)	12.95 (10.58 to 15.23)
Disease duration in years, median (IQR)	1.37 (0.13 to 4.27)	0.93 (0.10 to 3.59)	1.81 (0.16 to 4.67)
AMS, median (IQR)*	2.55 (0.68 to 4.99)	2.36 (0.57 to 5.13)	2.57 (0.71 to 4.55)
SLICC/ACR Damage Index (score >0), n (%)*	38 (17.1)	19 (20.7)	18 (13.7)
Arthritis	129 (57.6)	52 (56.5)	77 (58.3)
Nephritis	81 (36.2)	32 (34.8)	49 (37.1)
NPSLE	35 (15.6)	22 (23.9)	13 (9.85)
Psychosis	18 (51.4)	11 (50.0)	7 (53.8)
No psychosis	17 (48.6)	11 (50.0)	6 (46.2)
Glucocorticoid use (ever), n (%)	151 (67.4)	64 (69.6)	87 (65.9)
Any comorbid psychiatric diagnosis†, n (%)	45 (20.1)	24 (26.1)	21 (15.9)
Depressive disorder	27 (12.1)	16 (17.4)	11 (8.33)
Anxiety disorder	30 (13.4)	17 (18.5)	13 (9.85)
Other‡	7 (3.1)	5 (5.4)	2 (1.5)

^*Data were missing for the following variables: material resources score (n=3), AMS (n=2), SLICC Damage Index (n=2).

^†Psychiatric comorbidities that are not attributable to NPSLE.

^‡Other comorbid psychiatric diagnoses included: psychotic disorder (n=1), neurocognitive disorder (n=1), substance use disorder (n=1), eating/feeding disorder (n=4).

ACE, adverse childhood experience; AMS, adjusted mean SLEDAI; cSLE, childhood-onset SLE; NPSLE, neuropsychiatric SLE; SLEDAI, SLE Disease Activity Index; SLICC/ACR, Systemic Lupus International Collaborating Clinics/American College of Rheumatology.

Arthritis was present in 57.6% (n=129) of patients, nephritis in 36.2% (n=81) and NPSLE in 15.6% (n=35). Among patients with NPSLE, manifestations included psychosis (51.4%, n=18), cognitive dysfunction (25.7%, n=9), headaches (22.9%, n=8), seizures (22.9%, n=8), cerebrovascular disease (14.3%, n=5), acute confusional state (8.6%, n=3), movement disorder (5.7%, n=2), aseptic meningitis (2.9%, n=1) and other manifestations (20.0%, n=7) such as depersonalisation, cerebral oedema, encephalitis and aphasia.

ACEs prevalence and categories

Of the 224 patients, 41.1% (n=92) had at least one documented ACE and 58.9% (n=132) had no documented ACEs. Patient characteristics by binary ACEs status are shown in table 1. There were 169 ACEs events identified among the 92 patients, with the number of ACEs per patient as follows: 60.9% (n=56) with one ACE, 16.3% (n=15) with two ACEs, 9.8% (n=9) with three ACEs and 13.0% (n=12) with four or more ACEs. Among the 92 patients with ACEs, the most commonly reported types (figure 1) were having separated/divorced parents (50.0%, n=46 patients), being bullied (23.9%, n=22), having a parent with mental illness (17.4%, n=16), environmental/extrapersonal trauma (13.0%, n=12) and domestic violence (12.0%, n=11). Other traumatic experiences stemming from violence included physical abuse (10.9%, n=10), emotional abuse (8.7%, n=8), sexual abuse (8.7%, n=8), experiencing a violent act (4.3%, n=4) and witnessing violence (2.2%, n=2).

Figure 1. The frequency of 18 ACEs collected from a cSLE cohort of 224 patients. 169 ACEs were documented among n=92 patients with the most frequently reported ACEs being separated/divorced parents (n=46 patients), bullying (n=22) and parental mental illness (n=16). ACEs, adverse childhood experiences; cSLE, childhood-onset SLE.

Of the 92 patients with ACEs, 54.3% (n=50) had ACEs that occurred only before disease diagnosis, 15.2% (n=14) had ACEs that occurred only after disease diagnosis and 30.4% (28) had ACEs of non-specific timing in relation to cSLE diagnosis (including n=19 due to indeterminate timing, and n=9 with ACEs occurring both before and after cSLE diagnosis). Patient characteristics by timed ACEs exposure category are shown in table 2. For analyses modelling ACEs as a categorical exposure for timing in relation to cSLE diagnosis (Model 2), 196 patients included in the analysis, excluding the 28 patients with non-specific ACEs timing.

Table 2. Characteristics of adolescents with ACEs categorised by time relationship to SLE diagnosis.

Baseline demographic characteristics	ACEs=92
	ACEs pre-SLE diagnosis (n=50)	ACEs post-SLE diagnosis (n=14)	ACEs non-specific timing (n=28)
Age in years, median (IQR)	15.93 (13.93 to 17.01)	15.09 (13.55 to 16.30)	15.87 (13.49 to 17.53)
Female, n (%)	43 (86)	12 (85.7)	22 (78.6)
Ethnicity, n (%)
East Asian	8 (16.0)	1 (7.14)	5 (17.9)
South Asian	10 (20.0)	3 (21.4)	3 (10.7)
European	12 (24)	3 (21.4)	6 (21.4)
African/Caribbean	4 (8.0)	2 (14.3)	4 (14.3)
Multiethnic	8 (16.0)	1 (7.1)	4 (14.3)
Latin American	5 (10.0)	2 (14.3)	4 (14.3)
Arabic	2 (4.0)	2 (14.3)	0
Indigenous	1 (2.0)	0 (0)	1 (3.6)
Other	0 (0)	0 (0)	1 (3.6)
Material resources score (range −3 to +3), median (IQR)*	0.62 (−0.19 to 1.37)	−0.26 (−0.73 to −0.07)	0.15 (−0.28 to 0.73)
Disease characteristics
Age at disease diagnosis in years, median (IQR)	14.09 (12.22 to 15.95)	11.54 (9.26 to 13.01)	13.96 (10.57 to 15.15)
Disease duration in years, median (IQR)	0.59 (0.00 to 2.06)	2.52 (1.00 to 6.28)	1.92 (0.52 to 3.91)
AMS, median (IQR)*	3.19 (0.79 to 5.50)	1.86 (0.31 to 5.93)	1.73 (0.68 to 3.23)
SLICC/ACR Damage Index (score >0), n (%)*	11 (22.0)	2 (15.4)	6 (21.4)
Arthritis	28 (56.0)	8 (57.1)	16 (57.1)
Nephritis	19 (38.0)	5 (35.7)	8 (28.6)
NPSLE	10 (20.0)	3 (21.4)	9 (32.1)
Glucocorticoid use (ever), n (%)	35 (70.0)	9 (64.3)	20 (71.4)
Any comorbid psychiatric diagnosis†, n (%)	12 (24.0)	5 (35.7)	7 (25.0)
Depressive disorder	7 (14.0)	3 (21.4)	6 (21.4)
Anxiety disorder	9 (18.0)	4 (28.6)	4 (14.3)
Other‡	0 (0)	3 (21.4)	2 (7.1)

^*Data were missing for the following variables: material resources score (n=1), AMS (n=1), SLICC Damage Index (n=1).

^†Psychiatric comorbidities that are not attributable to NPSLE.

^‡Other comorbid psychiatric diagnoses included: neurocognitive disorder (n=1), substance use disorder (n=1), eating/feeding disorder (n=3).

ACE, adverse childhood experience; AMS, adjusted mean SLEDAI; NPSLE, neuropsychiatric SLE; SLEDAI, SLE Disease Activity Index; SLICC/ACR, Systemic Lupus International Collaborating Clinics/American College of Rheumatology.

The association between ACEs and cSLE outcomes

Table 3 shows results from multivariable regression analyses examining the association between two models of ACEs exposure (binary and timed in relation to cSLE diagnosis) and the disease outcomes. Model 1 showed no statistically significant associations between the presence of ACEs and disease outcomes; however, the association approached significance for disease damage (adjusted OR=2.19, 95% CI 1.00 to 4.87, p=0.050). Model 2 showed statistically significant associations between ACEs occurring before diagnosis and the presence of disease damage (adjusted OR=3.10, 95% CI 1.21 to 8.04, p=0.018) and between ACEs occurring after diagnosis and comorbid psychiatric diagnosis (adjusted OR=5.01, 95% CI 1.20 to 21.12, p=0.025). No significant associations were found between ACEs exposure and disease activity in either Model 1 or Model 2.

Table 3. Multivariable linear and logistic regression of the association between ACEs and clinical outcomes.

Model 1: Binary ACEs exposure (n=224)
	SLE disease activity*		SLE disease damage*		Comorbid psychiatric diagnosis†
Exposure category	Regression coefficient (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
No ACEs	Reference	-	Reference	-	Reference	-
ACEs	−0.18 (−1.07 to 0.72)	0.699	2.19 (1.00 to 4.87)	0.050	1.33 (0.58 to 2.99)	0.492

Model 2: Categorical ACEs exposure by time relationship to SLE diagnosis (n=196)
	SLE disease activity*		SLE disease damage*		Comorbid psychiatric diagnosis†
Exposure category	Regression coefficient (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
No ACEs	Reference	-	Reference	–	Reference	-
ACEs before cSLE	0.25 (−0.87 to 1.36)	0.663	3.10 (1.21 to 8.04)	0.018	0.86 (0.28 to 2.41)	0.790
ACEs after cSLE	0.19 (−1.66 to 2.04)	0.839	0.89 (0.12 to 4.31)	0.892	5.01 (1.20 to 21.12)	0.025

Shown are multivariable regression models examining the association between clinical outcomes and ACEs exposure modelled as a binary variable (Model 1) and categorical variable based on time relationship to SLE diagnosis (Model 2). Covariates for both Model 1 and Model 2 included: age at disease diagnosis, sex, ethnicity (categorised as European, African/Caribbean, East Asian, South Asian, Multiethnic and Other (includes Indigenous, Latin American, Arabic and Other)) and disease duration. Bold values indicate statistically significant findings.

^*The number of patients included in the disease activity and damage analyses was n=222 for Model 1 and n=194 for Model 2, due to missing disease and damage scores for n=2 patients.

^†The number of patients included in the comorbid psychiatric diagnosis analysis was n=189 for Model 1 and n=170 for Model 2, due to exclusion of n=35 and n=26 with NPSLE diagnosis, respectively. Both Model 1 and Model 2 analyses for comorbid psychiatric diagnosis included an additional covariate for glucocorticoid use (ever).

ACEs, adverse childhood experiences; cSLE, childhood-onset SLE; NPSLE, neuropsychiatric SLE.

The association between ACEs and disease manifestations

Table 4 shows results from multivariable regression analyses examining the association between two models of ACEs exposure (binary and timed in relation to cSLE diagnosis) and disease manifestations of NPSLE, nephritis and arthritis. For both Models 1 and 2, the presence of NPSLE showed statistically significant associations between the presence of ACEs (adjusted OR=3.40, 95% CI 1.55 to 7.78, p=0.003) and ACEs occurring before diagnosis (adjusted OR=2.67, 95% CI 1.02 to 6.98, p=0.043). No significant associations were found between ACEs exposure and nephritis or arthritis in either Model 1 or Model 2.

Table 4. Multivariable logistic regression of the association between ACEs and disease manifestations.

Model 1: Binary ACEs exposure (n=224)
	Nephritis		NPSLE		Arthritis
Exposure category	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
No ACEs	Reference	-	Reference	-	Reference	-
ACEs	1.02 (0.55 to 1.86)	0.958	3.40 (1.55 to 7.78)	0.003	0.89 (0.49 to 1.60)	0.688

Model 2: Categorical ACEs exposure by time relationship to SLE diagnosis (n=196)
	Nephritis		NPSLE		Arthritis
Exposure category	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
No ACEs	Reference	-	Reference		Reference	-
ACEs before cSLE	1.16 (0.56 to 2.42)	0.686	2.67 (1.02 to 6.98)	0.043	0.89 (0.43 to 1.82)	0.738
ACEs after cSLE	1.32 (0.36 to 4.44)	0.656	2.66 (0.52 to 10.85)	0.192	0.98 (0.30 to 3.35)	0.968

Shown are multivariable regression models examining the association between disease manifestations and ACEs exposure modelled as a binary variable (Model 1) and categorical variable based on time relationship to SLE diagnosis (Model 2). Covariates for Model 1 and Model 2 included: age at disease diagnosis, sex, ethnicity (categorised as European, African/Caribbean, East Asian, South Asian, Multiethnic and Other (includes Indigenous, Latin American, Arabic and Other)), disease duration, presence of nephritis (except for nephritis model), NPSLE (except for NPSLE model) and arthritis (except for arthritis model). Bold values indicate statistically significant findings.

ACEs, adverse childhood experiences; cSLE, childhood-onset SLE; NPSLE, neuropsychiatric SLE.

Discussion

This retrospective cohort study of children with cSLE characterised ACEs, we examined the association between ACEs exposure, disease outcomes and manifestations. We found that ACEs exposure was prevalent in 41% of the cohort and associated with NPSLE diagnosis. Specifically with regard to timing, children with documented ACEs pre-cSLE diagnosis were more likely to experience disease damage and NPSLE, and those with documented ACEs post-cSLE diagnosis were more likely to receive a comorbid psychiatric diagnosis. No associations were found between ACEs exposure and disease activity, arthritis and nephritis. Our study contributes to the limited literature on ACEs in cSLE and sheds light on their relationship to cSLE disease features.

We found that 41% of patients experienced at least one ACE, compared with previous studies that indicated 40% and 76% ACE exposure in other paediatric settings,^{33 34} with variation likely due to methodological differences in data collection. Notably, our reported prevalence aligns with our previous work in a research participant subset of our cohort with prospective psychiatric interviews focused on traumatic events revealing a prevalence of ACEs at 40%.³⁵ Recognising the health disparities among vulnerable patients with cSLE, we used the expanded ACEs framework to capture community-level adversity like bullying and discrimination,²⁶ alongside conventional household-level ACEs. Our findings highlight prevalent household ACEs, including separated/divorced parents, parental mental illness, domestic violence and death of a parent. Bullying, a community-level ACE, was the second most frequent event. Interestingly, previous studies focusing on children with ACEs have similarly identified divorce/separation, parental mental illness, various forms of abuse, poverty and bullying as the most reported ACEs, which closely corresponds to our findings.^{33 34}

Based on the literature indicating ACEs exposure to be a significant risk factor for developing adult-onset SLE in women,¹⁹ we examined the timing of ACEs relative to cSLE diagnosis. Expectedly, we found that most ACEs (78%) occurred before cSLE diagnosis. Given that the median age of disease diagnosis was 13 years old, and most patients with a history of early adversity experienced ACEs before diagnosis, it is plausible that the trauma occurred during their early childhood or adolescent years. Literature suggests that the adverse experiences faced during sensitive developmental periods in early life become biologically ingrained and manifest during adolescence.³⁶ While the precise mechanisms by which ACEs affect mental and physical health outcomes are not yet fully understood, there is a consensus that chronic, systemic inflammation plays a significant role in this process.¹³ Consequently, youth with a history of ACEs are at risk of experiencing prolonged, low-level inflammation and compromised immune systems.

Recent studies indicate a potential link between ACEs and increased SLE disease activity and flares in adults.^{20 21} Higher ACE scores were associated with increased self-reported disease activity and flares but not necessarily with physician-assessed disease activity in adult patients with SLE.²¹ Our findings did not support an association between ACE exposure (either binary or timed in relation to diagnosis) and physician-assessed disease activity (SLEDAI-2K) in cSLE; however, we did not examine self-reported disease activity. Additionally, many of the patients in our cohort had only one documented ACE and were still at an age where they could experience further adversity. As these patients continue to grow, any additional trauma they encounter during their formative years may exert a negative influence on their disease activity. Further longitudinal studies in young patients, considering both self-report and physician-report measures, are needed to understand the relationship between ACEs and SLE disease activity.

In contrast to our findings for disease activity, we found that children with ACEs prior to cSLE diagnosis exhibited a statistically significant higher odds of disease damage, and this finding neared statistical significance for the binary ACEs exposure. Patients with ACEs may experience potential delays in care or poor access to care due to socioenvironmental factors, which may lead to higher disease activity surrounding diagnosis and increased risk for disease damage. Additionally, research has established that ACEs elicit a maladaptive stress response, often contributing to chronic low-grade inflammation,³⁷ which may not be captured on the SLEDAI-2K, but could lead to organ damage. Further understanding of the interplay between ACEs, healthcare access and disease outcomes could provide valuable insights into the mechanisms driving disease progression in this vulnerable population.

In our secondary analysis examining the relationships between ACEs and disease manifestations, we found that patients who experienced ACEs (for both binary and timed exposure models) had an approximately threefold higher likelihood of having NPSLE. In adolescents, NPSLE has been associated with a decreased quality of life,³⁸ SLE disease damage accumulation³⁹ and increased morbidity and mortality.⁴⁰ Furthermore, as the aetiology of NPSLE remains unclear, it can be challenging to diagnose and treat.^{32 41} Our identification of ACEs as a potential risk factor for NPSLE is quite intriguing. Existing literature suggests that early traumatic experiences can significantly impact the neurological framework, potentially rendering patients vulnerable to psychosis.^{42 43} Of those with NPSLE in our cohort, psychosis was the most common syndrome in 51%; however, the proportion with ACEs exposure was similar to that in patients with NPSLE without psychosis. Further study examining ACEs and specific NPSLE syndromes in larger cohorts is needed. It is also worth noting that patients with NPSLE are seen more often by both social work and psychiatry, potentially providing greater opportunity to ascertain the presence of ACEs for the patients with NPSLE. Further investigation with prospective data collection is needed to confirm our findings and examine possible underlying mechanisms for the link between ACEs and NPSLE.

We did not find any associations between ACEs and lupus nephritis or arthritis. These findings contradict the existing literature, where ACEs were shown to increase the odds of arthritis in both children and adults, and nephritis in adults.^{20 44} These findings were based on a dose–response relationship, suggesting that individuals with more ACE exposures were more susceptible to developing arthritis and nephritis. However, most of our cohort did not have high counts of ACEs, which are implicated in negative health consequences. Furthermore, specific ACEs, such as physical and sexual abuse, which were not highly prevalent in the current cohort, have been linked with arthritis and nephritis in research studies.^{44 45} The nuanced nature of the impact of dose, type and timing of ACEs exposure on disease manifestations warrants further extensive research.

We also found that ACEs experienced after cSLE diagnosis conveyed a fivefold increase in the likelihood of having a co-occurring psychiatric disorder (not attributable to SLE). The prevalence of psychiatric comorbidities, particularly mood and anxiety disorders, is high among patients with SLE⁹ and associated with poor outcomes such as increased self-reported disease activity⁴⁶ and worse health-related quality of life.⁴⁷ Our results are consistent with extant literature that has established a link between ACEs and increased risk of mental health diagnoses like depression and anxiety among children and adolescents.^{17 18} However, additional considerations are raised in the cSLE population. The added strain of experiencing traumatic events after the onset of a chronic, autoimmune disease like cSLE could potentially exacerbate or induce psychopathology in those who are susceptible.⁴⁸ As the number of patients with ACEs after cSLE diagnosis was small, further study in larger, longitudinal cohorts is needed to examine the link between ACEs and comorbid psychopathology in cSLE.

Our findings emphasise an opportunity for timely access to appropriate mental health support to mitigate the effects of ACEs on a patient’s mental and physical well-being. While trauma-informed and resilience-building interventions exist for children and adults,⁴⁹ this has been understudied in the SLE population, and there might be particular benefit for those with cSLE who are more proximal to ACEs events and SLE diagnosis compared with their adult counterparts. Clinical screening for ACEs for all patients with SLE may be warranted to understand and mitigate their risk for poor clinical and psychosocial outcomes. As mental health resources are currently lacking for children and adults with SLE,³⁵ our findings also serve as a call to action for advocacy for mental health support as a critical component of comprehensive SLE care. Moreover, parental educational and support initiatives focused on reducing ACEs exposure could decrease the prevalence of concurrent mental health issues. Further research is needed to develop early targeted interventions, address ACEs and optimise health outcomes of individuals with cSLE.

Our study has limitations, being a single centre study with potential for limited generalisability. Our approach for retrospective ACEs data collection using the EHR is prone to ascertainment bias and may have led to the underidentification of ACEs due to general patient/caregiver under-reporting and/or lack of clinician questioning or documentation. Delays between symptom onset and cSLE diagnosis may have also affected the classification of ACEs timing, such that the number of patients with ACEs after disease onset may be underestimated (ie, ACEs occurred after symptom onset but before diagnosis); we were not able to determine the symptom onset reliably from our chart review. Additionally, patients with a more recent cSLE diagnosis had shorter follow-up duration, which inherently restricted the window of observation and group size for ACEs occurring after diagnosis. As a result, the range of stressors experienced over time may not be fully captured for these patients. Moreover, our analysis was only able to detect associations and not causal relationships. We also did not account for SES in our analyses as we did not have household-level income or education data available for the cohort; however, several studies indicate a strong link between ACEs and SES, which may act as a confounder and more likely mediator in the relationship to health outcomes.50,52 While we included descriptive data on neighbourhood-level marginalisation, we did not incorporate this variable in our analysis due to our intentional inclusion of expanded ACEs which are highly inter-related with household and neighbourhood SES characteristics.²⁶ Further study of the intersection of the ACEs, SES and health outcomes is warranted in cSLE using approaches such as mediation models. Despite these limitations, the strengths of our study include our structured, comprehensive and detailed approach to ACEs data collection, including frequency, types (both household-level and community-level) and timing related to cSLE diagnosis. Additionally, we used rigorously documented clinical disease data from a large and ethnically diverse cSLE cohort.

In conclusion, current literature lacks sufficient investigation into the impact of ACEs in cSLE. This study aimed to address this knowledge gap by examining the effects of ACEs on different aspects of the disease. The results revealed that ACEs were associated with NPSLE, and the timing of ACEs was differentially associated with risk of cSLE disease damage and psychiatric comorbidities. These findings highlight the complex relationship between ACEs and their implications for clinical and psychosocial outcomes. Further investigation is needed to substantiate these findings and elucidate the underlying mechanisms driving the effect of ACEs on cSLE health outcomes. Furthermore, the findings underscore the need for a comprehensive healthcare approach that addresses social determinants of health like ACEs, through targeted strategies for trauma-informed care.

Data availability statement

Data are available upon reasonable request.