Adverse Childhood Experiences and Primary Headache Disorders

Abstract

Background and Objectives

Headache disorders are among the leading causes of disability worldwide. While an association between adverse childhood experiences and primary headaches has been reported, the pooled magnitude across studies and pathways of the association are unknown. Our objectives were (1) to estimate the pooled effect of ≥1 adverse childhood experience (ACE) on primary headache disorders in adulthood and (2) to test the hypothesis that ACEs categorized as “threat” traumas or “deprivation” traumas have distinct effects on primary headaches based on a selected theory from our narrative review of how ACEs affect human development along the life course.

Methods

PubMed, EMBASE, MEDLINE, Web of Science, Google Scholar, Biological Psychiatry, and gray literature were searched up to March 16, 2023 (PROSPERO, CRD42020223403). Selected articles included (1) observational studies with a comparator group, (2) ACEs that occurred before 18 years of age, and (3) primary headaches occurring at or after 21 years of age. Pooled odds ratios (ORs) were calculated using multilevel linear random-effects modeling. The narrative review included theories that describe how ACEs affect human development and disease across the life course. We selected a theory from our narrative review and tested ACEs categorized according to this theory for any modification of point estimates.

Results

Our search identified 32 studies, of which 28 were eligible for meta-analysis (n = 154,739 participants, 19 countries). The occurrence of ≥1 adverse childhood experience(s) was associated with primary headaches (pooled OR = 1.48 [95% CI 1.36–1.61]; high-quality evidence, 134,696 participants). As the number of ACEs increased, the odds of primary headaches increased (range: 1 ACE OR = 1.24 [95% CI 1.14–1.35] to ≥4 ACEs OR = 2.09 [95% CI 1.83–2.38], p for trend <0.0001). From the narrative review, a neurodevelopmental theory that categorizes ACEs into threat or deprivation was tested, and both were independently associated with primary headaches (threat OR = 1.46 [95% CI 1.32–1.60] and deprivation OR = 1.35 [95% CI 1.23–1.49], respectively), accounting for heterogeneity (p = 0.021).

Discussion

This systematic review and meta-analysis confirm that ACEs are important risk factors of primary headache disorders in adulthood. Our findings provide epidemiologic support that ACEs categorized as threat and deprivation may manifest as distinct pathways of early adversity.

Introduction

Primary headache disorders are among the leading causes of all-age disability that impact quality of life, including family and social life, and result in productivity loss.^1-3 The World Health Organization recognizes that the burden of headache is underestimated, under-recognized, and undertreated,³ and this in part can be explained by gaps in knowledge on the causal pathways of primary headaches.

Similar to other chronic health conditions, the origins of headache disorders likely begin earlier in life than the manifestation of symptoms. Life course epidemiology is a framework that considers how the harmful effects of adverse childhood experiences (ACEs) or developmental trauma disorders (DTDs) contribute to the development of chronic conditions in adulthood.⁴ ACEs or DTDs are composite terms that capture an array of stressful exposures, such as sexual abuse, physical abuse, and neglect, and are consistently linked to chronic conditions, disability, and early death.^4-10

A growing body of evidence postulates that ACEs are an upstream risk factor of primary headaches, although the mechanism by which ACEs lead to headaches later in life is unclear. Some theoretical frameworks, such as critical period theory and cumulative risk theory, propose mechanisms of how ACEs alter neurodevelopment.^11-24 However, a limitation of these theories is that they methodologically assume each ACE, whether occurring directly to the child, or through parent/guardian proxy, elicits a similar degree of stress-response dysregulation. For instance, direct sexual abuse in childhood is methodologically considered equivalent to having a parent with a substance use disorder, neglecting any probability that they may have different neurodevelopmental processing and manifestations.^19,23,24

An emerging neurodevelopmental adversity theory suggests broadly conceptualizing ACEs into (1) “threat” traumas, a direct threat to the child, or (2) “deprivation” traumas, the child's needs are largely neglected.¹⁹ Correspondingly, threat ACEs involve the hippocampus, amygdala, and ventromedial prefrontal cortex, activating the hypothalamic-pituitary-adrenal (HPA) axis.¹⁹ Deprivation ACEs are characterized by an absence of typical sensory and cognitive environmental complexity, manifesting as structural changes in the brain to deal with low-complexity environments (e.g., reduction in thickness and volume of the association cortex from overpruning).¹⁹ While most studies seek to determine the association of an exposure (i.e., ACEs) to an outcome (i.e., headache), no studies have tested a theoretical biological framework to depict differences between ACEs. This neurodevelopmental adversity theory has not yet been applied in large epidemiologic studies to test the association of ACEs with chronic diseases, such as headaches. The specific objectives of this study are

1. To examine (a) the pooled association of the presence of at least one of any ACE in childhood and (b) the burden of different ACEs in childhood and their association with primary headaches in adulthood.

2. To test the hypothesis that ACEs broadly categorized as threat or deprivation may have different effect sizes on primary headaches, based on a neurodevelopmental adversity theory.¹⁹

Methods

Systematic Review and Meta-analysis of ACEs and Primary Headache Disorders

Standard Protocol Approvals, Registrations, and Patient Consents

The protocol was registered on PROSPERO (ID CRD42020223403) and presented according to MOOSE and PRISMA reporting guidelines.

Search Strategy and Selection Criteria

This systematic review and meta-analysis included all studies that assessed the relationship between at least one of any ACE that occurred before 18 years of age and primary headache cases in an adult population 21 years and older. We selected an adult population 21 years and older to increase the probability in the direction of the relationship and allow a latency period. Only observational studies investigating primary headache disorders, as determined by the authors of the included studies, with an appropriate comparative group were included. Primary headache disorders included migraine, tension-type headache, cluster headache, and chronic/severe headache. Headaches of stroke origin, injuries, wounds, brain disorders, infections, or medication overuse were excluded. Studies were compiled from PubMed, EMBASE, MEDLINE, Web of Science, Google Scholar, Biological Psychiatry, and gray literature. Searches were conducted up to March 16, 2023. For additional search details, refer to eTable 1 (links.lww.com/WNL/D157).

Data Collection

At least 2 authors (C.S., A.C.M., K.M., D.B.) reviewed all titles, abstracts, and full-text articles in stepwise stages, with kappa values 0.78–0.98. Data extraction of full-text articles was conducted by 2 authors (C.S., A.C.M., K.M., D.B.) using a standardized form (eTable 2, links.lww.com/WNL/D157). Discrepancies were resolved through group discussions involving all study authors. From each study, we extracted outcome point estimates including odds ratios, relative risks, hazard ratios, and mean scores, with corresponding 95% confidence intervals, the number of events in each group, and covariates included in the model(s). All ACEs were subcategorized according to threat or deprivation (Figure 1).

Figure 1. Original Adverse Childhood Experience (ACE) Terms Categorized as Threat or Deprivation for This Meta-analysis.

Risk of Bias

At least two authors (C.S., A.C.M., K.M., D.B.) independently assessed the risk of bias using the Quality in Prognostic Studies (QUIPS) tool,²⁵ with a group discussion to resolve discrepancies. We applied a scoring system to assess the risk of bias in our models (eTable 3, links.lww.com/WNL/D157).

Data Analysis

For the primary analysis, we calculated the odds and hazard ratios of headache among those with at least one of any ACE compared with those with no ACEs. Most of the eligible studies contained more than one effect size (e.g., one effect size for physical abuse and one effect size for sexual abuse), resulting in intracorrelations among effect sizes. A three-level mixed linear model was fitted with random effects for the variance in point estimates between and within studies to account for the complex variance structure of the data. To determine the heterogeneity in effect sizes, moderator analyses were conducted based on a priori selected variables. A secondary multilevel linear analysis was constructed using a subset of the studies that captured the burden of ACEs (e.g., 1 ACE, 2 ACEs, 3 ACEs, and ≥4 ACEs) as the fixed effect, and random effects accounted for the variance in effect sizes between and within each study. Sensitivity analyses estimated the effect size excluding studies with specific patient populations or imputations and among studies with a prospective study design. Publication bias was assessed using Egger's regression test for funnel plot asymmetry, adjusted for variance to account for the multilevel design in this meta-analysis.²⁶ To assess the quality of the evidence, we conducted a GRADE assessment (C.S., A.C.M.) with modifications for prognostic systematic reviews.²⁷ We implemented open-source software, resources, and tools. The statistical software package metafor in R version 4.0.2 was used.^28,29

Secondary Objectives

The secondary objective of this study applied results from our narrative review of theories that describe the pathophysiologic effects of ACEs in the development of chronic disease that apply a life course perspective to explain how disease manifests (eTable 4, links.lww.com/WNL/D157). Theories reviewed must have been based on human development, and theories based solely on animal models were excluded. Findings from the selected theory¹⁹ propose to categorize ACEs broadly as “threat” or “deprivation” (Figure 1), and we tested this theory for modification of point estimates by adding this categorization to the model and examining the change in heterogeneity.

Data Availability

Data will be available on reasonable request.

Results

The literature search returned 39,658 articles, of which 97% were indexed as journal articles (n = 38,487), 1% as web pages (n = 568), 1% as books (n = 424), and 0.5% as white papers (n = 179) (Figure 2). Overall, 32 studies were included in the systematic review (168,394 participants across 20 countries), of which 28 studies (26 unique studies) of 154,739 participants from 19 countries were eligible for meta-analysis (Table; eTables 5–7, links.lww.com/WNL/D157). Of the 154,739 participants in the meta-analysis, an estimated total of 48,625 participants (31%) reported at least one of any ACE, and a total of 24,956 (16%) participants were diagnosed with primary headaches. Among study participants with at least one of any ACE, 26% were diagnosed with primary headache (12,565/48,625), compared to 12% of participants that had no ACEs (12,391/106,114). Most studies were cross-sectional (81%), sampled from the general population (69%), and conducted in the United States (42%). The most commonly studied ACEs were physical abuse (77%), sexual abuse (73%), and exposure to family violence (38%) (Table). Eleven theories were identified in our narrative review of how ACEs affect neurodevelopment and can lead to disease later in life (eTable 4).

Figure 2. PRISMA Flowchart of Study Selection.

Table.

Characteristics of Studies Included in the Meta-analysis With Adverse Childhood Experiences Categorized as Threat and Deprivation

	Categorization of studies
Characteristica	Any adverse childhood experience(s) N = 26	Threatb,c adverse childhood experience(s) N = 19	Deprivationb,c adverse childhood experience(s) N = 10
Total sample size	154,739	127,590	91,813
>75% of sample female, n (%)	10 (38)	7 (37)	1 (10)
Study design, n (%)
Cross-sectional	21 (81)	16 (84)	7 (70)
Prospective cohort	3 (12)	1 (5)	2 (20)
Case-control	2 (8)	2 (11)	1 (10)
Population sampled, n (%)
Clinical	7 (27)	6 (32)	2 (20)
Community/general	18 (69)	13 (68)	8 (80)
Community and clinical sample	1 (4)	0 (0)	0 (0)
Regiond, n (%)
United States	11 (42)	8 (44)	3 (30)
Canada	2 (8)	2 (11)	2 (20)
Australia and New Zealand	3 (12)	2 (11)	0 (0)
Europee	5 (19)	3 (16)	3 (30)
Latin America (Mexico and Peru)	2 (8)	2 (11)	0 (0)
Asia (Philippines and Japan)	2 (8)	1 (5)	1 (10)
Multicountryf	1 (4)	1 (5)	1 (10)
Adverse childhood experiences, n (%)
Physical abuse	20 (77)	15 (79)	0 (0)
Sexual abuse	19 (73)	14 (74)	0 (0)
Exposure to intimate partner violence or family violence	10 (38)	7 (37)	0 (0)
Emotional abuse	8 (31)	3 (16)	0 (0)
Other household instabilitiesg	13 (50)	0 (0)	10 (100)
Assessment of primary headache disorders, n (%)
Self-report	13 (50)	9 (47)	6 (60)
Diagnostic criteriah	13 (50)	10 (53)	4 (40)
Type of primary headache disorderd, n (%)
Migraine (exclusively)	11 (42)	9 (47)	4 (40)
Headaches (nonspecific)	3 (12)	2 (11)	2 (20)
Severe/chronic headaches	12 (46)	8 (42)	4 (40)
Assessment of confounding in analysis, n (%)
Adjusted for age and sex	19 (73)	14 (74)	7 (70)
Adjusted for additional demographic and health-related characteristics (e.g., obesity, smoking)	11 (42)	8 (42)	3 (30)
Unadjusted analysis	7 (27)	5 (26)	3 (30)

^aStudies reporting odds ratios (n = 24 studies) and hazard ratios (n = 2) were eligible for meta-analysis. Studies with relative risks (n = 1) and studies reporting adverse childhood experiences as a mean score are not shown (n = 3) as were not eligible for meta-analysis.

^bA single study can report adverse childhood experiences of threat and deprivation; therefore, there is an overlap between the number of threat and number of deprivation studies.

^cSee Figure 1 for threat and deprivation categorization of adverse childhood experiences.

^dPercentages may not total 100 due to rounding.

^eGreat Britain, Finland, Poland, Germany, Norway, and the Netherlands.

^fColumbia, Mexico, the United States, Belgium, France, Germany, Italy, the Netherlands, Spain, and Japan (pooled).

^gAt least one of neglect, parental death, parental separation or divorce, serious household conflicts, economic adversities, household mental or chronic illness, household substance or alcohol abuse, institutionalization or incarceration of parent, and household criminal behavior.

^hWithin our study, “diagnostic criteria” were defined as the following: (a) headache diagnosed through a patient chart review, (b) diagnosis based on the use of ICD-9/ICD-10/ICHD-2 or ICHD-3 criteria, (c) headache diagnosed by a neurologist using IHS criteria, or (d) if participants were diagnosed by a physician, nurse, or other health care professional.

Pooling the data from the 24 unique studies reporting ORs resulted in a total study population of 134,696 participants (Figure 3). At least one of any ACE was associated with primary headaches (OR = 1.48, 95% CI 1.36–1.61), compared with no ACEs (Figure 3). As the number of ACEs increased, the strength of the association with primary headaches also increased in a dose-response relationship (p-trend <0.0001; Figure 4). Both threat and deprivation ACEs were associated with primary headaches, and heterogeneity was explained by categorization into threat or deprivation (F = 4.03, p = 0.021). The effect sizes of threat ACEs [OR=1.46 (95% CI 1.32–1.60), 17 studies] and deprivation ACEs [OR=1.35 (95% CI 1.23–1.49), 8 studies] were consistent with the direction and magnitude of the pooled main effect of ACEs associated with primary headaches (Figure 3), supporting the internal validity of our findings.

Figure 3. Forest Plot of Multilevel Meta-analysis: Pooled Associations of at Least 1 of Any Adverse Childhood Experience (1) and of Categorized Adverse Childhood Experiences (2–4) With Primary Headache Disorders.

ACE = adverse childhood experience; CI = confidence interval; k = number of effect sizes; n_p = number of participants; n_s = number of studies. ^aStudies where both threat and deprivation ACEs are aggregated, and it is impossible to distinguish between the 2 categorizations.

Figure 4. Dose-Response Relationship of Pooled Multilevel Effect Size Between the Number of Adverse Childhood Experiences and Primary Headache Disorders.

ACE = adverse childhood experience. The reference group in all groups is 0 ACE.

The overall effect size was not significantly different between studies that varied in headache-specific disorder, study design, population type, method of headache diagnosis, year of publication, adjustment for sociodemographic and health behaviors, or risk of bias (all not significant). After adjusting for ACEs categorized as threat or deprivation, the overall variance in effect size of primary headaches decreased from 0.072 to 0.062. Publication bias was not detected (z = 1.93, p = 0.05) (eFigure 1, links.lww.com/WNL/D157). Age and sex were included as confounders among 71% (n = 17) of studies that composed the pooled effect size. Among all studies, only 38% (n = 9) adjusted for additional demographic or health-related risk factors. Despite the inconsistent reporting of confounders among the included studies, we assessed the overall risk of bias as low (eTable 3; eFigure 2). Of 32 studies, 6 reported association measures other than ORs (eTable 7). The quality of evidence was high (eTable 8). Sensitivity analysis excluding population-specific studies did not meaningfully change the final point estimate (overall <6% change). Among the studies with a prospective study design (n = 3 studies, 10,417 participants), at least one of any ACE was associated with primary headaches (OR = 1.47, 95% CI 1.18–1.83), compared with no ACEs.

Discussion

While headache disorders are among the leading causes of disability worldwide, upstream risk factors of primary headaches remain largely unknown. The main pooled estimate from this systematic review and meta-analysis, collating the ACE histories of 134,969 individuals from 19 countries, indicates that the presence of at least one of any ACE is associated with primary headache disorders in adulthood. Among studies that report the number of ACEs, the risk of primary headaches among adults increases in a positive dose-response relationship, supporting the strength of the association. We purport that our results are a conservative estimate as ACEs are often underreported due to their sensitive nature. Despite this, the robustness of these findings cannot be underappreciated as the studies composing this meta-analysis represent diverse global regions and the findings supersede cultural contexts. Furthermore, individual study estimates align with the overall pooled estimate, and this consistency strengthens our findings, suggesting that there may be common underlying mechanisms.

Studying ACEs inherently recognizes that conditions in childhood/adolescence can have serious health implications later in life. We explored human developmental theories that incorporate a life course perspective and tested a neurodevelopmental theory that broadly categorized ACEs into experiences of threat (e.g., physical abuse) or deprivation (e.g., neglect) (Figure 1). This theory proposes that it is overly simplistic to assume that all ACEs affect neurodevelopment through the same mechanism.¹⁹ The pathway of ACEs to chronic disease is undoubtedly complex and likely involves multiple pathways over the life course. Our study findings indicate that ACEs categorized as threat or deprivation are independently associated with primary headache disorders in adulthood. This novel application of a biological theory on an epidemiologic population supports that there is likely more than one neurodevelopmental pathway of ACE to primary headache in adulthood (Figure 5). These findings have implications within the life course to develop multipronged primary and secondary prevention and treatment strategies tailored to the underlying ACE.

Figure 5. Conceptualization of Adverse Childhood Experiences Categorized as Threat and Deprivation and Their Influence on Health Outcomes Across the Life Course.

This conceptualization builds on the work of Felitti et al.⁵ by recognizing the potentially distinct pathways of adverse childhood experiences categorized as threat and deprivation. Felitti is historically considered the grandfather of ACEs.

Studies on primary headache disorders often focus on identifying and describing precipitating triggers (e.g., menstruation, visual stimuli, weather changes, nitrates, wine, and fasting). However, upstream risk factors that increase the likelihood of developing primary headaches are not well-known. Patients often report “stress” as the most common trigger^30,31; however, “stress” is an umbrella term that can include and blur between stress exposures and stress responses across the life course.³² In this study, we propose that stress exposures, such as ACEs that occur during a critical period in human development, may alter downstream stress-response mechanisms (Figure 5). We, along with many others, produce a dose-response relationship, confirming that as the burden of ACEs increases, the probability of adverse health conditions also increases.^7,8 This relationship is so robust that it is consistent across a multitude of health conditions over various social and cultural contexts from different regions around the world.^7,8,33 This persistent association also suggests that there may be a common underlying developmental biological mechanism. It is hypothesized that the main mechanism in which ACE exposure affects health conditions later in life is primarily attributed to sustained dysregulation of the HPA axis, resulting in allostatic overload.^14,22,23 However, more recently, this mechanism has been criticized as oversimplistic and unable to account for the breadth and complexity of the pathways influencing structural and functional remodeling during neurodevelopment.¹⁹

Traditionally, ACEs are categorized according to their burden (e.g., 1, 2, 3, and 4+ ACEs); however, we challenged this approach by grouping ACEs into broad categories that reflect their distinct structural and functional effects during neurodevelopment. Despite a greater representation of studies focusing on threat ACEs (17 threat studies vs eight deprivation studies), we found that both threat and deprivation ACEs are independently associated with primary headache disorders. Furthermore, by distinguishing between threat and deprivation ACEs, there is a 14% absolute reduction in variance, suggesting better signal detection in the noise (all ACEs σ = 0.072 vs ACEs categorized according to threat or deprivation σ = 0.062). ACEs characterized as threat (e.g., physical/emotional/sexual abuse) often involve activating the HPA axis and compromise neural circuits involved in emotional learning and fear conditioning.¹⁹ By contrast, deprivation ACEs (e.g., neglect or household substance abuse) are linked to impaired pathways that may not include the HPA axis (e.g., over-pruning of synaptic connections).¹⁹ Conventionally, there has been a greater emphasis on the role of threat ACEs and activation of the HPA pathway. This bias is a problem because (1) it reinforces the HPA axis pathway as the central mechanism in which ACEs manifest and ignores other stress responses generated, (2) combining both ACEs of threat and deprivation into one category does not provide the opportunity to disentangle independent effects or variability within an ACE category, (3) the effect of deprivation ACEs remains underappreciated, and (4) it does not allow for new theories of how ACEs affect neurodevelopment to emerge. However, we hypothesize that these 2 pathways of ACEs are likely not the only dimensions that lead to primary headaches in adulthood, and future studies should explore additional potential pathways. Moreover, ACEs of threat and deprivation often co-occur, potentially affecting the final pooled estimates of the effect size of threat and deprivation ACEs. Unfortunately, this is difficult to determine and disentangle because most studies do not provide data on the degree of overlap between different ACEs.

Our findings reflect a conservative estimate of the true effect size of ACEs on primary headaches as ACEs are commonly underreported because of their sensitive nature.³⁴ Previous studies rarely explore the duration, severity, frequency, or age of onset of ACE exposure, and we cannot exclude the possibility that these features have implications for primary headaches. Unfortunately, because of the nature of reporting within primary studies, we cannot statistically disentangle younger and older adults in our analyses; however, future studies on headache disorders should differentiate between young adults of 21–26 years of age (with a still-developing brain), and older adults (older than 26 years of age). Future studies should also collect data on the age of ACE onset and the age of first headache to understand the impact of timing and potential recall bias. While two-thirds of included studies adjust for age and sex, at a very minimum, we hope that all future studies will include this important information because age and sex are known risk factors of headaches. We suspect that additional sociodemographic factors and health risk behaviors such as smoking and alcohol use may modify the association between ACEs and primary headaches. Despite this, among the studies that did report and adjust for risk factors, the risk factors did not explain heterogeneity in the final point estimate. Furthermore, heterogeneity was not explained by study characteristics, including assessment of ACEs with different screening tools, methods of ascertaining primary headaches, types of primary headaches, study design, and publication year, supporting the internal validity of our findings. Beyond the scope of this study, we are aware that among individuals who have ACEs, it is possible to develop primary headache disorders in childhood and adolescence^35-38 and hope that future researchers can explore this important relationship prospectively and longitudinally. The relationship between ACEs and headache disorders and other common comorbidities (e.g., anxiety and depression) should also be further explored prospectively in large population studies.^39,40 Owing to the inherent nature of studies investigating ACEs, causation cannot be inferred. However, we triangulated our findings by testing an underlying biological theory that threat and deprivation ACEs may manifest differently in neurodevelopment with distinct effects on primary headaches. Despite the challenging nature of studies examining ACEs, our findings are externally valid. Representing results from 19 countries, we demonstrate (1) a dose-response relationship between the burden of ACEs and primary headache disorders consistent with others, (2) temporality of the effect as ACE(s) occur well before the headache, and (3) plausibility of biological support for these findings.⁴¹

While a common criticism of ACEs is that they are considered “nonreversible” and “nonmodifiable”, multipronged primary and secondary prevention, public health, and clinical strategies tailored to the underlying ACE can help mitigate the effects of ACEs on adult health outcomes.⁴² It is important to prevent and treat ACEs as soon as they are identified because early intervention may help to resolve ACE(s) and prevent related associations with chronic diseases in the future. Societally, awareness should be raised, with routine screening for ACEs among children, adolescents, and young adults, to prevent ACE-related health outcomes, including headaches, which can start developing from early childhood and adolescence.^35-38 While ACEs commonly refer to experiences occurring to an individual or within their household, future studies should explore the impact of experiences occurring at the community level (e.g., witnessing crime or riots) as potential ACEs.⁴³ Other primary prevention strategies, including evidence-based public health education, policies, and programs in schools, community settings, social care, criminal justice, and health care settings, can be leveraged. Developing and teaching resilience strategies to children and adolescents may mediate some of the health effects of ACEs.^44,45 From a clinical perspective, health care providers, such as neurologists and primary care physicians who treat primary headaches in adults, should routinely screen for ACEs,⁴⁶ educate patients on the connection between ACEs and health, and provide referrals for treatment strategies tailored to the respective ACE. Treatment strategies, especially if targeted to the differential pathways of threat and deprivation, may help to rewire parts of the brain that have been dysregulated.^19,47,48 For example, trauma-informed therapy is often used in patients with threat ACEs and attachment-based therapy for patients with deprivation ACEs.^19,49 Personalized management and treatment of primary headaches can be tailored by identifying ACE subgroups.

This systematic review and meta-analysis highlights that ACEs of threat and deprivation are important and distinct risk factors of adult primary headache disorders. Identifying threat and deprivation ACEs as risk factors may have implications for tailored primary and secondary prevention and treatment strategies for one of the leading disabling disorders worldwide. A multipronged public health and clinical strategy tailored to the underlying ACE is of high priority.

Glossary

ACE

adverse childhood experience

DTD

developmental trauma disorder

HPA

hypothalamic-pituitary-adrenal

OR

odds ratio

QUIPS

Quality in Prognostic Studies tool

Appendix. Authors

Name	Location	Contribution
Claudia Sikorski, PhD	Department of Health Research Methods, Evidence, and Impact, McMaster University, Hamilton, ON, Canada	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data
Anna C. Mavromanoli, MD	University Medical Center of the Johannes Gutenberg University, Mainz, Germany	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data
Karishma Manji, MD	Department of Family and Community Medicine, University of Toronto, ON, Canada	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data
Danial Behzad, BSc	Department of Health Sciences, Brock University, St. Catherines, ON, Canada	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data
Catherine Kreatsoulas, PhD	Health Policy and Management, Harvard T.H. Chan School of Public Health, Boston, MA	Drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data

Study Funding

The authors report no targeted funding.

Disclosure

C. Sikorski reports no disclosures relevant to the manuscript; A.C. Mavromanoli reports no disclosures relevant to the manuscript; K. Manji reports no disclosures relevant to the manuscript; D. Behzad reports no disclosures relevant to the manuscript; C. Kreatsoulas reports no disclosures relevant to the manuscript. Go to Neurology.org/N for full disclosures.