Adverse Childhood Experiences and Frequent Headache by Adolescent Self-Report

Abstract

Background:

The association between exposure to adverse childhood experiences (ACEs) and increased headache in adults has been well characterized. Childhood adversity and its effect on headache in children have not been as robustly investigated. This study examines the relationship of self-reported ACEs to frequent headache in an adolescent cohort.

Methods:

We performed a retrospective cohort study using data from the National Longitudinal Study of Adolescent to Adult Health Wave I (n = 20,745) to examine self-reported ACE exposures and their relationship to frequent headache.

Results:

The study population was composed of 20,745 participants; 50.6% male and 49.4% female. The mean age of respondents was 15.9 years (range 12 to 21 years, standard error: 0.12 years). Frequent headache was reported in 29.3% of respondents, and 45% of respondents reported one or more ACE exposures. For each increase in cumulative ACE score, odds of frequent headache increased by 1.22 (95% confidence interval [CI] 1.15 to 1.30). The ACEs that individually showed an association with frequent headache after adjusting for demographic factors were lack of maternal warmth (odds ratio [OR] 1.40, 95% CI 1.12 to 1.74, P = 0.002), lack of paternal warmth (OR 1.47, 95% CI 1.20 to 1.81, P < 0.001), paternal alcoholism (OR 1.21, 95% CI 1.05 to 1.40, P = 0.007), suicide attempt of family member (OR 1.51, 95% CI 1.22 to 1.87, P < 0.001), and living in an unsafe neighborhood (OR 1.22, 95% CI 1.06 to 1.39, P = 0.004).

Conclusions:

Several ACE exposures were associated with frequent headache in adolescents. An increase in cumulative ACE exposure increased the odds of having frequent headache.

Introduction

The effects of chronic headache are far-reaching. At an individual level, headache leads to personal disability and diminished quality of life.^1,2 At the family and population levels, this disability causes decreased productivity among the segment of society of young to middle-aged adults who should be at their most productive.³ This disability is mirrored in children and adolescents with headache, predisposing them to increased absenteeism and poor academic achievement.^4,5 Globally, among 10- to 14-year-olds and 15- to 16-year-olds, migraine is the fifth and sixth most disabling disease, respectively.³

Although we often think about the genetic aspects of migraine, prior work has demonstrated that adverse childhood experiences (ACEs) are associated with headache. ACEs are common—46.3% of children in the United States experience at least one ACE.⁶ These exposures may contribute to headache prevalence and disability through the impact of toxic stress on brain development.⁷ The foundational ACE Study, a large retrospective survey of adult patients from San Diego assessed exposure to ACEs and their relationship to health outcomes and demonstrated a dose response relationship between cumulative ACE score and risk/prevalence of headache in adults.^8,9

The National Longitudinal Study of Adolescent to Adult Health (Add Health) followed a representative sample of teenagers with the intent of assessing adolescent health behaviors and social issues.¹⁰ History of adverse events and health information, including headache frequency, was obtained directly from adolescents. We used Add Health data to examine ACE exposure, cumulative ACE score, and the relationship to frequent headache in this cohort of adolescents. Our objective is to understand the effects of specific ACEs and cumulative ACE exposure on risk for adolescent headache.

Methods

Add Health was approved by the Institutional Review Board of the University of North Carolina at Chapel Hill. Wave I data were collected in 1994 to 1995 from 20,745 seventh to twelfth grade students aged 11 to 17 years. Add Health utilized a stratified, random school selection process, and students were sampled from 145 middle and high schools in North Carolina. The present retrospective cohort study is composed of a secondary analysis of deidentified data from Add Health Wave I and is therefore exempt from review by the University of Pennsylvania Institutional Review Board. The following question was asked in Wave I to elicit headache history from the adolescent, “In the past 12 months how often have you had a headache?” Answer choices were never, rarely, once a week, almost daily, and daily. Answer choices were recoded into a dichotomous categorical variable: “1” yes frequent headache (reporting headache weekly, almost daily, or daily) versus “0” no frequent headache (reporting headache never or rarely). The following ACE variables were coded to represent either absence of exposure (0 = no exposure) or presence (1 = exposure): maternal alcoholism, paternal alcoholism, low parental warmth, witnessing a shooting or stabbing, experiencing a shooting or stabbing, having a knife or gun pulled on you, being cut or stabbed, being jumped or beaten up, living in an unsafe neighborhood, and having a family member attempt or complete suicide. See Table 1 for further information on the questions that elicited these ACEs and the recoding of the responses. All questions were answered by the participating adolescent except for parental alcoholism that was answered by the parents. The original ACE exposures in the Kaiser study that were elicited in Wave I of Add Health included lack of parental warmth, parental substance abuse, and household mental illness. Respondents were excluded from analysis of specific ACEs if a question was not answered. Odds ratios (OR) and 95% confidence intervals (CI) were obtained by performing logistic regression to identify which ACEs were individually associated with frequent headache while controlling for the following demographic factors: age, sex, race/ethnicity, and food and housing insecurity (which together served as proxies for low income status). Analyses were adjusted for the stratified and clustered sampling design of Add Health and included sampling weights that adjusted for differential response.

TABLE 1.

Adverse Childhood Experience Measure Information

ACE Category and Measure	Question	Response Choices	Recoded Variable	Percentage of Sample Reporting ACE in Recoded Variable
Victimization
Witnessing a shooting or stabbing	During the past 12 months, how often did the following happen? You saw someone shoot or stab another person.	0 = Never, 1 = once, 2 = more than once	0 = No exposure 1 = Any exposure	12.1
Having a knife or gun pulled on you	During the past 12 months, how often did the following happen? Someone pulled a knife or gun on you.	0 = Never, 1 = once, 2 = more than once	0 = No exposure 1 = Any exposure	12.5
Experiencing a shooting or stabbing	During the past 12 months, how often did the following happen? Someone shot or stabbed you.	0 = Never, 1 = once, 2 = more than once	0 = No exposure 1 = Any exposure	1.3
Experiencing being jumped or beaten up	During the past 12 months, how often did the following happen? you were beaten or jumped?	0 = Never, 1 = once, 2 = more than once	0 = No exposure 1 = Any exposure	11.1
Lack of parental warmth
Lack of maternal warmth	Most of the time your mother/father is warm and loving towards you?	1 = Strongly agree 5 = Strongly disagree	Recoded as 1 = disagree, strongly disagree 0 = strongly agree, agree	3.6
Lack of paternal warmth				4.7
Household substance abuse
Maternal alcoholism	Please tell me whether [participant’s] biological mother or his/her biological father has alcoholism now.	Yes/No	Recoded as No = 0, Yes = 1	2.6
Paternal alcoholism		Yes/No		12
Household mental illness
Family member suicide attempt	Have any of your family members tried to kill themselves during the past 12 months?	Yes/No	Recoded as No = 1, Yes = 0	4.8
Family member suicide	Have any of your family members succeeded in committing suicide in the past 12 months?	Yes/No	Recoded as No = 1, Yes = 0	1.1
Neighborhood safety
Not feeling safe in current neighborhood	Do you feel safe in your neighborhood?	Yes/No	Recoded as No = 1, Yes = 0	10.4

Abbreviation:

ACE = Adverse childhood experience

Table based on Brumley et al.’s representation of ACE measures.¹¹

We also created a cumulative ACE score adding individual ACE exposures (for participants who answered all ACE exposure questions) into a single measure (consistent with the methodology used by Brumley et al.).¹¹ Analysis of the relationship between cumulative ACE score and frequent headache therefore excluded participants with missing values for individual ACEs. Logistic regression was used to examine the relationship between cumulative ACE score and the binary outcome of frequent headache, again controlling for age, sex, race/ethnicity, and food and housing insecurity. Analyses were conducted using STATA version 14 (Statacorp, College Station, TX, USA).

Results

The study population was composed of 20,745 participants; 50.6% male and 49.4% female. The mean age of respondents was 15.9 years (range 12 to 21 years, standard error: 0.12 years), and 65.6% of participants identified their race as white (Table 2). Frequent headache was present in 29.3% of participants. The ACEs that individually showed an association with frequent headache after adjusting for demographic factors were lack of maternal warmth (OR 1.40, 95% CI 1.12 to 1.74, P = 0.002), lack of paternal warmth (OR 1.47, 95% CI 1.20 to 1.81, P < 0.001), = paternal alcoholism (OR 1.21, 95% CI 1.05 to 1.40, P = 0.007), suicide attempt of family member (OR 1.51, 95% CI 1.22 to 1.87, P < 0.001), and living in an unsafe neighborhood (OR 1.22, 95% CI 1.06 to 1.39, P = 0.004). See Table 3. Respondents who reported family member suicide = attempt were asked about family member suicide. History of attempted suicide increased the odds of headache. Among those with a history of family suicide attempt, completed suicide appears to potentially convey additional increase in odds of headache, but the findings did not meet statistical significance possibly due to a small subsample of respondents reporting this exposure.

TABLE 2.

Participant Characteristics (n = 20,745)

Characteristics	%
Sex
Boys	50.6
Girls	49.4
Race and ethnicity*
White	65.6
African American	15.1
Asian	3.1
Hispanic/Latino	11.0
Other	5.2

^*Participants could select multiple races/ethnicities.

TABLE 3.

Association of Frequent Headaches in Participants With Specific Adverse Childhood Experiences

ACE Category	Adjusted Odds Ratio*, 95% CI, P Value
Victimization
Witnessing a shooting or stabbing	1.12 (0.98–1.28), P = 0.086
Having a knife or gun pulled at you	1.03 (0.90–1.19), P = 0.595
Experiencing a shooting or stabbing	1.07 (0.73–1.58), P = 0.695
Experiencing being jumped or beaten up	0.96 (0.83–1.11), P = 0.621
Lack of parental warmth
Lack of maternal warmth	1.40 (1.12–1.74), P = 0.002
Lack of paternal warmth	1.47 (1.20–1.81), P < 0.001
Household substance abuse
Maternal alcoholism	1.29 (0.99–1.68), P = 0.059
Paternal alcoholism	1.21 (1.05–1.40), P = 0.007
Household mental illness
Family member suicide attempt	1.51 (1.22–1.87) P < 0.001
Family member suicide	1.18 (0.69–2.01) P = 0.541
Neighborhood safety
Not feeling safe in current neighborhood	1.22 (1.06–1.39), P = 0.004

Abbreviation:

ACE = Adverse childhood experience

^*Adjusted for age, sex, race/ethnicity, food, and housing insecurity.

For each additional ACE in the cumulative ACE score, the odds of frequent headache increased by 1.22 (OR 1.22, 95% CI 1.15 to 1.3, P-value < 0.001). In our adjusted model of cumulative ACE score, associations with frequent headache, sex, race, and ethnicity were found (Table 4). ACE scores were also significantly associated with frequent headache (Table 4). Compared with males, females had greater than double the odds of reporting frequent headache (OR 2.29, 95% CI 2.06 to 2.66, P value <0.001). White race when compared with non-white race was associated with increased reporting of frequent headache (OR 1.63, 95% CI 1.46 to 1.81, P < 0.001). Hispanic ethnicity when compared with non-Hispanic ethnicity was associated with decreased reporting of frequent headache (OR 0.63, 95% CI 0.52 to 0.82, P value < 0.001).

TABLE 4.

Association of Frequent Headaches With Cumulative ACE Score and Specific Demographic Characteristics

Characteristic	Adjusted Odds Ratio* 95% CI, P Value
Cumulative ACE score	1.22 (1.15–1.30), P < 0.001
Female	2.29 (2.06–2.66), P < 0.001
Food insecurity	1.12 (0.98–1.28), P = 0.086
Housing insecurity	0.91 (0.68–1.21), P = 0.531
White	1.62 (1.46–1.81), P < 0.001
African American	0.77 (0.69–0.85), P < 0.001
Asian	0.50 (0.40–0.61), P < 0.001
Hispanic/Latino	0.65 (0.55–0.77), P < 0.001
Other	0.65 (0.53–0.79), P < 0.001

Abbreviation:

ACE = Adverse childhood experience

^*Adjusted for age, sex, race/ethnicity, food, and housing insecurity.

Discussion

This study from adolescent report adds further evidence for the association between cumulative ACE exposure and frequent headache, which has been previously demonstrated in adults.^7,9,12,13 Few studies have investigated the potential association of ACE exposure and headache in childhood and adolescence.^14–16 Mansuri et al. examined this relationship in persons younger than 18 years using the 2016 and 2017 National Survey of Childhood Health¹⁶; it was a parental survey that asked questions about ACEs and health issues. The authors found an overall dose response relationship between cumulative ACE exposure and headache such that each additional ACE >4 increased the odds of headache by 3.40. Individually the only ACE that was associated with headache was adversity due to income (a nontraditional ACE). Our study did not use food and housing insecurity as nontraditional ACEs but rather used these measures as a proxy for income/control measure. Another study that has shown the relationship between childhood adversity and headache included ~4600 high school students in Taiwan. The adolescents in this study who met criteria for chronic daily headache (defined as 15 or more headache days per month lasting two or more hours per day) reported more childhood adversity than adolescents without chronic daily headache.¹⁴ In adults with migraine, three large population-based studies similarly showed dose-response associations between ACE exposure and relative odds of migraine.^7,12,13 A small case control study of premenopausal women with migraine demonstrated a positive correlation between ACE score and monthly headache frequency.¹⁷ No prior study has examined the burden of ACEs on headache prevalence during adolescence from adolescent self-report. The critical value of self-report in adolescents has been demonstrated in prior studies regarding quality of life in teenage patients with chronic health conditions.^6,18

We provide clear evidence of the association between self-reported ACE exposure and the prevalence of headaches in adolescents. Our study showed that multiple individual ACEs were independently associated with increased risk of frequent headache, including lack of parental warmth, suicide attempt in a family member, paternal alcoholism, and living in an unsafe neighborhood. Lack of parental warmth may suggest emotional neglect, which is defined in the original ACE study as feeling that “no one in your family loved you or thought you were special, or that your family didn’t look out for each other, feel close to each other, or support each other,”⁹ but without further information, this is uncertain. Lack of parental warmth may reflect a spectrum from normative dips in adolescents’ perceptions of their relationships with their parents to parental behavior that is genuinely abusive or neglectful. Emotional neglect in childhood has been associated with migraine in adulthood, including in the population-based American Migraine Prevalence and Prevention study.^7,17 Notably, the ACE, American Migraine Prevalence and Prevention, and Add Health studies in adults each reported that emotional abuse was the ACE most strongly associated with frequent headache or migraine.^7,9,13 Questions surrounding emotional abuse were not specifically asked during Wave I of Add Health.¹⁰

Interestingly, although we controlled for food insecurity and housing insecurity as proxies for income status in our analysis, we did not find that these factors were independently associated with ACE score and frequent headache. This contrasts with the study by Mansuri et al., which categorized difficulty due to family income as a nontraditional ACE and was the only individual ACE in their study to be independently associated with headache.¹⁶ This also conflicts with a study by Bigal et al.,¹⁹ which showed that socioeconomic status was associated with migraine diagnosis in adolescents without a family history of migraine; however, their study did not specifically assess ACE exposure. It is unclear why some studies have shown an association between income level and headache, whereas ours did not. Low income status was not categorized as an ACE in the original ACEs study, which was performed in a largely middle-class cohort and demonstrated significant associations between ACE exposures and health outcomes.⁸ Other studies have observed that associations between ACE exposure and health outcomes in children are distributed across the income gradient and not clustered at lower incomes.²⁰ More research must be conducted to fully understand the effects of income level on headache and headache disability.

In our study, only white race was associated with increased frequent headache and Hispanic ethnicity was associated with decreased frequent headache. Prior migraine-specific prevalence data in adults have shown nearly identical rates of migraine between whites, Hispanics, and blacks and demonstrated that the highest rates of chronic migraine were among Hispanic women (2.26%).²¹ This discrepancy in findings between our study and prior literature raises the question of variation in reporting patterns in adolescents from different racial and ethnic groups. It is worth noting that within Hispanic groups, there is great cultural and linguistic heterogeneity based on the country of origin and distribution of specific groups across the United States. A small focus group survey of foreign-born Spanish speaking Hispanic patients showed that reporting of pain may vary based on language and cultural differences between participants and health care providers.²² It is also possible that within different groups, headaches and migraines may be understood as two distinct entities, which may explain differences in reporting of both. Future studies should focus on differences in ACE exposure, headache prevalence, and health care access and utilization across different racial and ethnic groups.

Strengths of our study include the large, nationally representative sample of adolescents. The young age of the population may potentially decrease recall bias in answering questions about adverse childhood events and headache frequency. In addition, Wave I queries exposure to community violence, information that is missing from many other studies of the effects of childhood adversity on health.

Limitations include the restricted number of ACEs assessed in Wave I, particularly the absence of information on frequency, duration, and age at onset of each ACE exposure, as well as absence of information on age at onset of headache and of specific headache diagnosis and psychiatric comorbidities. As such, we were unable to control for these comorbidities that could potentially mediate the relationship between ACE and migraine. More specifically, we do not have data to corroborate self-report of ACE. In addition, our ACE scale is limited as it does not capture abuse and neglect ACEs. We note that emotional, physical, and sexual abuse were measured in Add Health starting at Wave 3, and further studies should include the evaluation of Add Health data from multiple waves to gain a fuller picture of the relationship of frequent headache in childhood to later migraine diagnosis.

One important limitation of this study is its retrospective design that can only illustrate associations between ACE exposure and headache frequency. Ideally, a prospective study designed to follow participants from early childhood that regularly analyzes both ACE exposures and health outcomes could better demonstrate potential causative health ramifications of toxic stress and specific ACEs across childhood and adolescence. It is worth noting that exposure to stress as early as in utero has been found to affect neurodevelopmental and neuropsychiatric outcome in children.²³ Studies have showed how maternal stress during pregnancy can act as a modulator of maternal immune activation increasing susceptibility to developing neurodevelopmental disorders as children.²³

The relationship between childhood adversity and headache in adults and children has now been found in multiple studies, although causality has not been determined. Preclinical evidence suggests that early toxic stress that results from adverse childhood experiences affects the development of the neuroendocrine hypothalamic-pituitary-adrenal axis and causes structural changes in the brain specifically affecting normal growth and development of the hippocampus, prefrontal cortex, and amygdala.²⁴ These structures play a key role in pain pathways and are especially vulnerable to the effects of toxic stress and inflammation early in life. A 2012 study of premenopausal women with migraine demonstrated a strong positive correlation between ACE score and vascular inflammatory biomarker levels, supporting the theory that toxic stress is linked to migraine and stroke via inflammation.¹⁷ Additional investigation in children and adolescents should examine ACE exposure, inflammatory markers, and brain morphometry through magnetic resonance imaging to better understand the physiologic and neuroanatomical effects of ACEs on headache and, specifically, migraine.

Conclusion

In this study, frequent headache in adolescents was associated with a variety of self-reported ACE exposures, most notably lack of parental warmth, family member suicide attempt, and lack of neighborhood safety. An increase in cumulative ACE exposure increased the odds of having frequent headache. This reaffirms what has been found in existing adult and pediatric literature regarding ACEs and their potential deleterious effects on headache burden. Further investigation including prospective cohort studies should be performed in children and adolescents to better clarify the relationship between childhood adversity and headache; this can hopefully elucidate possible mechanisms of pathophysiology and drive the development of new, more effective treatments.