Article Text
PDF
Abstract
Background Limited research has shown a possible association between exposure to physical or sexual abuse prior to age 18 and the risk of developing hypertension as an adult. The factors mediating this relationship are unknown.
Methods Questionnaire data from 68 505 female participants in the Nurses' Health Study II were analysed regarding exposure to physical and sexual abuse prior to age 18. Cox proportional hazards regression was used to assess the relationship between abuse exposure and hypertension.
Results 64% of the participants (n=41 792) reported physical and/or sexual abuse prior to age 18; 17% reported hypertension. All forms of abuse had a dose-response relationship with hypertension. Adjustments for smoking, alcohol, family history of hypertension, exercise and oral contraceptives did not alter risk estimates. Adjustment for body mass index (BMI) significantly attenuated the associations between abuse and risk of hypertension and accounted for approximately 50% of the observed association between abuse exposure and hypertension. Women experiencing forced sexual activity as a child and as an adolescent had a 20% increased risk for developing hypertension (95% CI 8% to 32%) that was independent of BMI. Similarly, women reporting severe physical abuse in childhood and/or adolescence had risk estimates ranging from 14% (95% CI 5% to 24%) to 22% (95% CI 11% to 33%).
Conclusion Early interpersonal violence may be a widespread risk factor for the development of hypertension in women. BMI is a significant mediator in the relationship between early abuse and adult hypertension.
- Child abuse
- children
- domestic violence
- hypertension
- violence RB
- women CG
Statistics from Altmetric.com
Background
The global burden of hypertension is extensive, affecting an estimated 972 million adults worldwide.1 2 Exposure to certain types of violence may increase the risk of developing hypertension.3 4 Community-wide violence (gunshot wounds, fractures, stabbings, burns and rape) has been shown to triple rates of hypertension in exposed populations.5 Exposure to terrorist violence, such as the September 11th attack on the USA, has been shown to increase rates of hypertension and stroke in adults for up to 3 years following the event.6
Exposure to interpersonal violence, particularly during childhood, may also be a risk factor for hypertension. The scope of this violence is staggering; in 2007 nearly 1 million cases of child abuse were verified by authorities in the USA.7 8 Early life exposure to violence has been shown to be a possible risk factor for a number of health conditions including chronic pain,9 autoimmune disease,10 ischaemic heart disease,11 headache,12 fibromyalgia13 and asthma.14 Findings from a small number of studies suggest that exposure to violence in childhood may also be a risk factor for the development of hypertension. One early study evaluated childhood trauma (parental death, parental psychosis, separation from a parent or being beaten by a parent) in 35 men, 17 of whom had a diagnosis of hypertension; 64% of the hypertensive men reported childhood trauma compared to 27% of the normotensives.15
The association of specific types of interpersonal violence with hypertension has not been clearly established. In a community-based study of 323 women evaluating the impact of psychosocial factors on health, childhood psychological and sexual abuse, but not physical abuse, were associated with hypertension.16 In 2009, the association between blood pressure, adverse environment (exposure to physical abuse, verbal abuse, sexual abuse or harsh parenting), and expression of negative emotions (depression, anxiety or anger) was measured in a group of 2739 adults. Only an indirect association, via the presence of negative emotions in study participants, was found between adverse environment and elevated blood pressure.17 Finally, when Nomura and colleagues prospectively followed 1704 children with a history of low birth weight, they found no significant relationship between a history of child abuse and adult-onset hypertension.18
Established risk factors for the development of hypertension include age, race, family history of hypertension, inactivity, overweight or obesity, tobacco use and alcohol use. Among these risk factors there is evidence that weight, alcohol use19 and smoking20 may be influenced by violence exposure. In particular, research on BMI indicates that adults with a history of child abuse are more likely to be overweight than adults with no abuse history.21–23 Although research has shown an association between child abuse and adult-onset hypertension, little is known about the magnitude of this association or the factors that may mediate it. It was hypothesised that exposure to physical and sexual violence prior to age 18, relative to being unexposed, is associated with increased risk of developing hypertension in adulthood. In addition, this study aimed to identify if established risk factors for hypertension, particularly smoking, alcohol use and weight explained any observed association between abuse and hypertension.
Methods
Participants in the Nurses' Health Study II (NHS II) were studied, a prospective cohort of 116 686 female nurses age 25–44 years at baseline, from 14 US states.24 The purpose of the NHS II was to investigate long-term effects of oral contraceptive use, diet and lifestyle. Since the inception of the cohort in 1989, participants were mailed biennial questionnaires, including sociodemographic, behavioural, comorbidity and dietary domains.
A violence questionnaire designed to ascertain abuse across the lifecourse was mailed to 91 286 study participants in 2001. Questionnaires were returned by 68 505 participants (75%). This study was approved by the Institutional Review Board at Brigham and Women's Hospital in Boston, Massachusetts, USA and the Human Subjects Committee at the Harvard School of Public Health, Boston, Massachusetts, USA. Completion and return of the supplementary questionnaire was an indicator of consent.
The violence questionnaire included assessments of abuse in childhood, adolescence and adulthood. For this analysis, physical and sexual abuse exposures reported for childhood (up to age 11 years) and adolescence (11–17 years) were examined.
Questions from the Revised Conflict Tactics Scale (CTS)25 were used to measure physical abuse. Participants were asked to indicate the type of abuse sustained from a parent, step-parent, or adult guardian, as well as the frequency of abuse (never, once, a few times or more than a few times). Physical abuse was categorised as: no abuse; mild to moderate abuse (hit, pushed, grabbed or shoved); severe abuse as child or adolescent (kicked, bitten, punched, choked, burned, or physically attacked as a child or adolescent); and severe abuse as child and adolescent (kicked, bitten, punched, choked, burned, or physically attacked as a child and as an adolescent).
Sexual abuse was measured using questions from a national telephone survey conducted by the Gallup Organisation in 1995.26 Questions were included on forced sexual touching, ‘Were you ever touched in a sexual way by an adult or an older child or were you forced to touch an adult or an older child in a sexual way when you did not want to?’, and forced sexual activity, ‘Did an adult or older child ever force you or attempt to force you into any sexual activity by threatening you, holding you down, or hurting you in some way when you did not want to?’ Participants answered ‘No, this never happened’, ‘Yes, this happened once’, or ‘Yes, this happened more than once’. Sexual abuse was categorised as: no abuse, mild to moderate abuse (touched in a sexual way as child and/or adolescent), severe child or adolescent abuse (forced sexual activity as a child or adolescent), and severe child and adolescent abuse (forced sexual activity as a child and as an adolescent).
As the co-occurrence of physical and sexual abuse is highly correlated in the literature,27 a cumulative abuse indicator was derived, which combined subtypes of physical and sexual abuse across childhood and adolescent periods. Categorisations of cumulative abuse, combining exposure type, severity, number of events and timing of abuse in either childhood or adolescence, are detailed in table 1.
Combined physical and sexual abuse in childhood and/or adolescence: distribution of women in cumulative abuse Levels 1–6
Hypertension diagnosis by a physician was self-reported biennially. The validity of self-reported hypertension was assessed by obtaining medical records from samples of women in the original Nurses' Health Study.28 29 Of 51 women who reported hypertension for whom we obtained medical records, hypertension (blood pressure greater than 140/90) was confirmed in all cases. In a second validation survey, blood pressure was measured in a sample of Boston-area NHS II participants; among the 161 participants who did not report hypertension, none had a blood pressure greater than 160/95 mm Hg and 6.8% had values between 140/70 and 160/95.
Covariates were considered for analysis based on their established association with hypertension and their role as potential confounders of the relationship between abuse and hypertension.30–33 Race/ethnicity was categorised as Caucasian, African–American, Hispanic, Asian, and other. Alcohol use was measured in grams consumed per day (no alcohol, 0.1 to <5 g/day, 5 to <15 g/day, and >15 g/day). Metabolic equivalents of exercise per week were calculated from self-reported activity.34 BMI (kg/m2) was calculated from height reported in 1989 and weight reported on each biennial questionnaire. A childhood somatogram score was included, derived from nine female body figures. Participants chose the diagram that best depicted their body type at age 5 (scored from 1 (very thin) to 9 (extremely obese)).35 Family history of hypertension in any first degree relative was assessed at baseline and was updated in 2003. Smoking, age and oral contraceptive use were updated biennially. For a subset of 29 100 nurses (45%) whose mothers reported parental occupations at the time of the participant's birth, adjustments were made for socioeconomic status.
The number of women who completed the violence questionnaire was 68 505; 3772 women who were diagnosed with hypertension prior to 1989 were excluded. Women were followed from 1989 until 2003, or until they developed hypertension, died or were lost to follow-up.
Missing indicators were used to model missing covariate status. Cox proportional hazards regression was used to estimate HRs and 95% CI for the relationship between abuse and hypertension. Covariates were added to the models based on their established relationship with hypertension. BMI was examined as a continuous covariate as well as in categories (<18.5 kg/m2, 18.5–24.9 kg/m2, 25–29.9 kg/m2, 30+ kg/m2). The effect of BMI on the association of abuse with hypertension was evaluated using BMI as a continuous variable. Proportion and 95% CI were estimated of the abuse association that was explained by adult BMI using the SAS mediation macro of Spiegelman and colleagues.36
The years 2001–2003 were examined separately, following administration of the 2001 violence questionnaire. All models were re-run, excluding the 45 635 women who reported physical or sexual abuse exposure as adults.
Results
Sixty-four per cent of the participants (n=41 792) reported abuse prior to age 18. Thirty-three per cent reported sexual abuse and 53% reported physical abuse. There were 4328 women (7%) who reported severe physical abuse by a parent, step-parent or guardian, and 11% who reported forced sexual activity.
There was very little difference across cumulative levels of abuse with respect to age, race/ethnicity, parity, exercise and family history of hypertension (table 2). Women exposed to violence prior to age 18 were less likely to use oral contraceptives or drink alcohol compared to those with no abuse. Women reporting exposure to violence were more likely to smoke cigarettes, had slightly increased mean BMI at age 18 and larger increases in BMI at baseline (table 2).
Population characteristics at baseline (1989) by cumulative early life abuse scale, Nurses' Health Study II
Between 1989 and 2003, 17.4% of the cohort reported incident hypertension. Sexual abuse had a dose-response relationship with risk of incident adult hypertension (age- and race-adjusted) ranging from a 5% (95% CI 0% to 9%) increased risk associated with forced sexual touching to 47% (95% CI 33% to 62%) with forced sexual activity as a child and adolescent (table 3). Findings were similar for physical abuse, ranging from a 6% (95% CI 2% to 10%) increased risk among those reporting mild to moderate physical abuse, to a 27% increased risk (95% CI 17% to 38%) among women reporting severe physical abuse (table 3).
Sexual abuse, physical abuse, and cumulative physical and sexual abuse in childhood and/or adolescence and the multivariate HR and 95% CI for adult hypertension
Adjustments for smoking, alcohol use, family history of hypertension, somatogram score, exercise and oral contraceptive use made no difference to the risk estimates associated with hypertension (table 3). Further adjustment for parental occupation made no difference to these estimates (data not shown).
Inclusion of adult BMI in the models attenuated associations of child and adolescent abuse with risk of adult hypertension (table 3); the results were similar whether BMI was parameterised as continuous or categorical. After adjustment for BMI, only severe physical abuse and forced sexual activity remained associated with statistically significant increases in the risk of hypertension (ranging from 14% to 20%) compared to women with no history of abuse.
The findings for cumulative abuse were similar to those for physical and sexual abuse. Compared to women who reported no abuse, the risk of hypertension (age- and race-adjusted) rose in a dose-response fashion with cumulative abuse exposure, from 4% (95% CI 1% to 8%) among women reporting Level 2 abuse, to 59% (95% CI 42% to 78%) among women reporting Level 6 abuse (table 3). BMI adjustment attenuated the estimates considerably, with increased risk of hypertension remaining significant for only Level 5 (16% (95% CI 5% to 28%)) and Level 6 (28% (95% CI 14% to 44%)) cumulative abuse.
Adult BMI accounted for 48% of the association of severe physical abuse with hypertension and for 60% of the association of severe sexual abuse with hypertension. The effects of BMI on cumulative abuse and risk of hypertension were similar, accounting for 59% (range 42–76%) of the association between cumulative abuse, Levels 3 through 6, and hypertension.
In the 2 years of follow-up after the administration of the 2001 abuse questionnaire there were 2283 incident reports of hypertension. The same associations were observed in this group of prospectively detected cases as in the cases reported between 1989 and 2001 (data not shown). Associations of abuse restricted to childhood and/or adolescence with risk of hypertension were similar to associations reported for the entire cohort (data not shown).
Discussion
The results of the present study suggest that women exposed to severe physical and/or sexual abuse prior to age 18 are more likely to develop hypertension than women with no abuse history or less severe forms of abuse exposure. Although adult adiposity mediated a large part of the association between early abuse and adult hypertension, there remained a 20% increased risk with the most severe abuse that was independent of BMI.
This study is by far the largest of its kind, including over 68 000 women from 14 US states. Notably, the prevalence of physical and sexual abuse reported by this cohort is similar to the prevalence reported in national surveys.37 38
Abuse exposure was recalled on a questionnaire administered in 2001, and hypertension was self-reported on questionnaires from 1991 to 2003. If administration of the violence questionnaire influenced patients' subsequent reporting of hypertension, it is likely that a difference would have been seen in the relationship between abuse and hypertension in the subset of women diagnosed with hypertension after 2001. This was not the case, as the relationship between abuse and hypertension was similar in the women with hypertension occurring after the violence questionnaire was administered.
There were no data on depression or antidepressant medications for the entire follow-up period. In women with a history of abuse, depression is more common than in women with no abuse history.33 Certain antidepressants can cause hypertension. Given that depression itself has been shown to be associated with lower blood pressure,39 it is unclear what impact, if any, the measurement of depression may have had on the present analysis. This warrants consideration in future research.
These results may not be generalisable to other groups, given the present cohort of nurses had a higher than average level of education and was 95% Caucasian.
Women with hypertension prior to the start of the NHS II were excluded from this analysis. These women could have developed hypertension earlier as a result of their exposure to violence prior to age 18. Although this issue was not addressed, it is suggested that the reported results may be a conservative estimate of the association between early abuse and adult hypertension.
The impact of the co-occurrence of physical and sexual abuse was estimated via the cumulative abuse variable. Children who experience one type of maltreatment are at increased risk of other types of violence, with reported coexistence of physical and sexual abuse ranging from 43% to 71%.27 40 Moreover, research suggests that experiencing multiple forms of abuse, compared to any one abuse subtype, may be more detrimental to health.41 This is corroborated in the present analysis as there was no significant difference between cumulative abuse estimates and those analysed separately for physical and sexual abuse, suggesting that abuse severity and frequency, and not subtype, may have the greatest impact on adult health.
Overweight and obesity are strong intermediate factors driving the association of child/adolescent abuse and risk of hypertension in the present cohort. Data in table 2 suggest that early abuse sets women on different weight trajectories in adulthood: whereas there was little difference in BMI at age 18 by abuse categories, there was a sizeable trend in increasing BMI with increasing abuse history by 1989, when the participants were age 25–44. In support of the proposition that abuse precedes weight gain, Noll and colleagues examined prospectively collected data on weight to report that the obesity rates of girls with and without sexual abuse histories were similar in childhood, began to diverge in adolescence, and were greater among young adults who had suffered sexual abuse as girls.21 Bentley et al reported that children with a history of physical abuse were more likely to have higher BMI as adults than those who experienced no physical abuse; this association remained despite adjustment for age, race, smoking and alcohol use.23 Thus, it seems plausible that child and adolescent abuse exposure increases the risk of overweight and obesity among young adult women, increasing their risk for hypertension.
After adjustment for BMI, an association of severe physical and/or sexual abuse with hypertension remained, suggesting that abuse prior to age 18 may affect the risk of adult hypertension through a mechanism independent of BMI. One possible mechanism for the development of hypertension in women with a history of abuse is fear-induced alterations to the developing hypothalamic-pituitary-adrenal (HPA) axis.42 Serum cortisol levels in patients placed in stressful situations correlate with blood pressure, obesity and other markers of cardiometabolic risk.43 Several studies have reported altered cortisol or diurnal patterns among children exposed to violence.44–47 Others have reported increased cardiac output, stroke volume and peripheral resistance among children exposed to community and/or interpersonal violence.48 Importantly, altered HPA axis function has been shown to persist long after abuse has ceased.49 In women with a history of child abuse, peak adrenocortictotropin levels and urinary cortisol levels are higher than levels in non-abused controls in response to stressful situations. These findings are consistent among populations suffering from mental illness or substance abuse.50
One in four women has experienced domestic violence at some point in her lifetime.38 The present findings, that severe physical and sexual abuse prior to age 18 are associated with increased risk of hypertension in adulthood, independent of BMI, suggest that early interpersonal violence is a risk factor for the development of hypertension in women. Roughly half of the association between early abuse and hypertension was explained by greater adiposity among women who had been abused as girls. This suggests that the prompt detection of early abuse, followed by psychological counselling, stress and weight management, may help to prevent adult hypertension.
What is already known on this subject
There have been a small number of studies in the past 10 years evaluating the possible role of violence exposure in the development of hypertension.
These studies suggest that exposure to childhood and adolescent interpersonal violence may increase the risk for developing hypertension in adulthood.
The number of participants in these studies has been small and the population characteristics vary widely. In addition, the types of violence and methods of documenting the violence exposure have been inconsistent.
What this study adds
This study is by far the largest of its kind, including over 67 000 women from 14 US states.
Its measurement of physical and sexual abuse, as well as severity and frequency of abuse, offers a greater level of detail than any prior analyses.
This study emphasises the role of BMI as a mediator in the relationship between abuse and risk of hypertension.
Despite the mediating effect of BMI, a statistically significant increase in the risk of developing adult onset hypertension remains in women with the most severe physical and/or sexual abuse exposures prior to age 18.
References
Footnotes
Funding This work was funded by National Institute of Health grants HL081557, CA50385, and HL64108. The study sponsors had no role in the study design, collection, analysis and interpretation of data, writing or editing of the manuscript, or in the decision to submit for publication.
Competing interests None.
Ethics approval This study was conducted with the approval of the Brigham and Women's Hospital Institutional Review Board and the Human Subjects Committee at Harvard School of Public Health.
Provenance and peer review Not commissioned; externally peer reviewed.