Abstract
Introduction:
Lifetime exposure to interpersonal violence or abuse (IVA) has been associated with several chronic diseases including adult-onset diabetes, yet this pattern has not been confirmed by sex and race within a large cohort.
Methods:
Data from the Southern Community Cohort Study collected 2002-2009 and 2012-2015 were used to explore the relationship between lifetime IVA and diabetes (n=25,251). Prospective analyses of lower income people living in the southeastern United States (U.S.) were conducted in 2022 to examine risk of adult-onset diabetes associated with lifetime IVA by sex and race. Lifetime IVA was defined as 1) physical or psychological violence, threats or abuse in adulthood (adult IVA), and/or 2) childhood abuse or neglect (CAN).
Results:
After adjustment for potentially confounding factors, adult IVA was associated with a 23% increased risk of diabetes (adjusted hazard ratio [aHR]=1.23; 95% confidence interval [CI]: 1.16-1.30). Diabetes risks associated with CAN were 15% (95% CI: 1.02-1.30) for neglect and 26% (95% CI: 1.19-1.35) for abuse. When combining adult IVA and CAN, the risk of diabetes was 35% higher (aHR=1.35; 95% CI: 1.26-1.45) than those experiencing no violence, abuse, or neglect. This pattern held among Black and White participants and among females and males.
Conclusions:
Both adult IVA and CAN increased the risk of adult-onset diabetes in a dose-dependent pattern for males and females and by race. Intervention and prevention efforts to reduce adult IVA and CAN could not only reduce risk of lifetime IVA but may also reduce one of the most prevalent chronic diseases, adult-onset diabetes.
Introduction
Diabetes affects approximately 11% of adults in the United States (U.S.), with people of color (Blacks 12.1% vs. Whites 7.4%) and lower income people (<100% federal poverty level [FPL] 14.1% vs. ≥500% FPL 5.6%) being disproportionately impacted.1 Lifetime interpersonal violence or abuse (IVA) has been directly2 and indirectly3 associated with adult-onset diabetes. Here, lifetime IVA is defined as 1) physical or psychological violence, threats or abuse in adulthood (adult IVA), and/or 2) childhood abuse or neglect (CAN), according to variables collected using the Adverse Childhood Experiences (ACE) questionnaire of occurrences prior to age 18 years.
Chronic stress, brought on by the trauma of lifetime IVA, has been linked with adult-onset diabetes among women.2 Re-occurring violence and accompanying stress may increase risk of diabetes via prolonged periods of elevated allostatic load.4–5 When the hypothalamus-pituitary-adrenal (HPA) axis is overactivated during allostasis, the body responds by elevating cortisol levels and suppressing insulin levels5 which directly impact diabetes risk through increased blood glucose levels.6–7 Findings from Pinna et al. (2014)3 establish a link between higher waking cortisol levels among females who had versus those who had not experienced adult IVA. Lifetime IVA can also indirectly contribute to diabetes risk through poor mental health outcomes. Females who experience lifetime IVA often report high rates of depression8–9 and post-traumatic stress disorder which are both also associated with dysregulation of the HPA axis6 and increased diabetes risk.7
Several studies report on the relationship between lifetime IVA and diabetes.8–14 Diabetes was associated with physical and sexual CAN in a dose-response fashion in the Nurses’ Health Study, a large cohort of female nurses.10 In agreement with a large cross-sectional study,8 a previous analysis of the Southern Community Cohort Study (SCCS) found a significant positive trend in diabetes associated with greater scores on the ACE questionnaire.11 A significantly elevated diabetes risk associated with any form of adult IVA was seen among female and male participants in the cross-sectional National Intimate Partner and Sexual Violence Survey (NISVS).9 Relatedly, Mason and colleagues (2013) found an increased risk of diabetes associated with adult physical and psychological IVA, but not sexual IVA in the large cohort of female nurses.12 A recent systematic review, of 52 studies conducted between 2012 and 2019 on adult IVA and physical health among females, observed an increased rate of diabetes associated with IVA.13 In a large cohort study of health care plan participants, adult IVA was associated with diabetes risk in men and women.14
The SCCS, a large prospective cohort study of people living in the southeastern U.S., is an ideal data source for the current analysis because it includes questions on both adult IVA and CAN. The community-engaged research strategy used to recruit the majority of SCCS participants, through community health centers, enabled access to a vulnerable southeastern population of racial minorities and people of lower socioeconomic status (SES), who are at greater risk of diabetes.15 Based on the above literature review, the hypothesis that both adult IVA and CAN would be associated with an increased risk of adult-onset diabetes was formed. Because adult IVA and CAN rates are higher among heterosexual females than males and the health impact of these forms of violence appear greater for females,8–9 an hypothesis was added to explore whether experiencing both adult IVA and CAN would have an increased risk of diabetes among females versus males. A significant contribution of this investigation was the ability to investigate the separate and combined effects of adult IVA and CAN associated with diabetes risk among Black and White residents of the southeastern U.S. Because race is a social construct and all members of this cohort are of lower socioeconomic status, no differences in the association between adult IVA and CAN among Black and White SCCS participants were anticipated. To the author’s knowledge, this is the first large prospective cohort analysis of the joint effects of adult IVA and CAN on the risk of diabetes by sex and race among those of similar income and education.
Methods
Study Sample
Detailed methods of the SCCS appear elsewhere.15–16 Briefly, the SCCS included approximately 85,000 males and females aged 40–79 who completed a baseline interview in 12 southeastern U.S. states between 2002 and 2009 after providing written informed consent. The majority (86%) were lower income persons enrolled at community health centers where computer-assisted in-person interviews were conducted. The remaining participants were sampled from the general population and completed telephone or mailed questionnaires.
Measures
Demographic characteristics collected at baseline included age, household income, highest level of education, and current marital status. Information on prior health conditions included self-reported diagnoses of a heart attack or myocardial infarction, any cancer, stroke, hypertension, and depressive symptoms. The 10-item Center for Epidemiological Studies-Depression scale ranging from 0 (rarely or none of the time) to 3 (most or almost all the time) was summed, and scores categorized as ≥10 symptoms as depressed and 0-9 as not depressed. The second follow-up survey, completed with 41,244 participants between 2012 and 2015 provided updated information on heart attack or myocardial infarction, any cancer, and stroke.
The primary outcome was adult-onset diabetes. In the baseline and second follow-up survey, participants were asked whether a doctor had ever told them they had diabetes; if yes, a follow-up question asked age at diagnosis. History of adult-onset diabetes was updated based on the second follow-up survey.
The primary exposure was lifetime IVA operationally defined as adult IVA and/or CAN. These experiences were included in the second follow-up survey only. New measures of adult IVA were included in the follow-up survey as three items measuring adult psychologic harm, physical violence, and threats using a gun or other weapon. These items were adapted from existing measures of adult IVA to query violence or abusive behaviors by a spouse, family member or close friend:17–18 ‘During your adult life, has your spouse, family member or close friend ever: 1) shouted, yelled, screamed, scolded, made fun of, severely criticized, said you were stupid or worthless, threatened, or psychologically harmed you? (n=6,350; 25.1%), 2) hit, punched, kicked, pushed, shoved, or otherwise physically hurt you? (n=7,792, 30.9%), and 3) threatened you with a gun or weapon? (n=3,297, 13.1%).’ These three adult interpersonal violence items were combined as one dichotomous variable; 9,040 or 35.8% experienced adult IVA.
CAN was defined from the appropriate ACE19 items and grouped using the following hierarchical scheme: 1) physical, sexual, or emotional abuse, with or without neglect, 2) physical or emotional neglect without any form of abuse, and 3) no abuse nor neglect (referent). Because adult IVA and CAN often co-occur and may magnify the stress exposure, hierarchically ordered variable of adverse experiences by combining those who experienced: 1) both adult IVA and CAN, 2) no adult IVA and CAN only, 3) adult IVA without CAN, 4) no adult IVA nor CAN (referent) was created.
Statistical Analysis
The mean follow-up time in this cohort analysis was 7.7 years. Participants were excluded from analysis for missing race or ethnicity (n=444), missing ACE questions with the exception of divorce or prison (n=3,881); missing all 3 adult IVA questions (n=994), and missing diabetes at baseline and follow-up (n=10). Participants of race/ethnicities other than non-Hispanic White and Black Americans (n=1,654) were excluded because there were too few for sub-analyses. To distinguish child versus adult-onset diabetes, those with diabetes diagnosed at less than age 30 (n=2,543) as a proxy for childhood onset diabetes, and participants whose diabetes was diagnosed prior to the baseline survey (n=6,467) were excluded. After these exclusions, 25,251 cases remained as the analytic sample.
Cox proportional hazards regression was used to estimate the adjusted hazard ratio of diabetes associated with adult IVA and CAN, with age used as the time metric of analysis. Person-time was accrued for each participant from age at study enrollment until the earliest of the following: age diagnosed with diabetes or December 31, 2015. An interaction term, the product of adult IVA and CAN and race (Table 3) or sex (Table 4), was added to Cox regression models and likelihood ratio tests were performed to test for effect modification. Statistical analyses were performed with SAS version 9.4 (Copyright © 2020, SAS Institute Inc., Cary NC, USA). The Institutional Review Boards of Vanderbilt University Medical Center (#010345) and Meharry Medical College (#14-03-132) approved this study’s protocol.
Table 3.
Lifetime Interpersonal Violence or Abuse (IVA) and Adult-Onset Diabetes Rates by Race (SCCS Participants)
| White SCCS Participants | Black SCCS Participants | Likelihood ratio test for race | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Lifetime IVA | Sample | Diabetes | aHRa | (95% CI) | Sample | Diabetes | aHRa | (95% CI) | |||||
| N | (Column %) | n | (row %) | N | (Column %) | n | (row %) | P-valueb | |||||
| All | 10,211 | 15,040 | |||||||||||
| Adult physical or psychological violence or abuse (IVA) | |||||||||||||
| No IVA (Referent, for all adult violence) | 6,284 | (61.5) | 703 | (11.2) | 1.00 | (REF) | 9,927 | (66.0) | 2,175 | (21.9) | 1.00 | (REF) | |
| Yes | 3,927 | (38.9) | 572 | (14.6) | 1.19 | (1.06-1.34) | 5,113 | (34.0) | 1,377 | (26.9) | 1.24 | (1.16-1.33) | 0.22 |
| By form | |||||||||||||
| Physical | 3,625 | (57.7) | 521 | (14.4) | 1.18 | (1.04-1.33) | 4,167 | (40.6) | 1,129 | (27.1) | 1.26 | (1.17-1.36) | 0.10 |
| Gun | 1,326 | (13.0) | 216 | (16.3) | 1.31 | (1.11-1.55) | 1,971 | (13.1) | 517 | (26.2) | 1.20 | (1.09-1.32) | 0.88 |
| Psychological | 2,676 | (26.2) | 420 | (15.7) | 1.27 | (1.12-1.45) | 3,674 | (24.4) | 995 | (27.1) | 1.25 | (1.16-1.35) | 0.59 |
| Adult violence hierarchical | 0.37 | ||||||||||||
| Physical or Gun | 3,713 | (36.4) | 538 | (14.5) | 1.19 | (1.05-1.34) | 4,501 | (29.9) | 1,207 | (26.8) | 1.24 | (1.16-1.33) | |
| Psychological alone | 214 | (2.1) | 34 | (15.9) | 1.28 | (0.91-1.81) | 612 | (4.1) | 170 | (27.8) | 1.23 | (1.05-1.43) | |
| CAN hierarchical | 0.34 | ||||||||||||
| None | 6,477 | (63.4) | 715 | (56.1) | 1.00 | (REF) | 10,678 | (71.0) | 2,392 | (67.3) | 1.00 | (REF) | |
| Any physical, sexual or emotional abuse with or without neglect | 3,347 | (32.8) | 499 | (39.1) | 1.28 | (1.13-1.44) | 3,579 | (23.8) | 952 | (26.8) | 1.27 | (1.17-1.37) | |
| Physical or emotional neglect alone | 387 | (3.8) | 61 | (4.8) | 1.35 | (1.04-1.76) | 783 | (5.2) | 208 | (5.9) | 1.11 | (0.96-1.28) | |
| Combined adult IVA and CAN hierarchical | 0.19 | ||||||||||||
| Neither adult IVA nor CAN | 4,916 | (48.1) | 532 | (41.7) | 1.00 | (REF) | 8,309 | (55.2) | 1,768 | (49.8) | 1.00 | (REF) | |
| Adult IVA and CAN | 2,366 | (23.2) | 389 | (30.5) | 1.37 | (1.19-1.58) | 2,744 | (18.2) | 753 | (21.2) | 1.34 | (1.23-1.46) | |
| CAN alone; No Adult IVA | 1,368 | (13.4) | 171 | (13.4) | 1.18 | (0.99-1.40) | 1,618 | (10.8) | 407 | (11.4) | 1.23 | (1.11-1.37) | |
| Adult IVA; No CAN | 1,561 | (15.3) | 183 | (14.4) | 1.05 | (0.89-1.25) | 2,369 | (15.8) | 624 | (17.6) | 1.23 | (1.12-1.35) | |
Hazard Ratios adjusted for sex (male, female), income group (<$25,000, $25,000-$49,999, ≥$50,000, missing) and age group (n=8) in 5-year intervals between 40-79.
P-value for interaction.
Boldface indicates statistical significance (p<0.05).
aHR = Adjusted hazard ratio
CAN = Childhood abuse or neglect
CI = Confidence interval
IVA = Interpersonal Violence or Abuse
REF = Referent
SCCS = Southern Community Cohort Study
Table 4.
Lifetime Interpersonal Violence or Abuse (IVA) and Adult-Onset Diabetes Rates by Sex (SCCS Participants)
| Female SCCS Participants | Male SCCS Participants | Likelihood ratio test for sex | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Lifetime IVA | Sample | Diabetes | aHRa | (95% CI) | Sample | Diabetes | aHRa | (95% CI) | |||||
| N | (Column %) | n | (row %) | N | (Column %) | n | (row %) | P-valueb | |||||
| All | 16,357 | 8,894 | |||||||||||
| Adult physical or psychological violence or abuse (IVA) | |||||||||||||
| No IVA (Referent, for all adult violence) | 9,434 | (57.7) | 1,764 | (18.7) | 1.00 | (REF) | 6,777 | (76.2) | 1,114 | (16.4) | 1.00 | (REF) | |
| Yes | 6,923 | (42.3) | 1,495 | (21.6) | 1.22 | (1.13-1.30) | 2,117 | (23.8) | 454 | (21.4) | 1.28 | (1.15-1.43) | 0.98 |
| By form | |||||||||||||
| Physical | 5,970 | (36.5) | 1,273 | (21.3) | 1.24 | (1.15-1.34) | 1,822 | (26.9) | 377 | (20.7) | 1.25 | (1.11-1.41) | 0.60 |
| Gun | 2,623 | (16.0) | 572 | (26.8) | 1.19 | (1.08-1.31) | 674 | (7.6) | 161 | (23.9) | 1.32 | (1.12-1.56) | 0.90 |
| Psychological | 5,238 | (32.0) | 1,170 | (22.3) | 1.24 | (1.15-1.34) | 1,112 | (12.5) | 245 | (22.0) | 1.34 | (1.16-1.54) | 0.94 |
| Adult violence hierarchical | 0.57 | ||||||||||||
| Physical or Gun | 6,256 | (38.2) | 1,335 | (21.3) | 1.22 | (1.13-1.31) | 1,958 | (22.0) | 410 | (20.9) | 1.26 | (1.12-1.41) | |
| Psychological alone | 667 | (4.1) | 160 | (24.0) | 1.18 | (1.01-1.39) | 159 | (1.8) | 44 | (27.7) | 1.52 | (1.13-2.06) | |
| CAN hierarchical | 0.33 | ||||||||||||
| None | 10,654 | (65.1) | 2,019 | (62.0) | 1.00 | (REF) | 6,501 | (73.1) | 1,088 | (69.4) | 1.00 | (REF) | |
| Any physical, sexual or emotional abuse with or without neglect | 4,937 | (30.2) | 1,057 | (32.4) | 1.27 | (1.18-1.37) | 1,989 | (22.4) | 394 | (25.1) | 1.27 | (1.13-1.42) | |
| Physical or emotional neglect alone | 766 | (4.7) | 183 | (5.6) | 1.21 | (1.04-1.40) | 404 | (4.5) | 86 | (5.5) | 1.08 | (0.87-1.35) | |
| Combined adult IVA and CAN hierarchical | 0.43 | ||||||||||||
| Neither adult IVA nor CAN | 7,649 | (46.8) | 1,409 | (43.2) | 1.00 | (REF) | 5,576 | (62.7) | 891 | (56.8) | 1.00 | (REF) | |
| Adult IVA and CAN | 3,918 | (23.9) | 885 | (27.2) | 1.36 | (1.24-1.48) | 1,192 | (13.4) | 257 | (16.4) | 1.35 | (1.17-1.56) | |
| CAN alone; No Adult IVA | 1,785 | (10.9) | 355 | (10.9) | 1.22 | (1.09-1.38) | 1,201 | (13.5) | 223 | (14.2) | 1.22 | (1.05-1.41) | |
| Adult IVA; No CAN | 3,005 | (18.4) | 610 | (18.7) | 1.15 | (1.05-1.27) | 925 | (10.4) | 197 | (12.6) | 1.30 | (1.12-1.52) | |
Hazard Ratios adjusted for sex (male, female), income group (<$25,000, $25,000-$49,999, ≥$50,000, missing) and age group (n=8) in 5 year intervals between 40-79.
P-value for interaction.
Boldface indicates statistical significance (p<0.05).
aHR = Adjusted hazard ratio
CI = Confidence interval
IVA = Interpersonal Violence or Abuse
REF = Referent
SCCS = Southern Community Cohort Study
Results
As anticipated, the majority of this lower income cohort of southerners (n=25,251) was female, Black, less than age 55, had incomes less than $15,000, had ≤high school education, and were currently married (Table 1). Hypertension and depressive symptoms were the most commonly reported chronic conditions. In this cohort analysis, 19.1% reported being diagnosed with diabetes in the period between the baseline and second follow-up survey. Demographic factors significantly associated with an increased incidence of adult-onset diabetes (p<0.0001) included being female, Black, less than age 55, having lower income and education, not being currently married, and having had a myocardial infarction, stroke, hypertension or depressive symptoms (Table 1).
Table 1.
Demographic Attributes of the SCCS Sample and Adult-onset Diabetes Incidence Rate
| Attribute | Sample | Diabetes | X2 | df | p-value | ||
|---|---|---|---|---|---|---|---|
| N | Column %) | N | (Row %) | ||||
| All | 25,251 | (100.0) | 4,827 | (19.1) | |||
| Sex | 19.61 | 1 | <0.0001 | ||||
| Male | 8,894 | (35.2) | 1,568 | (17.6) | |||
| Female | 16,357 | (64.8) | 3,259 | (19.9) | |||
| Race (from existing SCCS categories) | 487.31 | 1 | <0.0001 | ||||
| Black (African American) | 15,040 | (59.6) | 3,552 | (23.6) | |||
| White (Caucasian) | 10,211 | (40.4) | 1,275 | (12.5) | |||
| Age | 119.53 | 7 | <0.0001 | ||||
| 40-44 | 5,074 | (20.1) | 981 | (19.3) | |||
| 45-49 | 5,812 | (23.0) | 1,222 | (21.0) | |||
| 50-54 | 5,105 | (20.2) | 1,089 | (21.3) | |||
| 55-59 | 4,022 | (15.9) | 733 | (18.2) | |||
| 60-64 | 2,708 | (10.7) | 495 | (18.3) | |||
| 65-69 | 1,209 | (6.0) | 194 | (16.0) | |||
| 70-74 | 703 | (2.8) | 91 | (12.9) | |||
| 75-79 | 318 | (1.3) | 22 | (6.9) | |||
| Income | 264.48 | 4 | <0.0001 | ||||
| <$15,000 | 10,228 | (40.5) | 2,299 | (22.5) | |||
| $15,000-24,999 | 4,669 | (18.5) | 975 | (20.9) | |||
| $25,000-49,999 | 4,318 | (17.1) | 718 | (16.6) | |||
| $50,000+ | 4,521 | (17.9) | 528 | (11.7) | |||
| Missing | 1,515 | (6.0) | 307 | (20.3) | |||
| Education | 149.46 | 3 | <0.0001 | ||||
| <High school | 4,754 | (18.8) | 1,138 | (23.9) | |||
| High school | 7,737 | (30.6) | 1,572 | (20.3) | |||
| High school+ | 12,462 | (49.4) | 2,039 | (16.4) | |||
| Missing | 298 | (1.2) | 78 | (26.2) | |||
| Marital status | 67.21 | 4 | <0.0001 | ||||
| Married | 10,112 | (40.1) | 1,684 | (16.7) | |||
| Separated or divorced | 7,212 | (28.6) | 1,486 | (20.6) | |||
| Widowed | 3,306 | (13.1) | 682 | (20.6) | |||
| Single | 4,281 | (17.0) | 898 | (21.0) | |||
| Missing | 340 | (1.4) | 77 | (22.6) | |||
| Other conditions | |||||||
| Myocardial Infarction | 2,260 | (9.0) | 655 | (29.0) | 177.62 | 2 | <0.0001 |
| Cancer | 4,536 | (18.0) | 833 | (18.4) | 3.91 | 2 | NS |
| Stroke | 1,689 | (6.7) | 453 | (26.8) | 88.47 | 2 | <0.0001 |
| Hypertension | 11,633 | (46.1) | 2,834 | (24.4) | 417.06 | 2 | <0.0001 |
| Depression (CESD) | 8,143 | (32.3) | 1,797 | (22.1) | 77.63 | 2 | <0.0001 |
P-value comparing diabetics and non-diabetics.
Excluded those missing on 1) races other than White and Black,2) diabetes at baseline or follow-up, and adult or childhood adverse experiences except a parent being in prison or divorced.
Excluded from analyses those with a report age at diabetes diagnosis<30 and a prevalent diabetes diagnosis at baseline.
Boldface indicates statistical significance (p<0.05).
CESD = Center for Epidemiological Studies Depression
NS = Not significant
SCCS = Southern Community Cohort Study
After adjusting for confounders, both adult IVA and CAN were associated with an increased incidence of adult-onset diabetes (Table 2). Adult IVA was experienced by 35.8% of the SCCS participants. Those experiencing adult IVA had a 23% increased risk of diabetes (adjusted hazard ratio [aHR]=1.23; 95% confidence interval [CI]: 1.16-1.30). This statistically significant increased risk of diabetes associated with adult IVA was observed for each form of violence or abuse. CAN was experienced by 32.1% of the SCCS participants. Participants who experienced abuse with or without neglect had a 26% increased risk of diabetes (aHR=1.26; 95% CI: 1.19-1.35), while participants who experienced neglect alone had a smaller increase in diabetes risk (aHR=1.15; 95% CI: 1.02-1.30).
Table 2.
Lifetime Interpersonal Violence or Abuse (IVA) and Adult-onset Diabetes Rates (SCCS Participants)
| Lifetime IVA | Sample | Diabetes | aHRa | (95% CI) | ||
|---|---|---|---|---|---|---|
| N | (Column %) | N | (Row %) | |||
| All | 25,251 | 4,827 | ||||
| Adult physical or psychological violence or abuse (IVA) | ||||||
| No IVA (Referent, for all adult violence) | 16,211 | (64.2) | 2,878 | (17.8) | 1.00 | (REF) |
| Yes | 9,040 | (35.8) | 1,949 | (21.6) | 1.23 | (1.16-1.30) |
| By form | ||||||
| Physical | 7,792 | (30.9) | 1,650 | (21.2) | 1.24 | (1.17-1.32) |
| Gun | 3,297 | (13.1) | 733 | (22.2) | 1.22 | (1.13-1.33) |
| Psychological | 6,350 | (25.1) | 1,415 | (22.3) | 1.26 | (1.18-1.34) |
| Adult violence hierarchical | ||||||
| Physical violence or gun | 8,214 | (32.5) | 1,745 | (21.2) | 1.23 | (1.15-1.30) |
| Psychological alone | 826 | (3.3) | 204 | (24.7) | 1.24 | (1.08-1.44) |
| CAN hierarchical | ||||||
| None | 17,155 | (67.9) | 3,107 | (64.4) | 1.00 | (REF) |
| Any physical, sexual or emotional abuse with or without neglect | 6,926 | (27.4) | 1,451 | (30.0) | 1.26 | (1.19-1.35) |
| Physical or emotional neglect alone | 1,170 | (4.6) | 269 | (5.6) | 1.15 | (1.02-1.30) |
| Combined adult IVA and CAN hierarchical | ||||||
| Neither adult IVA nor CAN | 13,225 | (52.4) | 2,300 | (47.6) | 1.00 | (REF) |
| Adult IVA and CAN | 5,110 | (20.2) | 1,142 | (23.7) | 1.35 | (1.26-1.45) |
| CAN alone; No Adult IVA | 2,986 | (11.8) | 578 | (12.0) | 1.21 | (1.11-1.33) |
| Adult IVA; No CAN | 3,930 | (15.6) | 807 | (16.7) | 1.18 | (1.09-1.28) |
Hazard Ratios adjusted for sex (male, female), race (White or Black), income group (<$25,000, $25,000-$49,999, ≥$50,000, missing) and age group (n=8) in 5-year intervals between 40-79.
Boldface indicates statistical significance (p<0.05).
aHR = Adjusted hazard ratio
CAN = Childhood abuse or neglect
CI = Confidence interval
IVA = Interpersonal Violence or Abuse
REF = Referent
SCCS = Southern Community Cohort Study
When combining adult IVA and CAN to characterize lifetime IVA, those experiencing both adult IVA and CAN (20.2%) had the highest risk of diabetes (aHR=1.35; 95% CI: 1.26-1.45) relative to those experiencing no adult IVA or CAN (52.4%) (Table 2). Those experiencing CAN yet no adult IVA were 21% more likely to develop diabetes while those experiencing adult IVA yet no CAN were 18% more likely to develop diabetes.
Both adult IVA and CAN were associated with an increased risk of diabetes among Black and White participants (Table 3). When compared with participants who did not experience adult IVA or CAN (approximately 52%), adult IVA and CAN were associated with a 37% (95% CI: 1.19-1.58) increased risk of diabetes among Whites and a 34% (95% CI: 1.23-1.46) increase among Blacks. CAN but no IVA and adult IVA with no CAN were associated with significantly increased risks of diabetes among Black yet not among White participants. There was no evidence of effect modification of lifetime IVA on diabetes by race (p-value for interaction=0.19).
Focusing on the analyses combining adult IVA and CAN using a hierarchical comparison, no differences in diabetes risk were observed by sex for the grouping of adult IVA and CAN (aHR=1.36 for females and 1.35 for males, Table 4). A similar pattern was noted for CAN, yet no adult IVA (aHR=1.22 for both sexes). Adult IVA, but no CAN, was associated with an increased risk of diabetes among both sexes, but the magnitude of the associations appeared greater among males (aHR=1.30; 95% CI: 1.12-1.52) than females (aHR=1.15; 95% CI: 1.05-1.27). There was no evidence of effect modification of lifetime IVA and diabetes by sex (p for interaction=0.43).
Discussion
In this study, both adult IVA and CAN were associated with a significantly increased risk of diabetes in this large, lower-income cohort. As hypothesized, this pattern held for both Black and White and female and male SCCS participants. The latter finding contrasts with the hypothesized greater impact of adult IVA and CAN on diabetes risk for female relative to male participants.
Findings from this cohort analysis were most comparable to analyses from a large cohort of male and female health care plan participants.14 Significant associations were reported between diabetes and adult physical/psychological IVA (HR=1.68, 95% CI 1.14-2.48). While restricted to females, findings from the Nurses’ Health Study10, 12 are also comparable to those from the current SCCS cohort of females. Similar risk of diabetes was associated with adult physical IVA in the Nurses’ Health Study, (hazard ratio [HR]=1.18; 95% CI: 1.00-1.39), a stronger risk associated with severe psychological IVA (HR=1.78; 95% CI:1.21-2.61), yet no risk associated with sexual IVA (HR=1.08; 95% CI: 0.86-1.35).12 Among female nurses, diabetes risk increased with child physical abuse (aHR=1.26-1.54), sexual abuse (aHR=1.34-1.69)10 and were similar to the diabetes risk among females in the SCCS cohort for child physical, sexual or emotional abuse (aHR=1.33, 95% CI: 1.22-1.44).10
The authors are unaware of any prior studies that examined the impact of both adult and childhood violence or abuse on adult-onset diabetes. While CAN appeared to have a greater impact on diabetes risk than adult IVA in this study, the combined effect of both adult IVA and CAN had the strongest association. There was little difference in this association when stratified by race and when stratified by sex. The observation of a similar risk of diabetes associated with lifetime IVA by sex could be explained by the ability to investigate the temporal sequencing of lifetime IVA and incidence of adult-onset diabetes. Previous cross-sectional studies8–9 conducted among females and males could not establish temporality, and the only other large cohort study of adult IVA and adult-onset diabetes, the Nurses’ Health Study,12 was restricted to females.
This research adds to the existing literature in the following ways. Using the SCCS allowed an evaluation of the combined effect of IVA experienced across childhood and adulthood. These data also provided specific measures of adult IVA and CAN. This lifetime assessment of IVA can serve as a cumulative indicator of stress potentially affecting allostatic load. The latter has been associated with chronic, stress-associated diseases, including hypertension, angina, and adult-onset diabetes.4–5 The ability to estimate risk of adult-onset diabetes in this large prospective cohort is a significant contribution to the existing literature. The combined evaluation of adult IVA and CAN on diabetes incidence furthers the understanding of the potential etiology of stress-inducing violence and abuse on diabetes risk. One small cohort study (n=125) of pre-diabetics that examined associations between stress reactivity (based on salivary cortisol response) found that self-reported stress was associated with cortisol response among participants from high-SES neighborhoods only and there was no significant difference in cortisol response by race.20 This racially diverse study population of similarly lower income males and females provided evidence of the generalizability of these findings across sex and race in a lower income population who are often missing from cohort studies.
Limitations
One limitation of this study was the abbreviated measure of adult IVA, which did not include adult sexual violence. As indicated previously, the Nurses’ Health Study showed no association between adult sexual IVA and diabetes,12 thus the effect of excluding sexual violence from the measure of IVA may have been minimal. Adult IVA and CAN exposures and diabetes outcomes were based on self-report and could have resulted in misclassification. However, for lifetime IVA including CAN, self-report was the only viable source. No other reliable data source exists to validate these experiences. Another limitation of this study is the potential misclassification of self-reported diabetes. Although a validation sub-study conducted within the SCCS found that 96% of self-reported diabetes could be confirmed through medical records or elevated HbA1c measurements,15 approximately one-fourth of diabetes in the U.S. is undiagnosed.1 Thus our results could be partially explained if persons who experienced violence were more likely than persons who did not experience violence to have undiagnosed diabetes. Lastly, data characterizing childhood SES were not available and could have confounded observed associations between IVA and adult-onset diabetes.
Confounding bias was unlikely to explain the observed findings given the similar demographics of the SCCS population, and the ability to assess and control for confounders in these data. The likelihood of selection bias due to loss to follow-up was small evidenced by the identical percentage of baseline and second follow-up survey participants with prevalent diabetes (21.0%).
Conclusions
Both adult IVA and CAN increased the risk of adult-onset diabetes in a dose-dependent pattern for males and females and by race. The current study fills a gap in the diabetes literature and provides additional evidence to support assessments for current and lifetime violence, abuse and trauma. Trauma-informed practices21 have the potential to identify and link individuals to mental health care services, which have been connected to reductions in trauma response disorders.22 Efforts to support families, build resilience during childhood and adulthood, and address the mental and physical health effects of violence are essential to reducing the impact of violence and subsequent risk of diabetes and other chronic diseases.
Acknowledgements
Research reported in this manuscript was supported by the National Cancer Institute of the National Institutes of Health under award number U01CA202979. SCCS data collection was performed by the Survey and Biospecimen Shared Resource which is supported in part by the Vanderbilt-Ingram Cancer Center (P30CA068485). MS was partially supported by the National Cancer Institute (U54CA163069). MC was partially supported by the National Center for Advancing Translational Sciences (KL2TR002245). LLB was partially supported by the National Institute on Aging (P30AI110537) and the National Institute on Minority Health and Health Disparities (U54MD007586). ALC was partially supported by the National Institute on Drug Abuse (K12DA035150). The authors have no relevant financial or non-financial interests to disclose. No financial disclosures have been reported by the authors of this paper.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Credit Author Statement
Maureen Sanderson: Conceptualization, Methodology, Formal analysis, Writing – Original draft preparation, Writing – Reviewing and Editing. Ann L. Coker: Conceptualization, Methodology, Writing – Original draft preparation, Writing – Reviewing and Editing. Mekeila Cook: Writing – Reviewing and Editing. Writing – Reviewing and Editing. L. Lauren Brown: Writing – Original draft preparation, Writing – Reviewing and Editing. Veronica Mallett: Writing – Original draft preparation, Writing – Reviewing and Editing.
References
- 1.National Diabetes Statistics Report, 2022. By the Numbers: Diabetes in America. Atlanta, GA: CDC. https://www.cdc.gov/diabetes/health-equity/diabetes-by-the-numbers.html Accessed September 11, 2022. [Google Scholar]
- 2.Harris ML, Oldmeadow C, Hure A, et al. Stress increases the risk of type 2 diabetes onset in women: a 12-year longitudinal study using causal modelling. PLOS One. 2017;12(2):e0172126. 10.1371/journal.pone.0172126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Pinna KL, Johnson DM, Delahanty DL. PTSD, comorbid depression, and the cortisol waking response in victims of intimate partner violence: preliminary evidence. Anxiety Stress Coping. 2014;27(3):253–269. 10.1080/10615806.2013.852185. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Geronimus AT, Hicken M, Keene D, Bound J. “Weathering” and age patterns of allostatic load scores among blacks and whites in the United States. Am J Public Health. 2006;96(5):826–833. 10.2105/AJPH.2004.060749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.McEwen BS, Seeman T. Protective and damaging effects of mediators of stress: elaborating and testing the concepts of allostasis and allostatic load. Ann N Y Acad Sci. 1999;896:30–47. 10.1111/j.1749-6632.1999.tb08103.x. [DOI] [PubMed] [Google Scholar]
- 6.Heim C, Newport DJ, Heit S, et al. Pituitary-adrenal and autonomic responses to stress in women after sexual and physical abuse in childhood. JAMA. 2000;284(5):592–597. 10.1001/jama.284.5.592. [DOI] [PubMed] [Google Scholar]
- 7.Weitzman A, Goosby BJ. Intimate partner violence, circulating glucose, and non-communicable disease: adding insult to injury? SSM Popul Health. 2020;13:100701. 10.1016/j.ssmph.2020.100701. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Gilbert LK, Breiding MJ, Merrick MT, et al. Childhood adversity and adult chronic disease: an update from ten states and the District of Columbia, 2010. Am J Prev Med. 2015;48(3):345–349. 10.1016/j.amepre.2014.09.006. [DOI] [PubMed] [Google Scholar]
- 9.Gilbert LK, Zhang X, Basile KC, Breiding M, Kresnow MJ. Intimate partner violence and health conditions among U.S. adults-National Intimate Partner Violence Survey, 2010-2012. J Interpers Violence. 2022:8862605221080147. 10.1177/08862605221080147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rich-Edwards JW, Spiegelman D, Lividoti Hibert EN, et al. Abuse in childhood and adolescence as a predictor of Type 2 diabetes in adult women. Am J Prev Med. 2010;39(6):529–536. 10.1016/j.amepre.2010.09.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Sanderson M, Mouton CP, Cook M, et al. Adverse childhood experiences and chronic disease risk in the Southern Community Cohort Study. J Health Care Poor Underserved. 2021;32(3):1384–1402. 10.1353/hpu.2021.0139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mason SM, Wright RJ, Hibert EN, et al. Intimate partner violence and incidence of type 2 diabetes in women. Diabetes Care. 2013;36(5):1159–65. 10.2337/dcl2-1082. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Stubbs A, Szoeke C. The effect of intimate partner violence on the physical health and health-related behaviors of women: a systematic review of the literature. Trauma Violence Abuse. 2021;1524838020985541. 10.1177/1524838020985541. [DOI] [PubMed] [Google Scholar]
- 14.Pantell MS, Prather AA, Downing JM, Gordon NP, Adler NE. Association of social and behavioral risk factors with earlier onset of adult hypertension and diabetes. JAMA Netw Open. 2019;2(5):e193933. 10.1001/iamanetworkopen.2019.3933. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Signorello LB, Hargreaves MK, Blot WJ. The Southern Community Cohort Study: investigating health disparities. J Health Care Poor Underserved. 2010;21(1Suppl):26–37. 10.1353/hpu.0.0245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Hargreaves MK, Mouton CP, Liu J, et al. Adverse childhood experiences and health care utilization in a low-income population. J Health Care Poor Underserved. 2019;30(2):749–767. 10.1353/hpu.2019.0054. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Smith SG, Chen J, Basile KC, et al. The National Intimate Partner and Sexual Violence Survey (NISVS): 2010-2012 State Report. Atlanta, GA: National Center for Injury Prevention and Control, Centers for Disease Control and Prevention. https://www.cdc.gov/violenceprevention/pdf/nisvs-statereportbook.pdf. Published April 2017. Accessed September 11,2022. [Google Scholar]
- 18.Follingstad D, Coker AL, Lee E, et al. Validity and psychometric properties of the measure of psychologically abusive behaviors among young women and women in distressed relationships. Violence Against Women. 2015;21(7):875–896. 10.1177/1077801215584070. [DOI] [PubMed] [Google Scholar]
- 19.Felitti VJ, Anda RF, Nordenberg D, et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults: the Adverse Childhood Experiences (ACE) Study. Am J Prev Med. 1998;14(4):245–258. 10.1016/S0749-3797(98)00017-8. [DOI] [PubMed] [Google Scholar]
- 20.Kalesnikava VA, Clarke PJ, Mukherjee B, Sen S, Mezuk B. Psychosocial stress and hypothalamic-pituitary-adrenal axis stress reactivity: variations by race and socioeconomic status among adults at risk of diabetes. Psychosom Med. 2022;84(7):813–821. 10.1097/PSY.0000000000001112. [DOI] [PubMed] [Google Scholar]
- 21.Substance Abuse and Mental Health Service Administration. Trauma-informed care in behavioral health services. Treatment Improvement Protocol (TIP) Series 57. HHS Publication No. (SMA) 13-4801. Rockville, MD: Substance Abuse and Mental Health Services Administration. https://store.samhsa.gov/sites/default/files/d7/priv/sma14-4816.pdf Published 2014. Accessed September 11, 2022. [Google Scholar]
- 22.Han H-R, Miller HN, Nkimbeng M, et al. Trauma informed interventions: a systematic review. PLoS One. 2021;16(6):e0252747. 10.1371/iournal.pone.0252747. [DOI] [PMC free article] [PubMed] [Google Scholar]