Introduction
Individuals exposed to childhood maltreatment (CM) are at increased risk for psychological and global functioning deficits (Adams, Reference Adams2006; Jonas et al. Reference Jonas, Bebbington, McManus, Meltzer, Jenkins, Kuipers, Cooper, King and Brugha2011). In adolescence, these individuals are at increased risk for conduct disorder, early onset alcohol and substance use, depression, and suicidality (Kilpatrick et al. Reference Kilpatrick, Acierno, Saunders, Resnick, Best and Schnurr2000; Sartor et al. Reference Sartor, Waldron, Duncan, Grant, McCutcheon, Nelson, Madden, Bucholz and Heath2013). Maladaptive consequences persist into adulthood, with survivors of childhood physical and sexual abuse exhibiting higher rates of adverse mental health problems such as suicidality, post-traumatic stress disorder (PTSD), major depressive disorder (MDD), and substance dependence (Duncan et al. Reference Duncan, Saunders, Kilpatrick, Hanson and Resnick1996; Kendler et al. Reference Kendler, Bulik, Silberg, Hettema, Myers and Prescott2000; Molnar et al. Reference Molnar, Buka and Kessler2001). Additionally, studies have consistently found that exposure to CM increases the risk of adult sexual victimization (Fergusson et al. Reference Fergusson, Horwood and Lynskey1997; Roodman & Clum, Reference Roodman and Clum2001; Nelson et al. Reference Nelson, Heath, Madden, Cooper, Dinwiddie, Bucholz, Glowinski, McLaughlin, Dunne, Statham and Martin2002; Desai et al. Reference Desai, Arias, Thompson and Basile2002; Filipas & Ullman, Reference Filipas and Ullman2006). As maladaptive outcomes associated with CM – substance misuse, PTSD, MDD – have also been found to increase the risk of adult sexual victimization, it is not surprising that survivors of CM are at increased risk for revictimization (Classen et al. Reference Classen, Palesh and Aggarwal2005). Furthermore, compared to either CM or adult sexual trauma only, sexual revictimization is associated with substance misuse and multiple psychiatric diagnoses including MDD, PTSD, schizophrenia, and bipolar disorder (Classen et al. Reference Classen, Palesh and Aggarwal2005; Walsh et al. Reference Walsh, Resnick, Danielson, McCauley, Saunders and Kilpatrick2014) highlighting the importance of considering multiple trauma exposures in the psychiatric setting.
Long-term difficulties and sexual revictimization are particularly prevalent in childhood sexual abuse (CSA) survivors. Some have posited that CSA, particularly in women, is one of the strongest predictors of subsequent sexual victimization (Siegel & Williams, Reference Siegel and Williams2003; Casey & Nurius, Reference Casey and Nurius2005). Estimates from community samples reveal that a history of CSA is associated with a 2–5 times increased risk of adult sexual victimization (Fleming et al. Reference Fleming, Mullen, Sibthorpe and Bammer1999; Desai et al. Reference Desai, Arias, Thompson and Basile2002) with prevalence estimates of sexual revictimization in survivors of CSA as high as 35% in women and 24–37% in men (Aosved et al. Reference Aosved, Long and Voller2011; Black et al. Reference Black, Basile, Breiding, Smith, Walters, Merrick and Stevens2011; Artime et al. Reference Artime, McCallum and Peterson2014). Furthermore, a large community investigation observed a higher CSA-associated revictimization risk in male compared to female CSA survivors (with respective odds ratios of 5.5 and 1.8; Desai et al. Reference Desai, Arias, Thompson and Basile2002). Similar findings were reported by Nelson and colleagues suggesting that male CSA survivors represent a high-risk population that deserves further examination (Nelson et al. Reference Nelson, Heath, Madden, Cooper, Dinwiddie, Bucholz, Glowinski, McLaughlin, Dunne, Statham and Martin2002).
Despite the extent of our understanding of risk factors associated with revictimization in women, there is a dearth of literature examining risk in men. When revictimization in men is examined, the population is often not representative, focusing rather on specific subgroups such as gay or undergraduate populations (Kalichman et al. Reference Kalichman, Benotsch, Rompa, Gore-Felton, Austin, Luke, DiFonzo, Buckles, Kyomugisha and Simpson2001; Aosved et al. Reference Aosved, Long and Voller2011). The scant findings in community samples have found a history of CSA in men is associated with increased likelihood of adult sexual victimization (Coxell et al. Reference Coxell, King, Mezey and Gordon1999; Desai et al. Reference Desai, Arias, Thompson and Basile2002; Nelson et al. Reference Nelson, Heath, Madden, Cooper, Dinwiddie, Bucholz, Glowinski, McLaughlin, Dunne, Statham and Martin2002). However, further research is needed to extend findings by comparing risk factors of sexual revictimization among males and females.
Most investigations into the association between CM and adult sexual victimization only investigate the impact of CSA and childhood physical abuse (CPA), neglecting to account for other commonly co-occurring psychopathology associated with childhood and adult trauma. Specifically, substance use disorders (SUDs), PTSD, and depression have all consistently been linked to CM, but are rarely investigated as covariates when predicting adult revictimization. One study investigating revictimization in a population of female sexual abuse victims reported PTSD symptoms and alcohol use mediated the relationship between CSA and adult sexual assault, suggesting CSA may only indirectly increase adult sexual victimization risk (Ullman et al. Reference Ullman, Najdowski and Filipas2009). The impact of SUDs and other psychopathology on the association between CSA and adult victimization has yet to be addressed in male populations.
Methodological limitations also exist in the current revictimization literature. When investigating the impact of CM on later victimization, trauma exposure is often considered dichotomously (identifying the presence or absence of trauma), not addressing information pertaining to the severity or type of trauma exposure (Bernstein et al. Reference Bernstein, Ahluvalia, Pogge and Handelsman1997; Bulik et al. Reference Bulik, Prescott and Kendler2001; Sartor et al. Reference Sartor, McCutcheon, Nelson, Duncan, Bucholz and Heath2012b ). CSA is the most common CM examined, but other childhood interpersonal trauma such as CPA and witnessing parental partner abuse (PPA) also predict adult sexual victimization (Desai et al. Reference Desai, Arias, Thompson and Basile2002). Furthermore, considering these childhood interpersonal trauma factors dimensionally would allow for more robust and informative measure of childhood trauma.
The present study aimed to overcome these limitations through considering the impact of childhood interpersonal trauma factors of CSA, CPA, and PPA [identified through confirmatory factor analysis; see S. Kristjansson et al. unpublished data), SUDs, PTSD, and MDD on adult sexual victimization in three independent community samples (all including men and women). Furthermore, sex disparities in the association among childhood trauma factors, psychopathology and adult sexual victimization will be explored in each sample.
Method
Participants and procedures
Detailed descriptions of the Childhood Trauma Study (CTS; Nelson et al. Reference Nelson, Lynskey, Heath, Madden and Martin2010; Sartor et al. Reference Sartor, Grant, Lynskey, McCutcheon, Waldron, Statham, Bucholz, Madden, Heath, Martin and Nelson2012a ), Comorbidity and Trauma Study (CATS; Conroy et al. Reference Conroy, Degenhardt, Mattick and Nelson2009; Shand et al. Reference Shand, Degenhardt, Slade and Nelson2011; Nelson et al. Reference Nelson, Lynskey, Heath, Wray, Agrawal, Shand, Henders, Wallace, Todorov, Schrage, Saccone, Madden, Degenhardt, Martin and Montgomery2013), and the Australian Twin-Family Study of Alcohol Use Disorders (OZ-ALC; Heath et al. Reference Heath, Whitfield, Martin, Pergadia, Goate, Lind, McEvoy, Schrage, Grant, Chou, Zhu, Henders, Medland, Gordon, Nelson, Agrawal, Nyholt, Bucholz, Madden and Montgomery2011) methods have been previously reported; summaries pertaining to the current investigation are provided below. Although each of the samples was ascertained on the basis of either history of CM or substance use, each also included control populations from similar demographic background with a wide range of substance use and trauma exposure.
Primary sample: CTS
Participants were drawn from a large Australian twin cohort (born 1964–1971) based on the twins’ responses to childhood trauma screening questions included in a semi-structured psychiatric diagnostic assessment (conducted between 1996 and 2000 via telephone (Heath et al. Reference Heath, Howells, Kirk, Madden, Bucholz, Nelson, Slutske, Statham and Martin2001). From this cohort, high-risk families were identified if one or both twins endorsed a screening question on CSA (5 total) or CPA (4 total) while control families included those in which no interviewed twin reported either form of abuse. Verbal consent, obtained pre-interview, was confirmed through signed consent form allowing use of interview data; procedures were approved by the Queensland Institute of Medical Research (QIMR) Ethics Committee and the Washington University School of Medicine (WUSM) Human Research Protection Office. Follow-up telephone interviews were completed between 2003 and 2008 by 3434 respondents from 524 high-risk and 373 control families; data from 27 interviewed respondents who did not provide post-interview consent were not analyzed. Thirty individuals from the original sample were excluded from the current analyses for missing adult victimization assessment. Data are reported from 2564 respondents from comparable high-risk and control families with childhood trauma assessment available, including 1514 twins (981 female, 64.8%) and 1050 non-twin siblings (618 female, 58.9%). The mean age at interview was 37.2 years (s.d. = 2.3) for twins and 40.6 years (s.d = 6.3) for siblings.
Replication 1 sample: CATS
Participants in a case-control examination of genetic and environmental factors contributing to liability for opioid dependence (CATS) included 1468 cases (577 female, 39.3%) recruited from opioid replacement therapy (ORT) clinics in the greater Sydney, Australia region and 513 controls (284 female. 55.4%) from the geographic areas in proximity to ORT clinics who denied having used opioids recreationally more than ten times (Conroy et al. Reference Conroy, Degenhardt, Mattick and Nelson2009; Shand et al. Reference Shand, Degenhardt, Slade and Nelson2011; Nelson et al. Reference Nelson, Lynskey, Heath, Wray, Agrawal, Shand, Henders, Wallace, Todorov, Schrage, Saccone, Madden, Degenhardt, Martin and Montgomery2013). Written informed consent was obtained from all participants as per the institutional review board (IRB) approvals obtained from the University of New South Wales, WUSM, the QIMR, and the ethics committees governing the participating clinics. The mean age at interview was 36.4 years (s.d. = 8.6) for cases and 34.7 years (s.d. = 10.6) for controls.
Replication 2 sample: OZ-ALC
The Alcoholism Research Center (ARC) Project 7 re-assessed the sample from a genome-wide association study of alcohol dependence and heaviness of drinking (OZ-ALC), which partially overlaps with the CTS study sample. The complete sample was ascertained through index cases identified via surveys of two large general-population Australian twin cohorts [cohort 1, born 1890–1964 (mostly 1940–1964), and cohort 2 born 1964–1971], the spouses/partners of cohort 1 twins, and an Australian population-representative sample that ascertained families containing ⩾5 full siblings. For the current sample, priority was given to individuals from families where childhood trauma was reported at prior assessment, and re-interviews were completed with 1882 participants. Verbal consent was obtained from all participants in compliance with IRB approvals obtained from WUSM and QIMR. When those who participated in the CTS and individuals with missing data were excluded (n = 345), a sample of 726 alcohol-dependent cases (270 female, 36.2%) and 811 controls (476 female, 58.7%) were retained. The current sample (N = 1537) includes 791 men and 746 women with a mean age of 53.0 (s.d. = 8.3).
Assessments
Childhood trauma history
A computer-assisted diagnostic interview used in each study included the modified Christchurch Trauma Assessment (Fergusson et al. Reference Fergusson, Horwood, Shannon and Lawton1989; Fergusson & Lynskey, Reference Fergusson and Lynskey1997). To provide more robust, continuous measures of trauma in each sample, factor analyses were performed which distilled Christchurch Trauma Assessment component items into three first order factors: CSA, CPA, and PPA experiences prior to the age of 18. The assessment's detailed trauma questions included: CPA [14 items, asked separately about mother, father, and other adult members of the household, covering forms of severe physical punishment (including whether each endorsed form occurred occasionally or frequently) and consequences], CSA (17 items asking about non-contact and contact sexual abuse and additional questions to determine the frequency of abuse occurring during various age periods), and PPA (19 total items including questions about emotional and physical abuse of partner by each parent and additional items querying whether police visited the home, respondent and mother left the home, or if the respondent avoided spending time in the home). This assessment was further modified in the CATS assessment by combining maternal and paternal CPA assessment and combining some of the CSA screening questions (reducing the CSA total from 17 to 11 items).
Adult sexual victimization
Adult sexual victimization was defined as responding affirmatively to any of the following: ‘Since you turned 18, has anyone attempted to involve you in any sexual behavior by the use of (a) physical violence, (b) threats of physical violence?’ or ‘Since you turned 18 has anyone had sexual intercourse with you when you did not want to have intercourse and you made it clear that you did not want to have sex?’.
Psychopathology and substance use
Lifetime diagnoses of nicotine, alcohol, and illicit drug dependence, PTSD and MDD, were assessed via a modified version of the Semi-Structured Assessment for the Genetics of Alcoholism (SSAGA; Bucholz et al. Reference Bucholz, Cadoret, Cloninger, Dinwiddie, Hesselbrock, Nurnberger, Reich, Schmidt and Schuckit1994). All three studies queried participants using computer-assisted interviews. CTS and OZ-ALC interviews were performed via telephone; CATS interviews were done face-to-face. As all OZ-ALC participants had been assessed previously, lifetime diagnoses of MDD, licit and illicit drug dependence, and PTSD were obtained from previous interview data for some individuals in the current sample. When multiple assessments of an outcome were available on a participant, an affirmative response at any assessment was coded positive. To decrease respondent burden, sections of the interview that were identical in CTS and OZ-ALC (e.g. the depression section) were not re-administered to CTS participants who had been interviewed in OZ-ALC.
Socioeconomic and familial risk factors
A question on respondent's educational attainment that was asked in all three studies was used as a proxy for socioeconomic class (SEC). A binary variable was coded to include individuals who reported ⩽10 years education without an apprenticeship or diploma (combining the two lowest response categories). A variable for familial risk was coded to reflect not having been raised by both parents until the age of 16. Maternal and paternal alcohol misuse were independently assessed by querying: (a) Has drinking ever caused your biological mother/father to have problems with health, family, job or police, or other problems? and (b) Have you ever felt that your biological mother/father was an excessive drinker? Endorsement of either question for either parent was coded as a measure of parental alcohol misuse.
Statistical analyses
Prevalence rates of CSA, psychopathology, and adult sexual victimization were calculated for each sample. Separate continuous measures of CSA, CPA, and PPA were used in the current study. These CSA, CPA, and PPA factors were generated via confirmatory factor analyses performed in MPlus (Muthen & Muthen, Reference Muthen and Muthen1998) with each factor created in each sample from the construct-specific component items. Means were centered at zero for all factors (mean = 0, s.d. = 1), with the exception of CSA factor in women (S. Kristjansson et al. unpublished data); therefore odds ratios estimate the change in the risk associated with a standard deviation difference of 1 in the factor. Analyses were performed using SAS v. 9.4 (SAS Institute Inc., USA) for each study separately.
Logistic regression analyses predicting adult sexual victimization in each sample and stratified by sex were used to generate univariate and multivariate odds ratios for childhood trauma factors. The impact of the CSA factor on revictimization adjusting for possible confounding familial risks – parental presence, SEC, and parental alcohol misuse – and SUDs, PTSD, and MDD was studied with sex included as a covariate (with post-hoc analyses for the interaction of the CSA factor and sex) and in separate analyses stratified by sex. Family clustering was corrected through the Taylor series variance estimation method – which estimates the covariance matrix of the regression coefficients in clustered data, thus reducing sample bias – implemented in Proc Survey Logistic. To formally test for combined effect across studies, effect sizes for (a) the CSA factor and (b) CSA factor CM sex interaction on revictimization were separately meta-analyzed using an inverse-variance weighted model with fixed and random effects in the R package (R Foundation, Austria) for general meta-analysis.
Results
Primary sample: CTS
Of the CTS sample, lifetime prevalence of psychopathology was as follows: alcohol dependence (21.9%), nicotine dependence (31.1%), cannabis dependence (5.9%), non-cannabis illicit drug dependence (4.3%), MDD (33.7%), and PTSD (11.4%; see Table 1 for prevalence by sex). Thirty-four percent (19.3% of men and 43.1% of women) endorsed any CSA and 29.2% (16.8% of men and 36.6% of women) reported contact CSA. Eight percent of the total sample endorsed adult sexual victimization and women were more likely to experience adult sexual victimization [odds ratio (OR) 4.70, 95% confidence interval (CI) 3.09–7.14].
Table 1. Prevalence rates for adult victimization, substance dependence, and psychopathology by sample and sex
CTS, Childhood Trauma Study – Men (n = 965), Women (n = 1599); CATS, Comorbidity and Trauma Study – Men (n = 1120), Women (n = 861); OZ-ALC, Australian Twin-Family Study of Alcohol Use Disorders – Men (n = 791), Women (n = 746).
The CSA, CPA, and PPA factors were independently associated with significant risk for adult sexual victimization in both men and women (Table 2). When childhood trauma factors were simultaneously considered, only CSA and CPA remained significantly associated. With each unit increase in the CSA factor score, the odds of adult sexual victimization increased two times in men; in women, the corresponding increase in risk was 54% for each unit increase in CSA or CPA respectively.
Table 2. Logistic regression analyses examining risk a of adult sexual victimization associated with childhood trauma factors
aOR, Adjusted odds ratio; CI, confidence interval; CTS, Childhood Trauma Study; CATS, Comorbidity and Trauma Study; OZ-ALC, Australian Twin-Family Study of Alcohol Use Disorders; CSA, childhood sexual abuse factor; CPA, childhood physical abuse factor; PPA, parental partner abuse factor.
a Odds ratios estimate risk associated with a 1 s.d. change in the specific childhood trauma factor score.
b Adjusted for sex.
Multivariate model includes all trauma factors simultaneously.
Further analyses were conducted to examine the association of CSA with revictimization after adjusting simultaneously for co-occurring psychopathology: alcohol, nicotine, cannabis, and other illicit drug dependence, PTSD, and MDD. In the multivariate model, CSA score remained significantly associated with sexual victimization risk (Table 3). Sex was reported to be significantly associated with revictimization [adjusted odds ratio (aOR) 4.81, 95% CI 2.91–7.97]; however, post-hoc analyses did not reveal a significant interaction between CSA score and sex (aOR 0.82, 95% CI 0.57–1.18) in this sample. Increased risk of victimization was associated with CSA factor score in men (aOR 2.13, 95% CI 1.40–3.25) and women (aOR 1.32, 95% CI 1.09–1.59). MDD was associated with increased risk of victimization in both men and women; however, nicotine dependence and PTSD were only associated with adult sexual victimization in women.
Table 3. Multivariate regression analyses examining risk a of adult sexual victimization associated with childhood sexual abuse, substance dependence, depression and PTSD
Values given are adjusted odds ratios and 95% confidence intervals.
CTS, Childhood Trauma Study; CATS, Comorbidity and Trauma Study; OZ-ALC, Australian Twin-Family Study of Alcohol Use Disorders; CSA, childhood sexual abuse; AD, alcohol dependence; ND, Nicotine dependence; CD, cannabis dependence; IDD, non-cannabis illicit drug dependence; MDD, major depressive disorder; PTSD, post-traumatic stress disorder.
a All disorders included in the model simultaneously and all models adjusted for parental presence, socio-economic class, and parental alcohol misuse; odds ratios estimate risk associated with a 1 s.d. change in childhood trauma factor.
bAdjusted for sex.
Replication sample 1: CATS
As the CATS oversampled participants with opioid dependence, high rates of lifetime cannabis dependence (48.3%) and non-cannabis illicit drug dependence (78.7%) were reported (Table 1). A lifetime history of alcohol dependence was found in 36.7%, while 59.3% of individuals met criteria for lifetime nicotine dependence. Twenty-eight percent of individuals endorsed adult sexual victimization, over half endorsed MDD (57.4%), and a third met criteria for PTSD lifetime (32.9%). Forty-four percent (29.2% of men and 64.7% of women) endorsed any CSA and 39.8% (28.1% of men and 56.9% of women) reported contact CSA. Women were found to be at almost five times increased risk for adult sexual victimization (OR 4.84, 95% CI 3.91–5.99).
Univariate regression analyses revealed CSA, CPA, or PPA scores were independently associated with risk of adult sexual victimization in the total sample and when stratified by sex (Table 2). When considering childhood trauma factors in a multivariate model predicting adult victimization, CSA and CPA were significant predictors while PPA no longer contributed to risk. Both men and women with a history of CSA or CPA were revealed to be at increased risk of adult victimization when stratified by sex.
Multivariate analyses examining the unique risk associated with CSA for adult victimization while also considering substance dependence and psychopathological factors, revealed a significant contribution of CSA to adult victimization (aOR 1.63, 95% CI 1.45–2.84) (Table 3). Post-hoc analyses revealed significant interaction between CSA score and sex (aOR 0.72, 95% CI 0.57–0.92); thus men with a history of CSA were more likely to experience revictimization. Increased risk for adult victimization associated with CSA was elevated in men (aOR 2.14, 95% CI 1.73–2.65) and women (aOR 1.45, 95% CI 1.24–1.69). Nicotine dependence increased risk of victimization in both men and women; however, MDD was associated with increased risk of victimization in men only, while alcohol dependence, non-cannabis illicit substance dependence, and PTSD were only associated with adult victimization in women.
Replication sample 2: OZ-ALC
A large percentage of individuals in the OZ-ALC sample met criteria for a lifetime history of alcohol dependence (47.2%) or nicotine dependence (39.4%, Table 1). Lifetime diagnosis of cannabis dependence and other illicit drug dependence was found to be 6.1% and 2.4%, respectively. Of the total population, 7.2% had a history of adult sexual victimization, 32.8% endorsed a lifetime MDD, and 9.2% had a history of PTSD. Thirty-two percent (22.1% of men and 41.6% of women) endorsed any CSA and 25.6% (18.1% of men and 33.5% of women) reported contact CSA. Women were nine times more likely to experience adult sexual victimization (OR 8.94, 95% CI 4.94–16.19) than men.
CSA, CPA, and PPA were independently associated with significant risk for adult (re)victimization in the total sample when controlling for sex and in both men and women when considered separately (Table 2). When childhood trauma factors were considered in a multivariate model, CSA and CPA factors remained significant predictors in the complete sample. When analyses were conducted within sexes, the CSA factor was associated with increased risk of adult victimization in both sexes, but increased adult victimization risk was associated with the PPA factor only in men.
Multivariate analyses examining the unique association of CSA, substance dependence, and other psychopathology with adult victimization revealed CSA score remained associated with increased risk of victimization in the total sample and when men and women were considered separately (Table 3). Post-hoc analyses in this sample did not revealed significant interaction between CSA and sex (aOR 0.77, 95% CI 0.54–1.09). MDD was associated with increased risk of adult victimization in women only.
Meta-analysis
For the relationship between the CSA factor and revictimization, adjusting for sex and related psychopathology, a significant fixed effect was observed (OR 1.57, 95% CI 1.44–1.72; Fig. 1) indicating an overall effect for the CSA factor on revictimization. Likewise, across all samples, the fixed effect for CSA factor × sex interaction was significant (OR 0.75, 95% CI 0.64–0.90; Fig. 1) above and beyond sex alone and related psychopathology. There was no evidence for heterogeneity of effect sizes across samples (p > 0.10).
Fig. 1. Meta-analysis of (a) association of CSA factor and (b) CSA factor × sex interaction with revictimization across all three samples.
Discussion
This is the first investigation to examine and replicate the associated risk of dimensionally measured childhood trauma for adult sexual victimization in both men and women. We sought to assess the unique association of childhood trauma factors (CSA, CPA, PPA), SUDs (alcohol, nicotine, cannabis, and illicit other), PTSD, and MDD with the risk of adult sexual victimization for women and men. Across all three studies, multivariate regression analyses revealed that, of the childhood trauma factors, CSA and CPA exposure consistently increased risk for adult sexual victimization. When considering the impact of childhood trauma in men and women separately, only the CSA factor was associated with adult victimization in all three studies. Furthermore, across all studies, we found every one standard deviation increment increase in CSA factor score conferred an increased risk for adult sexual victimization in fully adjusted models.
Our findings across multiple populations are consistent with previous research and provide strong evidence for CSA as a significant risk factor for adult victimization in both men and women (Roodman & Clum, Reference Roodman and Clum2001; Classen et al. Reference Classen, Palesh and Aggarwal2005). Although CSA factor scores were associated with adult sexual victimization at a higher magnitude in men, greater overall risk for revictimization was observed for women. Two previous studies investigated the association between CSA (binary present/absent) and adult sexual victimization in both sexes reporting similar associations between CSA and adult sexual revictimization. In a partially overlapping sample, Nelson and colleagues found, in survival analyses, the hazard ratios for adult rape (age ⩾18) associated with a history of CSA were 3.6 in women and 26.7 in men (Nelson et al. Reference Nelson, Heath, Madden, Cooper, Dinwiddie, Bucholz, Glowinski, McLaughlin, Dunne, Statham and Martin2002). Furthermore, Desai et al. (Reference Desai, Arias, Thompson and Basile2002) reported CSA-associated risk of adult sexual victimization of 1.8 in women and 5.5 in men. Our results fully support and build upon those results as CSA score was found to increase risk for adult sexual victimization even after adjusting for co-occurring psychopathology and substance dependence. In addition to CSA contributing to increased risk for adult sexual victimization, Desai et al. (Reference Desai, Arias, Thompson and Basile2002) also found increased risk for adult sexual assault associated with physical abuse in both men and women. These findings were only partially supported in the current investigation; although CPA was associated with adult sexual victimization in univariate analyses, multivariate associations taking into account CSA and PPA revealed CPA was only associated with adult victimization in both men and women in the CATS sample. In the CTS sample multivariate regression, CPA was only associated with adult victimization in women and no significant association between CPA and adult victimization was revealed in the OZ-ALC sample.
With regards to the association of adult sexual victimization and psychopathology, findings are consistent with previous prospective research in women showing PTSD and substance abuse increase risk for revictimization (Messman-Moore et al. Reference Messman-Moore, Ward and Brown2009). In the CTS sample multivariate analyses, adjusting for psychiatric and substance co-morbidity, significant risk for adult sexual victimization was observed in women with alcohol and nicotine dependence, major depression, and PTSD. In men from this sample, MDD was the only disorder associated with significant risk for adult victimization. Men in the CATS with a history of MDD or nicotine dependence also conferred increased risk for adult victimization. Similar risk was found in the CATS women with alcohol, nicotine, and illicit drug dependence. In the OZ-ALC study, increased risk was only associated with MDD in women. In addition, the CSA factor score was associated with a significant increased risk of adult victimization in both men and women. These findings support discrete risk associated with psychopathology in men and women for adult sexual victimization. However, the directionality of these relationships remain unclear (i.e. whether substance dependence has a larger impact on revictimization in women than men or revictimization takes a larger toll on women who then turn to substances for self-medication).
A main strength of this study is consistency of results across different samples. By design, CTS oversampled individuals with childhood trauma, CATS oversampled individuals at high-risk of substance dependence, and OZ-ALC oversampled individuals at high-risk of alcohol dependence. Despite differences in sample size, population characteristics, and ascertainment strategies, meta-analysis overall revealed CSA was associated with increased risk of adult sexual victimization across the three studies. An additional strength is the inclusion of both men and women as very few investigations are able to contrast the impact of sexual abuse by sex. Male survivors of childhood and adult sexual victimization are notoriously underrepresented in this field. Of studies that have examined sexual revictimization in men, studies were limited by only including men (disallowing for sex comparisons) and mostly men who have sex with men (Classen et al. Reference Classen, Palesh and Aggarwal2005) which limits the generalizability of findings to community samples. Our meta-analysis for the interaction of the CSA factor and sex revealed men with a history of CSA are at increased risk for experiencing adult sexual revictimization. Additionally, our findings supports previous observations in a study of college men where revictimized men report significantly higher PTSD and depression than non-victimized men (Aosved et al. Reference Aosved, Long and Voller2011).
The current findings also have clinical implications. Those who experience childhood trauma – and more severe trauma – are more likely to experience revictimization which has been found to increase risk and be an associated outcome of psychopathology including PTSD, MDD, and SUDs (Breslau, Reference Breslau2002; Classen et al. Reference Classen, Palesh and Aggarwal2005; Collishaw et al. Reference Collishaw, Pickles, Messer, Rutter, Shearer and Maughan2007; Cinamon et al. Reference Cinamon, Muller and Rosenkranz2014; Walsh et al. Reference Walsh, Resnick, Danielson, McCauley, Saunders and Kilpatrick2014). Therefore, our findings point to the importance of considering CSA (and severity) as a risk factor and adult victimization as both a risk factor and possible outcome associated with psychopathology in community and psychiatric populations.
The current study is not without limitations. All samples included were Australian cohorts and results in other nations may differ. Self-report measures were utilized in all of the studies; therefore recall bias may impact our observations. Trauma exposure is particularly of concern as it is especially vulnerable to reporting biases, although research has shown that differences in prevalence of CM do not have a large impact on the strength of the association observed (Fergusson et al. Reference Fergusson, Horwood and Woodward2000). This study utilized logistic regression methodology and did not take into account the causal relationship of psychopathology to adult sexual victimization. By definition, CSA always predates adult victimization and is associated with downstream risk for the psychiatric disorders we included as covariates. Therefore, the reduction in CSA-associated risk with inclusion of these disorders in the models thus provides a conservative estimate of risk mediated via this route (i.e. psychopathology). Since additional trauma exposure (i.e. revictimization) further increases risk for these illnesses, survival analyses taking into account age of onset, time varying measures, and longitudinal methodology would be necessary to mitigate this limitation and confirm our results.
Conclusion
Despite these limitations, this research strengthens our understanding of the relationship between CSA and adult revictimization in women and adds unique impact to the literature in male survivors of CSA. CSA is directly associated with adult sexual trauma above and beyond childhood physical abuse and commonly associated substance dependence and psychopathology. The strength of the association between CSA and adult sexual victimization is more elevated in men than women, pointing to the need for further investigation of sexual victimization in male community samples. Furthermore, additional attention should be given to male survivors of CSA, in order to decrease risk of future victimization. Although CSA factor scores were consistently associated with adult sexual victimization, findings involving other psychopathology are less clear and thus more difficult to interpret. Additional research will be needed to understand the relationship between SUDs, PTSD, and MDD with adult victimization and the timing and interplay between and among these factors.
Acknowledgements
The authors thank Yi-Ling Chou for her statistical assistance with the meta-analysis. This work was primarily supported by National Institute of Health grants to E.C.N. (AA13446; DA017305; AA011998_5978); support for CATS was also received from the National Drug and Alcohol Research Centre (NDARC) and the Australian National Health and Medical Research Council (to L.D.). Additional support includes grants to A.C.H. (AA10249, AA07728, AA11998, AA13321, AA17688), N.G.M. (AA13326), P.F.M. (DA12854; DA027995; AA13446; DA017305; AA011998_5978), VVM (AA018146), A.A. (DA23668; DA32573), L.D. (Australian National Health and Medical Research Council Principal Research Fellowship no. 1041742), and K.B.W. (T32 DA15035).
Declaration of Interest
None.
