The role of childhood traumatization in the development of borderline personality disorder in Hungary

Eur. J. Psychiat. Vol. 29, N.° 2, (105-118) 2015 Keywords: Borderline personality disorder; Childhood traumatization; Physical abuse; Sexual abuse; Emotional abuse. The role of childhood traumatization in the development of borderline personality disorder in Hungary Katalin Merzaa,* Gábor Pappb Ildikó Kuritárné Szabó, PhDa aUniversity of Debrecen, School of Public Health, Department of Behavioral Sciences, Debrecen bUniversity of Debrecen, Institute of Psychology, Debrecen HUNGARY ABSTRACT – Background and Objectives: There is a growing body of evidence suggesting the role of childhood abuse in the etiology of borderline personality disorder (BPD). Studies found that complex traumatization related to BPD include emotional/physical/sexual abuse and neglect. This study examines self-reported experiences of childhood traumatization in Hungarian inpatients with a diagnosis of borderline personality disorder and reveal which etiological factors are most strongly associated with the development of BPD. Methods: Traumatic childhood experiences of 80 borderline inpatients, 73 depressed inpatients and 51 healthy controls were assessed with the Traumatic Antecedents Questionnaire and the Sexual Abuse Scale of Early Trauma Inventory. Results: Adverse childhood experiences (neglect, emotional abuse, physical abuse, sexual abuse, witnessing trauma) were more prevalent among borderline patients than among depressed and healthy controls. Borderline patients reported severe sexual abuse, characterized by incest, penetration and repetitive abuse. Sexually abused borderline patients experienced more physical and emotional abuse than borderlines who were not sexually abused. The strongest predictors of borderline diagnosis were sexual abuse, intrafamilial physical abuse and neglect by the caretakers. Conclusions: Overall, our results suggest that a reported childhood history of abuse and neglect are both common and highly discriminating for borderline patients in Hungary as well. Received: 1 December 2014 Revised: 3 March 2015 Accepted: 9 April 2015 106 KATALIN MERZA ET AL. Background and objectives The role of childhood traumatization in the etiology of borderline personality disorder (BPD) has been a focus of research for more than 30 years1-4. North American and Western European studies have provided a large body of evidence on the numerous intrafamilial pathological childhood experiences, such as a history of physical, emotional, sexual abuse and neglect that are commonly reported by borderline patients1,46. More specifically, 40-86% of subjects with BPD have reported being sexually abused during their childhood7-14, to a maximum of 94.7% reported by McLean and Gallop in 200315. Furthermore, 10-73% have reported being physically abused by parents or adult caretakers1,9,11,13,16,17 and 17-25% have reported being physically neglected11,13. Threequarters of borderline patients have reported emotional abuse and 70% have reported childhood emotional withdrawal13. In addition, findings of a large number of studies show that childhood experiences of both abuse and neglect are significantly more common among borderline patients than among comparisons with a diagnosis of Axis I or Axis II disorders2,11-13,18. Of all of the psychosocial factors, childhood sexual abuse is considered to be the most specific in the etiology of BPD19. Parameters of childhood sexual abuse highly discriminate between abused BPD patients and abused non-BPD patients20,21. Borderline patients compared to Axis I and II subjects have reported the most severe parameters of sexual abuse: significantly higher intra-familiar rates (72%)15, no single incidents of abuse (50%)11, multiple perpetrators (35-79%)21,22, early-onset abuse (13-60%)8,10,12,17,19, use of force (93%)21 and penetration (33-44%)8,11,21. Although studies have found that 40-75% of borderline patients are abused by a full-time adult caretaker9,23,24, reported childhood sexual abuse in borderline samples is perpetrated not only by parents. Moreover, reporting memories of sexual abuse by siblings, grandfathers or other family members could, in some cases, be a ‘screen’ for parental incest, which some may find intolerable to recall8. However, sexual abuse is neither necessary, nor sufficient for the development of BPD13. A meta-analytic study of the published literature (containing 19 North-American and two European studies) before 1999, conducted by Fossati, Mededdu and Maffei25, has found that in Western society there is only a moderate association between childhood sexual abuse and BPD diagnosis. Sexual abuse does not occur in a vacuum, it occurs in the context of other forms of abuse and dysfunctional parental behavior26,27 and contributes to the development of BPD through interaction with other pathological childhood experiences4,5. For example, sexually abused BPD patients are more likely to report physical neglect, emotional withdrawal, and inconsistent treatment by caretakers13. They usually come from disturbed families which do not or cannot protect their children and fail to meet their needs2,8. Thus BPD patients have also reported significantly higher rates of psychiatric disorders in their families, especially anxiety disorders, depression, separation from parents, unfavorable parental rearing styles1, antisocial disorders and substance use disorder28,29. The findings of previous researches suggest that studies should assess a range of pathological childhood experiences, rather than focus solely on the prevalence of sexual abuse13. Outside North-America and Western Europe, only a limited number of studies have examined the relationship between childhood abuse and borderline psychopathology. Recently, results emerging from studies conducted in Japan and China contribute to un- derstanding cross-cultural and culture-specific aspects of borderline etiology30,31,32. Consistent with prior North American and Western European research, both Japanese and Chinese studies have found that experiences of emotional, physical, sexual abuse, emotional and physical neglect are more prevalent among BPD patients than among non-BPD patients30,31,33. In these studies, the reported rate of childhood sexual abuse among borderline patients was lower than reported rates in North America, attributed by Huang30 to the lower rate of reported sexual abuse in the general population of China. In addition, in these studies only outpatient samples were used and the findings may not be generalizable to more severely disturbed inpatients30. The severity of childhood sexual abuse was also greater among BPD patients in China who had suffered incest, penetration and repetitive abuse30. Furthermore, Huang and colleagues30 used multivariate analyses to examine the variables which predict BPD. The results in China revealed that a combination of sexual abuse, maternal neglect, maternal physical abuse, and paternal antipathy were significant predictors of BPD, but none of these variables alone were predictive of BPD30. In the Japanese study, emotional abuse, emotional neglect and paternal overprotection were significant predictors of BPD31. The authors have assumed the results reflect differences between parenting styles in Western countries and the Far East, the latter being more collectivistic and the parenting style, more authoritarian, characterized by coercive control and low responsiveness31,32,34. Importantly, pathological childhood experiences are just a part of the more complex borderline etiology. The multifactorial model of the development of BPD suggests that the disorder results from the interaction and transaction of biological and environmental factors35,36. Examining biological factors is outside the scope of this study. In summary, childhood experiences of abuse are part of a cross-cultural etiology of BPD in North American and Western European countries as well as the Far East, but the relative contribution of particular forms of childhood abuse to the development of this disorder seems to vary across cultures. A literature search has revealed that in Hungary and in the European Post-Soviet states there have been no studies that have investigated the role of traumatic childhood experiences in the development of BPD. The purposes of this study are to assess retrospectively the self-reported experiences of childhood traumatization in Hungarian inpatients with a diagnosis of BPD and to determine which etiological factors are most strongly associated with the development of BPD. Furthermore, we seek to reveal the differences between Hungarian and North-American/Western European borderline samples in order to understand the relative contributions of etiological factors that are culture-specific and others that are cross-cultural. Hungary and China both being countries with only relatively recent Post-Soviet histories, parents may not consider corporal punishment as abuse and, therefore, this parenting style may be correlated with higher prevalence of child physical abuse. We hypothesized that in our study physical abuse as a culture-specific factor would increase the risk for BPD. Furthermore, we assume that the major predictors (neglect, sexual abuse) of BPD in the Western societies would also associated with the development of BPD in Hungary. In this study borderline inpatients were compared with depressed inpatients and a group of healthy individuals in the community. The control groups were initially chosen, because in sexual abuse research, chronic depression is often seen as a long-term result of sexual abuse23,37,38 and because, BPD patients are rarely compared with a healthy control group with regard to childhood trauma1. THE ROLE OF CHILDHOOD TRAUMATIZATION IN THE DEVELOPMENT... 107 Methods Participants and procedure The present study was conducted over the period from January 2013 to June 2013. The study followed the ethics declaration of Helsinki and prior to data collection the project was approved by the Hungarian Scientific and Medical Ethics Committee (ETT-TUKEB). 171 inpatients were recruited from eight Hungarian psychiatric hospitals. All patients were initially screened to determine that they 1) were between the ages of 18 and 50 years 2) had at least an average level of intellectual functioning and 3) had been given a definite or a probable clinical diagnosis of borderline personality disorder or major depressive disorder by a senior psychiatrist. Participants were excluded if they had current symptoms or a history of 1) bipolar mood disorder 2) major psychotic disorder or 3) cognitive impairment. After providing patients with a complete description of the study, written informed consent was obtained from each of them. The Hungarian version of Structured Clinical Interview for DSM-IV Axis I and Axis II disorders (SCID-I-II)39,40 was then administered to confirm the diagnosis of borderline personality disorder and major depressive disorder (MDD). 18 of 171 participants were eliminated by the diagnostic interview: 14 because they were found to have disorders excluded from this study, as above, and 4 because they met less than five BPD diagnostic criteria. Patients who met at least five DSM-IV criteria for BPD on the SCID-II were included in the borderline cohort. Patients who met the diagnosis of major depressive disorder without personality disorders were the non-borderline depressed controls. The interviews were conducted by the first author, a well-trained psychologist. Healthy controls were recruited for the study from employees of local companies in Debrecen. Out of a pool of 62 controls 51 were free of psychiatric disorders, as assessed by the SCID. Measures Experiences of childhood traumatization were assessed by the Traumatic Antecedents Questionnaire (TAQ) including in the Trauma Center Assessment Package41. The TAQ is a 42-item self-report questionnaire for gathering information about the frequency and severity of traumatic and adaptive experiences and it has been used by many researchers15,42. The process of the translation and adaptation of TAQ to Hungarian followed a standard procedure according to the guidelines for self-assessment instruments. Analysis of the psychometric properties of the TAQ is currently underway, and preliminary research showed satisfactory validity and reliability (with Cronbach alphas of 0.683-0.923). In this study only 6 subscales of the TAQ were used, selected for the purpose of assessing traumatic or adverse experiences. These subscales cover neglect, separation, emotional abuse and physical abuse by a caretaker or family member, sexual abuse by an adult and witnessing of domestic violence. These adverse experiences were assessed at four different developmental periods: early childhood (0-6 years), latency (7-12 years), adolescence (13-18 years), and adulthood (over 18 years). Responses can be scored 1 to 4, with 1 corresponding to ‘never or not at all’, 2 to ‘rarely or a little bit’, 3 to ‘occasionally or moderately’, 4 to ‘often or very much’. However, in this study the type of childhood traumatization was analyzed as a dychotomic variable (traumatized – non-traumatized), because we aimed to compare our results with the results of previous studies which also used dychotomic variables. The cut-off point was determined 2/3, thus sub108 KATALIN MERZA ET AL. jects who scored 1 (‘never or not at all’) or 2 (‘rarely or little bit’) in each item were considered non-traumatized, and subjects who scored 3 (‘occasionally or moderately’) or 4 (‘often or very much’) were considered traumatized. The 2/3 cut-off point was initially chosen to prevent the overestimation of traumatic experiences in the sample. Previous studies have indicated that the parameters of childhood sexual abuse have a specific relationship with the borderline diagnosis and highly discriminate abused BPD patients and abused non-BPD patients21. In view of this, we aimed to examine the parameters of sexual abuse in detail. In addition to the TAQ, the Sexual Abuse Scale of the Early Trauma Inventory (ETI)43 (with Cronbach alpha of 0.932) was used to assess childhood experiences of sexual abuse more accurately. The process of the translation of ETI Sexual Abuse Subscale to Hungarian language followed the standard protocol. Correlation analysis between the Sexual Abuse Scale of ETI and the sexual abuse subscale of TAQ showed that the scales measure the same construct (r= 0.88, p< 0.001). The ETI defines sexual abuse as unwanted sexual contact performed solely for the gratification of the perpetrator, for the purpose of dominating or degrading the victim. This 15item scale gathers information about frequency, age at onset, use of force, relationship to the perpetrator and nature of sexual abuse at 4 developmental stages (similar to the developmental stages of the TAQ). All assessments were done by the first author. Between-group comparison of age involving continuous data were computed by means of one-way analysis of variance. Chi-square tests were performed to compare the frequency distributions of certain demographic characteristics, traumatic experiences and the nature of sexual abuse between the patient and the control groups. When cell sizes were 5 or less Fischer’s exact test of probability was used. Logistic regression was used to test the research question regarding the possible predictors of BPD. A p value less than 0.05 was considered statistically significant. The statistical analyses were performed by the SPSS statistical package version 20.0. Results The final sample consisted of 204 participants, of whom 80 psychiatric inpatients were in the BPD group, 73 psychiatric inpatients were in the depressed comparison group and 51 people were in the healthy comparison group. Table 1 presents the comparison of de - mographic data among the three groups. To evaluate age differences among the groups one-way analysis of variance (ANOVA) was conducted. ANOVA revealed significant differences among the three groups (F= 49.42, p< 0.001), the mean ages of the BPD group (30.5 ± 10.87 years) and of the healthy comparison group (33.6 ± 8.71) were significantly lower than of the depressed comparison group (44.3 ± 5.91 years) according to the results of the Tukey test. Chi-square test for the categorical data showed significant differences in marital status ( χ 2= 88.72, p< 0.001) and employment ( χ 2= 130.1, p< 0.001) among the groups, with borderline patients less likely to be married and more likely to be unemployed than comparison subjects. The groups were found to be similar in sex distribution, with all groups containing significantly more females than males ( χ 2= 0.43, p= 0.808). There were no statistically significant differences among the groups for education ( χ 2= 9.00, p= 0.061). THE ROLE OF CHILDHOOD TRAUMATIZATION IN THE DEVELOPMENT... 109 Table 2 compares borderline patients and control subjects on overall rates of reported childhood traumatization, specifically, neglect by caretakers, separation from caretakers, intrafamilial emotional and physical abuse, intra/extrafamilial sexual abuse and witnessing family violence. A Chi-square test revealed significant differences among the groups for neglect ( χ 2= 35.88, p< 0.001), emotional abuse ( χ 2= 34.36, p< 0.001), physical abuse ( χ 2=51.58, p< 0.001), sexual abuse ( χ 2=45.52, p< 0.001) and witnessing trauma ( χ 2=45.02, p< 0.001) before the age of 18. There were no significant differences among the groups for separation ( χ 2= 5.19, p= 0.075). As seen in Table 3, subsequent comparisons indicated that the rates of childhood neglect ( χ 2= 24.19, p< 0.001; χ 2= 31.63, p< 0.001), emotional ( χ 2=17.47, p< 0.001; χ 2= 33.77, p< 0.001), physical ( χ 2 = 32.67, p< 0.001; χ 2 =35.78, p< 0.001), sexual abuse ( χ 2 = 20.42, p< 0.001; χ 2 = 37.07, p< 0.001) and witnessing trauma ( χ 2 =12.02, p< 0.001; χ 2 = 44.95, p< 0.001) were significantly higher in the BPD group than in depressed and healthy control groups. 110 KATALIN MERZA ET AL. Table 1 Demographic characteristics of the borderline and depressed groups. Analysis Demographic BPD Depressed Healthy characteristic (N = 80) (N = 73) (N = 51) χ 2P Age (Mean, SD)* 30.5 ± 10.87 44.3 ± 5.91 33.6 ± 8.71 <0.001 Sex Male 12 (15%) 13 (17.8%) 7 (13.7%) 0.43 (df = 2) 0.808 Female 68 (85%) 60 (82.2%) 44 (86.3%) Marital status Single 50 (62.5%) 6 (8.2%) 19 (37.0%) 88.72 (df = 8) <0.001 Married 3 (3.8%) 42 (57.5%) 12 (24.0%) Common-law marriage 10 (12.5%) 6 (8.2%) 16 (31.0%) Divorced 14 (17.5%) 14 (19.2%) 2 (4.0%) Relict 3 (3.8%) 5 (6.9%) 2 (4.0%) Education Primary school or below 19 (23.8%) 13 (17.8%) 5 (9.8%) 9.00 (df = 4) 0.061 Secondary school 51 (63.8%) 45 (61.6%) 30 (58.8%) University 10 (12.5%) 15 (20.5%) 16 (31.4%) Employment Unemployed 30 (37.5%) 14 (19.2%) 2 (3.9%) 130.1 (df = 12) <0.001 Disabled 13 (16.3%) 39 (53.4%) 0 (0.0%) Full-time emloyment 6 (7.5%) 6 (8.2%) 26 (51.0%) Part-time employment 6 (7.5%) 1 (1.4%) 2 (3.9%) Temporary employment 4 (5.0%) 0 (0%) 1 (2.0%) Student 16 (20.0%) 1 (1.4%) 19 (37.3%) Other 5 (6.3%) 12 (16.4%) 1 (2.0%) Note. * F = 49.42. Table 4 presents the prevalence of adverse experiences reported by the groups at the three age periods, early childhood (0-6 years), latency (7-12 years) and adolescence (13-18 years). A Chi-square test indicated significant differences among the groups for neglect, emotional, abuse, physical abuse, sexual abuse and witnessing trauma at each developmental period. There were no significant differences among the groups for separation at early childhood and adolescence. Table 4 shows the subsequent comparisons of the groups on the reported rates of childhood traumatization separately at each age period. Chi-square test indicated significant differences between the borderline and both control groups on most types of traumatization at each period. More specifically, a higher percentage of BPD patients than depressed and healthy comparisons reported having experienced neglect, emotional abuse, physical abuse, sexual abuse and witnessing trauma at early childhood, latency and adolescence. A significantly higher percentage of borderlines than depressed controls also reported separation during all three of these periods. However, at early childhood and adolescence borderline and healthy groups did not differ on reported rates of separation from caretakers. THE ROLE OF CHILDHOOD TRAUMATIZATION IN THE DEVELOPMENT... 111 Table 3 Subsequent comparisons of the groups on reported rates of childhood traumatization. BPD vs. Depressed BPD vs. Healthy Type of traumatization χ 2 (df = 2) P χ 2 (df = 1) P Neglect 24.19 <0.001 31.63 <0.001 Separation 3.94 0.047 0.003 0.956 Emotional abuse 17.47 <0.001 33.77 <0.001 Physical abuse 32.67 <0.001 35.78 <0.001 Sexual abuse 20.42 <0.001 37.07 <0.001 Witnessing 12.02 <0.001 44.95 <0.001 Table 2 Reported rates of childhood traumatization in the borderline group and the control groups before the age of 18. Analysis Type of BPD Depressed Healthy traumatization (N = 80) (N = 73) (N = 51) χ 2P Neglect 69 (86.3%) 36 (49.3%) 20 (39.2%) 35.88 <0.001 Separation 64 (80.0%) 48 (65.8%) 41 (80.4%) 5.19 0.075 Emotional abuse 70 (87.5%) 42 (57.5%) 20 (39.2%) 34.36 <0.001 Physical abuse 52 (65.0%) 14 (19.2%) 6 (11.8%) 51.58 <0.001 Sexual abuse 45 (56.3%) 15 (20.5%) 2 (3.9%) 45.52 <0.001 Witnessing 62 (77.5%) 37 (50.7%) 9 (17.6%) 45.02 <0.001 112 KATALIN MERZA ET AL. Table 4 Comparisons of the groups on reported rates of childhood traumatization in the three developmental stages. Analysis All Groups BPD vs. Healthy BPD vs. Depressed Type of traumatization BPD Depressed Healthy χ 2 χ 2 χ 2 (N = 80) (N = 73) (N = 51) (df = 2) P(df = 2) P(df = 1) P Neglect Early childhood (0-6 years) 50 (62.5%) 27 (37.0%) 10 (19.6%) 24.92 <0.001 23.08 <0.001 9.94 0.002 Latency (7-12 years) 56 (70.0%) 34 (46.6%) 13 (25.5%) 25.38 <0.001 24.75 <0.001 8.65 0.003 Adolescence (13-18 years) 68 (85.0%) 34 (46.6%) 20 (39.2%) 35.43 <0.001 29.61 <0.001 25.36 <0.001 Separation Early childhood (0-6 years) 32 (40.0%) 18 (24.7%) 19 (37.3%) 4.37 0.112 0.09 0.753 4.08 0.043 Latency (7-12 years) 48 (60.0%) 29 (39.7%) 29 (56.9%) 6.94 0.031 0.13 0.722 6.28 0.012 Adolescence (13-18 years) 61 (76.3%) 43 (58.9%) 32 (62.7%) 5.64 0.06 2.76 0.097 5.28 0.022 Emotional abuse Early childhood (0-6 years) 48 (60.0%) 20 (27.4%) 7 (13.7%) 32.98 <0.001 27.38 <0.001 16.43 <0.001 Latency (7-12 years) 59 (73.8%) 37 (50.7%) 13 (25.5%) 29.49 <0.001 29.30 <0.001 8.69 0.003 Adolescence (13-18 years) 68 (85.0%) 34 (46.6%) 20 (39.2%) 35.43 <0.001 29.61 <0.001 25.36 <0.001 Physical abuse Early childhood (0-6 years) 34 (42.5%) 6 (8.2%) 0 (0.0%) 45.04 <0.001 29.27 <0.001 23.23 <0.001 Latency (7-12 years) 40 (50.0%) 10 (13.7%) 2 (3.9%) 43.14 <0.001 30.36 <0.001 22.86 <0.001 Adolescence (13-18 years) 47 (58.8%) 10 (13.7%) 6 (11.8%) 47.94 <0.001 28.54 <0.001 33.14 <0.001 Sexual abuse Early childhood (0-6 years) 17 (21.3%) 4 (5.5%) 0 (0.0%) 18.08 <0.001 12.45 <.0001 8.02 0.005 Latency (7-12 years) 28 (35.0%) 7 (9.6%) 0 (0.0%) 31.42 <0.001 22.70 <0.001 13.97 <0.001 Adolescence (13-18 years) 42 (52.5%) 6 (8.2%) 2 (3.9%) 56.03 <0.001 32.95 <0.001 34.76 <0.001 Witnessing Early childhood (0-6 years) 45 (56.3%) 16 (21.9%) 3 (5.9%) 41.42 <0.001 34.03 <0.001 18.77 <0.001 Latency (7-12 years) 50 (62.5%) 22 (30.1%) 5 (9.8%) 39.61 <0.001 35.51 <0.001 16.05 <0.001 Adolescence (13-18 years) 55 (68.8%) 34 (46.6%) 8 (15.7%) 35.21 <0.001 35.13 <0.001 7.71 0.005 Table 5 compares borderline and depressed patients who reported childhood sexual abuse with respect to the parameters of sexual abuse. In this sample 45 borderline patients and 15 depressed comparisons reported having experienced sexual abuse before the age of 18. For patients who had been abused by more than one perpetrator, data were entered for the specific perpetrator whose abuse would be expected to be more traumatic, according to the Paris & Zweig-Frank’s21 hierarchy. A significantly higher percentage of borderline patients than depressed comparisons reported childhood sexual abuse perpetrated by a father/male caretaker ( χ 2 =5.93, p= 0.015). Moreover, the frequency of multiple perpetrators was also higher among BPD patients ( χ 2 = 9.27, p< 0.001). More BPD patients reported abuse by perpetrators familiar to them ( χ 2 =0.86, p = 0.353), and by brothers ( χ 2 = 0.69, p= 0.41), but these differences were not significant. Furthermore, the rate of childhood sexual abuse by a male stranger was significantly higher in the depressed control group ( χ 2 = 4.69, p= 0.03). There was a large, significant difference between the groups with respect to the nature of childhood sexual abuse. The prevalence of THE ROLE OF CHILDHOOD TRAUMATIZATION IN THE DEVELOPMENT... 113 Table 5 Parameters of childhood sexual abuse in the borderline group and in the depressed control group. Analysis Parameters BPD (N = 45) Depressed (N = 15) χ 2 (df = 1) P Perpetrator Father/Male caretaker 22 (48.9%) 2 (13.3%) 5.93 0.015 Brother 2 (4.4%) 0 (0.0%) 0.69 0.41 Other male 18 (40.0%) 4 (26.7%) 0.86 0.353 Other female 2 (4.4%) 0 (0.0%) 0.69 0.406 Male stranger 13 (28.9%) 9(60.0%) 4.69 0.03 Multiple 19 (42.2%) 0 (0.0%) 9.27 <0.001 Nature Fondling 43 (95.6%) 13 (86.7%) 1.43 0.232 Oral sex 23 (51.1%) 3 (20.0%) 4.43 0.035 Penetration 34 (75.6%) 3 (20.0%) 14.68 <0.001 Incomplete penetration 33 (73.3%) 6 (40.0%) 5.49 0.019 Anal penetration 2 (4.4%) 0 (0.0%) 0.69 0.406 Frequency Single incident 8 (17.8%) 9 (60.0%) 9.87 0.002 More than once a year 10 (22.2%) 5 (33.3%) 0.124 0.389 Monthly 18 (40.0%) 1 (6.7%) 5.77 0.016 Weekly 5 (11.1%) 0 (0.00%) 1.82 0.178 Daily 4 (8.9%) 0 (0.00%) 1.43 0.232 Age at onset Early childhood 0-6 years 17 (37.8%) 4 (26.7%) 0.61 0.435 Latency 7-12 years 18 (40.0%) 5 (33.3%) 0.212 0.646 Adolescence 13-18 years 10 (22.2%) 6 (40.0%) 0.182 0.178