Attitudes toward life and death and suicidality among inpatient female adolescents with eating disorders.

ORIGINAL ARTICLE

Attitudes Toward Life and Death and Suicidality Among Inpatient Female Adolescents With Eating Disorders Daniel Stein, MD,*Þ Dana Zinman, PhD,*þ Liron Halevy, MsC,* Amit Yaroslavsky, MD,* Eytan Bachar, PhD,§ Shulamit Kreitler, PhD,*þ and Israel Orbach, PhD|| Abstract: This study investigated whether attitudes about life and death are associated with suicidal behavior in eating disorders (EDs). We examined 43 nonsuicidal inpatients with EDs, 32 inpatients with EDs who attempted suicide, and 21 control participants with scales assessing attitudes to life and death, bodyrelated attitudes, core ED symptoms, depression, and anxiety. Both ED groups showed less attraction to life and more repulsion from life than did the control participants. The suicide attempters showed greater attraction to death, less repulsion from death, and more negative attitudes toward their body than did the nonsuicidal ED and control participants. Fear of life was associated with elevated depression, body-related problems, and childhood sexual abuse. Pathological attitudes toward death were associated with greater depression and body-related problems. Suicide attempts were found in the inpatients with EDs showing binge/ purge ED pathology and maladaptive attitudes toward death. This study suggests that whereas fear of life is a core feature of an ED, maladaptive attitudes toward death appear only in ED patients who have attempted suicide. Key Words: Anorexia nervosa, body image, bulimia nervosa, suicide. (J Nerv Ment Dis 2013;201: 1066Y1071)

S

uicidal behavior is an issue of great concern in anorexia nervosa (AN) and bulimia nervosa (BN; Franko and Keel, 2006). Several factors may be associated with suicidal behavior in patients with eating disorders (EDs), including binge/purge (B/P) type ED pathology, impulsivity, depression, and greater severity of the ED (Bulik et al., 2008; Corcos et al., 2002; Favaro and Santonastaso, 1997; Stein et al., 2004). Disturbances in body image and body dissatisfaction represent another factor likely associated with suicidal behavior in patients with EDs (Corcos et al., 2002). Still, according to Bruch (1973), it is not necessarily the fear of gaining weight and the pursuit of thinness that increase suicidality in EDs. Rather, incorrect interpretation of interoceptive stimuli, insensitivity to bodily functions, and lack of bodily control may eventuate in a detachment from the body, leading, in turn, to self-neglect and the facilitation of self-destructive behavior (Bruch, 1973). Orbach has theorized that basic feelings, attitudes, and experiences of the body may affect one’s attitudes toward life and death (Orbach and Mikulincer, 1998; Orbach et al., 2001) and that selfpreservation requires an active protecting and caring attitude toward the body (Orbach and Mikulincer, 1998; Orbach et al., 2001). Using a questionnaire assessing attitudes toward life and death, the MultiAttitude Suicidal Tendencies (MAST) scale (Orbach et al., 1991),

*Edmond and Lily Safra Children’s Hospital, The Chaim Sheba Medical Center, Tel Hashomer, Israel; †Sackler Faculty of Medicine, and ‡Department of Psychology, Tel Aviv University, Tel Aviv, Israel; §Departments of Psychology and Psychiatry, The Hebrew University, Jerusalem, Israel; and ||Department of Psychology, Bar Ilan University, Ramat Gan, Israel. Send reprint requests to Daniel Stein, MD, Pediatric Psychosomatic Department, Edmond and Lily Safra Children’s Hospital, The Chaim Sheba Medical Center, Tel Hashomer, 52621, Israel. E-mail: [email protected]. Copyright * 2013 by Lippincott Williams & Wilkins ISSN: 0022-3018/13/20112Y1066 DOI: 10.1097/NMD.0000000000000055

1066

www.jonmd.com

Orbach has shown that suicidal patients, regardless of their specific psychiatric diagnosis, are characterized by decreased attraction to life (AL), greater repulsion from life (RL), greater attraction to death (AD), and less repulsion from death (RD) in comparison with nonsuicidal patients (Orbach and Mikulincer, 1998; Orbach et al., 2001). A few studies have previously used the MAST in EDs. Bachar et al. (2002) have found that adult patients with EDs show less AL and more RL but no differences in the MAST death subscales in comparison with healthy control participants. Our group (Stein et al., 2003) has previously assessed attitudes to life and death in suicidal and nonsuicidal psychiatric adolescent inpatients, nonsuicidal adolescent inpatients with AN, and healthy control participants. In accordance with the findings of Bachar et al. (2002), we have shown that nonsuicidal patients with AN are similar to suicidal psychiatric patients and different from both other groups in showing less AL and more RL but different from the suicidal patients and similar to the other groups in showing no elevated maladaptive attitudes toward death. The AN and suicidal patients have also been different from either one or the two other groups in showing more negative attitudes and feelings toward their bodies. These findings have lead us to suggest that adolescent AN patients may go through an existential developmental crisis in which they are unable to carry out a basic age-specific function, the choice of life (Tyano and Vincent, 1997). This contention is supported by the elevated maturity fears of many patients with AN (Garner, 1991) and by their resistance to change (i.e., elevated harm avoidance; Fassino et al., 2002), both related to a considerable dysfunction in age-specific tasks such as the development of peer relationships (Strober et al., 1997), mature sexuality, and intimacy (Crisp, 1980). The aim of the present study was to investigate potential factors associated with maladaptive attitudes toward life and death and with a history of attempted suicide among female adolescents diagnosed with EDs. The following are our hypotheses: 1. ED patients with a history of attempted suicide and nonsuicidal patients with EDs will show less AL and more RL than healthy control participants. Greater fear of life will be associated with elevated anxiety, social insecurity, body image disturbances, and negative attitudes toward the body. 2. ED patients with a history of attempted suicide will show more AD and less RD than both the nonsuicidal and control participants. Maladaptive attitudes toward death will be associated with elevated depression, body image disturbances, and negative attitudes toward the body. 3. A history of attempted suicide in patients with EDs will be associated with B/P pathology, more severe ED symptoms, body image disturbances, negative attitudes toward the body, impulsivity, depression, and maladaptive attitudes toward death.

METHODS Participants We studied all female adolescents and young adults hospitalized because of an ED in the Pediatric Psychosomatic Department at the Sheba Medical Center, Tel Hashomer, Israel, between 2003 and 2007, who complied with this study’s inclusion and exclusion criteria. The

The Journal of Nervous and Mental Disease

& Volume 201, Number 12, December 2013

Copyright © 2013 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.



inclusion criteria were being of female sex between the ages of 14 and 19 years, having a good understanding of the Hebrew language, completion of inpatient treatment, and responding to this study’s questionnaires. The exclusion criteria consisted of current or lifetime bipolar disorder, schizophrenic spectrum disorder, organic brain syndrome, mental retardation, and any lifetime or current medical disorder with the potential to affect food consumption and weight (e.g., diabetes mellitus). Forty-five of the original group of 150 inpatients did not take part in this study, either because of not fulfilling this study’s inclusion or exclusion criteria or because of refusal to participate. No differences were found between the 105 patients who participated in this study and the 45 patients excluded from this study in any of the parameters assessed. The research group was divided into suicidalVhaving made at least one suicide attempt before or during hospitalization (n = 32), and nonsuicidalVhaving no evidence of current or lifetime suicidal behavior or self-injurious behavior (SIB; n = 43). A suicide attempt was defined as a nonfatal self-destructive act that was sufficiently serious to require medical evaluation and was carried out with the intent to die or with an expectation of death; SIB was defined as any selfdestructive act that was not life threatening and that was carried out with no/minimal intent to die and/or no/minimal expectation of death. These definitions were based on Linehan’s definitions of suicide attempts and SIBs (Brown et al., 2002), which were previously tested in other populations of ED patients (Stein et al., 2004). Use of medications to lose weight was not considered a suicidal behavior. The 10 inpatients who both attempted suicide and engaged in SIBs were included in the suicidal group because they were not different from the other 22 suicidal patients with no SIBs in any of the parameters introduced. The 30 inpatients with EDs with evidence of SIBs who did not attempt suicide were excluded from this study because of the controversy regarding whether SIBs represent suicidal behavior (Walsh and Rosen, 1998). Thirty-one of the 32 suicidal patients attempted suicide with medication overdose, and one patient attempted to drown herself. All suicidal inpatients with SIBs cut themselves with sharp objects. The control participants included 21 female volunteers between the ages of 14 and 19 years recruited from families of the staff of the hospital who were matched with the research patients on age, years of schooling (the participants of all groups were high school students), and parents’ country of birth and years. The control participants were required to have no lifetime or current psychiatric disorder, medical disorder, or long-term use of medications; no stigmata indicative of an ED; and no evidence of suicide attempts and SIBs. The weight of these participants had to be more than 85% of ideal body weight since puberty according to the 2000 sex-specific growth charts from the Centers for Disease Control and Prevention (www.cdc.gov/growthcharts), which were found adequate for Israeli youngsters (Goldstein et al., 2001). The control participants were also required to have regular menses since menarche. The participants and their parents, in the case of minors younger than 18 years, signed a written informed consent after receiving an explanation about the goals and the methodology of this study. This study was approved by the ethics review board of the Sheba Medical Center.

Instruments Diagnosis of an ED was obtained using the Eating Disorders Family History Interview (EDFHI; Strober, 1987). This semistructured clinical interview is designed to collect data about eating and weight history and to determine the presence of a Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994), ED diagnosis. The EDFHI was previously used in patients with EDs (Pollice et al., 1997), including in Israeli samples (Yackobovitch-Gavan et al., 2009). ED not otherwise specified (ED-NOS) was diagnosed when the patients fulfilled all but * 2013 Lippincott Williams & Wilkins

Life and Death and Suicide in ED

one DSM-IV criterion for AN or BN and the disturbance was deemed clinically significant. This category was divided into restricting (EDNOS-R)and B/P type ED-NOS (ED-NOS-B/P). Other psychiatric morbidity was determined using the Structured Clinical Interview for DSM-IV Axis I DisordersYPatient Edition (SCID-I/P, version 2.0; First et al., 1995). We used the following self-rating scales: MAST Scale (Orbach et al., 1991): This 30-item scale provides four independent scores, AL, RL, AD, and RD. Higher scores indicate a greater value of the respective dimension. High AL, low RL, low AD, and high RD reflect low suicidal tendencies; the opposite constellation reflects high suicidal tendencies. The validity of the MAST in discriminating suicidal from nonsuicidal adolescents and its test-retest reliability have been determined previously (Orbach et al., 2001). In the present study, the internal consistencies of the MAST scales are AL, > = 0.90; RL, > = 0.83; AD, > = 0.84; and RD, > = 0.85. Body Investment Scale (BIS; Orbach and Mikulincer, 1998): This 24-item scale measures body-related attitudinal-behavioral parameters. It includes four factors: attitudes and feelings toward the body, comfort in touch, body care, and body protection. The BIS has been found to successfully differentiate between suicidal and nonsuicidal adolescents (Orbach and Mikulincer, 1998; Orbach et al., 2001; Stein et al., 2003) and between nonsuicidal AN patients and nonED controls (Stein et al., 2003). The internal consistencies of the BIS scales in the present study are > = 0.89 for attitudes and feelings toward the body, > = 0.75 for comfort in touch, > = 0.75 for body care, and > = 0.76 for body protection. Maladaptive eating-related parameters have been assessed using the Eating Attitudes TestY26 (EAT-26; Garner et al., 1982) and the Eating Disorder InventoryY2 (EDI-2; Garner, 1991). Both scales have been shown to successfully differentiate Israeli ED patients from controls (Koslowsky et al., 1992; Yackobovitch-Gavan et al., 2009). In the present study, we have used 6 of the 11 EDI-2 subscales, hypothesized to be of relevance for our purposes: drive for thinness, body dissatisfaction, perfectionism, interoceptive awareness, impulse regulation, and social insecurity. The internal consistency of the EAT-26 in the present study is > = 0.93; and that of the 11 subscales and the combined total EDI-2 score, > = 0.89 to 0.91. Depression was assessed using the 21-item Beck Depression Inventory (BDI; Beck et al., 1961). Anxiety was assessed using the 40item State-Trait Anxiety Inventory (STAI; Spielberger et al., 1970), which measures the severity of anxiety at the time of examination (STAI-State) and the general tendency to display anxiety (STAI-Trait). Both scales were previously used in ED patients (Pollice et al., 1997), including in Israeli samples (Yackobovitch-Gavan et al., 2009). In the present study, the internal consistencies of the BDI, the STAI-State, and the STAI-Trait were > = 0.89, > = 0.94, and > = 0.95, respectively.

Procedure The participants responded to a structured questionnaire providing relevant demographic data. Diagnoses of lifetime and current EDs and comorbid DSM-IV axis I psychiatric disorders were achieved independently by two experienced psychiatrists (D. S. and A. Y.) with the EDFHI and the SCID-I/P, respectively. The degree of interrater reliability (according to the correlation coefficient procedure) between the two interviewers for the EDFHI and the SCID-I/P version 2.0 was r = 0.91 and r = 0.89, respectively. Thereafter, diagnoses were confirmed in clinical meetings of the department. Only patients for whom there was a unanimous agreement about their ED diagnosis could enter the study. The control participants were interviewed with the EDFHI and the SCID-I/P by master’s level psychologists trained by the principal investigator (D. S.). The degree of interrater reliability between these interviewers and the principal investigator for the EDFHI and the SCID-I/P was r = 0.90 and r = 0.87, respectively. The patients and the www.jonmd.com


1067


Stein et al.


TABLE 1. Between-Group Differences in Attitudes Toward Life and Death and in Body-Related Attitudes Suicidal Patients (n = 32)

MAST-AL MAST-RL MAST-AD MAST-RD BISYattitudes and feelings BIS-touch BISYbody care BISYbody protection

25.7 T 19.9 T 21.9 T 19.4 T 2.0 T 2.8 T 2.2 T 2.8 T

Nonsuicidal Patients (n = 43)

a

6.62 5.68a 6.21a 7.85a 1.1a 0.8a 0.7 0.8a

28.7 T 17.8 T 17.2 T 24.1 T 2.6 T 2.8 T 2.5 T 3.7 T

a

6.82 6.25a 6.67b 7.91b 1.1b 0.7a 0.9 0.8b

Controls (n = 21) b

32.7 T 5.64 13.6 T 4.52b 17.4 T 3.88b 21.9 T 7.85 3.7 T 0.8c 3.5 T 0.5b 2.8 T 0.6 4.1 T 0.7b

F(2,93)

7.4** 7.6** 6.1** 3.6* 18.2*** 7.04** 2.3 (NS) 4.5***

Means with different letters are different from each other at p G 0.05; means with no letters are not different from any other mean in that row. *p G 0.05. **p G 0.01. ***p G 0.001.

controls were administered the questionnaires by another team of master’s level psychologists within 2 weeks of the psychiatric interview. Data for suicide attempts, SIBs, and trauma-related events occurring before or during hospitalization were collected with the SCID-I/P and from the patients’ files. Final decisions with respect to attempted suicide and self-harm were confirmed in clinical meetings of the department. Only those patients for whom there was unanimous agreement about the occurrence of attempted suicide were enrolled in this study. The body mass index (BMI; weight per height squared; Bray, 1992) of the research group was calculated from their weight and height at admission. The BMI of the controls was achieved by self-report.

Statistical Analysis Between-group comparison for age, BMI, the participants’ and the parents’ years of education, and the psychometric scales was achieved using one-way analysis of variance and Tukey’s post hoc analysis. Comparison among the three groups for country of birth of the participants and the parents and between the two research groups for ED subtype, DSM-IV comorbid diagnoses, and trauma-related events was achieved using chi-square tests. Comparison between the suicidal and nonsuicidal patients for duration of illness and hospitalization was achieved with t-tests for independent variables. We used Pearson’s correlation coefficients to assess the correlations between the four MAST attitudes and all other variables introduced. Clinical Variables showing significant correlations (p G 0.05) in individual analyses were entered into a hierarchal regression analysis for each MAST subscale. Psychometric variables showing significant correlations (p G 0.05) in individual analyses were entered into a stepwise regression analysis for each MAST subscale. Chi-square tests for dichotomous variables and t-tests for continuous variables were used for the comparison of the suicidal and nonsuicidal patients. All variables showing significant differences between the two groups (p G 0.05) in individual analyses were entered into a stepwise regression analysis.

RESULTS No significant differences were found among the three groups for the demographic variables introduced. The mean T SD age of the nonsuicidal patients, the patients attempting suicide, and the control participants was 16.0 T 1.7, 16.2 T 1.3, and 16.5 T 1.4 years, respectively (F[2,93] = 0.33, p value not significant [NS]), and more than 90% of all participants were born in Israel. Similarly, there were no differences for the parents’ country of birth, with 68% to 85% of the fathers (W22 = 0.35, NS), and 70% to 81% of the mothers (W22 = 0.90, NS) being Israeli born. The mean T SD years of education of the fathers of the nonsuicidal patients, the suicide attempters, and the control participants was 13.6 T 3.2, 12.4 T 3.2, and 14.4 T 2.7 years, respectively (F[2,93] = 2.85, NS]); and that of the mothers, 13.5 T 2.6, 12.5 T 1068

www.jonmd.com

1.5, and 14.0 T 2.5 years, respectively (F[2,93] = 3.15, p G 0.05]). No significant correlations were found between maternal years of education and any of the clinical and psychometric variables introduced. Significant between-group differences were shown for BMI, with the mean BMI of the suicide attempters (18.1 T 3.53) and the nonsuicidal patients (16.4 T 2.55) being significantly lower than the BMI of the control participants (20.8 T 2.56; F[2,93] = 17.2, p G 0.001). No differences were shown between the BMI of the two ED groups in a three-group analysis. Still, a significant difference was found in BMI when comparing only between the two ED groups (t = 2.5; p G 0.5). With respect to ED subtype, we found that among the suicide attempters, 10 patients were diagnosed with BN; 6, with AN-B/P; 7, with AN-R; 8, with ED-NOS-B/P; and 1, with ED-NOS-R. Among the nonsuicidal group, 4 patients were diagnosed with BN; 6, with AN-B/P; 27, with AN-R; and 6, with ED-NOS-B/P. Thus, 24 of the 32 patients attempting suicide (75%) were diagnosed with B/P type EDs; and 8 patients (25%), with restricting-type EDs. The respective numbers and percentages for the nonsuicidal patients were 16 (37%) and 27 (63%). The difference in the ED subtype (B/P versus restricting) between the suicidal and the nonsuicidal group was significant (W21 = 15.8, p G 0.001). Table 1 summarizes the between-group comparison of the MAST and BIS dimensions. Both groups with EDs showed less AL (MAST-AL) and greater RL (MAST-RL) compared with the control participants. No differences were found in these parameters between the two groups with EDs. The suicide attempters showed greater AD (MAST-AD) than did both the control and nonsuicidal participants, who were not different from each other, and less RD (MAST-RD) than did the nonsuicidal group. With respect to the BIS, we found that the two groups with EDs showed less favorable results in comparison with the control group on attitudes and feelings toward the body and on comfort in touch. The patients attempting suicide had more pathological scores on the BISYattitudes and feelings toward the body than those of the nonsuicidal group and lower scores on the BIS-protection than those of both other groups. No between-group differences were found for the BISYbody care (see Table 1). Table 2 summarizes the between-group differences in eatingrelated pathology, depression, and anxiety. Both groups with EDs scored higher than did the control participants on all dimensions and did not differ from each other, except for the EDI-2 impulse regulation, in which only the suicide attempters fared worse than did the controls. Still, although the BDI scores of the two groups with EDs were not different from each other when both were compared with the non-ED control participants, comparison of only the two research groups revealed a significantly higher BDI score in the suicide attempters (31.1 T 13.9) versus the nonsuicidal group (24.1 T 12.7; t[73] = 5.64, p G 0.001). * 2013 Lippincott Williams & Wilkins





TABLE 2. Between-Group Differences in Eating-Related Pathology, Depression, and Anxiety Suicidal Patients (n = 32)

EAT-26 EDI-2-DT EDI-2-BD EDI-2-P EDI-2-IA EDI-2-IR EDI-2-SI BDI STAI-S STAI-T

Nonsuicidal Patients (n = 43)

a

45.7 T 19.5 14.1 T 6.3a 19.2 T 8.3a 7.7 T 4.9a 8.7 T 6.2a 7.6 T 5.3a 8.1 T 4.3a 31.1 T 13.9a 58.1 T 13.0a 54.9 T 10.0a

41.6 T 12.3 T 15.5 T 7.4 T 7.1 T 5.2 T 6.3 T 24.1 T 53.1 T 50.5 T

a

22.4 7.3a 9.5a 4.6a 5.7a 5.7 5.3a 12.7a 14.6a 12.4a

F(2,93)

Controls (n = 21)

7.8 3.4 6.0 4.7 2.0 2.3 2.2 4.5 34.4 35.0

b

T 8.7 T 4.7b T 6.1b T 3.4b T 2.6b T 3.3b T 2.1b T 6.3b T 10.8b T 9.4b

28.6*** 19.1*** 16.1*** 3.2* 10.3*** 6.5** 11.5*** 31.7*** 20.8*** 21.4***

Means with different letters are different from each other at p G 0.05; means with no letters are not different from any other mean in that row. *p G 0.05. **p G 0.01. ***p G 0.001. BD indicates body dissatisfaction; DT, drive for thinness; IA, interoceptive awareness; IR, impulse regulation; P, perfectionism; SI, social insecurity; STAI-S, State-Trait Anxiety InventoryYState; STAI-T, State-Trait Anxiety InventoryYTrait.

Twenty-eight (88%) of the 32 patients attempting suicide had a history of at least one comorbid DSM-IV axis I disorder in comparison with 31 (72%) of the 43 nonsuicidal patients (W21 = 0.11, NS). Twentysix suicide attempters (81%) had evidence of current or lifetime depressive disorder (major depressive disorder, dysthymic disorder, or depressive disorder NOS) in comparison with 23 nonsuicidal patients (54%; W21 = 6.2; p G 0.05). No significant between-group differences were reported for comorbid anxiety disorders (including panic disorder, social phobia, obsessive-compulsive disorder, generalized anxiety disorder, and posttraumatic stress disorder; W21 = 0.62, NS). Accordingly, 8 patients attempting suicide (25%) and 13 nonsuicidal patients (30%) showed evidence of one or more anxiety disorders. Duration of illness was significantly longer among the suicide attempters versus the nonsuicidal group (38.1 T 21.5 months vs. 25.2 T 15.0 months, respectively, t[73] = 4.5, p G 0.05). By contrast, no significant difference was shown in duration of inpatient treatment (7.4 T

4 months among the suicide attempters vs. 6.3 T 2.5 months in the nonsuicidal group, t[73] = 1.46, NS; inpatient treatment in this department includes full hospitalization and a subsequent day treatment program). All traumatic events found in our research sample were related to the occurrence of sexual abuse during childhood. Thus, six suicide attempters (19%) but none of the nonsuicidal group had evidence of childhood sexual abuse (W21 = 8.8; p G 0.01). Many of the variables assessed were correlated with one or more of the MAST at p G 0.05 in individual analyses (see Table 3). The use of hierarchal regression analyses for clinical variables (comorbidity with affective disorders, childhood sexual abuse, and duration of illness) and of stepwise regression analyses for psychometric variables (EAT-26, EDI-2, BIS, BDI, and STAI) revealed the following associations: Two variables were associated with MAST-AL (r2 = 0.539, p G 0.001): BDI (B = j0.283, p G .0.001) and BISYbody care (B = 0.484,

TABLE 3. Correlations of Psychometric and Clinical Variables With Attitudes to Life and Death EAT-26 EDI-2-DT EDI-2-BD EDI-2-IA EDI-2-SI EDI-2-IR BIS-AF BIS-BC BIS-BP BDI STAI-S STAI-T Comorbidity with affective disorders Childhood sexual abuse Duration of illness

MAST-AL

MAST-RL

MAST-AD

MAST-RD

j0.41*** j0.45*** j0.54*** j0.48*** j0.68*** j0.43*** 0.59*** 0.48*** 0.36*** j0.68*** j0.65*** j0.65*** j0.24* j0.17 (NS) j0.15 (NS)

0.46*** 0.47*** 0.55*** 0.49*** 0.72*** 0.52*** j0.61*** j0.25* j0.42*** 0.74*** 0.69*** 0.67*** 0.19 (NS) 0.25* 0.12 (NS)

0.33** 0.29** 0.37*** 0.33*** 0.47*** 0.41*** j0.40*** j0.21* j0.46*** 0.52*** 0.40*** 0.44*** 0.10 (NS) 0.20 (NS) 0.29*

0.08 (NS) 0.05 (NS) j0.03 (NS) j0.01 (NS) j0.14 (NS) 0.03 (NS) j0.08 (NS) 0.36*** 0.27** j0.1 (NS) j0.04 (NS) 0.01 (NS) j0.16 (NS) j0.20 (NS) j0.09 (NS)

*p G 0.05. **p G 0.01. ***p G 0.001. AF indicates attitudes and feelings; BC, body care; BD, body dissatisfaction; BP, body protection; DT, drive for thinness; IA, interoceptive awareness; IR, impulse regulation; SI, social insecurity; STAI-S, State-Trait Anxiety InventoryYState; STAI-T, State-Trait Anxiety Inventory–Trait

* 2013 Lippincott Williams & Wilkins

www.jonmd.com


1069


Stein et al.

TABLE 4. Stepwise Regression Model Based on Variables Differentiating Between the Suicidal and Nonsuicidal ED Patients B/P type EDs MAST-AD MAST-RD

OR

95% CI for OR

p

11.76 1.119 0.901

3.17Y43.48 1.016Y1.232 0.835Y0.973

0.0001 0.023 0.008

CI indicates confidence interval; OR, odds ratio.

p G 0.01). Specifically, reduced AL appeared in the context of greater depression and less care toward the body. Two variables were associated with MAST-RL (r2 = 0.575, p G 0.001): childhood sexual abuse (B = 4.242; p G 0.05) and BDI (B = 0.356, p G 0.001). Specifically, a history of childhood sexual abuse and elevated depression were associated with greater RL. MAST-AD has been associated with two variables (r2 = 0.472, p G 0.001): BDI (B = 0.229, p G 0.001) and BISYbody protection (B = j0.322, p G 0.05). This suggests that greater AD in patients with EDs may be found in the context of elevated depression and reduced protection of the body. Lastly, one variable, BISYbody care, has been associated with MAST-RD (r2 = 0.205, p G 0.001; B = 0.757, p G 0.01). Accordingly, lower RD appears in individuals with EDs who take less care of their body. Ten variables distinguished between the suicide attempters and the nonsuicidal group in individual analyses: BMI, duration of illness, ED type, BDI, MAST-AD, MAST-RD, BSIYattitudes and feelings, BISYbody protection, comorbid depressive disorders, and the occurrence of childhood sexual abuse. All these variables were entered into a stepwise regression analysis except for sexual abuse because it was found only in the suicidal group. According to Table 4, three variables were significantly associated with attempted suicide in EDs: the presence of B/P type EDs (i.e., AN-B/P, BN, or EDNOS-B/P), elevated MAST-AD, and reduced MAST-RD.

DISCUSSION The aim of the present study was to investigate the factors associated with the turning of ED patients from being afraid of life to showing pathological attitudes toward death and to attempting suicide. In line with our first hypothesis, both the ED patients attempting suicide and the nonsuicidal patients with EDs showed lower AL and elevated RL in comparison with the healthy control participants. Pathological attitudes to life were associated with elevated depression, reduced body care, and a history of childhood sexual abuse. In line with our second hypothesis, the ED patients who attempted suicide were different from either one or the two other groups in showing greater AD and less RD. Pathological attitudes toward death were associated with elevated depression and reduced body protection and body care. Lastly, in line with our third hypothesis, elevated MAST-AD and reduced MAST-RD were associated with attempted suicide in the patients with EDs, alongside one other variable, B/P type ED. The finding of maladaptive attitudes toward life in the present cohort of adolescent patients with AN-R, AN-B/P, BN, and ED-NOS is similar to the results of Bachar et al. (2002) in adult patients with AN and BN and to our previous findings in adolescent patients with AN (Stein et al., 2003). From a self-psychological perspective, the basic problem of patients with EDs may be conceptualized as a retreat from life in general, not merely from a specific potentially problematic life task (Bachar, 2001). This elevated fear of life may result from other characteristics interfering with choosing life as an age-appropriate developmental task that are found in both restricting and B/P ED patients, including elevated social anxiety (Hinrichsen et al., 2007), harm 1070

www.jonmd.com


avoidance (i.e., resistance to change; Fassino et al., 2002), and interpersonal distrust (Clausen et al., 2009). The association found in the present study in female adolescents and young adults with EDs between sexual abuse during childhood and RL during adolescence and young adulthood emphasizes the enormous influence of childhood abuse events on the overall development and well-being of young girls exposed to these events (Murray et al., 2008). Interestingly, we found that maladaptive attitudes to both life and death are associated with elevated depression and lower body care and/ or body protection. Still, the BDI score of the suicide attempters is significantly greater than that of the nonsuicidal group if only these two groups are compared, and more patients attempting suicide have a comorbid depressive disorder in comparison with nonsuicidal patients. Hence, elevated depression may increase the likelihood of ED patients, who are basically afraid of life, to attempt suicide (Bulik et al., 2008). With respect to body-related influences, it is of note that attitudes toward life and death in our study have not been associated with EDI-2 drive for thinness and EDI-2 body dissatisfaction. In line with the contention of Bruch (1973), this suggests that body-related attitudes and behaviors other than the core ED pursuit of thinness and preoccupation with weight and shape may be of relevance in the manner in which young women with EDs are involved with, or reject, basic life- and death-related wishes and behaviors. Rather, it is the lack of active care and protection of the body that may reduce adaptive and increase maladaptive attitudes to life and death. Similar to our findings in EDs, Orbach (1996) has shown in other psychiatric disturbances that a reactive distancing from the body in the form of negative feelings, attitudes, and behaviors toward the body, including reduced body care and body protection, may differentiate suicide attempters from nonsuicidal patients, irrespective of the psychiatric diagnosis (Orbach and Mikulincer, 1998; Orbach et al., 2001). Hence, overall negative relations toward the body may play an important role in facilitating suicidal behavior. Only two parameters, attitudes toward death and B/P ED type, have been associated with the occurrence of suicide attempts in ED patients (see Table 4). The greater occurrence of attempted suicide in ED patients with B/P spectrum disorders in comparison with restricting patients has been reported previously (Corcos et al., 2002; Favaro and Santonastaso, 1997; Franko and Keel, 2006; Stein et al., 2004). Still, it is of note that depression and bodily-related disturbances, previously shown to be associated with attempted suicide in EDs (Bulik et al., 2008; Corcos et al., 2002), seem in this study to be associated with it only indirectly, via their association with maladaptive death-related attitudes. Several limitations of our study should be noted. Firstly, because our design is cross-sectional, we can relate at present only to associations between maladaptive attitudes to life and death and suicidal behavior and not to directions of influence. Secondly, we have not included a psychiatric comparison group. Thirdly, because the number of participants is relatively small and our sample includes only inpatients, our findings cannot be generalized to other populations of ED patients. In conclusion, the present study identifies increased fear of life as a core determinant of EDs, regardless of their specific subtype, that is associated with other characteristics of these disorders, including elevated depression, reduced body care, and a history of childhood sexual abuse. On the other hand, only patients with EDs who have attempted suicide show maladaptive attitudes toward death, associated mainly with pathological bodily-related attitudes. The conceptualization of EDs as including a basic fear of life, with death wishes appearing only in suicidal patients, stands in sharp contrast to theories suggesting that unconscious wishes to die in the form of refusal to eat may be involved in the very development and maintenance of EDs, particularly AN (Jackson and Davidson, 1986). Moreover, the present study proposes an important addition to our previous study assessing attitudes to life and death in suicidal * 2013 Lippincott Williams & Wilkins




psychiatric patients, nonsuicidal psychiatric patients, nonsuicidal AN patients, and healthy controls (Stein et al., 2003). This study has shown that suicidal psychiatric patients differ from nonsuicidal psychiatric patients in their attitudes to both life and death. Accordingly, changes in how the individual perceives his/her life combined with elevated attraction and reduced fear toward death are necessary determinants in these patients in increased risk for suicide, irrespective of their psychiatric diagnosis (Orbach et al., 1991, 2001). The picture seems to be different in patients with EDs because only attitudes toward death differ between suicidal and nonsuicidal patients. Maladaptive attitudes to life, on the other hand, emerge as a vulnerability factor in ED patients of all subtypes, interfering with their overall development during adolescence, rather than increasing their inclination toward suicidal behavior. Future research of the role of attitudes to life and death in the course and outcome of EDs should be longitudinal, carried out in a larger cohort of ambulatory ED patients from the identification of the ED until achieving recovery. Such a design would enable to determine the role of these constructs not only with respect to suicidal behavior in EDs but also in the overall prognosis of these disorders. DISCLOSURE The authors declare no conflict of interest. REFERENCES American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders (4th ed). Washington, DC: American Psychiatric Association. Bachar E (2001) The fear of occupying space. Jerusalem: The Hebrew University Magnes Press. Translated from Hebrew by M. Schwartz. Bachar E, Latzer Y, Canetti L, Gur E, Berry EM, Bonne O (2002) Rejection of life in anorexic and bulimic patients. Int J Eat Disord. 31:43Y48. Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J (1961) An inventory for measuring depression. Arch Gen Psychiatry. 4:561Y571. Bray GA (1992) An approach to the classification and evaluation of obesity. In Bjorntorp P, Brodoff BN (Eds), Obesity (pp 294Y308). Philadelphia: Lippincott. Brown MZ, Comtois KA, Linehan MM (2002) Reasons for suicide attempts and nonsuicidal self-injury in women with borderline personality disorder. J Abnorm Psychol. 111:198Y202. Bruch H (1973) Eating disorders: Obesity, anorexia nervosa and the person within. New York: Basic Books. Bulik CM, Thornton L, Pinheiro AP, Plotnicov KH, Klump KL, Brandt H, Crawford S, Fichter MM, Halmi KA, Johnson C, Kaplan AS, Mitchell J, Nutzinger D, Strober M, Treasure J, Woodside DB, Berrettini WH, Kaye WH (2008) Suicide attempts in anorexia nervosa. Psychosom Med. 70:378Y383. Clausen L, Rokkedal K, Rosenvinge JH (2009) Validating the Eating Disorder Inventory (EDI-2) in two Danish samples: A comparison between female eating disorder patients and females from the general population. Eur Eat Disord Rev. 17:462Y467. Corcos M, Taieb O, Benoit-Lamy S, Paterniti S, Jeammet P, Flament MF (2002) Suicide attempts in women with bulimia nervosa: Frequency and characteristics. Acta Psychiatr Scand 106:381Y386. Crisp AH (1980) Anorexia nervosa: Let me be. London: Academic Press. Fassino S, Abbate-Daga G, Amianto F, Leombruni P, Boggio S, Rovera GG (2002) Temperament and character profile of eating disorders: A controlled study with the Temperament and Character Inventory. Int J Eat Disord. 32:412Y425.

* 2013 Lippincott Williams & Wilkins


Favaro A, Santonastaso P (1997) Suicidality in eating disorders: Clinical and psychological correlates. Acta Psychiatr Scand. 95:508Y514. First MB, Spitzer RL, Gibbon M, Williams JB (1995) Structured clinical interview for axis-I, DSM-IV disorders: Patient edition (SCID-I/P, Version 2.0). New York: New York State Psychiatric Institute, Biometrics Research Department. Franko DL, Keel PK (2006) Suicidality in eating disorders: Occurrence, correlates, and clinical implications. Clin Psychol Rev. 26:769Y782. Garner DM (1991) The Eating Disorder Inventory-2 (EDI-2). Odessa, FL.: Psychological Assessment Resources Inc. Garner DM, Olmsted MP, Bohr Y, Garfinkel PE (1982) The Eating Attitudes Test: Psychometric features and clinical correlates. Psychol Med. 12:871Y878. Goldstein A, Haelyon U, Krolik E, Sack J (2001) Comparison of body weight and height of Israeli schoolchildren with the Tanner and Centers for Disease Control and Prevention growth charts. Pediatrics. 108:E108. Hinrichsen H, Sheffield A, Waller G (2007) The role of parenting experiences in the development of social anxiety and agoraphobia in the eating disorders. Eat Behav. 8:285Y290. Jackson C, Davidson GP (1986) The anorexic patient as a survivor: The denial of death and death themes in the literature on anorexia nervosa. Int J Eat Disord. 5:821Y835. Koslowsky M, Scheinberg Z, Bleich A, Mark M, Apter A, Danon Y, Solomon Z (1992) The factor structure and criterion validity of the short form of the Eating Attitudes Test. J Pers Assess. 58:27Y35. Murray CD, Macdonald S, Fox J (2008) Body satisfaction, eating disorders and suicide ideation in an Internet sample of self-harmers reporting and not reporting childhood sexual abuse. Psychol Health Med. 13:29Y42 Orbach I (1996) The role of body experience in self-destruction. Clin Child Psychol Psychiatry. 1:607Y619. Orbach I, Mikulincer M (1998) The Body Investment scale: Construction and validation of a body experience scale. Psychol Assess. 10:415Y425. Orbach I, Milstein I, Har-Even D, Apter A, Tyano S, Elizur A (1991) A MultiAttitude Suicide Tendency scale for adolescents. Psychol Assess. 3:398Y404. Orbach I, Stein D, Shani-Sela M, Har-Even D (2001) Body attitudes and body experiences in suicidal adolescents. Suicide Life Threat Behav. 31:237Y249. Pollice C, Kaye WH, Greeno CG, Weltzin TE (1997) Relationship of depression, anxiety, and obsessionality to state of illness in anorexia nervosa. Int J Eat Disord. 21:367Y376. Spielberger CD, Gorsuch R, Lushene R (1970) The State Trait Anxiety Inventory (STAI). Palo Alto, CA: Consulting Psychologists Press. Stein D, Lilenfeld LR, Wildman PC, Marcus MD (2004) Clinical features that distinguish parasuicidal from non-parasuicidal eating disorders patients: A retrospective chart design. Compr Psychiatry. 45:447Y451. Stein D, Orbach I, Shani-Sela M, Har-Even D, Yaroslavsky A, Roth D, Meged S, Apter A (2003) Suicidal tendencies and body image and experience in suicidal and anorexia nervosa female adolescent inpatients. Psychother Psychosom. 72:16Y25. Strober M (1987) The Eating Disorders Family History Interview. Los Angeles: University of California. Strober M, Freeman R, Morrell W (1997) The long-term course of severe anorexia nervosa in adolescents: Survival analysis of recovery, relapse and outcome predictors over 10-15 years in a prospective study. Int J Eat Disord. 22;339Y360. Tyano S, Vincent M (1997) L’adolescent, la vie et la mort, le quatrieme organisateur. Psychiatr Fr. 4:131Y150. Walsh WB, Rosen PM (1998) Self mutilation. New York: Guilford Press. Yackobovitch-Gavan M, Golan M, Valevski A, Kreitler S, Bachar E, Lieblich A, Mitrani E, Weizman A, Stein D (2009) An integrative quantitative model of factors influencing the course of anorexia nervosa over time. Int J Eat Disord. 42:306Y317.

www.jonmd.com


1071

Health-related quality of life among adolescents with eating disorders.

Family meal frequency among children and adolescents with eating disorders.

Risk of suicide among female adolescents with eating disorders: a longitudinal population-based study.

Attitudes toward lesbians and gays among American and Dutch adolescents.

Hypnotizability, attitudes toward eating, and concern with body size in a female college population.

Attitudes towards psychotherapy manuals among clinicians treating eating disorders.

Attitudes Toward Male and Female Sexuality Among Men and Women With Intellectual Disabilities.

Body Image Satisfaction, Eating Attitudes and Perceptions of Female Body Silhouettes in Rural South African Adolescents.

Picking and nibbling in children and adolescents with eating disorders.

Adolescents' stigma attitudes toward internalizing and externalizing disorders: Cultural influences and implications for distress manifestations.

Attitudes toward psychotropic medications among patients with chronic psychiatric disorders and their family caregivers.

Purging using the Heimlich maneuver among children and adolescents with eating disorders.

Comparison between the SCOFF Questionnaire and the Eating Attitudes Test in patients with eating disorders.

Attitudes of medical professionals towards patients with eating disorders.

Sex differences in attitudes towards females with eating disorders.

Attitudes of adolescents toward the aged.

Eating Disorders in Late-life.

Adolescents with sickle-cell anemia deal with life and death.

Prevalence and severity of categorical and dimensional personality disorders in adolescents with eating disorders.

Factors associated with substance use in adolescents with eating disorders.

Attitudes toward homosexuality among West Indian male and female college students.

Adverse childhood experiences as risk factors for serious mental disorders and inpatient hospitalization among adolescents.

Eating Disorders and Disordered Eating Symptoms in Adolescents with Type 1 Diabetes.

Attitudes toward dying and death: a comparison of recreational groups among older men.