اختلالات شخصیت در یک گروه مجموعه جمعیت دوقلو با اختلالات خوردن

Abstract Findings regarding the occurrence of personality disorders (PDs) in eating disorders (EDs) have been contradictory. Most previous studies have been clinic-based. The aims of the current study were to assess the prevalence of PD in ED in a population-based twin group and to establish the distribution of PD in three subgroups of ED. A two-step screening and diagnostic study of ED was performed in a large Danish twin population. Axis I and axis II DSM-III-R and DSM-IV ED diagnoses were made on the basis of results obtained at clinical investigations and interviews. Forty-nine percent of the participants with ED had at least one PD, compared to 26% in those with no ED (P < .001). Cluster C PD was the most common type of PD in all subgroups of ED, and cluster B PD was found only in participants with bulimic symptoms. Genetic factors appeared to contribute significantly to the variance of cluster C PD in ED, which was evaluated as a possibly important background factor in ED. THE COMORBIDITY between personality disorders (PDs) and eating disorders (EDs) has been described in numerous studies. Table 1 provides a summary of some of the most important studies in the field. There is limited evidence that cluster C PD may be particularly common in anorexia nervosa (AN), whereas cluster B PD may be more characteristic of bulimia nervosa (BN). The interpretation of this comorbidity is complicated by the use of different methods and study populations.1, 2, 3 and 4 In general, the studies are of limited size, and comorbidity varies form one study to another whether or not the entire spectrum of ED has been included.

Results The prevalence of PD was established in 187 twins, including 63 participants with ED. Details are given in Table 2. Table 2. DSM-III-R Personality Disorders in a Twin Cohort With Eating Disorders (N = 187) AN (n = 20) AN/BN (n = 20) BN (n = 23) ED (n = 63) No ED (n = 124) Thr Subthr Total % Thr Subthr Total % Thr Subthr total % Total % Thr Subthr Total % Any PD 10 50∗ 10 50∗ 11 47.8 (NS) 31 49.2∗ 32 25.8 Any cluster A 2 10 (NS) 2 10 (NS) 4 17.4 (NS) 8 12.7 (NS) 8 6.5 Paranoid 1 1 1 1 2 1 2 3 6 2 4 6 Schizotypal Skizoid 1 1 1 1 1 1 3 1 2 3 Any cluster B 0 0 (NS) 4 20 (NS) 4 17.4 (NS) 8 12.7 (NS) 9 7.3 Histrionic 1 1 1 1 2 1 1 Narcissistic 1 1 1 1 2 3 1 1 Borderline 2 2 4 1 2 3 7 5 5 Antisocial 3 3 Any cluster C 9 45† 7 35∗ 9 39.1∗ 25 40.0† 20 16.1 Avoidant 4 4 3 3 3 3 10 3 4 7 Dependent 1 1 3 3 1 4 5 9 2 2 Obsessive-compulsive 6 6 3 3 2 5 7 16 10 10 Passive 2 2 4 Aggressive Self-defeating 0 0 1 1 4.3 1 1.6 3 3 2.4 Abbreviations: Thr, threshold; Subthr, subthreshold; NS, not significant. ∗ P = .05 † P = .01; ED subgroups v no ED. Table options In total, PD was found in 49% of individuals with ED (31 of 63) and in 26% of individuals without ED (32 of 124, P < .001). It is important to stress that, due to the selection procedure, the group without ED cannot be seen as a “control” group; they were not randomly selected, but instead were included because they and their cotwin were suspected of having an ED. Cluster C PD was the most frequently occurring PD in all three subgroups of ED, with a tendency towards more obsessive-compulsive PD, followed by avoidant and dependent PD. No cases of passive-aggressive PD were found. Cluster B PD was only found in the two subgroups with bulimic symptoms. There was no borderline or other cluster B PD in the AN-only subgroup. In DSM-IV ED subgroups, 17 of 20 with AN only were diagnosed as AN-restricting type and the remaining three as AN-binge-eating/purge type. Three of the AN/BN subgroup was diagnosed as having AN-restricting with later normal-weight BN. The remaining 17 were diagnosed as AN-binge-eating/purge type, a subgroup consisting of 20 individuals. There were no cluster B PDs in the AN-restricting group, and the group with AN-restricting and later normal-weight BN was too small to determine whether it differs from the AN-binge-eating/purge type. In both the AN-restricting and AN-binge-eating/purge type, cluster C PD was the most frequent. For individuals with ED, the OR for any PD was 2.91 (95% confidence interval [CI], 1.46 to 5.80; df = 1, P < .005) There was no significant difference between twins with and without ED for cluster A and cluster B PD, whereas for cluster C PD, a significant difference between individuals with and without ED was found, with ORs ranging from 2.80 to 4.25 depending on the subgroup of ED (Fig 2). OR for DSM-III-R cluster C PD in EDs as compared with non-Eds (95% CI). Fig 2. OR for DSM-III-R cluster C PD in EDs as compared with non-Eds (95% CI). Figure options The prevalence of cluster B PD in the AN/BN and BN-only subgroups was not significantly elevated compared to the group without ED, when tested separately. However, in the collapsed group of individuals with bulimic symptoms, a higher prevalence of cluster B PD, compared to the group without ED, was found (χ2 = 4.47, df = 1, P = .002). In the ED group, all with multiple PDs (n = 6) had a cluster B PD, and five of six had borderline PD. Conversely, in the group without ED, neither of the individuals with multiple PDs had a cluster B PD. All participants in the ED group with multiple PD had bulimic symptoms (Table 3). Table 3. Numbers of Personality Disorders in Twins With and Without Eating Disorder (N = 187) No. of PDs ED (n = 63) No ED (n = 124) No. % No. % 0 32 50,8 92 74,2 1–2 25 39,7 30 24,2 3–5 6 9,5 2 1,6 Table options In three males definite/probable BN was found. One of these also had cluster C PD. No ED was found in the remaining males. In the total group of twin pairs there was a correlation of cluster C personality traits of 0.47, (P = .00002) in the MZ pairs and of 0.13 (P = .281) in the DZ pairs. In the group of pairs with at least one twin partner affected with ED, the correlation in MZ twins was 0.46 (P = .00059) and in DZ twins 0.07 (P = .716). In the group of twins without ED the correlation in MZ twins was 0.49 (P = .02) and in DZ pairs 0.14 (P = .403). The correlations were substantially higher in MZ than in DZ twin pairs, although the 95% CIs did overlap to some degree (Table 4). Table 4. Correlation Coefficient of Cluster C Personality Traits in Twins With and Without ED Zygocity Correlation Coefficient (r) 95% CI P Total MZ 0.47 0.27 to 0.63 .00002 DZ 0.13 −0.11 to 0.36 .281 Pairs with ED MZ 0.46 0.21 to 0.65 .00059 DZ 0.07 −0.29 to 0.41 .716 Pairs without ED MZ 0.49 0.09 to 0.76 .020 DZ 0.14 −0.19 to 0.45 .403 Table options When we looked at twins discordant for ED we found, in accordance with our hypothesis, that in MZ but not in DZ pairs both twins were equally affected by cluster C personality traits. Figure 3 shows that in 20 MZ pairs discordant for ED, the twins were equally affected by pathologic personality traits, but in the 16 DZ twin pairs discordant for ED, the cluster C personality traits usually were seen only in the twin affected by ED. Cluster C personality traits in twins discordant for ED. Fig 3. Cluster C personality traits in twins discordant for ED. Figure options Attrition analyses The validity of the screening questionnaire was examined. Register linkage to the National Register of Medical Admissions (NMR) and the National Register of Psychiatric Admissions (PNR) was performed. It was estimated that in the total twin population 61% (NMR) and 68% (PNR), respectively, had escaped detection. (J. Joergensen et al., submitted paper). In a small group of patients with a clinic discharge diagnosis of ED who were not identified via the screening process, there was a tendency towards increased occurrence of PD, social problems and psychoses (L. Kortegaard, personal communication, 2000), which might have contributed to withdrawal of a small number of severely impaired individuals with ED from the present study. The twins who had been admitted according to NMR or PNR with a diagnosis ED, and who also participated in the study, were all screened positive and were clinically diagnosed as having an ED. There was no significant difference in the rate of participation in the clinical study across ED screen-positive and ED screen-negative individuals.