|Home | About | Journals | Submit | Contact Us | Français|
Assessment of personality disorders during the acute phase of major depression may be invalidated by the potential distortion of personality traits in depressed mood states. However, few studies have tested this assumption. We examined the stability of personality disorder diagnoses during and then after a major depressive episode (MDE). Subjects with major depression (N = 82) completed the 17-item Hamilton Depression Scale (HAM-17) and the Structured Clinical Interview for Axis II both at baseline during an MDE and at 3-month follow-up. We compared subjects who continued to meet DSM-IV criteria for the same Axis II diagnoses with patients whose diagnosis changed and patients with no DSM-IV personality disorder to determine the relationship to major depression and its severity. Sixty-six percent of subjects met DSM-IV criteria for at least one Axis II diagnosis at baseline and 80% had the same personality disorder diagnoses at follow-up. Thirty-four percent had a full remission of MDE at 3-month follow-up. Instability of Axis II diagnosis was associated with number of Axis II diagnoses at baseline (p = .036) and Hispanic ethnicity (p = .013). HAM-17 score change was unrelated to differences in the number of symptoms of personality disorders from baseline to follow-up, nor was remission from MDE on follow-up. Axis II diagnoses in acutely depressed patients reassessed after 3 months are often stable and not associated with remission of or improvement in major depression.
Personality disorders (PDs) frequently coexist with mood disorders. Reported rates of comorbidity vary between 12% and 95% (Bagby et al., 2008; Fan and Hassell, 2008), with the average prevalence of PDs being approximately 50% in major depressive disorder (MDD) (George et al., 2003; Grant et al., 2005) and 40% in bipolar disorder (BD) (Fan and Hassell, 2008; George et al., 2003). Comorbid PDs are associated with poorer prognosis of mood disorders, including increased risk for suicidal behavior (Rihmer, 2007), higher rates of relapse and chronicity, and worse treatment response (George et al., 2003; Newton-Howes et al., 2006). Diagnosis of comorbid PDs in individuals with mood disorders may improve treatment outcome and risk management (Leichsenring and Leibing, 2003; Newton-Howes and Tyrer, 2003).
Making a diagnosis of a PD during a major depressive episode (MDE) is debated in terms of validity (De Fruyt et al., 2006; Santor et al., 1997); it is common practice to defer assessment and diagnosis of PDs during acute depression. Depressed states may alter both the patients’ and the clinicians’ perception of personality features (Hirschfeld et al., 1983), and as stated in DSM-IV-TR, “the enduring pattern” of a PD should not be “better accounted for as a manifestation or consequence of another mental disorder” (American Psychiatric Association, 2000, p. 689). Moreover, interpersonal functioning is a domain frequently compromised among depressed patients, and the emphasis placed on interpersonal relationships as key to the diagnosis of Axis II disorders has reinforced this practice of deferring Axis II diagnosis.
Earlier research has examined this issue. Studies based on self-rating scales reported that fewer patients meet criteria for PD diagnoses after treatment for depression with antidepressants (Fava et al., 1994; Joffe and Regan, 1988) or electroconvulsive therapy (Blais et al., 1998). However, diagnoses based on self-reported questionnaires may be more dependent on clinical state than clinical interviews are (Zimmerman et al., 1991), and results from studies of stability of PDs using clinical interviews are inconsistent. Loranger et al. (1991) found that anxiety and depression did not influence the diagnoses of PDs when a clinical interview based on DSM-III-TR was used. On the other hand, Fava et al. (2002) reported a consistent trend for lower Structured Clinical Interview for Axis II (SCID-II) scores across all PDs after an 8-week follow-up with antidepressant treatment. However, when they corrected for multiple comparisons, significance was lost, except for avoidant PD. In a subsample of 75 subjects, they found that comorbid PD diagnoses remained stable throughout the continuation phases of treatment even after remission (Farabaugh et al., 2002). More recently, Morey et al. (2010) used 6-year outcome data in a large sample of patients with and without comorbid MDD to study the stability of borderline, schizotypal, obsessive-compulsive, and avoidant PDs. They found no significant differences in the stability of personality assessments for PD patients with and without MDD at baseline.
We examined baseline and follow-up Axis I and Axis II diagnostic data to investigate the stability of Axis II diagnoses made during a baseline acute MDE on reassessment 3 months later. The present work continues previous research studying a) the stability of PD diagnoses and symptom scores made in acute depression as measured by SCID-II and b) the degree to which changes in either presence or severity of depression are related to symptoms and diagnostic changes in PD diagnoses from baseline to follow-up.
A total of 82 subjects with a current MDE participated in the study after giving written informed consent as required by the institutional review board. Subjects had been referred to a research clinic for evaluation of MDEs or other major psychiatric disorders. Half of the subjects were outpatients on admission to the study. Subjects were assessed by psychiatrists or clinical psychologists at baseline and at 3-month follow-up. At baseline and follow-up, subjects were diagnosed according to DSM-IV criteria for Axis I diagnoses using the SCID-I (Patient or Non-Patient edition; First et al., 2002) and for PDs using SCID-II (First et al., 1996). A best estimate diagnosis procedure was used to determine both Axis I and II diagnoses. Consensus conference meetings composed of senior clinical investigators and master’s level or above clinical raters considered all data from the patient, medical records, and family informants. Primary lifetime Axis I diagnoses were 57.3% (n = 47) MDD and 42.7% (n = 35) BD. Secondary lifetime diagnoses were 42.7% (n = 35) anxiety disorders, 24.4% (n = 20) posttraumatic stress disorder (PTSD), 17.1% (n = 14) dysthymia, 12.2% (n = 10) obsessive-compulsive disorder, and 39.0% (n = 32) other. Severity of depression was rated at baseline and follow-up using the clinician-administered 17-item Hamilton Depression Scale (HAM-17) (Hamilton, 1960). The same clinician conducted baseline and follow-up assessments in 87.8% of the cases. Of the subjects, 65.8% (n = 54) had at least one Axis II diagnosis at baseline. At follow-up, 34.1% (n = 28) of subjects no longer met diagnostic criteria for MDE. Of the participants, 67.1% were female, and average age was 38.6 ± 11.6 years.
We defined stability as the maintenance of all Axis II diagnoses within each subject from baseline to follow-up, using only a forward measure. This offers a more conservative measure of stability when compared with previous reports (Farabaugh et al., 2002; Fava et al., 2002) that considered stability as maintaining at least one stable diagnosis at follow-up. Depression remission was defined as a HAM-17 score of 7 or lower at follow-up.
We classified subjects into three groups: no Axis II diagnosis at baseline or follow-up (no Axis II group, n = 28), stable Axis II diagnosis (i.e., Axis II baseline diagnosis present at follow-up assessment, n = 43), and at least one Axis II diagnosis changed from baseline to follow-up assessment (nonstable Axis II group, n = 11). The latter included six subjects who had three Axis II diagnoses at baseline (one of them lost two and one lost three diagnoses at follow-up), two subjects who had two Axis II diagnoses at baseline (both lost one diagnosis at follow-up), and three subjects with one diagnosis at baseline. We compared baseline demographics and baseline and follow-up clinical characteristics between the three groups using Fisher’s exact test for categorical variables because of cell counts smaller than 5 (the Freeman-Halton extension was used for 2 × 3 and 3 × 3 tables) and analysis of variance for continuous variables. Severity of PD diagnoses was examined according to Tyrer and Johnson (1996). Depressive PD was not included in the analyses.
For subjects with at least one PD diagnosis, we applied Fisher’s exact test and chi-square test to examine if stability of Axis II diagnoses varied with depression remission at follow-up and number of PDs at baseline. Lastly, to evaluate the relationship between depression scores and personality assessment, we compared the change from baseline to follow-up in HAM-17 scores and in the number of PD symptoms using a Spearman nonparametric correlation test because PD score differences were not normally distributed. Change from baseline to follow-up in the number of PD symptoms was compared using a Mann-Whitney U test between patients whose depression remitted and patients whose depression did not remit. For 21 cases, only positive baseline diagnoses or subthreshold scores of PDs were reassessed at follow-up. PD score differences were not considered when follow-up assessment was missing.
Table 1 provides details of demographic comparisons between the stable Axis II, at least one unstable Axis II, and no Axis II groups. Stability did not vary depending on whether the assessment was done by the same rater or not (p = .910). No differences in sex, age, marital status, education, or household income were found between the groups. Subjects were mostly unmarried (83%) non-Hispanic Whites (64%), with an average of 15.6 years of education. Hispanic ethnicity was associated with a higher risk of unstable PD diagnoses (p = .013).
Table 2 provides details of the comparison of baseline clinical characteristics and change in clinical status at follow-up assessment of the three groups. Mean number of previous MDE was 2.6 ± .7 and mean number of PDs was 1.0 ± .9. No significant differences were found between the groups in terms of comorbid substance use disorders, inpatient status, number of previous MDE, current diagnosis, and HAM-17 scores at baseline or follow-up. Subjects with PDs were more likely to have additional comorbid Axis I diagnoses (p = .015), in particular anxiety disorders (p = .001), although not PTSD.
At follow-up, 40% of the subjects had one (n = 33), 17% had two (n = 14), and 7% had three (n = 6) Axis II diagnoses. Thirty-five had no Axis II diagnosis at follow-up (n = 29). Among those who met criteria for a minimum of one PD, 20.4% (n = 11) had at least one Axis II diagnosis that changed from baseline to follow-up assessment. The rate of instability was significantly different among subjects who had three (100%; 6/6), two (12.5%; 2/16), or one (9.3%; 3/32) baseline PD diagnoses (Fisher’s exact test, p = .001). No significant differences in instability appeared when subjects who had one or two baseline PD diagnoses were compared (Fisher’s exact test, p = 1.00). Results on Axis II diagnoses by cluster at baseline and follow-up can be seen in Table 3.
A total of 28 subjects (28/82; 34%) had remitted from baseline MDE at 3-month follow-up. Improvement in the HAM-17 score from baseline (18.7 ± 5.2) to follow-up (11.4 ± 7.2) was, on average, 7.3 ± 8.3 points (Table 2). The number of PDs at baseline did not affect the risk of the subject remitting afterward (χ2 = .229, df = 1, p = .632). Correlations between HAM-17 score differences and differences in the number of SCID-II criteria met by diagnosis from baseline to follow-up were nonsignificant for all PDs. However, the number of criteria met decreased from baseline to follow-up for every PD except not otherwise specified PD, in particular for borderline PD and obsessive-compulsive PD (average of .68 and .56, respectively).
The present study examined the pattern of change in PD diagnoses and diagnostic criteria being met during an MDE and then after 3 months of treatment. In agreement with previous studies (Fava et al., 2002; Hirschfeld, 1999), depressed subjects were frequently assigned at least one Axis II diagnosis at baseline. However, the results show that PD diagnoses were, by and large, stable between baseline and follow-up. Furthermore, the remission of the MDE at the 3-month assessment was not associated with instability of PD diagnoses. However, greater score changes in the HAM-17 from baseline to follow-up was associated with meeting fewer Axis II diagnostic criteria. Interestingly, Hispanic ethnicity was a significant factor in the diagnostic stability of PDs, suggesting that cultural expression may mediate the way depressive episodes alter personality assessment.
Similar to previous studies (Fava et al., 2002; Loranger et al., 1991; Melartin et al., 2010; Mulder et al., 2010), an overall decrease in number of PD diagnostic criteria met was found at follow-up. No correlation was found between HAM-17 score change and PD criteria met change from baseline to follow-up. It should be noted that PDs are a heterogeneous group defined by amalgams of traits and symptoms (Skodol et al., 2005), and some PDs and personality traits are more associated with depressed mood than others are (Fava et al., 2002; Vliegen et al., 2010). Our results are in agreement with those of Morey et al. (2010), although they studied only four PDs and used different assessment instruments. They contradict, however, using a more conservative measure of stability, the study by Fava et al. (2002), which reported that changes in personality traits after antidepressant treatment were associated with a decline in HAM-17 scores in some PDs and that a significant proportion of the subjects were unstable. Low to moderate stability of PD diagnoses was also reported by Melartin et al. (2010) and Mulder et al. (2010) in samples of depressed patients reassessed after longer interval periods (up to 18 months). Besides, and in contrast with an earlier study (Candrian et al., 2008), we observed no differences between clusters in the accuracy of PD diagnosis during an MDE. Of note, diagnostic stability may be related to the qualification of the raters (Candrian et al., 2008; Fava et al., 2002; Loranger et al., 1991). On the other hand, the limitations of DSM-IV criteria to characterize severe PDs (Tyrer and Johnson, 1996) could be reflected in the association of instability with multiple PD diagnoses. In addition, severity of PDs was close to being significantly associated with instability in our sample, and this association may have been significant in later assessments. However, the evolution of personality features across the lifespan should be considered in longer-term assessments, which found PD diagnoses to be generally unstable (Tyrer et al., 2007).
This study has several limitations. The same assessor made baseline and 13-week reassessments in most of the cases. This strategy minimizes interrater variance but may introduce a bias to maintain previous diagnoses. Personality evaluations used the SCID-II, and the findings may not generalize to other instruments. Furthermore, the small sample size may have increased the risk of type 2 error. It should also be considered that subjects may underreport maladaptive behavior in repeated measures because of knowledge of the questions and interest in accelerating the interview (Loranger et al., 1991). Finally, results were not adjusted by type of pharmacological treatment because of its heterogeneity.
Most mood disorder patients are able to provide reliable and accurate information regarding the cognitive, behavioral, and affective domains used to assess Axis II diagnostic criteria. The evaluation of PDs during acute depressive episodes might provide valuable information and promote more effective treatment.
Paloma Aroca has provided editorial assistance for the publication of this manuscript.
Dr Lopez-Castroman was supported by Koplowitz Foundation and Instituto de Salud Carlos III research grants. The study was funded by the Centro de Investigación Biomédica en Red de Salud Mental, CIBERSAM. The Spanish Ministry of Health and the organizations cited above had no further role in study design; in the collection, analysis, and interpretation of data; in the writing of the report; and in the decision to submit the paper for publication.
Dr Mann has received unrelated grants from GSF and Novartis. Dr Oquendo has received funding from National Institute of Mental Health, National Institute on Alcohol Abuse and Alcoholism, American Foundation for Suicide Prevention, Moody’s Foundation, as well as an unrestricted educational grant from Eli Lilly, and has served as a consultant to Pfizer.