مقایسه طولی اختلال شخصیت افسرده و اختلال افسرده خویی

Abstract Background Few studies have compared the related diagnostic constructs of depressive personality disorder (DPD) and dysthymic disorder (DD). The authors attempted to replicate findings of Klein and Shih in longitudinally followed patients with personality disorder or major depressive disorder (MDD) in the Collaborative Longitudinal Personality Disorders Study. Methods Subjects (N = 665) were evaluated at baseline and over 2 years (n = 546) by reliably trained clinical interviewers using semistructured interviews and self-report personality questionnaires. Results Only 44 subjects (24.6% of 179 DPD and 49.4% of 89 early-onset dysthymic subjects) met criteria for both disorders at baseline. Depressive personality disorder was associated with increased comorbidity of some axis I anxiety disorders and other axis II diagnoses, particularly avoidant (71.5%) and borderline (55.9%) personality disorders. Depressive personality disorder was associated with low positive and high negative affectivity on dimensional measures of temperament. Depressive personality disorder subjects had lower likelihood of remission of baseline MDD at 2-year follow-up, whereas DD subjects did not. The DPD diagnosis appeared unstable over 2 years of follow-up, as only 31% (n = 47) of 154 subjects who had DPD at baseline and also had follow-up assessment met criteria on blind retesting. Limitations Results from this sample may not generalize to other populations. Conclusions Depressive personality disorder and dysthymic disorder appear to be related but differ in diagnostic constructs. Its moderating effect on MDD and predicted relationship to measures of temperament support the validity of DPD, but its diagnostic instability raises questions about its course, utility, and measurement.

Results The CLPS sample has been described previously (eg, [19], [21], [23] and [27]). Table 1 indicates that subjects with DPD were mainly female, White, with mean 31.6 (SD, 8.2) years, characteristics similar to non-DPD subjects in the sample. Compared to other subjects in the CLPS cohort, DPD subjects did not differ in gender or race, but were 1.6 years younger (t = 2.22, df = 663, P = .03). Table 1. Demographics of subjects with and without DPD Subjects with DPD Subjects without DPD All subjects N % N % N % Age (mean)* 179 31.6 (SD, 8.2); range, 18-45 486 33.2 (SD, 8.0); range, 18-45 665 32.8 (SD, 8.1) Gender Female 110 61.5 312 64.2 422 63.5 Male 69 38.6 174 35.8 243 36.5 Ethnicity Euro-American 129 72.1 377 77.6 506 76.1 African American 24 13.4 54 11.1 78 11.7 Hispanic 21 11.7 40 8.2 61 9.2 Asian 4 2.2 9 1.9 13 2.0 Other 1 0.6 6 1.2 7 1.1 * t = 2.22, P = .027. Table options 3.1. Diagnostic overlap of DPD and DD Data were available on 665 CLPS subjects, of whom 179 met definite DIPD criteria for DPD and 89 met criteria for early onset DD. Diagnostic overlap was relatively modest: only 44 subjects, representing 24.6% of subjects with DPD and 49.4% of subjects with DD, met criteria for both definite DIPD-IV DPD and early-onset DD (κ = 0.18; χ2 = 26.5, df = 1, both P < .0001). The alternative definitions of DPD and DD described above did not alter these fundamental relationships, showing marginally lower statistical associations. For example, for definite DPD criteria and DD of any age of onset (n = 115), κ = 0.16, χ2 = 19.4, df = 1 (both P < .001). A t test with unequal variances to compare subjects with and without early-onset DD with respect to the continuous DPD measure was significant (t = −6.77, df = 146, P < .001): early-onset DD subjects scored a mean 10.9 (SD, 2.9) and non-DD subjects a mean 8.5 (SD, 4.0) of a total possible score of 0 to 14 for 7 DPD items. 3.2. Comorbidity of DPD with other axis II disorders Assessing other axis II comorbidity using definite DIPD-IV diagnoses yielded statistically significant associations of DPD with 9 of the 11 axis II diagnoses, excepting only narcissistic and obsessive-compulsive personality disorders. Table 2 presents the proportion of subjects with personality disorders among subjects with and without definite DPD. As in the study of Klein and Shih [9] and similar to that of McDermut et al [10], DPD had strongest comorbidity with avoidant (71.5%) and borderline (55.9%) personality disorders. Results using definite or probable DPD were similar (available upon request). Table 2. Comorbidity of DPD with other axis II diagnoses at baseline Axis II diagnosis DPD present (%) DPD absent (%) χ2 (df = 1) Odds ratio (95% CI) κ Paranoid 22.4 8.4 24.2 (P < .0001) 3.16 (1.96-5.08) 0.17 Schizotypal 20.0 12.0 5.43 (P < .02) 1.71 (1.09-2.70) 0.08 Schizoid 5.6 1.6 7.9 (P < .005) 3.57 (1.39-9.20) 0.06 Histrionic 3.91 1.22 4.9 (P = .03) 3.29 (1.09-9.93) 0.04 Borderline 55.9 28.7 42.0 (P < .0001) 3.14 (2.21-4.48) 0.24 Narcissistic 6.15 5.09 0.29 (P = .6) 1.22 (0.59-2.54) 0.01 Antisocial 12.9 5.3 11 (P = .0009) 2.64 (1.46-4.76) 0.10 Avoidant 71.5 39.7 53.1 (P < .0001) 3.81 (2.63-5.52) 0.25 Dependent 14.5 4.7 18.7 (P < .0001) 3.46 (1.92-6.24) 0.13 Obsessive/compulsive 44.7 36.7 3.6 (P = .06) 1.40 (0.99-1.98) 0.07 Passive/aggressive 17.9 4.7 30.3 (P < .0001) 4.43 (2.51-7.81) 0.17 Table options We also examined axis I comorbidity of DPD and DD subjects. Neither diagnosis was associated with an increased prevalence of lifetime MDD relative to the remainder of the CLPS cohort (DPD = 78%, DD = 82%, others = 78%). Depressive personality disorder subjects had elevated lifetime rates of bipolar II disorder (7.3% vs 4.2% for the cohort, P < .02), panic disorder (32.4% vs 26.2%, P < .03), social phobia (33.0% vs 22.9%, P = .001), and anorexia nervosa (9.5% vs 6.0%, P = .02), whereas dysthymic subjects did not. 3.3. Positive and negativity affectivity As hypothesized, definite DPD was associated with lowered levels of positive affectivity (DPD mean, 11.96; non-DPD mean, 13.77; t = 3.20, df = 668, P = .001) and elevated negative affectivity (DPD mean, 21.89; non-DPD mean, 18.98; t = −6.01, df = 408, P < .0001) on the SNAP. Interestingly, no significant difference appeared on the NEO for extraversion as an analog of positive affectivity or neuroticism as a measure of negative affectivity. Multiple logistic regression modeling presence or absence of DPD as a function of both positive and negative affectivity confirmed these analyses. The odds of having a DPD diagnosis increased as negative affectivity increased (Wald = 21.7, P < .0001) and positive affectivity decreased (Wald = 4.98, P < .03) on the SNAP. Controlling for the potential confounding variables of avoidant and borderline personality disorders, the 2 most prevalent comorbid axis II diagnoses, did not affect the significance of these findings. A comparable analysis of subjects with early-onset DD similarly found lower positive affectivity (mean, 11.26 vs 13.06; t = 3.17, df = 676, P = .001) and higher negative affectivity (21.30 vs 19.53; t = 2.45, df = 676, P = .01) for these subjects relative to the larger cohort. 3.4. Diagnostic stability of DPD Both κ and ICCs were used to assess the stability of DPD diagnosis from baseline to 2-year follow-up. Follow-up data were available for 547 subjects. Using the definite DPD definition, κ equaled 0.29 (P < .0001). Diagnostic stability needs to be adjusted for rater reliability on the basis of immediate test-retest reliability; incorporating the immediate test-retest coefficient of 0.62 [27], the adjusted 2-year κ was 0.47 (see Table 3.) Using the dimensional measure of DPD criteria to compare baseline and 2-year ratings, the ICC was 0.41 (P < .0001; adjusting for rater reliability r = 0.71, adjusted ICC = 0.57); there was a decrease of 2.85 DPD symptom items over 2 years. Table 4 indicates that only 47 (31%) of 154 subjects who initially met DIPD-IV criteria for DPD at baseline retained that diagnosis at 2-year follow-up, whereas 25 (6.4%) of 392 subjects without baseline DPD received a definite DPD diagnosis from blind raters 2 years later. Table 3. κ and ICCs for DPD and DD Diagnosis κ/ICC Acute test-retest Adjusted κ/ICC DPD Categorical (definite) 0.29 0.62 0.29/0.62 = 0.47 Dimensional 0.41 0.72 0.41/0.72 = 0.57 Dysthymic disorder 80% of time 0.17 0.35 0.17/0.35 = 0.49 50% of time 0.11 0.35 0.11/0.35 = 0.31 DD is defined by symptoms 80% of time and 50% of time. Table options Table 4. Stability of DPDa over 2 years DPD at year 2 Baseline DPD No Yes Total No 367 25 392 Yes 107 47 154 Total 474 72 546 a “Definite ” DPD defined as definite as rated on the DIPD-IV. Table options By comparison, early-onset DD was stable in 45.3% (34/75) of subjects (κ = 0.17; adjusted for immediate test-retest coefficient of .35, adjusted 2-year κ = 0.49) using a criterion of ongoing dysthymic symptoms 80% of the time, and in 57.5% (46/80; κ = 0.11; adjusted 2-year κ = 0.31) using the criterion of 50% of the time. 3.5. Effect of DPD and DD on remission of MDD Survival analysis using a Cox-proportional hazards model indicated that subjects with baseline MDD (n = 294) who met baseline criteria for definite DPD at baseline had lower hazard of having MDD remit at 2 years, whereas subjects with early onset DD did not. The ratio of the hazard for MDD remission in the 2-year follow-up among subjects with baseline DPD relative to subjects without baseline DPD was 0.67 [95% confidence interval (CI) (0.54, 0.84), Wald = 12.07, P < .001]; that is, subjects with baseline DPD had a 33% lower likelihood of remission than subjects without baseline DPD. This finding remains significant after controlling for avoidant [hazard ratio 0.70, 95% CI (0.56, 0.86), Wald = 10.72, P = .001] and borderline [hazard ratio 0.74, 95% CI (0.59, 0.91), Wald = 8.05, P = .005] personality disorders. By contrast, dysthymic subjects with baseline MDD carried no heightened risk for nonremission [hazard ratio 1.25, 95% CI (0.93, 1.66), Wald = 2.2, P = .14].