Journal of Affective Disorders 174 (2015) 83–88
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
A diagnosis of bipolar spectrum disorder predicts diagnostic conversion from unipolar depression to bipolar disorder: A 5-year retrospective study Young Sup Woo a, In Hee Shim b, Hee-Ryung Wang a, Hoo Rim Song a, Tae-Youn Jun a, Won-Myong Bahk a,n a b
Department of Psychiatry, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea Department of Psychiatry, Cancer Center, Dongnam Institute of Radiological & Medical Sciences, Busan, Republic of Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 13 May 2014 Received in revised form 18 November 2014 Accepted 18 November 2014 Available online 26 November 2014
Background: The major aims of this study were to identify factors that may predict the diagnostic conversion from major depressive disorder (MDD) to bipolar disorder (BP) and to evaluate the predictive performance of the bipolar spectrum disorder (BPSD) diagnostic criteria. Methods: The medical records of 250 patients with a diagnosis of MDD for at least 5 years were retrospectively reviewed for this study. Results: The diagnostic conversion from MDD to BP was observed in 18.4% of 250 MDD patients, and the diagnostic criteria for BPSD predicted this conversion with high sensitivity (0.870) and specificity (0.917). A family history of BP, antidepressant-induced mania/hypomania, brief major depressive episodes, early age of onset, antidepressant wear-off, and antidepressant resistance were also independent predictors of this conversion. Limitations: This study was conducted using a retrospective design and did not include structured diagnostic interviews. Conclusions: The diagnostic criteria for BPSD were highly predictive of the conversion from MDD to BP, and conversion was associated with several clinical features of BPSD. Thus, the BPSD diagnostic criteria may be useful for the prediction of bipolar diathesis in MDD patients. & 2014 Elsevier B.V. All rights reserved.
Keywords: Bipolar depression Unipolar depression Diagnostic conversion Bipolar spectrum disorder
1. Introduction The under-diagnosis of bipolar disorder (BP) is common, and in fact, BP is often either not diagnosed at all or falsely diagnosed as major depressive disorder (MDD; (De Fruyt and Demyttenaere, 2007). The rate of misdiagnosis or under-diagnosis was approximately 80% in a community sample (Hirschfeld et al., 2003) and 40% in a hospitalized sample (Ghaemi et al., 1999). This may be partly due to the diagnostic criteria for BP because the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), and DSM-5 criteria require the occurrence of a manic or hypomanic episode for a diagnosis of BP. It has been suggested that BP patients presenting with an episode of depression without a distinct history of mania/hypomania are practically impossible to diagnose correctly (Grotegerd et al., 2013), and approximately half of BP patients present with a major depressive episode as their first mood episode n Correspondence to: Department of Psychiatry, Yeouido St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, 10, 63-ro, Yeongdeungpo-gu, Seoul, 150-713, Korea. Tel.: þ 82 2 3779 1250; fax: þ 82 2 780 6577. E-mail address: [email protected] (W.M. Bahk).
http://dx.doi.org/10.1016/j.jad.2014.11.034 0165-0327/& 2014 Elsevier B.V. All rights reserved.
(Tondo et al., 2010; Etain et al., 2012). BP patients are three times more likely to suffer from depressive symptoms than from manic or hypomanic symptoms (Judd et al., 2002, 2003), and many BP-II patients seek help when in the depressed phase (Akiskal et al., 2000). Patients who begin BP with depressive episodes appear to be at increased risk for long-term morbidity, disability, and suicide (Baldessarini et al., 2010, 2012). The misdiagnosis and under-diagnosis of BP are due in part to the ‘soft’ symptoms of bipolarity that characterize patients with nonclassical BP (Katzow et al., 2003). Consequently, the search for bipolar signatures in the symptoms and course of MDD has intensified. Several clinical markers of soft bipolarity have been suggested, including mixed anxiety/depressive symptoms (Katzow et al., 2003), conditions associated with impulsivity such as substance abuse, borderline personality, bulimia, and attention deficit disorder (Regier et al., 1990; Katzow et al., 2003), atypical depressive features (Perugi et al., 1998; Mitchell et al., 2001), early age of onset (Neuman et al., 1997; Benazzi, 2009), recurrent depressive episodes (Benazzi, 2009), a family history of BP in first-degree relatives (Benazzi, 2009), brief major depressive episodes (De Fruyt and Demyttenaere, 2007), antidepressant-induced mania/hypomania (Hirschfeld et al., 2003),
84
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
post-partum depression (Judd et al., 2002), treatment resistance (Woo et al., 2008), and specific affective temperament types. i.e., depressive, cyclothymic, hyperthymic, irritable, and anxious (Rihmer et al., 2010). The concept of ‘bipolar spectrum’ arose from the work of Dunner et al. (1970), whose definition of bipolarity included varying degrees of mania and hypomania. This concept was developed by Klerman (1981) and later by Akiskal (1996) to include the continuum from psychotic mania through other expressions of bipolar I disorder and bipolar II disorder, to soft subsyndromal manifestations of bipolarity, i.e., soft or subthreshold bipolarity (Nusslock and Frank, 2011). Later, Ghaemi et al. (2001, 2002) proposed an approach to the spectrum concept that focuses on how to distinguish BP from unipolar depression. He suggested that the following 11 features may predict the diagnosis of BP in patients with depressive symptoms and proposed that the definition of bipolar spectrum disorder (BPSD) is a potential sign of bipolarity: family history of BP in first-degree relatives; antidepressant-induced mania or hypomania; hyperthymic personality; recurrent major depressive episodes (43); brief major depressive episodes (lasting an average of less than 3 months); atypical depressive symptoms; psychotic major depressive episodes; early age of onset of major depressive episodes (o25 years of age); postpartum depression; antidepressant “wearoff;” and treatment resistance (lack of response to Z3 antidepressant treatment trials). Ghaemi et al. (2004) demonstrated that the proposed criteria can distinguish patients with bipolar depression from those with unipolar depression, but the proposed diagnostic criteria were not tested regarding their ability to predict the conversion from unipolar depression to BP. A large proportion of patients suffering from MDD exhibit underlying soft bipolarity, and the identification of antecedents that predict the conversion from unipolar depression to BP would be highly beneficial for these patients. Thus, a set of patients diagnosed with MDD were retrospectively investigated in an attempt to identify the features and diagnostic criteria proposed for BPSD by Ghaemi et al. (2002) to analyze the predictive performance of the diagnostic criteria overall and of each clinical feature.
2. Methods 2.1. Patients and assessments This study was a retrospective investigation that reviewed the medical records of patients who were hospitalized in the psychiatric ward of Yeouido St. Mary's Hospital, College of Medicine at The Catholic University of Korea in Seoul, Korea, between January 1, 2005 and December 31, 2008, with a diagnosis of MDD without prior history of mania or hypomania. The patients were clinically diagnosed with MDD according to the DSM-IV criteria at the index hospitalization. Those who were admitted for an adverse event or other non-psychiatric diagnostic or environmental etiology were excluded from the analyses. To be eligible for inclusion in this study, the patients were required to have had at least 5 years of follow-up care and to have been diagnosed with MDD or BP at the 5-year follow-up. Patients with insufficient data and those who had a severe comorbid medical or neurological condition that could contribute to depressive symptoms, a diagnosis of a mood disorder due to a general medical condition, or a diagnosis that converted from MDD to another diagnosis other than BP were excluded from the analyses. A diagnosis of BPSD was made based on the proposed diagnostic criteria of Ghaemi et al. (2001), with the exception that postpartum depression was excluded from the criteria to avoid a gender bias that may affect the predictive ability of the criteria. The diagnoses of BPSD at index hospitalization and/or an Axis I
disorder during the follow-up period were independently made by two physicians (IHS and HRW) who were blind to the purpose of the study and who separately evaluated the medical records. Based on the DSM-IV which noted that “manic-like episodes that are clearly caused by somatic antidepressant treatment (e.g., medication, electroconvulsive therapy, and light therapy) should not count toward a diagnosis of bipolar I disorder”, patients who exhibited mania/hypomania only during periods of treatment with causative medications, such as antidepressants, were regarded as patients with MDD when we made the DSM-IV diagnosis. However, those subjects with sufficient manic/hypomanic symptoms to be diagnosed as bipolar disorder during the follow-up period were categorized in the conversion group irrespective of a history of antidepressant-induced mania/hypomania at the index hospitalization or during the follow-up period. Clinical data from the index hospitalization included each item of the diagnostic criteria for BPSD as well as age at index hospitalization and sex. A history of the clinical characteristics, such as age at first psychiatric treatment, age at onset, and number of past major depressive episodes before the subjects visited our hospital, which were investigated at the time of the index hospitalization retrospectively, was also included in the analysis. As this was a retrospective study, all data were obtained during routine psychiatric examinations and treatment. 2.2. Statistical analysis The clinical and demographic variables of patients with a diagnosis of MDD at index hospitalization and the 5-year followup (unipolar group) were compared with the variables of patients whose diagnosis changed from MDD at index hospitalization to BP during the follow-up period (conversion group). A Chi-squared test or Fisher's exact test was used to analyze categorical variables and an independent t-test or Mann–Whitney U-test was used to test continuous variables. To identify the independent predictors of the conversion from MDD to BP, a multinomial logistic regression analysis was conducted for which the independent variables were sex, age, psychiatric comorbidity, and the clinical features of BPSD. Statistical analyses were performed using SAS for Windows (ver. 9.2), and two-tailed p-Values with an alpha level of 0.05 were considered significant, except that a Bonferroni-corrected significance level of po0.0045 (0.05/11 tests) was applied to the univariate analysis that compared the groups in terms of the proportion of subjects with each of the 10 proposed bipolar features in the modified BPSD diagnostic criteria who met the modified BPSD criteria. 2.3. Ethics This study was approved by the institutional review board of Yeouido St. Mary's Hospital in Seoul, Korea, and was conducted according to the Declaration of Helsinki. The institutional review board approved the exemption for informed consent because this was a retrospective chart review study.
3. Results During the study period, 448 patients were discharged with a diagnosis of MDD. Of these patients, 250 (55.8%) fulfilled the eligibility criteria for the study. 3.1. Diagnostic conversion from MDD to BP Of the original 250 patients initially diagnosed with MDD, the conversion rate to BP was 18.4%. The demographic and clinical characteristics of these patients at the index hospitalization are
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
85
Table 1 Characteristics of evaluable patients at index episode. Total (n ¼250) n (%), mean 7SD
Age (years) Sex (female) Age at onset (years) Total number of MDE Psychiatric comorbidity
Unipolar group (n¼204) n (%), mean 7 SD
Conversion group n (%), mean 7 SD
p-Value
BP-I or II (n¼ 46)
BP-I (n¼ 25)
UP vs. BP conversion
BP-II (n ¼21)
UP vs. BP-I
UP vs. BP-II
BP-I vs. BP-II
55.3 7 15.5 192 (76.8%) 48.47 16.0
57.8 7 14.9 157 (77.0%) 51.0 7 15.4
44.1 713.3 35 (76.1%) 36.9 713.5
41.8 7 12.2 46.9 7 14.4 o 0.001 18 (72.0%) 17 (81.0%) 0.899 35.9713.3 38.17 14.0 o 0.001
o 0.001 0.581 o 0.001
0.004 0.677 0.001
0.704 0.478 40.999
1.9 7 2.8
1.6 7 2.7
3.0 73.3
2.0 7 2.0
4.17 4.1
0.011
40.999
o0.001
0.027
17 (6.8%)
10 (4.9%)
7 (15.2%)
4 (16.0%)
3 (14.3%)
0.02
0.052
0.109
40.999
Abbreviations: n; number; SD; standard deviation; MDE; major depressive episode, BP-I; bipolar I disorder, BP-II; bipolar II disorder. Unipolar group; the group of patients whose diagnosis remained major depressive disorder at the 5-year follow-up. Switch group; the group of patients whose diagnosis was changed from major depressive disorder to bipolar disorder during the 5-year follow-up period.
presented (Table 1). There were 35 female patients (76.1%) in the conversion group and 157 female patients (77.0%) in the unipolar group (p¼ 0.899). The mean age at index hospitalization (conversion group¼ 44.1 713.3, unipolar group¼57.8714.9; po 0.001) and the age of onset (conversion group¼ 36.9 713.5, unipolar group¼51.0 715.4; p o0.001) were significantly lower in the conversion group compared with the unipolar group (Table 1). The frequency of occurrence of each proposed bipolar feature and the results of the univariate analyses are provided (Table 2). The proportion of patients with reporting BP in first-degree relatives was significantly higher in the conversion group (52.2%, n¼24) than in the unipolar group (11.3%, n¼23; po0.001, odds ratio [OR]¼8.585, 95% confidence interval [CI]¼4.166–17.693), and a history of antidepressant-induced mania or hypomania was significantly more common in the conversion group (60.9%, n¼28) than in the unipolar group (4.9%, n¼ 10; po0.001, OR¼ 30.178, 95% CI¼ 12.660–71.935. Recurrent major depressive episodes (43) occurred in a significantly greater proportion of patients in the conversion group (45.7%, n¼21) than the unipolar group (15.2%, n¼31; po0.001, OR¼ 4.688, 95% CI¼2.340–9.391), and likewise, brief major depressive episodes (lasting an average of less than 3 months) were also significantly more common in the conversion group (23.9%, n¼11) than the unipolar group (3.9%, n¼ 8; po0.001, OR¼ 7.700, 95% CI¼2.892– 20.499). A significantly greater proportion of patients in the conversion group (58.7%, n¼ 27) exhibited atypical depressive features at index hospitalization relative to the unipolar group (18.1%, n¼ 37; po0.001, OR¼6.414, 95% CI¼3.228–12.743), and every each atypical symptoms were also significantly frequent in conversion group than unipolar group (Table 2). The psychotic features were more common in the conversion group (45.7%, n¼ 21) compared with the unipolar group (7.4%, n¼ 15; po0.001, OR¼10.584, 95% CI¼ 4.837– 23.157). Moreover, early age of onset (prior to the age 25) was significantly more common in the conversion group (43.5%, n¼20) compared with the unipolar group (6.4%, n¼13; po0.001, OR¼ 11.302, 95% CI¼5.030–25.395), and antidepressant wear-off (po0.001, OR¼6.276, 95% CI¼3.086–12.764) and antidepressant resistance (po0.001, OR¼22.121, 95% CI¼10.092–48.486) were also significantly different between the unipolar and conversion group. Only a hyperthymic temperament was not significantly different between the two groups (p¼0.465). Accordingly, the proportion of patients who met the criteria for BPSD was significantly higher in the conversion group (87.0%, n¼40) than the unipolar group (8.3%, n¼17; po0.001, OR¼73.333, 95% CI¼27.212–197.625). The sensitivity of the BPSD diagnosis was 0.870, the specificity was 0.917; the positive predictive value of the BPSD diagnosis was 0.702, and the negative predictive value was 0.969. The multivariate logistic regression included sex, age at index episode, and each of the 11 proposed bipolar features (except for
hyperthymic temperament, which was not significantly different between the groups, and postpartum depression, which was excluded from the analysis) as variables. Psychiatric comorbidity revealed several independent factors correlated with the conversion to BP including a family history of BP in first-degree relatives (p¼ 0.046, OR¼4.304, 95% CI¼1.025–18.061), antidepressant-induced mania or hypomania (po0.001, OR¼42.915, 95% CI¼8.886–207.273), brief major depressive episodes (p¼0.004, OR¼18.660, 95% CI¼2.478– 140.497), early age of onset (p¼0.040, OR¼5.146, 95% CI¼1.077– 27.229), antidepressant wear-off (p¼0.001, OR¼8.967, 95% CI¼ 2.364–34.009), and antidepressant treatment resistance (p¼0.006, OR¼6.751, 95% CI¼1.722–26.465; Table 3). 3.2. Predictors of the diagnostic conversion from MDD to BP-I or BP-II Further subanalyses were conducted to investigate the predictors of conversion from MDD to either BP-I or BP-II. The conversion rates from MDD to BP-I and BP-II were 10.0% (n ¼25) and 8.4% (n ¼21), respectively. The results from the univariate analysis are presented (Table 2); the predictive clinical features for the conversion from MDD to BP-I or BP-II were the same as the predictors of the conversion from MDD to BP. Psychotic features were less common in the BP-II (28.6%, n ¼6) than in the BP-I group (60.0%, n ¼15), but the difference did not reach Bonferroni-corrected significance (p ¼0.033). In the multivariate analysis (Table 4), antidepressant-induced mania/hypomania (po0.001, OR¼59.753, 95% CI¼ 6.870–519.696) and antidepressant wear-off (p¼ 0.014, OR¼ 7.497, 95% CI¼1.496– 37.568) significantly predicted the conversion to BP-I. In addition, age (p¼0.080, OR¼ 0.943, 95% CI¼0.882–1.007), family history of BP (p¼ 0.090, OR¼ 5.312, 95% CI¼0.770–36.628), brief major depressive episodes (p¼0.075, OR¼12.787, 95% CI¼0.776–210.679), and atypical features (p¼0.069, OR¼5.368, 95% CI¼ 0.876–32.893) tended to predict conversion. Predictors of the conversion to BP-II included antidepressant induced mania/hypomania (po0.001, OR¼178.666, 95% CI¼11.035–2892.825), brief major depressive episodes (p¼ 0.013, OR¼28.407, 95% CI¼2.017–399.988), early age at onset (p¼ 0.023, OR¼33.690, 95% CI¼ 1.618–701.580), antidepressant wear-off (p¼0.004, OR¼26.951, 95% CI¼2.892–251.182), and treatment resistance (p¼ 0.005, OR¼13.803, 95% CI¼ 2.164–251.182). In addition, family history of BP showed a trend toward significance (p¼ 0.060, OR¼ 7.750, 95% CI¼ 0.917–65.493). The sensitivity of the BPSD diagnosis for the conversion to BP-I was 0.880, and the specificity was 0.917. The positive predictive value of the BPSD diagnosis for the conversion to BP-I was 0.564, and the negative predictive value was 0.984. The sensitivity of the BPSD diagnosis for the conversion to BP-II was 0.857, and the
86
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
Table 2 Comparisons of characteristics and 10 proposed bipolar features in the bipolar spectrum disorder diagnostic criteria except postpartum depression between unipolar group and conversion group. Unipolar group Total (n ¼204) n (%), (n ¼250) mean 7SD n (%), mean7 SD
47 (18.8%) Family history of bipolar disorder in a 1st-degree relatives 38 (15.2%) Antidepressantinduced mania/ hypomania Hyperthymic 40 (16.0%) temperament
Conversion group n (%), mean 7 SD
UP vs. BP-I
UP vs. BP-II
BP-I vs. BP-II
BP-II p-Value OR BP-I or BP-I (n¼ 25) (n¼ 21) (95% CI) II (n¼ 46)
p-Value OR (95% CI)
p-Value OR (95% CI)
p-Value OR (95% CI)
o 0.001 8.585 (4.166– 17.693) o 0.001 30.178 (12.660– 71.935) 0.465 1.357 (0.596– 3.090) o 0.001 4.688 (2.340– 9.391) o 0.001 7.700 (2.892– 20.499) o 0.001 6.414 (3.228– 12.743) o 0.001 5.118 (2.599– 10.075) o 0.001 9.308 (4.542– 19.077) o 0.001 4.667 (2.362– 9.220) o 0.001 6.009 (2.966– 12.176) 0.001 3.214 (1.601– 6.455) o 0.001 10.584 (4.837– 23.157) o 0.001 11.302 (5.030– 25.395) o 0.001 6.276 (3.086– 12.764) o 0.001 22.121 (10.092– 48.486) o 0.001 73.333 (27.212– 197.625)
o 0.001 7.264 (2.964– 17.806) o 0.001 29.100 (10.474– 80.850) 0.256 1.762 (0.652– 4.763) 0.005 3.720 (1.533– 9.031) o 0.001 9.528 (3.098– 29.300) o 0.001 8.024 (3.292– 19.558) o 0.001 6.375 (2.597– 15.650) o 0.001 5.626 (2.302– 13.750) o 0.001 7.000 (2.911– 16.832) 0.002 4.370 (1.783– 10.715) 0.029 2.813 (1.148– 6.889) o 0.001 18.900 (7.255– 49.234) o 0.001 13.562 (5.167– 35.598) o 0.001 6.320 (2.606– 15.327) o 0.001 22.408 (8.386– 59.880) o 0.001 80.667 (21.886– 297.312)
o 0.001 10.493 (3.989– 27.598) o 0.001 31.525 (10.641– 93.400) 40.999 0.930 (0.258– 3.347) o 0.001 6.139 (2.403– 15.681) 0.017 5.765 (1.573– 21.121) 0.001 4.965 (1.964– 12.552) 0.002 4.000 (1.593– 10.043) o 0.001 17.900 (6.358– 50.396) 0.039 2.872 (1.109– 7.436) o 0.001 8.741 (3.365– 22.701) 0.007 3.750 (1.466– 9.594) 0.007 5.040 (1.707– 14.885) o 0.001 9.041 (3.181– 25.699) o 0.001 6.224 (2.407– 16.095) o 0.001 21.786 (7.632– 62.190) o 0.001 66.000 (17.644– 246.880)
0.536
23 (11.3%)
24 12 12 (52.2%) (48.0%) (57.1%)
10 (4.9%)
28 15 13 (60.9%) (60.0%) (61.9%)
31 (15.2%)
9 6 3 (19.6%) (24.0%) (14.3%)
Recurrent MDEs ( 43)
52 (20.8%) 31 (15.2%)
21 10 11 (45.7%) (40.0%) (52.4%)
Brief MDEs (on average, o3 months) Atypical feature
19 (7.6%)
11 7 4 (23.9%) (28.0%) (19.0%)
8 (3.9%)
64 (25.6%) 37 (18.1%)
27 16 11 (58.7%) (64.0%) (52.4%)
Mood reactivity
80 (32.0%) 51 (25.0%)
29 17 12 (63.0%) (68.0%) (57.1%)
Weight gain or increase in appetite
51 (20.4%)
25 (12.3%)
26 11 15 (56.5%) (44.0%) (71.4%)
Hypersomnia
59 (23.6%) 36 (17.6%)
23 15 8 (50.0%) (60.0%) (38.1%)
Leaden paralysis
49 (19.6%)
22 (47.8%)
Interpersonal rejection sensitivity
52 (20.8%) 34 (16.7%)
18 9 9 (39.1%) (36.0%) (42.9%)
Psychotic feature
36 (14.4%)
15 (7.4%)
21 15 6 (45.7%) (60.0%) (28.6%)
Early age at onset of MDE ( o 25 years)
33 (13.2%)
13 (6.4%)
20 12 8 (43.5%) (48.0%) (38.1%)
Antidepressant wearoff
48 (19.2%)
26 (12.7%)
22 (47.8%)
12 10 (48.0%) (47.6%)
Lack of response to Z 3 antidepressant treatment trials Met the diagnostic criteria of bipolar spectrum disorder
54 (21.6%)
21 (10.3%)
33 (71.7%)
18 15 (72.0%) (71.4%)
57 (22.8%)
17 (8.3%)
40 (87.0%)
22 18 (88.0%) (85.7%)
27 (13.2%)
10 12 (40.0%) (57.1%)
UP vs. BP conversion
1.444 (0.450– 4.641) 0.895 1.083 (0.330– 3.561) 0.478 0.528 (0.114– 2.434) 0.401 1.650 (0.511– 5.327) 0.478 0.605 (0.150– 2.444) 0.425 0.619 (0.189– 2.020) 0.447 0.627 (0.188– 2.095) 0.062 3.182 (0.927– 10.917) 0.139 0.410 (0.125– 1.348) 0.246 2.000 (0.616– 6.494) 0.635 1.333 (0.406– 4.379) 0.033 0.267 (0.077– 0.921) 0.500 0.667 (0.205– 2.169) 0.979 0.985 (0.308– 3.146) 0.966 0.972 (0.268– 3.524) 40.999 0.818 (0.147– 4.557)
Abbreviations: n; number, SD; standard deviation, OR; odds ratio, CI; confidence interval, MDE; major depressive episode, BP-I; bipolar I disorder, BP-II; bipolar II disorder. Unipolar group; the group of patients whose diagnosis remained major depressive disorder at the 5-year follow-up. Conversion group; the group of patients whose diagnosis was changed from major depressive disorder to bipolar disorder during the 5-year follow-up period.
specificity was 0.917. The positive predictive value of the BPSD diagnosis for the conversion to BP-II was 0.514, and the negative predictive value was 0.984.
4. Discussion To our knowledge, this is among the first studies to evaluate the predictive performance of the BPSD diagnostic criteria proposed by Ghaemi et al. (2001) regarding the diagnostic conversion of MDD to BP. As expected, the BPSD diagnostic criteria were highly predictive of the conversion from MDD to BP (OR ¼73.333), MDD to BP-I (OR ¼80.667), and MDD to BP-II (OR ¼66.000) during the
5-year follow-up period. The diagnostic conversion rates found in this study (BP-I: 2.0% per year; BP-II: 1.7% per year) are comparable to the findings of several previous studies in which the conversion rate from MDD to BP-I was 0.4–1.3% per year, and that for MDD to BP-II was 0.5–1.8% per year (Akiskal et al., 1995; Coryell et al., 1995; Goldberg et al., 2001; Angst et al., 2005; Holma et al., 2008). Moreover, the predictive clinical features identified here, including family history of BP, antidepressant-induced mania/hypomania, brief major depressive episodes, early age at onset, antidepressant wear-off, and antidepressant resistance, are in agreement with the findings of previous studies (Neuman et al., 1997; Ghaemi et al., 2002; Hirschfeld et al., 2003; De Fruyt and Demyttenaere, 2007; Benazzi, 2009). However, although these factors were predictive in
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
87
Table 3 Final model of multivariate logistic regression among the unipolar group and conversion group. Significance
Sex (male) Age Psychiatric comorbidity Family history of bipolar disorder in a 1st-degree relatives Antidepressant-induced mania/hypomania Recurrent MDEs ( 43) Brief MDEs (on average, o 3 months) Atypical feature Psychotic feature Early age at onset of MDE ( o25 years) Antidepressant wear-off Lack of response to Z 3 antidepressant treatment trials
OR
0.315 0.108 0.629 0.046 o 0.001 0.720 0.004 0.334 0.981 0.040 0.001 0.006
95% CI
0.429 0.960 0.545 4.304 42.915 1.304 18.660 1.901 1.018 5.416 8.967 6.751
Lower
Upper
0.082 0.913 0.046 1.026 8.886 0.306 2.478 0.516 0.247 1.077 2.364 1.722
2.232 1.009 6.407 18.061 207.273 5.565 140.497 7.000 4.196 27.229 34.009 26.465
Abbreviations: OR; odds ratio, CI; confidence interval, MDE; major depressive episode. Unipolar group; the group of patients whose diagnosis remained major depressive disorder at the 5-year follow-up. Conversion group; the group of patients whose diagnosis was changed from major depressive disorder to bipolar disorder during the 5-year follow-up period. Table 4 Final model of multivariate logistic regression among the unipolar group and bipolar I and bipolar II disorder conversion group. MDD to BP-I conversion Significance
Sex (male) Age Psychiatric comorbidity Family history of bipolar disorder in a 1st-degree relatives Antidepressant-induced mania/hypomania Recurrent MDEs ( 43) Brief MDEs (on average, o 3 months) Atypical feature Psychotic feature Early age at onset of MDE ( o25 years) Antidepressant wear-off Lack of response to Z 3 antidepressant treatment trials
0.354 0.080 0.674 0.090 0.000 0.791 0.075 0.069 0.223 0.315 0.014 0.287
OR
MDD to BP-II conversion 95% CI
0.374 0.943 0.476 5.312 59.753 1.288 12.787 5.368 2.791 2.677 7.497 2.690
Significance
Lower
Upper
0.047 0.882 0.015 0.770 6.870 0.199 0.776 0.876 0.536 0.392 1.496 0.435
2.988 1.007 15.242 36.628 519.696 8.346 210.679 32.893 14.538 18.275 37.568 16.636
0.908 0.642 0.959 0.060 0.000 0.893 0.013 0.330 0.244 0.023 0.004 0.005
OR
0.856 0.983 1.095 7.750 178.666 0.862 28.407 0.341 0.273 33.690 26.951 13.803
95% CI Lower
Upper
0.061 0.917 0.034 0.917 11.035 0.100 2.017 0.039 0.031 1.618 2.892 2.164
12.006 1.055 34.912 65.493 2892.825 7.464 399.988 2.967 2.430 701.580 251.182 88.028
Abbreviations: MDD; major depressive disorder, BP-I; bipolar I disorder, BP-II; bipolar II disorder, OR; odds ratio, CI; confidence interval, MDE; major depressive episode.
the univariate analysis, recurrent major depressive episodes, atypical features, and psychotic features were not significantly predictive in the multivariate analysis. Psychotic features were found to be an accurate predictor of diagnostic conversion by Coryell et al. (1995), Goldberg et al. (2001), and Othmer et al. (2007). However, these findings were based on univariate analyses in two studies (Coryell et al., 1995; Goldberg et al., 2001), whereas Othmer et al. (2007) found only three major risk factors using logistic regression analysis; this may explain this discrepancy among these studies. Psychotic features were also found to be a significant predictor in univariate, but not multivariate analyses, conducted by Holma et al. (2008). In addition, a number of studies have found that atypical features were more common in patients with BP than in patients with MDD (Quitkin et al., 2003; Angst et al., 2006; Akiskal and Benazzi, 2008), but these features were not predictive of a diagnostic conversion in the present study. For instance, Gan et al. (2011) found that some atypical features, such as hypersomnia and irritability, were associated with BP, but others, such as weight gain, were not. Since our study defined atypical features according to DSM-IV criteria, the predictive value of each atypical feature could not be evaluated, which might explain the discrepancy in the results. The presence of recurrent major depressive episodes or a high number of past episodes was shown to be a significant risk factor for the diagnostic conversion from MDD to BP in some studies (Angst et al., 2005; Dudek et al., 2013; Ostergaard et al., 2014), but
not others (Coryell et al., 1995; Gan et al., 2011). Several of these studies (Coryell et al., 1995; Dudek et al., 2013; Ostergaard et al., 2014) compared the mean number of depressive episodes between BP converters and non-converters, whereas recurrent depression was defined as six or more depressive episodes by Angst et al. (2005) and two or more depressive episodes by Gan et al. (2011). In concurrence with Ostergaard et al. (2014) and Dudek et al. (2013), we found that the mean number of major depressive episodes was higher in the BP conversion group than the MDD group. However, in the subanalysis that divided the conversion group into BP-I and BP-II patients, this difference in recurrent depression was significant in the BP-II, but not the BP-I, group. When the definition for recurrent major depressive episodes (43) used in the present study was applied, a significant difference was found for both the BP-I and BP-II groups in the univariate analysis, similar to Angst et al. (2005), but none in the multivariate analysis, as reported by Gan et al. (2011). Hence, the definition of ‘recurrent’ should be carefully considered in the interpretation of the results from relevant studies. Several possible limitations in the present study design should be considered when interpreting these findings, including the retrospective design and small sample. Much of the clinical information was obtained during routine clinical practice, and the possibility of recall bias and reviewer bias needs to be considered. Although the data were collected prospectively after the subjects began psychiatric treatment at our hospital, the clinical characteristics before the
88
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
treatment, including retrospective information from the subjects such as age at first psychiatric treatment, age at onset, and number of past major depressive episodes, could be affected by recall bias. The diagnoses of MDD and BP were made clinically without use of the structured diagnostic interview for Axis I or II disorders or of a scale/ questionnaire for the evaluation of hyperthymic temperament. In addition, because the diagnostic conversion from MDD to BP was observed in only 46 patients due to the small sample size, and only patients with at least 5 years of follow-up care were included in this study, the results are most applicable to patients who are largely adherent to their treatment plan. Moreover, the severity of the depressive or subthreshold hypomanic symptoms was not evaluated; these factors may be another predictor of the conversion to BP (Holma et al., 2008; Fiedorowicz et al., 2011). And the exclusion of MDD patients who were admitted due to an adverse event or other non-psychiatric diagnostic or environmental etiologies and who converted to a diagnosis other than BP during follow-up could decrease the generalizability. Finally, the follow-up period of 5 years is only moderately long for this type of investigation because the diagnostic conversion from MDD to BP may still occur after decades. The present data demonstrate that a substantial portion of MDD patients ultimately convert to BP and that the BPSD diagnostic criteria may predict this conversion with high sensitivity and specificity. In addition, the predictive value of BPSD and several of its clinical features, including a family history of BP, antidepressantinduced mania/hypomania, brief major depressive episodes, early age of onset, antidepressant wear-off, and antidepressant resistance, may be considered independent risk factors for this diagnostic conversion. These predictors should be taken into account in clinical practice because the diagnostic conversion from MDD to BP has clinical importance for the treatment and prognosis of patients.
Role of funding source Nothing declared.
Conflict of interest No conflict declared.
Acknowledgments The authors had no conflict of interest in conducting this study or preparing the manuscript.
References Akiskal, H.S., 1996. The prevalent clinical spectrum of bipolar disorders: beyond DSM-IV. J. Clin. Psychopharmacol. 16, 4S–14S. Akiskal, H.S., Benazzi, F., 2008. Continuous distribution of atypical depressive symptoms between major depressive and bipolar II disorders: dose–response relationship with bipolar family history. Psychopathology 41, 39–42. Akiskal, H.S., Bourgeois, M.L., Angst, J., Post, R., Moller, H., Hirschfeld, R., 2000. Reevaluating the prevalence of and diagnostic composition within the broad clinical spectrum of bipolar disorders. J. Affect. Disord. 59 (Suppl. 1), S5–S30. Akiskal, H.S., Maser, J.D., Zeller, P.J., Endicott, J., Coryell, W., Keller, M., Warshaw, M., Clayton, P., Goodwin, F., 1995. Switching from ‘unipolar’ to bipolar II. An 11-year prospective study of clinical and temperamental predictors in 559 patients. Arch. Gen. Psychiatry 52, 114–123. Angst, J., Gamma, A., Benazzi, F., Silverstein, B., Ajdacic-Gross, V., Eich, D., Rossler, W., 2006. Atypical depressive syndromes in varying definitions. Eur. Arch. Psychiatry Clin. Neurosci. 256, 44–54. Angst, J., Sellaro, R., Stassen, H.H., Gamma, A., 2005. Diagnostic conversion from depression to bipolar disorders: results of a long-term prospective study of hospital admissions. J. Affect. Disord. 84, 149–157. Baldessarini, R.J., Salvatore, P., Khalsa, H.M., Gebre-Medhin, P., Imaz, H., GonzalezPinto, A., Perez, J., Cruz, N., Maggini, C., Tohen, M., 2010. Morbidity in 303 firstepisode bipolar I disorder patients. Bipolar Disord. 12, 264–270. Baldessarini, R.J., Undurraga, J., Vazquez, G.H., Tondo, L., Salvatore, P., Ha, K., Khalsa, H.M., Lepri, B., Ha, T.H., Chang, J.S., Tohen, M., Vieta, E., 2012. Predominant recurrence polarity among 928 adult international bipolar I disorder patients. Acta Psychiatr. Scand. 125, 293–302.
Benazzi, F., 2009. Classifying mood disorders by age-at-onset instead of polarity. Prog. Neuropsychopharmacol. Biol. Psychiatry 33, 86–93. Coryell, W., Endicott, J., Maser, J.D., Keller, M.B., Leon, A.C., Akiskal, H.S., 1995. Longterm stability of polarity distinctions in the affective disorders. Am. J. Psychiatry 152, 385–390. De Fruyt, J., Demyttenaere, K., 2007. Bipolar (spectrum) disorder and mood stabilization: standing at the crossroads? Psychother. Psychosomat. 76, 77–88. Dudek, D., Siwek, M., Zielinska, D., Jaeschke, R., Rybakowski, J., 2013. Diagnostic conversions from major depressive disorder into bipolar disorder in an outpatient setting: results of a retrospective chart review. J. Affect. Disord. 144, 112–115. Dunner, D.L., Gershon, E.S., Goodwin, F.K., 1970. Heritable factors in the severity of affective illness. Sci. Proc. Am. Psychiatr. Assoc. 123, 187–188. Etain, B., Lajnef, M., Bellivier, F., Mathieu, F., Raust, A., Cochet, B., Gard, S., M’Bailara, K., Kahn, J.P., Elgrabli, O., Cohen, R., Jamain, S., Vieta, E., Leboyer, M., Henry, C., 2012. Clinical expression of bipolar disorder type I as a function of age and polarity at onset: convergent findings in samples from France and the United States. J. Clin. Psychiatry 73, e561–e566. Fiedorowicz, J.G., Endicott, J., Leon, A.C., Solomon, D.A., Keller, M.B., Coryell, W.H., 2011. Subthreshold hypomanic symptoms in progression from unipolar major depression to bipolar disorder. Am. J. Psychiatry 168, 40–48. Gan, Z., Diao, F., Wei, Q., Wu, X., Cheng, M., Guan, N., Zhang, M., Zhang, J., 2011. A predictive model for diagnosing bipolar disorder based on the clinical characteristics of major depressive episodes in Chinese population. J. Affect. Disord. 134, 119–125. Ghaemi, S.N., Ko, J.Y., Goodwin, F.K., 2001. The bipolar spectrum and the antidepressant view of the world. J. Psychiatr. Pract. 7, 287–297. Ghaemi, S.N., Ko, J.Y., Goodwin, F.K., 2002. “Cade's disease” and beyond: misdiagnosis, antidepressant use, and a proposed definition for bipolar spectrum disorder. Can. J. Psychiatry 47, 125–134. Ghaemi, S.N., Sachs, G.S., Chiou, A.M., Pandurangi, A.K., Goodwin, K., 1999. Is bipolar disorder still underdiagnosed? Are antidepressants overutilized?. J. Affect. Disord. 52, 135–144. Ghaemi, S.N., Hsu, D.J., Ko, J.Y., Baldassano, C.F., Kontos, N.J., Goodwin, F.K., 2004. Bipolar spectrum disorder: a pilot study. Psychopathology 37, 222–226. Goldberg, J.F., Harrow, M., Whiteside, J.E., 2001. Risk for bipolar illness in patients initially hospitalized for unipolar depression. Am. J. Psychiatry 158, 1265–1270. Grotegerd, D., Suslow, T., Bauer, J., Ohrmann, P., Arolt, V., Stuhrmann, A., Heindel, W., Kugel, H., Dannlowski, U., 2013. Discriminating unipolar and bipolar depression by means of fMRI and pattern classification: a pilot study. Eur. Arch. Psychiatry Clin. Neurosci. 263, 119–131. Hirschfeld, R.M., Calabrese, J.R., Weissman, M.M., Reed, M., Davies, M.A., Frye, M.A., Keck Jr., P.E., Lewis, L., McElroy, S.L., McNulty, J.P., Wagner, K.D., 2003. Screening for bipolar disorder in the community. J. Clin. Psychiatry 64, 53–59. Holma, K.M., Melartin, T.K., Holma, I.A., Isometsa, E.T., 2008. Predictors for switch from unipolar major depressive disorder to bipolar disorder type I or II: a 5-year prospective study. J. Clin. Psychiatry 69, 1267–1275. Judd, L.L., Akiskal, H.S., Schettler, P.J., Coryell, W., Endicott, J., Maser, J.D., Solomon, D.A., Leon, A.C., Keller, M.B., 2003. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch. Gen. Psychiatry 60, 261–269. Judd, L.L., Akiskal, H.S., Schettler, P.J., Endicott, J., Maser, J., Solomon, D.A., Leon, A.C., Rice, J.A., Keller, M.B., 2002. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch. Gen. Psychiatry 59, 530–537. Katzow, J.J., Hsu, D.J., Ghaemi, S.N., 2003. The bipolar spectrum: a clinical perspective. Bipolar Disord. 5, 436–442. Klerman, G.L., 1981. The spectrum of mania. Compr. Psychiatry 22, 11–20. Mitchell, P.B., Wilhelm, K., Parker, G., Austin, M.P., Rutgers, P., Malhi, G.S., 2001. The clinical features of bipolar depression: a comparison with matched major depressive disorder patients. J. Clin. Psychiatry 62, 212–216 (quiz 217). Neuman, R.J., Geller, B., Rice, J.P., Todd, R.D., 1997. Increased prevalence and earlier onset of mood disorders among relatives of prepubertal versus adult probands. J. Am. Acad. Child Adolesc. Psychiatry 36, 466–473. Nusslock, R., Frank, E., 2011. Subthreshold bipolarity: diagnostic issues and challenges. Bipolar Disord. 13, 587–603. Ostergaard, S.D., Straszek, S., Petrides, G., Skadhede, S., Jensen, S.O., MunkJorgensen, P., Nielsen, J., 2014. Risk factors for conversion from unipolar psychotic depression to bipolar disorder. Bipolar Disord. 16, 180–189. Othmer, E., Desouza, C.M., Penick, E.C., Nickel, E.J., Hunter, E.E., Othmer, S.C., Powell, B.J., Hall, S.B., 2007. Indicators of mania in depressed outpatients: a retrospective analysis of data from the Kansas 1500 study. J. Clin. Psychiatry 68, 47–51. Perugi, G., Akiskal, H.S., Lattanzi, L., Cecconi, D., Mastrocinque, C., Patronelli, A., Vignoli, S., Bemi, E., 1998. The high prevalence of “soft” bipolar (II) features in atypical depression. Compr. Psychiatry 39, 63–71. Quitkin, F.M., McGrath, P.J., Stewart, J.W., Klein, D.F., 2003. A reappraisal of atypical depression. Am. J. Psychiatry 160, 798–800 (author reply 800–791). Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd, L.L., Goodwin, F.K., 1990. Comorbidity of mental disorders with alcohol and other drug abuse. Results from the Epidemiologic Catchment Area (ECA) Study. JAMA 264, 2511–2518. Rihmer, Z., Akiskal, K.K., Rihmer, A., Akiskal, H.S., 2010. Current research on affective temperaments. Curr. Opin. Psychiatry 23, 12–18. Tondo, L., Lepri, B., Cruz, N., Baldessarini, R.J., 2010. Age at onset in 3014 Sardinian bipolar and major depressive disorder patients. Acta Psychiatr. Scand. 121, 446–452. Woo, Y.S., Chae, J.H., Jun, T.Y., Kim, K.S., Bahk, W.M., 2008. The bipolar diathesis of treatment-resistant major depressive disorder. Int. J. Psychiatry Clin. Pract. 12, 142–146.
Contents lists available at ScienceDirect
Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad
Research report
A diagnosis of bipolar spectrum disorder predicts diagnostic conversion from unipolar depression to bipolar disorder: A 5-year retrospective study Young Sup Woo a, In Hee Shim b, Hee-Ryung Wang a, Hoo Rim Song a, Tae-Youn Jun a, Won-Myong Bahk a,n a b
Department of Psychiatry, College of Medicine, The Catholic University of Korea, Seoul, Republic of Korea Department of Psychiatry, Cancer Center, Dongnam Institute of Radiological & Medical Sciences, Busan, Republic of Korea
art ic l e i nf o
a b s t r a c t
Article history: Received 13 May 2014 Received in revised form 18 November 2014 Accepted 18 November 2014 Available online 26 November 2014
Background: The major aims of this study were to identify factors that may predict the diagnostic conversion from major depressive disorder (MDD) to bipolar disorder (BP) and to evaluate the predictive performance of the bipolar spectrum disorder (BPSD) diagnostic criteria. Methods: The medical records of 250 patients with a diagnosis of MDD for at least 5 years were retrospectively reviewed for this study. Results: The diagnostic conversion from MDD to BP was observed in 18.4% of 250 MDD patients, and the diagnostic criteria for BPSD predicted this conversion with high sensitivity (0.870) and specificity (0.917). A family history of BP, antidepressant-induced mania/hypomania, brief major depressive episodes, early age of onset, antidepressant wear-off, and antidepressant resistance were also independent predictors of this conversion. Limitations: This study was conducted using a retrospective design and did not include structured diagnostic interviews. Conclusions: The diagnostic criteria for BPSD were highly predictive of the conversion from MDD to BP, and conversion was associated with several clinical features of BPSD. Thus, the BPSD diagnostic criteria may be useful for the prediction of bipolar diathesis in MDD patients. & 2014 Elsevier B.V. All rights reserved.
Keywords: Bipolar depression Unipolar depression Diagnostic conversion Bipolar spectrum disorder
1. Introduction The under-diagnosis of bipolar disorder (BP) is common, and in fact, BP is often either not diagnosed at all or falsely diagnosed as major depressive disorder (MDD; (De Fruyt and Demyttenaere, 2007). The rate of misdiagnosis or under-diagnosis was approximately 80% in a community sample (Hirschfeld et al., 2003) and 40% in a hospitalized sample (Ghaemi et al., 1999). This may be partly due to the diagnostic criteria for BP because the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV), and DSM-5 criteria require the occurrence of a manic or hypomanic episode for a diagnosis of BP. It has been suggested that BP patients presenting with an episode of depression without a distinct history of mania/hypomania are practically impossible to diagnose correctly (Grotegerd et al., 2013), and approximately half of BP patients present with a major depressive episode as their first mood episode n Correspondence to: Department of Psychiatry, Yeouido St. Mary’s Hospital, College of Medicine, The Catholic University of Korea, 10, 63-ro, Yeongdeungpo-gu, Seoul, 150-713, Korea. Tel.: þ 82 2 3779 1250; fax: þ 82 2 780 6577. E-mail address: [email protected] (W.M. Bahk).
http://dx.doi.org/10.1016/j.jad.2014.11.034 0165-0327/& 2014 Elsevier B.V. All rights reserved.
(Tondo et al., 2010; Etain et al., 2012). BP patients are three times more likely to suffer from depressive symptoms than from manic or hypomanic symptoms (Judd et al., 2002, 2003), and many BP-II patients seek help when in the depressed phase (Akiskal et al., 2000). Patients who begin BP with depressive episodes appear to be at increased risk for long-term morbidity, disability, and suicide (Baldessarini et al., 2010, 2012). The misdiagnosis and under-diagnosis of BP are due in part to the ‘soft’ symptoms of bipolarity that characterize patients with nonclassical BP (Katzow et al., 2003). Consequently, the search for bipolar signatures in the symptoms and course of MDD has intensified. Several clinical markers of soft bipolarity have been suggested, including mixed anxiety/depressive symptoms (Katzow et al., 2003), conditions associated with impulsivity such as substance abuse, borderline personality, bulimia, and attention deficit disorder (Regier et al., 1990; Katzow et al., 2003), atypical depressive features (Perugi et al., 1998; Mitchell et al., 2001), early age of onset (Neuman et al., 1997; Benazzi, 2009), recurrent depressive episodes (Benazzi, 2009), a family history of BP in first-degree relatives (Benazzi, 2009), brief major depressive episodes (De Fruyt and Demyttenaere, 2007), antidepressant-induced mania/hypomania (Hirschfeld et al., 2003),
84
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
post-partum depression (Judd et al., 2002), treatment resistance (Woo et al., 2008), and specific affective temperament types. i.e., depressive, cyclothymic, hyperthymic, irritable, and anxious (Rihmer et al., 2010). The concept of ‘bipolar spectrum’ arose from the work of Dunner et al. (1970), whose definition of bipolarity included varying degrees of mania and hypomania. This concept was developed by Klerman (1981) and later by Akiskal (1996) to include the continuum from psychotic mania through other expressions of bipolar I disorder and bipolar II disorder, to soft subsyndromal manifestations of bipolarity, i.e., soft or subthreshold bipolarity (Nusslock and Frank, 2011). Later, Ghaemi et al. (2001, 2002) proposed an approach to the spectrum concept that focuses on how to distinguish BP from unipolar depression. He suggested that the following 11 features may predict the diagnosis of BP in patients with depressive symptoms and proposed that the definition of bipolar spectrum disorder (BPSD) is a potential sign of bipolarity: family history of BP in first-degree relatives; antidepressant-induced mania or hypomania; hyperthymic personality; recurrent major depressive episodes (43); brief major depressive episodes (lasting an average of less than 3 months); atypical depressive symptoms; psychotic major depressive episodes; early age of onset of major depressive episodes (o25 years of age); postpartum depression; antidepressant “wearoff;” and treatment resistance (lack of response to Z3 antidepressant treatment trials). Ghaemi et al. (2004) demonstrated that the proposed criteria can distinguish patients with bipolar depression from those with unipolar depression, but the proposed diagnostic criteria were not tested regarding their ability to predict the conversion from unipolar depression to BP. A large proportion of patients suffering from MDD exhibit underlying soft bipolarity, and the identification of antecedents that predict the conversion from unipolar depression to BP would be highly beneficial for these patients. Thus, a set of patients diagnosed with MDD were retrospectively investigated in an attempt to identify the features and diagnostic criteria proposed for BPSD by Ghaemi et al. (2002) to analyze the predictive performance of the diagnostic criteria overall and of each clinical feature.
2. Methods 2.1. Patients and assessments This study was a retrospective investigation that reviewed the medical records of patients who were hospitalized in the psychiatric ward of Yeouido St. Mary's Hospital, College of Medicine at The Catholic University of Korea in Seoul, Korea, between January 1, 2005 and December 31, 2008, with a diagnosis of MDD without prior history of mania or hypomania. The patients were clinically diagnosed with MDD according to the DSM-IV criteria at the index hospitalization. Those who were admitted for an adverse event or other non-psychiatric diagnostic or environmental etiology were excluded from the analyses. To be eligible for inclusion in this study, the patients were required to have had at least 5 years of follow-up care and to have been diagnosed with MDD or BP at the 5-year follow-up. Patients with insufficient data and those who had a severe comorbid medical or neurological condition that could contribute to depressive symptoms, a diagnosis of a mood disorder due to a general medical condition, or a diagnosis that converted from MDD to another diagnosis other than BP were excluded from the analyses. A diagnosis of BPSD was made based on the proposed diagnostic criteria of Ghaemi et al. (2001), with the exception that postpartum depression was excluded from the criteria to avoid a gender bias that may affect the predictive ability of the criteria. The diagnoses of BPSD at index hospitalization and/or an Axis I
disorder during the follow-up period were independently made by two physicians (IHS and HRW) who were blind to the purpose of the study and who separately evaluated the medical records. Based on the DSM-IV which noted that “manic-like episodes that are clearly caused by somatic antidepressant treatment (e.g., medication, electroconvulsive therapy, and light therapy) should not count toward a diagnosis of bipolar I disorder”, patients who exhibited mania/hypomania only during periods of treatment with causative medications, such as antidepressants, were regarded as patients with MDD when we made the DSM-IV diagnosis. However, those subjects with sufficient manic/hypomanic symptoms to be diagnosed as bipolar disorder during the follow-up period were categorized in the conversion group irrespective of a history of antidepressant-induced mania/hypomania at the index hospitalization or during the follow-up period. Clinical data from the index hospitalization included each item of the diagnostic criteria for BPSD as well as age at index hospitalization and sex. A history of the clinical characteristics, such as age at first psychiatric treatment, age at onset, and number of past major depressive episodes before the subjects visited our hospital, which were investigated at the time of the index hospitalization retrospectively, was also included in the analysis. As this was a retrospective study, all data were obtained during routine psychiatric examinations and treatment. 2.2. Statistical analysis The clinical and demographic variables of patients with a diagnosis of MDD at index hospitalization and the 5-year followup (unipolar group) were compared with the variables of patients whose diagnosis changed from MDD at index hospitalization to BP during the follow-up period (conversion group). A Chi-squared test or Fisher's exact test was used to analyze categorical variables and an independent t-test or Mann–Whitney U-test was used to test continuous variables. To identify the independent predictors of the conversion from MDD to BP, a multinomial logistic regression analysis was conducted for which the independent variables were sex, age, psychiatric comorbidity, and the clinical features of BPSD. Statistical analyses were performed using SAS for Windows (ver. 9.2), and two-tailed p-Values with an alpha level of 0.05 were considered significant, except that a Bonferroni-corrected significance level of po0.0045 (0.05/11 tests) was applied to the univariate analysis that compared the groups in terms of the proportion of subjects with each of the 10 proposed bipolar features in the modified BPSD diagnostic criteria who met the modified BPSD criteria. 2.3. Ethics This study was approved by the institutional review board of Yeouido St. Mary's Hospital in Seoul, Korea, and was conducted according to the Declaration of Helsinki. The institutional review board approved the exemption for informed consent because this was a retrospective chart review study.
3. Results During the study period, 448 patients were discharged with a diagnosis of MDD. Of these patients, 250 (55.8%) fulfilled the eligibility criteria for the study. 3.1. Diagnostic conversion from MDD to BP Of the original 250 patients initially diagnosed with MDD, the conversion rate to BP was 18.4%. The demographic and clinical characteristics of these patients at the index hospitalization are
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
85
Table 1 Characteristics of evaluable patients at index episode. Total (n ¼250) n (%), mean 7SD
Age (years) Sex (female) Age at onset (years) Total number of MDE Psychiatric comorbidity
Unipolar group (n¼204) n (%), mean 7 SD
Conversion group n (%), mean 7 SD
p-Value
BP-I or II (n¼ 46)
BP-I (n¼ 25)
UP vs. BP conversion
BP-II (n ¼21)
UP vs. BP-I
UP vs. BP-II
BP-I vs. BP-II
55.3 7 15.5 192 (76.8%) 48.47 16.0
57.8 7 14.9 157 (77.0%) 51.0 7 15.4
44.1 713.3 35 (76.1%) 36.9 713.5
41.8 7 12.2 46.9 7 14.4 o 0.001 18 (72.0%) 17 (81.0%) 0.899 35.9713.3 38.17 14.0 o 0.001
o 0.001 0.581 o 0.001
0.004 0.677 0.001
0.704 0.478 40.999
1.9 7 2.8
1.6 7 2.7
3.0 73.3
2.0 7 2.0
4.17 4.1
0.011
40.999
o0.001
0.027
17 (6.8%)
10 (4.9%)
7 (15.2%)
4 (16.0%)
3 (14.3%)
0.02
0.052
0.109
40.999
Abbreviations: n; number; SD; standard deviation; MDE; major depressive episode, BP-I; bipolar I disorder, BP-II; bipolar II disorder. Unipolar group; the group of patients whose diagnosis remained major depressive disorder at the 5-year follow-up. Switch group; the group of patients whose diagnosis was changed from major depressive disorder to bipolar disorder during the 5-year follow-up period.
presented (Table 1). There were 35 female patients (76.1%) in the conversion group and 157 female patients (77.0%) in the unipolar group (p¼ 0.899). The mean age at index hospitalization (conversion group¼ 44.1 713.3, unipolar group¼57.8714.9; po 0.001) and the age of onset (conversion group¼ 36.9 713.5, unipolar group¼51.0 715.4; p o0.001) were significantly lower in the conversion group compared with the unipolar group (Table 1). The frequency of occurrence of each proposed bipolar feature and the results of the univariate analyses are provided (Table 2). The proportion of patients with reporting BP in first-degree relatives was significantly higher in the conversion group (52.2%, n¼24) than in the unipolar group (11.3%, n¼23; po0.001, odds ratio [OR]¼8.585, 95% confidence interval [CI]¼4.166–17.693), and a history of antidepressant-induced mania or hypomania was significantly more common in the conversion group (60.9%, n¼28) than in the unipolar group (4.9%, n¼ 10; po0.001, OR¼ 30.178, 95% CI¼ 12.660–71.935. Recurrent major depressive episodes (43) occurred in a significantly greater proportion of patients in the conversion group (45.7%, n¼21) than the unipolar group (15.2%, n¼31; po0.001, OR¼ 4.688, 95% CI¼2.340–9.391), and likewise, brief major depressive episodes (lasting an average of less than 3 months) were also significantly more common in the conversion group (23.9%, n¼11) than the unipolar group (3.9%, n¼ 8; po0.001, OR¼ 7.700, 95% CI¼2.892– 20.499). A significantly greater proportion of patients in the conversion group (58.7%, n¼ 27) exhibited atypical depressive features at index hospitalization relative to the unipolar group (18.1%, n¼ 37; po0.001, OR¼6.414, 95% CI¼3.228–12.743), and every each atypical symptoms were also significantly frequent in conversion group than unipolar group (Table 2). The psychotic features were more common in the conversion group (45.7%, n¼ 21) compared with the unipolar group (7.4%, n¼ 15; po0.001, OR¼10.584, 95% CI¼ 4.837– 23.157). Moreover, early age of onset (prior to the age 25) was significantly more common in the conversion group (43.5%, n¼20) compared with the unipolar group (6.4%, n¼13; po0.001, OR¼ 11.302, 95% CI¼5.030–25.395), and antidepressant wear-off (po0.001, OR¼6.276, 95% CI¼3.086–12.764) and antidepressant resistance (po0.001, OR¼22.121, 95% CI¼10.092–48.486) were also significantly different between the unipolar and conversion group. Only a hyperthymic temperament was not significantly different between the two groups (p¼0.465). Accordingly, the proportion of patients who met the criteria for BPSD was significantly higher in the conversion group (87.0%, n¼40) than the unipolar group (8.3%, n¼17; po0.001, OR¼73.333, 95% CI¼27.212–197.625). The sensitivity of the BPSD diagnosis was 0.870, the specificity was 0.917; the positive predictive value of the BPSD diagnosis was 0.702, and the negative predictive value was 0.969. The multivariate logistic regression included sex, age at index episode, and each of the 11 proposed bipolar features (except for
hyperthymic temperament, which was not significantly different between the groups, and postpartum depression, which was excluded from the analysis) as variables. Psychiatric comorbidity revealed several independent factors correlated with the conversion to BP including a family history of BP in first-degree relatives (p¼ 0.046, OR¼4.304, 95% CI¼1.025–18.061), antidepressant-induced mania or hypomania (po0.001, OR¼42.915, 95% CI¼8.886–207.273), brief major depressive episodes (p¼0.004, OR¼18.660, 95% CI¼2.478– 140.497), early age of onset (p¼0.040, OR¼5.146, 95% CI¼1.077– 27.229), antidepressant wear-off (p¼0.001, OR¼8.967, 95% CI¼ 2.364–34.009), and antidepressant treatment resistance (p¼0.006, OR¼6.751, 95% CI¼1.722–26.465; Table 3). 3.2. Predictors of the diagnostic conversion from MDD to BP-I or BP-II Further subanalyses were conducted to investigate the predictors of conversion from MDD to either BP-I or BP-II. The conversion rates from MDD to BP-I and BP-II were 10.0% (n ¼25) and 8.4% (n ¼21), respectively. The results from the univariate analysis are presented (Table 2); the predictive clinical features for the conversion from MDD to BP-I or BP-II were the same as the predictors of the conversion from MDD to BP. Psychotic features were less common in the BP-II (28.6%, n ¼6) than in the BP-I group (60.0%, n ¼15), but the difference did not reach Bonferroni-corrected significance (p ¼0.033). In the multivariate analysis (Table 4), antidepressant-induced mania/hypomania (po0.001, OR¼59.753, 95% CI¼ 6.870–519.696) and antidepressant wear-off (p¼ 0.014, OR¼ 7.497, 95% CI¼1.496– 37.568) significantly predicted the conversion to BP-I. In addition, age (p¼0.080, OR¼ 0.943, 95% CI¼0.882–1.007), family history of BP (p¼ 0.090, OR¼ 5.312, 95% CI¼0.770–36.628), brief major depressive episodes (p¼0.075, OR¼12.787, 95% CI¼0.776–210.679), and atypical features (p¼0.069, OR¼5.368, 95% CI¼ 0.876–32.893) tended to predict conversion. Predictors of the conversion to BP-II included antidepressant induced mania/hypomania (po0.001, OR¼178.666, 95% CI¼11.035–2892.825), brief major depressive episodes (p¼ 0.013, OR¼28.407, 95% CI¼2.017–399.988), early age at onset (p¼ 0.023, OR¼33.690, 95% CI¼ 1.618–701.580), antidepressant wear-off (p¼0.004, OR¼26.951, 95% CI¼2.892–251.182), and treatment resistance (p¼ 0.005, OR¼13.803, 95% CI¼ 2.164–251.182). In addition, family history of BP showed a trend toward significance (p¼ 0.060, OR¼ 7.750, 95% CI¼ 0.917–65.493). The sensitivity of the BPSD diagnosis for the conversion to BP-I was 0.880, and the specificity was 0.917. The positive predictive value of the BPSD diagnosis for the conversion to BP-I was 0.564, and the negative predictive value was 0.984. The sensitivity of the BPSD diagnosis for the conversion to BP-II was 0.857, and the
86
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
Table 2 Comparisons of characteristics and 10 proposed bipolar features in the bipolar spectrum disorder diagnostic criteria except postpartum depression between unipolar group and conversion group. Unipolar group Total (n ¼204) n (%), (n ¼250) mean 7SD n (%), mean7 SD
47 (18.8%) Family history of bipolar disorder in a 1st-degree relatives 38 (15.2%) Antidepressantinduced mania/ hypomania Hyperthymic 40 (16.0%) temperament
Conversion group n (%), mean 7 SD
UP vs. BP-I
UP vs. BP-II
BP-I vs. BP-II
BP-II p-Value OR BP-I or BP-I (n¼ 25) (n¼ 21) (95% CI) II (n¼ 46)
p-Value OR (95% CI)
p-Value OR (95% CI)
p-Value OR (95% CI)
o 0.001 8.585 (4.166– 17.693) o 0.001 30.178 (12.660– 71.935) 0.465 1.357 (0.596– 3.090) o 0.001 4.688 (2.340– 9.391) o 0.001 7.700 (2.892– 20.499) o 0.001 6.414 (3.228– 12.743) o 0.001 5.118 (2.599– 10.075) o 0.001 9.308 (4.542– 19.077) o 0.001 4.667 (2.362– 9.220) o 0.001 6.009 (2.966– 12.176) 0.001 3.214 (1.601– 6.455) o 0.001 10.584 (4.837– 23.157) o 0.001 11.302 (5.030– 25.395) o 0.001 6.276 (3.086– 12.764) o 0.001 22.121 (10.092– 48.486) o 0.001 73.333 (27.212– 197.625)
o 0.001 7.264 (2.964– 17.806) o 0.001 29.100 (10.474– 80.850) 0.256 1.762 (0.652– 4.763) 0.005 3.720 (1.533– 9.031) o 0.001 9.528 (3.098– 29.300) o 0.001 8.024 (3.292– 19.558) o 0.001 6.375 (2.597– 15.650) o 0.001 5.626 (2.302– 13.750) o 0.001 7.000 (2.911– 16.832) 0.002 4.370 (1.783– 10.715) 0.029 2.813 (1.148– 6.889) o 0.001 18.900 (7.255– 49.234) o 0.001 13.562 (5.167– 35.598) o 0.001 6.320 (2.606– 15.327) o 0.001 22.408 (8.386– 59.880) o 0.001 80.667 (21.886– 297.312)
o 0.001 10.493 (3.989– 27.598) o 0.001 31.525 (10.641– 93.400) 40.999 0.930 (0.258– 3.347) o 0.001 6.139 (2.403– 15.681) 0.017 5.765 (1.573– 21.121) 0.001 4.965 (1.964– 12.552) 0.002 4.000 (1.593– 10.043) o 0.001 17.900 (6.358– 50.396) 0.039 2.872 (1.109– 7.436) o 0.001 8.741 (3.365– 22.701) 0.007 3.750 (1.466– 9.594) 0.007 5.040 (1.707– 14.885) o 0.001 9.041 (3.181– 25.699) o 0.001 6.224 (2.407– 16.095) o 0.001 21.786 (7.632– 62.190) o 0.001 66.000 (17.644– 246.880)
0.536
23 (11.3%)
24 12 12 (52.2%) (48.0%) (57.1%)
10 (4.9%)
28 15 13 (60.9%) (60.0%) (61.9%)
31 (15.2%)
9 6 3 (19.6%) (24.0%) (14.3%)
Recurrent MDEs ( 43)
52 (20.8%) 31 (15.2%)
21 10 11 (45.7%) (40.0%) (52.4%)
Brief MDEs (on average, o3 months) Atypical feature
19 (7.6%)
11 7 4 (23.9%) (28.0%) (19.0%)
8 (3.9%)
64 (25.6%) 37 (18.1%)
27 16 11 (58.7%) (64.0%) (52.4%)
Mood reactivity
80 (32.0%) 51 (25.0%)
29 17 12 (63.0%) (68.0%) (57.1%)
Weight gain or increase in appetite
51 (20.4%)
25 (12.3%)
26 11 15 (56.5%) (44.0%) (71.4%)
Hypersomnia
59 (23.6%) 36 (17.6%)
23 15 8 (50.0%) (60.0%) (38.1%)
Leaden paralysis
49 (19.6%)
22 (47.8%)
Interpersonal rejection sensitivity
52 (20.8%) 34 (16.7%)
18 9 9 (39.1%) (36.0%) (42.9%)
Psychotic feature
36 (14.4%)
15 (7.4%)
21 15 6 (45.7%) (60.0%) (28.6%)
Early age at onset of MDE ( o 25 years)
33 (13.2%)
13 (6.4%)
20 12 8 (43.5%) (48.0%) (38.1%)
Antidepressant wearoff
48 (19.2%)
26 (12.7%)
22 (47.8%)
12 10 (48.0%) (47.6%)
Lack of response to Z 3 antidepressant treatment trials Met the diagnostic criteria of bipolar spectrum disorder
54 (21.6%)
21 (10.3%)
33 (71.7%)
18 15 (72.0%) (71.4%)
57 (22.8%)
17 (8.3%)
40 (87.0%)
22 18 (88.0%) (85.7%)
27 (13.2%)
10 12 (40.0%) (57.1%)
UP vs. BP conversion
1.444 (0.450– 4.641) 0.895 1.083 (0.330– 3.561) 0.478 0.528 (0.114– 2.434) 0.401 1.650 (0.511– 5.327) 0.478 0.605 (0.150– 2.444) 0.425 0.619 (0.189– 2.020) 0.447 0.627 (0.188– 2.095) 0.062 3.182 (0.927– 10.917) 0.139 0.410 (0.125– 1.348) 0.246 2.000 (0.616– 6.494) 0.635 1.333 (0.406– 4.379) 0.033 0.267 (0.077– 0.921) 0.500 0.667 (0.205– 2.169) 0.979 0.985 (0.308– 3.146) 0.966 0.972 (0.268– 3.524) 40.999 0.818 (0.147– 4.557)
Abbreviations: n; number, SD; standard deviation, OR; odds ratio, CI; confidence interval, MDE; major depressive episode, BP-I; bipolar I disorder, BP-II; bipolar II disorder. Unipolar group; the group of patients whose diagnosis remained major depressive disorder at the 5-year follow-up. Conversion group; the group of patients whose diagnosis was changed from major depressive disorder to bipolar disorder during the 5-year follow-up period.
specificity was 0.917. The positive predictive value of the BPSD diagnosis for the conversion to BP-II was 0.514, and the negative predictive value was 0.984.
4. Discussion To our knowledge, this is among the first studies to evaluate the predictive performance of the BPSD diagnostic criteria proposed by Ghaemi et al. (2001) regarding the diagnostic conversion of MDD to BP. As expected, the BPSD diagnostic criteria were highly predictive of the conversion from MDD to BP (OR ¼73.333), MDD to BP-I (OR ¼80.667), and MDD to BP-II (OR ¼66.000) during the
5-year follow-up period. The diagnostic conversion rates found in this study (BP-I: 2.0% per year; BP-II: 1.7% per year) are comparable to the findings of several previous studies in which the conversion rate from MDD to BP-I was 0.4–1.3% per year, and that for MDD to BP-II was 0.5–1.8% per year (Akiskal et al., 1995; Coryell et al., 1995; Goldberg et al., 2001; Angst et al., 2005; Holma et al., 2008). Moreover, the predictive clinical features identified here, including family history of BP, antidepressant-induced mania/hypomania, brief major depressive episodes, early age at onset, antidepressant wear-off, and antidepressant resistance, are in agreement with the findings of previous studies (Neuman et al., 1997; Ghaemi et al., 2002; Hirschfeld et al., 2003; De Fruyt and Demyttenaere, 2007; Benazzi, 2009). However, although these factors were predictive in
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
87
Table 3 Final model of multivariate logistic regression among the unipolar group and conversion group. Significance
Sex (male) Age Psychiatric comorbidity Family history of bipolar disorder in a 1st-degree relatives Antidepressant-induced mania/hypomania Recurrent MDEs ( 43) Brief MDEs (on average, o 3 months) Atypical feature Psychotic feature Early age at onset of MDE ( o25 years) Antidepressant wear-off Lack of response to Z 3 antidepressant treatment trials
OR
0.315 0.108 0.629 0.046 o 0.001 0.720 0.004 0.334 0.981 0.040 0.001 0.006
95% CI
0.429 0.960 0.545 4.304 42.915 1.304 18.660 1.901 1.018 5.416 8.967 6.751
Lower
Upper
0.082 0.913 0.046 1.026 8.886 0.306 2.478 0.516 0.247 1.077 2.364 1.722
2.232 1.009 6.407 18.061 207.273 5.565 140.497 7.000 4.196 27.229 34.009 26.465
Abbreviations: OR; odds ratio, CI; confidence interval, MDE; major depressive episode. Unipolar group; the group of patients whose diagnosis remained major depressive disorder at the 5-year follow-up. Conversion group; the group of patients whose diagnosis was changed from major depressive disorder to bipolar disorder during the 5-year follow-up period. Table 4 Final model of multivariate logistic regression among the unipolar group and bipolar I and bipolar II disorder conversion group. MDD to BP-I conversion Significance
Sex (male) Age Psychiatric comorbidity Family history of bipolar disorder in a 1st-degree relatives Antidepressant-induced mania/hypomania Recurrent MDEs ( 43) Brief MDEs (on average, o 3 months) Atypical feature Psychotic feature Early age at onset of MDE ( o25 years) Antidepressant wear-off Lack of response to Z 3 antidepressant treatment trials
0.354 0.080 0.674 0.090 0.000 0.791 0.075 0.069 0.223 0.315 0.014 0.287
OR
MDD to BP-II conversion 95% CI
0.374 0.943 0.476 5.312 59.753 1.288 12.787 5.368 2.791 2.677 7.497 2.690
Significance
Lower
Upper
0.047 0.882 0.015 0.770 6.870 0.199 0.776 0.876 0.536 0.392 1.496 0.435
2.988 1.007 15.242 36.628 519.696 8.346 210.679 32.893 14.538 18.275 37.568 16.636
0.908 0.642 0.959 0.060 0.000 0.893 0.013 0.330 0.244 0.023 0.004 0.005
OR
0.856 0.983 1.095 7.750 178.666 0.862 28.407 0.341 0.273 33.690 26.951 13.803
95% CI Lower
Upper
0.061 0.917 0.034 0.917 11.035 0.100 2.017 0.039 0.031 1.618 2.892 2.164
12.006 1.055 34.912 65.493 2892.825 7.464 399.988 2.967 2.430 701.580 251.182 88.028
Abbreviations: MDD; major depressive disorder, BP-I; bipolar I disorder, BP-II; bipolar II disorder, OR; odds ratio, CI; confidence interval, MDE; major depressive episode.
the univariate analysis, recurrent major depressive episodes, atypical features, and psychotic features were not significantly predictive in the multivariate analysis. Psychotic features were found to be an accurate predictor of diagnostic conversion by Coryell et al. (1995), Goldberg et al. (2001), and Othmer et al. (2007). However, these findings were based on univariate analyses in two studies (Coryell et al., 1995; Goldberg et al., 2001), whereas Othmer et al. (2007) found only three major risk factors using logistic regression analysis; this may explain this discrepancy among these studies. Psychotic features were also found to be a significant predictor in univariate, but not multivariate analyses, conducted by Holma et al. (2008). In addition, a number of studies have found that atypical features were more common in patients with BP than in patients with MDD (Quitkin et al., 2003; Angst et al., 2006; Akiskal and Benazzi, 2008), but these features were not predictive of a diagnostic conversion in the present study. For instance, Gan et al. (2011) found that some atypical features, such as hypersomnia and irritability, were associated with BP, but others, such as weight gain, were not. Since our study defined atypical features according to DSM-IV criteria, the predictive value of each atypical feature could not be evaluated, which might explain the discrepancy in the results. The presence of recurrent major depressive episodes or a high number of past episodes was shown to be a significant risk factor for the diagnostic conversion from MDD to BP in some studies (Angst et al., 2005; Dudek et al., 2013; Ostergaard et al., 2014), but
not others (Coryell et al., 1995; Gan et al., 2011). Several of these studies (Coryell et al., 1995; Dudek et al., 2013; Ostergaard et al., 2014) compared the mean number of depressive episodes between BP converters and non-converters, whereas recurrent depression was defined as six or more depressive episodes by Angst et al. (2005) and two or more depressive episodes by Gan et al. (2011). In concurrence with Ostergaard et al. (2014) and Dudek et al. (2013), we found that the mean number of major depressive episodes was higher in the BP conversion group than the MDD group. However, in the subanalysis that divided the conversion group into BP-I and BP-II patients, this difference in recurrent depression was significant in the BP-II, but not the BP-I, group. When the definition for recurrent major depressive episodes (43) used in the present study was applied, a significant difference was found for both the BP-I and BP-II groups in the univariate analysis, similar to Angst et al. (2005), but none in the multivariate analysis, as reported by Gan et al. (2011). Hence, the definition of ‘recurrent’ should be carefully considered in the interpretation of the results from relevant studies. Several possible limitations in the present study design should be considered when interpreting these findings, including the retrospective design and small sample. Much of the clinical information was obtained during routine clinical practice, and the possibility of recall bias and reviewer bias needs to be considered. Although the data were collected prospectively after the subjects began psychiatric treatment at our hospital, the clinical characteristics before the
88
Y.S. Woo et al. / Journal of Affective Disorders 174 (2015) 83–88
treatment, including retrospective information from the subjects such as age at first psychiatric treatment, age at onset, and number of past major depressive episodes, could be affected by recall bias. The diagnoses of MDD and BP were made clinically without use of the structured diagnostic interview for Axis I or II disorders or of a scale/ questionnaire for the evaluation of hyperthymic temperament. In addition, because the diagnostic conversion from MDD to BP was observed in only 46 patients due to the small sample size, and only patients with at least 5 years of follow-up care were included in this study, the results are most applicable to patients who are largely adherent to their treatment plan. Moreover, the severity of the depressive or subthreshold hypomanic symptoms was not evaluated; these factors may be another predictor of the conversion to BP (Holma et al., 2008; Fiedorowicz et al., 2011). And the exclusion of MDD patients who were admitted due to an adverse event or other non-psychiatric diagnostic or environmental etiologies and who converted to a diagnosis other than BP during follow-up could decrease the generalizability. Finally, the follow-up period of 5 years is only moderately long for this type of investigation because the diagnostic conversion from MDD to BP may still occur after decades. The present data demonstrate that a substantial portion of MDD patients ultimately convert to BP and that the BPSD diagnostic criteria may predict this conversion with high sensitivity and specificity. In addition, the predictive value of BPSD and several of its clinical features, including a family history of BP, antidepressantinduced mania/hypomania, brief major depressive episodes, early age of onset, antidepressant wear-off, and antidepressant resistance, may be considered independent risk factors for this diagnostic conversion. These predictors should be taken into account in clinical practice because the diagnostic conversion from MDD to BP has clinical importance for the treatment and prognosis of patients.
Role of funding source Nothing declared.
Conflict of interest No conflict declared.
Acknowledgments The authors had no conflict of interest in conducting this study or preparing the manuscript.
References Akiskal, H.S., 1996. The prevalent clinical spectrum of bipolar disorders: beyond DSM-IV. J. Clin. Psychopharmacol. 16, 4S–14S. Akiskal, H.S., Benazzi, F., 2008. Continuous distribution of atypical depressive symptoms between major depressive and bipolar II disorders: dose–response relationship with bipolar family history. Psychopathology 41, 39–42. Akiskal, H.S., Bourgeois, M.L., Angst, J., Post, R., Moller, H., Hirschfeld, R., 2000. Reevaluating the prevalence of and diagnostic composition within the broad clinical spectrum of bipolar disorders. J. Affect. Disord. 59 (Suppl. 1), S5–S30. Akiskal, H.S., Maser, J.D., Zeller, P.J., Endicott, J., Coryell, W., Keller, M., Warshaw, M., Clayton, P., Goodwin, F., 1995. Switching from ‘unipolar’ to bipolar II. An 11-year prospective study of clinical and temperamental predictors in 559 patients. Arch. Gen. Psychiatry 52, 114–123. Angst, J., Gamma, A., Benazzi, F., Silverstein, B., Ajdacic-Gross, V., Eich, D., Rossler, W., 2006. Atypical depressive syndromes in varying definitions. Eur. Arch. Psychiatry Clin. Neurosci. 256, 44–54. Angst, J., Sellaro, R., Stassen, H.H., Gamma, A., 2005. Diagnostic conversion from depression to bipolar disorders: results of a long-term prospective study of hospital admissions. J. Affect. Disord. 84, 149–157. Baldessarini, R.J., Salvatore, P., Khalsa, H.M., Gebre-Medhin, P., Imaz, H., GonzalezPinto, A., Perez, J., Cruz, N., Maggini, C., Tohen, M., 2010. Morbidity in 303 firstepisode bipolar I disorder patients. Bipolar Disord. 12, 264–270. Baldessarini, R.J., Undurraga, J., Vazquez, G.H., Tondo, L., Salvatore, P., Ha, K., Khalsa, H.M., Lepri, B., Ha, T.H., Chang, J.S., Tohen, M., Vieta, E., 2012. Predominant recurrence polarity among 928 adult international bipolar I disorder patients. Acta Psychiatr. Scand. 125, 293–302.
Benazzi, F., 2009. Classifying mood disorders by age-at-onset instead of polarity. Prog. Neuropsychopharmacol. Biol. Psychiatry 33, 86–93. Coryell, W., Endicott, J., Maser, J.D., Keller, M.B., Leon, A.C., Akiskal, H.S., 1995. Longterm stability of polarity distinctions in the affective disorders. Am. J. Psychiatry 152, 385–390. De Fruyt, J., Demyttenaere, K., 2007. Bipolar (spectrum) disorder and mood stabilization: standing at the crossroads? Psychother. Psychosomat. 76, 77–88. Dudek, D., Siwek, M., Zielinska, D., Jaeschke, R., Rybakowski, J., 2013. Diagnostic conversions from major depressive disorder into bipolar disorder in an outpatient setting: results of a retrospective chart review. J. Affect. Disord. 144, 112–115. Dunner, D.L., Gershon, E.S., Goodwin, F.K., 1970. Heritable factors in the severity of affective illness. Sci. Proc. Am. Psychiatr. Assoc. 123, 187–188. Etain, B., Lajnef, M., Bellivier, F., Mathieu, F., Raust, A., Cochet, B., Gard, S., M’Bailara, K., Kahn, J.P., Elgrabli, O., Cohen, R., Jamain, S., Vieta, E., Leboyer, M., Henry, C., 2012. Clinical expression of bipolar disorder type I as a function of age and polarity at onset: convergent findings in samples from France and the United States. J. Clin. Psychiatry 73, e561–e566. Fiedorowicz, J.G., Endicott, J., Leon, A.C., Solomon, D.A., Keller, M.B., Coryell, W.H., 2011. Subthreshold hypomanic symptoms in progression from unipolar major depression to bipolar disorder. Am. J. Psychiatry 168, 40–48. Gan, Z., Diao, F., Wei, Q., Wu, X., Cheng, M., Guan, N., Zhang, M., Zhang, J., 2011. A predictive model for diagnosing bipolar disorder based on the clinical characteristics of major depressive episodes in Chinese population. J. Affect. Disord. 134, 119–125. Ghaemi, S.N., Ko, J.Y., Goodwin, F.K., 2001. The bipolar spectrum and the antidepressant view of the world. J. Psychiatr. Pract. 7, 287–297. Ghaemi, S.N., Ko, J.Y., Goodwin, F.K., 2002. “Cade's disease” and beyond: misdiagnosis, antidepressant use, and a proposed definition for bipolar spectrum disorder. Can. J. Psychiatry 47, 125–134. Ghaemi, S.N., Sachs, G.S., Chiou, A.M., Pandurangi, A.K., Goodwin, K., 1999. Is bipolar disorder still underdiagnosed? Are antidepressants overutilized?. J. Affect. Disord. 52, 135–144. Ghaemi, S.N., Hsu, D.J., Ko, J.Y., Baldassano, C.F., Kontos, N.J., Goodwin, F.K., 2004. Bipolar spectrum disorder: a pilot study. Psychopathology 37, 222–226. Goldberg, J.F., Harrow, M., Whiteside, J.E., 2001. Risk for bipolar illness in patients initially hospitalized for unipolar depression. Am. J. Psychiatry 158, 1265–1270. Grotegerd, D., Suslow, T., Bauer, J., Ohrmann, P., Arolt, V., Stuhrmann, A., Heindel, W., Kugel, H., Dannlowski, U., 2013. Discriminating unipolar and bipolar depression by means of fMRI and pattern classification: a pilot study. Eur. Arch. Psychiatry Clin. Neurosci. 263, 119–131. Hirschfeld, R.M., Calabrese, J.R., Weissman, M.M., Reed, M., Davies, M.A., Frye, M.A., Keck Jr., P.E., Lewis, L., McElroy, S.L., McNulty, J.P., Wagner, K.D., 2003. Screening for bipolar disorder in the community. J. Clin. Psychiatry 64, 53–59. Holma, K.M., Melartin, T.K., Holma, I.A., Isometsa, E.T., 2008. Predictors for switch from unipolar major depressive disorder to bipolar disorder type I or II: a 5-year prospective study. J. Clin. Psychiatry 69, 1267–1275. Judd, L.L., Akiskal, H.S., Schettler, P.J., Coryell, W., Endicott, J., Maser, J.D., Solomon, D.A., Leon, A.C., Keller, M.B., 2003. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch. Gen. Psychiatry 60, 261–269. Judd, L.L., Akiskal, H.S., Schettler, P.J., Endicott, J., Maser, J., Solomon, D.A., Leon, A.C., Rice, J.A., Keller, M.B., 2002. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch. Gen. Psychiatry 59, 530–537. Katzow, J.J., Hsu, D.J., Ghaemi, S.N., 2003. The bipolar spectrum: a clinical perspective. Bipolar Disord. 5, 436–442. Klerman, G.L., 1981. The spectrum of mania. Compr. Psychiatry 22, 11–20. Mitchell, P.B., Wilhelm, K., Parker, G., Austin, M.P., Rutgers, P., Malhi, G.S., 2001. The clinical features of bipolar depression: a comparison with matched major depressive disorder patients. J. Clin. Psychiatry 62, 212–216 (quiz 217). Neuman, R.J., Geller, B., Rice, J.P., Todd, R.D., 1997. Increased prevalence and earlier onset of mood disorders among relatives of prepubertal versus adult probands. J. Am. Acad. Child Adolesc. Psychiatry 36, 466–473. Nusslock, R., Frank, E., 2011. Subthreshold bipolarity: diagnostic issues and challenges. Bipolar Disord. 13, 587–603. Ostergaard, S.D., Straszek, S., Petrides, G., Skadhede, S., Jensen, S.O., MunkJorgensen, P., Nielsen, J., 2014. Risk factors for conversion from unipolar psychotic depression to bipolar disorder. Bipolar Disord. 16, 180–189. Othmer, E., Desouza, C.M., Penick, E.C., Nickel, E.J., Hunter, E.E., Othmer, S.C., Powell, B.J., Hall, S.B., 2007. Indicators of mania in depressed outpatients: a retrospective analysis of data from the Kansas 1500 study. J. Clin. Psychiatry 68, 47–51. Perugi, G., Akiskal, H.S., Lattanzi, L., Cecconi, D., Mastrocinque, C., Patronelli, A., Vignoli, S., Bemi, E., 1998. The high prevalence of “soft” bipolar (II) features in atypical depression. Compr. Psychiatry 39, 63–71. Quitkin, F.M., McGrath, P.J., Stewart, J.W., Klein, D.F., 2003. A reappraisal of atypical depression. Am. J. Psychiatry 160, 798–800 (author reply 800–791). Regier, D.A., Farmer, M.E., Rae, D.S., Locke, B.Z., Keith, S.J., Judd, L.L., Goodwin, F.K., 1990. Comorbidity of mental disorders with alcohol and other drug abuse. Results from the Epidemiologic Catchment Area (ECA) Study. JAMA 264, 2511–2518. Rihmer, Z., Akiskal, K.K., Rihmer, A., Akiskal, H.S., 2010. Current research on affective temperaments. Curr. Opin. Psychiatry 23, 12–18. Tondo, L., Lepri, B., Cruz, N., Baldessarini, R.J., 2010. Age at onset in 3014 Sardinian bipolar and major depressive disorder patients. Acta Psychiatr. Scand. 121, 446–452. Woo, Y.S., Chae, J.H., Jun, T.Y., Kim, K.S., Bahk, W.M., 2008. The bipolar diathesis of treatment-resistant major depressive disorder. Int. J. Psychiatry Clin. Pract. 12, 142–146.
