<<

. 24
( 68 .)



>>

suggesting that a strong bipolarity (long episodes of hypomania) may not be
required to have more AD in BP-II.


Is AD a predictor of BP-II?
Specificity of AD for predicting BP-II was found to be 82.8% (sensitivity 45.3%)
(n ¼ 161, BP-II þ UP). In the discriminant analysis of some predictor variables for
the diagnosis of BP-II (many MDEs, early onset, interpersonal rejection sensitivity,
depressive mixed state, and AD), AD was found to be still significant, suggesting
that it was a strong predictor compared to the other variables (Benazzi, 2001d).
Cross-sectional clinical markers of BP-II could reduce the high underdiagnosis of
BP-II (Ghaemi et al., 2000; Benazzi, 2001e).


Can AD increase the probability of UP switching into hypomania?
UP MDE switchers into hypomania during antidepressant treatment had features
similar to the BP-II MDE switchers (age of onset, frequency of AD: 50.0%; Benazzi,
1997b).


Is there any difference in AD frequency in early-onset versus late-onset BP-II?
In early- (n ¼ 99) versus late-onset (n ¼ 80) BP-II MDE, there was a significant AD
frequency difference (73.7% versus 51.2%), which however was not significant,
controlling for age. This finding suggests that the difference was related to age
(Benazzi, 2000b).


Is there any difference between BP-II AD and UP AD?
BP-II AD (n ¼ 79) versus BP-II non-AD (n ¼ 53) had a significantly lower age of onset.
BP-II AD (n ¼ 79, n ¼ 90) versus UP (n ¼ 42, n ¼ 50) AD had significantly lower age of
onset. Two-way analysis of variance (ANOVA) found that difference in age of onset
between BP-II AD and UP AD was related to diagnosis, and not to AD (Benazzi,
1999a, b, 2000c). A different age of onset (McMahon et al., 1994; Robins and Guze,
1970) runs against the view of AD as a distinct mood disorder (Rabkin et al., 1996), and
is in line with DSM-IV, where AD is a specifier of BP and depressive disorders.


Is there a link between BP-II and UP AD?
BP-II AD (n ¼ 124) versus UP AD (n ¼ 38) did not have a significantly different
age of onset and MDE recurrences, but did have a significantly higher frequency of
137 Atypical depression and bipolar spectrum


depressive mixed-state and BP-II family history. BP-II non-AD (n ¼ 110) versus
UP AD did not have a significantly different age of onset, MDE recurrences,
depressive mixed-state frequency, and significantly higher frequency of a BP-II
family history. UP AD versus UP non-AD (n ¼ 120) had a significantly lower age
of onset (and not significantly different MDE recurrences, depressive mixed-state
frequency, and BP-II family history). These findings support a link between BP-II
and UP AD (Benazzi, 2003a).


What is the relationship between AD and age?
In a sample of 525 (BP-I, BP-II, UP) MDE outpatients (43.6% BP-II, 52.3% UP),
BP-II with age less than 50 years had significantly more AD than BP-II with age
50 years or more (60.9% versus 26.1%, P ¼ 0.0000), and UP with age less than 50
years had significantly more AD than UP with age 50 years or more (34.2%
versus 18.1%, P ¼ 0.0025). Findings suggest that AD frequency decreases with
age (as also shown by Angst et al., 2002). AD frequency was significantly higher
in BP-II MDE versus UP MDE with age less than 50 years (P ¼ 0.0000), but not in
BP-II MDE versus UP MDE with age 50 years or more, suggesting that the higher
frequency of AD in BP-II versus UP is mainly due to young patients. BP-II
frequency was higher in younger compared with older patients (53.4% versus
32.9%, P ¼ 0.0000), and UP frequency was higher in older compared with
younger patients (67.0% versus 46.5%, P ¼ 0.0000; Benazzi, 2001a). Findings
suggest a change in the clinical picture of depression related to age. In a 358 MDE
sample (BP-II þ UP), AD was present in 55.0% of patients under age 60 years
and in 28.1% of patients age 60 years and over (P ¼ 0.0000). BP-II frequency
significantly decreased between the two age groups (56.4% versus 23.9%,
P ¼ 0.0000), while UP frequency significantly increased (P ¼ 0.0000; Benazzi,
2000d). When AD in patients aged 60 years or more was compared to AD in
those under age 60 years, controlling for age, few significant differences were
found.


Are there differences between AD and non-AD?
AD frequency was 38.1% in 467 MDE outpatients. Comparison between AD
(n ¼ 121) and non-AD (n ¼ 133) (BP-II þ UP sample) found that AD had sig-
nificantly more BP-II (65.2% versus 39.8%, P ¼ 0.0000), lower age of onset
(P ¼ 0.0059), more females (76.8% versus 61.6%, P ¼ 0.0089), more axis I co-
morbidity (not including substance abuse: 74.3% versus 57.1%, P ¼ 0.0039;
Benazzi, 1999b). Findings are in line with previous reports (Horwath et al., 1992;
Agosti and Stewart, 2001; Angst et al., 2002; Posternak and Zimmerman, 2002).
138 F. Benazzi



Is there any difference between early-onset and late-onset AD?
Early-onset (before 18 years) versus late-onset AD (BP-II þ UP sample) was
significantly associated with female gender (þ), number of MDE recurrences
(þ), BP-II (þ), UP (À). Early-onset BP-II AD versus late-onset LO BP-II AD
was significantly associated with female gender (þ), number of MDE recurrences,
depression chronicity (þ). The same analysis in UP did not find these associations.
Findings suggest that there are differences between BP-II AD and UP AD (Benazzi,
2000e), further supporting the distinction between BP-II AD and UP AD.


What is the relationship between AD and chronic depression?
No significant difference in AD frequency was found between BP-II (n ¼ 67) and
UP (n ¼ 69) chronic depression (chronic MDE and MDE without full interepisode
recovery, lasting more than 2 years from index MDE). BP-II (chronic and non-
chronic: n ¼ 95) depression versus non-chronic UP depression (n ¼ 81) had sig-
nificantly more AD. Chronic UP could be a link between BP-II and non-chronic
UP, a finding in line with reports suggesting that chronic depression may be
related to BP-II (Akiskal et al., 1995; Coryell et al., 1995). Frequency of depression
chronicity was not significantly different (46.3% versus 40.5%) between AD
(n ¼ 164, BP-II 64.6% of the sample) and non-AD (n ¼ 162, BP-II 35.8% of the
sample). UP AD had significantly more depression chronicity than UP non-AD
and BP-II AD. Chronicity was not significantly different in BP-II AD compared
with BP-II non-AD. Chronic AD versus non-chronic AD had significantly more
UP. Findings suggest that depression chronicity in AD is mainly related to UP, and
may explain why AD was often reported to be chronic (as most previous studies
mainly included UP samples; Rabkin et al., 1996). In BP-II, depression chronicity
findings related to AD seem different in comparison to UP. In BP-II chronic
depression (n ¼ 87), early onset was associated with higher AD frequency. Chronic
depression in old (> 60 years) versus young (< 60 years) (199 BP-II þ 200 UP)
found the frequency of AD to be significantly higher in young patients (59.5%
versus 27.6%), but when the comparison was controlled for age (ANCOVA), the
difference was no longer significant, suggesting that it was related to age (Benazzi,
1999c, d, 2000d, f, 2001f).


Are females more common in AD versus non-AD?
In BP-II MDE (n ¼ 251) versus UP (n ¼ 306) MDE, AD was present in 45.4% versus
25.4% (P ¼ 0.0000). AD was significantly more common in BP-II females versus
males, in UP females versus males, in BP-II females versus UP females, and in
139 Atypical depression and bipolar spectrum


BP-II males versus UP males. Female gender was significantly associated with AD
but not with diagnosis (by logistic regression). The higher frequency of AD in
females versus males was not related to the higher frequency of AD in BP-II versus
UP, but to an association between female gender and AD (Benazzi, 1999e).
Findings are in line with studies showing more females than males in AD
(Agosti and Stewart, 2001; Angst et al., 2002, Posternak and Zimmerman, 2002,
and other studies reviewed by Angst et al., 2002).


Is AD a moderate-severity depression?
In a 536 MDE sample (BP-II n ¼ 241, UP n ¼ 295), severe MDE (n ¼ 219) (defined as
a Global Assessment of Funtioning (GAF) scale score of 50 or less) had AD in 34.7%,
and non-severe MDE had AD in 37.5% (P ¼ 0.5022; Benazzi, 1999f). No BP-II versus
UP differences were found. The GAF severity cut-off of 50 followed the depression
severity definition of Elkin et al. (1989). This finding runs against the report that AD is
often a moderate-severity depression (Kendler et al., 1996), but is in line with the
report that severe AD is not uncommon (35.6% of AD; Sullivan et al., 1998).


Is there a link between depressive mixed state and AD?
Depressive mixed state (defined as a BP-II and UP MDE plus more than two
concurrent hypomanic symptoms) was significantly more common (50.0% versus
20.3%) in AD versus non-AD (BP-II þ UP sample, n ¼ 87), and the association
persisted when controlled for BP-II by logistic regression (BP-II could be a
confounding factor because it was associated with both AD and depressive
mixed state). Among the DSM-IV hypomanic symptoms, AD had significantly
more talkativeness, distractibility, and psychomotor agitation (Benazzi, 2001g).
In a second, larger study, depressive mixed state was highly associated with AD
(odds ratio ¼ 3.1, P ¼ 0.000). UP depressive mixed state (n ¼ 35) versus BP-II
MDE (n ¼ 226) did not have a significantly different age of onset, frequency of
AD, or BP-II family history, while UP depressive mixed states versus UP non-
depressive mixed state MDE (n ¼ 116) had a significantly higher frequency of BP-
II family history. These findings suggest a link between BP-II and UP depressive
mixed state, including a similar frequency of AD (Akiskal and Benazzi, 2003).


Is there a link between female gender and AD in depressive mixed state?
In a depressive mixed-state outpatient sample (n ¼ 205), female gender was sig-
nificantly associated with atypical features (odds ratio ¼ 2.1, P ¼ 0.021; Benazzi,
2003d).
140 F. Benazzi



Is psychomotor retardation more common in AD versus non-AD?
In 95 AD patients (80% BP-II, 20% UP), 21.0% had psychomotor agitation, while
0.0% had retardation. The result may be related to the finding that melancholic
features were not more common in BP-II MDE versus UP MDE outpatients
(19.2% versus 22.6%, n ¼ 182; 20.6% versus 25.0%, n ¼ 161). MDE with psycho-
motor agitation (BP-II þ UP: n ¼ 85) versus MDE without psychomotor agitation
(n ¼ 292) had a significantly higher frequency of AD (51.7% versus 37.3%), while
AD frequency was not significantly different in BP-II-agitated MDE versus UP
agitated MDE (Benazzi, 2000g, 2002b, c; Benazzi, et al., 2002).
These results run against previous studies reporting that psychomotor retardation
was more common in AD versus non-AD (Horwath et al., 1992; Kendler et al., 1996;
Posternak and Zimmerman, 2002), and that melancholic features were more com-
mon in BP versus UP depression (Parker et al., 2000). Different samples (inpatient
versus outpatient, community versus clinical), different study settings (tertiary care
versus non-tertiary care), BP-I and BP-II lumping together, and mainly UP samples
in some studies, may be related to the different findings. Himmelhoch (1999)
reported that melancholic features were only present in 12 of 1100 BP patients.
Prevalence of melancholic features was reported to be higher in the inpatient severe
and psychotic MDE (American Psychiatric Association, 2000).


Is there a link between recurrences and AD?
Comparisons among BP-II MDE (n ¼ 151), highly recurrent UP MDE (> 4
MDEs: n ¼ 57), and low recurrent UP MDE (< 5 MDEs: n ¼ 32) found that AD
was significantly more common in BP-II MDE versus highly recurrent UP MDE,
but also significantly more common in highly recurrent UP MDE versus low
recurrent UP MDE. These findings suggest that recurrences may be related to
increased frequency of AD in UP MDE (Benazzi, 2003c).


What is the relationship between AD and psychotic features?
Frequency of AD was much lower in psychotic BP (I þ II) MDE (6.6%) compared to
the frequency of AD found in other BP-II samples of the author. Psychotic versus non-
psychotic MDE (BP-I þ BP-II þ UP) had significantly less AD (Benazzi, 1999g, h).


What is the relationship between AD and menopause?
Female MDE (BP-I þ BP-II þ UP) with age before 40 years (n ¼ 283), versus
female MDE with age at or after 40 years (n ¼ 63) (the 40-year age cut-off gave a
141 Atypical depression and bipolar spectrum


median age of onset of 52, which is near the median age of onset of menopause, 51
years) found more AD (53.2% versus 31.6%, P ¼ 0.0019) and BP-II (47.7% versus
26.9%, P ¼ 0.0026) before 40 years (in males there was no significant difference in
the BP-II frequency, while AD frequency was significantly reduced after age 40),
suggesting that menopause may change the picture of depression (Benazzi,
2000h).
More recent, published, studies of the author on AD have focused on a defini-
tion of AD based only on the reversed vegetative symptoms (oversleeping,
overeating, weight gain: Benazzi, 2002f), on the impact of AD on trials of anti-
depressants in BP-II (Benazzi, 2004a), on supporting the subtyping of AD into an
early-onset, chronic subtype versus a non-early-onset, non-chronic subtype
(Benazzi, 2004b), on the normal-like distribution of atypical symptoms between
BP-II and UP MDE, supporting a continuity between the two disorders by finding
no zone of rarity (which would be expected because AD is more common in BP-II:
Benazzi, 2003c), on testing the DSM-IV definition of AD (Benazzi, 2003d), and on
further testing the predictive power for BP-II of AD versus other bipolor validators
such as BP family history and depressive mixed state (Benazzi, 2003e).


The author™s last sample study on atypical depression

Study methods
Interviewer
The interviewer was a senior clinical and mood disorder research psychiatrist.

Study setting
The study was carried out in a private outpatient psychiatry center (a University of
California in San Diego (USA) collaborating center). Private practice is more
representative of mood-disorder patients in Italy, where it is the first (or the second,
after family doctors) line of treatment of mood disorders, and national mental
health services and university centers usually treat the most severe patients. Most
individuals can be visited by a private psychiatrist in Italy (reducing a possible
selection bias). Authorities believe that mood-disorder patients in tertiary-care
centers may not be representative of patients who are usually treated in clinical
practice (Akiskal and Pinto, 1999; Goldberg and Kocsis, 1999; Ghaemi et al., 2000;
Post et al., 2001).


Patients and interview
Consecutive UP (major depressive disorder (MDD), MDD superimposed on
dysthymic disorder) and BP-II outpatients, presenting spontaneously for MDE
142 F. Benazzi


treatment, were interviewed. MDD and MDD superimposed on dysthymic
disorder were combined in one group (Angst et al., 2000; Judd and Akiskal,
2000). No psychopharmacotherapy before evaluation avoided the inclusion of
antidepressant-induced mixed states (Akiskal and Pinto, 1999). Current substance
abuse and patients with severe personality disorder were not included (Benazzi,
2000i), as this would confound the diagnosis of BP-II and mixed states (Akiskal
et al., 2000). Clinically significant general medical illness and dementia patients
were not included. All patients were interviewed during the first visit with the
SCID-CV. The SCID-CV is partly semistructured and is based on clinical evalu-
ation (not on simple yes/no answers to structured questions). Clinical evaluation
by clinicians trained in BP-II diagnosis using semistructured interviews resulted

<<

. 24
( 68 .)



>>