Journal of Affective Disorders 204 (2016) 4853

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Research paper

Discrepancy between subjective and objective severity as a predictor

of response to chronotherapeutics in bipolar depression
Masahiro Suzuki a,b,n, Sara Dallaspezia a, Clara Locatelli a, Makoto Uchiyama b,
Cristina Colombo a, Francesco Benedetti a
a
b

Department of Clinical Neurosciences, Scientic Institute and University Vita-Salute San Raffaele, Milan, Italy
Department of Psychiatry, Nihon University School of Medicine, Tokyo, Japan

art ic l e i nf o

a b s t r a c t

Article history:
Received 27 October 2015
Received in revised form
7 June 2016
Accepted 11 June 2016
Available online 15 June 2016

Background: Chronotherapeutic techniques (sleep deprivation and light therapy) are effective treatments
for bipolar depression, but viable predictors of response for the daily clinical practice have not yet been
established. The discrepancy between subjective and objective severity of the depressive syndrome has
been proposed as a possible predictor of treatment outcome in depression. This study examined whether
this discrepancy could predict response to chronotherapeutics in bipolar depression.
Method: We studied 149 consecutively admitted inpatients with a major depressive episode in course of
bipolar disorder. Patients were treated with the combination of repeated sleep deprivation and bright
light therapy. Severity of depression was evaluated using self-rated (Beck Depression Inventory: BDI) and
observer-rated (Hamilton Depression Rating Scale: HDRS) measures. BDI-HDRS discrepancy score at
baseline was calculated, and its associations with clinical response and with depressive cognitive distortions, as measured on the Cognitions Questionnaire, were examined.
Results: Among the 147 completers, 66% responded to treatment (50% reduction of HDRS score). The
response rate in patients with low discrepancy scores and in patients with high discrepancy scores were
80.2% and 48.5%, respectively. High BDI-HDRS discrepancy predicted negative response to treatment with
odds ratio of 3.79 (95%CI: 1.618.93). BDI-HDRS discrepancy was positively associated with depressive
cognitive distortions.
Limitations: Potential factors affecting the discrepancy and outcome other than cognitive distortion were
not examined in this study.
Conclusion: Higher BDI-HDRS discrepancy can predict poorer response to chronotherapeutics in bipolar
depression. The tendency to generalize hopelessness may be a factor inuencing the link between the
discrepancy and outcome.
& 2016 Elsevier B.V. All rights reserved.

Keywords:
Bipolar disorder
Chronotherapeutics
Sleep deprivation
Prediction of outcome
Cognitive distortion
Clinical rating scales

1. Introduction
Chronotherapeutic techniques (total sleep deprivation and light
therapy) can rapidly reverse depressive symptoms in 60% of patients with bipolar depression (Benedetti and Colombo, 2011;
Wirz-Justice et al., 2009), including life-threatening suicidal
symptoms (Benedetti et al., 2014). This approach has very few side
effects (Benedetti and Colombo, 2011; Wirz-Justice et al., 2009),
and switch rate into mania is lower than that observed with antidepressant drugs (Colombo et al., 1999; Leverich et al., 2006).
Chronotherapeutics has then been proposed as a possible rst-line
treatment for bipolar depression (Benedetti et al., 2007; Bunney
n
Correspondence to: Department of Psychiatry, Nihon University School of
Medicine, 30-1 Oyaguchi-Kamicho, Itabashi-ku, Tokyo 173-8610, Japan.
E-mail address: suzuki.masahiro94@nihon-u.ac.jp (M. Suzuki).

http://dx.doi.org/10.1016/j.jad.2016.06.044
0165-0327/& 2016 Elsevier B.V. All rights reserved.

and Bunney, 2013; Dallaspezia and Benedetti, 2011; Wirz-Justice

et al., 2009), and as a model for studying the neurobiological
correlates of rapid antidepressant effects (Gillin et al., 2001;
Schloesser et al., 2012).
In order to optimize treatment options for patients, it is important to establish valid predictors of response to treatment.
Several biological markers have been reported to predict response
to sleep deprivation, a core component of psychiatric chronotherapeutics: an abnormal dexamethasone suppression test
(Kasper et al., 1983); low interleukin-6 levels (Benedetti et al.,
2002); high delta sleep ratio (quotation of slow-wave activity in
the rst to the second non-REM episode) (Nissen et al., 2001);
higher relative metabolic rates in the ventral anterior cingulate,
medial prefrontal cortex, and posterior subcallosal cortex (Wu and
Bunney, 1990); higher frontal cortical excitability (Canali et al.,
2014); and a lower degree of white matter abnormalities (Bollettini et al., 2015). Moreover, response to sleep deprivation was

M. Suzuki et al. / Journal of Affective Disorders 204 (2016) 4853

found to be associated with gene variants affecting the promoter

of the serotonin transporter (Benedetti et al., 1999b), the serotonin
receptor 2A (Benedetti et al., 2008b), the catechol-O-methyl
transferase (Benedetti et al., 2010), and the glycogen synthase kinase-3b promoter (Benedetti et al., 2004).
Although these biological ndings have deepened our understanding of the pathophysiology of depression and of the mechanisms of chronotherapeutics, it is not easy to use them as
predictors in daily clinical practice due to uncertain availability of
laboratory methods, costs, technical requirements, invasiveness,
and time. With regards to clinical predictors, positive diurnal
mood uctuations (Haug, 1992; Reinink et al., 1993), and endogenous subtype (Vogel, 1975) have been reported to predict
response to sleep deprivation, but it takes several days to reliably
evaluate diurnal mood uctuations, and the concept of endogenous subtype is not currently used in most countries.
The Hamilton Depression Rating Scale (HDRS) (Hamilton,
1960), an observer-rated scale, and the Beck Depression Inventory
(BDI) (Beck et al., 1961), a self-report inventory, are the most
commonly used rating scales of depression. Both scales have good
reliability and validity (Rush et al., 2008), but discrepant scores on
these two scales are often observed (Moller and von Zerssen, 1995;
Richter et al., 1998), with many patients under- or over-reporting
depressive symptoms on self-ratings relative to observer ratings.
Rane et al. (2010) examined the predictive effects of this BDI-HDRS
discrepancy on response to further treatment in drug-resistant
depression, and found that high BDI-HDRS discrepancy predicted
delayed treatment response, with an odds ratio of 5.40 (95%CI:
1.6817.35), in a mixed sample of patients with major depressive
disorder (70%) or bipolar depression (19%). The BDI-HDRS discrepancy was then proposed as a predictor of treatment outcome.
Cognitive distortion is a core feature of bipolar depression
(Goodwin and Jamison, 1990). Systematic errors in the perception
and processing of information (Beck, 2011) lead to overestimation
of negative events and underestimation of positive events. It can
be hypothesized that this cognitive bias could inuence BDI-HDRS
discrepancy, but this hypothesis has never been tested. Here we
studied the discrepancy between subjective and objective severity
as a predictor of response to chronotherapeutics, and whether the
typical depressive cognitive distortion associates with BDI-HDRS
discrepancy and with clinical outcome.

2. Methods
2.1. Patients
We studied 149 patients, consecutively admitted to the mood
disorder unit of San Raffaele Hospital in Milano, who met criteria
for a major depressive episode without psychotic features in
course of bipolar disorder type I (DSM-IV criteria, SCID-I interview) (First et al., 1995). All patients had been referred for hospitalization by the psychiatrists in charge, because of treatment-resistant and/or severe depression. The inclusion criterion was Hamilton Depression Rating Scale score (Hamilton, 1960) Z18 at
baseline. Exclusion criteria were other axis I diagnoses, mental
retardation in axis II diagnoses, pregnancy, history of epilepsy,
major medical or neurologic disorders, treatment with long-acting
neuroleptic drugs in the last 3 months before admission, and
history of drug or alcohol dependency or abuse within the last
6 months. After a complete description of the study was given to
the subjects, their written informed consent was obtained. The
study was approved by the local ethical committee.

49

2.2. Treatment
All patients were treated with a combination of total sleep
deprivation and light therapy (Wirz-Justice et al., 2009). According
to an established protocol (Benedetti et al., 2005, 2014; Martiny
et al., 2012), patients were administered 3 consecutive total sleep
deprivation cycles (day 07); each cycle was composed of a 36-h
period of wakefulness. On days 0, 2, and 4, patients were totally
sleep deprived from 07:00 until 19:00 of the following day. They
were then allowed to sleep during the night in the sleep window
19:00 to 08:00 of days 1, 3, and 5. Patients were administered light
therapy (exposure for 30 min to a 10,000 lux bright white light,
color temperature 4600 K) at 03:00 during the total sleep deprivation night and in the morning after recovery sleep, half an hour
after awakening, between 08:00 and 09:00. Light therapy in the
morning was then continued for two weeks. At admission, patients who had been taking lithium (n 74) continued it, and those
who had not been taking lithium (n 73) started it together with
the chronotherapeutic procedure to enhance its effect and prevent
relapse (Baxter, 1985; Benedetti et al., 2008a, 1999a). No other
antidepressant was administered.
2.3. Measures
Severity of depression was rated in the mornings of days 0, 1, 2,
and 6 with the 21-item version of Hamilton Depression Rating
Scale (HDRS-21) and the 13-item version of Beck Depression Inventory (BDI-13). The HDRS-21 was rated by psychiatrists in
charge. FB and CC take full responsibility for the accuracy of all
clinical data. The therapeutic effect of the treatment was assessed
using a modied version of the HDRS-21 (HDRS-NOW) (Leibenluft
et al., 1993), from which items that could not be meaningfully
rated due to the total sleep deprivation procedure were excluded
(i.e., weight changes and insomnia: item numbers 4, 5, 6, and 16).
The HDRS delta score was calculated by subtracting HDRS-NOW
score on day 6 from HDRS-NOW score on day 0. Clinical response
was dened as a Z50% reduction in HDRS-NOW scores on day 6.
BDI-HDRS discrepancy raw score was calculated by the following formula: (BDI-13 total score on day 0)/(39, maximum score
of the BDI-13)  100  (HDRS-21 total score on day 0)/(63, maximum score of the HDRS-21)  100. Using this formula, subjective
severity and objective severity are equally weighted, and each
severity is shown on a scale of one hundred. Hence, the range of
scores on BDI-HDRS discrepancy is  100 to 100. Moreover, given
that the scores of BDI and HDRS are on two different scales, a
standardized score was calculated by performing a z linear
transformation of BDI and HDRS scores, and then subtracting the
standardized HDRS score from the standardized BDI score. This
procedure allowed to minimize the case that one score had a
larger impact on the composite score due purely to its scale. Since
the main purpose of this study is to establish a viable predictor of
response to chronotherapeutics for the daily clinical practice, main
analyses were done using the raw score to ensure comparability to
previous research. The standardized score was used only to conrm the predictive effect of the raw score on outcome.
The depressive cognitive distortion was rated using the Cognitions Questionnaire (CQ) on Day 0 (Fennell and Campbell, 1984).
The CQ is a self-rating scale that assesses specic dimensions of
negative thinking in relation to a number of hypothetical events. In
the CQ, ve dimensions of negative thinking are measured: emotional impact, causal attribution, generalization across time, generalization across situations, and perceived uncontrollability. It
also gives a total score of cognitive distortion and enables the
identication of specic cognitive distortions in response to positive, negative and neutral events.

50

M. Suzuki et al. / Journal of Affective Disorders 204 (2016) 4853

2.4. Statistical analyses

3. Results

Students t-tests were used to compare continuous variables,

and chi-square tests were used to compare categorical variables.
The association between baseline BDI-HDRS discrepancy and
clinical improvement (HDRS-NOW delta score) was determined
using a correlation analysis. Predictive effects of baseline BDIHDRS discrepancy on response to treatment were examined using
a logistic regression analysis. This analysis was performed, both
using raw scores and using standardized scores for discrepancy.
We then performed a receiver operating characteristics (ROC)
analysis on raw discrepancy scores to obtain an unbiased estimate
of the area under the curve (AUC) and to dene the optimal cut-off
point when equal weight is given to sensitivity and specicity
(Youden Index) (Ruopp et al., 2008). Subsequently, the patients
were divided into two groups, above and below the ROC-derived
cut-off point. As mentioned later (see Section 3), there was a signicant difference between the two groups in the duration of
current depressive episode. Therefore, the difference in clinical
course between the two groups was analyzed using a two-way
(time  group) repeated measures ANCOVA on HDRS-NOW, controlling for duration of current episode. A Huynh-Feldt adjustment
was employed when a Mauchly's sphericity test reached signicance. In order to explore the clinical utility of this cut off point
for a prediction of response to chronotherapeutics, we also carried
out a logistic regression, using dichotomous discrepancy variable
(low or high).
The association between BDI-HDRS discrepancy and cognitive
distortion (total score and the 5 dimensions) was determined
using a correlation analysis. The association between cognitive
distortion and clinical improvement (HDRS-NOW delta score) was
also examined using a correlation analysis. Statistical signicance
was set at p o0.05. All statistical analyses were performed using
SPSS version 22.0 (IBM Corporation, Armonk, NY, USA).

One hundred forty-seven patients completed the treatment.

Two patients (1.3%) discontinued the treatment because of
switching polarity, and were then treated with benzodiazepine
and a second mood stabilizer, thus restoring euthymia. The clinical
and demographic characteristics of the 147 completers are summarized in Table 1. The response (50% reduction in HDRS-NOW
score on day 6) rate was 66.0%, conrming previous results
(Benedetti and Colombo, 2011; Colombo et al., 1999; Wirz-Justice
et al., 2009). The response rate in patients who had been taking
lithium before admission (63.5%) was not signicantly different
from that in patients who started lithium after admission (68.5%)
(2 0.41, p 0.602).
3.1. BDI-HDRS discrepancy and clinical outcome
No patient had missing data for BDI or HDRS. Baseline BDI
scores showed a negative association with the HDRS-NOW delta
score (r 0.33, p o0.001), while baseline HDRS scores showed a
positive association with the HDRS-NOW delta score (r 0.35,
po 0.001). The BDI-HDRS discrepancy scores ranged from  34.6
to 59.9, and were normally distributed (Shapiro-Wilk W0.99).
They directly correlated with BDI scores (r 0.94, p o0.001), and
inversely correlated with HDRS scores (r 0.17, p 0.040).
The correlation analysis showed a signicant negative association between the BDI-HDRS discrepancy score and the HDRSNOW delta score (r 0.44, p o0.001) (Fig. 1): the higher the
discrepancy score, the poorer the clinical improvement. In the
logistic regression analysis, adjusting for demographic variables
(age, gender, education, age at onset, duration of current episode,
number of previous depressive episodes, and number of previous
manic episodes), raw BDI-HDRS discrepancy scores predicted
clinical response (odds ratio: 0.97; 95%CI 0.940.99; p 0.007).
When using standardized discrepancy scores instead of raw scores,
the same result was obtained (odds ratio: 0.72; 95%CI 0.550.96;
p 0.025).

Table 1
Clinical and demographic characteristics of the high discrepancy group and the low discrepancy group.
Demographic details

Age, y
Gender (M/F)
Education, y
Age at onset, y
Duration of illness, y
Duration of current episode, wk
No. of previous depressive episodes
No. of previous manic episodes
Total no. of previous recurrences
Previous lithium treatment (Yes/No)
21-item HDRS baseline score
HDRS-NOW day 0
HDRS-NOW day 1
HDRS-NOW day 2
HDRS-NOW day 6
13-item BDI baseline score
BDI-HDRS discrepancy score
CQ total score

Mean (SD) or n
Whole sample (n 147)

High discrepancy (n 66)

Low discrepancy (n 81)

47.9 (11.5)
56/91
11.4 (4.2)
32.3 (10.4)
15.6 (10.2)
24.0 (26.9)
5.9 (6.0)
3.3 (3.5)
9.2 (8.4)
74/73
23.9 (4.2)
20.6 (4.1)
13.8 (6.3)
12.2 (6.2)
8.3 (6.8)
16.2 (7.6)
3.7 (19.5)
25.1 (11.0)a

47.4 (10.8)
20/46
11.7 (4.1)
31.6 (10.6)
15.7 (11.4)
30.1 (32.6)
5.1 (4.5)
2.7 (2.0)
7.9 (5.9)
36/30
23.4 (4.1)
20.0 (3.9)
14.4 (6.2)
13.5 (5.5)
11.0 (6.2)
22.6 (4.9)
20.8 (11.8)
30.1 (9.8)b

48.4 (12.2)
36/45
11.2 (4.2)
32.9 (10.2)
15.4 (9.1)
19.0 (20.0)
6.6 (7.0)
3.8 (4.3)
10.4 (9.9)
38/43
24.3 (4.2)
21.0 (4.3)
13.3 (6.4)
11.1 (6.6)
6.1 (6.5)
11.0 (5.0)
 10.3 (11.8)
21.0 (10.3)c

t or 2

 0.52
3.08
0.84
 0.76
0.18
2.20
 1.51
 1.88
 1.86
0.85
 1.37
 1.47
1.06
2.29
4.60
14.03
5.23
15.89

NS
NS
NS
NS
NS
0.031
NS
NS
NS
NS
NS
NS
NS
0.024
o 0.001
o 0.001
o 0.001
o 0.001

Abbreviations: HDRS-NOW modied version of Hamilton Depression Rating Scale, BDI Beck Depression Inventory, CQ Cognitions Questionnaire, NS non-signicant,
SD Standard Deviation, M male, F female, y years, wk weeks.
a
b
c

n 135.
n 62.
n 73.

M. Suzuki et al. / Journal of Affective Disorders 204 (2016) 4853

51

Fig. 1. Correlation between BDI-HDRS discrepancy and clinical improvement.

3.2. Low discrepancy group vs. high discrepancy group

The estimated AUC at the ROC analysis was 0.684, with an
optimal cut-off point of 6.78, yielding a sensitivity/specicity of
0.68/0.67. Duration of current episode in patients with high discrepancy scores was signicantly longer than that in patients with
low discrepancy scores (p 0.03), but other demographic variables
were not signicantly different between the two groups (Table 1).
Fig. 2 shows the time course of the mean HDRS-NOW scores in
the two groups. The two-way repeated measures ANCOVA on
HDRS-NOW scores, controlling for duration of current episode,
showed a signicant effect of time  group interaction (F8.13;
df 2.92, 321.02; p o0.001), meaning that patients with low discrepancy scores (r 6.78) showed greater improvement than patients with high discrepancy scores ( 46.78). The response rate in
patients with low discrepancy scores and in patients with high
discrepancy scores were 80.2% and 48.5%, respectively (2 16.35,
p o0.001). In the logistic regression analysis, adjusting for demographic variables (age, gender, education, age at onset, duration of
current episode, number of previous depressive episodes, and
number of previous manic episodes), low BDI-HDRS discrepancy
(r 6.78) predicted response to treatment with odds ratio of 3.79
(95%CI 1.618.93; p 0.002).
3.3. BDI-HDRS discrepancy and cognitive distortion
One-hundred and thirty-ve patients completed the CQ. The
correlation analysis revealed a positive association between the
BDI-HDRS discrepancy score and the CQ total score (r 0.41,
p o0.001). Furthermore, all ve dimensions of the CQ were signicantly correlated with the BDI-HDRS discrepancy score
(emotional impact: r 0.28, p 0.001; causal attribution:
r 0.20, p 0.023; generalization across time: r 0.41, po 0.001;
generalization across situations: r 0.40, po 0.001; perceived
uncontrollability: r 0.22, p 0.012).
3.4. Cognitive distortion and clinical improvement
The CQ total score did not show signicant association with the
HDRS-NOW delta score (r  0.16, p 0.057). In the sub-scale
analysis, the dimension generalization across situations was negatively associated with improvement (r  0.24, p 0.006), but
the other dimensions did not show signicant associations with

Fig. 2. Time course of the mean HDRS-NOW scores in the two groups based on the
ROC-derived cut-off point of the BDI-HDRS discrepancy score. Data are expressed
as means 7 SD.

the HDRS-NOW delta score (emotional impact: r  0.11,

p 0.204; causal attribution: r  0.13, p 0.878; generalization
across time: r  0.15, p 0.083; perceived uncontrollability:
r 0.09, p 0.330).

4. Discussion
This is the rst study to report the association of baseline discrepancy between subjective and objective severity of depression,
with response to chronotherapeutics in bipolar depression. The
BDI-HDRS discrepancy was signicantly associated with improvement of depressive symptoms: the higher the discrepancy
score, the smaller the clinical improvement. Consistent with this
correlation, response rate of the low discrepancy ( r6.78) group
was signicantly higher than that of the high discrepancy ( 46.78)
group (80.2% vs. 48.5%). High BDI-HDRS discrepancy predicted
negative response to treatment with odds ratio of 3.79.
The BDI-HDRS discrepancy has clear advantages over other
predictors previously reported, being a low cost, quickly obtained,
relatively non-invasive measure. Several measures of neuroendocrinology, neurophysiology, neuroimaging, genetics, and other
clinical characteristics can predict response to sleep deprivation
(Dallaspezia and Benedetti, 2015), but in common clinical settings
have proven rather unwieldy because of costs, technical requirements, invasiveness, and time. The BDI-HDRS discrepancy can
then be proposed as a candidate predictor of antidepressant response to be used in a wide variety of clinical settings.
Furthermore, we found that all the dimensions of cognitive
distortions measured by the CQ (emotional impact, causal attribution, generalization across time, generalization across situations,

52

M. Suzuki et al. / Journal of Affective Disorders 204 (2016) 4853

and perceived uncontrollability) associated with discrepancy between subjective and objective severity of depression. Cognitive
distortions are a regular core symptom of bipolar depression,
leading to mood-congruent biases in information processing that
inuence evaluative processes, social judgments, decision-making,
attention and memory (Murphy and Sahakian, 2001). Our present
observation is compatible with previous reports, which showed
that the item content of the BDI and the HDRS considerably differ,
with the BDI weighing depressive cognitions and the HDRS
weighting somatic/vegetative symptoms (Uher et al., 2008).
The literature suggests that cognitive distortions affect the
clinical course of depression. Williams et al. (1990) showed that
individuals with higher baseline levels of dysfunctional attitudes
were more likely to have persistent depressive symptoms at
6 weeks. Bothwell and Scott (1997) also found that higher levels of
dysfunctional attitudes signicantly predicted the chronicity of
depression. In this study, the CQ total score did not signicantly
correlate with the HDRS-NOW delta score (r  0.16, p 0.057),
but in the sub-scale analysis, the dimension generalization across
situations showed negative correlation with the HDRS-rated
symptom improvement (r  0.24, p 0.006). Generalization
across situations relates to hopelessness, and our result suggests
that this dimension of negative thinking might be related to antidepressant response. However, this correlation was milder than
that of the BDI-HDRS discrepancy (r  0.24 vs r  0.44, twosided p 0.053). Therefore, we speculate that the BDI-HDRS discrepancy could catch something more, in relation to antidepressant response, than the CQ.
To date, a number of demographic and clinical factors have
been reported to be associated with discrepancy between subjective and objective severity of depression. Younger age and
higher educational attainment are predictive of higher self-ratings
of depression relative to observer-ratings (Enns et al., 2000). Patients who over-report the severity of their depression show
personality traits of low extraversion (Enns et al., 2000; Schneibel
et al., 2012), low agreeableness (Corruble et al., 1999; Enns et al.,
2000), low openness to experience (Duberstein and Heisel, 2007),
low self-esteem (Domken et al., 1994), and high neuroticism (Enns
et al., 2000; Paykel and Prusoff, 1973; Schneibel et al., 2012).
Moreover, patients with non-endogenous or neurotic depression
tend to rate their depression as more severe (Domken et al., 1994;
Rush et al., 1987), while patients showing melancholic features
tend to rate their depression as less severe (Prusoff et al., 1972;
Rush et al., 1987) than the observers. To elucidate why the discrepancy is predictive of response to chronotherapeutics, these
factors need to be comprehensively examined in future studies.
Also, further study is needed to evaluate the possible relationship
between BDI-HDRS discrepancy and the functional and structural
MRI predictors of antidepressant response (see Section 1).
Our results must be viewed in light of some methodological
limitations. First, there was no control group, because it is difcult
to design a control condition for chronotherapeutics, and a placebo
effect cannot then be excluded. Second, a lack of detailed data
about previous medications, other than lithium, prevented us from
assessing their possible effects on response to chronotherapeutics
and on clinical ratings. Third, according to current literature, we
rated response on HDRS, which became then part of both the
predictor (HDRS-BDI discrepancy) and the outcome (HDRS delta).
Future research will test the predictive value of discrepancy ratings in respect to other outcome measure. Finally, potential factors
affecting the discrepancy and outcome other than cognitive distortion were not examined, and further study is then required to
clarify what factors contribute to the predictive ability of BDIHDRS discrepancy. It is expected that such ndings would improve
our understanding of treatment-resistant depression.
In conclusion, BDI-HDRS discrepancy is a useful predictor of

response to combination chronotherapeutics (repeated sleep deprivation and light therapy) in bipolar depression, and can be used
in daily clinical practice. Cognitive distortion is associated with the
discrepancy, and the tendency to generalize hopelessness might
be a factor inuencing the link between BDI-HDRS discrepancy
and outcome.

References
Baxter Jr., L.R., 1985. Can lithium carbonate prolong the antidepressant effect of
sleep deprivation? Arch. Gen. Psychiatry 42, 635.
Beck, A.T., Ward, C.H., Mendelson, M., Mock, J., Erbaugh, J., 1961. An inventory for
measuring depression. Arch. Gen. Psychiatry 4, 561571.
Beck, J.S., 2011. Cognitive Behavior Therapy: Basics And Beyond. Guilford Press,
New York.
Benedetti, F., Barbini, B., Bernasconi, A., Fulgosi, M.C., Campori, E., Colombo, C.,
Dallaspezia, S., Lorenzi, C., Pontiggia, A., Radaelli, D., Smeraldi, E., 2008a. Lithium overcomes the inuence of 5-HTTLPR gene polymorphism on antidepressant response to sleep deprivation. J. Clin. Psychopharmacol. 28,
249251.
Benedetti, F., Barbini, B., Bernasconi, A., Fulgosi, M.C., Colombo, C., Dallaspezia, S.,
Gavinelli, C., Marino, E., Pirovano, A., Radaelli, D., Smeraldi, E., 2008b. Serotonin
5-HT2A receptor gene variants inuence antidepressant response to repeated
total sleep deprivation in bipolar depression. Prog. Neuropsychopharmacol.
Biol. Psychiatry 32, 18631866.
Benedetti, F., Barbini, B., Colombo, C., Smeraldi, E., 2007. Chronotherapeutics in a
psychiatric ward. Sleep Med. Rev. 11, 509522.
Benedetti, F., Barbini, B., Fulgosi, M.C., Colombo, C., Dallaspezia, S., Pontiggia, A.,
Smeraldi, E., 2005. Combined total sleep deprivation and light therapy in the
treatment of drug-resistant bipolar depression: acute response and long-term
remission rates. J. Clin. Psychiatry 66, 15351540.
Benedetti, F., Colombo, C., 2011. Sleep deprivation in mood disorders. Neuropsychobiology 64, 141151.
Benedetti, F., Colombo, C., Barbini, B., Campori, E., Smeraldi, E., 1999a. Ongoing lithium treatment prevents relapse after total sleep deprivation. J. Clin. Psychopharmacol. 19, 240245.
Benedetti, F., Dallaspezia, S., Colombo, C., Lorenzi, C., Pirovano, A., Smeraldi, E., 2010.
Effect of catechol-O-methyltransferase Val(108/158)Met polymorphism on antidepressant efcacy of uvoxamine. Eur. Psychiatry 25, 476478.
Benedetti, F., Lucca, A., Brambilla, F., Colombo, C., Smeraldi, E., 2002. Interleukin-6
serum levels correlate with response to antidepressant sleep deprivation and
sleep phase advance. Prog. Neuropsychopharmacol. Biol. Psychiatry 26,
11671170.
Benedetti, F., Riccaboni, R., Locatelli, C., Poletti, S., Dallaspezia, S., Colombo, C., 2014.
Rapid treatment response of suicidal symptoms to lithium, sleep deprivation,
and light therapy (chronotherapeutics) in drug-resistant bipolar depression. J.
Clin. Psychiatry 75, 133140.
Benedetti, F., Serretti, A., Colombo, C., Campori, E., Barbini, B., di Bella, D., Smeraldi,
E., 1999b. Inuence of a functional polymorphism within the promoter of the
serotonin transporter gene on the effects of total sleep deprivation in bipolar
depression. Am. J. Psychiatry 156, 14501452.
Benedetti, F., Serretti, A., Colombo, C., Lorenzi, C., Tubazio, V., Smeraldi, E., 2004. A
glycogen synthase kinase 3-beta promoter gene single nucleotide polymorphism is associated with age at onset and response to total sleep deprivation in
bipolar depression. Neurosci. Lett. 368, 123126.
Bollettini, I., Poletti, S., Locatelli, C., Vai, B., Smeraldi, E., Colombo, C., Benedetti, F.,
2015. Disruption of white matter integrity marks poor antidepressant response
in bipolar disorder. J. Affect. Disord. 174, 233240.
Bothwell, R., Scott, J., 1997. The inuence of cognitive variables on recovery in
depressed inpatients. J. Affect. Disord. 43, 207212.
Bunney, B.G., Bunney, W.E., 2013. Mechanisms of rapid antidepressant effects of
sleep deprivation therapy: clock genes and circadian rhythms. Biol. Psychiatry
73, 11641171.
Canali, P., Sferrazza Papa, G., Casali, A.G., Schiena, G., Fecchio, M., Pigorini, A.,
Smeraldi, E., Colombo, C., Benedetti, F., 2014. Changes of cortical excitability as
markers of antidepressant response in bipolar depression: preliminary data
obtained by combining transcranial magnetic stimulation (TMS) and electroencephalography (EEG). Bipolar Disord. 16, 809819.
Colombo, C., Benedetti, F., Barbini, B., Campori, E., Smeraldi, E., 1999. Rate of switch
from depression into mania after therapeutic sleep deprivation in bipolar depression. Psychiatry Res. 86, 267270.
Corruble, E., Legrand, J.M., Zvenigorowski, H., Duret, C., Guel, J.D., 1999. Concordance between self-report and clinicians assessment of depression. J. Psychiatr. Res. 33, 457465.
Dallaspezia, S., Benedetti, F., 2011. Chronobiological therapy for mood disorders.
Expert Rev. Neurother. 11, 961970.
Dallaspezia, S., Benedetti, F., 2015. Sleep deprivation therapy for depression. Curr.
Top. Behav. Neurosci. 25, 483502.
Domken, M., Scott, J., Kelly, P., 1994. What factors predict discrepancies between
self and observer ratings of depression? J. Affect. Disord. 31, 253259.
Duberstein, P.R., Heisel, M.J., 2007. Personality traits and the reporting of affective
disorder symptoms in depressed patients. J. Affect. Disord. 103, 165171.

M. Suzuki et al. / Journal of Affective Disorders 204 (2016) 4853

Enns, M.W., Larsen, D.K., Cox, B.J., 2000. Discrepancies between self and observer
ratings of depression. The relationship to demographic, clinical and personality
variables. J. Affect. Disord. 60, 3341.
Fennell, M.J., Campbell, E.A., 1984. The cognitions questionnaire: specic thinking
errors in depression. Br. J. Clin. Psychol. 23 (Pt 2), 8192.
First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B., 1995. Structured Clinical Interview for DSM-IV Axis I Disorders Patient Edition (SCID I/P, Version 2.0).
Biometrics Research Department, New York State Psychiatric Institute, New
York.
Gillin, J.C., Buchsbaum, M., Wu, J., Clark, C., Bunney Jr., W., 2001. Sleep deprivation
as a model experimental antidepressant treatment: ndings from functional
brain imaging. Depress. Anxiety 14, 3749.
Goodwin, F.K., Jamison, K.R., 1990. Manic-Depressive Illness. Oxford University
Press, New York.
Hamilton, M., 1960. A rating scale for depression. J. Neurol. Neurosurg. Psychiatry
23, 5662.
Haug, H.J., 1992. Prediction of sleep deprivation outcome by diurnal variation of
mood. Biol. Psychiatry 31, 271278.
Kasper, S., Moises, H.W., Beckmann, H., 1983. Dexamethasone suppression test
combined with total sleep deprivation in depressed patients. Psychiatr. Clin. 16,
1725.
Leibenluft, E., Moul, D.E., Schwartz, P.J., Madden, P.A., Wehr, T.A., 1993. A clinical
trial of sleep deprivation in combination with antidepressant medication.
Psychiatry Res. 46, 213227.
Leverich, G.S., Altshuler, L.L., Frye, M.A., Suppes, T., McElroy, S.L., Keck Jr., P.E., Kupka,
R.W., Denicoff, K.D., Nolen, W.A., Grunze, H., Martinez, M.I., Post, R.M., 2006.
Risk of switch in mood polarity to hypomania or mania in patients with bipolar
depression during acute and continuation trials of venlafaxine, sertraline, and
bupropion as adjuncts to mood stabilizers. Am. J. Psychiatry 163, 232239.
Martiny, K., Refsgaard, E., Lund, V., Lunde, M., Sorensen, L., Thougaard, B., Lindberg,
L., Bech, P., 2012. A 9-week randomized trial comparing a chronotherapeutic
intervention (wake and light therapy) to exercise in major depressive disorder
patients treated with duloxetine. J. Clin. Psychiatry 73, 12341242.
Moller, H.J., von Zerssen, D., 1995. Self-rating procedures in the evaluation of antidepressants. Psychopathology 28, 291306.
Murphy, F.C., Sahakian, B.J., 2001. Neuropsychology of bipolar disorder. Br. J. Psychiatry 178, S120S127.
Nissen, C., Feige, B., Konig, A., Voderholzer, U., Berger, M., Riemann, D., 2001. Delta
sleep ratio as a predictor of sleep deprivation response in major depression. J.
Psychiatr. Res. 35, 155163.
Paykel, E.S., Prusoff, B.A., 1973. Relationships between personality dimensions:
neuroticism and extraversion against obsessive, hysterical and oral personality.

53

Br. J. Soc. Clin. Psychol. 12, 309318.

Prusoff, B.A., Klerman, G.L., Paykel, E.S., 1972. Concordance between clinical assessments and patients self-report in depression. Arch. Gen. Psychiatry 26,
546552.
Rane, L.J., Fekadu, A., Wooderson, S., Poon, L., Markopoulou, K., Cleare, A.J., 2010.
Discrepancy between subjective and objective severity in treatment-resistant
depression: prediction of treatment outcome. J. Psychiatr. Res. 44, 10821087.
Reinink, E., Bouhuys, A.L., Gordijn, M.C., Van Den Hoofdakker, R.H., 1993. Prediction
of the antidepressant response to total sleep deprivation of depressed patients:
longitudinal versus single day assessment of diurnal mood variation. Biol.
Psychiatry 34, 471481.
Richter, P., Werner, J., Heerlein, A., Kraus, A., Sauer, H., 1998. On the validity of the
Beck Depression Inventory. A review. Psychopathology 31, 160168.
Ruopp, M.D., Perkins, N.J., Whitcomb, B.W., Schisterman, E.F., 2008. Youden Index
and optimal cut-point estimated from observations affected by a lower limit of
detection. Biom. J. 50, 419430.
Rush, A.J., First, M.B., Blacker, D., 2008. Handbook of Psychiatric Measures. American Psychiatric Pub, Washington, DC.
Rush, A.J., Hiser, W., Giles, D.E., 1987. A comparison of self-reported versus clinician-related symptoms in depression. J. Clin. Psychiatry 48, 246248.
Schloesser, R.J., Martinowich, K., Manji, H.K., 2012. Mood-stabilizing drugs: mechanisms of action. Trends Neurosci. 35, 3646.
Schneibel, R., Brakemeier, E.L., Wilbertz, G., Dykierek, P., Zobel, I., Schramm, E.,
2012. Sensitivity to detect change and the correlation of clinical factors with the
Hamilton Depression Rating Scale and the Beck Depression Inventory in depressed inpatients. Psychiatry Res. 198, 6267.
Uher, R., Farmer, A., Maier, W., Rietschel, M., Hauser, J., Marusic, A., Mors, O., Elkin,
A., Williamson, R.J., Schmael, C., Henigsberg, N., Perez, J., Mendlewicz, J., Janzing, J.G., Zobel, A., Skibinska, M., Kozel, D., Stamp, A.S., Bajs, M., Placentino, A.,
Barreto, M., McGufn, P., Aitchison, K.J., 2008. Measuring depression: comparison and integration of three scales in the GENDEP study. Psychol. Med. 38,
289300.
Vogel, G.W., 1975. A review of REM sleep deprivation. Arch. Gen. Psychiatry 32,
749761.
Williams, J.M., Healy, D., Teasdale, J.D., White, W., Paykel, E.S., 1990. Dysfunctional
attitudes and vulnerability to persistent depression. Psychol. Med. 20, 375381.
Wirz-Justice, A., Benedetti, F., Terman, M., 2009. Chronotherapeutics for Affective
Disorders. A Clinicians Manual for Light and Wake Therapy. Karger, Basel.
Wu, J.C., Bunney, W.E., 1990. The biological basis of an antidepressant response to
sleep deprivation and relapse: review and hypothesis. Am. J. Psychiatry 147,
1421.