Just Accepted by The World Journal of Biological Psychiatry

C-reactive protein: a differential biomarker for major de-

pressive disorder and bipolar II disorder
Hui Hua Chang, Tzu-Yun Wang, I Hui Lee, Sheng-Yu Lee, Kao Chin Chen, San-Yuan
Huang, Yen Kuang Yang, Ru-Band Lu, Po See Chen

Doi: 10.3109/15622975.2016.1155746

Abstract
Objective. We aimed to examine whether the C-reactive protein (CRP) level could
be used to differentiate between major depressive disorder (MDD) and bipolar II dis‐
order (BD II).
Methods. Ninety-six healthy controls, 88 BD II and 72 MDD drug-naïve patients in
their major depressive episodes were enrolled. The fasting plasma level of high-sen‐
sitivity CRP was assessed at baseline and after treatment.
Results. The BD II patients presented significantly higher HDRS scores and CRP
levels at baseline when adjustment for age, gender, and body mass index (BMI)
(p<0.001 and p<0.001, respectively). After treatment the CRP levels remained significantly different (p<0.001), thought the
HDRS score was not significantly different between the BD II and MDD patients. A receiver-operating characteristic analysis
showed that a baseline CRP level of 621.6 ng/mL could discriminate between BD II and MDD, with an area under the curve
of 0.816 and a sensitivity and specificity of 0.699 and 0.882, respectively. Furthermore, the baseline CRP level greater than
621.6 ng/mL had 28.2 higher odds of a diagnosis of BD II (p<0.001, 95% CI: 10.96–72.35).
Conclusions. The level of CRP plays a role of biomarker to differentiate between MDD and BD II depression in both their
depressed and euthymic state.

© 2016 Taylor & Francis. This provisional PDF corresponds to the article as it appeared upon acceptance. Fully formatted
PDF and full text (HTML) versions will be made available soon.
DISCLAIMER: The ideas and opinions expressed in the journal's Just Accepted articles do not necessarily reflect those of Taylor & Francis (the Publisher),
the Editors or the journal. The Publisher does not assume any responsibility for any injury and/or damage to persons or property arising from or related to
any use of the material contained in these articles. The reader is advised to check the appropriate medical literature and the product information currently
provided by the manufacturer of each drug to be administered to verify the dosages, the method and duration of administration, and contraindications. It is
the responsibility of the treating physician or other health care professional, relying on his or her independent experience and knowledge of the patient, to
determine drug dosages and the best treatment for the patient. Just Accepted articles have undergone full scientific review but none of the additional editorial
preparation, such as copyediting, typesetting, and proofreading, as have articles published in the traditional manner. There may, therefore, be errors in
Just Accepted articles that will be corrected in the final print and final online version of the article. Any use of the Just Accepted articles is subject to the
express understanding that the papers have not yet gone through the full quality control process prior to publication.
C-reactive protein: a differential biomarker for major depressive disorder and
bipolar II disorder
Running head: CRP level in MDD and BD II

Hui Hua Changa,b, Tzu-Yun Wangc, I Hui Leec,d,e, Sheng-Yu Leef, Kao Chin Chenc,d,
San-Yuan Huangg, Yen Kuang Yangc,d,e,h, Ru-Band Luc,d, Po See Chenc,d*
a
Institute of Clinical Pharmacy and Pharmaceutical Sciences, National Cheng Kung
University, Tainan, Taiwan
b
School of Pharmacy, College of Medicine, National Cheng Kung University, Tainan,
Taiwan
c
Department of Psychiatry, National Cheng Kung University Hospital, College of
Medicine, National Cheng Kung University, Tainan, Taiwan
d
Addiction Research Center, National Cheng Kung University, Tainan, Taiwan
e
Department of Psychiatry, National Cheng Kung University Hospital, Dou-Liou

D
Branch, Yunlin, Taiwan
f
Department of Psychiatry, Kaohsiung Veterans General Hospital, Kaohsiung, Taiwan
g

Center, Taipei, Taiwan

h
TE
Department of Psychiatry, Tri-Service General Hospital, National Defense Medical

Institute of Behavioral Medicine, College of Medicine, National Cheng Kung

EP
University, Tainan, Taiwan
C

Total number of words: 2980

AC

Table: 2
Figure: 3

Corresponding author: Po See Chen

ST

Department of Psychiatry, National Cheng Kung University Hospital,

138 Sheng Li Road, North Dist., Tainan 70403, Taiwan.
Tel.: +886 6 2353535x5189; fax: +886 6 2759259.
JU

E-mail: chenps@mail.ncku.edu.tw
Abstract

Objective. We aimed to examine whether the C-reactive protein (CRP) level could be

used to differentiate between major depressive disorder (MDD) and bipolar II

disorder (BD II).

Methods. Ninety-six healthy controls, 88 BD II and 72 MDD drug-naïve patients in

their major depressive episodes were enrolled. The fasting plasma level of

high-sensitivity CRP was assessed at baseline and after treatment.

Results. The BD II patients presented significantly higher HDRS scores and CRP

D
TE
levels at baseline when adjustment for age, gender, and body mass index (BMI)

(p<0.001 and p<0.001, respectively). After treatment the CRP levels remained
EP
significantly different (p<0.001), thought the HDRS score was not significantly
C

different between the BD II and MDD patients. A receiver-operating characteristic

AC

analysis showed that a baseline CRP level of 621.6 ng/mL could discriminate between

BD II and MDD, with an area under the curve of 0.816 and a sensitivity and
ST

specificity of 0.699 and 0.882, respectively. Furthermore, the baseline CRP level
JU

greater than 621.6 ng/mL had 28.2 higher odds of a diagnosis of BD II (p<0.001, 95%

CI: 10.96–72.35).

Conclusions. The level of CRP plays a role of biomarker to differentiate between

MDD and BD II depression in both their depressed and euthymic state.

Key words: biomarker; bipolar disorder; C-reactive protein; inflammation; major

depressive disorder.

Introduction

Depressive episodes are more numerous and last longer than manic or

hypomanic episodes in bipolar disorder (BD) (Bowden 2010). Misdiagnosis of major

depressive disorder (MDD) delays the initiation of appropriate therapy, leading to a

D
poor outcome (Culpepper 2014). Bipolar II disorder (BD II) depression, a distinct

TE
category within mood disorders, is especially difficult to diagnose accurately
EP
(Daigneault et al. 2015). A systematic search found that 40% of major depressive

episodes could be classified as BD II; however, only half of these were recognized by
C

clinicians (Hantouche et al. 1998). At present, few phenotypes have been confirmed
AC

to differentiate BD II depression from unipolar depression (Angst et al. 2011). Among

ST

them, somatic and cognitive symptoms have been considered to be indicators of

potential endophenotypes in bipolar spectrum disorders (Lovdahl et al. 2014). The

JU

number of prior episodes, feelings of worthlessness, age of onset, sleep patterns and

mixed features could also help to distinguish between MDD and BD II depression

(Frankland et al. 2015). Sub-threshold manic symptoms have been linked to the

progression to BD (Fiedorowicz et al. 2011).

 In practice, a poor response to antidepressants could be a useful predictor of

bipolar spectrum disorders (Li et al. 2012). Therefore, predictors of a poor treatment

response in major depressive episodes might also be biomarkers for BD II depression

(Benazzi 2007). Among the predictors of a poor treatment response, the level of

C-reactive protein (CRP) has been found to be a differential predictor of the outcome

of MDD (Vogelzangs et al. 2014). CRP is a sensitive marker of inflammation,

synthesized by the liver, in response to interleukin-6 released by macrophages, T cells,

D
and adipocytes. The elevated plasma levels of CRP have significant correlation with

TE
increased risk of cardiovascular diseases and metabolic disturbances (Pearson et al.
EP
2003; Nordestgaard and Zacho 2009; Wei et al. 2014). Moreover, the odds ratios for

antidepressant use, hospitalization and depression recurrence are significantly

C

increased with increasing CRP levels (Halder et al. 2010). In MDD patients, the poor
AC

improvement of cognitive function after antidepressant treatment was also found to be

ST

associated with the CRP levels (Chang et al. 2012). Importantly, the CRP level is

significantly associated with sub-threshold manic symptoms in MDD patients

JU

(Dickerson et al. 2007; Becking et al. 2013).

Elevated CRP level was an important risk factor for the onset of manic

symptoms in depressed patient during 2 years of follow-up (Becking et al. 2013). In

BD patients, peripheral immune markers have been found to be associated with mood
status, metabolic status and cognitive function (Dickerson et al. 2013; Bai et al. 2014;

Tsai et al. 2014). In addition, the levels of CRP in individuals with BD have been

found to be significantly associated with their disease severity (Dickerson et al. 2007).

One pilot study revealed that patients with BD I have significantly higher CRP levels

than healthy controls, but this was not the case in MDD patients (Huang and Lin

2007). A meta-analysis study has reported that BD was associated with inflammation

and that CRP could be a marker of mood state (Dargél et al. 2015; Wysokinski et al.

D
2015).

TE
The differentiation of BD II depression from MDD is of clinical importance.
EP
Furthermore, the identification of biomarkers that represent the pathophysiologic

status of BD II in different mood phases can enable the staging of BD II and provide
C

biological targets for treatment (Dargel et al. 2015). Therefore, in this study, we aimed
AC

to examine whether the CRP level could be a biomarker to differentiate MDD and BD
ST

II in both their depressed and remitted states.

JU
Material and methods

Participants

The Institutional Review Board for the Protection of Human Subjects at National

Cheng Kung University Hospital approved the research protocols (HR-95-06,

HR-98-025, and BR-100-035). All participants signed written informed consent forms

after the procedures have been fully explained. Nninety six healthy controls were

recruited from the community after exclusion of individuals with mental illnesses by a

D
senior psychiatrist using the Chinese version of the Mini International

Neuropsychiatry Interview.
TE
EP
Patients were recruited from outpatient settings at the National Cheng Kung

University Hospital. All BD-II patients recruited were drug-naïve and first diagnosed
C

without a history of taking mood stabilizers. We used the 2-day minimum for
AC

hypomania in the diagnosis of BD II (Lee et al. 2013). All BD II patients were

ST

initially evaluated in an interview by an attending psychiatrist and followed-up with a

more detailed interview by a clinical psychologist using the Chinese Version of the
JU

Modified Schedule of Affective Disorder and Schizophrenia–Life Time, which has a

good inter-rater reliability, to determine the Diagnostic and Statistical Manual of

Mental Disorders, Fourth Edition diagnoses (DSM-IV). At time of study entry, all of

the BD-II patients were at major depressive status, which the 17-item Hamilton
Depression Rating Scale (HDRS) scores > 15. After screening, the BD-II patients

received valproate (Valproate Chrono [VPA], 500 to 1000 mg daily with the mean

plasma concentration of VPA at 79.3±15.8 µg/m) for 12 weeks. Twenty mg

fluoxetine per day was permitted their depressive symptoms. In terms of

antidepressant choice, only one antidepressant was permitted for their depressive

symptoms for BD II patients. Fluoxetine was prescribed because previous studies

indicated that it appeared to be safe and effective in BD II patients with major

D
depressive episodes (Amsterdam and Shults 2010), while venlafaxine is not.

TE
Concomitant benzodiazepine medication (lorazepam <8 mg) was used for night-time
EP
sedation and to treat agitation and insomnia during the study.

The drug-naïve MDD patients enrolled were all meeting diagnostic criteria of
C

MDD via the DSM-IV criteria at time of study entry and the 17-item Hamilton
AC

Depression Rating Scale (HDRS) scores > 15, as described before (Chang et al. 2012).
ST

In the current study, both MDD and BD II patients were diagnosed for the first time.

They were randomly assigned to either the fluoxetine or the venlafaxine treatment
JU

group and treated for 6 weeks. The initial dose of fluoxetine was 20 mg once daily,

which could be increased by 20 mg in divided doses to a maximal daily dose of 80 mg.

The initial dose of venlafaxine was 37.5 mg once daily for 4 days titrated to 75 mg

once daily, which could be increased by 75 mg in divided doses to a maximal daily

dose of 225 mg. Lorazepam was the only concomitant drug in MDD patients to a

maximal daily dose of 6 mg.

To avoid confounding factor interfering the baseline CRP level, all of the

subjects meeting the following criteria were excluded through chart reviews and

patient-reported questionnaires: (i) a serious surgical condition or physical illness

(including all chronic diseases, heart disease, stroke, kidney dialysis, transplant, etc.);

(ii) patients who were pregnant or breastfeeding; (iii) a DSM-IV diagnosis of

D
substance abuse within the past three months; (iv) previous use of any psychotropic

TE
agent; (v) taking any anti-inflammatory medication before, during the trial period,
EP
having autoimmune disorder, or at infectious status (CRP > 7mg/L); and (vi) an

organic mental disease, mental retardation or dementia.

C
AC

C-reactive protein measurement

ST

Fasting blood samples were collected between 0800 and 1000 hours. Ten

milliliters of whole blood were withdrawn from the antecubital vein of each patient.
JU

Plasma, which was isolated from whole blood after centrifugation at 3000 ×g for 15

minutes at 4°C, was immediately stored at –80°C. The fasting plasma CRP level was

assessed using a high-sensitivity CRP ELISA kit (Bender MedSystems, USA), for

which the limit of detection was 3 pg/mL and the intra- and inter-assay coefficients of
variation were 6.9% and 13.1%, respectively. An independent assistant who was blind

to the allocation of the patients examined the patients’ assessments and CRP levels.

Also, all of the patients were measured their body mass index (BMI) at this time

point. BMI was calculated as weight (kg) divided by height squared (m2), and waist

circumference was measured at the level midway between the lateral lower rib margin

and the superior anterior iliac crest.

D
Statistical analysis

TE
We analyzed the data using the Statistical Package for Social Sciences 16.0 for
EP
Windows (SPSS Inc, Chicago, IL). Categorical variables were expressed as numbers

and percentages, and continuous variables as means ± SD unless otherwise specified.

C

The one-way ANOVA, t test, and chi-square analyses were used to compare the
AC

demographic and clinical characteristics between groups. In addition,

ST

receiver-operating characteristic (ROC) curves were used for discrimination of BD II

or MDD based on the baseline CRP level. The gold standard is referred to the
JU

diagnosis of BD II or MDD. Since age and gender were significantly different

between these two groups at baseline, covariate-adjusted ROC curve, by age and

gender, was determined by R 2.13.1 software in pcvsuite package (Pepe et al. 2009).

Furthermore, the values of odds ratio were determined for all the cutoff points
corresponding to the baseline CRP levels by R 2.13.1 software in ROCR package

(Sing et al. 2005). Logistic regression has been used to estimate the odd ratio of the

cut-off value of the baseline CRP level adjusted by age and gender. The level of

significance was set at 0.05.

Results

Ninety-six healthy controls, 88 BD II and 72 MDD drug-naïve patients in their

D
major depressive episodes were enrolled. The demographic characteristics of the

TE
subjects are shown in Table I. The age and gender of the patients were significantly
EP
different among the BD II, MDD, and healthy controls groups (p<0.001 and p=0.002,

respectively). The mean educational levels (13.5 and 12.5 years) and the past suicide
C

attempts (35.0% and 32.8%) were also not significantly different between BD II and
AC

MDD patients. At baseline, the BD II patients presented significantly higher HDRS

ST

scores and CRP levels than those of the MDD patients and the healthy controls

(19.41±5.46, 15.48±7.37, and 3.46±1.48, p<0.001; 1981.91±1771.93, 323.10±431.72,

JU

and 177.40±215.02, p<0.001, respectively). The results were consistent when adjusted

by age, gender, and the baseline BMI (Table I). Moreover, the CRP level was

significantly correlated with the HDRS score (r= 0.347, p<0.001).

 After treatment, the HDRS score was not significantly different between the BD

II and MDD patients. However, the CRP level remained significantly different

between the two groups (p<0.001), even following adjustment by age, gender, the

baseline BMI, and HDRS scores (Table I and Figure 1). Moreover, the CRP level

increased significantly in the MDD patients (p<0.001) but remained unchanged in the

BD II patients (p=0.238) during the treatment period. Both fluoxetine (N=39,

388.52±480.92 to 582.00±682.89 ng/mL, p=0.038) and venlafaxine (N=33,

D
245.77±357.06 to 583.31±840.38 ng/mL, p=0.005) treated MDD patients showed

TE
increased CRP levels. In addition, after treatment only in BD II patients had a
EP
significantly increased BMI (22.83±4.42 to 13.42±3.96, p< 0.001), while the MDD

patients did not.

C

To test if the CRP levels could be a differential biomarker for MDD and BD II
AC

disorder in major depressive episode, we further estimated the cut-off value of the
ST

baseline CRP level that could discriminate between BD II and MDD. ROC analysis

was performed in the two groups, and the results showed that a baseline CRP level of
JU

621.6 ng/mL could discriminate between BD II and MDD, with an area under the

curve (AUC) of 0.816 (95% confidence interval: 0.760–0.872) (Figure 2a). The

sensitivity and specificity were 0.699 and 0.882, respectively. The covariate-adjusted

ROC curve by age, gender, and the baseline BMI showed a consistent result with an
AUC of 0.856 (95% confidence interval: 0.793–0.903) (Figure 2b). Furthermore,

logistic regression was performed and the results showed that baseline CRP level

greater than 621.6 ng/mL had 28.2 higher odds of a diagnosis of BD II adjusted by

age, gender, and the baseline BMI (p<0.001) (Table II). The values of odds ratio for

each cutoff point corresponding to the baseline CRP level have also been calculated

(Figure 3).

D
Discussion

TE
The BD II depression population was our key focus in the current study as
EP
identifying biomarkers before the onset of mood switching and comorbid medical

conditions is especially important. Previously, different versions of hypomanic

C

checklists have been developed for the detection of hypomania in patients with mood
AC

disorders (Chou et al. 2012). However, their screening properties require

ST

improvement for clinical diagnosis. The results of this study suggested that the level

of CRP could be used as a biomarker to differentiate between MDD and BD II

JU

depression in both their depressed and euthymic state. Moreover, the higher level of

CRP could be considered as evidence that link peripheral immune dysfunction to BD

susceptibility in mood disorder patients (Anand et al. 2009; Hodes et al. 2014; Dargel

et al. 2015; Harrison et al. 2015). Because patients with elevated inflammation are
more likely to be obese and meet the criteria for metabolic syndrome, the difference

in the CRP level might also be linked to the differing prevalence of metabolic

syndrome among MDD, BD II and BD I patients (Mansur et al. 2015). Our previous

studies have provided evidence to support this point (Hung Chi et al. 2013).

Evidence that supports the use of the CRP level as a stable biomarker to

differentiate bipolar depression from MDD also supports the concept of categorical

diagnosis of BD (Blacker and Tsuang 1992). A previous study in Taiwan including

D
BD I patients currently experiencing a depressive episode also identified persistent

TE
inflammation with elevated levels of CRP from the acute phases to full remission
EP
(Tsai et al. 2014). Moreover, the CRP level reported in BD I patients was significantly

higher than we have reported here for BD II patients.

C

However, although the current study was a prospective one, the observation
AC

duration was relatively short. Whether a polarity switch would occur in certain MDD
ST

patients with a change in the CRP level remains to be investigated (Dudek et al. 2013).

If the polarity switch from depression to hypomania or mania accompanies an

JU

increase in the CRP level, then the level of CRP might be considered a state marker of

bipolarity of depression. If this is the case, the changes of CRP levels in depressed

patients could be considered in a dimensional approach to better define BD II

depression. Interestingly, we have found a significant increase in the CRP level in the
MDD patients after receiving antidepressant treatment, either serotonin-specific

reuptake inhibitor (SSRI) fluoxetine or the serotonin–norepinephrine reuptake

inhibitor (SNRI) venlafaxine. Population-based studies also confirmed that the use of

antidepressants, including SSRI and tricyclic antidepressant, is associated with a

subsequently elevated CRP level independent of the symptoms of depression (Hamer

et al. 2011). Moreover, compared with SSRI treatment, tricyclic antidepressant and

SNRI treatment have been found to be related to a risk of induction of mood switch

D
(Leverich et al. 2006; Amsterdam and Shults 2010). In contrast to our results, one

TE
study with limited patients showed a significant drop in the CRP level after SSRI
EP
treatment in MDD patients (O'Brien et al. 2006). Whether the changes in CRP level in

MDD patients linked to the increased possibility of mood switch is of clinical

C

importance. Because now patients’ histories as well as their symptomatic and genetic
AC

factors are the only predictors. Long term follow-up study is required to answer this
ST

question.

After VPA/fluoxetine treatment, the HDRS scores reduced significantly in the

JU

BD II depression patients. Our results also showed a trend of a decrease in the level of

CRP in the BD II depression patients after treatment. This indicated that the

therapeutic effect of VPA might be correlated partially with its anti-inflammation

action (Ximenes et al. 2013). In addition, the anti-inflammatory effect is achieved by

histone deacetylase inhibition in glia and immune cells, which also play a key role in

BD (Watkins et al. 2014). Although whether the anti-inflammatory effect of VPA is

associated with its therapeutic effect in bipolar depression remains unclear, however,

it has been shown that lithium has anti-inflammatory effects that may contribute to its

therapeutic efficacy (Beurel and Jope 2014; Fond et al. 2014). It has been proposed

that the pathophysiology of BD may not be limited to neurotransmitter abnormalities,

but may reflect multifunctional system dysfunction that includes the immune system

D
(Leboyer et al. 2012; Rosenblat and McIntyre 2015). Adjunct medications with

TE
anti-inflammatory activity might help to improve the treatment outcome and also
EP
reduce co-morbidity in BD II depression (Vogelzangs et al. 2012).

Nevertheless, our findings were interpreted in some limitations. First, there were
C

small sample size, and the short duration of treatment and measurement of the change
AC

in CRP level after treatment. Whether the CRP level could be a differential diagnostic
ST

biomarker remains further larger sample sizes. Second, non-medication factors that

may have confounded the results of the study, such as diet, alcohol, exercise, and
JU

comorbidities were not accounted. In addition, although smoking could be a

confounding factor that strongly influences CRP levels, we did not collect it in the

beginning of the study. Although these factors could influence CRP level, we have

controlled age, gender and BMI to explain the association between mood disorders
and CRP level. Third, in BD II and MDD patients, the treatment duration and the time

point to measure CRP level after treatment were not the same. Fourth, though we

recruited drug-naïve participants, but the CRP levels might be influenced by on-going

or previously administered medications in certain patients. Further prospective studies

with a long follow-up period could provide a chance to test the result and confirm the

role of the CRP levels as a differential biomarker.

In summary, our study suggested that the level of CRP could be a potential

D
biomarker to differentiate between MDD and BD II depression in both their depressed

TE
and euthymic state. The level of CRP might be integrated with genetic, molecular,
EP
neural circuitry, and behavioral level biomarkers in BD by using a dimensional

approach to represent the underlying pathophysiological processes in BD (Almeida et

C

al. 2013; Lee et al. 2013; Hirschfeld 2014). Whether the increase in the CRP level is
AC

associated with mood switch after long-term antidepressant use in certain patients
ST

merits further investigation. Adjunct treatments focusing on the inflammatory process

might help to improve the treatment outcome and also reduce co-morbidity in both
JU

MDD and BD II patients.

Acknowledgments

This study was financially supported by the National Science Council of Taiwan

(NSC 93-2314-B-006-107, NSC 97-2314-B-006-006-MY3, NSC 98-2627-B-006-016,

NSC 99-2627-B-006-014, NSC 100-2314-B-006-041-MY3, NSC

100-2627-B-006-012, NSC 101-2314-B-006-064-MY3, and NSC

101-2314-B-006-065), Department of Health, Taiwan (DOH96-TD-D-113-041), and

National Cheng Kung University Hospital (NCKUH-10301003). This research also

D
received funding (D102-35001 and D103-35A09) from the Headquarters of

TE
University Advancement at the National Cheng Kung University, which is sponsored
EP
by the Ministry of Education, Taiwan, ROC. The authors wish to thank Mr. Chien

Ting Lin for their administrative support.

C
AC

Statement of Interest
ST

The authors declare that they have no conflicts of interest in relation to this work.

The funders had no role in study design, data collection and analysis, decision to
JU

publish, or preparation of the manuscript.

References

Almeida, JR, Mourao-Miranda, J, Aizenstein, HJ, Versace, A, Kozel, FA, Lu, H, et al.

2013. Pattern recognition analysis of anterior cingulate cortex blood flow to

classify depression polarity. Br J Psychiatry 203:310-311.

Amsterdam, JD, Shults, J. 2010. Efficacy and safety of long-term fluoxetine versus

lithium monotherapy of bipolar II disorder: a randomized, double-blind,

placebo-substitution study. Am J Psychiatry 167:792-800.

D
Anand, A, Li, Y, Wang, Y, Lowe, MJ, Dzemidzic, M. 2009. Resting state

TE
corticolimbic connectivity abnormalities in unmedicated bipolar disorder and
EP
unipolar depression. Psychiatry Res 171:189-198.

Angst, J, Azorin, JM, Bowden, CL, Perugi, G, Vieta, E, Gamma, A, et al. 2011.
C

Prevalence and characteristics of undiagnosed bipolar disorders in patients with a

AC

major depressive episode: the BRIDGE study. Arch Gen Psychiatry 68:791-798.
ST

Bai, YM, Su, TP, Tsai, SJ, Wen-Fei, C, Li, CT, Pei-Chi, T, et al. 2014. Comparison of

inflammatory cytokine levels among type I/type II and

JU

manic/hypomanic/euthymic/depressive states of bipolar disorder. J Affect Disord

166:187-192.

Becking, K, Boschloo, L, Vogelzangs, N, Haarman, BC, Riemersma-van der Lek, R,

Penninx, BW, et al. 2013. The association between immune activation and manic
 symptoms in patients with a depressive disorder. Transl Psychiatry 3:e314.

Benazzi, F. 2007. Bipolar disorder--focus on bipolar II disorder and mixed depression.

Lancet 369:935-945.

Beurel, E, Jope, RS. 2014. Inflammation and lithium: clues to mechanisms

contributing to suicide-linked traits. Transl Psychiatry 4:e488.

Blacker, D, Tsuang, MT. 1992. Contested boundaries of bipolar disorder and the limits

of categorical diagnosis in psychiatry. Am J Psychiatry 149:1473-1483.

D
Bowden, CL. 2010. Diagnosis, treatment, and recovery maintenance in bipolar

depression. J Clin Psychiatry 71:e01.

TE
EP
Chang, HH, Lee, IH, Gean, PW, Lee, S-Y, Chi, MH, Yang, YK, et al. 2012. Treatment

response and cognitive impairment in major depression: Association with

C

C-reactive protein. Brain Behav Immun 26:90-95.

AC

Chou, CC, Lee, IH, Yeh, TL, Chen, KC, Chen, PS, Chen, WT, et al. 2012.
ST

Comparison of the validity of the Chinese versions of the Hypomania Symptom

Checklist-32 (HCL-32) and Mood Disorder Questionnaire (MDQ) for the

JU

detection of bipolar disorder in medicated patients with major depressive disorder.

Int J Psychiatry Clin Pract 16:132-137.

Culpepper, L. 2014. Misdiagnosis of bipolar depression in primary care practices. J

Clin Psychiatry 75:e05.

Daigneault, A, Duclos, C, Saury, S, Paquet, J, Dumont, D, Beaulieu, S. 2015.

Diagnosis of bipolar disorder in primary and secondary care: What have we

learned over a 10-year period? J Affect Disord 174:225-232.

Dargél, AA, Godin, O, Kapczinski, F, Kupfer, DJ, Leboyer, M. 2015. C-Reactive

Protein Alterations in Bipolar Disorder: A Meta-Analysis. J Clin Psychiatry

76:142-150.

Dargel, AA, Godin, O, Kapczinski, F, Kupfer, DJ, Leboyer, M. 2015. C-reactive

D
protein alterations in bipolar disorder: a meta-analysis. J Clin Psychiatry

76:142-150.
TE
EP
Dickerson, F, Stallings, C, Origoni, A, Boronow, J, Yolken, R. 2007. Elevated serum

levels of C-reactive protein are associated with mania symptoms in outpatients

C

with bipolar disorder. Prog Neuropsychopharmacol Biol Psychiatry 31:952-955.

AC

Dickerson, F, Stallings, C, Origoni, A, Vaughan, C, Khushalani, S, Yolken, R. 2013.

ST

Elevated C-reactive protein and cognitive deficits in individuals with bipolar

disorder. J Affect Disord 150:456-459.

JU

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.

Fiedorowicz, JG, Endicott, J, Leon, AC, Solomon, DA, Keller, MB, Coryell, WH.
 2011. Subthreshold hypomanic symptoms in progression from unipolar major

depression to bipolar disorder. Am J Psychiatry 168:40-48.

Fond, G, Hamdani, N, Kapczinski, F, Boukouaci, W, Drancourt, N, Dargel, A, et al.

2014. Effectiveness and tolerance of anti-inflammatory drugs' add-on therapy in

major mental disorders: a systematic qualitative review. Acta Psychiatr Scand

129:163-179.

Frankland, A, Cerrillo, E, Hadzi-Pavlovic, D, Roberts, G, Wright, A, Loo, CK, et al.

D
2015. Comparing the phenomenology of depressive episodes in bipolar I and II

TE
disorder and major depressive disorder within bipolar disorder pedigrees. J Clin
EP
Psychiatry 76:32-39.

Halder, I, Marsland, AL, Cheong, J, Muldoon, MF, Ferrell, RE, Manuck, SB. 2010.
C

Polymorphisms in the CRP gene moderate an association between depressive

AC

symptoms and circulating levels of C-reactive protein. Brain Behav Immun

ST

24:160-167.

Hamer, M, Batty, GD, Marmot, MG, Singh-Manoux, A, Kivimaki, M. 2011.

JU

Anti-depressant medication use and C-reactive protein: results from two

population-based studies. Brain Behav Immun 25:168-173.

Hantouche, EG, Akiskal, HS, Lancrenon, S, Allilaire, JF, Sechter, D, Azorin, JM, et al.

1998. Systematic clinical methodology for validating bipolar-II disorder: data in

 mid-stream from a French national multi-site study (EPIDEP). J Affect Disord

50:163-173.

Harrison, NA, Cercignani, M, Voon, V, Critchley, HD. 2015. Effects of inflammation

on hippocampus and substantia nigra responses to novelty in healthy human

participants. Neuropsychopharmacology 40:831-838.

Hirschfeld, RM. 2014. Differential diagnosis of bipolar disorder and major depressive

disorder. J Affect Disord 169 Suppl 1:S12-16.

D
Hodes, GE, Pfau, ML, Leboeuf, M, Golden, SA, Christoffel, DJ, Bregman, D, et al.

TE
2014. Individual differences in the peripheral immune system promote resilience
EP
versus susceptibility to social stress. Proc Natl Acad Sci U S A 111:16136-16141.

Huang, TL, Lin, FC. 2007. High-sensitivity C-reactive protein levels in patients with
C

major depressive disorder and bipolar mania. Prog Neuropsychopharmacol Biol

AC

Psychiatry 31:370-372.
ST

Hung Chi, M, Hua Chang, H, Tzeng, NS, Huang, SY, Chou, KR, Chun Tsai, H, et al.

2013. The prevalence of metabolic syndrome in drug-naive bipolar II disorder

JU

patients before and after twelve week pharmacological intervention. J Affect

Disord 146:79-83.

Leboyer, M, Soreca, I, Scott, J, Frye, M, Henry, C, Tamouza, R, et al. 2012. Can

bipolar disorder be viewed as a multi-system inflammatory disease? J Affect

 Disord 141:1-10.

Lee, S-Y, Chen, S-L, Chang, Y-H, Chen, PS, Huang, S-Y, Tzeng, N-S, et al. 2013.

Inflammation’s Association with Metabolic Profiles before and after a

Twelve-Week Clinical Trial in Drug-Naïve Patients with Bipolar II Disorder.

PLoS ONE 8:e66847.

Leverich, GS, Altshuler, LL, Frye, MA, Suppes, T, McElroy, SL, Keck, PE, Jr., et al.

2006. Risk of switch in mood polarity to hypomania or mania in patients with

D
bipolar depression during acute and continuation trials of venlafaxine, sertraline,

TE
and bupropion as adjuncts to mood stabilizers. Am J Psychiatry 163:232-239.
EP
Li, CT, Bai, YM, Huang, YL, Chen, YS, Chen, TJ, Cheng, JY, et al. 2012. Association

between antidepressant resistance in unipolar depression and subsequent bipolar

C

disorder: cohort study. Br J Psychiatry 200:45-51.

AC

Lovdahl, H, Boen, E, Malt, EA, Malt, UF. 2014. Somatic and cognitive symptoms as
ST

indicators of potential endophenotypes in bipolar spectrum disorders: An

exploratory and proof-of-concept study comparing bipolar II disorder with

JU

recurrent brief depression and healthy controls. J Affect Disord 166:59-70.

Mansur, RB, Brietzke, E, McIntyre, RS. 2015. Is there a "metabolic-mood syndrome"?

A review of the relationship between obesity and mood disorders. Neurosci

Biobehav Rev 52:89-104.

Nordestgaard, BG, Zacho, J. 2009. Lipids, atherosclerosis and CVD risk: is CRP an

innocent bystander? Nutr Metab Cardiovasc Dis 19:521-524.

O'Brien, SM, Scott, LV, Dinan, TG. 2006. Antidepressant therapy and C-reactive

protein levels. Br J Psychiatry 188:449-452.

Pearson, TA, Mensah, GA, Alexander, RW, Anderson, JL, Cannon, RO, 3rd, Criqui,

M, et al. 2003. Markers of inflammation and cardiovascular disease: application

to clinical and public health practice: A statement for healthcare professionals

D
from the Centers for Disease Control and Prevention and the American Heart

Association. Circulation 107:499-511.

TE
EP
Pepe, M, Longton, G, Janes, H. 2009. Estimation and Comparison of Receiver

Operating Characteristic Curves. Stata J 9:1-1.

C

Rosenblat, JD, McIntyre, RS. 2015. Are medical comorbid conditions of bipolar
AC

disorder due to immune dysfunction? Acta Psychiatr Scand:doi:

ST

10.1111/acps.12414.

Sing, T, Sander, O, Beerenwinkel, N, Lengauer, T. 2005. ROCR: visualizing classifier

JU

performance in R. Bioinformatics 21:3940-3941.

Tsai, SY, Chung, KH, Huang, SH, Chen, PH, Lee, HC, Kuo, CJ. 2014. Persistent

inflammation and its relationship to leptin and insulin in phases of bipolar

disorder from acute depression to full remission. Bipolar Disord 16:800-808.

Vogelzangs, N, Beekman, AT, van Reedt Dortland, AK, Schoevers, RA, Giltay, EJ, de

Jonge, P, et al. 2014. Inflammatory and metabolic dysregulation and the 2-year

course of depressive disorders in antidepressant users.

Neuropsychopharmacology 39:1624-1634.

Vogelzangs, N, Duivis, HE, Beekman, AT, Kluft, C, Neuteboom, J, Hoogendijk, W, et

al. 2012. Association of depressive disorders, depression characteristics and

antidepressant medication with inflammation. Transl Psychiatry 2:e79.

D
Watkins, CC, Sawa, A, Pomper, MG. 2014. Glia and immune cell signaling in bipolar

TE
disorder: insights from neuropharmacology and molecular imaging to clinical
EP
application. Transl Psychiatry 4:e350.

Wei, W, Yang, S, Qiu, Y, Wang, H, Zhao, X, Zhao, Y, et al. 2014. CRP gene
C

polymorphism contributes genetic susceptibility to dyslipidemia in Han Chinese

AC

population. Mol Biol Rep 41:2335-2343.

ST

Wysokinski, A, Margulska, A, Strzelecki, D, Kloszewska, I. 2015. Levels of

C-reactive protein (CRP) in patients with schizophrenia, unipolar depression and

JU

bipolar disorder. Nord J Psychiatry 69:346-353.

Ximenes, JC, de Oliveira Goncalves, D, Siqueira, RM, Neves, KR, Santos Cerqueira,

G, Correia, AO, et al. 2013. Valproic acid: an anticonvulsant drug with potent

antinociceptive and anti-inflammatory properties. Naunyn Schmiedebergs Arch

Pharmacol 386:575-587.

D
TE
EP
C
AC
ST
JU
Table I. The demographic characteristics of the BD II, MDD patients, and healthy

controls.

BD II MDD Controls p Adjus Adjus Adjus Adjus

(N=88) (N=72) (N=96) value ted p ted p ted p ted p
value value value value
1 2 3 4

Baseli
ne
Ag 31.74±11.7 40.72±12. 33.30±12. <0.0 - - - -
e (yr) 8 45 11 01*
Ge 44 (50.0%) 51 55 (57.3) 0.00 - - - -
nder, (70.8%) 2*

D
femal
e (%)

MI
B

(kg/m
22.83±4.42 22.25±4.1 22.78±3.6 0.51
1 8 4 TE
0.287 - - -
EP
2
)
H 19.41±5.46 15.48±7.3 3.46±1.48 <0.0 <0.00 <0.00 <0.00 -
DRS 7 01* 1* 1* 1*
C

scores
AC

CR 1981.91±1 323.10±4 177.40±2 <0.0 <0.00 <0.00 <0.00 <0.00

P 771.93 31.72 15.02 01* 1* 1* 1* 1*
(ng/m
L)
ST

After
treat
ment
JU

H 8.9±5.17 9.82±7.57 - 0.35 0.42 0.604 0.632 -

DRS
scores
CR 1784.10±1 582.60±7 - <0.0 <0.00 <0.00 <0.00 0.001
P 841.55 53.59 01* 1* 1* 1* *
(ng/m
L)
BD II, bipolar II; MDD, major depressive disorder; H, healthy controls; BMI, body
mass index; HDRS, 17-item Hamilton Depression Rating Scale; CRP, C-reactive
protein.
* p<0.05; 1 adjusted for age and gender; 2 adjusted for baseline BMI; 3 adjusted for
age, gender, and baseline BMI; 4 adjusted for age, gender, baseline BMI, and HDRS
scores

D
TE
EP
C
AC
ST
JU
Table II. Logistic regression for discriminating bipolar II depression and major
depressive disorder.
B Odds ratio p values 95% CI
Model 1
Constant -6.29 0.002 <0.001* –
Age 0.69 1.09 <0.001* 1.04–1.11
Gender 0.74 2.09 0.064 0.96–4.59
BMI 0.55 1.06 0.287 0.95-1.17
CRP (ng/mL) 3.34 28.16 <0.001* 10.96–72.35
(≥621.6 vs. <621.6)
Model 2
Constant -4.67 0.01 0.004 -
Age 0.08 1.08 <0.001* 1.04-1.12
Gender 0.55 1.73 0.192 0.76-3.99

D
BMI 0.03 1.03 0.561 0.93-1.15
HDRS scores -0.07 0.93 0.022 0.88-0.99
CRP (ng/mL)
(≥621.6 vs. <621.6)
2.91 18.42
TE
<0.001* 6.92-49.01
EP
BMI, body mass index; HDRS, 17-item Hamilton Depression Rating Scale; CRP,
C-reactive protein.
* p<0.05
C
AC
ST
JU
Figure 1. The CRP levels in BD II and MDD patients.

D
TE
The C-reactive protein (CRP) levels were significantly different across the treatment
EP
period between the bipolar II (BD II) and major depressive disorder (MDD) patients

(p< 0.001).
C
AC
ST
JU
Figure 2. ROC curve analysis.

(a)

D
TE
EP
(b)
C
AC
ST
JU
(a) Receiver-operating characteristic (ROC) analysis showed that a baseline

C-reactive protein (CRP) level of 621.6 ng/mL could discriminate between bipolar

II (BD II) and major depressive disorder (MDD) with an area under the curve

(AUC) of 0.816 (95% CI: 0.760–0.872). The sensitivity and specificity were 0.699

and 0.882, respectively.

(b) ROC analysis was performed after adjusted by age, gender, and the baseline body

D
mass index (BMI) with an AUC of 0.856 (95% CI: 0.793–0.903).

TE
EP
C
AC
ST
JU
Figure 3. The odds ratio for all the cutoff points corresponding to the baseline CRP

level.

D
TE
To discriminate bipolar II from major depressive disorder, the values of odds ratio

were performed for each of the cutoff point corresponding to the baseline C-reactive
EP
protein (CRP) level.
C
AC
ST
JU