SEARCH

SEARCH BY CITATION

Keywords:

  • bipolar disorder;
  • brain-derived neurotrophic factor;
  • biomarkers;
  • treatment response;
  • quetiapine

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Objective

Brain-derived neurotrophic factor (BDNF) is consistently associated with acute mood episodes in bipolar disorder, but there is a lack of longitudinal data to support this hypothesis. In this 16-week open-label clinical trial, we tested the predictive role of BDNF Val66Met polymorphism on serum BDNF levels and the relationship of serum BDNF and clinical response in people with bipolar disorder during an acute illness episode.

Method

Sixty-four people with bipolar disorder who were medication-free at baseline and in an acute mood episode were recruited. They were matched with 64 healthy controls. Clinical evaluation, serum BDNF, and BDNF Val66Met polymorphism were determined at baseline, and change in serum BDNF was assessed in patients at weeks 2, 4, 8 and 16.

Results

There were no differences between patients and controls in serum BDNF or in frequencies of the BDNF Val66Met polymorphism genotype at baseline. The multivariable model showed that Met carriers had a significantly different change in BDNF levels compared with Val homozygotes. Not achieving a complete remission was also associated with lower prospectively assessed BDNF levels.

Conclusion

This study provides the first longitudinal evidence that both the BDNF Val66Met polymorphism and remission status predict change in circulating BDNF levels.

Significant Outcomes
  • People with bipolar disorder had a differential change in peripheral levels of brain-derived neurotrophic factor along the treatment of mood episodes depending on their Val66Met genotype.
  • Met carriers of Val66Met genotype had a significantly different trajectory and a tendency for a less substantial increase in peripheral levels than Val homozygotes.
  • Change in brain-derived neurotrophic factor levels was associated with treatment response.
Limitations
  • There were no differences in brain-derived neurotrophic factor levels between patients and controls at baseline; this is possibly due to a floor effect.
  • Owing to the naturalistic design of the study, we could not control for the effect of each particular treatment.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Synaptic plasticity and resilience are increasingly recognized as a dimension of vital importance in mood disorders [1, 2]. Neuronal survival and apoptosis are regulated by key molecules, among them the class of neurotrophins [3]. A neurotrophic model of mood disorders has been actively investigated in the last decade [4]. Brain-derived neurotrophic factor (BDNF) has enjoyed particular prominence, mediating neural processes involved in synaptic efficacy and neuroplasticity [5]. Consistent study data indicate that BDNF is associated with mood disorders and with the mechanism of action of antidepressants, mood stabilizers, and antipsychotics [6, 7].

Secretion of the BDNF protein can be constitutive, but it is mainly regulated by stimuli [8]. Certain functional polymorphisms on BDNF gene and epigenetic regulation may determine BDNF secretion [9, 10]. Specifically, the valine (Val) to methionine (Met) substitution in the 5' proregion of the human BDNF gene has been shown to influence the activity-dependent BDNF secretion and signaling [11]. This same polymorphism has been associated with the diagnosis of bipolar disorder [12], rapid cycling [13], and early age of onset [14].

In bipolar disorder, available data largely indicate an association between circulating BDNF levels and acute mood episodes. There is consistent cross-sectional evidence, including meta-analytic data, that episodes of mania and depression are associated with low peripheral BDNF levels [15-17]. Considering that serum BDNF levels are decreased during manic and depressive episodes and that the normalization of BDNF levels may be associated with clinical stabilization, we have argued that BDNF is a potentially relevant biomarker of illness activity in BD [18]. Nevertheless, these hypotheses have been largely based on case–control studies. To date, only three longitudinal studies to our knowledge have investigated longitudinal BDNF levels in BD. In two studies [19, 20], BDNF levels of patients with BD were lower than those of healthy controls, and the differences in baseline BDNF levels vanished after successful treatment. In the other study [21], the trajectory of BDNF levels during treatment seemed to depend on the polarity of the mood episode.

It is clear from the extant information that longitudinal studies are needed to understand how BDNF changes during treatment and what factors influence this change. Such knowledge is relevant given the effects of neurotrophins on neuroplasticity and ultimately on cognition and functioning [18].

Aims of the study

In this study, we prospectively investigated the effects of treatment and Val66Met polymorphism on serum brain-derived neurotrophic factor levels in a sample of drug-free patients with BD during an acute mood episode along 16 weeks.

Material and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Participants

This is an open-label, longitudinal trial with individuals diagnosed with BD and a group of healthy control subjects matched for age, gender, and ethnicity. Recruitment of patients was performed in the Bipolar Disorders Program of Hospital de Clínicas de Porto Alegre and the Hospital Universitario de Santa Maria in Brazil. The recruitment of controls was carried out in the hospital catchment area of Hospital de Clínicas de Porto Alegre. The study was conducted from April 2009 to December 2011. Inclusion criteria were as follows: 18 years of age or older, BD diagnosis, and current manic, mixed, or depressive episode according to DSM-IV-TR criteria. The confirmation of the diagnosis was established by board-certified psychiatrists based on the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I). Patients additionally had to be medication-free, except for benzodiazepines, for at least 2 weeks (6 weeks in the case of fluoxetine or depot antipsychotics). The exclusion criteria were as follows: any DSM-IV disorders not defined in the inclusion criteria comprising drug abuse, history of neurodegenerative disorders and mental retardation, as well as current cancer, chronic or acute infection, unstable medical illness, glucocorticoid treatment, pregnancy or lactation, and participation in a drug trial within 4 weeks prior to enrollment. Controls were 18 years of age or older and were never diagnosed with a psychiatric disorder. Absence of psychiatric morbidity was confirmed by the non-patient version of SCID-I (SCID/NP). Subjects on current pharmacological treatment, smokers, or those with a family history of psychiatric disorder in first-degree relatives were ruled out as controls.

All participants provided written informed consent after receiving a complete description of the study and adequately understanding it. Procedures were approved by the Institutional Review Board of Hospital de Clínicas and were conducted in accordance with the Declaration of Helsinki.

A total of 64 patients with BD diagnosis and 64 healthy controls were enrolled. Patients were recruited from two similar, but distinct treatment protocols. Patients in an open-label treatment group (n = 31) were medication-free at intake and were medicated at the discretion of the therapist. Patients recruited from a quetiapine extended-release monotherapy open-label trial (n = 33) were also medication-free at intake and received exclusively 300 mg/day of quetiapine extended release. In the event that patients developed severe adverse effects or were clinically deemed to need other treatment, they were withdrawn from the study (ClinicalTrial.gov identifiers: NCT00879632 and NCT00879307).

Procedure and outcome

A baseline assessment was carried out to obtain sociodemographic, clinical, therapeutic, and functional data using semistructured interview. Symptomatology was assessed with clinical rating scales: the Young Mania Rating Scale (YMRS) for mania, the Hamilton Depression Rating Scale 21-item version (HDRS-21) for depression, and the Clinical Global Impression (CGI) severity scale for overall clinical judgment. Functioning was evaluated with the Global Assessment of Functioning (GAF) scale [22-25]. A register of pharmacological treatment was conducted as well.

Longitudinal assessments were performed at weeks 2, 4, 8, and 16. We classified patients as responders if they showed a ≥50% reduction in YMRS and/or HDRS scores from baseline to the final assessment and as remitters those with YMRS and/or HDRS scores ≤8 at the final assessment [26].

Blood and genetic analysis

Ten milliliters of blood was collected from each subject by venipuncture at baseline and weeks 2, 4, 8, and 16 of the treatment. Five milliliters was introduced into a free-anticoagulant vacuum tube for serum BDNF level analysis and immediately centrifuged at 2000 × g for 10 min. Five milliliters was placed into an ethylenediaminetetraacetic acid (EDTA) vacuum tube for DNA analysis. Serum and total blood were kept frozen at −80°C until further procedures.

Serum BDNF levels were measured with sandwich ELISA, using a commercial kit according to the manufacturer's instructions (Millipore, Temecula, CA, USA). Briefly, 96-well, flat-bottomed microtiter plates were coated for 24 h with the samples diluted 1 : 100. Plates were then washed four times with wash buffer, and monoclonal anti-BDNF rabbit antibody was added (diluted 1 : 1000 with sample diluents) and incubated for 3 h at room temperature. After washing, a second incubation with peroxidase-conjugated anti-rabbit antibody (diluted 1 : 1000) was carried out for 1 h at room temperature. After addition of streptavidin enzyme, substrate, and stop solution, the amount of BDNF was determined (absorbance set at 450 nm). The standard curve demonstrated a direct relationship between optical density and BDNF concentrations. The assay sensitivity for BDNF was 7.8 pg/ml. BDNF values are presented as ng/ml.

Genomic DNA was extracted using standard procedures [27]. The genotyping of the BDNF Val66Met polymorphism was performed using a 5' nuclease TaqMan allelic discrimination assay on the Applied Biosystems 7500 Real-Time PCR Systems (Applied Biosystems, Carlsbad, CA, USA). We grouped Met allele carriers (Val/Met and Met/Met genotypes) together for analyses because the rarity of the Met/Met genotype prevents meaningful analysis.

The investigators who performed the laboratory assays were blinded to the clinical status of the subjects.

Statistical analysis

Twenty-seven patients failed to complete all assessments. All but one had at least one postbaseline BDNF level, with an average of 3.6 observations. BDNF levels were log-transformed. Categorical variables were analyzed using chi-square analysis. The distribution of continuous variables was examined using univariate procedures. We constructed random slope and random intercept mixed effect models to investigate longitudinal changes in BDNF levels and longitudinal clinical variations [28]. Residuals were investigated for normality. Mixed models with BDNF polymorphism as a predictor of clinical change in depressive and manic symptoms were used to test these effects controlling for baseline episode. All P values reported were two-tailed. Statistical significance was defined as < 0.05.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Demographic and clinical data

Sixty-four patients were included in this trial, of which 48% were in the open-label treatment protocol and other 52% were in the quetiapine extended-release monotherapy protocol. Thirty per cent were in a manic episode, 45% in a depressive episode, and 25% in a mixed episode. Table 1 depicts the demographic profile of the participants at baseline. No differences were found between patients and controls according to age, gender, and ethnicity. Two patients were dropped out immediately after baseline and were excluded from posterior analyses. Mean body mass index at endpoint was 27.88 with a standard deviation of 4.76 kg/m2.

Table 1. Baseline characteristics of patients
VariablePatients with BD (= 64)
  1. a

    Values are indicated as n (%).

  2. b

    Values are indicated as mean (SD).

  3. BD, Bipolar disorder; NOS, not otherwise specified; YMRS, Young Mania Rating Scale; HDRS, Hamilton Depression Rating Scale; CGI, Clinical Global Impression scale; GAF, Global Assessment of Functioning scale; ECT, electroconvulsive therapy.

Sociodemographic features
Gender (woman)a4062.5
Age at enrolment (years)b37.211.5
Ethnicity (Caucasian)a5687.5
Marital Statusa
Married/Stable partner3250.0
Single1929.7
Other situation1218.8
Educational level (years of study)b11.84.3
Occupational statusa
Working3250.0
Student710.9
Unemployed2539.1
Body Mass Index (kg/m2)b26.55.0
Clinical features
Bipolar disordera
Type I4671.9
Type II or NOS1828.1
Any psychiatric comorbiditya2539.1
Family historya3859.4
First episode
Age (years)b23.410.5
 Psychotic featuresa1828.1
Substance-induced mood episodea46.3
Duration of disorder (years)b13.510.3
Total number of episodesb10.110.5
Depressive episodes5.57.9
Manic episodes4.65.2
Scale assessmentb
YMRS13.913.0
HDRS16.310.4
CGI-BD general4.51.5
GAF48.818.9
Episode at study entrya
Depressive type2945.3
Manic type1929.7
Mixed type1625.0
Protocol followeda
Treatment as usual3148.4
Quetiapine monotherapy3351.6
Treatmenta
Mood stabilizers
Lithium2031.3
Valproate1015.6
Lamotrigine11.6
Antipsychotics
Quetiapine extended release3351.6
Antidepressants46.3
Benzodiazepines2640.6

Treatment response

There was a significant clinical improvement from baseline to endpoint according to the CGI; that was seen in the whole sample (Z = 6.10, < 0.001) and divided according to the polarity of the index episode: mania (Z = 3.78, < 0.001), depression (Z = 3.36, < 0.001), and mixed (Z = 3.30, = 0.001). Figure 1 shows symptom improvement according to the polarity of the index episode. Seventy-seven per cent of those on treatment and 53% of those on quetiapine extended-release monotherapy were responders. Seventy per cent of those on treatment and 38% of those on quetiapine extended-release monotherapy achieved clinical remission.

image

Figure 1. Clinical evolution of the patients according to the polarity of the index episode. YMRS, Young Mania Rating Scale, HAMD, Hamilton Depression Rating Scale.

Download figure to PowerPoint

The BDNF Val66Met genotype frequencies were in Hardy–Weinberg equilibrium in the patient and control groups (χ2 = 1.57, = 0.2109; χ2 = 0.33, P = 0.565 respectively). No significant differences were found in the frequencies of the BDNF Val66Met genotypes, in the allele distribution, or in the serum BDNF levels between patients and controls (Table 2). Considering the clinical response, serum BDNF levels were significantly higher at baseline in non-responders compared with controls (= 2.11, = 0.044) and in non-responders compared with responders (= 2.76, = 0.008).

Table 2. Baseline peripheral brain-derived neurotrophic factor (BDNF) levels and BDNF Val66Met polymorphism in patients with BD and control subjects
VariablePatients with BD (= 62)aControls (= 62)a Stat b P-value
  1. a

    Two patients and two controls have missing BDNF genotype because of difficulties in the technique.

  2. b

    Stats: statistical analysis: t-test or chi-square, as appropriate.

  3. c

    Values are indicated as n (%).

  4. d

    Values are indicated as mean (SD).

  5. BD, Bipolar disorder; BDNF, brain-derived neurotrophic factor; Val, valine; Met, methionine.

BDNF Val66Met polymorphism
Genotype distributionc    1.0740.584
Val66Val4572.64367.21.0740.584
Val66Met1727.41828.1
Met66Met11.6
Allele distributionc    0.1270.722
Val allele10786.310483.9  
Met allele1713.72016.1
Serum BDNF levelsd41.529.836.421.21.2430.219

Along the follow-up, no general changes were detected in serum BDNF levels (Z = 0.67, = 0.501) (Fig. 2). BDNF levels were consistently lower in patients under the naturalistic protocol compared with the quetiapine extended-release protocol (Z = 4.32, < 0.001) and in the Met carrier patients compared with patients with the BDNF Val66Val genotype (Z = 2.35, = 0.019). These were the baseline predictors found for BDNF levels. We further tested whether treatment response and remission were associated with changes in BDNF levels. Interestingly, in addition to both being associated with serum BDNF levels, remission was more strongly associated, and the use of remission as a predictor yielded a better model, with significant effects for time and time vs. remission interaction (Table 3). We built a final multivariate model that included BDNF Val66Met polymorphism, protocol followed, remission status, time on treatment, time vs. remission status interaction, and polarity of the index episode. This model revealed significant effects for the BDNF Val66Met genotype (Z = 2.35, = 0.019), the protocol followed (Z = 2.85, = 0.004), the remission status (Z = 2.85, = 0.004), the time (Z = 2.08, = 0.038), as well as the remission vs. time interaction (Z = −2.34, = 0.019), but not for the polarity of the index episode (Z = 0.30, = 0.766).

Table 3. Fixed effects and interactions with time of selected predictors of change in serum brain-derived neurotrophic factor (BDNF) levels
 Fixed effectTime interaction
  1. a

    Controlled for treatment protocol.

Treatment responseZ = 2.99, = 0.003Z = −2.07, = 0.038
Full remissionZ = 3.82, < 0.001Z = −2.19, = 0.028
Treatment protocolZ = 3.51, < 0.001Z = −0.17, = 0.862
AntidepressantsaZ = 0.23, = 0.816Z = −0.58, = 0.563
LithiumaZ = 0.30, = 0.761Z = 0.32, = 0.746
BDNF Val66Met polymorphismZ = 2.11, = 0.035Z = 0.28, = 0.776
Bipolar disorder type I Z = −0.97, = 0.330Z = −0.25, = 0.800
SexZ = 0.27, = 0.788Z = −1.25, = 0.210
Polarity of index episodeZ = 0.99, = 0.324Z = −0.11, = 0.914
image

Figure 2. Log-transformed serum brain-derived neurotrophic factor (BDNF) levels along the 16-week follow-up according to clinical remission and BDNF Val66Met polymorphism.

Download figure to PowerPoint

Finally, we tested the effect of the BDNF Val66Met polymorphism on clinical change. In the multivariate models, BDNF Val66Met polymorphism was not associated with changes in depressive (Z = 0.81, = 0.416) or manic symptoms (Z = 0.51, = 0.613) or with significant time interactions.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

People with bipolar disorder had a differential change in peripheral levels of BDNF along the treatment of mood episodes depending on their genotype. Met carriers had a significantly different trajectory and a tendency for a less substantial increase in peripheral levels than Val homozygotes. This is the first clinical trial with frequent assessments and a relatively large sample demonstrating such effects in BD.

The BDNF Val66Met polymorphism plays a critical role in the expression of BDNF and prompts cerebral modifications. In a recent study, Matsuo et al. [29] studied the differences between patients with BD and healthy volunteers according to memory function, diagnosis, and BDNF Val66Met polymorphism. Patients with BD showed smaller regional brain volumes than healthy controls, and Met carriers within the same population were likely to have smaller regional brain volumes as compared with Val homozygotes. In our study, the Met carriers had lower serum BDNF levels at baseline compared with Val homozygotes. This is in line with a study of Rybakowski et al. [30] who assessed the neurocognitive performance in patients with BD compared with schizophrenic and control subjects. Interestingly, the only difference related to the BDNF Val66Met polymorphism in the performance of the Wisconsin Card Sorting Test was determined in patients with BD, depicting a significantly better accomplishment in subjects with the BDNF Val66Val genotype compared with the BDNF Val66Met genotype. These results may be caused by the effect of the BDNF Val66Met polymorphism on the activity-dependent secretion of BDNF [11], which might be particularly evident during acute affective episodes rather than in euthymia. The dysregulation of BDNF release may also help to explain the evolution of the serum BDNF levels in our study. They varied notably showing an initial slump in the group of Met carriers. All in all, activity-dependent secretion on BDNF might be crucial for clinical response in BD unlike what has been described in unipolar depression in a recent meta-analysis [31].

Previous pilot longitudinal studies of BDNF levels in BD reported an increase in BDNF levels after the patients had received pharmacological treatment. Tramontina et al. [20] assessed serum BDNF levels in 10 manic in-patients before and after a mean of 52 days of treatment. Palomino et al. [19] studied plasma BDNF levels in 14 patients with BD undergoing their first psychotic episode. Both studies reported lower baseline BDNF levels during acute episode compared with controls, with no differences in BDNF levels between patients and controls after acute treatment and after treatment for 6 months respectively. These results are consistent with the literature on major depression, in which an increase in serum BDNF is usually seen with antidepressant treatment [32]. In our study, we did not find initial differences in serum BDNF levels between patients and controls as in the study of Rybakowski et al. [33].

Despite the advances achieved in the field of biomarkers in BD, their usefulness related to clinical improvement and treatment response still remains to be elucidated. In a recent open-label study, Grande et al. [21] found a time per mood polarity interaction. Increasing BDNF levels were associated with clinical improvement from a depressive episode, and decreasing BDNF levels were related to clinical improvement from a manic/mixed episode. A dose effect could have influenced outcome as patients were treated with quetiapine extended-release 300 mg/day monotherapy. Higher quetiapine doses may need to be used for manic or mixed episodes. In the present study, achieving both treatment response and full remission determined an increase in serum BDNF levels, regardless of the polarity of the index episode. Similar results have been described in individuals with major depressive disorder. In a 12-week follow-up study, Kurita et al. [34] found increased plasma BDNF levels in patients with depressive syndrome who underwent remission in contrast with non-responders, showing a decrease in plasma BDNF levels during monitoring. Likewise, in a cross-sectional study, Molendijk et al. [35] compared the close link between clinical response and serum BDNF levels in patients with major depressive disorder. Serum BDNF levels were low in patients with a current depressive episode regardless of receiving antidepressant treatment compared with healthy controls. Nevertheless, serum BDNF levels were higher in patients in an acute depression with antidepressants compared with those without. Remarkably, the difference between patients in a current depressive episode and controls became imperceptible after 6 months of having attained sustained full remission, which may indicate the prolonged biological impact of affective symptomatology on BDNF.

Different variables that could have confounded the presented results were inspected. As it has been reported that BDNF decreases along the neuroprogression of the disorder [36], chronicity of the disorder was taken into account. Considering the systemic toxicity of the affective episodes, the total number of episodes was evaluated [37, 38]. The high affinity of quetiapine for the serotonin-2A receptor and a subsequent increase in BDNF levels could also be argued as an explanation for the higher BDNF levels in this group [35, 39, 40]. Considering the influence of other treatments on the BDNF levels in our study, it is worth mentioning that neither lithium nor antidepressants showed any direct repercussion on these levels. The limited sample size is a plausible cause.

In this study, some limitations should be taken into account. A possible caveat is related to the neuroprogression hypothesis. At this time, there are data suggesting that the biology of BD may vary according to illness stage [41, 42]. This may signify that BDNF levels only drop significantly in certain stages. This may have acted as a floor effect in the current trial, as BDNF levels were not different from healthy controls at baseline. Moreover, assessment of BDNF levels was carried out in the peripheral blood. Despite evidence of correlation between cortical BDNF and serum BDNF and that BDNF crosses the blood–brain barrier, the presumption that such findings might occur in the human brain needs to be confirmed [43]. In this regard, studies about postmortem brain tissue and cortical integrity are providing more evidence in this framework. As this is a naturalistic study, diversity in treatment may have had an effect on BDNF levels. Mood stabilizers, antidepressants, and atypical antipsychotics are reported to interfere with pathways related to BDNF. Despite having assessed the treatment with quetiapine extended release, lithium, and antidepressants, other pharmacological strategies were difficult to evaluate due to the reduced sample under these treatments. Nevertheless, the fact that patients were medication-free at baseline mitigates chronic treatment effects, which strengthens the results presented here.

There is growing evidence linking circulating BDNF and systemic toxicity, and perhaps as one of the most relevant mediators of allostatic overload in BD [2, 18, 44]. There are several interesting competing hypotheses of how clinical improvement and serum BDNF levels are linked and how polymorphisms can influence these levels. This study suggests that assessment of peripheral BDNF levels may be useful as biomarkers of treatment response. A deeper understanding of the molecular determinants involved in BDNF-signaling cascades may provide a means for monitoring treatment response. Finally, the finding that Met carriers have less pronounced increase in BDNF levels is consistent with previous biological hypotheses and may provide future pharmacogenetic opportunities.

Declaration of interest

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References

Dr. I. Grande has received a research grant Río Hortega Contract (CM12/00062), Instituto de Salud Carlos III, Spanish Ministry of Economy and Competiveness, Barcelona, Spain, and has served as a speaker for AstraZeneca.

G.R. Fries has a doctoral scholarship from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), Brazil.

Dr. B.N. Frey has received research support from CIHR, Father Sean O'Sullivan Research Centre, Hamilton Health Sciences Foundation, Stanley Medical Research Foundation, Eli Lilly, Pfizer, and Bristol-Myers Squibb; has received speaker fees/travel support from AstraZeneca, Lundbeck, and Pfizer.

Professor E. Vieta has received research grants and served as consultant, advisor, or speaker for the following companies: Alexza, AstraZeneca, Bristol-Myers Squibb, Cephalon, Eli Lilly, Forest Research Institute, Gedeon Richter, GlaxoSmithKline, Janssen-Cilag, Jazz, Lundbeck, Merck Sharp & Dohme, Novartis, Otsuka, Pfizer Inc, Sanofi-Aventis, Servier, Takeda, and UBC.

Professor Kapczinski has received grant/research support from Astra-Zeneca, Eli Lilly, Janssen-Cilag, Servier, CNPq, CAPES, NARSAD, and Stanley Medical Research Institute; has been a member of the speakers' boards for Astra-Zeneca, Eli Lilly, Janssen, and Servier; and has served as a consultant for Servier.

The other authors declare no conflict of interests.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Material and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Declaration of interest
  9. References