1. Introduction
Bipolar disorder (BD) is a recurrent mood disorder affecting 1–3% of the world population [Reference Merikangas, Jin, He, Kessler, Lee and Sampson1]. It has a long-term outcome with incomplete recovery between episodes, cognitive impairment, and functional decline [Reference Merikangas, Jin, He, Kessler, Lee and Sampson1–Reference Grande, Berk, Birmaher and Vieta6]. Its chronic course is associated with high rates of morbidity and mortality, making bipolar disorder one of the main causes of disability among young and working-age people [Reference Vieta, Salagre, Grande, Carvalho, Fernandes and Berk7]. Cognitive dysfunction is the key feature of functional impairment throughout periods of mania, depression and euthymia in BD [Reference Tse, Chan, Ng and Yatham8]. The neurobiological underpinnings of cognitive dysfunction remain unknown in psychiatric disorders [Reference Millan, Agid, Brüne, Bullmore, Carter and Clayton9]. Cognition is affected by a range of medical issues and neurochemical mechanisms [Reference Bauer, Pascoe, Wollenhaupt-Aguiar, Kapczinski and Soares10]. There is evidence that endocrine (such as obesity) and cardiovascular issues which are more common in BD than in general population [Reference McIntyre, Woldeyohannes, Soczynska, Miranda, Lachowski and Liauw11, Reference Calkin and Alda12], may be negatively associated with cognitive progression [Reference Strejilevich, Samamé and Martino13, Reference Mora, Portella, Martinez-Alonso, Teres, Forcada and Vieta14].
Cognition is better understood in terms of complex networks operating over multiple temporal scales and incorporating various dimensions: from cellular cascades to cerebral circuits and, ultimately, society [Reference Millan, Agid, Brüne, Bullmore, Carter and Clayton9]. Moreover, a diverse palette of neuromodulators including acetylcholine, cytokines and neurotrophic proteins such as brain-derived neurotrophic factor (BDNF) influence cognitive performance [Reference Millan, Agid, Brüne, Bullmore, Carter and Clayton9]. In the same line, there are some neurobiological factors such as BDNF, immune-inflammatory and oxidative stress markers that have been consistently reported to be associated with brain structure and function and to be relevant to physiological and pathological neurodevelopment [Reference García-Bueno, Caso and Leza15, Reference Pfaffenseller, Fries, Wollenhaupt-Aguiar, Colpo, Stertz and Panizzutti16]. Associations between alterations in these systems have been reliably described in children and adults, across different mental disorders [Reference Fernandes, Berk, Turck, Steiner and Gonçalves17, Reference Mansur, Cunha, Asevedo, Zugman, Levandowski and Gadelha18]. In particular, compelling evidence about these neurobiological factors in bipolar disorder has been found [Reference Modabbernia, Taslimi, Brietzke and Ashrafi19–Reference Fernandes, Molendijk, Köhler, Soares and CMGS21]. Indeed, several pathophysiological mechanisms have been proposed and investigated, to understand the interaction between these neurobiological factors and mood symptomatology in BD [Reference Pfaffenseller, Fries, Wollenhaupt-Aguiar, Colpo, Stertz and Panizzutti16, Reference Rosa, Singh, Whitaker, De Brito and Lewis22, Reference Rosenblat, Brietzke, Mansur, Maruschak, Lee and McIntyre23]. However, significantly less investigation has been conducted to elucidate the effect of inflammation, oxidative stress and neuroplasticity on cognition in BD [Reference Barbosa, Bauer, Machado-Vieira and Teixeira24, Reference Rosenblat and McIntyre25].
The main aim of the current study was to investigate the association of cognitive performance with neuroprotective and neurodegenerative mechanisms (neurotrophins, and inflammation and oxidative stress markers) in patients with BD. We hypothesized that (1) bipolar individuals (euthymic and manic) would display a different pattern of BDNF, inflammatory-cytokine and oxidative stress markers as compared to healthy controls; and (2) cognitive performance would be associated with these neurobiological factors regardless of mood state.
2. Methods
2.1. Participants
A total of 133 individuals were enrolled in the study: 49 healthy controls, 52 euthymic bipolar and 32 manic bipolar individuals. All of them were evaluated by clinical interview with a psychiatrist, and underwent neuropsychological and biochemical tests.
Euthymic bipolar patients were recruited from the Outpatient Lithium Clinic at Hospital Universitari Santa Maria, Lleida, from 2003 until 2011. All of them met Diagnostic and Statistical Manual of Mental Disorders, fourth edition, text revision (DSM-IV-TR) criteria for BD, as was determined by a Psychiatrist using the Structured Clinical Interview for Axis I Disorders [Reference First, Spitzer, Gibbon and Structured26]. They were aged between 18 and 65 years old. Euthymia state was determined when bipolar individuals obtained a total 17-item Hamilton Rating Scale for Depression (HAMD) [Reference Hamilton27] score below 8 and a total Young Mania Rating Scale (YMRS) [Reference Young, Biggs, Ziegler and DA28] score below 6 for at least 3 months prior to the assessment [Reference Mur, Portella, Martínez-Arán, Pifarré and Vieta29]. Significant non-psychiatric illness, substance abuse or dependence or electroconvulsive therapy during the preceding year, were the exclusion criteria.
Manic bipolar patients were recruited from the Inpatient Psychiatric Unit, at Hospital Universitari Santa Maria, Lleida, during a manic phase, from 2012 until 2014. Inclusion and exclusion criteria were the same as for euthymic patients, except for the criteria of clinical stability: manic patients were included if they had a total score of 14 or above in the YMRS [Reference Tohen, Frank, Bowden, Colom, Ghaemi and Yatham30]. Blood extraction was performed at the beginning of the hospitalization in the acute phase of the disorder. Cognitive assessment was performed just prior to discharge when major manic symptoms were partially remitted, in order to guarantee collaboration of manic patients, without biasing cognitive performance, even though it meant gathering information at two separate points in time.
Forty-nine healthy controls were enrolled with advertisements and from non-medical hospital staff. Controls had no current or past psychiatric history, as determined by the Structured Clinical Interview for DMS-IV Axis I Disorders [Reference First, Spitzer, Gibbon and Structured26]. Additionally, healthy subjects were excluded if there was a family history of any Axis I disorder in a first-degree relative. Healthy subjects underwent the same exclusion criteria as the patients, and were assessed at the same full clinical and demographics interview by a trained psychiatrist. The Local Ethics Committee approved the study and written informed consent was obtained from all participants.
2.2. Demographic, clinical and pharmacological data
Demographic variables included age, gender, years of education and current work status. Body mass index (BMI) was also calculated for each participant. The established criteria for BMI were: normal weight, BMI of 18.5–24.9 kg/m2; overweight, BMI of 25.0–29.9 kg/m2; and obese, BMI > 30 kg/m2 (National Heart LaBI NI of D and D and KD [31]. Psychiatric variables were obtained from the sample of bipolar patients, including: age at onset of illness, number of prior manic episodes and hospitalizations, period of stabilization (years), history of psychotic symptoms, seasonal pattern, suicide attempts and bipolar subtype (I or II) during the psychiatric interview. Other physical and medical issues (e.g. cardiovascular, neurological, gastrointestinal, haematological, renal, hepatic, respiratory or endocrine illnesses) and concurrent psychiatric and non-psychiatric medications were recorded in the same interview. Biochemical tests were performed in all patients, including thyroid function, lipid profile, serum lithium levels and urine drug testing.
2.3. Neuropsychological assessment
To characterize the cognitive functioning, a selected battery that included neuropsychological tasks covering the most impaired cognitive domains in BD, i.e. executive and memory functioning [Reference Mora, Portella, Forcada, Vieta and Mur3, Reference Mur, Portella, Martínez-Arán, Pifarré and Vieta29, Reference Martinez-Aran, Torrent, Tabares-Seisdedos, Salamero and Daban32] was administered to all participants. The estimated mean intelligence quotient (IQ) of the subjects was obtained from the weighted scores of the Vocabulary and Block Design subtests of the Wechsler Adult Intelligence Scale (WAIS-III) [Reference Wechsler33], on the basis that these two scores are highly correlated with total IQ.
The instruments administered were:
i Vocabulary, Block Design and Digits Subtests from WAIS-III [Reference Wechsler33];
ii Wisconsin Card Sorting Test (WCST) [Reference Heaton34], to assess executive function and perseverative behaviour;
iii Stroop Color and Word Test [Reference Golden and Stroop Color35], to evaluate selective attention and inhibition capacity;
iv FAS verbal fluency task of the Controlled oral Word Association Test/Categories [Reference Benton and Hamsher36], to assess executive function;
v Trail making Test (TMT), to evaluate processing speed parts A (TMT-A) and cognitive flexibility parts B (TMT-B) [Reference Reitan37];
vi Conners’ continuous Performance Test II (CPT-II) [Reference Conners38], to evaluate sustained attention, processing speed, and perseverative behavior;
vii The California Verbal Learning Test (CVLT) [Reference Delis39], to evaluate verbal learning, recall, and recognition;
viii Rey-Osterrieth Complex figure (RCFT) [Reference Meyers and Meyers40]; to evaluate visual memory.
2.4. Biochemical measures
For each participant (patients and healthy controls), blood samples were collected between 8:00 and 9:00 to avoid variations due to the circadian rhythm. Ten millilitres of blood were withdrawn from all participants by venipuncture into a free-anticoagulant vacuum tube. Serum was separated within 2 h by centrifugation at 3500 g during 15 min at room temperature. All samples were stored at ―80 °C until assayed in the IRBLleida Biobanc (B.0000682) and PLATAFORMA BIOBANCOS PT13/0010/0014. Afterwards, neurobiological factors were determined and all samples were assayed in duplicates. The neurobiological factors related to neurotrophins, inflammation and oxidative stress were examined as follows:
• Brain-derived neurotrophic factor (BDNF) serum levels were measured with sandwich-ELISA, using a commercial kit according to the manufacturer’s instructions (Chemicon, Temecula, CA, USA), as previously described [Reference Kauer-Sant’Anna, Kapczinski, Andreazza, Bond, Lam and Young41]. All samples were assayed in duplicates. The results were expressed as ng/ml. Intra- and inter-assay coefficients of variation were <12%.
• Serum pro-inflammatory cytokines, such as interleukin 6 (IL-6) and tumour necrosis α (TNF-α) and an anti-inflammatory cytokine, interleukin 10 (IL-10) were measured according to the procedures supplied by the manufacturer using highly sensitive sandwich-ELISA kits for TNF-α, IL-6 and IL-10 (Quantikine, R&D Systems, Minneapolis, Minn., USA). All samples were assayed in duplicates. The results were expressed as ng/ml. Intra- and inter-assay coefficients of variation were <10%.
• Oxidative damage was measured by the levels of lipid peroxidation using the thiobarbituric acid reactive substances (TBARS) method described by Wills [Reference Wills42]. All samples were assayed in duplicates and the results were expressed as nmol/mL.
2.5. Statistical procedure
Data analyses were carried out with the statistical package SPSS for Windows, version 22.0 (SPSS Inc., USA) and R statistical software (version 3.2.2) [Reference R Core Team43]. Demographics, clinical, and pharmacological characteristics of groups were compared with analysis of variance (ANOVAs) for continuous variables and Chi-square (or Fisher’s exact test) for categorical variables as descriptive statistics analyses. All variables were assessed for normality and those variables that were not normally distributed were log-transformed. To counteract the problem of multiple comparisons, Bonferroni correction was performed on variables that displayed statistical differences between the three groups in the ANOVAs. Performance on neuropsychological tasks was compared among three groups using a multivariate analysis of variance (MANOVA), and post-hoc analyses were then performed to establish pairwise differences. Neurobiological variables were also analysed with MANOVA, and post-hoc analyses. Parametric and non-parametric correlations were carried out to explore associations between clinical, demographical and neurobiological variables to be included in further analyses. Multiple linear regression models were built to study the association between different variables (demographics and neurobiological factors) and cognitive functioning. For this purpose, neuropsychological tasks were z-transformed and grouped by cognitive domains (i.e., executive function, processing speed, inhibition, attention, and verbal and visual memories) based on our previous work [Reference Mur, Portella, Martínez-Arán, Pifarré and Vieta44]. A composite score was created for each cognitive domain, which included all the tests encompassed within each domain (see Table 2). The composite score was calculated as an arithmetic mean and it was used to avoid redundant information of separate tests. On measures of reaction time (low scores indicating good performance), z-scores were reversed before forming the composite score. Each cognitive composite score was then used as a dependent variable in the regression models. Multicollinearity was checked for each model using tolerance and variance inflation factor (VIF) criteria. Significant or relevant demographic variables, such as current age, premorbid IQ, BMI and neurobiological variables were included as predictive factors. The obtained results will be explained using standardized coefficient (β) and p values.
Firstly, in order to elucidate the role of clinical and neurobiological factors on cognitive functioning, Pearson’s correlations were performed between these variables. Further regression models were then carried out in the sample of patients (euthymic and manic individuals) including the independent factors mentioned above together with duration of illness, number of manic and depressive episodes and number of hospitalizations.
3. Results
3.1. Demographic and clinical characteristics of the sample
Demographics, clinical features and neurobiological factors levels of the 133 participants are depicted in Table 1. Euthymic bipolar patients and healthy controls were comparable in age and gender. There were statistical differences between groups in terms of years of education, IQ, BMI, HAMD and YMRS.
With respect to pharmacological variables, most of euthymic bipolar patients (81%) were on lithium treatment (15 patients were on lithium monotherapy and 27 on combination treatment -plus antidepressant or antipsychotic-). Off the manic bipolar patients group, half of them (16 individuals) were on lithium combination treatment and the other half (16 individuals) with another mood stabilizer plus antipsychotic (i.e., valproate; see Table 1).
In terms of medical issues, there was a history of cardiovascular and endocrine issues in all groups without statistical differences (χ2 = 11.8, df = 16, p = 0.76). Particularly, they had high blood pressure (9 healthy controls, 4 euthymic and 2 manic bipolar individuals), history of diabetes mellitus type 2 (4 healthy controls, 3 euthymic and 1 manic bipolar individuals), and subclinical (4 euthymic bipolar individuals) and clinical (2 manic) hypothyroidism. Regarding body mass index (BMI), euthymic bipolar individuals were more overweight than manic individuals and healthy controls (F = 4.8; p = 0.01).
3.2. Neuropsychological performance
All neuropsychological variables included in each cognitive domain and univariate effects are listed in Table 2. As expected, in the post-hoc analyses, euthymic and manic individuals performed worse compared to healthy control group (p < 0.0001) in executive functioning, inhibition, processing speed, verbal and visual memory (see Fig. 1). The only domain that displayed statistical differences between euthymic and manic groups was verbal memory (p < 0.0001), in which euthymic patients showed better performance than manic patients, but worse than healthy controls.
3.3. Descriptive analyses of neurobiological factors and associations with demographic, clinical and pharmacological variables
BDNF levels in both groups of bipolar patients (euthymic and manic individuals; p = 0.039 and p < 0.0001 respectively) were significantly lower compared to healthy controls. Conversely, the pro-inflammatory cytokine interleukin 6 (IL-6) levels were significantly higher in the group of manic patients compared to healthy controls (p = 0.019), and the anti-inflammatory interleukin 10 (IL-10) levels were higher in both groups of patients compared to healthy controls (both comparisons with statistical significance; p < 0.0001). Levels of TBARS were higher in euthymic bipolar group compared to healthy controls (p = 0.003); but there were not statistical differences between manic patients group and the other groups.
* =<0.0001.
Values represent mean (SD) unless otherwise specified
Statistics of between-subject effects from the multivariate analyses of variance. Abbreviations: GAF = Global Assessment of Functioning, YMRS = Young Mania Rating Scale, HAM-D = Hamilton Rating Scale for Depression, NA = Not applicable, NS = Not significant. HC=Healthy Controls, Eu=Euthymic Patients, Ma=Manic patients.
Correlations between neurobiological factors and demographic variables showed that there was an association between IL-6 and BMI (r = 0.515; p < 0.0001) and with current age (r = 0.313; p < 0.0001) in all participants. Premorbid IQ displayed a significant relationship with IL-10 (r=―0.268; p = 0.002) and with levels of oxidative damage (TBARS; r=―0.247; p = 0.005).
BDNF showed a negative correlation with HDRS (r=―0.251, p = 0.004) and YMRS (r=―0.284, p = 0.001) in all participants. Conversely, as it was expected, IL-6 displayed a positive association with the same scales (HDRS: r = 0.176, p = 0.045; and YMRS: r = 0.192, p = 0.029). IL-10 did not show any association with these scales. When analysing patients alone (euthymic and manic), correlations between clinical variables and neurobiological factors showed that only IL-6 was associated with duration of illness (r = 0.245; p = 0.025).
Regarding pharmacological treatment, most bipolar euthymic patients and half of the manic patients were on lithium treatment. There was not any association between lithium and neurobiological factors levels, with the exceptions of IL-6, which correlated positively with lithium levels (r = 0.33; p = 0.011); and T-BARS, where patients who were on lithium combination treatment displayed higher levels (F = 4.7; p = 0.035) compared to patients who were on lithium monotherapy.
Bipolar type I and type II euthymic individuals did not display any statistical differences in neurobiological factors levels (data not shown).
3.4. Association of neurobiological factors, demographics and BMI with cognitive functioning
Six regression models were performed, one for each neuropsychological domain in order to explore the association of neurobiological factors and other demographic variables with cognitive functioning (see Table 3). Only the models of executive functioning and verbal memory were significant, with a R2 = 0.56 and 0.58, respectively. In particular, BDNF (β = 0.01, p = 0.02) was the only neurobiological factor that explained executive functioning together with other variables: being manic (β=―0.4, p = 0.012), current age (β=―0.023, p < 0.0001) and premorbid IQ (β = 0.03, p < 0.0001), and explained a 56% of variance (F = 16.5; df = 9115; p < 0.0001) (see Fig. 2). Again BDNF (β = 0.013, p = 0.005) together with BMI (β=-0.4, p = 0.005) and other variables [being euthymic (β=―0.345, p = 0.019), being manic (β=―1.07, p < 0.0001), current age (β=-0.015, p = 0.002) and premorbid IQ (β = 0.018, p < 0.0001)] predicted worse performance in verbal memory, and explained a 58% of variance (F = 17.2; df = 9.115; p < 0.0001) (see Fig. 2). The rest of neurobiological factors did not predict any other cognitive domain, although the models were significant.
a Values shown as mean (SD).
Statistics of between-subject effects from the multivariate analyses of variance.
Abbreviations: CVLT = California Verbal Learning Test, TMT = Trail Making Test, RCFT=Rey Complex Figure Test, HC=Healthy Controls, Eu=Euthymic Patients, Ma=Manic patients.
3.5. Association of neurobiological factors and clinical variables with cognitive functioning in Bipolar Patients
Again six regression models were performed only analysing euthymic and manic bipolar individuals (see Table 4). Current age and premorbid IQ were predictor factors in all models (p < 0.05). Being a manic bipolar individual was associated with worse performance in executive functioning (β=―0.33, p = 0.029), verbal memory (β=―0.864, p < 0.0001) and visual memory (β=―0.459, p = 0.039).
Neuropsychological performance adjusted by age, premorbid IQ and years of education from all participants.
Abbreviations: IQ: intelligence quotient; BMI: body mass index.
Of the clinical variables, the number of manic episodes (β=-0.066, p = 0.011) was the only clinical variable that was associated with worse performance on the attention domain, together with current age (β=―0.025, p = 0.001) and premorbid IQ (β = 0.014, p = 0.015), and explained a 32% of variance (F = 2.7; df = 12, 70; p = 0.004). The rest of clinical variables were not associated with any of the other domains.
Verbal memory was associated with BDNF levels (β = 0.023, p = 0.001), BMI (β=―0.05, p = 0.017), together with being manic (β=―0.864, p < 0.0001), current age (β=―0.018, p = 0.023) and premorbid IQ (β = 0.022, p = 0.001), and explained a 54% of variance (F = 7; df = 12, 70; p < 0.0001).
4. Discussion
The most remarkable finding of this study was the significant association of neurotrophins with illness phases and their significant impact on cognitive dysfunction in bipolar individuals, affecting particularly executive functioning and verbal memory. This represents valuable information as few studies have explored the role of neurobiological factors on cognitive symptoms in BD, and results had been contradictory.
Scatter plots displaying predicted executive functioning (a) and verbal memory (b) by BDNF levels clustered by euthymic and manic bipolar individuals and healthy controls.
Linking neurobiological factors and cognition, our results are partially consistent with the few previous studies that focus on BNDF and cognitive outcome in BD, specifically in executive functioning [Reference Lee, Wang, Chen, Chang and Chen45] and verbal memory [Reference Dias, Brissos, Frey, Andreazza, Cardoso and Kapczinski46, Reference Cao, Bauer, Sharma, Mwangi, Frazier and Lavagnino47]. In the same line, TNF-α, was found to be negatively correlated with accuracy on the delayed memory component on the Rey Auditory Verbal Learning Test (RAVLT) in euthymic type I individuals [Reference Doganavsargil-Baysal, Cinemre, Aksoy, Akbas, Metin and Fettahoglu48]. Remarkably, the presence of verbal memory impairment across mood phases may indicate that these deficits are trait markers of bipolar illness [Reference Gualtieri and Johnson49].
Furthermore, current evidence indicates that the BD phenotype is heterogeneous and there may be different illness trajectories that could at least in part be explained by distinct underlying pathophysiology [Reference Berk, Kapczinski, Andreazza a, Dean and Giorlando50, Reference Birmaher, Gill, Axelson, Goldstein, Goldstein and Yu51]. The current study provides evidence that other conditions such as obesity can play a role in defining verbal memory as a trait marker, as has been observed in previous studies [Reference Yim, Soczynska, Kennedy, Woldeyohannes, Brietzke and McIntyre52–Reference Lackner, Bengesser, Birner, Painold, Fellendorf and Platzer54]. In a recent study from our group, the interaction of bipolar disorder and obesity was found to impact cognitive dysfunction at a single point in time and long-term [Reference Mora, Portella, Martinez-Alonso, Teres, Forcada and Vieta14].
Strikingly, the relation between BDNF and verbal memory found in our study has not been replicated in healthy population in a recent study [Reference Wilkosc, Markowska, Zajac-Lamparska, Skibinska, Szalkowska and Araszkiewicz55]. There are also previous data showing a positive [Reference Erickson, Prakash, Voss, Chaddock, Heo and McLaren56], a negative [Reference Niitsu, Shirayama, Matsuzawa, Hasegawa, Kanahara and Hashimoto57] or no correlation [Reference Kim, Fagan, Goate, Benzinger, Morris and Head58] between serum BDNF and memory functions in healthy subjects. All these conflicting results could be explained by the fact that the relationship between cognition and BDNF may be affected by many other factors such as mood state, sociodemographic and lifestyle factors (for example: physical exercise, BMI and others) [Reference Wilkosc, Markowska, Zajac-Lamparska, Skibinska, Szalkowska and Araszkiewicz55]. In the same line, BDNF Val66Met genotype has been described as a potential risk factor for obesity and insulin resistance measures in patients with BD who are also receiving antipsychotic medication [Reference Bonaccorso, Sodhi, Li, Bobo, Chen and Tumuklu59].
Regarding the association between clinical variables and neurobiological factors, duration of illness was the only clinical variable related with an inflammatory marker (IL-6) in the current study. In the same line, another study found that IL-6 levels showed significant differences between early and late stages of BD [Reference Grande, Magalh??es, Chendo and Stertz60]. Consistent with our results, in a recent meta-analysis [Reference Fernandes, Molendijk, Köhler, Soares and CMGS21], duration of illness in euthymia was not associated with BDNF levels, suggesting that it may not be a useful marker of illness stage but it could be a marker of illness activity instead. Furthermore, it has been reported that BDNF levels did not differ between bipolar patients, unaffected first-degree relatives and healthy controls. Thus, BDNF levels may not reflect high genetic risk for BD, acting as state marker rather than trait marker for the disease [Reference Nery, Gigante, Amaral, Fernandes, Berutti and Almeida61]. The correlation of BDNF levels with mood psychometric scales and mood phases may corroborate this hypothesis.
In terms of the association between psychopharmacological treatment and neurobiological factors, in our study lithium variables were not significantly associated with BDNF levels which may indicate that lithium did not influence the relationship between neurotrophins and cognition. In this paper the impact of psychopharmacological treatment on the association between neurobiological factors and cognitive dysfunction was not directly assessed because it was not methodologically designed for this purpose. Focusing on lithium, it has been reported in the current literature that excellent lithium responders (ELR) had higher plasma BDNF levels and performed better on all neuropsychological tests than the remaining lithium patients [Reference Rybakowski and Suwalska62] but worse than healthy controls in the long-term [Reference Mora, Portella, Forcada, Vieta and Mur63]. Treatment with mood stabilizers (i.e. lithium) was associated with lower levels of DNA methylation of BDNF promoter [Reference D’Addario, Dell’Osso, Palazzo, Benatti, Lietti and Cattaneo64]. Further analyses of neurobiological factors by selecting individuals taking lithium are warranted to understand the effects of lithium on cognition.
In our study, neither inflammatory nor oxidative stress markers were associated with cognitive performance, which differ from previous studies [Reference Bauer, Pascoe, Wollenhaupt-Aguiar, Kapczinski and Soares10, Reference Doganavsargil-Baysal, Cinemre, Aksoy, Akbas, Metin and Fettahoglu48, Reference Carvalho, Köhler, Fernandes, Quevedo, Miskowiak and Brunoni65]. Of note, many other difficult-to-control factors related with mood symptoms which may result in inflammation and oxidative stress, were not assessed such as: undiagnosed inflammatory medical comorbidities, history of early childhood adversity [Reference Jiménez, Solé, Arias, Mitjans, Varo and Reinares66], dysfunctional gut-microbiota, dietary patterns, and low-grade idiopathic systemic inflammation [Reference Bercik67–Reference Jacka, Neil, Opie, Itsiopoulos, Cotton and Mohebbi72].
The reported discrepancies could arguably be attributed to some methodological limitations of the current study that warrant acknowledgement: the variety of techniques to measure neurobiological factors, demographics and clinical characteristics of the sample, such as age and gender-related brain characteristics [Reference Blumberg, Kaufman, Martin, Whiteman, Zhang and Gore73], BD subtype [Reference Strasser, Lilyestrom, Ashby, Honeycutt, Schretlen and Pulver74] and medication history [Reference Yucel, McKinnon, Taylor, Macdonald, Alda and Young75]. Although the sample size of the present study is acceptable given the type of population, the study should be seen as exploratory given the number of statistical analyses conducted. In any case, these analyses served to cover different aspects of the intricate variables involved in cognitive performance of bipolar patients. In our sample euthymic bipolar disorder individuals might represent mid-late stage of the illness with low rates of recurrence which might not be representative of the majority of bipolar patients, being biased to patients with better outcomes [Reference Goodwin and Vieta76]. But including manic patients groups in the sample has allowed us to demonstrate the cognitive consequences of neurotrophins in different mood phases. Moreover, our sample was balanced in terms of gender and there were no statistical differences between groups according to BD type I and type II in terms of neurobiological factors levels. The cross-sectional nature of the study and the methodological limitations described above preclude the establishment of more conclusive evidence about the neuroprogression or neurodevelopmental nature of the illness or about the causality and direction of the associations. But the correlation findings shed light on this intriguing field. Some authors state that there may be a subgroup of patients not exceeding 25–40% that could show phenomena of cycle acceleration [Reference Baldessarini, Salvatore, H-MK, Imaz-Etxeberria and Gonzalez-Pinto77]. Moreover, there is growing evidence that about one third of euthymic BD patients have more severe cognitive deficits than usually reported in the literature, while a similar proportion are indistinguishable from healthy controls in terms of cognitive functioning [Reference Burdick, Russo, Frangou, Mahon, Braga and Shanahan78]. Clinical evidence supporting the concept of neuroprogression in BD is scarce and limited [Reference Martino, Samamé, Marengo, Igoa and Strejilevich79], but the longitudinal studies that have been published to date are too small and too short-term to prove that there is no progression [Reference Vieta, Berk, Schulze, Carvalho, Suppes and Calabrese80]. This controversy is currently still under debate [Reference Vieta, Salagre, Grande, Carvalho, Fernandes and Berk7]. Future investigation using larger samples, drug-naïve patients, longitudinal designs incorporating repeated measures of these markers [Reference Frey, Andreazza, Houenou, Jamain, Goldstein and Frye81], selecting homogenous treatment response like excellent lithium responders and assessing the impact of childhood trauma [Reference Jiménez, Solé, Arias, Mitjans, Varo and Reinares66] and lifestyle factors (for example: physical exercise and BMI) [Reference Wilkosc, Markowska, Zajac-Lamparska, Skibinska, Szalkowska and Araszkiewicz55] are needed to reveal the specific connections between BD, neurobiological factors and neurocognitive performance. A careful attention to attrition in those cohorts will be critical.
Abbreviations: IQ: intelligence quotient; BMI: body mass index.
4.1. Conclusion
In conclusion, our findings support the hypothesis that neurotrophic mechanisms correlate with clinical variables and cognitive functioning, specifically in executive functioning and verbal memory. Other variables such as mood state and obesity may underlie the link between neurotrophins and cognition. Additional work is needed to understand how pro-inflammatory and oxidative damage processes affect brain function and induce cognitive impairment [Reference Bauer, Pascoe, Wollenhaupt-Aguiar, Kapczinski and Soares10]. Monitoring neurobiological factors levels at the time of assessment in clinical psychiatry could help to tailor specific and individualized treatment interventions not only to treat mood symptoms but also to revert biological changes associated with the illness [Reference Pfaffenseller, Fries, Wollenhaupt-Aguiar, Colpo, Stertz and Panizzutti16] such as cognitive and functional decline.
Role of funding source
This study was financially supported by the following: an IRBLleida (Biomedicine Research Institute) Contest for a research project for Medical Registers (P10062, 2010); and a Spanish FIS-MSC Grant (number PI11/01956, 2011). M.J.P. is funded by the Ministerio de Ciencia e Innovación of the Spanish Government and by the Instituto de Salud Carlos III through a ‘Miguel Servet’ research contract (CP16-0020, 2016); National Research Plan (Plan Estatal de I + D+I 2016–2019); and co-financed by the European Regional Development Fund (ERDF). The supporters had no role in the design, analysis, interpretation, or publication of this study.
Conflict of interest
E.V. has received research grants and served as consultant, advisor or speaker for the following companies: AB-Biotics, Almirall, Allergan, AstraZeneca, Bristol-Myers Squibb, Dainippon Sumitomo Pharma, Elan, Eli Lilly, Ferrer, Forest Research Institute, Geodon Richter, Glaxo-Smith-Kline, Janssen-Cilag, Jazz, Lundbeck, Merck, Novartis, Organon, Otsuka, Pfizer Inc., Roche, Sanofi-Aventis, Servier, Solvay, Schering-Plough, Shire, Sunovion, Takeda, United Biosource Corporation, and Wyeth. E.V. has received research funding from the Spanish Ministry of Science and Innovation, the Stanley Medical Research Institute and the 7th Framework Programme of the European Union. The rest of the authors have no conflicts of interest to declare.
Acknowledgements
We thank the staff of the Department of Psychiatry of Hospital Universitari Santa Maria, Lleida; Núria Vidal D.Clin.Psy, (funded by a Spanish FIS-MSC Grant [PI11/01956]), from Hospital FREMAP Barcelona, Catalonia, Spain, who performed cognitive assessments; E. Vieta thanks the support of the Spanish Ministry of Economy and Competitiveness (PI15/00283) integrated into the Plan Nacional de I + D+I and cofinanced by ISCIII-Subdirección General de Evaluación y el Fondo Europeo de Desarrollo Regional (FEDER); CIBERSAM; and the Comissionat per a Universitats i Recerca del DIUE de la Generalitat de Catalunya to the Bipolar Disorders Group (2017 SGR 1365). M.J. Portella thanks the suport of the Catalan Department of Health (SLT006_17_177); and the Agència de Gestió d’Ajuts Universitaris i de Recerca de la Generalitat de Catalunya to the Research Group in Psychiatric Disorders (2017 SGR 1343). We also thank the patients and healthy controls who participated in the study for their kind cooperation. We also thank to Rebecca Oglesby MD, who kindly helped us with the language editing.
Comments
No Comments have been published for this article.