RNA editing-based biomarker blood test for the diagnosis of bipolar disorder: protocol of the EDIT-B study

Abstract

Introduction

Misdiagnosis of bipolar disorder (BD) can lead to ineffective treatment, increased risk of manic episodes, and increased severity. Objective diagnostic tests or precise tools to diagnose BD and distinguish it from major depressive disorder (MDD) in depressed patients are lacking.

Aim

To assess the external diagnostic validity of a blood-based test using an RNA epigenetic signature for the differential diagnosis of BD versus MDD in patients with depression.

Methods and analysis

Multicentre cross-sectional study including an adult sample of inpatients or outpatients diagnosed with BD or MDD, currently treated for a major depressive episode. A structured diagnostic interview based on validated scales will be conducted. Sociodemographic variables, clinical history, toxic consumption, current treatment and quality of life will be assessed. Blood samples will be obtained and stored at −80 °C until RNA sequencing analysis. The EDIT-B is a blood-based test that combines RNA editing biomarkers and individual data (e.g., age, sex, and tobacco consumption). The clinical validation performance of the EDIT-B will be evaluated using the area under the curve, sensitivity, specificity, positive and negative predictive values, and likelihood ratios.

Ethics and dissemination

The principles of the Declaration of Helsinki 2013, precision psychiatry research and good clinical practice will be followed. The Research Ethics Committees of the participating centres approved the study. Participants will receive an information sheet and must sign the informed consent before the interview. Participants’ data will be pseudonymized at the research sites. Any publication will use fully anonymized data. Publications with the final study results will be disseminated in international peer-reviewed journals and presented at international conferences.

Study registration: This study has been registered on clinicaltrials.gov (NCT05603819). Registration date: 28-10-2022.

Introduction

Bipolar disorder (BD) is a common and disabling condition characterized by episodes of either mania or hypomania and depression. Depressive episodes are frequently the initial manifestation of BD because patients are more likely to seek professional help during these episodes than during manic, hypomanic, or mixed episodes. This tendency is often attributed to a lack of insight during the latter types of episodes [1], as well as a higher frequency of depression as the initial polarity [2], particularly among women [3]. BD leads to significant impairments in personal, social, and occupational areas, placing substantial burdens on individuals and healthcare systems [4, 5].

The clinical presentation of depressive episodes in individuals with BD closely resembles that of individuals with major depressive disorder (MDD). Currently, the diagnosis of a depressive episode relies on clinical interviews and the use of psychometric instruments. The overlap in clinical presentation between BD and MDD, combined with the higher prevalence of MDD, often leads to inadequate screening and, consequently to misdiagnosis of BD [1, 6]. The estimated mean delay between the first depressive episode and BD diagnosis is approximately 9 years [7], which is likely greater for individuals with depressive polarity [8]. Nonetheless, certain clinical features—such as an age of onset between 15 and 25 years, subthreshold mania, frequent insomnia or hypersomnia, cyclothymic traits, recurrent suicidal ideation, and having a first-degree relative with BD—have shown promise in predicting the onset of BD, particularly among young adults. Early differentiation is important as pharmacological interventions seem to be more effective at an early illness course [9]. Although these features do not provide a definitive solution, they are valuable clinical indicators that can help identify high-risk cases and improve early detection efforts.

However, early differentiation between BD and MDD remain challenging often leading to a wrong diagnosis. Misdiagnosis can lead to detrimental outcomes, such as reduced treatment efficacy, clinical progression of the illness [10], heightened risk of manic episodes, and increased severity. Both inaccurate diagnosis and inadequate treatment of BD may contribute to elevated rates of hospitalization, suicide attempts, and suicide deaths. Furthermore, it can worsen the risk of manic episodes and lead to more frequent rapid cycling [11, 12]. The economic burden associated with misdiagnosis or inadequate treatment of BD encompasses not only the direct expenses of treatment but also the significantly greater indirect costs, such as decreased productivity, elevated unemployment, and increased mortality [13].

A critical factor that substantially adds to the burden linked with BD is the absence of objective diagnostic tests and of dependable and precise tools to aid in the differential diagnosis between BD and MDD, enabling the implementation of suitable treatments for these disorders. In this respect, although there is evidence that immune biomarkers and lipid peroxidation are associated with certain major depression phenotypes [14] and literature have showed possible distinct immune biomarkers profiles for MDD and BD [15], characterization is still unclear. Thus, when assessing specifically the differences among activated immune and oxidative pathways between MDD and BD no biomarker differences have been showed [14]. This is an emerging field of “precision psychiatry”, which may radically change psychiatric practices worldwide [11]. Even so, the comprehensive analysis of immune, neurochemical, neurobiological, and metabolic biomarkers influencing mood disorders remains highly complex and will require time. Moreover, the absence of validated biological markers (biomarkers) to delineate boundaries between various subtypes of depression poses a significant challenge. In this context, blood biomarkers stand out as a promising tool, being reliable, minimally invasive, and easy to implement, potentially paving the way toward precision medicine in psychiatry.

Epigenetics regulates gene expression [16]. Various epigenetic modifications, such as deoxyribonucleic acid (DNA) methylation, histone modifications, ribonucleic acid (RNA) editing, and miRNA dysregulation, have been linked to psychiatric disorders [17]. Specifically, RNA editing enables the generation of multiple protein variants within cells and tissues by modifying ribonucleotides according to the corresponding DNA sequence, which involves substitutions, deletions, or insertions [18]. Alterations in neurotransmitter receptors caused by RNA editing have been found in postmortem studies of suicide victims [19, 20], patients with schizophrenia [21], patients with BD [22], and patients with psychotic disorders [20, 23]. RNA editing has also been associated with autism and postpartum psychosis [18]. Furthermore, prior findings have demonstrated the potential of RNA editing in identifying depression in postmortem sections [19] and in blood after interferon treatment [24].

A recent study showed that a blood-based signature comprising specific RNA editing biomarkers exhibits high accuracy in the biological detection of BD [25]. This signature encompasses biomarkers from eight genes implicated in various mood and inflammation regulation pathways, alongside individual data such as age, which is analysed by an artificial intelligence algorithm. This signature effectively distinguishes between BD and MDD within a sample of individuals experiencing moderate to severe depression (n = 245). The differential diagnosis of BD from MDD was replicated in an external independent clinical validation sample (n = 143), which showed high sensitivity (86.4%) and specificity (80.8%) (area under the receiver operating characteristic curve (AUC-ROC) 0.904) [25]. This biomarker signature, as well as two supplementary signatures showing high accuracy, have been selected to be validated in a new clinical sample.

This protocol outlines the rationale, methodology and expected outcomes of the EU-supported EDIT-B project (Grant number 220628/230125), which aims to identify among three previously tested and selected blood-based RNA epigenetic signatures the best signature in terms of external diagnostic validity (sensitivity, specificity, AUC, etc.) for the differential diagnosis of BD versus MDD in individuals with acute major depressive episodes.

Methods and analysis

Study design and setting

The present trial is a multicentre cross-sectional study including an adult sample diagnosed with bipolar disorder (BD) or major depressive disorder (MDD). Participants are recruited from four clinical sites: two in Barcelona, Spain; one in Paris, France; and one in Copenhagen, Denmark (Fig. 1). Different centres around Europe have been selected to include populations with different characteristics to ensure further applicability and generalizability of the expected results. This study has been registered on clinicaltrials.gov under the number NCT05603819 (Table 1). The SPIRIT reporting guidelines were used for reporting the study and the checklist has been included [26] (Table 2).

Fig. 1

EDIT-B study design. The EDIT-B study is a multicentre study recruiting participants in four different centres in Spain, France and Denmark, respectively. Participants are assessed for eligibility with several inclusion and exclusion criteria during the inclusion visit

Table 1 EDIT-B study summary

Table 2 Reporting checklist for the EDIT-B study protocol

Eligibility criteria

Eligible participants are between 18 and 80 years old diagnosed with BD or MDD and currently treated for a major depressive episode (MDE), both of which were verified using the Mini-International Neuropsychiatric Interview (MINI) [27] for the Diagnostic and Statistical Manual of Mental Disorders 5 (DSM-5). Both inpatients and outpatients are eligible for recruitment. The Montgomery–Åsberg Depression Rating Scale (MADRS) [28] total score should be equal to or greater than 20 at the time of assessment. The BD participants must have experienced at least one manic or hypomanic episode, while the MDD participants must have had at least one MDE. All participants must provide signed informed consent following oral and written study information.

The exclusion criteria for MDD participants are having first-degree family history of BD (e.g., parents, siblings or children), a total score on the Young Mania Rating Scale (YMRS) [29] equal to or greater than 12 at the time of assessment, pregnancy, BD or MDD secondary to major central nervous system affect or a diagnosis of schizoaffective disorder (Fig. 1).

EDIT-B assay test

The EDIT-B assay is a blood-based test system that combines sequencing data related to RNA editing biomarkers and individual data, including age, sex, tobacco consumption, alcohol abuse and treatment (antiepileptic, antipsychotic, anxiolytic, hypnotic/sedative and antidepressant data). The biomarkers originate from different genes: GAB2 (growth factor receptor bound protein 2-associated protein 2); IFNAR1 (interferon alpha/beta receptor 1); LYN (tyrosine-protein kinase Lyn); MDM2 (E3 ubiquitin-protein ligase Mdm2); PRKCB (protein kinase C beta type); IL17RA (interleukin 17 receptor A), PTPRC (protein tyrosine phosphatase receptor type C, also called CD45 antigen) and ZNF267 (zinc finger protein 267) [25]. These biomarkers have been described in previous work [25, 30] and have been selected using multiple criteria. First, an editome analysis was performed in a discovery cohort (n = 57 with 31 controls and 26 patients with depression) identifying several hundred potential targets. Then, stringent quality inclusion criteria (no location in intergenic regions or near SNPs; editing present in at least 25% of samples; median/mean coverage ≥30×; RNA Editing 0.8 ≥ Fold Change ≥ 1.20; p value < 0.05 and AUC ROC > 0.7) were applied following biological process analyses (with Gene Ontology and Reactome Pathways) and gene disease association analysis (DisGeNET) leading to a total of 8 genes. The biomarkers were recently validated in two independent clinical studies showing that they can be used to significantly differentiate patients with BD from patients with MDD [25].

As described in this previous work, an ensemble machine learning method, called ExtraTrees method—is derived from randomForest—was used for unipolar and bipolar depression differentiation, including RNA editing biomarkers, psychiatric treatment classes, and demographic variables (age, sex, substance use) as features. The ExtraTrees model learned effectively the complex relationships among these factors and their combined impact on diagnostic outcomes.

To execute EDIT-B assay test, blood samples are extracted using routine methods, and biological analyses are carried out by laboratories accredited for RNA sequencing (EN UN ISO 15189:2023). RNA is extracted (QIAsymphony, QIAGEN) and reverse transcribed (PrimeScript, TAKARA), and cDNA is subsequently amplified with specific primers (Q5 Hot Start High Fidelity enzyme, New England Biolabs), purified with magnetic beads (SPRIselect, Beckman Coulter) and indexed (Nextera XT, Illumina). After sample pooling and purification (SPRIselect, Beckman Coulter), the resulting library is sequenced (NextSeq 500/550 Mid-Output, Illumina). Detailed information about the biological and sequencing protocols used has been published elsewhere [25]. The EDIT-B assay is intended to assist physicians in diagnosing BD and differentiating between BD and MDD.

Enrolment

Eligible participants are invited to participate by psychiatrists or psychologists from the clinical centres. The information sheet is provided during the inclusion visit, and all participants’ questions regarding their study participation are answered by the professional. Before the interview, the participants must sign the written informed consent. If they are unable to act on their own behalf, a witness or guardian will sign the consent on their behalf. Once the consent is signed, the interview will be conducted, including: (a) sociodemographic variables (age, sex, ethnicity, race, civil status, coexistence, academic level, years of education, labour qualifications, current occupation); (b) clinical history (type of BD (I vs. II), duration of the illness, age of onset of the first depressive, manic or hypomanic episode, number of depressive, manic, hypomanic or mixed episodes); (c) personal psychiatric history (history of psychotic symptoms, delusions, hallucinations, catatonia, seasonal pattern, rapid cycling, melancholy, atypical symptoms, psychotic depression, post-partum depression); (d) family psychiatric history (any history of affective disorders, history of suicide, and any other mental disorders for relatives of first, second, and/or third degree of consanguinity); (e) current and previous interventions received (e.g., medical procedures, psychological therapy, alternative treatments); (f) current toxic consumption (e.g., tobacco, nicotine, caffeine, cannabis); and (g) current medication for the diagnosed disorder and any other concomitant medications, start date, dose and frequency are registered. Moreover, MADRS [28], YMRS [29], MINI scale [27], and the European Quality of Life 5 Dimensions Questionnaire (EQ-5D) [31] (before and after diagnosis) will be completed. The first participant was recruited in the summer of 2022, and the last site initiation occurred at the beginning of 2023. The enrolment has been completed up to 80%. We plan to complete the study at the end of 2024.

All clinical and biological data is collected during a unique visit. Thus, participation ends after completing the interview and blood sample collection. Blood samples are collected in two PAXgene^® RNA collection tubes (2.5 mL each), shipped to a central laboratory and stored at −80 °C until analysis. Sample analysis will be performed in batches. The extracted RNA from the blood and remaining blood samples will be stored at −80 °C for 5 years and may be used for additional studies if the patient agrees to participate in the consent form. Otherwise, the samples will be destroyed. Neither global nor individual results of the signatures will be communicated to the investigators before the end of the study. Figure 2 summarizes the participant flow from inclusion with blood sampling until the final analysis.

Fig. 2

Summary of the participant flow in the EDIT-B study. (1) The clinician includes the participant in EDIT-B study and evaluate the diagnosis; (2) After inclusion, a blood sample is taken from the participant at the respective centre; (3) Blood sample is sent to the central laboratory that performs the biological analysis; (4) Data analysis is done using EDIT-B signature before clinical and EDIT-B results are compared for final analysis

The sample size was fixed to provide sufficient precision (±5%) in the estimates of the diagnostic performance of the EDIT-B test in terms of sensitivity, specificity, accuracy and area under the receiver operating characteristic curve (AUC-ROC). Based on the literature [25], the estimated values of sensitivity, specificity and AUC-ROC are expected to be close to 86%, 80%, and 90%, respectively. A total sample size of 436 participants with MDE with a 1:1: ratio of MDD (n = 218) vs. BD (n = 218) will be included. A maximal 10% attrition rate was considered. The Buderer and Jones methods [32, 33] were used for these calculations.

Patient and public involvement

Patients were not involved when designing the study, but the study investigators actively participated in the study design and setting of the research question. We carefully assessed the burden of the clinical study intervention on participants. Before and throughout the study, the general public was informed in press releases about the study advancement. At the end of the study, the study investigators will be informed of the global and individual results of their participants. Study participants will be informed of both upon their request to their investigator. Study results are intended to be disseminated in scientific articles, at conferences and to the general public through press releases and social media. For dissemination plans and methods, patients are not involved directly, but dissemination is discussed with the investigators of the study and partially performed by them.

Outcomes

The primary aim of this study is to identify among three RNA biomarker signatures the best in terms of external validation of the diagnostic performance (sensitivity, specificity, accuracy and AUC-ROC) of BD versus MDD differentiation in participants with current MDE. The endpoint is the categorization of the patient according to the signatures as patient with MDD or BD compared to the diagnosis made by the treating psychiatrist using the MINI. Secondary analyses will be performed depending on the available information and statistical power of the collected data.

Data management plan

The data is collected via paper and later translated into an electronic CRF (e-CRF) by the investigator and/or study coordinator and the online operation of the data management checks. A clinical research assistant (CRA) monitors and validates the data continuously during the study period. During the recording process, the investigator or the study coordinator will also respond to online controls in case of inconsistencies or to queries raised by the e-CRF data manager/CRAs. At the end of the study, the investigator must sign the e-CRFs. Four validation batches are planned to check for discrepancies in the database. Validation batches were conducted at the beginning of the study.

Data checks will be performed to identify any errors, inconsistencies or missing data within the e-CRF by predefined computerized and manual checks. All systematic edit checks (manual, automatic, reconciliations) will be applied to the database to detect any discrepancies. The data manager reviews each discrepancy to validate the relevance of the issue and possibly raises a query. Manual queries can be generated according to the data manager evaluation, upon medical coder request, or following review of the medical listings.

After the recruitment of the last participant and after all queries have been solved, the data manager will proceed to the database lock of the e-CRF. A final extraction will be performed, and the final datasets will be produced and locked with a database lock certificate. A second database, consisting of the EDIT-B result (BD profile/MDD profile), will be locked when the last participant is analysed. Unlock and relock databases can be constructed by the data manager at the study coordinator’s request; in the case of unlock, the data manager provides a list of data modifications made on the unlocked database.

Only personal data needed for the objective of the study will be collected. To protect personal data, participants’ data will be pseudonymized at the research sites. Only the principal investigator and study coordinator at each site have access to the encryption table. During and after the study, appropriate technical and organizational measures will be implemented to ensure the processing of personal data in accordance with the European General Data Protection Regulation. Pseudonymized participants’ data will be archived for a maximum of 25 years after the end of the study on health data hosting servers by Euris and located in France. The final study results will be archived, and any publication will use fully anonymized data.

Data analysis plan

Data from all included participants who provided signed informed consent and who did not withdraw from the study will be analysed. Descriptive analysis of the samples will be performed.

The clinical validation performance of the three signatures will be evaluated using the following matrices: AUC-ROC, sensitivity, and specificity. Additionally, the positive predictive value, negative predictive value, positive and negative likelihood ratios, and odds ratios will be calculated. The categorization of disorder type (BD vs. MDD) by experts relies on the MINI, which is considered the gold standard for diagnosis.

To determine the best signature among the three, the performances of the tested signatures will be compared on a two-by-two basis by assessing sensitivity and specificity utilizing methods based on the McNemar test, as proposed by Hawass [34]. Furthermore, AUC-ROC values of the three signatures will be compared using the DeLong method, which is akin to accuracy assessment through the no informative rate test [34, 35].

ExtraTrees models (as used for EDIT-B assay) are capable to handle complex and non-linear interactions frequently encountered in clinical studies. Additionally, their capacity to assess feature importance enhances transparency by providing clinicians with insights into variables influencing the most significantly predictions. In this study, the importance of features will be evaluated using randomForestExplainer package [36]. This tool makes it possible to explore deeply various importance metrics of the artificial intelligence model, such as total number of trees and nodes, mean minimal depth, node impurity and frequency as root variables. To further dissect interactions, relationships between RNA editing biomarkers and psychiatric treatment ATC classes (e.g., antipsychotics, antidepressants, and antiepileptics) will be investigated. Specifically, the occurrence of splits involving RNA editing biomarkers within maximal subtrees corresponding to each treatment class will be examined. This analysis will provide deeper insights into how these treatment categories influence RNA editing biomarkers and contribute to the model’s decision process. Finally, conditional minimal depth will be calculated, a metric that quantifies the strength of interactions between features, to rank and evaluate the interaction dynamics between RNA editing biomarkers and other variables. This systematic approach will shed light on the interplay of treatment effects and RNA editing biomarkers within the predictive framework, enhancing the understanding of their combined diagnostic power.

Considering the nature of the study, all calculations will be conducted on valid data without missing value imputation. However, the percentage of participants for whom the test has been unfeasible will be calculated since it is also an additional important parameter of test performance under real-life conditions.

The analysis will be conducted using SAS version Viya 3.05.

Ethics and dissemination

This study is conducted in accordance with the ethical principles set out in the latest version of the World Medical Association’s Declaration of Helsinki 2013, attending to all the nuances involved in precision psychiatry research [37] and following the requirements of good clinical practice.

Prior to starting the study, approval was obtained from several research ethics committees, as the study is performed in different countries. For French approval, the documents were reviewed, and the study was approved under ID 2021-A-03194-37 by the research ethics committee Ile-de-France X. In Spain, the research ethics committees of the Hospital Clinic of Barcelona and the Sant Joan de Déu Foundation reviewed and approved the study under ID HCB/2022/0042 and PS-05-22, respectively. For Danish approval, the documents were addressed to De Videnskabsetiske Komiteer for Region Hovedstaden (the Scientific Ethics Committees for the Capital Region of Denmark) and approved under project ID H-22010899. Any amendments to the study will be resubmitted to the ethics committee.

A copy of the consent form, signed and dated by the participant and the principal investigator or the physician representing the investigator, will be given to the participant prior to the screening visit. One copy will be retained at the clinical site. At the end of the study, another copy will be sent, in a temper-proof sealed envelope, to the coordinator of the study. In addition, the investigator will specify in the research participants’ medical files all the procedures conducted for the study.

A publication with the final study results will be disseminated in an international peer-reviewed journal and presented at international conferences.

Discussion

The current diagnostic assessment of bipolar disorder (BD) includes a subjective component that complicates timely diagnosis. General practitioners are the first to see patients who present with a major depressive episode (MDE), and they find it challenging to suspect a diagnosis of BD and therefore to refer patients to a specialist. The training and expertise of psychiatrists are essential for accurate evaluation. However, trials for the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) demonstrated an intraclass kappa value of 0.28 [38]. This represents a minimal agreement and means that, under the study conditions, highly trained psychiatrists agreed on the diagnosis between 4 and 15% of the time. Healthcare systems often lack the resources and tools to ensure responsiveness across all levels of care. Moreover, the existence of various subtypes of the disorder, each exhibiting distinct dysfunctions, further complicates the process of achieving a precise diagnosis. In this context, standardized screening tools become crucial to complement clinical evaluations. For instance, the Mood Disorder Questionnaire (MDQ) has demonstrated a sensitivity slightly above 70% in outpatient psychiatric clinics [39], as well as good accuracy for screening BD in primary care [40]. However, some studies indicate that in patients with bipolar spectrum illness, the sensitivity reaches 70% only for BD type I [41]. Meanwhile, scales such as the 32-item Hypomania symptom check-list, first revision (HCI-32-R1), have shown utility in detecting hypomanic symptoms [42], which can further aid in identifying at-risk patients. Although diagnostic scales are useful tools to support screening and diagnosis, their results are often highly heterogeneous and frequently based on small sample sizes. Therefore, their use should be approached with an awareness of their limitations. On the other hand, the diversity inherent in BD diagnoses, as classified by existing systems, may have impeded progress in psychiatric treatment, despite contributing to diagnostic consistency in the field. Furthermore, once patients are diagnosed, few receive evidence-based care, and those who do receive such care have a low probability of recovery [43].

Studies suggest that biomarkers could aid in diagnosing BD. Metabolomics biomarker signatures may improve diagnostic accuracy [44], and biochemical biomarkers combined with decision trees might differentiate BD from MDD [45]. However, further validation or improvement of their performance is necessary. Thus, there remains a need to identify and validate reliable biomarkers with strong performance that can effectively distinguish between BD and MDD [13, 46].

Advances in biomarker development hold promise for revolutionizing treatment guidance and identifying new targets for early intervention and recurrent episode prevention [47]. Recognizing early signs of recurrence may be indicative of noncompliance [48], warning healthcare providers and enlightening patients. The ongoing progress in biomarker measurement technologies represents an innovative leap that could forecast and prevent adverse clinical outcomes by furnishing real-time updates on the patient’s current clinical condition, along with predictive insights suggestive of future relapses, ultimately enhancing patient well-being. Furthermore, it will empower patients to supervise their own health conditions, aiding them in making well-informed choices regarding treatment alternatives. The ultimate objective is to enhance patient outcomes through accessible technology.

Groups of patients with different BD subtypes or dysfunctions may show divergent responses to treatment [49, 50]. This diversity suggests several underlying causes and physiological processes linked to the same disorder, which could result in varying illness courses and treatment outcomes even among individuals with identical diagnoses [51]. Therefore, reducing the significant variability in psychiatric conditions through a deeper comprehension of their biological basis might enable progress in treatment by identifying accurate biomarkers for diagnosis, prognosis, and predicting treatment responses. On the other side, psychiatric disorders remain a complex interplay of biological, psychological and social factors, needing a global evaluation that considers all dimensions.

Future approaches should therefore combine clinical evaluation with biomarkers permitting to establish a diagnosis that is based on different criteria. This helps to find a balance between variation and one-sided input that could lead to misdiagnosis. Introducing biomarkers into clinical routine is crucial keeping them as a complementary diagnostic tool and not as a stand-alone method. This opens a new era advancing precision psychiatry and improving patient outcomes.

Longitudinal studies should be conducted to examine the dynamics of fluctuating episodes, particularly in BD patients, and to assess biomarkers over time, possibly establishing specific biomarker patterns for diseases onset, relapse, or treatment responses. This might help to understand the progression of the diseases.

BD and MDD are heterogeneous disorders, encompassing bipolar disorder types I and II, melancholic depression psychotic depression and various comorbidities such as personality disorders and Attention-Deficit/Hyperactivity Disorder (ADHD). To include populations with diverse characteristics, aiming to ensure applicability and generalizability of the expected results, the study employs only a few exclusion criteria. Consequently, the findings will reflect the biological variability present in real-life heterogenous populations. However, this extensive heterogeneity in the study population may reduce the discriminatory power of the biomarker analyses and their results.

Risk and benefit

Participants benefit from a thorough diagnostic evaluation, including a MINI interview, by participating in the study. Since participants’ diagnoses are established according to the highest clinical standards before any procedures related to the clinical study, the foreseeable risk associated with participating in the research is negligible. Moreover, sample extraction is akin to a routine blood test, and no adverse effects are anticipated.

From a society perspective, the EDIT-B clinical study will contribute to the development of innovative solutions that lead to earlier and personalized management of patients with mood disorders, including BD and MDD, by improving the diagnostic accuracy and differential diagnosis of BD. Based on a blood test complementary to clinical evaluation, the treatment will be more effective, and the risk of antidepressant-induced mania [52] may be diminished. This may also improve the definition of mixed states with MDE ⁵³, reducing hospitalizations and costs. Given the urgent need for accurate diagnostic methods for the earlier management of patients, this study will contribute to the development of innovative solutions for timely diagnosis and appropriate treatment, reducing avoidable side effects and hospitalization rates and completely improving patient outcomes. Moreover, the use of EDIT-B will reduce the economic burden and optimize the use of medical and financial resources.