Improving early recognition and intervention in people at increased risk for the development of bipolar disorder: study protocol of a prospective-longitudinal, naturalistic cohort study (Early-BipoLife)

Bipolar disorder (BD) is a severe and disabling mental disorder, characterized by an early onset and recurrent, severe episodes or a chronic course with poor psychosocial functioning in a proportion of patients (Ferrari et al. 2016; Fagiolini et al. 2013). Although efficacious psychopharmacological and psychosocial treatments for manifest BD exist (da Silva Lima et al. 2017; Pfennig et al. 2012a), BD is often diagnosed and treated with a significant delay (up to 10 years on average) (Baldessarini et al. 2003; Pfennig et al. 2011; Dagani et al. 2017). Treatment delay is associated with a worse functional outcome, an elevated risk of suicide (Chen and Dilsaver 1996; Miller et al. 2014; Post et al. 2010) and an inferior response to mood stabilizing drug treatment (Kessing et al. 2014). Improved diagnostic instruments for early recognition and guidelines for early-targeted intervention have the potential to enhance overall disease management or even to prevent conversion to manifest BD (Lish et al. 1994).

Recent evidence from early recognition centers has shown that help-seeking persons at-risk for BD are often already affected by a substantial impairing subsyndromal symptomatology (Leopold et al. 2013a, 2014; Pfennig et al. 2012b). Given the low specificity of individual precursors and early symptoms of BD prior to the first manifest manic episode (Berk et al. 2007), comprehensive criteria defining high-risk profiles need to be assembled. Due to its high heritability (Craddock and Sklar 2009), a positive family history for BD is one of the major risk factors for BD (Duffy et al. 2010, 2014; Smoller and Finn 2003; Mendlewicz and Rainer 1977). Moreover, evidence from prospective clinical and observational studies indicates that depressive and subthreshold (hypo-)manic symptoms may be antecedents or rather first noticeable symptoms of BD (Beesdo et al. 2009; Duffy et al. 2014, 2017; Bechdolf et al. 2012; Mesman et al. 2013; Ryles et al. 2017).

There is a vivid discussion about the association between attention-deficit/hyperactivity disorder (ADHD) and BD. At least a subgroup of patients with ADHD may be at risk to develop BD (Leopold et al. 2012; Faedda et al. 2014; Wang et al. 2016). Additionally, the following factors are discussed as further potential risk factors for BD: (a) a history of critical/stressful life events (Garno et al. 2005; Kessing et al. 2004), (b) childhood anxiety disorders or being anxious/worried/fearful, hyper-alert or sensitive (Lenox et al. 2002; Egeland et al. 2003; Tillman et al. 2003), (c) mood swings and impaired emotional regulation (frequent “ups and downs”) (Thompson et al. 2003), (d) changes in sleep and circadian rhythm (Lenox et al. 2002), (e) somatic complaints or medical/physical problems (Egeland et al. 2003), (f) substance (ab-)use or dependence (Rush 2003), and (g) particular personality, temperament and character traits (e.g., extraversion, novelty seeking, creativity, high reward responsiveness and ambitious goal-striving, cyclothymic/hyperthymic temperament) (Duffy et al. 2010; Mesman et al. 2013; Leopold et al. 2012; Pfennig et al. 2017; Egeland et al. 2000; Correll et al. 2014a; Alloy et al. 2012; Kwapil et al. 2000; Blechert and Meyer 2005).

Based on this knowledge, structured instruments have been developed by different research groups to identify persons at risk for BD (Leopold et al. 2012; Bechdolf et al. 2014; Correll et al. 2014b) and to assess initial subsyndromal symptomatology (Bipolar Prodrome Symptom Scale, Correll et al. 2014b), EPIbipolar, Leopold et al. 2012), Bipolar at-risk criteria (Bechdolf et al. 2012)/extended BAR criteria (Fusar-Poli et al. 2018). First prospective studies revealed promising results on the psychometric properties of single instruments and the predictive validity regarding the conversion to BD (Bechdolf et al. 2014; Correll et al. 2014b; Fusar-Poli et al. 2018). Bechdolf and colleagues showed a conversion rate to BD of 14.3% within 12 months in people presenting with BAR criteria (Bechdolf et al. 2014), and Birmaher and colleagues found a conversion rate of 25% to BD in children and adolescents with clinically relevant bipolar symptoms that did not fulfil the DSM-IV criteria for bipolar-I or bipolar-II disorder (Birmaher et al. 2006). Conversion rates of patients with unipolar depression range in different studies from 4 to 49% (Geller et al. 2001; Gilman et al. 2012). However, previous samples were small, and the comparison of study results is hampered by heterogeneous risk definition/description and differences in study methodology (e.g., diverse follow-up duration). Standardization of the diagnostic process across specialized early recognition facilities is not formed yet. Moreover, at present protective/resilience factors (Stange et al. 2013) and potential biomarkers (Duffy et al. 2014; Ritter et al. 2016; Kapczinski et al. 2009) are not part of the existing assessment tools (Leopold et al. 2012; Bechdolf et al. 2014; Correll et al. 2014b).

Besides the challenges of early and appropriate diagnostics, prevention as well as early-targeted intervention of BD is crucial, because treatment delay is associated with a worse social adjustment, an increased risk for suicide, increased comorbidity rates, more hospitalizations, and an impaired age-appropriate development (Goldberg and Ernst 2002; Matza et al. 2005). Pharmacological treatment approaches with mood stabilizing agents (lithium and divalproex) in at-risk patients have only been investigated in underpowered studies not showing effectiveness (Findling et al. 2007; Geller et al. 1998). Regarding monotherapy with antidepressants, study data in bipolar at-risk patients are lacking, and there might be a risk for inducing mania. However, in case of depressive, anxiety and/or obsessive symptomatology, antidepressant treatment under close monitoring is in line with guideline recommendations for both, unipolar (DGPPN 2015) and bipolar (DGBS and DGPPN 2019) disorders. A recent review of psychotherapeutic interventions in young at-risk patients (Pfennig et al. 2017) showed promising results. Three studies on the efficacy of early family-focused approaches suggest favorable outcomes such as improved symptoms, a longer duration in remission and a better psychosocial functioning (Miklowitz et al. 2011, 2013; Nadkarni and Fristad 2010). A updated systematic review (Blum 2018) provided evidence for the efficacy of Interpersonal and social rhythm therapy (IPSRT, Goldstein et al. 2018) and Mindfulness-based cognitive therapy (MBCT, Cotton et al. 2016) with improvement in sleep patterns, emotion regulation and a decrease of anxiety, compared to baseline.

Based on the knowledge of risk factors and early clinical features of BD and of the promising research on early targeted interventions, pilot clinical category models and treatment guidance (Berk et al. 2007; Kapczinski et al. 2009; Leopold et al. 2013b; Schneck et al. 2017) have been formulated. Primary treatment recommendations for early (at-risk) stages of BD are psychoeducation, preventive strategies to halt conversion to BD, and strategies on preserving the young person’s ability to meet age-appropriate developmental tasks (Berk et al. 2007; Kapczinski et al. 2009; Leopold et al. 2013b). At later stages, the focus changes to symptomatic treatment, and the establishment of adherence and relapse prevention of BD (Berk et al. 2007; Kapczinski et al. 2009; Leopold et al. 2013b). The verification of the suggested models and guidance could enable evidence-based treatment approaches in early stages of BD. A validated model would provide greater utility for testing the efficacy, cost-effectiveness, risk–benefit ratios and feasibility of available interventions (McGorry et al. 2006).

In summary, early recognition of at-risk states for BD is an important clinical field, but the diagnostic process across specialized early recognition facilities has not been standardized yet. First approaches of clinical categorization and treatment guidance have been formulated (Berk et al. 2007; Kapczinski et al. 2009; Leopold et al. 2013b), but no common diagnostic and/or treatment guidelines are available supporting the individual decision-making processes for those in the early stages of BD. This situation impedes the accumulation of superior clinical-epidemiological knowledge regarding the core predictors for conversion to manifest BD and prevention of incident BD.

To address these areas of unmet needs, a prospective, naturalistic cohort study in the age group of 15–35 years was designed to clarify the following research questions, with (a) being the primary question:

The Improving early recognition and intervention in people at-risk for development of bipolar disorder (Early-BipoLife) study is a multicentre study conducted by a research consortium of ten university and teaching hospitals across Germany with early recognition centres/facilities and specialised in- and outpatient care. Early-BipoLife is funded by the Federal Ministry of Education and Research (BMBF, Grant number: 01EE1404A) and is part of the BipoLife consortium described elsewhere (Ritter et al. 2016; Mühlbauer et al. 2018). The study is conducted according to good clinical practice (GCP) standards and has been approved by the responsible Ethics Committee of the Medical Faculty of the Technische Universität Dresden (No: EK290082014) and all local ethics committees. All participants provided written informed consent after comprehensive information about study aims and procedures.

Early-BipoLife is a naturalistic prospective-longitudinal observational cohort study of 1419 participants who are repeatedly assessed during their most vulnerable and formative years (Leopold et al. 2014; Pfennig et al. 2012b; Lish et al. 1994). This comprehensive design has been developed in order (a) to determine the natural course of potential risk and resilience factors and early symptoms/precursors of BD; (b) to observe the change in severity of early symptomatology of BD as primary outcome (including conversion to manifest BD); (c) to evaluate the predictive power of potential risk factors/precursors of BD and biomarkers as well as psychometric properties of recently introduced instruments for the early identification of persons at-risk for BD; and (d) to investigate currently recommended treatment advices in daily care and compare these with recently proposed clinical categorization and treatment guidance (Berk et al. 2007; Leopold et al. 2012; Kapczinski et al. 2009; Schneck et al. 2017).

To the best of our knowledge, this is the first investigation of a help-seeking cohort of at-risk persons for BD sampled from early recognition centres/facilities as well as in-/outpatient settings. Other research on risk factors for BD mainly focused on the investigation of offspring of persons with manifest BD (Duffy et al. 2007, 2009). For a discussion on differences of studies regarding samples, focus, outcomes and statistical approaches see (Geoffroy and Scott 2017).

Major strengths of the Early-BipoLife study are:

The study has the following limitations. First, this is a naturalistic follow-up study. As such, we will not be able to test the effectiveness of individual components of the treatments that were recommended and delivered. We consider this a second-generation question for further efficacy studies and wish to focus on more generalizable outcomes to study risk and protective factors for BD without the restrictions of an efficacy design and restricted treatment approaches. Second, adherence to the prescribed treatments is not assessed. The study sample includes patients with depression, ADHD, as well as subjects at increased risk for the development of bipolar disorder. In some of these, psychotropic medication is applied out of which the risk for mania might arise. Individual medication is assessed at baseline and FU and will be attended to in the analysis as effectively as possible. Third, we include mostly university centres and hospitals with specialized BD services. While this approach reduces the generalizability of the findings, it provides a yardstick against which other more usual care settings can be compared. Fourth, the study is conducted in Germany and under German health care conditions, with less generalizability of the pathways to care and treatment approaches as well as related illness trajectories in systems with less access to care. Fifth, not all patients enrolling in the Early-BipoLife study will have a complete neuroimaging, electrophysiology and genetics battery. However, given the large number of expected participants, the generated data will be highly informative. We acquire data on somatic diseases and BMI, but no further information on the subject`s physical health. Sixth, the study sample includes patients with depression, ADHD, as well as subjects at increased risk for the development of bipolar disorder. In some of these, psychotropic medication is applied out of which the risk for mania might arise. Individual medication is assessed at baseline and FU and will be attended to in the analysis as effectively as possible. Seventh, with applying a comprehensive FU assessment, especially hypomanic episodes could be detected that otherwise had been overlooked, or had not been classified as such with coarser questioning. Eighth, the control sample is representative for the young population in Dresden and Berlin, was, however, recruited under different conditions as the risk group subjects, which might affect comparability.

Nevertheless, despite these limitations, the Early-BipoLife study will be one of the largest prospective studies to comprehensively characterize a sample of adolescents and young adults at risk for the development of BD. Results are expected to inform patients, families, clinicians, guideline developers, payers and policy makers alike.

Long-term follow-up beyond the visits predefined in the study is planned to increase confidence in the evaluation of the proposed risk and protective factors and lead to a deeper understanding of early stages of BD. At present, all study centers follow patients up to month 36, an extension study will be applied for at a public sponsor.

The Early-BipoLife study will provide unprecedented and detailed information about the relationship between recently discussed risk and protective factors and the onset and natural course of BD. Findings expected from the Early-BipoLife study will significantly contribute novel insights into pathomechanisms of disease and beneficial treatment algorithms. This information is crucial to further develop early-targeted primary and secondary prevention and intervention strategies to reduce the risk for conversion to BD and improve outcomes when BD has developed. A validated clinical categorization and treatment guidance/staging model will be able to inform patients, clinicians and researchers additionally about the prognosis and treatment response. It is also of particular interest, which assessment tool(s) or which parts of those will perform best to allow a valid categorization and staging of BD risk.