Staging Models in Bipolar Disorder

Bipolar disorder (BD) affects about 2% of the world’s population, with subthreshold forms affecting another 2% (1). Treatment in BD conventionally focuses on acute stabilization of the mood episode and on maintenance, in which the goals are relapse prevention and reduction of subthreshold symptoms (2). Even with syndromal treatment, about 37% of patients relapse into depression or mania within 1 year and 60% within 2 years (3). The rates of completed suicide in BD patients are 7.8% in men and 4.9% in women (4).

Staging models have become an important contribution to understanding disease progression, functional outcome, and treatment response in medicine (5). Physicians and researchers from different specialties rely on staging models to improve available treatments and to generate novel interventions for medical conditions such as heart failure (6), sepsis (7), chronic obstructive pulmonary disease (8), and cancer (9). Staging models in psychiatry assume that some mental illnesses may progress from an at-risk stage to a treatment-refractory end stage (10). Moreover, staging systems also emphasize that while some patients may have a more benign course of illness, others may present with a higher propensity for episode recurrence and refractoriness (11, 12). Table 1 shows staging models proposed for bipolar disorder.

Accordingly, recent studies suggest that staging may help tailor treatment in a subset of BD patients (13). In view of that, we aim to review the literature related to staging models in BD and their potential clinical use.

BD patients may present different courses of illness progression (11). Trauma exposure, increased number of episodes, and comorbidity are commonly associated with unfavorable clinical outcomes, such as reduced interepisode intervals, rapid cycling, and cognitive and functional impairments, as well as augmented rates of hospitalizations and suicide (14).

Cross-sectional and longitudinal studies assessed the impact of trauma exposure on BD severity. A 24-month follow-up study of 131 bipolar I disorder patients has shown that trauma exposure predicted greater severity of interpersonal chronic stressors (15). In addition, among 587 BD patients, consistent associations between childhood trauma (emotional abuse, sexual abuse, and emotional neglect) and more severe clinical outcomes (suicide attempts, rapid cycling, and an increased number of depressive episodes) have been reported (16). A history of verbal abuse in 634 BD outpatients was related to an earlier age of onset of disease and other poor prognosis characteristics, including rapid cycling, and a deteriorating illness course as reflected in ratings of increasing frequency or severity of mania and depression (17).

Increased number of episodes may also have a predictive effect on the risk of recurrence in BD, since vulnerability and latency to relapse vary directly as a function of number of prior hospital admissions (18). In the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD), patients with multiple previous episodes presented with worse functioning and lower quality of life. In addition, in the STEP-BD study, patients with multiple episodes had more disability, as well as more chronic and severe symptoms (19). Moreover, evidence suggests that younger patients with fewer episodes present a better response to treatment with conventional mood stabilizers and psychoeducation (20). Of note, increasing episode number is linked to a reduction in the likelihood of response to lithium (21) and cognitive behavioral therapy (CBT) (22).

Regarding co-occurring disorders, a systematic review with meta-regression has reported a significant correlation of comorbid substance use disorders and suicide attempt (23). Within this same meta-analysis, substance abuse disorder in the preceding 8-week period predicted greater likelihood of suicide attempt in a 5-year follow-up prospective study with 413 youths with BD (24). Current or past substance use disorders have also been associated with greater risk of switch into manic, mixed, or hypomanic states (25). Specifically, cannabis and alcohol use disorders were associated with higher symptomatology and lower functioning in a longitudinal study of 137 BD patients (26). Moreover, BD patients with attention deficit hyperactivity disorder (ADHD) presented an earlier onset and had a chronic course, irritable mood, and greater clinical and functional impairment, and treatment resistance (27). In a meta-regression analysis, comorbid ADHD was significantly associated with suicide attempts (23). Comorbid PTSD also was associated with poor outcomes, including significantly worse social functioning (28), interepisode depression, and quality of life (29). In addition, there is a body of literature suggesting that predictors of poor prognosis (stressors, increased episode number, and comorbidity) may show sensitization to themselves and cross-sensitization to the others, contributing to greater illness burden and treatment resistance (14).

The term neuroprogression is used to describe the biological underpinnings of illness progression in the context of BD (11) and may shed some light on how such predictors interact with the pathophysiology of BD (30). Such a process can further disrupt the brain circuits that are responsible for mood regulation and cognition, increasing the vulnerability to illness (30). It is now becoming apparent that the biochemical foundation of neuroprogression is multifactorial and interactive, not only between pathways, but via stress sensitization from the environment stressors and comorbidities (11). As shown in recent meta-analyses and postmortem studies, the core components of this biochemical process of neuroprogression are abnormal levels of inflammatory cytokines (31–33), markers of oxidative stress (34, 35), and neurotrophins, including brain-derived neurotrophic factor (BDNF) (36, 37). There seems to be a significant reduction in BDNF expression and IL-6 levels along with significantly higher levels of tumor necrosis factor–alpha (TNF–α) as patients progress to more severe stages (38). In the same vein, differential changes in oxidative stress activity levels have also been reported as patients progress to later stages of BD (39). A recent PET scan study showed microglial activation in the right hippocampus of BD patients (40). Moreover, there is evidence for the involvement of epigenetic changes, particularly histone and DNA methylation (41, 42) and acetylation (43) leading to long-acting effects on gene expression, which may contribute to neuroprogression.

Noteworthy is the severity of comorbidity between PTSD and BD, since reduced BDNF function from several contributing sources, including the met variant of the BDNF val66met (rs6265) single-nucleotide polymorphism, trauma-induced epigenetic regulation, and current stress, is associated with core characteristics of both disorders (44). Trauma exposure is also associated with significantly lower BDNF levels in BD patients (45).

Structural neuroimaging studies have reported an association between the number of episodes and the enlargement of the lateral ventricles (46, 47) as well as decreases in cerebellar vermal volume (48). While changes during the first episode of mania seem to be related specifically to white matter pathology (49, 50), progressive loss of gray matter volume in prefrontal areas (51, 52) has been reported in late-stage BD (53, 54). For instance, in contrast to findings in schizophrenia, which have shown that hippocampal volume loss and ventricular dilatation may occur prior to the first episode, the literature suggests that in BD, gross brain structure is relatively preserved during its early phases (53). BD patients have shown hippocampal subfield volume reductions in cornu ammonis (CA) subfields CA2/3, CA4/dentate gyrus, subiculum, and right CA1 (55). Moreover, lithium treatment has been shown to prevent such reduction (56, 57).

Psychosocial functioning describes a person’s ability to function socially and occupationally and to live independently (58). Patients with BD may suffer from functional impairment, even when euthymic, and may experience serious dysfunction in distinct life domains, such as work productivity, social activities, and autonomy (59). A 24-month follow-up study found that almost 100% of first-episode manic patients with BD had syndromal recovery within 2 years, but only one-third achieved functional recovery (60). Therefore, functional and symptomatic recoveries are not always associated and may need different therapeutic approaches.

The ability to function in daily-life activities has also been proposed as an important correlate of staging and illness progression in BD (61, 62) and may provide clinicians with a proxy of staging (63). By applying latent class analysis in a sample of 106 remitted adults with BD, a recent study identified two subtypes of patients presenting “good” and “poor” functional outcome (61). Episode density, level of residual depressive symptoms, estimated verbal intelligence, and inhibitory control emerged as the most significant predictors of such subtype membership (61). An interesting point of this study is that functional outcome was not predicted by illness duration, since the two groups were comparable in age, age of onset, and illness duration (61). Prior studies have also supported number of hospitalizations and comorbidity as determinants of poor functional outcome (26, 64, 65).

BD patients may suffer from marked cognitive impairment even when euthymic (66, 67), which could also limit long-term psychosocial functioning (68, 69). Beyond estimated verbal intelligence and inhibitory control, executive and memory dysfunctions tend to show greater impairment in daily-life activities (70, 71). In comparisons of neurocognitive performance between patients with low and high functioning scores, patients with poor functioning had significantly more severe impairment in memory tasks, inhibitory control, and working memory (72).

Functional staging in BD may provide a practical model to evaluate the progressive course of illness, as well as guide therapy to improve patient’s quality of life, the ultimate goal of treatment. In this regard, a strong linear association was found between functioning assessment short test scores and the clinical stages described by Kapczinski (73), suggesting a progressive functional decline from stage I through stage IV of BD (63). Although significant differences in functional status were found between patients in all stages, only patients in severe stages (III and IV) were more impaired than healthy subjects (63).

It has been widely accepted that earlier interventions in BD are likely to be associated with a better response to treatment and lower rates of both functional and cognitive impairment (63, 73). Conversely, interventions for late-stage BD patients usually require more complex treatment regimens (73, 74). While treatments such as lithium monotherapy, CBT, and psychoeducation seem to be better suited to prevent further recurrences in early-stage patients, complex pharmacological strategies along with functional and cognitive remediation may be considered optimal to alleviate functional changes and prevent further deterioration in late-stage patients.

Clinical staging as a means to guide and tailor treatments according to patient’s needs may provide an important tool for clinicians working with BD patients. However, the potential value of staging systems with regard to their specificity and validity remains to be clarified by longitudinal studies. In the meantime, targeting treatment for each BD patient from a staging perspective may provide clinicians with a heuristic model as to how to approach BD patients according to their individual needs.