Abstract
Background
Bipolar disorder (BD) is a chronic episodic illness, with patients spending about two-fifths of their lifetime in illness episodes. 1 Further, BD is associated with increased disability, 2 elevated suicide risk, 3 social exclusion, 4 high rates of medical comorbidity and substance use, 5 reduced quality of life, and high economic costs. 6 Hence, it is vital to reduce the time spent in episodes. Achieving this goal requires a good understanding of the key principles of BD management.
The treatment of BD is typically divided into three phases: acute, continuation, and maintenance phases. In comparison to other psychiatric disorders, BD management involves a broader range of pharmacologic treatments, including mood stabilizers, antipsychotics, antidepressants, and benzodiazepines, with polypharmacy being recommended by many clinical guidelines. 2 Additional treatment modalities for managing BD include light therapy, sleep deprivation, dark therapy, and brain stimulation (including electroconvulsive therapy, transcranial direct current stimulation, and repetitive transcranial magnetic stimulation). Finally, evidence-based psychosocial interventions for managing BD include interpersonal and social rhythm therapy, cognitive behavior therapy, and family therapy. 7
The maintenance treatment of BD typically involves long-term use of psychotropics, often lifelong, with the goal of preventing further episodes; available treatment guidelines widely support this approach. 8 Given that patients with BD may need to take medications for more than 50 years (assuming age of onset at 20 years and a lifespan of 70 years), and given the long-term risks associated with both illness and medication usage, it is important to critically evaluate the evidence for and against lifelong prophylaxis. In this article, we discuss the pros and cons of recommending lifelong pharmacoprophylaxis for BD. We primarily focus on the efficacy and effectiveness of lithium as it has the longest history and largest evidence; however, where relevant, we also discuss insights from the literature on prophylaxis with valproate and antipsychotics.
Discussion
Arguments Against Long-Term Pharmacoprophylaxis
Historically, the concept of “prophylactic” lithium originated from observations of patients maintained on the drug for extended periods. Hartigan 9 initially collated the data of 45 patients from admissions to a busy mental hospital over six years and concluded that “prophylactic lithium” should be considered an “essential therapeutic weapon”. However, many authors later contradicted this viewpoint, with some early studies showing minimal differences between the course of BD under lithium treatment and the natural illness course. 10
Prophylaxis should be recommended for any disorder only if the treatment alters the course and outcome of the illness. In the case of BD, this would imply that pharmacoprophylaxis reduces the frequency and duration of episodes and improves outcomes such as disability, dysfunction, residual symptoms, and recovery. Understanding the natural course of BD is essential to determining whether pharmacoprophylaxis, particularly with lithium, offers these benefits.
To understand this better, we must examine the natural course of BD in community samples, as clinic-based samples tend to identify a more severe subpopulation. Naturalistic studies examining the course of BD should be representative and employ a longitudinal design. Such studies would provide valuable insights into the natural course of patients receiving professional care and those in the community who remain untreated. 11 Additionally, these studies should account for time and age-related effects, capturing changes that might occur over time as a result of improvement in care provisions and quality. Unfortunately, such studies are lacking for BD.
There are significant gaps in our knowledge about early-onset BD characteristics, long-term course and syndromal shifts in untreated community cases, and the role of comorbidity and specific functional impairments in driving illness course. 11 Long-term studies suggest that 0%–55% of patients with BD have only one episode in their lifetime. 12 Further, data needs to be more consistent concerning the number of episodes with time. Earlier studies suggest that the inter-episode interval reduces with time; however, more recent studies found that the majority of BD-I patients do not demonstrate cycle acceleration. 13
Some evidence suggests that inter-episode duration may shorten initially, particularly in the first three episodes; however, after this point, the pattern is unpredictable. In contrast, other studies indicate that the interval between episodes either increases or remains unchanged over time. 14 In summary, findings from available naturalistic studies on the inter-episode duration in BD are inconclusive. Therefore, a lifelong prophylaxis recommendation might not benefit all patients.
An additional consideration here is the limited duration of follow-up in available BD pharmacoprophylaxis studies; most studies followed up patients for around two years and recruited patients who have experienced at least two episodes in the last two years. In other words, these studies have focused on short-term pharmacoprophylaxis in more severely ill patients. A five-year follow-up study on lithium prophylaxis indicated that, despite sustained treatment, the likelihood of experiencing at least one recurrence exceeded 70% within five years of recovery. 15 These findings suggest that the risk of recurrence remains relatively high even with lithium prophylaxis.
Similarly, a study examining recurrence rates with other medications reported a recurrence rate of 73% over 4.3 years of follow-up. Additionally, two-thirds of those who experienced a recurrence had multiple episodes. 16 The Zurich study, which evaluated the lifetime outcome of BD with medications, found that only about one-sixth of BD patients exhibited recovery.17,18 Some authors have argued that naturalistic evidence does not support a reduction in the rate of hospital admissions for mania with lithium; on the contrary, admissions tended to increase following the widespread use of lithium. 19
The promise of pharmacoprophylaxis also comes with certain challenges. Adherence to long-term medications is a universal problem, with around 50% rates of non-adherence in BD. Moreover, abrupt cessation of lithium has been linked to a high risk of relapse, particularly in the first three months, with relapse rates reaching up to 50%. Due to this, it has been argued that the use of lithium for less than two years in BD may even be counterproductive. 20
Arguments for Long-Term Pharmacoprophylaxis
Several meta-analyses21,22 have affirmed lithium’s efficacy in preventing illness relapses. Lithium also offers additional benefits for BD, including reducing suicidality, preventing dementia, attenuating the risk of osteoporosis, cancer risk reduction, and lowering all-cause mortality.23–26 Consequently, lithium is widely recommended as a first-line maintenance treatment for BD in several international guidelines.7,27–29
Major concerns surrounding long-term lithium use include the risk of renal and thyroid adverse effects. In this context, a causal association between lithium use and kidney damage remains to be definitively established. The risk of serious renal adverse events with lithium may have been overestimated in some of the earlier studies due to issues like detection bias and confounding by indication.30,31 Weight gain with lithium is also a concern but is generally less significant compared to other psychotropic medications and can be managed through low-intensity interventions. 32
Lithium is also known to be associated with thyroid dysfunction, mainly hypothyroidism, but overt hypothyroidism occurs in only 14%–17% of patients and is usually reversible. 33 Meta-analyses of long-term bipolar disorder (BD) treatment with valproate 34 and second-generation antipsychotics 35 indicate lower medication discontinuation rates compared to placebo, suggesting that the benefits of these treatments outweigh concerns regarding potential side effects. This parallels the case with long-term antipsychotic treatment in schizophrenia, which, despite its known risks, is justified by the benefits it offers. In addition, maintenance lithium treatment has been shown to have a favorable effect on several patient-reported outcomes, such as social functioning and quality of life. 36
Though long-term trials with lithium in BD are limited, randomized controlled trials (RCTs) have demonstrated the efficacy of lithium maintenance treatment for up to four years, 37 with uncontrolled observations suggesting benefits for up to 50 years.38,39 A common criticism against long-term lithium prophylaxis in BD is its relatively lower efficacy for preventing depressive than manic relapses. However, there are two key counterarguments: (a) The prophylactic effect of lithium against depression may manifest more slowly than mania, 40 necessitating longer trials, which are scarce, and (b) the perception of lithium’s lower efficacy against depression may be due to methodological issues in earlier studies that included rapid discontinuation protocols. Such protocols may predispose patients to manic relapses rather than depressive ones. 41 Indeed, studies that excluded rapid discontinuation have not supported the claim that lithium is less effective against depressive relapses. 36
Evidence for Long-Term Treatment with Other Agents
The 2018 Canadian Network for Mood and Anxiety Treatments (CANMAT) guidelines for the management of patients with BD 27 position divalproex as a first-line maintenance treatment for BD based on results from several meta-analyses demonstrating greater efficacy and acceptability versus placebo. Critics of valproate pharmacoprophylaxis in BD often cite the trial by Bowden et al. 42 to support their assertions. In this large, randomized, double-blind controlled trial of divalproex monotherapy in BD, it failed to outperform placebo in preventing a relapse of any mood episode. However, lithium, the established standard of care, also failed to distinguish itself from placebo, indicating a failed trial rather than a negative result for divalproex. Notably, a post hoc analysis revealed that divalproex was superior to placebo in preventing relapses in a subgroup that had responded well to acute treatment. This finding supports the maintenance efficacy of valproate in BD, which is consistent with clinical practice where treatment choices often align with acute phase responses.
With regard to other “first” and “second-generation” mood stabilizers, relatively limited positive evidence exists for the prophylactic efficacy of carbamazepine, with preferential response among those with atypical illness, schizoaffective disorder, or bipolar II. With oxcarbazepine, the evidence for long-term prophylactic efficacy is weak and insufficient.27,43 Finally, lamotrigine is a first-line treatment option for the maintenance treatment of bipolar I and II disorder with stronger evidence for the prevention of depressive than manic episodes. 27
A meta-analysis of 15 RCTs 35 found that olanzapine, quetiapine, and risperidone were more effective than placebo in reducing relapse risk, with higher all-cause discontinuation rates in the placebo group, suggesting better acceptability of the active treatments. These findings also receive support from a prior, larger network meta-analysis, 44 that examined 17 different treatments, including mood stabilizers and second-generation antipsychotics, for their efficacy in preventing mood relapse in BD. Authors found that most treatments analyzed outperformed placebo and were largely well-tolerated; the largest body of supporting evidence was for lithium. Trial durations ranged from 17.3 to 171.4 weeks; the average duration of follow-up was 74.0 weeks.
Conclusions
From the available evidence, it is clear that long-term evidence to support pharmacoprophylaxis that alters the course and outcome of BD is limited, particularly with lithium. Nonetheless, maintenance treatment confers not only clinical benefits but also a range of other neurocognitive, physical, and psychological benefits in BD. While the potential for adverse effects is a valid concern, these are generally manageable with lithium, and major side effects are rare. If this is accepted, then we should not deny the benefits of treatment by stopping medications that evidently help people take control of their disorder, live a life with dignity, and not be subservient to illness.
However, it is important to recognize that many factors may influence relapse/recurrence risk and individual responses to treatment. Consequently, a one-size-fits-all approach may not be appropriate, and decisions on lifelong pharmacoprophylaxis should be individualized, taking into account the patient’s specific context and needs. Future studies should account for the heterogeneity of BD when determining the benefits of pharmacoprophylaxis. There is a clear need for long-term naturalistic studies to better understand the course of BD—both with and without pharmacoprophylaxis—across cultures and settings to discern the profile of patients who may benefit the most from long-term pharmacoprophylaxis. But who will fund such studies, especially for a drug like lithium that is out of patent and has lower profit margins?
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.