Introduction

Following the rediscovery of its therapeutic effects by John Cade in 1949, lithium soon became a focus of clinical and basic science research. As its mood-stabilizing properties became evident, its use in the treatment of manic–depressive illness and then bipolar disorder gradually increased. However, Cade was not the only person to have taken an interest in lithium and promulgation of its properties was facilitated by the untiring efforts of Mogens Schou, Erik Strömgren, Paul Baastrup and Samuel Gershon [1]. Collectively, these clinical scientists along with many others demonstrated lithium’s effectiveness in treating acute mania and providing prophylaxis against manic or depressive relapses. An understanding of its mechanisms of action has lagged behind its clinical utility, which in recent decades declined because of perceived barriers to its use and the promotion of ‘newer’ alternatives. Findings from contemporary research should begin to reverse this trend given that it has been shown to have significant long-term efficacy [2••], and, in addition to providing neuroprotection, may even be able to reverse the putative neuropathology that underpins bipolar disorder. Thus, it shows no indication of relinquishing its bipolar throne and will likely remain at the forefront of mood disorders therapy, both alone and as an adjunct. Equally importantly, while its administration is undoubtedly sophisticated, its risks have been found to be exaggerated and hence its true value needs to be both recognized and reinstated [3••].

Lithium has a variety of actions on neurotransmitter systems and second messenger cascades. It inhibits apoptosis, reduces oxidative stress and promotes the transcription of neuroprotective proteins in the brain [4•]. At therapeutic levels of lithium, there are two particular targets of importance: glycogen synthase kinase 3ß and inositol monophosphate (IMPase) [5, 6]. Lithium’s actions on neurotransmission of cell function are likely key to understanding the neuropathology of bipolar disorder, which is increasingly emerging as an emotion and cognition neural network disorder underpinned by neuroinflammatory dysfunction in key brain regions. This perhaps partly explains the observed delay in the response to lithium and the need for ongoing treatment to prevent recurrences and progression of the illness. Empirically, its long-term efficacy is further indirectly underscored by the increased likelihood of relapse following abrupt cessation. Clinical trials of ‘newer’ therapies, which initially trump the effects of lithium in the short term by virtue of anxiolytic and sedative actions, fail to allow for these differential mechanisms of action and, therefore, often incorrectly conclude that lithium lacks efficacy and effectiveness because of delay. In practice, lithium can be used alone or in combination with neuroleptics, anticonvulsants and antidepressants, and remains an optimal strategy for the treatment of both acute and maintenance phases of mood disorders.

The evidence for lithium in the management of mood disorders is sufficient and substantial and commensurate with its standing in current clinical practice guidelines. This brief review paper reappraises its current therapeutic status, especially in the management of bipolar disorder.

Current Status

Bipolar Disorder

Lithium is the archetypal treatment for bipolar disorder, but bipolar disorder can manifest in a variety of alternate forms, which tend to be less sensitive to lithium and more difficult to treat. Despite taxonomic revisions and modifications, bipolar disorder remains a challenging diagnosis – partly because it usually first emerges forme fruste as ‘major depression’. It can also present as an admixture of manic and depressive mood symptoms superadded to which it is often accompanied by anxiety and coloured by personality. Within the spectrum of bipolar disorders, lithium is best suited to ‘classic’ manic–depressive patterns characterized clinically by recurrence with cyclical episodes of depression and mania punctuating sustained periods of remission and return to function [7].

The strongest predictive factors of response to lithium include a family history of bipolar or major depression; evidence of lithium response in relatives with mood disorders; later age of onset at first episode; fewer hospitalizations prior to commencing treatment; a history of recurrence with multiple past episodes; an MDI (mania followed by depression and then a well interval) pattern of illness with inter-episode recovery and an absence of mixed episodes or rapid cycling [8]. Consequently, lithium response is regarded as a marker of ‘classic’ bipolar disorder (or, more correctly, manic–depressive illness) and perhaps reflects the existence of a separate subset within the bipolar spectrum [9]. Duffy et al. recently found that offspring of bipolar disorder patients who developed bipolar themselves were more likely to experience an episodic course if their parents responded to lithium treatment even though lithium responsivity did not confer a higher risk of developing mood disorders overall [10••]. Bipolar I Disorder is generally more responsive to lithium therapy than Bipolar II Disorder [7], but patients in the latter group may be more generally treatment resistant, rather than specifically to lithium therapy, possibly because of greater psychosocial problems [11].

Remarkably, a third of lithium-treated patients achieve ‘excellent responder’ status [12], and maintain remission without relapse or recurrence for 10 or more years. These robust responders posit the possibility of distinct subtypes of bipolar disorder, and support the notion that “every bipolar disorder patient deserves a trial of lithium somewhere early in their course of illness to find out if they are in this category” [13].

Acute Mania

Evidence for lithium monotherapy in the treatment of acute mania is overwhelming [14] and hence its primacy in clinical practice guidelines [15–17]. Studies since the 1950s have consistently reinforced lithium’s effectiveness compared with placebo, and indeed most other medications including neuroleptics [18]. A meta–analysis of acute mania treatment trials with lithium reported a superior response versus placebo [18]. But despite this evidence, because the effects of lithium in the treatment of a manic episode may take 6–10 days after initiation or dose increase, it is often administered alongside a neuroleptic for prompt relief from symptoms such as agitation [19]. Interestingly, an early treatment response to lithium may predict whether it will be effective longer term [20]. In sum, lithium is an effective antimanic agent.

Acute Depression

The evidence for lithium’s efficacy in bipolar depressive episodes is less robust in comparison with its therapeutic effects in mania, and it is therefore not considered to be first choice as monotherapy. This phase of bipolar disorder usually presents greater treatment challenges overall, partly due to the more complex causes of depressive episodes [7]. Unfortunately, in most patients, bipolar depression is often the predominant phase of the illness, and it is the phase that confers greatest functional impairment and risk of suicide. Lithium can take 6–8 weeks before antidepressant effects achieve remission, which reduces its suitability as monotherapy in urgent cases. This delay is a distinct disadvantage in clinical trials, which often last only a matter of weeks, and this has contributed to insufficient controlled studies, and inconclusive systematic reviews of its efficacy [21, 22]. In this context, the BALANCE (Bipolar Affective disorder: Lithium/ANti-Convulsant Evaluation) study is seminal as it managed to demonstrate unequivocally that lithium monotherapy reduces the likelihood of relapse in depression and reduces the number of depressive episodes [2••]. But in practice, lithium is often used in combination with other agents for the treatment of acute bipolar depressive episodes, because of its anti-suicidal properties and mood stabilizing effects, which also decreases the likelihood of switching to mania [15–17, 23].

Prophylaxis

A large body of evidence accumulated over 60 years supports lithium prophylaxis in bipolar disorder [2••, 19, 24–26]. A significant decrease in risk of relapse has been shown time and time again, along with prevention of suicidal behaviours and long-term maintenance of euthymic mood. A recent study by Berghöfer et al. followed 346 patients for up to 10 years and demonstrated that the prophylactic effects of lithium were stable over time [27]. Lithium is particularly potent in preventing the recurrence of mania [28, 29], but its prophylaxis does extend to depression. For example, the BALANCE study underscores lithium’s effectiveness across both poles of the illness [2••]. This study also demonstrated its superiority over valproate though this was not its primary focus. The active phases of bipolar disorder have been described as ‘battles’ but it is successful maintenance and prophylaxis that determine the outcome of the ‘war’. In this regard, lithium reigns supreme and remains the most effective agent.

Unipolar Depression

Acute Depression

Adjunctive lithium therapy is useful for treating refractory major depression, and clinical trials support its efficacy in combination with tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs), tetracyclics and monoamine oxidase inhibitors (MAOIs) [30–32]. Patients with treatment-resistant depression are more likely to respond to lithium augmentation if they possess a family history of bipolar or unipolar depression, and if they have a history of more than three previous episodes of depression [33, 34]. In practice, lithium can be added to any antidepressant, whereas combination therapies with other agents are often complicated. While the extant evidence favours the use of lithium in combination with other antidepressant medications [30, 35], a recent study evaluating lithium monotherapy for unipolar depression found that 50 % responded to treatment in 4 weeks, and two thirds eventually achieved remission [36]. This is particularly noteworthy in the context of subsyndromal manic symptoms in major depression [37].

Clearly, undiagnosed bipolar disorder is a potential confounder for unipolar depression studies into lithium response, although the extent of this is difficult to determine. Sugawara and colleagues retrospectively reviewed adjunctive lithium for treatment-resistant depressed patients in a study that originally included 22 bipolar patients and an additional five from 57 patients who transmuted diagnosis from major–depressive disorder to bipolar disorder by study endpoint. Predictably, bipolar patients were more likely to achieve a superior response [33]. Corroborating this further, Inoue and colleagues demonstrated that presumed unipolar respondents often underwent a diagnostic change to bipolar disorder after follow up [38•]. This lack of longitudinal diagnostic fidelity makes it difficult to determine outcomes with confidence and suggests that lithium responsivity may be governed by ‘bipolarity’ per se rather than efficacy derived from major depression characteristics.

Prophylaxis

Interestingly, even though there is evidence to suggest that adjunctive lithium confers some prophylaxis against depressive episodes [39], it is seldom continued in patients with major depression much beyond remission of their acute episode, even though preventative antidepressant therapy is advocated especially in those prone to relapse [40].

Dysthymia and Cyclothymia

In addition to its use in bipolar disorder and major depression, adjunctive lithium therapy also has a role in the treatment of other mood disorders, such as dysthymia and cyclothymia (Table 1). Versiani et al. have reported on studies supporting the use of lithium in dysthymia [41], and similarly, Akiskal recommends low-dose lithium for cyclothymia [42]. Early studies suggested a significant preventative effect on depressive episodes in 26–36 % of cyclothymic patients over 2 years [43], but in recent years both cyclothymia and dysthymia have received less attention, possibly because of increasing diagnostic uncertainty, but also in a general move away from lithium use in ‘milder’ forms of mood disorders [44, 45].

Table 1 Summary of strength of evidence for lithium treatment
Full size table

Delivery and Dosing Considerations

The widespread use of lithium has been stifled by the increasingly popular view that it is difficult to administer. Lithium dosing has been discussed in detail elsewhere [46, 47], and despite the misconceptions regarding its administration, its efficacy demands that it is used more widely. To this end, tools such as the ‘lithiumeter’ (Fig. 1) can be employed to facilitate understanding and safe prescription.

Fig. 1
figure 1

‘Lithiumeter’. A schematic that can serve as a useful clinical tool to assist with understanding the appropriate therapeutic plasma concentrations of lithium indicated for different phases of illness. Reproduced with permission from Malhi GS, Berk M (2012) Is the safety of lithium no longer in the balance? Lancet 379:690–2. doi: 10.1016/S0140-6736(11)61703-0 [48].

Full size image

Side Effects and Contraindications

There are a number of side effects to consider when treating with lithium, both short and long term. When lithium is initiated or increased markedly, side effects can include gastrointestinal upsets, tremor, polydipsia, polyuria, headache and fatigue; these tend to be transient. Long-term side effects are of more concern. Long-term lithium use can impact the renal system (less common but high burden), thyroid and parathyroid functioning (moderately common but also in bipolar disorder more generally and lower burden), weight (common but mild, especially compared with other agents), and other systems more rarely. While long-term lithium use can impact various bodily systems, the population incidence of conditions in bipolar patients and the rarity of many of these effects should be taken into consideration, especially in balance with the treatment benefit in excellent responders or treatment-resistant cases [3••].

Lithium toxicity can result in serious long-term damage to the body, and may occur when kidney excretion of lithium is insufficient. An adverse reaction to high concentrations of lithium (>1.2 mmol/L) is known as acute intoxication. Chronic toxicity involves an adverse reaction despite lithium remaining at therapeutic doses, and can result after chronic lithium use and neural build-up. Haemodialysis is often used to promptly normalize lithium levels and avoid permanent kidney damage.

Lithium can increase the risk of heart defects in the foetus, though this has been overestimated previously [3••]. Effects through transfer in breast milk are unknown. Patients who become pregnant or are breastfeeding should consider the risks and benefits, as well as manage changes in filtration involved in pregnancy which may increase recurrence risk. The psychosocial effect of relapse, especially in responsive patients, should certainly be factored into management [47]. Older patients can be at higher risk of toxicity due to age-related changes in renal function and the impact of other comorbidities. This typically means that required dosages are lower than in the greater population [47]. Finally, those patients with existing kidney problems of higher severity may not be suitable for lithium therapy, while those with more manageable conditions may be suitable with more regular monitoring.

Dosing Schedules

Whether lithium should be administered twice daily or at a lower frequency continues to be a matter for debate [46]. While less frequent administration may improve compliance, reduce or delay risk of kidney damage, and minimize other adverse events, multiple dose schedules are currently recommended in many practice guidelines. Neural levels of lithium differ in their degradation to plasma levels, meaning that even as plasma levels are sometimes lower than optimal between daily doses, lithium may persist and continue to be effective in the brain. Alternate day dosing has also been shown to yield some benefits, such as decreases in some adverse events. However, the risk of suboptimal doses is higher with recurrence more likely [46], which may be alleviated with higher daily doses.

Therapeutic Plasma Concentrations

Optimal lithium concentrations in the blood depend upon which phase is being treated, although different recommendations vary substantially [46]. For initiating treatment and maintenance in bipolar disorder, lower levels of lithium between 0.6 and 0.8 mml/L are recommended. For depressive symptoms, levels of 0.4–0.8 mml/L are considered appropriate. For acute mania, greater levels of 0.6–1.0 mml/L are indicated; however, this increases risk of chronic toxicity (see Fig. 1). Appropriate levels must balance the competing concerns of ensuring efficacy while avoiding complications and minimizing risk.

Future Directions and Novel Findings

Suicidality

By definition and due to the nature of the illness, mood disorder patients are prone to suicidal thinking and are therefore at a greater risk of suicide; this is particularly true for bipolar disorder [49]. In this regard, lithium’s anti-suicidal properties are particularly helpful, but research into this somewhat unique trait has inconsistent results [50]. Cipriani and colleagues (2005 and 2013 update) have established its efficacy for unipolar and bipolar patients, in both a systematic review and a sophisticated meta-analysis [51, 52••], as has Baldessarini and colleagues, who showed a 5-fold decrease in suicide risk for bipolar patients [53]. In their meta-analysis of 349 patients, Guzzetta and colleagues found reduced suicide risk in major depressive disorder patients [54]. Lithium alone or combined with neuroleptics was found to lower the risk of suicide as compared with neuroleptics alone [55].

The mechanism of lithium’s anti-suicidal action remains unknown but is likely to involve moderation of impulsivity [49]. Notably, this anti-suicidal effect is independent of lithium’s effect on mood and even patients with a poor response to lithium and ongoing mood instability benefit with respect to suicidal thinking [56]. Furthermore, this effect extends across different patient populations, and can also be seen at a population level in extremely low doses – for example, when found in groundwater – as illustrated by a number of epidemiological studies [57, 58, 59•, 60, 61]. Hence, the suggestion of trace levels of lithium being added to the water supply in order to achieve population-level benefits with a low risk of side effects has been posited [58].

Neuroprotection

Mood disorders have been shown to be associated with neurological dysfunction, oxidative stress, impaired cellular mechanisms and decreases in grey matter volume and neuronal cell loss. Similar changes, though usually involving different brain regions, have also been reported in bipolar disorder. Research using advanced neuroimaging techniques has shown that lithium is neuroprotective and preserves emotion-related brain regions such as the amygdala, hippocampus, and other areas in the fronto-limbic network, with increased or preserved neural volume or density reported [4•, 62•]. Again, akin to its anti-suicidal actions, this neuroprotective effect appears to occur independently of its mood-modulating actions, such that even in those that are not classified as ‘excellent responders’, lithium can preserve and enhance grey matter volume [63•]. In comparison, the neuroprotective effects of lithium on white matter are less clear [64•]. Regardless, these tantalizing findings point to profound properties and further reinforce lithium’s suitability as an adjunctive agent, and possibly provide clues as to the functional mechanisms underpinning its clinical efficacy.

Dementia

Cognition is invariably compromised by mood disorders, and though the pattern of change is often mood-state related, naturally over time the brain receives repeated insults and accumulates trait changes. Lithium’s ‘neuroregenerative’ actions can therefore be extended and its use in dementia holds promise. For example, lithium has been shown to reduce the risk of dementia in bipolar patients [65], and may reduce the risk of Alzheimer’s disease [66•]. There are tentative indications that lithium may slow or reverse cognitive changes, such as verbal learning, verbal memory, and psychomotor speed [67]. These benefits may, however, depend upon responsivity to lithium [68]. Some studies have found no impairment or benefit in healthy volunteers at least over the short term [69]. Lithium also causes some cognitive impairments in mood-disordered patients, which makes the effects of lithium difficult to disentangle from the impact of the disorder [69]. Therefore, more research is needed to disentangle these complex effects on cognition.

Conclusion

Despite lithium’s resounding efficacy, its principal limitation remains that of tolerability, as it is toxic acutely at high doses, and chronically at therapeutic doses. The mechanisms of action of lithium suggest that second messenger systems are a key target and thus need to be better defined so as to identify opportunities for specific agents that are able to replicate its benefits but obviate its side effects. Such lithium mimetics have been designed and seem to offer promise in animal models.