Lithium the magic ion: Restoring and preventing?

Lithium is one of the oldest treatments in psychiatry. Over the past six decades, it has totally transformed the prognosis of bipolar disorder, but even so its prescription has dramatically decreased in contrast to the use of atypical antipsychotics. This change in practice is partly linked to concerns about the long-term side effects of lithium. A recent meta-analysis (McKnight et al., 2012) revised downward the side-effects risks of lithium intake, though according to naturalistic studies and clinical experience monitoring the renal effects, which can lead to renal failure, remain important. In addition to reducing urinary concentrating ability, lithium can also cause hypothyroidism, hyperparathyroidism and weight gain (Malhi et al., 2012), as well as compromised cognition.

This possibility of cognitive deficits has been the subject of extensive debate with mixed views from both patients and doctors. Most formularies inform clinicians that lithium can produce cognitive ‘slowing’: a complaint often echoed by patients treated with lithium. On the other hand, lithium can induce a persistent flattening of emotions, termed ‘the psychological syndrome’ in the French literature (Rosier et al., 1973). This is often construed by patients to be a side effect of lithium whereas psychiatrists are more inclined to view it as a positive consequence of its clinical effect on emotional reactivity – the core of manic-depressive illness. But the putative effects on cognition are difficult to distinguish from the cognitive effects of the illness itself. Interestingly, and in contrast, healthy volunteers report quite consistently that attention, memory and executive functions are relatively preserved with long-term lithium administration (Tsaltas and Kontis, 2008).

Mauer et al. (2014) discuss these points and invite us to think about the long-term effects of lithium on cognition. Thymic disorder is a recognised risk factor associated with dementia, particularly in Alzheimer disease (AD). In this systematic review, the authors synthesize evidence of lithium use for the prevention of AD in patients with mood disorders. Interestingly, they report that bipolar patients receiving chronic lithium therapy seem to have a smaller incidence of AD than those who are not on lithium maintenance therapy.

However, the authors also challenge the potential benefit of lithium in restoring cognitive function in AD in the absence of comorbid affective disorder. Indeed, to date the evidence in this regard is somewhat disappointing. How can this be explained? An obvious consideration is the duration of treatment. In bipolar disorder retrospective studies, lithium has usually been prescribed for a decade or more in contrast to several weeks or months in the majority of AD prevention trials. These clinical findings are further at odds with the astounding neuroprotective effects of lithium in vitro; notably on extracellular beta-amyloid (Aβ) deposits and intracellular Tau neurofibrillary tangles – two key molecules in the neuropathology of AD. This difference has prompted the hypothesis that, in clinical use, lithium could be better suited to prevention than to arresting the degenerative processes in AD per se.

Naturally, to properly appraise the value of lithium as a putative cognitive enhancer it is not sufficient to rely on clinical reports of beneficial effects. Proof of concept could be tested within amyotrophic lateral sclerosis (ALS). Many in vitro studies and studies of animal models have found marked neuroprotective effects of lithium, which delayed disease onset in duration and augmented lifespan. A novel trial of lithium in patients with ALS that suggested its clear therapeutic efficacy was regarded as a substantial advance by patients and the medical community (Fornai et al., 2008). Such encouraging results lead to the hypothesis being tested in a rigorous 18-month, multicentre, randomised, double-blind, placebo-controlled trial. Alas, LiCALS (lithium carbonate in patients with amyotrophic lateral sclerosis) demonstrated no advantage for the active agent over placebo at the primary endpoint (survival rate at 18 months), quashing perhaps the hopes of lithium being an effective treatment (Morrison et al., 2013).

Nevertheless, a growing literature base reports promising preclinical data for the use of lithium in many diseases, both acute, such as stroke or brain injury, as well as neurodegenerative (Young, 2011), such as Huntington disease or schizophrenia. Theoretically, lithium remains a powerful neuroprotective and neurotrophic agent and there is an urgent need for clinical trials in humans to test effects in a range of diseases. However, in order to establish lithium as a cognitive enhancer for neurodegenerative diseases, a number of points need to be taken into consideration.

First, the precise dose needed to induce a neuroprotective effect in vivo is unknown and this needs to be determined. In bipolar disorder maintenance therapy, the standard dose of lithium is high, but as suggested by Mauer et al. (2014) lithium even at trace levels appears to have benefits for suicide and other behavioural outcomes in primary prevention. As yet, there are no data for trace levels of lithium in the prevention of dementia – possibly because of multiple confounding factors.

Second, it is important to limit our considerations in the first instance to well-characterised neurodegenerative diseases. Even though there are biochemical changes in specific proteins that produce characteristic inclusion bodies within glia, neurons or both, the same populations of neurons can be affected by different disorders, and therefore there is a mismatch between the clinical presentation of dementia and its molecular pathology leading to significant heterogeneity in clinical trials. Biomarkers such as those in cerebrospinal fluid (Aβ_1–42 and phospho-tau) have 95% sensitivity for recognising prodromal AD (Hampel et al., 2014). But in familial AD it has been suggested that decrements in cerebrospinal fluid (CSF) Aβ_1–42 concentrations may begin up to 25 years before any clinical manifestations in familial AD mutation carriers. Therefore, it is perhaps not surprising that lithium is unable to reverse CSF-based biomarker concentrations in AD with 10 weeks of treatment. Again, it would be interesting to know if lithium use for a longer period could act upstream and prevent protein accumulation in the first place.

Sadly, most major pharmacological discoveries have been serendipitous and currently there are few promising molecules progressing through strategic drug development pipelines in the field of neuroscience. It is therefore important to turn to agents that are already tried and tested. Psychotropic drugs have multiple effects, many of which are not fully understood and remain untapped. For example, the FLAME (fluoxetine for motor recovery of patients with acute ischaemic stroke) trial identified selective serotonin reuptake inhibitors as having neurotrophic effects that can improve the prognosis of stroke (Chollet et al., 2011). Similarly, our understanding of the mechanisms of action of lithium therapy and its effects on the brain remains incomplete and so perhaps it is time that this becomes a focus for future studies, especially in the field of neurodegenerative diseases.

See Key Review by Mauer et al., 2014, 48(9): 809–818.