By Professor Guy Goodwin


Lithium is recommended as the first line choice for mood stabilisation in bipolar disorder by all the most credible medical guidelines. Unfortunately, lithium is not the first choice for every individual with bipolar disorder. Is that because some people are lithium responders and others not? The simple answer is probably 'no'. The cost benefit of taking lithium is a continuum from very good to very bad. We tend to prefer binary categories so we invent responders and non-responders as the ends of this range. For many people the advantages are something of a mixture or even rather uncertain: they are worth understanding.

How do we define lithium response?

Lithium was originally discovered to work as a mood stabiliser when it was given to people with a defined illness course – they had regular relapse into mania and or depression. After starting lithium, this pattern was interrupted and the best responders seemed to enter effective remission from severe episodes. This kind of observation, while important, is not fully trusted in medicine because placebo effects can be powerful, so it is crucial that the positive effect has been confirmed in randomised clinical trials. However, for the individual patient, we are still stuck back at the beginning of the story. What is the illness course before and after starting lithium? In the past, it was recommended to wait and see at least two severe episodes before starting lithium. That made it a little easier to judge subsequently how much advantage there was after starting lithium.

Nowadays, we are concerned that the more episodes you have the more you are likely to have, so starting lithium earlier in the illness course seems a better decision. However, this creates difficulties. First, it means we are uncertain what the illness course would have been before we start lithium: we don't have much to compare the ‘lithium response’ to. Second, lithium usually carries a certain side-effect burden which it is tough to ask anyone to accept, plus the blood tests to check lithium levels and metabolic function, when the benefit is hard to measure in the short term. That means a lot of people start on lithium and then discontinue it. And a sudden discontinuation of lithium risks provoking a manic episode which further confounds the lithium experience. Such a decision is understandable because it weighs need, which is hard to define early in the illness course, against the costs as described. 

So, I am often unsure a person is really a non-responder without a lengthy trial of consistent treatment. Most patients do not take lithium for more than 6 months when it is first prescribed. I recommend that initial treatment should be for at least 2 years and my challenge to patients is to accept that only when they relapse will we judge whether lithium has failed. 

Even if a lot of ‘non-response’ is really uncertain response, of course there are lengthy tests of lithium that do fail and the benefit of treatment is clearly outweighed by the cost. Lithium has a number of known biochemical effects which appear to require entry into nerve cells. Lithium in solid form lives in salts. When salts dissolve, lithium exists as a positively charged particle and it can only enter cells through specialised protein engines that propel it across the cell membrane. So, for lithium to work, the patient has to take it; it has to be absorbed from the gut and transported into nerve cells and there it has to be present at just the right concentration to block the target enzymes. Too much and it makes you sick: too little and it will not work. Even if we all had identical physical make up, getting the dose of lithium right might be difficult. But we don’t have an identical physiology, so we should expect variation in how we transport lithium into cells, the balance of how our enzymes react to lithium and then how the biochemical changes get translated into the way our brains work. In other words, there are a lot of ways that our bodies might cause lithium treatment failure. Research has shown that excellent lithium response tends to run in families. That usually implies that our genes determine whether lithium works or not: there are ongoing efforts to decide which genes in particular may be involved. It is an exciting example of where genetic studies could really move us forward in understanding how lithium works and how we could design something better. However, translating this new knowledge into clinical practice may take a long time. 

For the present, the optimal way to use lithium is to start it with a clear and understood need: to prevent relapse of bipolar disorder. The patients most likely to respond are those who have clear-cut episodes of euphoric mania (rather than more mixed irritable or paranoid experiences). There should be no doubt about the diagnosis in the mind of the doctor who prescribes it and as little ambivalence as possible about the need on the part of the patient being invited to take it. Lithium is not for the faint hearted! The dose should be raised quite slowly to the highest well tolerated dose, which can only be determined by trial and error. Usually this will give lithium levels around 0.7 mmol/litre in blood. Pushing for higher levels at the cost of side-effects is almost always counter-productive. Then, once on the chosen dose, have patience and plan to evaluate things over the long-term (at least two years). Modern mood monitoring on a digital platform can help. Use relapse (especially to mania) as the most important measure of treatment failure. 

In summary, if you do experience obvious benefit from lithium, you are lucky, use it wisely. If the benefit is less obvious, be patient, don't give up on it until you are sure it is really not for you and that the cost/benefit is negative. Its effective use requires the doctor and patient to work together to optimise the setting and the dose. Lithium remains a remarkable semi-accidental discovery with unique benefits for the majority of bipolar patients: remember it literally saves lives.