The more statistical view of group effects is that lithium will reduce the rate of relapse by 20-30% (Fig. 5.1). The relapse rates in most trials are about 50% over a year on placebo and this is reduced to about 30% on lithium. However, it is difficult to translate this figure to individuals.

Fig. 5.1 Lithium reduces the relapse rate in manic depression; stopping lithium precipitates relapse, beyond natural history.

(Based on data from Schou 1993.)

5.5 What are the indications for lithium treatment?

The most common reason for prescribing lithium is as a prophylactic treatment in bipolar illness. Lithium is more usually used for the prevention of bipolar illness with mania rather than hypomania though it is effective in both disorders. It is also effective in the treatment of mania and bipolar depression. It can be used on its own for the treatment of bipolar depression but is more commonly prescribed in combination with an antidepressant.

Lithium is also useful in the acute treatment of unipolar depression in combination with an antidepressant. Lithium is usually prescribed for those patients who are suffering from depression but have had only a partial response to an antidepressant; lithium is added to augment this treatment. It is the only well-validated augmenting strategy for unipolar depression (Anderson et al 2000, Goodwin 2003). Lithium also has a place in the prevention of recurrent unipolar depression.

Lithium is occasionally prescribed for certain forms of headache (e.g. cluster headaches).

Lithium has also been tried in a number of other disorders but no method of treatment is proven–for example it has been tried as a way of reducing aggression in those with a learning disability or helping alcoholics to reduce their intake or prevent relapse.

5.6 Who should be taking long-term preventive treatment for bipolar disorder?

If someone has had even a single, severe, highly disruptive episode of mania or severe depression then it is tempting to suggest that they should take long-term treatment and indeed a few patients do this. However, after one episode it is very uncertain when the next episode will occur (though it almost certainly will occur at some time, see Q 2.18) and most people will not want to undertake a very long-term treatment at this stage.

For those who experience a recurrent bipolar illness the decision to take long-term treatment is based on the frequency of relapse, the severity and length of episodes and the impact of the illness on normal life.

One estimate is that if someone has just recovered from an episode of illness and had a previous one in the last 4 years, then that person should probably take a preventive treatment. This is based on the expectation that lithium will be likely to prevent at least one major relapse over the next 5 years.

5.7 What are the common side-effects of lithium treatment?

Most people taking lithium should expect to experience an increase in thirst and, accompanying this, an increase in frequency of urine. Often when patients initially start treatment they experience a diuresis.

The other common symptom is tremor which is an exaggeration of the normal essential tremor. It is very likely that patients will experience these two side-effects but usually at a mild level.

Another side-effect which is commonly reported is weight gain. It is difficult to know why patients gain weight. For some it is clearly related to them drinking a lot of sweet fluids which appear to slake the thirst better than water. In others the cause is rather obscure and they tend to gain weight even though they don’t feel that they are eating or drinking any more.

Many patients with manic depression, both when they are taking lithium and when they are not, complain of concentration and memory problems. It is not clear if lithium at therapeutic levels causes cognitive problems. Depression even at a low level can also affect cognition and disentangling these effects for individual patients is difficult.

5.8 Can lithium have long-term harmful effects?

The two major concerns about long-term lithium treatment are permanent effects on the thyroid and kidney. Hypothyroidism occurs in about 10% of those who take lithium and will usually persist even if the lithium is stopped.

Lithium can also affect the ability of the body to concentrate urine, which is the mechanism behind the excessive thirst and drinking that many patients experience. The biochemical explanation is that the kidney is not responsive to antidiuretic hormone (ADH). This effect usually reverses when lithium is stopped but can occasionally be persistent. There should be no permanent damage to the kidney if there are no episodes of lithium toxicity.

5.9 How dangerous is it?

Lithium poisoning through very high blood levels (3 mmol/l and higher) can cause permanent damage to the brain and kidneys and is potentially fatal. Damage to the brain often affects the cerebellum in particular and can leave patients permanently disabled. Likewise renal function can be permanently compromised by toxicity.

However, when the levels are kept below 1.2 mmol/l then lithium is very unlikely to cause long-term harmful effects on either the brain or the kidney. Usually those that suffer from toxicity do so because the levels have slowly gone up rather than in an acute overdose. Although lithium is obviously known to be dangerous at high levels it is surprising how rare overdoses of lithium seem to be, especially given that it is prescribed to a group that is exceptionally vulnerable to suicide (see Q 5.26). This may be because it is a very effective treatment and also that it may have a specific antisuicidal effect. However, the most serious adverse effects of lithium toxicity do seem to occur among those who deliberately take a large overdose of slow release tablets where the levels continue to rise long after ingestion even if they have vomitted.

5.10 What are the symptoms of lithium toxicity?

As the blood level rises above 1.5 mmol/l people start to feel generally physically unwell, go off their food, feel nauseous and may start to vomit. They will also be likely to experience diarrhoea. In fact gastrointestinal symptoms are very useful warning symptoms of impending toxicity. As the blood level continues to rise above 2.0-3.0 mmol/l neurological symptoms appear with stiffness, ataxia and confusion. If the levels continue to rise confusion becomes more prominent and coma, cardiovascular collapse and convulsions will appear.

Levels can reach as high as 5.0 mmol/l and fatalities and permanent physical damage can be expected.

It is important that patients know the early symptoms of toxicity, particularly the gastrointestinal ones, so that if they do experience these symptoms they can take action early. Obviously a lot of people experience occasional diarrhoea and vomiting from food poisoning but consideration that this may be a symptom of lithium toxicity is important.

However, the levels can be quite confusing, as patients can still have prominent symptoms of toxicity even though the serum level is no longer high. This is because the lithium level in the tissues is still high.

5.11 What are the common reasons for patients becoming toxic?

Toxicity often creeps up on patients rather than suddenly hitting them. This can be because there is a gradual decrease in renal clearance with age and if the lithium levels are not monitored regularly the level can slowly increase with only a gradual onset of warning symptoms. Iatrogenic factors are also important and the commonest reason for toxicity is the addition of a thiazide diuretic to lithium treatment. This reduces clearance of lithium and toxicity follows.

Other commonly prescribed drugs that can interact with lithium to increase levels are non-steroidal anti-inflammatory drugs such as ibuprofen and so particular care needs to be taken when treating bipolar patients who also have arthritis.

It is important to warn patients that they must check that any medicines they are going to take are compatible with lithium.

5.12 What needs to be done prior to initiating lithium treatment?

The most important test is to ensure that renal function is normal. In physically healthy patients this can be done by checking their creatinine levels; in those whose renal function is impaired or likely to be impaired (e.g. the elderly) it is necessary to do a creatinine clearance test. Lithium can affect thyroid function (and abnormal thyroid function can also affect the course of bipolar illness) so checking thyroid function prior to initiation is also necessary.

Ensure that the patient has been given a leaflet outlining lithium treatment (see model in Appendix 3) as they will not remember enough from a conversation. Stress the symptoms of toxicity that need to be recognised, and also that careful use of lithium and good monitoring will prevent toxicity occurring. An example of specific instructions from the hospital about how to start treatment are shown in Figure 5.2.

Fig. 5.2 Example of a letter from a consultant to both GP and patient about starting lithium treatment.

The other important piece of information the patient should know is that this is a long-term treatment and sudden cessation of lithium can lead to rebound mania and so a worsening of the illness course (see Q 5.27). This means that the illness can be made worse by taking lithium intermittently than it would have been if the patient had never taken the treatment at all!

5.13 How is lithium treatment initiated?

Assuming that preliminary checks have been made (see Q 5.12) and renal function is normal, the usual routine for beginning lithium treatment would be to start at 400 mg at night and then check a blood lithium level at 5 days. It is likely that the initial target blood level will be 0.6 mmol/l and benefit will be judged at this level. If the blood level at 5 days comes back at less than 0.6 mmol/l the lithium dosage should be increased by 200 mg and checked again at 5 days. This process should be repeated until the desired level is reached. When a stable dose has been reached, continue to monitor every 2 weeks for 1 month, then monthly for a further 3 months and 3-monthly thereafter.

5.14 Why is lithium usually prescribed to be taken all at night?

There is no good therapeutic reason for taking lithium at night–it is no more effective then than if taken at any other time. However, side-effects are generally more prominent at higher blood levels which tend to be in the first few hours after taking the tablet (see Fig. 5.3), so it can be preferable for the patient to be asleep at this time. For compliance reasons taking tablets once a day is usually better (see Q 5.43).

Fig. 5.3 Daily variation in lithium level.

(Based on data from Bennie et al 1977.)

However, the main issue is arranging serum lithium estimations. These need to be taken 12 hours after the last dose so that if the lithium is taken in the evening it is relatively easy and convenient to do a blood test the following morning. The alternative of taking the tablets in the morning and doing blood tests in the evening causes not only logistic problems but also often results in samples being kept overnight which can affect other blood tests (e.g. potassium levels) that are done at the same time.

Some people find that splitting the doses of lithium between morning and evening is preferable, usually because this can reduce side-effects. This can make monitoring confusing if a blood test is taken only a couple of hours after the last dose which shows a high (peak) level. Patients should be asked to delay the morning dose until after the blood test, so that a trough level is measured.

5.15 What formulation of lithium should be prescribed?

Lithium is the same whatever formulation it comes in–tablets, slow release versions or liquid–carbonate or citrate. However, the dose is not interchangeable between different formulations and one cause of toxicity is a change of formulation. Different formulations have different bioavailabilities and if the formulation is changed then move to a lower dose and then monitor on a weekly basis until the levels are stable. The most common reason for doing this is changing to a liquid if the patient is not getting on well with tablets, as they are large tablets which can be difficult to swallow.

Currently in the UK lithium carbonate (Priadel) is the most commonly prescribed tablet. It is probably a good idea for all the patients in a practice to be taking the same formulation to reduce the possibility of prescribing errors.

5.16 What monitoring needs to take place while on lithium treatment?

When the patient is on a stable dose of lithium, checking lithium levels every 3 months is adequate; checking renal and thyroid function every year is also needed. Weight gain can be a problem with lithium and keeping an eye on this from the start rather than waiting until the patient has gained weight is worthwhile. There have been reports of hyperparathyroidism with lithium and some advocate checking calcium levels annually as well (Drug and Therapeutics Bulletin 1999).

5.17 Are lithium cards useful?

It is very helpful if patients keep a note of all their blood tests, so that whichever doctor they see there is a readily available record. This needs to include doses of drugs, dates of lithium levels, and also renal and thyroid function, thus allowing blood tests done either in the hospital or in general practice to be easily accessed.

Some pharmacies and clinics give out cards that are marked up to include spaces for the range of monitoring that is needed. Sometimes patients keep a notebook with various details about their treatment; no doubt some now use electronic means!

5.18 What blood levels should I be aiming for?

The therapeutic range that is routinely aimed for is 0.4-0.8 mmol/l in the serum at 12 hours after the last dose (Fig. 5.3). When lithium was first introduced into psychiatry, higher levels of up to 1.2 mmol/l were commonly used but over the years the level aimed for has tended to be lower. This is probably because doctors have become more concerned about toxicity, even though with care higher levels can be maintained safely.

Levels need to be at least 0.4 mmol/l in order for lithium to be an effective treatment, and no patient should be maintained below 0.4 mmol/l. Even levels of 0.4 mmol/l may not be enough and the usual practice would be to aim for a level of between 0.6 and 0.8 mmol/l. If this is not well tolerated a trial at lower levels down to 0.4 mmol/l is reasonable. If lithium is well tolerated at levels between 0.6 and 0.8 mmol/l but is only partially relieving or preventing symptoms, then higher levels up to 1.0 mmol/l (or even up to 1.2 mmol/l in acute treatment) may be needed and can be safely used with appropriate monitoring.

5.19 How should I react to a high blood lithium level?

If a lithium level comes back between 1.0 and 2.0 mmol/l it is important to consider the possible reasons for this and to ascertain the patient’s physical and mental state. Often higher readings come back because the blood sample has been taken less than 12 hours after the last dose (see Fig. 5.3). This may be because the patient is taking the lithium in the morning and has had a blood test only a few hours later. It is important to establish first whether this might be the case.

Next: How is the patient? The clinical assessment is more important than the blood level as patients can be toxic at apparently regular levels (<1.0 mmol/l) and not toxic at apparently high levels well over 1.0 mmol/l. If the patient has no symptoms of toxicity and seems well otherwise it is usually safe to continue on the lithium and have another estimation of the lithium level, aiming for as accurately at 12 hours after the last dose as possible. However, if there are symptoms of lithium toxicity (e.g. diarrhoea and vomiting) it is necessary to stop the lithium temporarily and then take another level. If the level is only marginally high (e.g. 1.0 mmol/l when the usual level is 0.8 mmol/l) and there are only mild changes (e.g. loose stools but not vomiting or neurological symptoms) then a reduction of the dose and a further blood test may be appropriate.

If the patient has prominent neurological symptoms (including confusion) or cardiovascular symptoms this would indicate the need for hospital admission. The usual treatment is intravenous infusion of normal saline but very high levels (e.g. 4.0 mmol/l) will be likely to require dialysis.

Overall, the level of symptoms, together with the clinical assessment, is more important than the blood levels. It is easy to keep changing the lithium dose frequently in response to blood levels but then to find it difficult to reach a stable dose. It is generally better to re-check levels rather than changing doses if there are no symptoms of toxicity. Dose changes in long-term treatment should be made only if two or three levels indicate the need for this (see Case vignette 5.1) unless there are symptoms of toxicity or the clinical state requires it.

5.20 How should levels above the usual therapeutic range be dealt with?

Case vignette 5.1 illustrates the difficult balance that needs to be made between clinical state, side-effects, lithium level and dose. Sometimes it is difficult to understand what is happening but the focus on the physical side-effects, mental state and the timing of the levels is the main guide to treatment. Even high levels do not immediately lead to a change in dose, but if the clinical state is satisfactory then a repeat level is the appropriate action.

Having a good degree of concordance with the patient described in Case vignette 5.1 about treatment and monitoring, along with her giving a high value to the benefit of lithium in treating the depression (because of the severity of the symptoms prior to treatment) enabled us to persevere with this treatment when other patients may have decided to try an alternative, even though the potential benefits would be uncertain.

CASE VIGNETTE 5.1 BALANCING LITHIUM LEVELS

A woman in her forties has a history of hypomania but has now had chronic depression for more than a year which has prevented her from working though she has maintained basic domestic tasks. Treatment with amitriptyline up to 225 mg daily has led to only partial improvement but there has been no recurrence of manic symptoms. It is agreed to add lithium to the antidepressant because of the need to augment the antidepressant treatment and also to prevent recurrence of hypomania. Her renal and thyroid functions are normal.

She reaches stable levels of 0.6-0.7 mmol/l on a dose of 600 mg and there is a substantial improvement in her depression over 6 months. However, there are occasional levels that are higher (e.g. 1.2 mmol/l) but further investigation shows that this was a 10 hour level when she usually gets levels at 14 hours.

However, her depressive symptoms then worsen and the dose is increased to 800 mg. The levels increase to 0.9-1.1 mmol/l. She has some tremor and thirst but the improvement in depression is more than compensating for this. One level is as high as 1.2 mmol/l but her side-effects are not worse and she continues on 800 mg; on repeat testing the level has dropped to 0.9 mmol/l.

However, 3 months later she develops diarrhoea and her level is measured at 1.2 mmol/l. Because the physical symptoms are more prominent the dose is decreased to 600 mg. The level drops to 0.6 mmol/l on the two follow-up samples over the next 2 weeks, and the diarrhoea resolves. Her depressive symptoms become more prominent and so an increase is made in the lithium to 700 mg (levels of 0.8 and 0.9 mmol/l) and then to 800 mg. However, the levels now go up to 1.5 mmol/l, though again this is at 10 hours. Because she has prominent tremor and thirst the lithium dose is reduced to 700 mg and levels of 0.9 mmol/l are then maintained.

5.21 Is there a lack of lithium in the blood of manic depressives?

No, everyone has a trace of lithium in their body and we probably all need a tiny daily intake of lithium, as we do with many other trace elements which we get from our food. However the amount of lithium needed is about 1 mg a day whereas the smallest tablet of lithium is a hundred times more than this. Those who have manic depression are no different from anyone else in the amount of lithium they naturally have in their body.

Sometimes patients see lithium treatment as ‘correcting a chemical imbalance’. It is often unclear exactly what they mean by this although it is likely that there are some neurotransmitter changes in the brain of a person who is manic or depressed. It is possible that these changes reverse with effective treatment and so lithium is ‘correcting an imbalance’. However, the chemical imbalance is not a constitutional lack of lithium. These phrases are often used as useful metaphors that can be agreed with the patient even if neither of you actually knows clearly what the other means!

5.22 What effect does lithium have on the developing foetus?

There is a definite risk of abnormalities in the foetus when the mother has been taking lithium during the pregnancy. This is an effect in the developmental part of the pregnancy, i.e. the first trimester. The best recognised abnormality is Ebstein’s anomaly, which is a severe and potentially fatal cardiac abnormality. However, the relationship of this to lithium has been questioned in recent years.

The risk of any congenital abnormality in ‘normal’ pregnancies is about 3% and in those taking lithium about 11%. This means that the great majority of babies will not have any recognisable abnormality, but this level of congenital abnormality is not usually an acceptable risk for mothers to take.

Apart from congenital abnormalities which occur in the first trimester, the effects on the foetus in the latter part of pregnancy can also be significant (e.g. on the thyroid or heart rhythm) and there may still be some risk in taking lithium even at this stage.

5.23 What is the management for a bipolar woman who has been successfully treated with lithium but now wants to become pregnant?

Decisions around treatment of bipolar women in regard to reproduction is an exceptionally difficult area and expert advice is needed. Several factors need to be balanced.

1. There are small but definite risks of lithium to the foetus (see Q 5.22). A judgement needs to be made whether it would be possible for the potential mother to come off lithium and remain well off this treatment prior to becoming pregnant. This might be possible for a woman who has been stable for several years where the risk of relapse is judged to be low. Unfortunately the common scenario is that when you try to slowly reduce the lithium she experiences a relapse and you are back to where you started. Some women opt to continue with the lithium while they become pregnant even knowing the risks to the baby and then ensure that the pregnancy is closely monitored and the foetus assessed for abnormalities at an early stage.

2. An alternative is to try a different medication during pregnancy. Older antipsychotic drugs such as chlorpromazine may be effective in the prevention of mania and are not thought to carry an increased risk of foetal abnormality. This is a possible option during pregnancy instead of lithium. Treatment with alternative preventive medications, particularly the anticonvulsants carbamazepine and valproate, run similar overall risks of malformation (though the particular malformations are different) to the growing baby.

3. The immediate postpartum period is an exceptionally vulnerable time for women and there is about a one in three chance of relapse, mania being more common than depression postpartum. Most doctors would recommend that if the woman has gone through pregnancy without lithium that she restarts this treatment immediately after childbirth and not breast feed.

5.24 Is it appropriate to breast feed while taking lithium?

No. There would have to be a very good reason why women are feeding their babies in this way if they are taking lithium as there is a substantial chance that the baby will experience toxic symptoms.

5.25 For how long should a bipolar patient continue to take lithium?

The first goal of any long-term treatment is to judge whether the drug is acceptable. Some patients are particularly sensitive to lithium and cannot take it even at therapeutic levels. For example a female bipolar patient found that she was getting diarrhoea even at levels of 0.4 mmol/l and so could not continue with a trial of treatment for longer than a month.

If the treatment is acceptable and reasonably tolerated then the next aim is to judge whether lithium is beneficial. This usually needs to be done over 6 months to 1 year, so that even if the patient suffers a relapse in the first few months it is usually worth continuing the treatment to make a longer term judgement as to whether lithium is providing a reduction in severity or frequency of relapse compared to the period prior to taking it. There is also a judgement to be made of balancing efficacy against tolerability.

If lithium does prove beneficial it needs to be taken long term. The propensity to recurrence in manic depression never goes away and if someone is doing well while on lithium there needs to be a very good reason to stop it. The experience of seeing a patient who has been well for 10 years or more while on lithium and then deciding to stop it and rapidly experiencing a relapse is not uncommon (Case vignette 5.2).

Statistically the time needed to prevent a relapse is 1-2 years so that patients who take lithium for a shorter period may be getting no benefit from it and if they stop it suddenly may well be making their illness worse through rebound effects. Taking the treatment for 6 months and suddenly stopping may mean that the illness course is worse than never having taken it at all. (see Q 5.27).

It is worthwhile discussing with patients that if lithium is beneficial they should continue taking it unless there is a very good reason to stop. Some patients may ask if they are taking it for life–a reasonable response is that they should be taking it for several years but with review as to how they are, how helpful the treatment is, what problems they are getting with the medicine and then make a decision as to whether to continue. There may, of course, be other treatments that emerge in the coming years which will replace lithium or indeed provide a permanent cure–but currently lithium is still the treatment most likely to be of benefit to manic depressives.

CASE VIGNETTE 5.2 STOPPING LITHIUM SUDDENLY

Peter was ill as a student (probably mania followed by depression), reading law at university. His father took him away for an extended holiday in Spain where he built walls, as Churchill had recommended. He got back to university and then qualified as a barrister. He had been brought up abroad and returned to Uganda to work and built up a sound legal practice. However, the arrival of Idi Amin put paid to that and he returned to England in a very depressed state from which he took a year to recover. He then worked in the Caribbean as a government lawyer for 10 years but this ended when he became manic, spent all his savings on yachts and was sacked. A further period of depression back home in England with family followed and he then moved to Australia. This was the first time that he had been prescribed lithium and his wife ensured that he took it every night. He worked successfully for 15 years and restored his affluent lifestyle. He had a talk with his doctor about the lithium and they both decided that he had been well for a long time and could reasonably give it up now. Within a month he had redecorated and re-carpeted his office in yellow. He attended court, not in his usual sober suit, but in a tee-shirt with a cork hanging hat. He was quickly suspended and flew off to Penang first class to gamble, drink and carouse. He was arrested and deported for not paying his hotel bill. On return to Australia he was committed to hospital, for which he never forgave his wife and they separated. He lost his job, his home and his wife and still lives in a hostel near the beach–he has started to take the lithium again.

5.26 Does lithium really reduce the rate of suicide?

There are two reasons why we think that lithium prevents suicides:

1. Lithium is effective in reducing the burden of illness in manic depression by both treating and preventing mania and depression. A reduction in the frequency and severity of depression would be likely to reduce the rate of suicidal acts.

2. There may be a specific effect of lithium on reducing suicidal thoughts and impulses which is why it has been tried in other groups who are known to be suicidal (e.g. those who frequently self-harm such as those with borderline personality disorder).

The research on non-bipolars is unclear. Anecdotally lithium does seem to be effective for some patients but many doctors feel it is very risky giving a medication which is known to be very dangerous in overdose to patients who regularly take overdoses!

Even among bipolars it is difficult to interpret the studies which indicate that the suicide risk is much lower on lithium. The patients in these studies who are most compliant and reliable in taking lithium may also be those who are least likely among the manic depressives to be impulsive and suicidal, so that comparison of those on and off lithium is not comparing like with like. However, post mortem inquiries into patients with manic depression who have committed suicide show that they are unlikely to have been taking or even been prescribed lithium. Overall, it is likely that lithium does have a strong benefit in reducing the risk of suicide if it is taken in the long term (Schou 1998).

5.27 Does stopping lithium suddenly lead to relapse?

Probably the best way to make bipolar patients manic is to treat them with lithium for a few months then suddenly stop it. About a third of patients will relapse, usually into mania rather than depression, in the following few weeks (Fig. 5.4). This is a rebound effect in that patients do not return to the underlying natural history of manic depression but ‘rebound’ beyond this to a much higher chance of recurrence.

Fig. 5.4 Sudden lithium withdrawal causes rapid relapse in bipolar disorder (usually into mania).

(Based on data from Suppes et al 1991.)

If lithium is stopped gradually over 4-6 weeks then the risk of relapse is less than with sudden cessation and patients return to the risk associated with the natural history and avoid the rebound. However, many doctors would stop the treatment even more slowly, so that any emerging symptoms can be responded to by increasing the dose of lithium again.

If a patient is well on lithium there needs to be a very good reason for stopping treatment–manic depression does not go away. If the patient is well it is likely that this is because the treatment is working effectively. There is a strange paradox in the treatment of many medical conditions in that those who are doing worst (and presumably getting the least benefit from the treatment) are those that we most encourage to continue and take higher doses, whereas those who are well (and presumably getting the most benefit) are often encouraged (or not discouraged) to stop (see Case vignette 5.2)

5.28 If the patient stops lithium and suffers a relapse, does it work as well the second time around?

There has been concern that the effect of lithium is reduced when it has been stopped suddenly and a relapse has occurred. This is a difficult question to answer as we know that the more episodes a patient has, and the shorter the well interval, the worse the prognosis. If a patient had an extra episode through stopping lithium then their prospects over the next year would be less good than if they had avoided this relapse. However, if there has previously been a good response to lithium then it is definitely worth going back to this treatment again after a relapse, and not shifting on principle to a different medication.

5.29 Why do manic depressives who are doing well on lithium stop taking it?

Doctors usually expect that the main reason that patients stop taking medications is either because the treatment is not effective or that side-effects are unacceptable. Given that everyone on lithium can expect to experience some side-effects and treatment is only partially effective, the assumption is that this is why some patients stop. However, the common scenario is of a bipolar patient who is doing well on the treatment and having only minor side-effects deciding to stop because he does not think that he is going to be ill again. It is very hard for patients to accept that they have little control over their mood and that even if they have their life on track and no serious stressors, they are still liable to a recurrence. It is common to hear comments such as: ‘When I got ill before I was having particular problems at work and I’ve sorted them out now, so I just don’t think it will happen again.’ It may well be true that they are helping to protect themselves by sorting out the problems at work but it is likely that it is not sufficient to prevent a relapse.

Of course the usual reasons (lack of efficacy and side-effects) apply not only to lithium but also to other long-term treatments and there are other associations with non-compliance (see Q 5.43) including personality factors and substance misuse.

5.30 Are there particular illness patterns which are likely to improve with treatment?

Finding patients who are particularly likely to benefit from treatment with lithium is not generally possible. Family history can be useful; if you know that someone else in the family has done well with lithium then it would in all probability be beneficial to move the patient to this treatment sooner rather than later. The pattern of illness–mania into depression and then recovery–is thought to be a pattern particularly responsive to treatment with lithium.

OTHER LONG-TERM TREATMENTS

5.31 Is valproate a useful long-term treatment in bipolar illness?

Mania can be effectively treated with valproate (see Q 4.4). It relieves the symptoms of mania as effectively as any other treatment and because it is generally well tolerated the dose can be increased rapidly even in patients who have never taken it before. Valproate may be the best treatment to give to those who are in a mixed episode (a simultaneous combination of mania and depression–see Q 1.11).

Many patients are treated with valproate as a long-term preventive measure. However, it is much less certain if this is an effective treatment and as the evidence currently stands it should not be given as a first line preventive treatment. On the other hand, it should be considered in those who are unresponsive to lithium and may be more effective than lithium in those who are in a rapid cycling phase of the illness. The doses needed for long-term treatment are similar to those used in the acute treatment of mania, but it is usually introduced more slowly as the clinical situation is usually less urgent. The medication is usually started at 200 mg twice a day and increased gradually over a month to 1200 mg/day. Sometimes higher doses of up to 2 g/day are needed. Although it is recommended that it is taken twice a day this can limit compliance and for those who have great difficulty in taking it this often, consideration should be given to prescribing it all in one dose of a slow-release version of valproate.

5.32 Are there any problems with taking valproate?

Valproate is generally a well-tolerated drug. The only common side-effects are gastrointestinal with nausea and occasionally vomiting. Less common side-effects are neurological with tremor and unsteadiness. Rarely it can cause liver dysfunction, though this seems to occur mainly in children, and also rarely pancreatitis.

5.33 Is there a difference between sodium valproate and valproate semi-sodium (Depakote)

The basic drug in these two formulations is the same: the semi-sodium formulation (Depakote) is a combination of valproic acid with sodium valproate, hence the semi- or half-sodium. It is absorbed as a combined molecule and later broken up into valproate. Virtually all of the trials of valproate in manic depression have been with Depakote (Bowden et al 2000, NICE Technology Appraisal 66). Therefore we can be confident that Depakote is an effective drug whereas the same cannot be said of other formulations of valproate. Valproate semi-sodium is a licensed treatment for mania although it currently does not have a licence in the longer term treatment of bipolar illness.

To date there have been no head-to-head trials in treating manic depression between valproate and semi-sodium valproate and so we do not know if they have the same clinical efficacy.

Valproate semi-sodium is better tolerated by the stomach than other formulations though in fact the slow-release version of sodium valproate is also a well-tolerated drug. The fact that Depakote is particularly well tolerated can be very useful in the initial stages of the treatment of mania because the dose can be escalated rapidly. Dose for dose Depakote will produce a higher blood level than other formulations of valproate.

5.34 Is carbamazepine useful in the treatment of bipolar illness?

Several anticonvulsants have been tried in the treatment of manic depression and carbamazepine has been used extensively over the last 10 years for this indication. Carbamazepine does have a licence in the UK in the treatment of manic depressive illness unresponsive to lithium. Carbamazepine is an effective antimanic treatment and is also quite sedative, particularly when it is first given, which can be an additional benefit in the early treatment of mania. The usual starting dose is 200 mg twice a day increasing up to 1 g, although sometimes higher doses are needed. Some studies showed carbamazepine to be an effective longer term preventive treatment but recent research showed that it was a less effective treatment than lithium for standard manic depressive illness (Greil et al 1997). It still has a place in the acute and long-term treatment of manic depression but should not be used in preference to lithium. However, it can be a useful alternative if lithium is neither suitable nor effective. The study by Greil et al (1997) showed that it was not as good as lithium generally, but did suggest that it may be a better treatment for schizoaffective manic depression (see Q 1.15)

5.35 Are there any problems in using carbamazepine?

The common side-effects of carbamazepine, particularly in the early stages of treatment, are sedation and gastrointestinal problems such as nausea. However, in the longer term it is usually a well-tolerated drug. One problem is that it induces liver enzymes: this means that carbamazepine will induce its own metabolism and sometimes the benefits of carbamazepine can wear off because of this. Commonly the enzyme induction is apparent in liver function tests so that the gamma glutamyl transferase (γGT) levels are raised. It is easy to attribute rises in γGT to being alcohol induced which is also a common complication of bipolar illness, but beware of implying this in those who are actually abstinent!

Carbamazepine also induces the metabolism of other drugs. The one that most commonly causes problems is the oral contraceptive as carbamazepine can render the oral contraceptive ineffective.

Carbamazepine also induces allergic reactions, most commonly rashes, in about 15% of those who take it. If this occurs it is usually impossible to continue with the drug and is also potentially dangerous as more serious allergic reactions can also occur including hepatitis and agranulocytosis.

5.36 Does lamotrigine have a place in the long-term treatment of manic depression?

Recent studies on lamotrigine in manic depression have been very encouraging and it seems likely that this will prove to be a more frequently used and beneficial treatment (Calabrese et al 2003). One of the main problems in treating bipolar illness is that depression is often chronic, pervasive and disabling. Lamotrigine has been shown to be an effective antidepressant in bipolar illness with only a limited propensity for switching into mania. The usual treatment dose is 200-300 mg but it can only be very slowly increased up to this level (see Q 5.37). This means that it is difficult to use lamotrigine in very acute depression but it can still be very useful in tackling longer term depressive symptoms. It is likely that it also has some benefit in the prevention of depression though probably not in the prevention of mania. For this reason, for those patients who have predominantly a depressive picture it is useful to consider this treatment.

5.37 Are there any problems in using lamotrigine?

Lamotrigine has a considerable propensity for inducing allergic reactions and these are potentially life threatening. It is for this reason that a very slow introduction of the drug is needed, the usual regimen being 25 mg daily for 2 weeks, increasing to 50 mg daily for a further 2 weeks, then increase by 50 mg daily every 2 weeks up to 200 mg. Sometimes higher doses are needed and can be effective, at least up to 300 mg daily. The usual allergic reaction is a rash, though sometimes this can be of the Stevens-Johnson type which includes inflammation of the mucous membranes including the eyes. This rash can be so severe that hospitalisation and treatment with steroids is required. Very rarely fatalities have occurred. It does seem a pity that of all the anticonvulsants lamotrigine seems to be the most helpful for depression but is potentially the most dangerous.

In terms of the more predictable side-effects lamotrigine is relatively free of these and is a well-tolerated drug; however, neurological side-effects such as headache and double vision can occur.

5.38 Have any other anticonvulsants been found to be helpful in manic depression?

Because some anticonvulsants do appear to be helpful in manic depression many have been tried but none can currently be recommended for use apart from in exceptional circumstances. It is relatively common to find in manic depression (which is such a changeable disorder) that there are initial reports of benefit with new treatments but careful controlled testing does not bear out these out. Trials of gabapentin are a good example of initial reports of success which turned out not to be confirmed. Lithium also tends to come out badly in initial comparisons because the patients who are tried on new treatments are usually those who are not getting on well and so are a preselected group of lithium non-responders. Lithium has stood the test of time and it will be interesting to see which of the currently used treatments are still recommended 40 years hence!

5.39 Should anyone be taking two drugs to prevent recurrence?

If you have treated many people with manic depression you will know that there is a substantial proportion of patients who are continuing to experience recurrences or chronic symptoms despite being on one or other of the long-term preventive treatments. If this is happening then it is likely that you will be considering combining medications. Probably the most common combination is lithium with an antipsychotic, though it may be that the antipsychotic is being used as symptomatic relief for anxiety or insomnia.

There is very little information currently about whether combinations of long-term treatment are better than monotherapy. It could be that actually we should be treating bipolars with combination treatment from early on as this may lead to a better outcome, but at the moment we just do not know (see Balance Trial http://www.psychiatry.ox.ac.uk/balance/index.html).

A common tactic is to add valproate to lithium if there has been a partial response to lithium but the patient still has continuing symptoms or episodes. If this leads to a major improvement then consider withdrawing the lithium to get back to a single medication again. However, there is little scientific basis for this decision and it would be as reasonable to continue with both medications if they were both well tolerated.

5.40 Are antipsychotics useful in the long term?

A large proportion of patients who suffer from manic depression are treated with antipsychotic drugs in the long term. The reasons for this are not entirely clear, but presumably the clinicians prescribing them and the patients taking them are finding them helpful. It seems common sense that a treatment that relieves the symptoms of acute mania should also be able to prevent these symptoms returning, but there have been very few studies to test out this theory, probably because we have focused on lithium treatment.

The reasons why antipsychotics are given are probably to relieve anxiety or improve sleep, through their sedative or calming effect. Those who suffer from psychotic symptoms during mania or depression, or those with a schizoaffective disorder who experience psychotic symptoms between episodes, may find antipsychotic drugs helpful in the long term.

A few people with manic depression have very little insight and do not take regular oral medication and so, as a last resort, a depot antipsychotic is prescribed. It is likely that this does have some benefit but it would not be recommended unless other treatments were not suitable or effective.

Olanzapine is one of the few antipsychotics that has been studied in the longer term in bipolar illness (Tohen et al 2004). It has been shown to be effective in reducing manic relapse when it is continued after it has been successfully used to treat mania, and may also have some benefit in preventing depression. Some people gain weight when they are taking olanzapine and this should be tackled right from the start of treatment. It seems to cause weight gain mainly through an increase in appetite and warning patients about this and encouraging them to limit their intake of sweet foods is important. If it were not for this problem it is likely that olanzapine would be even more widely prescribed to bipolars than it already is.

Other atypical antipsychotics (e.g. quetiapine, risperidone) have proved beneficial in the acute treatment of mania and it is likely that they will also prove effective in longer term treatment.

5.41 Are there any problems with taking antipsychotics in the long term?

Antipsychotics do cause side-effects in the long as well as the short term. The traditional antipsychotics cause anticholinergic effects including dry mouth and constipation. They also increase prolactin levels which affect menstruation, libido and bone density. Additionally, they cause movement disorders and the major problem with the long-term use of antipsychotics is the development of persistent abnormal movements–tardive dyskinesia. These are repetitive movements, usually of the face and particularly the mouth, but which can also affect other parts of the body. They are not usually distressing but are very noticeable to others so that they can be very stigmatising. The movements may reduce when the drugs are stopped but often they do not go away.

The newer (atypical) antipsychotics are much less likely to cause tardive dyskinesia; however they can still affect prolactin levels. Weight gain can be a problem with both traditional and atypical antipsychotics.

5.42 Are there special considerations for long-term treatment with lithium in the elderly?

The same long-term treatments are as appropriate for the elderly as for other adults. Complications can arise with lithium treatment because renal function slowly decreases with age. Because lithium is a salt it is not metabolised by the body but is entirely excreted by the kidney (with a small amount lost through sweat). Before starting lithium in the elderly you need to be confident that renal function is adequate, usually by undertaking a creatinine clearance test. However, the only real test is to start a small dose of lithium and check the levels while monitoring for adverse effects. The dosage of lithium is almost invariably lower, usually starting at 200 mg at night rather than 400 mg. The eventual dose is often 400-600 mg, whereas younger adults usually take 800-1200 mg daily.

Side-effects of lithium (see Q 5.7) are more common in the elderly, but it is as effective in this age group as for others. Similar considerations are needed for other long-term medications, where doses usually need to be lower and side-effects are often more evident.