Treatment of acute episodes

There is good evidence that lithium is efficacious in the treatment of the acute phase of mania, although the addition of major tranquillisers may be clinically indicated for more severely ill patients (American Psychiatric Association, 1994b ). Although less stringently evaluated than lithium, the current evidence suggests that carbamazepine and valproate have equivalent efficacy to lithium in the treatment of manic episodes. Research on pharmacotherapy for acute bipolar depression is less common and harder to interpret because of methodological difficulties (Reference Prien and PotterPrien & Potter, 1990).

Relapse prevention

What about the prevention of relapse and recurrence? How well do mood stabilisers prevent further bipolar episodes? Studies in the 1960s and 1970s suggested that lithium was an effective form of prophylaxis. However, reviews of more recent randomised controlled double-blind trials of mood stabilisers have suggested that these medications are not as effective at preventing relapse as initially reported (Reference MoncrieffMoncrieff, 1995; Reference Solomon, Keitner and MillerSolomon et al, 1995; Reference Silverstone, McPherson and HuntSilverstone et al, 1998). Whereas the original studies in the 1960s reported that patients on lithium experienced relapse rates of 35% over 2 years, more recent studies find average relapse rates of 50% over a 2-year follow-up. Furthermore, a significant proportion of patients with bipolar disorder do not seem to benefit from mood stabiliser medication. Reference Prien and PotterPrien & Potter (1990) estimated that lithium was ineffective in up to 20–40% of cases, either because of inadequate response or poor compliance due to the side-effects of medication. Although further research trials are required to enable a reliable determination of the degree of efficacy for other medications such as carbamazepine and valproate, their efficacy appears to be similar to that of lithium.

Several particular subtypes of bipolar depression have been identified as being less responsive to lithium. The rapid-cycling subtype, that is, having four or more mood episodes per year, seems to be less responsive to lithium treatment. Similarly, mixed affective state, that is, the simultaneous occurrence of both manic and depressed symptoms, is a predictor of poor response to lithium. Patients with bipolar disorder with comorbid personality disorders or high levels of neuroticism have also been found to respond poorly to the medication. Patients with higher numbers of previous episodes are also less responsive to lithium treatment. It is still an open question whether the factors that predict poor response to lithium treatment are specific predictors of poor prognosis to this particular medication or general predictors of a poor response to any treatment. If bipolar disorder has any similarities to unipolar depression, some of these factors might be expected to predict poor response to any treatment, in the same way that chronicity of depression and number of previous episodes in unipolar depression predict poor outcome for both pharmacotherapy and psychotherapy.

It is worth noting that combined medication (e.g. Reference Calabrese, Markovitz and KimmelCalabrese et al, 1992; Reference Denicoff, Smith-Jackson and DisneyDenicoff et al, 1997; Reference Frye, Ketter and LeverichFrye et al, 2000), i.e. lithium plus other mood stabilisers such as carbamazepine or valproate, appears to be superior to monotherapy in achieving prophylaxis. Reduced relapse for adjunctive medication on top of lithium points to a biological basis for the improvement in outcome, consistent with the view that the origins of bipolar depression are biological. Nonetheless, there are still relatively high rates of relapse in those on combined medication. For example, Reference Denicoff, Smith-Jackson and DisneyDenicoff et al(1997) reported that combined lithium and carbamazepine had some success in only 50% of patients with rapid-cycling bipolar disorder.

Thus, although pharmacotherapy for bipolar disorder has produced significant benefits both in the treatment of acute episodes and in relapse prevention, there are still high rates of relapse and a clear need for the development of improved treatment approaches for bipolar disorder.

Bipolar prodromes

Prodromes are the early signs and symptoms that can precede a full-blown episode. Recent research has suggested that there are common prodromes for both mania and depression and that patients with bipolar disorder can reliably identify their prodromes (Reference Molnar, Feeney and FavaMolnar et al, 1988; Reference Smith and TarrierSmith & Tarrier, 1992; Reference Lam and WongLam & Wong, 1997; Boxes 1 and 2). It is worth noting that patients find it easier to recognise prodromes for mania than for depression, possibly because the onset of mania is often more acute and more dramatic, whereas bipolar depression seems to develop more insidiously. There appears to be a relationship between residual symptoms and prodromal symptoms (Reference FavaFava, 1999) which, given that residual symptoms predict future outcome in depression, further indicates the value of identifying and intervening at the level of these symptoms (e.g. Reference Fava, Rafanelli and GrandiFava et al, 1998).

Importantly, patients with bipolar disorder display different coping strategies to deal with their prodromes, with the type of coping approaches chosen in response to manic prodromes predicting the level of social functioning and number of relapses 18 months later (Reference Lam, Wong and ShamLam et al, 2001). In general, helpful coping strategies involve behavioural activities that act against the prevailing prodrome, whereas unhelpful coping strategies further reinforce the initial stages of the bipolar episode, fuelling more-extreme mood swings. Thus, for mania prodromes, engaging in calming activities, increasing rest, reducing stimulation and decreasing activity would be useful strategies, whereas increasing activity levels, enjoying the ‘high’, and ‘making up for lost time’ would be unhelpful strategies likely to increase the risk of a full-blown manic episode. Similarly, for depression prodromes, keeping busy and maintaining routines are associated with better outcomes, whereas cutting down on activities, withdrawing from other people and going to bed are associated with worse outcomes. For both manic and depressed prodromes, seeing a doctor and increasing the appropriate medication are also useful coping strategies.

Each patient will have his or her own idiosyncratic prodromes so, clinically, it is advantageous to review the weeks preceding each distinct bipolar episode to determine the signs and symptoms that occur as the episode develops. Separating the prodromes into early, middle and late signs is a useful way of structuring the chronology of an episode's development, with late signs reflecting the symptoms of the full-blown episode and early signs reflecting the earliest possible indication of a change in mood state. For example, for one patient, an increase in witticisms and puns may be an early sign of an impending manic episode, while for another it may be that colours appear sharper and more vivid. Because a number of prodromes occur in a social context, involving friends and families in their identification and monitoring can be very productive. The research evidence suggests that helping patients with bipolar disorder to identify their prodromes and teaching them straightforward responses to counteract them can be beneficial in reducing the risk of relapse. Since the prodromes for mania can precede a full bipolar syndrome by several weeks (Reference Molnar, Feeney and FavaMolnar et al, 1988; Reference Smith and TarrierSmith & Tarrier, 1992) early detection followed by effective coping strategies can be very helpful in ‘nipping an episode in the bud’.

Life events and routine

There is evidence that life events and routine play an important role in the period prior to the onset of a bipolar episode. Reference Malkoff-Schwartz, Frank and AndersonMalkoff-Schwartz et al(1998) reported that in the 8 weeks prior to manic episodes, a significantly greater proportion of patients had experienced events that disrupted social rhythm than in an 8-week episode-free control period. Such events involve changes to daily or social routine, in particular disruptions of sleep or changes in working schedule. Disruptions to circadian rhythms have been implicated in the onset of bipolar episodes (Reference Wehr, Sack and RosenthalWehr et al, 1983; Reference Healy and WilliamsHealy & Williams, 1989), and there are reports of patients who relapsed after long-distance travel or jet lag (Reference Jauhar and WellerJauhar & Weller, 1982). The American Psychiatric Association (1994b) recommends regular social and sleep routines within their treatment guidelines for bipolar illness.

Modifying cognitive therapy for bipolar disorder

The psychosocial literature reviewed above suggests that helping patients with bipolar disorder to monitor their prodromes, develop new coping strategies and improve their daily routine, as well as using appropriate psychotropic medication, would help to reduce their risk of relapse. Structured, short-term, problem-focused therapies such as cognitive therapy are particularly suited to developing these skills. Cognitive therapy is built on the view that thinking, mood and behaviour all have a reciprocal influence on each other. Therapists aim to teach patients to become more aware of their thoughts and behaviours and to then alter dysfunctional thoughts and behaviours. For example, in cognitive therapy for depression, patients are taught to challenge their self-critical thoughts such as ‘I am a failure’ by looking for evidence and alternative explanations that do not fit the negative interpretations (Box 3). Similarly, patients with depression are encouraged to increase their levels of pleasurable and achievement-related activities, have more structured routines and reduce unhelpful patterns of behaviour such as social withdrawal and avoidance. This combination of monitoring and behavioural change is complementary to the psychosocial research on prodromes and routine in patients with bipolar disorder, indicating that cognitive therapy may be a useful approach in such cases.

Attempts to develop cognitive–behavioural approaches for bipolar disorder have been based within a diathesis–stress model. This approach proposes that there is an inherent biological vulnerability in bipolar illness, as a function of instability of circadian rhythms and of impaired regulation of the motivational system controlling approach and reward (Reference Depue and IaconoDepue & Iacono, 1989). It is proposed that disruption of sleep and routine can lead to dysregulation of circadian rhythms and the motivational system, leading to prodromal symptoms. The exact nature of life events and social disruptions may determine the specific prodromal symptoms that occur. For example, events involving goal attainment are associated with the development of manic symptoms (Reference Johnson, Meyer and WintersJohnson et al, 2000a ), whereas negative life events, low social support and low self-esteem predict the development of depressive symptoms (Reference Johnson, Meyer and WinettJohnson et al, 2000b ). The cognitive–behavioural approach to bipolar disorder then predicts that the particular coping strategies adopted in response to prodromal symptoms will determine whether a full bipolar episode will occur, which is more likely if coping strategies are poor. Finally, the consequences of a bipolar episode may further contribute to the maintenance of the episode. For example, impulsive spending may lead to financial problems; irritability coupled with poor concentration may lead to problems at work or the loss of employment; and promiscuous behaviour may lead to problems in intimate relationships. All of these episode-related difficulties could then act as further stressors to interact with the underlying biological vulnerability, to further generate bipolar symptoms.

The recognition of a biological vulnerability within the cognitive–behavioural approach to bipolar disorder, coupled with the identified need to improve relapse prevention, means that cognitive therapy for bipolar disorder has been conceptualised only as an adjunct to medication rather than as a replacement for it. Furthermore, cognitive therapy has been designed as a relapse prevention treatment, not as an acute treatment. Thus, attempts to modify cognitive therapy for bipolar disorder have concentrated on working with patients with the illness in a relatively euthymic state, outside of a major depression or manic episode. Clinical experience suggests that attempting cognitive therapy with patients with bipolar disorder who are in a manic episode is difficult and unlikely to be effective. Therefore, for maximum benefit, the cognitive therapy approach needs to be applied when patients are in remission between episodes.

To date, there have been several attempts to examine the efficacy of adding cognitive therapy to medication for bipolar disorder. Reference Lam, Bright and JonesLam et al(2000) and Reference Scott, Garland and MoorheadScott et al(2001) have reported that the addition of cognitive therapy to routine mood stabilisers produced significantly reduced relapse compared with routine mood stabilisers alone. However, both of these were pilot studies with small samples, such that any conclusions from them need to be tentative. Reference Fava, Bartolucci and RafanelliFava et al(2001) reported that cognitive–behavioural therapy reduced residual symptoms in patients with bipolar disorder who relapsed on lithium and may improve lithium prophylaxis. However, this study had a small sample size (n=15) and there was no comparison treatment group such as treatment as usual. In a randomised controlled trial, Reference Perry, Tarrier and MorrissPerry et al(1999) taught patients with bipolar disorder to detect early warning signs and seek medical help earlier than those on treatment as usual. Although not a trial of a full cognitive therapy package, this study did demonstrate that teaching improved monitoring and better coping skills could significantly reduce bipolar relapse. Reference Zaretsky, Segal and GemarZaretsky et al(1999) found that cognitive therapy could be helpful in treating episodes of bipolar depression, although this study was also limited by having only a small sample.

More recently, a full-scale randomised controlled trial of cognitive therapy for bipolar disorder has been completed (Reference Lam, Watkins and HaywardLam et al, 2003). This study investigated the efficacy of cognitive therapy in conjunction with routine mood stabilisers in the prevention of bipolar episodes over a 12-month period. It recruited 103 patients with bipolar I disorder, who were being prescribed prophylactic medication at a recommended dose and who were not currently in an acute bipolar episode. To ensure that the patients treated were a group that was highly vulnerable to relapse/recurrence despite the provision of adequate levels of mood stabilisers, all those included in the trial had to have experienced at least two bipolar episodes in the previous 2 years or three episodes in the previous 5 years.

Patients were randomly allocated either to a treatment-as-usual group or to a combined medication and cognitive therapy group. The treatment-as-usual group received mood stabilisers at a recommended level with regular psychiatric follow-up as out-patients. The combined medication and cognitive therapy group received treatment as usual plus cognitive therapy.

The cognitive therapy was based on a previously developed treatment manual (Reference Lam, Jones and BrightLam et al, 1999) and consisted of 12–18 weekly sessions, followed by 2 booster sessions each after 3-month intervals. The cognitive therapy package was developed from the classic cognitive therapy approach for depression (Reference Beck, Rush and Shaw1Beck et al, 1979), adapted to take into account psychosocial research into bipolar disorder (Box 4). Teaching better coping responses to prodromes is one core element of cognitive therapy for bipolar disorder. Identifying and challenging dysfunctional thoughts and beliefs is another. Depression-related thoughts and beliefs are challenged as per the standard cognitive therapy for unipolar depression.

In addition to challenging negative thoughts, patients can challenge dysfunctional positive thoughts associated with mania. For example, individuals can be taught to examine grandiose thoughts such as ‘I know better than everyone else’ and check whether these accurately reflect their past experience and whether such thoughts are dependent on their mood. Similarly, patients can be encouraged to examine the longer-term costs and benefits of their more-impulsive thoughts (e.g. ‘I need to buy that now’), to prevent overimpulsive behaviour. Patients can also be encouraged to examine grandiose and impulsive plans across a longer time frame (‘How many of your ideas still seem good a week later? If this is genuinely a good idea, it should still be a good idea next week. Can you try to leave it for a week?’). Highly driven and extreme goal-attainment beliefs, e.g. ‘I should be happy all the time’, ‘If I put in enough effort, I should be able to achieve everything I want’ are hypothesised to be vulnerability factors for mania. Patients with bipolar disorder who have these attitudes are more likely to engage in extreme goal-pursuing behaviour likely to disrupt their sleep and daily routines, precipitating further episodes. Cognitive therapy can involve identifying such beliefs in collaboration with patients and then exploring how realistic and useful these beliefs are.

Overall, Reference Lam, Watkins and HaywardLam et al(2003) found that, at 12 months, the combined medication and cognitive therapy group had significantly fewer bipolar episodes compared with the treatment-as-usual group (cumulative relapse 44% v. 75%), fewer days in bipolar episodes (27 v. 88) and significantly fewer hospitalisations. Those in the combined medication and cognitive therapy group were also significantly better than the treatment-as-usual group at coping with mania prodromes at 12 months, although there was no significant difference between the groups in coping with depression prodromes.

One area that requires future research concerns the impact of cognitive therapy on medication compliance. After 6 months, a significantly greater proportion of patients in the combined medication and cognitive therapy group (88%) than in the treatment-as-usual group (67%) reported missing their medication fewer than three times in the previous month. However, serum levels were available for only 50% of the sample, limiting the power of any calculation to test whether differences in compliance were reflected in serum levels. Never the less, more patients in the combined medication and cognitive therapy group (93%) met the criterion of an adequate serum level of mood stabilisers than in the treatment-as-usual group (78%). At this stage, it is not possible to tell to what degree cognitive therapy reduced bipolar relapse simply by improving compliance with mood stabiliser medication or by other mechanisms such as improving coping skills.

The principal limitation of the Reference Lam, Watkins and HaywardLam et al(2003) study was that there was no control treatment for the increased level of face-to-face contact that the cognitive therapy patients received compared with the treatment-as-usual group. Therefore, it is not possible to determine whether the improved outcome was due to specific aspects of cognitive therapy or due to non-specific therapy effects such as increased attention and support. Nonetheless, the study demonstrated that the addition of cognitive therapy to medication for bipolar disorder has benefits in reducing the frequency and durations of relapse/recurrence and in reducing the extent of hospitalisation.