REVIEW ARTICLE
Drugs 44 (3): 326-335. 1992 0012-6667/92/0009-0326/$05.00/0 © Adis International Limited. All rights reserved. DRU1 186
New Uses of Anticonvulsant Drugs in Psychosis Charles Van Valkenburg, 1 John C Kluznik 2 and Robert MerrilP 1 Nevada Department of Prisons, Carson City, Nevada, USA 2 Minnesota Security Hospital, St Peter, Minnesota, USA 3 Saint Peter Regional Treatment Center, St Peter, Minnesota, USA
Contents 326 327 327 327 328 328 32 328 328 329 330
331 331 332 332 332 332 332
Summary
Summary I. Schizophrenia 1.1 C10nazepam 1.2 Carbamazepine J.3 Valproic Acid 1.4 Phenytoin 1.5 Water Intoxication 2. Bipolar Disorders 2.1 C10nazepam 2.2 Carbamazepine 2.3 Valproic Acid and Phenytoin 3. Psychoses Secondary to Epilepsy 4. Other Organic Psychoses 5. Borderline Disorder 6. Anticonvulsant Drug-Induced Psychoses 7. Anticonvulsant Drug Withdrawal Psychoses 8. Neuroleptic-Induced Seizures 9. Conclusions
Psychotic patients not adequately relieved by neuroleptic drugs often improve when anticonvulsants are added. In bipolar disorders and organic psychoses, anticonvulsants can sometimes be used to replace neuroleptics. No individual anticonvulsant is clearly, consistently superior. Patients who fail on one agent may improve on the next. Clonazepam is an excellent adjunct to neuroleptic therapy, but there is little evidence that it is effective as monotherapy. However, it is safe, sedates rapidly, and has an excellent patient tolerability profile. Carbamazepine is the best established drug for patients with bipolar disorders, particularly for rapid cyclers, and is often effective monotherapy. The therapeutic profile of valproic acid (sodium valproate) is similar to that of carbamazepine, but its side effects are quite different and are often preferred. Other anticonvulsants are little studied, but might be chosen to avoid certain side effects, or after better-studied drugs have failed. The pharmacological basis behind using anticonvulsants in psychoses is primarily empirical. In almost every case it has been clinicians who have first noted the beneficial effects of these drugs. Theories such as that of Post have followed.
New Uses for Anticonvulsants
327
We have come about as far as we can with neuroleptics alone. Apart from some possible reported benefits of clozapine in refractory cases of schizophrenia, we are little better off than we were in the 1950s. It has now also been adequately established that neuroleptic medications have therapeutic ranges (van Putten et al. 1991), and that exceeding these ranges offers no benefit, and may in fact make patients worse. This leaves us with patients whose responses to antipsychotic medications are inadequate. We have learned that responses to electroconvulsive therapy (ECT), antidepressants and lithium are far less specific than nosologists would wish, and that they leave many patients still psychotic. Anticonvulsant drugs, usually added to neuroleptics or to lithium, sometimes used alone, have us on the move again. This article reviews the evidence of anticonvulsant drug efficacy in various psychoses (table I). Evidence is stronger for some than for others, but no systematic direct comparison has shown one agent superior to another. There still appears no good way to predict which anticonvulsant will be best for an individual patient. We have often seen one drug work after another failed.
1. Schizophrenia Treating schizophrenia with neuroleptic agents rarely results in complete remission, but that does not mean that schizophrenic patients cannot be considerably helped with other approaches. 1.1 Clonazepam Clonazepam augments the effects of neuroleptics in schizophrenia, makes them work faster, and reduces their most unpleasant side effects such as akathisia, rigidity and dystonia (Altamura et al. 1987). It also reduces neuroleptic-related postural myoclonus (Fukuzako et al. 1990), occulogyric crises (Horiguchi & Inami 1989), akathisia (Kutcher et al. 1989), blephaspasm (Ballard 1989), and 'painful legs and moving toes' (Sandyk 1990). Clonazepam monotherapy of schizophrenia appears ineffective (Karson et al. 1982). 1.2 Carbamazepine Carbamazepine has usually been given in combination with a neuroleptic rather than as a sole treatment. It appears to augment the effects of neuroleptics, reduce side effects and allow a lower neu-
Table I. Evidence of anticonvulsant efficacy in various psychoses. This table does not reflect relative efficacy, which is unknown,
but rather reflects the strength of current evidence of efficacy Clonazepam
Carbamazepine
Valproic acid
Phenytoin
Schizophrenia Symptoms most likely to respond
N:+++ Excitement, negative symptoms, dystonia, rigidity, akathisia
N:++ Psychosis, hostility
N:+ EXCitement, hostility, polydypsia
Manic disorder Rapid-cycling bipolar Bipolar prophylaxis Epileptic psychosis Supersensitivity psychosis
La: ++++ ? ? M:++ ?
N:+++ Excitement, tension, negative symptoms, suspiciousness, uncooperativeness M:++++ M:++++ M:++++ M:+++ M:++
M:++++ M:++++ M:++++ M:+++ ?
M:++ ? ? M:+ ?
a
Reported as effective monotherapy during first week. = effective as monotherapy; N = effective as adjunct to neuroleptic; L = effective as adjunct to lithium; ++++ = excellent efficacy; +++ = good efficacy; ++ = moderate efficacy; + = poor efficacy; ? = unknown, or equivocal results.
Abbreviations and symbols: M
Drugs 44 (3) 1992
328
nowsky & Putman 1943), but no other SUbtypes. Kubanek and Rowell (1946) found that one-third of patients were greatly improved by phenytoin. In a comparison between phenytoin and placebo in 76 chronic schizophrenics whose neuroleptic treatment had been stopped, phenytoin tended to diminish hostility and unusual thought content in the fIrst month only, and increase motor retardation and blunted affect in the second (Simopoulos et al. 1974). Phenytoin indeed contributes, with other organic factors, to cognitive decline and intellectual deterioration (table II) [Trimble 1988].
roleptic dose. Wetterling (1987) found that carbamazepine improved Brief Psychiatric Rating Scale items of excitement and tension, but not the schizophrenic defect, in 14 male chronic schizophrenics. However, Herrera et al. (1987) found that carbamazepine significantly improved negative (or defect) symptoms in 6 schizophrenics who had failed to improve on at least 2 chemically dissimilar neuroleptics. Similarly, Okuma et al. (1989a,b) demonstrated that 29 of 54, and.61 of 127 patients improved markedly or moderately on carbamazepine; suspiciousness, uncooperativeness, and excitement were particularly improved. Carbamazepine reduced neuroleptic side effects of haloperidol in 10 paranoid schizophrenics, without significantly changing clinical response (Martin-Munoz et al. 1989). Acute. schizophrenic patients given adjunctive carbamazepine· needed less haloperidol and experienced fewer side effects when given carbamazepine, as an adjunct to other therapy (Dose et al. 1987). These patients deteriorated if their carbamazepine was stopped. However, Carpenter et al. (1991) did not find carbamazepine monotherapy to be superior to placebo in the maintenance of 13 outpatient schizophrenics.
1.5 Water Intoxication Water intoxication is often a complication of schizophrenia. In severe cases it can lead to a syndrome of psychosis, intermittent hyponatraemia and polydipsia (PIP) which can cause seizures, coma, and aggressive behaviour. It is hard to manage, and can be life-threatening. Preliminary evidence shows that clozapine is effective in reducing water intoxication (Kluznik 1991). Phenytoin with lithium improves morning basal sodium levels, and is better tolerated than lithium alone (Vieweg et al. 1988). Carbamazepine and c10zapine both can cause agranulocytosis (table II). Giving them together would be expected to be particularly hazardous, and has been associated with 1 death (Gerson et al. 1991). It should also not be forgotten that anticonvulsants frequently cause birth defects, neonatal toxicity and abstinence syndromes (Sitland-Marker et al. 1989).
1.3 Valproic Acid Several early studies investigating valproic acid (sodium valproate) for schizophrenia were negative or inconclusive. Morinigo et al. (1989) added valproic acid to neuroleptic therapy in 4 treatmentresistant schizophrenics, and after 3 months noted a reduction in positive symptoms and reduced hostile disruptive behaviour. The improvement persisted at follow-up of 4 to 12 months. Sultan et al. (1990) described supersensitivity psychosis as a possible long term side effect of neurolepetic treatment, behaviourally analogous to tardive dyskinesia. They treated 5 -such patients with carbamazepine or phenytoin; all improved, 3 persistently. 1.4 Phenytoin Early studies reported that phenytoin benefitted catatonic schizophrenics (Freyhan 1945), and also 73% of schizophrenics who were 'excited' (Kalli-
2. Bipolar Disorders I
Much more attention has been focused the use of anticonvulsants in bipolar disorders than in schizophrenia. Bipolar disorders usually have a better outcome, and clinical improvement is more obvious. 2.1 Clonazepam Chouinard (1987) noted the advantages of clonazepam over neuroleptics in the acqte and maintenance treatment of bipolar disorder: rapid action,
329
New Uses for Anticonvulsants
Table II. Side effects of anticonvulsants
Common and bothersome
Rare and bothersome
Rare and deadly
Clonazepam
Drug abuse Initial sedation
None
Carbamazepine
Dry mouth Initial sedation Hypotension Weight gain Blurred vision Upset stomach Weight gain Tremor Headache Cognitive impairment Dizziness
Disinhibition Confusion Dizziness Rash Nausea Dizziness
Valproic acid
Phenytoin
a b
Aplastic anaemia Allergy
Balding a Vomiting Drowsiness
Hepatitis b Allergy
Irritability Overgrown gums Slurred speech
Allergy
Usually temporary; and may be reversible with zinc supplement. May be limited to infants also taking phenobarbital.
relative lack of toxicity, and no risk of tardive dyskinesia. He suggested treating acute mania with clonazepam alone for the first week, then adding lithium, so that neuroleptics could be avoided (Chouinard 1988). Edwards et al. (1991) showed clonazepam to have a specific anti manic effect, unrelated to its sedative effect, but no specific antipsychotic effect in acute manic patients. De la Gandara Martin et al. (199 I) found adding clonazepam to be beneficial in 85% of cases. However, Aronson et al. (1989) found clonazepam ineffective in maintenance treatment of bipolar disorder, although Mauri et al. (1990) suggested that clonazepam prophylaxis was effective in patients who could not take lithium. After sedation cleared, side effects were minimal. As with all sedative hypnotics, there is a risk of behavioural disinhibition with clonazepam (table II) [Marchevsky et ai. 1988], and clonazepam has been associated with emergence of manic symptoms in a single case (Dorevitch 1991), although improved behavioural control is a more common outcome. Clonazepam can also cause depression and sexual dysfunction (Cohen & Rosenbaum 1987). Brandwejn et al. (1990) found lorazepam su-
peri or to clonazepam in 24 patients with acute mania. Lorazepam also has the advantage of being well absorbed from intramuscular injection. Adjunctive benzodiazepines added to lithium, carbamazepine or neuroleptic regimens were useful in managing manic disorder (Cohen et a1. 1987). 2.2 Carbamazepine Carbamazepine has a 55 to 65% clinical response rate in mania, and is as fast acting and effective as lithium (Ballenger 1988; Okuma et al. 1990; Stuppaeck et a1. 1990). Kostiukova (1989) found carbamazepine superior in all respects, especially in unipolar depressed patients. Carbamazepine causes rapid clinical improvement in patients with more severe mania and dysphoria, in rapid cyclers with more previous admissions, and in patients without a family history of mania (Liu et a1. 1989; Post et a1. 1987). These are all predictors of poor response to lithium. The onset of efficacy of the drug is not as rapid for manic disorder as it is for epilepsy (Post 1988). Carmabazepine induces its own metabolism and that of tricyclic antidepressants (Leinonen et a1. 1991), so the initially adequate dosage must be in-
Drugs 44 (3) 1992
330
creased after a week (Tran-Johnson et al. 1987). A serum concentration in the 34 to 51 JoLmol/L range appears necessary (table III) [Chen et al. 1990]. Lithium and carbamazepine are more effective together (Kramlinger & Post 1989) and induce a more rapid response (di Costanzo & Schifano 1991), and carbamazepine augments the efficacy of haloperidol (Moller et al. 1989). Post (1990) noted that successive episodes ofaffective illness often show an accelerating frequency of recurrence, analogous to 'kindling' in epilepsy. Thus, prevention of affective episodes should improve the long term outcome in affective disorders. Post et al. (1990) followed 24 patients with refractory affective disorders for a mean of 4 years, and found that carbamazepine reduced the number and severity of manic and depressive episodes. However, while half the patients followed for over 2 years continued to improve, the other half got worse and showed 'escape' from prophylaxis. Vovin et al. (1987) followed 73 patients with phasic affective psychoses for 5 years. Carbamazepine benefited 83% of the patients, and prevented relapse entirely in 27.7%. Small et al. (1991) found carbamazepine and lithium equivalent in the prophylaxis of bipolar disorder. Mosolov (1991) found carbamazepine su-
perior in atypical effective or schizoafIective psychoses and depressions with anxiety. Coxhead et al. (1992) found carbamazepine monotherapy prophylaxis comparable with that of lithium, with less weight gain. There were relapses in the first month of change from lithium to carbamazepine, probably because the lithium was withdrawn too rapidly. 2.3 Valproic Acid and Phenytoin Valproic acid appears effective for the treatment and prophylaxis of acute mania (Fawcett 1989) and schizoafIective disorder (Brown 1989). For example, Calabrese et al. (1992) followed 78 rapid cycling bipolar patients for 15.8 months, and found valproic acid effective in 54% of patients with mania, 87% of those with mixed states, and 19% of those with depression; as prophylaxis it was 72, 33 and 94% effective, respectively, in these states. Mosolov (1991) found valproic acid equal to lithium for prophylaxis, and superior in the treatment of rapid cycling patients, and those with circadian variation and endogenomorphic features. About two-thirds of refractory bipolar patients respond acutely to valproic acid (Brown 1989). Valpromide (a valproic acid precursor) added ,to neuroleptic regimens decreases agitation, shortens hospitalis-
Table III. Typical dosages and effective serum concentrations of anticonvulsants
Clonazepam Carbamazepine Valproic acid Phenytoin a b
Therapeutic serum concentration [!
Drugs 44 (3): 326-335. 1992 0012-6667/92/0009-0326/$05.00/0 © Adis International Limited. All rights reserved. DRU1 186
New Uses of Anticonvulsant Drugs in Psychosis Charles Van Valkenburg, 1 John C Kluznik 2 and Robert MerrilP 1 Nevada Department of Prisons, Carson City, Nevada, USA 2 Minnesota Security Hospital, St Peter, Minnesota, USA 3 Saint Peter Regional Treatment Center, St Peter, Minnesota, USA
Contents 326 327 327 327 328 328 32 328 328 329 330
331 331 332 332 332 332 332
Summary
Summary I. Schizophrenia 1.1 C10nazepam 1.2 Carbamazepine J.3 Valproic Acid 1.4 Phenytoin 1.5 Water Intoxication 2. Bipolar Disorders 2.1 C10nazepam 2.2 Carbamazepine 2.3 Valproic Acid and Phenytoin 3. Psychoses Secondary to Epilepsy 4. Other Organic Psychoses 5. Borderline Disorder 6. Anticonvulsant Drug-Induced Psychoses 7. Anticonvulsant Drug Withdrawal Psychoses 8. Neuroleptic-Induced Seizures 9. Conclusions
Psychotic patients not adequately relieved by neuroleptic drugs often improve when anticonvulsants are added. In bipolar disorders and organic psychoses, anticonvulsants can sometimes be used to replace neuroleptics. No individual anticonvulsant is clearly, consistently superior. Patients who fail on one agent may improve on the next. Clonazepam is an excellent adjunct to neuroleptic therapy, but there is little evidence that it is effective as monotherapy. However, it is safe, sedates rapidly, and has an excellent patient tolerability profile. Carbamazepine is the best established drug for patients with bipolar disorders, particularly for rapid cyclers, and is often effective monotherapy. The therapeutic profile of valproic acid (sodium valproate) is similar to that of carbamazepine, but its side effects are quite different and are often preferred. Other anticonvulsants are little studied, but might be chosen to avoid certain side effects, or after better-studied drugs have failed. The pharmacological basis behind using anticonvulsants in psychoses is primarily empirical. In almost every case it has been clinicians who have first noted the beneficial effects of these drugs. Theories such as that of Post have followed.
New Uses for Anticonvulsants
327
We have come about as far as we can with neuroleptics alone. Apart from some possible reported benefits of clozapine in refractory cases of schizophrenia, we are little better off than we were in the 1950s. It has now also been adequately established that neuroleptic medications have therapeutic ranges (van Putten et al. 1991), and that exceeding these ranges offers no benefit, and may in fact make patients worse. This leaves us with patients whose responses to antipsychotic medications are inadequate. We have learned that responses to electroconvulsive therapy (ECT), antidepressants and lithium are far less specific than nosologists would wish, and that they leave many patients still psychotic. Anticonvulsant drugs, usually added to neuroleptics or to lithium, sometimes used alone, have us on the move again. This article reviews the evidence of anticonvulsant drug efficacy in various psychoses (table I). Evidence is stronger for some than for others, but no systematic direct comparison has shown one agent superior to another. There still appears no good way to predict which anticonvulsant will be best for an individual patient. We have often seen one drug work after another failed.
1. Schizophrenia Treating schizophrenia with neuroleptic agents rarely results in complete remission, but that does not mean that schizophrenic patients cannot be considerably helped with other approaches. 1.1 Clonazepam Clonazepam augments the effects of neuroleptics in schizophrenia, makes them work faster, and reduces their most unpleasant side effects such as akathisia, rigidity and dystonia (Altamura et al. 1987). It also reduces neuroleptic-related postural myoclonus (Fukuzako et al. 1990), occulogyric crises (Horiguchi & Inami 1989), akathisia (Kutcher et al. 1989), blephaspasm (Ballard 1989), and 'painful legs and moving toes' (Sandyk 1990). Clonazepam monotherapy of schizophrenia appears ineffective (Karson et al. 1982). 1.2 Carbamazepine Carbamazepine has usually been given in combination with a neuroleptic rather than as a sole treatment. It appears to augment the effects of neuroleptics, reduce side effects and allow a lower neu-
Table I. Evidence of anticonvulsant efficacy in various psychoses. This table does not reflect relative efficacy, which is unknown,
but rather reflects the strength of current evidence of efficacy Clonazepam
Carbamazepine
Valproic acid
Phenytoin
Schizophrenia Symptoms most likely to respond
N:+++ Excitement, negative symptoms, dystonia, rigidity, akathisia
N:++ Psychosis, hostility
N:+ EXCitement, hostility, polydypsia
Manic disorder Rapid-cycling bipolar Bipolar prophylaxis Epileptic psychosis Supersensitivity psychosis
La: ++++ ? ? M:++ ?
N:+++ Excitement, tension, negative symptoms, suspiciousness, uncooperativeness M:++++ M:++++ M:++++ M:+++ M:++
M:++++ M:++++ M:++++ M:+++ ?
M:++ ? ? M:+ ?
a
Reported as effective monotherapy during first week. = effective as monotherapy; N = effective as adjunct to neuroleptic; L = effective as adjunct to lithium; ++++ = excellent efficacy; +++ = good efficacy; ++ = moderate efficacy; + = poor efficacy; ? = unknown, or equivocal results.
Abbreviations and symbols: M
Drugs 44 (3) 1992
328
nowsky & Putman 1943), but no other SUbtypes. Kubanek and Rowell (1946) found that one-third of patients were greatly improved by phenytoin. In a comparison between phenytoin and placebo in 76 chronic schizophrenics whose neuroleptic treatment had been stopped, phenytoin tended to diminish hostility and unusual thought content in the fIrst month only, and increase motor retardation and blunted affect in the second (Simopoulos et al. 1974). Phenytoin indeed contributes, with other organic factors, to cognitive decline and intellectual deterioration (table II) [Trimble 1988].
roleptic dose. Wetterling (1987) found that carbamazepine improved Brief Psychiatric Rating Scale items of excitement and tension, but not the schizophrenic defect, in 14 male chronic schizophrenics. However, Herrera et al. (1987) found that carbamazepine significantly improved negative (or defect) symptoms in 6 schizophrenics who had failed to improve on at least 2 chemically dissimilar neuroleptics. Similarly, Okuma et al. (1989a,b) demonstrated that 29 of 54, and.61 of 127 patients improved markedly or moderately on carbamazepine; suspiciousness, uncooperativeness, and excitement were particularly improved. Carbamazepine reduced neuroleptic side effects of haloperidol in 10 paranoid schizophrenics, without significantly changing clinical response (Martin-Munoz et al. 1989). Acute. schizophrenic patients given adjunctive carbamazepine· needed less haloperidol and experienced fewer side effects when given carbamazepine, as an adjunct to other therapy (Dose et al. 1987). These patients deteriorated if their carbamazepine was stopped. However, Carpenter et al. (1991) did not find carbamazepine monotherapy to be superior to placebo in the maintenance of 13 outpatient schizophrenics.
1.5 Water Intoxication Water intoxication is often a complication of schizophrenia. In severe cases it can lead to a syndrome of psychosis, intermittent hyponatraemia and polydipsia (PIP) which can cause seizures, coma, and aggressive behaviour. It is hard to manage, and can be life-threatening. Preliminary evidence shows that clozapine is effective in reducing water intoxication (Kluznik 1991). Phenytoin with lithium improves morning basal sodium levels, and is better tolerated than lithium alone (Vieweg et al. 1988). Carbamazepine and c10zapine both can cause agranulocytosis (table II). Giving them together would be expected to be particularly hazardous, and has been associated with 1 death (Gerson et al. 1991). It should also not be forgotten that anticonvulsants frequently cause birth defects, neonatal toxicity and abstinence syndromes (Sitland-Marker et al. 1989).
1.3 Valproic Acid Several early studies investigating valproic acid (sodium valproate) for schizophrenia were negative or inconclusive. Morinigo et al. (1989) added valproic acid to neuroleptic therapy in 4 treatmentresistant schizophrenics, and after 3 months noted a reduction in positive symptoms and reduced hostile disruptive behaviour. The improvement persisted at follow-up of 4 to 12 months. Sultan et al. (1990) described supersensitivity psychosis as a possible long term side effect of neurolepetic treatment, behaviourally analogous to tardive dyskinesia. They treated 5 -such patients with carbamazepine or phenytoin; all improved, 3 persistently. 1.4 Phenytoin Early studies reported that phenytoin benefitted catatonic schizophrenics (Freyhan 1945), and also 73% of schizophrenics who were 'excited' (Kalli-
2. Bipolar Disorders I
Much more attention has been focused the use of anticonvulsants in bipolar disorders than in schizophrenia. Bipolar disorders usually have a better outcome, and clinical improvement is more obvious. 2.1 Clonazepam Chouinard (1987) noted the advantages of clonazepam over neuroleptics in the acqte and maintenance treatment of bipolar disorder: rapid action,
329
New Uses for Anticonvulsants
Table II. Side effects of anticonvulsants
Common and bothersome
Rare and bothersome
Rare and deadly
Clonazepam
Drug abuse Initial sedation
None
Carbamazepine
Dry mouth Initial sedation Hypotension Weight gain Blurred vision Upset stomach Weight gain Tremor Headache Cognitive impairment Dizziness
Disinhibition Confusion Dizziness Rash Nausea Dizziness
Valproic acid
Phenytoin
a b
Aplastic anaemia Allergy
Balding a Vomiting Drowsiness
Hepatitis b Allergy
Irritability Overgrown gums Slurred speech
Allergy
Usually temporary; and may be reversible with zinc supplement. May be limited to infants also taking phenobarbital.
relative lack of toxicity, and no risk of tardive dyskinesia. He suggested treating acute mania with clonazepam alone for the first week, then adding lithium, so that neuroleptics could be avoided (Chouinard 1988). Edwards et al. (1991) showed clonazepam to have a specific anti manic effect, unrelated to its sedative effect, but no specific antipsychotic effect in acute manic patients. De la Gandara Martin et al. (199 I) found adding clonazepam to be beneficial in 85% of cases. However, Aronson et al. (1989) found clonazepam ineffective in maintenance treatment of bipolar disorder, although Mauri et al. (1990) suggested that clonazepam prophylaxis was effective in patients who could not take lithium. After sedation cleared, side effects were minimal. As with all sedative hypnotics, there is a risk of behavioural disinhibition with clonazepam (table II) [Marchevsky et ai. 1988], and clonazepam has been associated with emergence of manic symptoms in a single case (Dorevitch 1991), although improved behavioural control is a more common outcome. Clonazepam can also cause depression and sexual dysfunction (Cohen & Rosenbaum 1987). Brandwejn et al. (1990) found lorazepam su-
peri or to clonazepam in 24 patients with acute mania. Lorazepam also has the advantage of being well absorbed from intramuscular injection. Adjunctive benzodiazepines added to lithium, carbamazepine or neuroleptic regimens were useful in managing manic disorder (Cohen et a1. 1987). 2.2 Carbamazepine Carbamazepine has a 55 to 65% clinical response rate in mania, and is as fast acting and effective as lithium (Ballenger 1988; Okuma et al. 1990; Stuppaeck et a1. 1990). Kostiukova (1989) found carbamazepine superior in all respects, especially in unipolar depressed patients. Carbamazepine causes rapid clinical improvement in patients with more severe mania and dysphoria, in rapid cyclers with more previous admissions, and in patients without a family history of mania (Liu et a1. 1989; Post et a1. 1987). These are all predictors of poor response to lithium. The onset of efficacy of the drug is not as rapid for manic disorder as it is for epilepsy (Post 1988). Carmabazepine induces its own metabolism and that of tricyclic antidepressants (Leinonen et a1. 1991), so the initially adequate dosage must be in-
Drugs 44 (3) 1992
330
creased after a week (Tran-Johnson et al. 1987). A serum concentration in the 34 to 51 JoLmol/L range appears necessary (table III) [Chen et al. 1990]. Lithium and carbamazepine are more effective together (Kramlinger & Post 1989) and induce a more rapid response (di Costanzo & Schifano 1991), and carbamazepine augments the efficacy of haloperidol (Moller et al. 1989). Post (1990) noted that successive episodes ofaffective illness often show an accelerating frequency of recurrence, analogous to 'kindling' in epilepsy. Thus, prevention of affective episodes should improve the long term outcome in affective disorders. Post et al. (1990) followed 24 patients with refractory affective disorders for a mean of 4 years, and found that carbamazepine reduced the number and severity of manic and depressive episodes. However, while half the patients followed for over 2 years continued to improve, the other half got worse and showed 'escape' from prophylaxis. Vovin et al. (1987) followed 73 patients with phasic affective psychoses for 5 years. Carbamazepine benefited 83% of the patients, and prevented relapse entirely in 27.7%. Small et al. (1991) found carbamazepine and lithium equivalent in the prophylaxis of bipolar disorder. Mosolov (1991) found carbamazepine su-
perior in atypical effective or schizoafIective psychoses and depressions with anxiety. Coxhead et al. (1992) found carbamazepine monotherapy prophylaxis comparable with that of lithium, with less weight gain. There were relapses in the first month of change from lithium to carbamazepine, probably because the lithium was withdrawn too rapidly. 2.3 Valproic Acid and Phenytoin Valproic acid appears effective for the treatment and prophylaxis of acute mania (Fawcett 1989) and schizoafIective disorder (Brown 1989). For example, Calabrese et al. (1992) followed 78 rapid cycling bipolar patients for 15.8 months, and found valproic acid effective in 54% of patients with mania, 87% of those with mixed states, and 19% of those with depression; as prophylaxis it was 72, 33 and 94% effective, respectively, in these states. Mosolov (1991) found valproic acid equal to lithium for prophylaxis, and superior in the treatment of rapid cycling patients, and those with circadian variation and endogenomorphic features. About two-thirds of refractory bipolar patients respond acutely to valproic acid (Brown 1989). Valpromide (a valproic acid precursor) added ,to neuroleptic regimens decreases agitation, shortens hospitalis-
Table III. Typical dosages and effective serum concentrations of anticonvulsants
Clonazepam Carbamazepine Valproic acid Phenytoin a b
Therapeutic serum concentration [!
