HISTORICAL NOTES As with many other psychotropic drugs, lithium was discovered by serendipity. In a short essay entitled, “The history of lithium”, Cade (1949) describes the background against which he made the discovery of lithium's specific effects in bipolar illness, his working hypothesis, as well as the unexpected way in which these effects were found.
Lithium has an erratic history in medicine… The alkali itself was discovered by Arfvedson in 1817…. Lithium salts were introduced to medicine by A.B. Garrod in 1859 for the treatment of gout… Culbreth in 1927 stated that lithium bromide is the most hypnotic of all bromides. My discovery of the specific antimanic effect of lithium ion was unexpected but to be retrospectively percipient for a moment, inevitable by-product of experimental work I was doing to test a hypothesis regarding the etiology of manic depressive illness. Could mania be a state analogous to thyrotoxicosis and myxoedema, mania being a state of intoxication by a normal product of the body circulating in excess, whilst melancholia is the corresponding deprivative condition? ...for this purpose guinea pigs were used and fresh urine (of manic patients) was injected intraperitonealy… it soon became evident that some specimens of urine from manic patients were far more toxic than any of control specimens from normal persons... In an attempt to uncover such a toxic metabolite, Cade examined first the role of creatinine and urea, although he hypothesized that a different, third metabolite may be the toxic one. Cade planned to inject his laboratory animals with an aqueous solution of uric acid. However as uric acid is not soluble in water, Cade used the most soluble urate - an aqueous solution of a lithium salt. While examining the effect of the lithium salt itself on the animals, to his surprise he noticed that “…although fully conscious (the animals), they became extremely lethargic and unresponsive to stimuli... it may seem a long way from lethargy in guinea pigs to the control of manic patients but…” and here Cade describes in length his experiments with lithium in manic patients. Interestingly Cade never found the hypothesized toxic metabolite he was looking for, but he opened the way for what is still considered the “gold standard” treatment of bipolar illness. Cade (1970) in a small series of cases reported remarkable therapeutic effects for the manic patients, while schizophrenic and chronic depressive patients failed to demonstrate any changes. Thus, Cade was the first to notice a specific effect for lithium in bipolar disorder.
Cade later studied the effect of other related alkali metals such as rubidium and cesium along with cerium, lanthanum, neodymium and strontium and praseodymium. Cade even examined the effect of strontium in three manic, seven acute schizophrenic, two psychotic depressed, and 10 chronic schizophrenic patients. Four of the acute schizophrenic patients showed rapid improvement within a few days, and this was true for two out of the three manic patients. None of the chronic schizophrenic patients or the psychotic depressive patients improved. Cade concludes his historical lecture (Cade 1949) stating, “That lithium, a simple inorganic ion, can reverse a major psychotic reaction (manic episode) must have, quite apart from its substantial therapeutic value, profound theoretical significance in unraveling the mystery of the so-called functional psychoses. It must be regarded as a major research tool. Strontium may also prove to have a similar value for research in this field even if it has only minor therapeutic value.” Cade, using the terminology of his time (i.e. psychotic reaction), was right about the major role of lithium in a host of studies trying to unravel the mechanisms underlying bipolar illness. Strontium, however, raised no interest, and no further research was done with this agent. LITHIUM IN ACUTE MANIA
The study of lithium in acute mania included several phases, some of them overlapping. The drug was first studied openly in small-scale studies, many of them lacking clear diagnostic criteria and well defined rating scales. After initial positive results were reported, placebo controlled double blind studies followed. As these studies demonstrated evidence for the efficacy of lithium in bipolar mania, a series of controlled studies compared it with traditional antipychotics, and also reported antimanic effects. After a lag of about two decades (the 1970’s and the 1980’s) the marketing of the new antiepileptic and antipsychotic agents, spurred new double-blind studies comparing lithium to these drugs. In the large majority of studies, lithium was shown to be highly effective in acute mania, although more recent naturalistic studies reported lower improvement rates, and some authors even suggested that the benefit of response is compromised by side effects appearing in the early phases of treatment.
One of the main problems in comparing and assessing the results of various studies is that different diagnostic criteria and rating scales were used. We will present here the results of the first two open studies conducted prior to the publication of the Diagnostic Statistical Manual (DSM) I (American Psychiatric Association 1952) to be followed by double blind studies during DSM I era, 1952-1967 (American Psychiatric Association 1952); DSM II era, 1968-1979 (American Psychiatric Association 1968); DSM III era, 1980-1986 (American Psychiatric Association 1980); DSM-III-R era, 1987-1993 (American Psychiatric Association 1987); DSM IV era; 1994-1999 (American Psychiatric Association 1994a) and the DSM IV-TR era; 2000- (American Psychiatric Association 2000). While not all the double blind studies used this diagnostic system, and some studies published during a certain period were conducted in a previous one, we felt that such an organization may serve as an appropriate framework enabling follow-up and assessment of response rates over these time periods. Pre-DSM era open studies
Two pioneering open studies were conducted during the pre-DSM era. The first one by Cade (1949) included 10 manic patients, all of whom improved following lithium treatment. No clear diagnostic and assessment criteria were presented. Cade’s description of the first case of mania ever to be treated with lithium illustrates the magnitude of the breakthrough associated with this new treatment:
this was a little wizened man of 51 who had been in a state of chronic manic excitement for 5 years. He was amiably restless, dirty, destructive, mischievous and interfering. He enjoyed preeminent nuisance value in a backward for all those years and bid fair to remain there for the rest of his life. He commenced treatment with lithium citrate 1200 mg tid on 29 March 1948. On the fourth day the optimistic therapist thought he saw some change... by the fifth day it was clear he was in fact more settled, tidier, less disinhibited and less distractible. From that day on there was a steady improvement... He remained well and left hospital (after housing was arranged) on 9 July, 1948... taking maintenance dose of lithium carbonate 300mg bid. The second open study was that of Noack & Trautner (1951). These authors studied openly 30 manic patients. Twenty-five patients responded well (83%) to lithium treatment. Both of these pioneering studies reported very high response rates to lithium. When comparing their results to results of more recent studies, one should bear in mind the expectation associated with these investigations, the open design utilized, and the fact that patients were lithium naïve.
Lithium vs. Placebo
Schou et al (1954) studied 38 manic (30 typical and eight atypical) patients in Denmark in a partially open and partially placebo-controlled double-blind design. The dose of lithium used resulted in serum lithium levels of 0.5-2.0 meq/L. Clinical response was determined by a three point global severity scale. Fourteen patients showed a positive response (37%), 18 a possible response (47%), and six had no response (16%). Schou (1959) also conducted a large open study in which 91 of 119 (76%) manic patients improved. On the role of Schou in introducing lithium treatment for mania, Cade (1970) wrote: “The person who has done most to achieve this recognition by validating and extending my original observation has been Mogens Schou in Denmark”. Maggs et al (1963) studied 28 acute manic patients. These authors were not clear as to the diagnostic criteria used for patient selection. Lithium or placebo was each given for two weeks in a cross-over design. Two weeks of no drug elapsed between crossing from one drug to another. The Wittenborn Scale for Manic State and Schizophrenic Excitement was used to assess the results. Ten patients dropped out of the study; eight due to uncooperativeness and two due to lithium toxicity. Lithium was found to be superior to placebo. The strengths of this study were the use of a rating scale and its blindness. Its weakness was the high drop-out rate and the cross-over design, which might have biased its results and the short duration of the treatment period.
Bunney et al (1968) studied two patients in a longitudinal double blind fashion demonstrating the sensitivity of manic symptoms to short periods of withdrawal from lithium treatment. The rating scale used was a 24 item mania scale measuring the core as well as other symptoms associated with mania. Diagnosis was made by the Mayer Gross definition of mania (Gershon and Shopsin 1973). This longitudinal follow-up demonstrated the reappearance or increase of manic symptoms within 24 hours of lithium discontinuation. Goodwin et al (1969) studied 30 manic-depressive patients in a double-blind placebo-controlled fashion. This group included 12 manic patients. Doses of lithium used were 900-1800mg daily. Patients’ serum lithium level was above 0.8 and below 1.2meq/L. Nurses rating scales were employed. Eight of the 12 manic patients had a complete response (67%), one had partial response, and three worsened. The symptomatology of the three patients who worsened seemed to correspond more closely with schizoaffective disorder than with bipolar mania. Stokes et al (1971), studied 38 manic patients in a double-blind, placebo-controlled cross-over design. Patients were selected for the study after diagnostic consensus was obtained by three psychiatrists. Lithium was given at 0.5meq/kg/day in four divided doses. Mean serum lithium was 0.93meq/L. Data was analyzed for 98 manic periods (56 on lithium and 42 on placebo), each of seven to ten days. Seventy-five percent of the 56 lithium-treated episodes responded to lithium treatment, while only 40.5% of the 42 placebo-treated episodes responded to placebo. However, this study has significant flaws. The period chosen for each period of treatment may be too short to examine lithium anti-manic effect. There was no washout period between the placebo and lithium periods that may attenuate a carry-over or withdrawal effect, and the rating scale used was not well defined. Nevertheless, this study added to the growing body of studies on lithium efficacy in acute mania. 1980-1993 (DSM-III & DSM-III-R):
The above studies established the role of lithium as the gold standard treatment for acute mania. No placebo-controlled studies of lithium effect in acute mania were performed during these years.
Bowden et al (1994) conducted a double blind placebo-controlled, randomly assigned prospective parallel group study comparing lithium, valproic acid, and placebo in acute mania. No neuroleptics were used, and rescue medication was lorazepam allowed in the first week of the study. One hundred and seventy-nine patients were treated for three weeks with either lithium, divalproex, or placebo in a 1:2:2 ratio. Dosage of lithium was increased if tolerated, resulting in up to 1.5mmol/L. The primary outcome measure was the Mania Rating Scale derived from the Schedule for Affective Disorders and Schizophrenia (SADS). Intent to treat analysis included 35 patients on lithium, 68 patients on divalproex, and 73 patients on placebo. Thirty-three percent of the lithium treated patients, 30% of the divalproex treated patients, and 51% of the placebo treated patients terminated prematurely due to lack of efficacy. More than 50% improvement was noticed in 49%, 48% and 25% of patients on lithium, divalproex, and placebo, respectively. Lithium vs. antipsychotic drugs 1968-1979 (DSM-II):
Johnson et at (1968; 1971) compared the effect of lithium versus chlorpromazine in 21 manic-depressive patients and 13 excited schizoaffective patients. Lithium blood levels were maintained above 1.0 meq/L and medications were given for 14-21 days. Rating scales included the Brief Psychiatric Rating Scale, and Clinical Global Impression Scale. Specific features of the manic symptomatology including excitability were more responsive to lithium, while motor activity was more responsive to chlorpromazine, suggesting that lithium treatment may be more specific to mania. Spring et al (1970) studied in a double-blind design 14 acute manic patients randomly assigned to treatment with lithium or chlorpromazine. Lithium treatment in the first week was 1800mg daily, and could subsequently be raised up to 3000mg daily. Chlorpromazine dose could be adjusted up to 1600mg daily. Six of seven patients who started on lithium responded, while only three of the five patients starting with chlorpromazine had a therapeutic response. Interestingly, the two chlorpromazine failures were switched to lithium and also responded. While the results are impressive for such a small study, regrettably these authors did not report how they defined response. Although not statistically significant, lithium was found to be more effective than chlorpromazine for the treatment of typical manic symptoms such as motor hyperactivity, flight of ideas, euphoria, expansiveness and pressured speech. Platman et al (1970) studied 23 manic patients in double-blind fashion. Patients were randomized to either lithium or chlorpromazine treatment, each after two weeks of lead in placebo treatment. Plasma lithium levels were maintained at 0.8 mEq/L. The mean daily dose of lithium was 1800 mg, and for chlorpramazine it was 870 mg. Thirteen patients were treated with lithium and ten with chlorpromazine. Ten patients dropped out of the study. Clinical change was evaluated blindly with the Psychiatric Evaluation Form. After three weeks of treatment, lithium seemed to be superior to chlorpromazine, but statistical significance was not reached. The Veterans Administration and the National Institute of Mental Health initiated a collaborative project on lithium involving 18 US centers (Prien et al 1972). One of its objectives was to compare the efficacy of lithium and chlorpromazine in the treatment of bipolar manic and schizoaffective patients. After a lead-in period of three to five days, patients were randomly assigned to lithium or chlorpromazine for three weeks. The Brief Psychiatric Rating Scale (BPRS) and Multidimensional Psychiatric Scale were used to evaluate the results. Data from 225 patients was analyzed. No major differences between the two drugs were found among study completers, however, a large proportion of the patients treated with lithium terminated prematurely because of lack of cooperativeness, lack of improvement (at least some of it may be attributed to delay in lithium therapeutic effect), or toxicity, while this was true for only eight percent of the chlorpromazine-treated patients. Patients defined as highly active were improved by both drugs. Shopsin et al (1975) compared lithium, haloperidol and chlorpromazine in severely ill hospitalized manic patients, in a double-blind study design, with random treatment assignment. All drugs showed similar efficacy on most clinical measures. The authors suggest however, that the scales used were not sensitive enough to measure the manic psychopathlogy, as the majority of lithium treated patients met discharge criteria at the end of the trial, but this was not true for patients receiving anti-psychotic drugs. The authors suggested that lithium and haloperidol significantly improved the manic symptoms without sedation, while chlorpromazine, on the other hand, seemed to produce considerable sedation, contributing less to improving the underlying manic symptoms. They also suggested that while haloperidol had rapid impact on behavioral-motor activity, lithium tended to act more evenly on the entire manic clinical picture. Takahashi et al (1975) conducted a multi-center trial comparing lithium and chlorpromazine in a double-blind controlled design in a series of 80 Japanese patients with endogenous manic psychosis. The dosages employed were at a ratio of 4:1 (lithium: chlorpromazine). Lithium was found to be superior to chlorpromazine in physicians’ overall ratings. The onset of therapeutic effects of lithium in 65% of the patients occurred within 10 days. Lithium, but not chlorpromazine, was found to improve mood and pressured speech. 1980-1986 (DSM-III):
Garfinkel et al (1980) studied lithium plus placebo vs. haloperidol plus placebo vs. the combination of lithium and haloperidol in a double-blind study with random treatment assignment. Each treatment was administered for three weeks, in 21 severely ill manic patients. The subjects on placebo and haloperidol and the subjects on lithium and haloperidol significantly improved after seven days of treatment in comparison to the lithium and placebo treated group. The authors concluded that haloperidol is superior to lithium in the treatment of acutely severe hospitalized manic patients, and mentioned that the combination of haloperidol and lithium was not superior to haloperidol and placebo. 1987-1993 (DSM-III-R):
No studies comparing lithium and neuroleptics were reported during these years. Such studies had to wait until the development and marketing efforts of the atypical antipsychotics.
Segal et al (1998) studied 45 inpatients diagnosed by DSM-IV as bipolar manic patients in a randomized, controlled, double-blind study of either six mg daily of risperidone, 10 mg daily of haloperidol or 800-1200 mg daily of lithium. Similar improvement was reported for all the three groups after four weeks of treatment as determined by Global Assessment of Functioning (GAF), Clinical Global Impression (CGI), and BPRS (mean BPRS improvement scores: lithium =9, haloperidol =5, risperidone =6.5) scales. Berk et al (1999) compared lithium to olanzapine in mania in a double-blind randomized controlled trial. Thirty patients meeting DSM-IV criteria for mania were randomly assigned to receive either lithium or olanzapine for 4 weeks. No significant differences were found between these treatments. Olanzapine was however, significantly superior to lithium on the CGI-severity scale at week 4. The authors suggest that olanzapine appears to be at least as effective as lithium in the treatment of mania.
Bowden et al (2005) evaluated the efficacy of lithium versus quetiapine versus placebo in the treatment of hospitalized patients with mania in a multi-center, double blind, parallel-group, and 12-week study. Patients with diagnosis of bipolar I disorder in their manic episode were randomly assigned to receive quetiapine (titrated up to 800 mg/day based on tolerance), placebo, or lithium. Results showed that more patients in the quetiapine (72/107) and lithium (67/98) treatment arms completed the study compared with the placebo group (35/97). Improvement in YMRS score was significantly greater for quetiapine than placebo at day 7 and the difference between groups continued to increase over time to day 21 and to endpoint at day 84. Lithium-treated patients improved significantly compared with placebo patients similar to quetiapine treated patients. The authors concluded that quetiapine and lithium demonstrated superior efficacy to placebo in patients with bipolar mania.
Comparison of these various studies is not an easy task, as different time points and rating scales were used. However, since some of the above studies suggested lithium to be superior to antipsychotic agents, and others suggested the superiority of antipsychotic agents, while still others found no difference, it seems logical to sum up the reviewed studies by stating that; overall, there are no overwhelming differences between these drugs in the treatment of bipolar mania. It appears, however, that typical neuroleptics tend to influence primarily the behavioral motor signs (haloperidol) or produce sedation (chlorpromazine), while lithium tends to affect the core manic symptoms in a more even way. This may have lead to the observation of Segal et al (1998) that haloperidol may have an advantage over lithium in the first week of treatment for hospitalized severely ill manic patients, and the clinical impression is` that these neuroleptic agents may help in decreasing combativeness and agitation, and other signs of motoric over activity. As to atypical antipsychotic agents it seems that they are at least as effective in treating acute mania as lithium and clinical trial data suggest that the combination of lithium with an atypical antipsychotic agent (i.e, quetiapine or olanzapine or risperidone) may be superior to lithium administered alone in the treatment of acute mania (see Sachs et al, 2002; Tohen et al, 2002, Sachs et al, 2004). Lithium versus Anticonvulsants 1987-1993 (DSM-III-R):
Lerer et al (1987) conducted a double-blind study comparing lithium and carbamazepine in acute mania. Fourteen patients were assigned randomly to each group. BPRS, CGI, and Beigel-Murphy Manic State Rating Scales were employed. A more consistent level of improvement across patients was found for lithium compared to a minority of good responders in the carbamazepine group. Small et al (1991) studied 52 hospitalized manic patients who were randomized to either carbamazepine or lithium treatment after two weeks of drug withdrawal. The subjects were followed for eight weeks. No difference was found between the two groups. The researchers had the impression that patients treated with carbamazepine were more manageable than those treated with lithium, while lithium-treated patients remained longer in a follow-up phase. The authors concluded that monotherapy with either drug may not be sufficient for hospitalized manic patients. Freeman et al (1992) studied 27 DSM-III-R manic patients in a three week randomized, double-blind parallel study comparing the efficacy of valproate and lithium. BPRS, GAF and SADS-C were used as rating scales. Nine of the 14 (64%) patients treated with valproate and 12 out of the 14 (86%) patients treated with lithium responded favorably, suggesting that both drugs are effective in the treatment of acute mania. 1994-1999 (DSM-IV):
Emilien et al (1996) conducted a meta-analysis comparing the efficacy of lithium, valproic acid and carbamazepine in the treatment of mania. The analysis included only randomized double-blind controlled clinical trials in which the therapeutic plasma level of lithium was 0.4-1.5 mmol/L. Effect sizes were measured by the odds ratio using the Mantel-Haenszel method. No significant difference was observed among the three drugs. Bowden et al (1994) conducted a double blind placebo-controlled, randomly assigned study comparing lithium, valproic acid, and placebo in acute mania on one hundred and seventy-nine patients. The dosage of lithium was increased, if tolerated, to 1.5mmol/L (A detailed description of this study appears in the section of lithium versus placebo). Bowden analyzed the effect size for the efficacy of each treatment arm in a later publication (Bowden et al 1997). Effect size for the improvement score in the Manic Syndrome Score was 0.79 for lithium and 1.01 for divalproex, compared with only 0.37 for placebo. Effect size for the improvement score at the Behavior and Ideation Score (BIS) was 0.62 for lithium, and 0.67 for divalproex, compared with only 0.25 for placebo. These results suggest that both drugs are effective in the treatment of mania compared to placebo. 2000-2006 (DSM-IV-TR): Hirschfeld et al (2003) evaluated the early efficacy of oral-loaded divalproex in acute mania versus standard-titration divalproex, lithium, olanzapine, or placebo by the use of a pooled analysis of 348 subjects from 3 randomized, double-blind, parallel-group, active- or placebo-controlled studies. Subjects were inpatients diagnosed with bipolar I disorder-acute manic phase (DSM-III-R or -IV and SADS-Change Version). Patients were administered oral-loaded divalproex (20 or 30 mg/kg/day on days 1 and 2 followed by 20 mg/kg/day, and increased at physician's discretion), standard-titration divalproex initiated at 250 mg t.i.d. and titrated to 40-150 microg/mL, lithium (300 mg t.i.d. initial dose) titrated to 0.4 to 1.5 mEq/L, olanzapine (10 mg q.d. initial dose) up to 20 mg/day, or placebo. The results suggested the oral loading of divalproex leads to a more rapid antimanic effect when compared with standard-titration divalproex, lithium, or placebo and is better tolerated than olanzapine and as well tolerated as lithium or standard-titration divalproex.
In summary, lithium was demonstrated to be effective in the treatment of mania. Some clinicians treat manic patients with lithium alone, while others may start with combined lithium and antipsychotic treatment, or may add lithium after the first few days of neuroleptic treatment. To some extent this depends on the acuity of the presenting manic episode and the symptom clusters that are behavioral issues (for instance, agitation, combativeness, extreme motor activity) or the need to target other symptoms such as lack of need for sleep, suicidality and anxiety among others. Alternatively, clinicians may use divalproex or carbamazepine, or in certain cases, combine each one of them with lithium, i.e. lithium and divalproex or lithium and carbamazepine.
LITHIUM IN BIPOLAR DEPRESSION
In a review of the treatment of acute bipolar depression, Srisurapanont et al (1995) reviewed 10 studies evaluating lithium, 1 open and 9 double-blind studies. In 7 of 8 placebo-controlled cross-over studies, the authors noted lithium was superior to placebo with a mean response rate of 76%, and 38 to 70% of patients (mean: 52%) relapsed when lithium was substituted for by placebo. The authors note in only one study did imipramine do better than lithium (response rate of 58% vs. 32%).In two augmentation studies where lithium was added to anti-depressant resistant patients, or to carbamazepine resistant patients response rates increased by 36 and 46% respectively. The authors concluded that the benefits of lithium monotherapy or lithium augmentation are impressive for bipolar depression, and that no hypomanic or manic switching was noted.
Nemeroff et al ( 2001) compared the efficacy and safety of paroxetine and imipramine with that of placebo in the treatment of bipolar depression in adult outpatients stabilized on lithium in a double-blind, placebo-controlled study, where 117 depressed bipolar patients were randomly assigned to each of thee treatment for 10 weeks. Some patients may have received either carbamazepine or valproate in combination with lithium for control of manic symptoms. No statistically significant difference was noticed among the three groups. However, both paroxetine and imipramine were superior to placebo for patients with low serum lithium levels (=0.8 meq/liter) but notfor higher serum lithium levels. Compared to imipramine, paroxetine resulted in a lower incidence of adverse events, most notably emergence of manic symptoms. The authors concluded that antidepressants may not be useful adjunctive therapy for bipolar depressed patients with high serum lithium levels. However, the authors suggest that antidepressant therapy may be beneficial for patients who cannot tolerate high serum lithium levels or who have symptoms that are refractory to the antidepressant effects of lithium. Although the hallmark of bipolar illness may be its manic episode, this disorder is also characterized by depressive episodes. Six placebo-controlled trials examined lithium efficacy in bipolar depression (Goodwin and Jamison 1990). Goodwin et al (1969) studied the role of lithium in 13 bipolar depressed patients, and reported that in 10 cases there was indication of some response. Stokes et al (1971) (described above) reported that lithium administration for bipolar depressive episodes of up to 10 days resulted in a significant trend towards response compared with placebo. Eighteen of their 38 patients demonstrated depressive episodes. Eleven patients with 17 depressive episodes were treated with lithium, while 15 patients with 21 depressive episodes were treated with placebo. Fifty-nine percent of the lithium treated episodes improved as compared to 48% of the placebo treated episodes. Goodwin et al (1972) studied 40 bipolar depressed patients reporting that 80% of these patients responded. Noyes et al (1974) reported that six of six bipolar depressed patients responded to lithium. Baron et al (1975) studied nine bipolar depressed patients of whom seven responded, and Mendels et al (1975) reported that nine of 13 bipolar depressed patients responded to lithium treatment. Overall, response rates of bipolar depression to lithium treatment seemed to exceed 70%. All these studies suggested that lithium has antidepressant effects in bipolar depression. Interestingly, some of these studies also reported lithium's effects on unipolar depression which were lower than in bipolar depression, and in general fell under 50% of response rate. Other studies in depressed individuals compared lithium to the classical tricyclic antidepressants. Fieve et al (1968), Mendels et al (1975), Watanabe et al (1975) and Worrall et al (1979) compared the efficacy of lithium to either imipramine (three studies) or desipramine in samples containing bipolar depressed and unipolar depressed patients. Although the majority of these studies suggested that lithium is at least as effective as the tricyclic antidepressants (see also review by Srisurapanont et al (1995)), the study of Fieve et al (1968) - the only one comparing both lithium and tricyclic drug to placebo - reported only a mild antidepressant effect for lithium. These results, and the fact that lithium seemed to be more effective in the prevention of mania than in the prevention of bipolar depression, suggested that lithium may have a greater role in the treatment of bipolar mania than bipolar depression. LITHIUM IN MIXED AND DYSPHORIC MANIC EPISODES
Only a few studies examined the role of lithium in mixed and dysphoric mania. Clotheir et al (1992) reported that both lithium and valproate were found to be equally effective in the control of depressed mood or anxiety associated with a manic episode. Swann et al (1997) reported that in a parallel group, placebo controlled, double-blind, multi-center study of lithium and divalproex for mania (Bowden et al, 1994), depressive symptoms were associated with poor antimanic response to lithium and better response to divalproex. Mixed patients were also suggested to have a poor response to lithium treatment varying between 29-42% (Himmelhoch et al 1976, de Montigny et al 1981, Secunda et al 1985, Swann et al 1986, Prien 1988, Dilsaver et al 1993b, Soares and Gershon, 1998). Interestingly although the existence of depressive symptoms during mania seems to lower responsivity to lithium, lithium has a role in the treatment of resistant depression and its addition to ongoing antidepressant treatment was suggested to reverse refractoriness to treatment (Prien et al 1972, Heninger et al 1983). In this regard, lithium has been added to anti-depressant resistant bipolar depressed patients or to carbamazepine resistant bipolar depressed patients as noted earlier and shown benefits”” (Srisuapanont et al, 1995). In summary, lithium seems to have a modest therapeutic effect in the treatment of bipolar depression. It is also generally less effective in dysphoric and mixed manic states. In such instances, other anticonvulsant mood stabilizers may be more effective, and atypical antipsychotic agents have also shown efficacy in mixed mania.
LITHIUM MAINTENANCE THERAPY OF BIPOLAR DISORDER
The issue of long-term prevention or prophylaxis of illness episodes is of utmost importance for this cycling disorder. A variety of studies to be detailed here corroborate the effectiveness of lithium as an important prophylactic agent for this disorder. These studies will be grouped as open versus double-blind, or prospective versus discontinuation studies.
Three open studies suggested that lithium may be effective in the prophylaxis of bipolar disorder (Hartigan 1963, Baastrup 1964, Baastrup and Schou 1967). 1968-1979 (DSM-II):
During the 1970s a series of double-blind studies were conducted to examine the possible efficacy of lithium in the prophylaxis of bipolar disorder. Goodwin (1994) summarized these studies calculating the recurrence rates over one a year period for placebo to be 81%, but only 34% for lithium. These included both prospective and discontinuation studies.
Coppen et al (1971) conducted a parallel group, randomized, prospective trial that included 17 patients on lithium and 21 on placebo, who were followed for four to six months. The recurrence rates were 18% in the lithium group and 95% in the placebo one. Prien et al (1973a) conducted a parallel group, randomized, prospective trial with patients admitted to hospital for the treatment of a manic episode. One hundred and one patients were on lithium, and 104 patients were on placebo. The duration of the study was 24 months. The recurrence rates were 43% in the lithium group and 80% on placebo. The recurrence rates of manic and depressive episodes were 32% and 16% on lithium, and 68% and 26% on placebo. Prien et al (1973b) conducted a parallel group, randomized, prospective trial in patients admitted for hospitalization due to the occurrence of depressive episodes. Eighteen patients were on lithium, 13 patients on placebo. The duration of the study was five to twenty-four months with a recurrence rate on lithium of 28% and on placebo 77%. Percent recurrence of manic and depressive episodes for lithium was 11% and 22%, for placebo 38% and 62% respectively. Stallone et al (1973) conducted a parallel group, randomized, prospective trial. Twenty-five patients were on lithium, and 27 patients were on placebo, for a study duration of up to 28 months. The recurrence rates on lithium were 44%, and on placebo 93%. The recurrence rates for manic and depressive episodes were 20% and 28% in the lithium group, and 56% and 48% in the placebo group, respectively. Dunner et al (1976) conducted a parallel group, randomized, prospective trial in bipolar II patients. Sixteen patients were on lithium and 24 patients on placebo. The study duration was up to 36 months. Recurrence rates for manic and depressive episodes on lithium were 6% and 56%, and on placebo 25% and 50%, respectively. Fieve et al (1976) conducted a parallel group, randomized, prospective trial in bipolar patients. Seventeen patients were on lithium and 18 patients on placebo. They were followed for up to 53 months. The recurrence rates for manic and depressive episodes on lithium were 59% and 29%, while on placebo they were 94% and 44%, respectively. Fieve et al (1976) also studied bipolar II patients. Seven patients were on lithium and 11 patients on placebo, and those were followed for up to 53 months. The recurrence rates for manic and depressive episodes on lithium were 0% and 57%, and on placebo were 9% and 64%, respectively.
B. Discontinuation trials
Baastrup et al (1970) studied 22 patients on placebo, and 28 patients on lithium. The duration of the study was up to five months. The recurrence rates of illness episodes were 0% on lithium and 55% on placebo. The recurrence rates of manic episodes for lithium and placebo were 0% and 27% respectively. The recurrence rates for depressive episodes for lithium and placebo were 0% and 23% respectively. Melia (1970) studied seven patients on lithium and eight patients on placebo for a period of 24 months. Percent recurrence on lithium was 57 %, and on placebo 78%. Cundall et al (1972) studied 12 patients on lithium and 12 patients on placebo, who were followed for six months. The recurrence rates on lithium were 33 %, and on placebo 83%. The recurrence rates of manic and depressive episodes in the lithium group were 8% and 25%, and in the placebo group 75% and 42%, respectively. Hullin et al (1972) studied 18 patients on lithium and 18 patients on placebo, who were followed for six months. The recurrence rates were six percent on lithium, and 33% on placebo.
A. Prospective trials
Coppen et al (1983) studied 72 patients in a prospective double-blind fashion. Patients continued with their ongoing lithium treatment or received a 25-50% reduction in their lithium dosage resulting in plasma levels of 0.45-0.79mmol/L. The patients on the reduced dose had fewer side effects, including hand tremors, and decreased thyroid stimulating hormone (TSH) blood levels. They also had significantly decreased affective morbidity. The authors suggested that there may be an advantage in keeping maintenance blood lithium levels at about 0.6mml/L.
Prien et al (1984) studied 117 bipolar patients receiving either lithium, imipramine, or both in a long-term double-blind study. Lithium carbonate and the combination were found superior to imipramine in preventing the manic recurrence, and equally effective to imipramine in preventing depressive episodes.