47 Lithium and Its Role in Psychiatry Roy H

47 Lithium and Its Role in Psychiatry Roy H. Perlis, MD, MSc, and Michael J. Ostacher, MD, MPH, MMSc

KEY POINTS formation. This effect is reversed by addition of myoinositol, providing some support for the importance of InsP3 in lithi- • Lithium remains a first-line treatment for all phases um’s effect. In the second, lithium inhibits glycogen synthesis kinase of bipolar disorder, including mania, depression, and 4–7 prevention of recurrence. 3-beta (GSK3B), an important enzyme in pathways including the Wnt signaling cascade.8 ,Notably mice expressing lower • While not examined in a controlled trial, abundant levels of GSK3B exhibit behaviors similar to mice treated evidence supports a role for lithium in decreasing the chronically with lithium.9 Signaling through the GSK3B risk of suicide. pathway may also be central to the observed neuroprotective 10 • Lithium has a narrow therapeutic window, effects of lithium. ,Of course both InsP3 and GSK3B pathways necessitating careful titration and close monitoring of have convergent effects—both influence the serine/threonine plasma levels. kinase Akt-1,4 for example. Expression of other genes (typi- • Lithium toxicity may cause confusion and ataxia. cally in lymphocytes) has been shown in single studies to be influenced by lithium administration, though the relevance of • Drugs that affect lithium levels include non-steroidal these effects to lithium’s effects on mood or other phenotypes anti-inflammatory drugs (NSAIDs) and diuretics. is unknown.

PHARMACOKINETICS AND PHARMACODYNAMICS HISTORICAL CONTEXT Lithium is absorbed through the gut and distributes rapidly The history of lithium’s use in psychiatry parallels the develop- through body water, achieving peak plasma concentrations 1 ment of modern psychopharmacology. The first specific to 2 hours after a single dose. (Slower-release forms may description of the application of lithium to treat mania require 4 to 5 hours to reach peak concentration, because of occurred in 1949, by an Australian named John Cade, who transit time through the gut.) As a monovalent cation like observed that lithium had calming effects on animals and sodium, lithium’s clearance relies entirely on renal function. then treated a series of 10 agitated manic patients. In fact, It is not metabolized by the liver, nor is it significantly protein however, descriptions of lithium treatment date back to at bound while circulating. In general, the half-life for renal least the United States Civil War. An 1883 textbook by Union excretion is approximately 24 hours, so steady-state serum Army Surgeon General William Hammond recommended the levels are typically reached after 5 days. For this reason, lithium use of lithium bromide to treat manic or agitated patients, levels are typically checked about 5 days after initiation or though he later downplayed the importance of lithium. In the dose change. Because lithium distributes throughout the body, early 1900s, a Danish physician, Lange, published a case series it is influenced by lean body mass—for example, among geri- reporting the treatment of manic patients with lithium car- atric patients, lithium levels for a given dose tend to be greater bonate. There is little evidence that lithium was studied than among younger patients with greater lean body (includ- further, however, until Garrod proposed that lithium urate ing muscle) mass. Magnetic resonance spectroscopy studies could be used to treat gout, opening the door to its broader suggest that brain lithium levels are highly correlated with therapeutic application. plasma levels, though less so in patients at the extremes of Unfortunately, despite early studies by Mogen Schou and age—that is, it is possible to have supra-therapeutic levels in others, lithium’s wider adoption in the US was hindered by the central nervous system (CNS) while maintaining a normal concerns about lithium toxicity. Lithium chloride had been plasma lithium level. used as a sodium substitute in the 1940s, until several deaths Drugs that affect renal function, particularly re-absorption, were reported from lithium toxicity among hyponatremic can have profound effects on lithium clearance. Perhaps patients. Thus, lithium was initially perceived as too danger- most notably from a clinical perspective, non-steroidal anti- ous for clinical application, and it was only in 1970 that inflammatory drugs (NSAIDs) or other COX-2 inhibitors may lithium was approved by the United States Food and Drug decrease renal blood flow and thereby increase lithium levels Administration (FDA) for the treatment of mania.1 by up to 25%. Diuretics likewise affect lithium levels, though the nature of their effect depends on their site of action. In the LITHIUM’S MECHANISM OF ACTION kidney, lithium is primarily re-absorbed in the proximal renal tubules, with some subsequent absorption in the loop of The mechanisms by which lithium exerts its therapeutic effects Henle. Importantly, in contrast to sodium, no significant are not entirely clear, but the signaling pathways with which absorption occurs in the distal tubules. Therefore, thiazide it interacts are becoming better understood. Two major diuretics, which act distally, will tend to increase lithium levels pathways are influenced by lithium. In the first, re-cycling of by up to 50%, while those that act more proximally generally inositol is inhibited by lithium, which influences inositol have less of an effect on lithium levels. 2–4 1,4,5-triphosphate (InsP3)-dependent signaling. InsP3 sign- More broadly, hydration status can affect lithium levels: aling acts in part by regulating intra-cellular calcium release individuals who become salt-avid (e.g., because of hypovo- and protein kinase activation, with broad effects. At a neuro- lemia or hyponatremia, perhaps in the context of vomiting nal level, lithium, like valproate, has been shown to increase and diarrhea or self-induced injury, such as long-distance- the spread of growth cones, which are necessary for synapse running) will cause their lithium levels to increase.

525

EVIDENCE FOR LITHIUM’S EFFICACY While these trials showed a positive benefit for lithium, none was a randomized, parallel-group study; instead, they used a Lithium in Acute Mania within-subject design in which each subject was started on Beginning with Schou’s study of lithium versus placebo for lithium or placebo and then switched to the other. It is acute mania, lithium has repeatedly shown efficacy for the unlikely, unfortunately, that there will be any large, well- treatment of mania,11 with the first large randomized study of designed trials of lithium to answer this question, most promi- lithium treatment of acute mania finding lithium comparably nently because there is no pharmaceutical manufacturer effective to the antipsychotic chlorpromazine.12 Since then, producing lithium that has any financial incentive to under- multiple studies have found lithium to be superior to placebo take such a study. and comparable or superior to other agents in the acute man- Lithium, used as monotherapy, appears to be as effective agement of bipolar mania; few studies have found superiority for the treatment of bipolar depression as the combination of for other drugs over lithium. A systematic review found 12 lithium and an antidepressant. In a study comparing imi- acute mania trials comparing lithium to placebo or another pramine, paroxetine, and placebo added to lithium carbonate agent that met their criteria for data-pooling. This review of for the treatment of a major depressive episode in bipolar disorder, neither antidepressant added benefit beyond lithium studies of lithium for acute mania found superiority for 20 lithium over placebo and chlorpromazine, with equivalence alone. Response rates (defined as a Hamilton Depression to valproate and carbamazepine.13 Rating Scale [HAM-D] score of 7 or lower) were 35% for As pointed out by Grunze,14 however, few studies of lithium lithium alone, compared to 39% for imipramine, and 46% in acute mania were undertaken with the methodological for paroxetine. In a secondary analysis, subjects with lower rigor required today for regulatory approval of a drug’s use. By lithium levels (less than 0.8 mEq/L) had a lower response coincidence, the first rigorously designed study to demonstrate rate compared to the adjunctive antidepressant group, suggest- the efficacy of lithium for acute mania was Bowden’s seminal ing, perhaps, that higher lithium levels are as effective as study of divalproex sodium for the treatment of acute mania lithium plus an antidepressant in the treatment of bipolar in 1994, which was designed to study that drug for FDA depression. approval;y b including lithium as an active comparator, the study also demonstrated lithium’s efficacy.15 This study was Lithium for Maintenance Treatment and Relapse adequately powered (i.e., it included a large enough sample Prevention of Bipolar Disorder of patients to find a high probability of finding a statistically significant difference between treatments with a low probabil- Lithium is the archetypal maintenance treatment for bipolar ity of error), compared a drug to an agent known to be effec- disorder. From Prien’s first maintenance study of lithium tive (lithium), and it included a placebo arm. Additionally, it (comparing it to chlorpromazine), to more recent studies was not biased by inclusion based on prior response to using lithium as a comparator for maintenance studies of other, drugs lithium has clear benefit for maintaining response lithium. In this 3-week study, lithium was superior to placebo 12,21 and equivalent in efficacy to divalproex sodium with a 50% and preventing relapse in bipolar disorder. sLithium’ clear- response rate (defined as a 50% drop in mania scale scores) est benefit in long-term use is in the prevention of relapse to for lithium compared to 26% for placebo.15 Since that study, mania, although relapse to depression is more common in pooled data from trials of topiramate for mania failed to patients with bipolar disorder. As is the case with lithium in demonstrate a benefit for that drug, but it did re-confirm the acute mania, lithium’s efficacy compared to placebo was only efficacy of lithium for acutely manic patients.16 The percentage confirmed in later studies designed to establish regulatory of patients with a 50% or greater reduction in the Young approval for newer drugs, including divalproex sodium and Mania Rating Scale (YMRS) at day 21 was 28% with placebo lamotrigine. Earlier studies were beset with methodological (n = 427), 27% with topiramate (n = 433), and 46% with problems, including on–off rather than parallel group designs, lithium (n = 227). Lithium was statistically better than placebo lack of diagnostic clarity (e.g., the inclusion of unipolar and topiramate on all psychometric measures other than the patients), and rapid or abrupt lithium discontinuation in Montgomery-Asberg Depression Rating Scale (MADRS). stable patients. Concerns about sudden discontinuation of Monotherapy treatment in any phase of bipolar disorder, lithium are genuine, as there is a high rate of manic relapse in however, is increasingly rare, and is especially so in the treat- these patients; inclusion of patients from these studies might ment of acute mania.17 It appears that mania outcomes, in artificially inflate the difference between lithium and placebo in maintenance treatment.22,23 terms of time to response and proportion of patients who 24 remit,y ma be improved with the addition of antipsychotics Geddes and colleagues have completed the definitive sys- to lithium.18 The adjunctive use of typical (including haloperi- tematic review of lithium for maintenance treatment in bipolar dol) and atypical (e.g., aripiprazole, asenapine, olanzapine, disorder. Having reviewed 300 studies, they included only five quetiapine, and risperidone) antipsychotics with lithium in their meta-analysis, limiting inclusion to randomized, carbonate has been found to improve outcomes compared to double-blind, placebo-controlled trials. They found that lithium alone. The atypical antipsychotic ziprasidone, however, lithium was more effective than placebo in preventing relapses did not improve outcomes significantly compared to placebo toy an mood episode (random effects relative risk = 0.65, 95% when added to lithium. confidence interval [CI] = 00.5 to 0.84) and to mania (relative risk = 0.62, 95% CI = 0.40 to 0.95), with a non-significant Lithium in Acute Bipolar Depression effect on relapse to depression (relative risk = 0.72, 95% CI = 0.49 to 1.07).24 The average risk of relapse of any kind in The options for the pharmacological treatment of major 1 to 2 years of follow-up was 60% for placebo, compared to depressive episodes in bipolar disorder, unlike those for 40% for lithium; this can be understood as lithium preventing mania, remain few. In spite of being recommended as first-line one relapse for every five patients treated compared to placebo. treatment in recent bipolar treatment guidelines, there are few Relapse rates to mania were 14% for lithium compared to 24% data to support the use of lithium as an acute antidepressant for placebo, while relapse rates to depression were 25% for in bipolar disorder. A comprehensive review by Bauer and placebo compared to 32% for lithium. There are some limita- Mitchner19 identified only three placebo-controlled trials of tions and criticisms of this study, however. The outcomes were lithium for bipolar depression (with a total of 62 subjects). not defined uniformly across the included studies; one study

Downloaded for Rohul Amin ([email protected]) at Uniformed Services Univ of the Health Sciences from ClinicalKey.com by Elsevier on September 21, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. Lithium and Its Role in Psychiatry 527 included in the analysis had exclusively bipolar II subjects, and sodium) for this specific course of bipolar disorder, but an the follow-up period of 1 to 2 years is too short to adequately ambitious clinical trial and a large body of naturalistic data 47 evaluate the benefit of lithium (as some have argued that the suggest that lithium is no less ineffective than other com- maintenance benefit of lithium is only apparent after 2 years pounds for rapid-cycling.29–32 Calabrese and colleagues31 com- of treatment).23,25 pared lithium to divalproex sodium in rapid-cycling patients Lithium was compared to olanzapine for the prevention of stabilized on both drugs and found no difference between relapse of bipolar I disorder in a randomized, controlled, drugs on time-to-episode-recurrence. As testament to the dif- double-blind trial.26 Bipolar I patients were stabilized on a ficulty of treating rapid-cycling, only 60 of the original 254 combination of lithium and olanzapine, randomized to one subjects, who were randomized to the two study conditions, or the other drug, and followed for 12 months. There was no achieved stabilization. In a cohort of 360 patients treated for difference between drugs on the primary outcome measure or bipolar disorder in Sardinia, time to recurrence was no differ- time to symptomatic relapse (YMRS or HAM-D scores of 15 ent for the patients with or without a rapid-cycling course.32 or greater), although there were fewer relapses to mania/mixed (but not depressive) episodes in the olanzapine-treated group. Lithium in Suicide Prevention A number of studies have examined outcomes for subjects stabilized on an antipsychotic added to lithium or valproate Lithium may have anti-suicide effects in patients with mood and then randomized to remain on lithium or valproate and disorders. While there are no prospective, randomized studies the antipsychotic or lithium or valproate plus placebo. designed to examine lithium’s potential to reduce suicide and Notably, these studies include aripiprazole, quetiapine, and suicide attempts, a number of meta-analyses, smaller inde- ziprasidone; they are enriched designs intended to study the pendent studies, and a study from two large health insurance impact of the antipsychotic primarily, but do suggest that databases generally substantiate lithium’s value as a prophy- those patients who are stabilized on lithium or valproate plus lactic agent against suicidal behavior in bipolar disorder. one of those antipsychotics remain on both drugs. The strongest evidence for decreased suicide in patients A study was completed specifically to examine whether treated with lithium comes from a meta-analysis by Cipriani combination treatment with lithium and valproate together is and co-workers33 of all randomized studies of lithium (either more effective than either of those two drugs as monotherapy versus placebo or another drug) in mood disorders (including to prevent recurrence in bipolar I disorder. BALANCE is a bipolar disorder and major depressive disorder [MDD]).33 randomized, open-label trial of lithium, divalproex sodium, They found that lithium-treated patients had significantly or the combination for maintenance treatment in bipolar dis- fewer suicides and deaths from all causes. In an examination order. Participants were stabilized on both drugs during a 4 to of 32 trials, 1,389 patients were randomly assigned to lithium 8-week open-label run-in phase (to screen for tolerability), treatment and 2,069 to other compounds. Seven trials reported then randomized to continue on lithium (titrated to at any deaths by suicide; subjects treated with lithium were less 0.4–1.0 mmol/L), divalproex sodium (750 mg, 1,250 mg, or likely to commit suicide (2 versus 11 suicides; odds ratio = 0.26; valproic acid serum concentration at least 50 µg/ml), or the 95% CI = 90.0 to 0.77). When suicides plus suicidal behavior combination, with the primary outcome measure being time (i.e., deliberate self-harm) were examined, the results also to intervention for a mood episode. While combination treat- favored the lithium-treated group (odds ratio = 0.21; 95% ment was superior to divalproex sodium (hazard ratio 0.59, CI = 0.08 to 0.50). In the 11 trials reporting any deaths, all- 95% CI 0.42–0.83) and lithium was also superior to dival- cause mortality was lower in the lithium group (data from 11 proex (HR 0.71, 95% CI 0.71–1.00), combination therapy was trials; 9 versus 22 deaths; odds ratio = 0.42, 95% CI = 10.2 to not superior to lithium (HR 0.82, 95% CI 0.58–1.17). This 0.87). suggests that the role for valproate monotherapy (i.e., not in In an analysis of databases from two large health mainte- combination with lithium) is limited, and that lithium alone nance organizations in the US, Goodwin and Goldstein34 or in combination with valproate is the preferred treatment. found a strong effect favoring lithium compared to divalproex There remains some controversy about what adequate sodium or other anticonvulsants. The incidence of emergency maintenance lithium levels should be. In order to minimize department admissions for suicide attempts (31.3 versus 10.8 adverse effects and to increase patient acceptance of lithium per 1,000 person-years; P < 0.001), suicide attempts resulting treatment, lowest effective doses should be the goal. A ran in hospitalization (10.5 versus 4.2 per 1,000 person-years; domized, double-blind study by Gelenberg and co-workers27 P < 0.001), and death by suicide (1.7 versus 0.7 per 1,000 stabilized patients on a standard serum level of lithium (0.8 person-years; P = 0.04) was lower in the group receiving at to 1 mmol/L), then assigned them to either remain at that least one prescription for lithium. When adjusted for a number level or to be maintained with a lower serum lithium level of demographic factors (including age and psychiatric and (0.4 to 0.6 mmol/L). Patients in the higher lithium level group medical co-morbidity), they found that the risk of death by had fewer relapses than those randomly assigned to lower suicide was 2.7 times that for patients prescribed divalproex lithium levels.27 A re-analysis of the data, however, controlling for a diagnosis of bipolar disorder compared to those pre- for the rate at which the lithium dose was lowered, found no scribed lithium (95% CI = 1.1 to 6.3; P = 0.04).34 The non- difference between groups, suggesting that lower maintenance randomized nature of the sample, however, leaves open the lithium levels may be adequate for some patients.28 concern that the groups were clinically different, and the results confounded by indication.35 For instance, it is not Lithium in Rapid-cycling Bipolar Disorder knownw ho many of the patients in the divalproex group had previously failed to respond to lithium and were Rapid cycling is no longer included as a course specifier in thus a treatment-resistant group, and whether there were DSM-5, but continues to be used conceptually by clinicians. co-morbidities (such as anxiety disorders, personality disor- Rapid-cycling is defined in DSM-IV-TR as four or more distinct ders, or substance use disorders) that were present to a greater mood episodes (either of opposite poles, or of the same pole degree in the non–lithium-treated subjects. In any case, the after at least 8 weeks of partial or full recovery) within a results are strongly in favor of lithium and are consistent with 12-month period; patients with this course are notoriously other examinations of the effect of lithium on suicide. difficult to treat and to stabilize. Some have concluded that Another meta-analysis of 33 studies investigating long- lithium is less effective than other drugs (e.g., divalproex term lithium treatment between the years 1970 and 2000

Downloaded for Rohul Amin ([email protected]) at Uniformed Services Univ of the Health Sciences from ClinicalKey.com by Elsevier on September 21, 2018. For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved. 528 PART X Treatment Approaches yields a result that favors lithium as a potential means of potentially limiting the utility of the treatment, and the suicide prevention.36 Of the 19 studies comparing groups with follow-up period was short. and without lithium treatment, 18 found a lower risk of suicide in the treatment group and one had no suicides in Other Uses of Lithium either group.36 Overall, the meta-analysis demonstrated a 13-fold reduction in suicidality for patients with an affective While well validated for use in the treatment of bipolar disor- illness, leading to a largely reduced suicide risk (which never- der, lithium has been studied with greater or lesser success theless remained larger than that estimated for the general through randomized trials in the treatment of other psychiat- population). The rates of suicide associated with lithium treat- ric illnesses, including unipolar MDD, schizophrenia, and ment (0.109% to 0.224% annually) are 10 times greater than alcohol dependence. the international base rate (0.017%).36 Lithium discontinuation itself may increase suicide risk. Augmentation of Antidepressants in Treatment- Rapid or accelerated lithium discontinuation (as may be prac- ticedy b non-compliant individuals who decide to simply stop refractory Major Depressive Disorder taking their medications) may increase risk for suicidal behav- The use of lithium as an agent to prevent relapse in MDD has ior. In a sample of 165 patients who decided to discontinue been somewhat controversial, although the accumulation of lithium for a variety of reasons (whether electively or for some evidence suggests that it may be effective in a small number medical reason), there was a 14-fold increase in all suicidal of difficult-to-treat, refractory patients. While some of the acts following discontinuation of lithium.37 It is unclear earlier placebo-controlled trials found lithium augmentation whether the risk of suicide following lithium discontinuation to be of no benefit, a few larger studies with improved meth- exceeds that found in untreated affective illness. Lithium dis- odology suggested a benefit for lithium over placebo; a meta- continuation may increase suicidal behavior due to higher analysis that included nine double-blind, placebo-controlled relapse rates, higher than would be expected even if subjects trials found a statistically significant difference in response had been treated with placebo or had been on no medication rates to lithium augmentation compared to placebo in trials at all.23 ,Ultimately although the effects of lithium are promis- that used a minimum lithium carbonate dose of 800 mg/day ing in the realm of suicide prevention, they have not yet been or a serum level of 0.5 mEq/L or greater.42 More recently, definitively determined (and are likely never to be). however, a small, double-blind, placebo-controlled trial of lithium augmentation found no benefit for lithium, and a Lithium in Children and Adolescents recent large head-to-head comparison between lithium carbonate and triiodothyronine (T3) found that only a small Pediatric Bipolar Disorder proportion of patients improved with lithium.43,44 43 There are no randomized, controlled, parallel group trials of Nierenberg and associates found no benefit for lithium lithium treatment of acute mania in children or adolescents. in a placebo-controlled trial. Thirty-five non-responders to 6 This is unfortunate, as the use of lithium in children without weeks of treatment with nortriptyline were treated with clear benefit may be inappropriate due to its known side lithium carbonate or placebo; only 12.5% of the lithium- effects. Open-label data are suggestive of an anti-manic effect, treated subjects improved, compared to 20% on placebo. As but without randomization or a control group, these data are part of the Sequential Treatment Alternatives to Relieve difficult to interpret.38 Kafantaris and co-workers39 published Depression (STAR*D) study, lithium carbonate was compared a discontinuation study of adolescents with acute mania who to T3 as an augmentation strategy in a 14-week randomized, were stabilized for 4 weeks on lithium, then randomly assigned open-label trial for 142 patients who had failed to improve to double-blind discontinuation over 2 weeks. They found no on citalopram followed by a second treatment (either a switch to another antidepressant or augmentation with another differences in rate of symptom-worsening between the group 44 continued on lithium (10 of 19, 52.6%) versus the group agent), and found similarly low response rates for lithium. switched to placebo (13 of 21, 61.9%), but their follow-up Remission rates were 15.9% with lithium augmentation periody ma have been too short to detect a meaningful (mean dose = 859.8 mg, SD = 373.1) and 24.7% with T3 aug- difference.39 mentation (mean dose = 45.2 µg, SD = 11.4) after a mean A small, randomized, placebo-controlled, 6-week study of of 9.6 weeks of treatment, although the difference between lithium in adolescents (n = 25, average age 16.3 years) with treatments was not statistically significant. Lithium, however, bipolar disorder and substance dependence disorder (includ- was more frequently associated with side effects than was ing alcohol, cannabis, stimulants, and sedative/hypnotics) T3 (P = 0.045), and more participants in the lithium group showed that lithium (average serum level 0.9 mEq/L) appears left treatment because of side effects (23.2% versus 9.6%; to improve both disorders.40 Urine screens and measures of P = 0.027). psychopathology improved in this group, although the results were preliminary and have yet to be replicated in a larger Relapse Prevention in Major Depressive Disorder sample using more rigorous methodology. Several efforts have been made to examine the potential Conduct Disorder benefit of lithium in preventing relapse in MDD. In an early study, Prien and colleagues45 found no additional benefit of Lithium has been of some interest for use in treating symp- the combination of lithium and imipramine over imipramine toms of aggression associated with conduct disorder in chil- monotherapy. Lithium monotherapy was less successful than dren. In the largest study of this, 40 children (33 boys, 7 girls, imipramine or combination treatment in this study. with an average age of 12.5 years) were randomly assigned to In a small cohort (n = 29) of patients who responded to lithium or placebo for 4 weeks.41 Sixteen of 20 subjects in the lithium augmentation in a 6-week open-label treatment phase lithium-treated group were considered responders on consen- and remained well over a 2- to 4-week stabilization period, sus ratings compared to 6 of 20 in the placebo group those randomized to continue lithium had lower relapse (P = 0.04), while Overt Aggression Scale scores decreased sig- rates in 4 months of follow-up (0 of 14) compared to nificantly for the lithium group compared to the placebo those on placebo (7 of 15, including one suicide).42 Serum group (P = 0.04). There were significant side effects, however, lithium levels were moderate, averaging between 0.65 and

0.72 mmol/L. One methodological limitation of the study is TABLE 47-1 Medications That Commonly Interact with Lithium that lithium was discontinued over a 1-week washout period, 47 raising the possibility that rapid lithium discontinuation may Medication Effect have contributed to the high relapse rate (and perhaps to the COMMON suicide in the placebo arm). Diuretics: thiazide, loop (e.g., furosemide) ↑ ACE inhibitors ↑ NSAIDs ↑ Lithium in Psychotic Disorders Metronidazole ↑ Tetracycline ↑ Lithium is ineffective in the treatment of psychosis, but it has Diuretics: osmotic, potassium-sparing, ↓ been studied as an adjunct to antipsychotics in the treatment acetazolamide of patients with schizophrenia and schizoaffective disorders. Theophylline/aminophylline ↓ In the last published randomized trial of lithium in schizo- Drugs that alkalinize urine ↓ phrenia, a double-blind, placebo-controlled, crossover design Caffeine ↓ study1 of 2 patients with schizophrenia, little benefit was RARE BUT NOTABLE found for lithium compared to placebo on most measures of Antipsychotics Risk of neurotoxicity 46 illness severity. The benefit for lithium in patients with Bupropion Increased seizure risk schizophrenia or schizoaffective disorder has only been SSRIs Serotonin syndrome studied in samples of a similar size, and results are not defini- Iodide salts Hypothyroidism 47–50 tively in favor of its use. ACE, Angiotensin-converting enzyme; NSAIDs, non-steroidal anti-inflammatory drugs; SSRIs, selective serotonin reuptake Lithium in Alcohol Dependence inhibitors. A Veterans Administration study of lithium in depressed and non-depressed alcoholics failed to find a benefit for the drug.51 In a year-long study of 286 alcoholics without depression and1 17 alcoholics with depression, no significant differences Lithium Dosing between alcoholics who took lithium and those who took placebo were found on any study measure (including number Lithium carbonate is available in 300 mg capsules or tablets, of alcoholics who became abstinent, number of days of drink- as well as 300 and 450 mg slow-release forms. A liquid, ing, number of alcohol-related hospitalizations, changes in lithium citrate, is also available, in 300 mg/5 ml form. For rating of severity of alcoholism, and change in severity of most patients, the immediate-release lithium carbonate is ini- depression). No significant differences were found when the tiated first, with a switch to another form only to maximize data from medication-compliant subjects were reviewed. tolerability if necessary. Thew narro therapeutic window for lithium treatment (see Pharmacokinetics and Pharmacodynamics, earlier in this PRINCIPLES OF LITHIUM TREATMENT chapter) complicates lithium dosing. Indeed, in its early appli- Predictors of Lithium Response cation lithium levels of 0.8 to 1.2mEq/L were advised, while more recently levels of 0.6 to 0.8—and perhaps even lower in Fewy if an reliable predictors of lithium response have been some cases—have been advocated. The optimal dose is driven, identified. Indeed, on closer look many of the purported pre- not merely by considerations of efficacy, but by tolerability as dictors have proven to be untrue. Thus, the bulk of evidence well:, that is the “best” dose for prevention of recurrence may suggests that patients with rapid-cycling bipolar disorder be too high for some patients to tolerate. Drug–drug interac- respond well to lithium. Likewise, there is little convincing tions must be considered, and a list of selected interactions is evidence that discontinuation of lithium decreases the likeli- provided in Table 47-1. hood of subsequent response to lithium. There is no evidence Typically, the clinician will begin with 600 mg once daily that males or females respond differentially to lithium. at bedtime and then increase it to 900 mg at bedtime. After 5 Individuals with a positive family history of bipolar disor- days, a lithium level is checked and the dose is increased or der (rather than schizophrenia) do appear to respond better decreased as required to attain a level in the 0.6 to 0.8 mEq/L to lithium. This is often interpreted, incorrectly, to mean range. While nomograms exist for predicting the necessary that lithium response itself is familial, which has not been lithium dose based on an initial test dose, because of concerns established. about lithium toxicity, the authors generally prefer to observe the level that results from 900 mg before titrating further. Laboratory Monitoring Some data suggest that lithium may be better tolerated by patients, and possibly less likely to cause renal complications, The standard evaluation (i.e., with evaluation of renal func- when dosed once daily. If tolerability of a single dose becomes tion, thyroid function, and cardiac rhythm) for a patient a problem—for example, in terms of sedation or gastrointes- beginning lithium flows from its major potential toxicities. tinal distress—divided dosing may be used. This need not be Specifically, standard guidelines advise checking electrolytes, morning and night; some patients prefer dinnertime and blood urea nitrogen (BUN) and creatinine, thyroid-stimulating bedtime. hormone (TSH), and an electrocardiogram (ECG) in individuals 40 years or older. Adverse Effects and Their Management Lithium levels should be checked at least 5 days after each dose change, or any time a patient reports the new onset of Hypothyroidism has been reported to be common in patients symptoms potentially suggestive of lithium toxicity. treated with lithium, with some studies describing a prevalance Guidelines also recommend following electrolytes, BUN, of up to 35%. However, more recent studies suggest that it is and creatinine up to every 2 months for the first 6 months, substantially less common, and is seen more often among with a lithium level, BUN, creatinine, and thyroid function women than men. When hypothyroidism occurs, it is typically tests every 6 months to 1 year thereafter. managed with the addition of thyroid (T4) treatment.

Lithium has a number of renal effects, which range from bothersome to potentially life-threatening. Perhaps the most BOX 47-1 Signs of Lithium Toxicity common, polyuria (including nocturia) and consequent poly- LEVELS ABOVE 1.5 mEq/L* dipsia, arise when lithium prevents distal tubule re-absorption of free water, antagonizing the actions of anti-diuretic hormone. Lethargy/fatigue When necessary, they may be addressed with careful addition Coarse tremor of a diuretic. Amiloride is often used first, as it has little effect Nausea/vomiting on lithium levels. Thiazide diuretics may also be used if Diarrhea required, though they will increase lithium levels and often Visual changes (blurring) require a dose decrease of 50% and close monitoring of lithium Vertigo levels. When edema (particularly common in the lower Hyperreflexia extremities) is seen in a patient with normal renal function, Dysarthria the diuretic spironolactone may be used, though again lithium Ataxia levels and renal function will require close monitoring. Confusion Of greater concern, long-term treatment with lithium is LEVELS ABOVE 2.5–3 mEq/L associated with reduction in creatinine clearance in about 10% Seizure of, patients indicating a decrease in glomerular filtration rate Coma (GFR). Small studies suggest that glomerulosclerosis and Arrhythmia interstitial fibrosis are seen among some patients receiving long-term lithium treatment. Beyond a certain point, this dete- *Absence of significant symptoms does not rule out toxicity; likewise, rioration may accelerate and dramatically increase the risk of some symptoms may be apparent at levels as low as 1 mEq/L. lithium toxicity, as small perturbations in GFR can have large effects on lithium levels. When creatinine rises above baseline by 25% or more, consultation with a nephrologist may be BOX 47-2 Management of Lithium Toxicity helpful to rule out other contributors to reduction in GFR (e.g., renal complications of diabetes mellitus [DM]), and If patient is comatose/obtunded, protect airway. discontinuation of lithium often becomes necessary once GFR In cases of suspected overdose, consider gastric lavage. For decreases beyond a certain point. lithium levels above 4 mEq/L,* or above 3 mEq/L and severely While, rare lithium can cause a depression of firing at the symptomatic, or in other cases where volume load will not be sino-atrial node, contributing to sinus arrhythmias. Lithium tolerated, initiate dialysis. may also cause benign ECG abnormalities, most typically an Hold further lithium doses. appearance resembling hypokalemia, including widening of Begin IV normal saline, 150–200 ml/h (as long as able to the QRS complex and increased PR interval. tolerate volume load). Weight gain is common among lithium-treated patients, Address other electrolyte abnormalities. and while the mechanism is not known, it is not merely a Follow lithium levels approximately every 2–3 h. Initiate work-up result of lithium-induced edema. Some studies suggest that up to determine cause of toxicity. to half of lithium-treated patients will experience a 5% to 10% increase in weight. As with any medication-induced weight *Recommendations vary about when to initiate dialysis, though most gain, early and aggressive intervention is warranted, focusing sources agree that levels above 4 mEq/L merit immediate dialysis. on both diet and exercise. A specific concern among lithium- treated patients with polydipsia is the consumption of sodas or juices high in sugar. Cognitive complaints are common among lithium-treated and when necessary by adding a beta-adrenergic blocker. This patients, sometimes characterized as feeling “foggy” or may be used on either an as-needed or a standing basis, “cloudy.” Adjunctive thyroid hormone is sometimes used in depending on patient preference. Propranolol is often initi- an attempt to ameliorate these symptoms; however, there is ated first to establish an effective dose, then changed to aten- no well-established means of treating such complaints, other olol for the convenience of once-daily dosing. than decreasing lithium dose. Both psoriasis and acne exacerbations have been associated with lithium treatment. Typically these do not require treat- Other Bothersome Adverse Effects ment discontinuation as they can be controlled with derma- tological treatments. Gastrointestinal adverse effects are common among lithium- Presentation of lithium toxicity develops as lithium levels treated patients, and may include upper (nausea, vomiting, rise above 1, and particularly beyond 1.2 mEq/L; initial symp- and dyspepsia) and lower (diarrhea and cramping) gastroin- tomsy ma include slurring of speech, ataxia/unsteady gait, testinal symptoms. Because slow-release formulations are confusion, and agitation. Box 47-1 shows signs and symptoms absorbed lower in the gut, they are more often associated with of lithium toxicity, and Box 47-2 shows management of acute the latter symptoms, while immediate-release formulations toxicity. are more often associated with the former symptoms. In many cases, dividing dosages or dosing with food can minimize or eliminate these symptoms. If this is not the case, a switch to Lithium in Pregnancy and Breast-feeding lithium citrate (in liquid form) may be helpful, though some Lithium use during the first trimester of pregnancy has been patients object to the taste of this preparation. associated with a potentially life-threatening cardiac abnor- Tremor is common with lithium treatment, even at thera- mality, known as Ebstein’s anomaly, a spectrum of changes peutic levels and particularly following each dose when peak that typically includes insufficiency of the tricuspid valve and levels are achieved. This tremor, which resembles a benign hypoplasia of the right ventricle. The risk appears to be about physiological tremor, may be exacerbated by caffeine and by 10-fold greater among children of lithium-treated patients anxiety. It is typically managed by changing the timing of the compared to those in the general population, though the lithium dose (to ensure that peak levels occur during sleep) absolute risk is still quite low (on the order of 1 in 2,000).

This risk must be balanced against the substantial dangers to 18. Perlis RH, Welge JA, Vornik LA, et al. Atypical antipsychotics in the fetus of a mother with uncontrolled bipolar disorder the treatment of mania: a meta-analysis of randomized, placebo- 47 during pregnancy. controlled trials. J Clin Psychiatry 67(4):509–516, 2006. The changes in maternal fluid status during pregnancy typi- 20. Nemeroff CB, Evans DL, Gyulai L, et al. Double-blind, placebo- cally cause a larger volume of distribution for lithium, which controlled comparison of imipramine and paroxetine in the treatment of bipolar depression. Am J Psychiatry 158(6):906–912, can lead to a decrease in lithium levels, entailing closer moni- 2001. toring of lithium during pregnancy and in some cases a dose 21. Prien RF, Caffey EM, Klett CJ. Prophylactic efficacy of lithium increase.t A the time of delivery, fluid shifts may likewise lead carbonate in manic-depressive illness: report of the Veterans to an increase in lithium levels. However, discontinuing Administration and National Institute of Mental Health Collabo- lithium in the peripartum period is no longer routinely rative Study Group. Arch Gen Psychiatry 28:337–341, 1973. advised, as it has become clear that the risk of bipolar recur- 22. Suppes T, Baldessarini RJ, Faedda GL, et al. Risk of recurrence rence is extremely high in this period. following discontinuation of lithium treatment in bipolar disor- Levels of lithium in breast milk are substantial, approach- der. Arch Gen Psychiatry 48:1082–1088, 1991. ing 50% of maternal plasma levels. For this reason, lithium- 23. Goodwin GM. Recurrence of mania after lithium withdrawal: implications for the use of lithium in the treatment of bipolar treated mothers should not breast-feed, as the effects of affective disorder. Br J Psychiatry 164:149–152, 1994. lithium on newborn development are not well studied and 24. Geddes JR, Burgess S, Hawton K, et al. Long-term lithium therapy the risk for toxicity is high. for bipolar disorder: systematic review and meta-analysis of randomized controlled trials. Am J Psychiatry 161:217–222, 2004. 25. Young AH, Newham JI. Lithium for maintenance therapy in Lithium Adherence bipolar disorder. J Psychopharmacol 20:17–22, 2006. Recent studies document very low rates of lithium adherence 26. Tohen M, Greil W, Calabrese JR, et al. Olanzapine versus lithium in bipolar disorder, particularly in a cohort of patients partici- in the maintenance treatment of bipolar disorder: a 12-month, pating in a health maintenance organization.52 These studies randomized, double-blind, controlled clinical trial. Am J Psychia- try 162(7):1281–1290, 2005. highlight the importance of careful lithium dosing and aggres- 27. Gelenberg AJ, Kane JM, Keller MB, et al. Comparison of standard sive management of adverse effects. andw lo serum levels of lithium for maintenance treatment of bipolar disorder. N Engl J Med 321:1489–1493, 1989. 28. Perlis RH, Sachs GS, Lafer B, et al. Effect of abrupt change from CURRENT CONTROVERSIES standard to low serum levels of lithium: a reanalysis of double- AND FUTURE DIRECTIONS blind lithium maintenance data. Am J Psychiatry 159:1155–1159, 2002. Several recent reviews have raised concern that lithium is 30. Bowden CL, Calabrese JR, McElroy SL, et al. A randomized, being under-used relative to other, potentially less effective placebo-controlled 12-month trial of divalproex and lithium in treatments. This has been attributed to lithium receiving less treatment of outpatients with bipolar I disorder. Divalproex attention in medical education, and more generally because Maintenance Study Group. Arch Gen Psychiatry 57:481–489, 2000. of the marketing efforts for newer, on-patent alternatives for 31. Calabrese JR, Bowden CL, Sachs G, et al. A placebo-controlled the treatment of bipolar disorder. 18-month trial of lamotrigine and lithium maintenance treat- Ifs lithium’ mechanism of action can be better understood, ment in recently depressed patients with bipolar I disorder. J Clin it might facilitate the identification of targets for drug therapy Psychiatry 64:1013–1024, 2003. in bipolar disorder—that is, proteins that might be targeted 32. Baldessarini RJ, Tondo L, Floris G, et al. Effects of rapid cycling on response to lithium maintenance treatment in 360 bipolar I by drugs that could exert similar effects to lithium, without its and II disorder patients. J Affect Disord 61:13–22, 2006. toxicity. Alternatively, it may be possible to explain why some 33. Cipriani A, Wilder H, Hawton K, et al. Lithium in the prevention patients fail to respond to lithium, and “sensitize” them by of suicidal behaviour and all-cause mortality in patients with combining lithium with other agents. Further mechanistic and mood disorders: a systematic review of randomized trials. Am J pharmacogenetic studies of lithium will be crucial in this Psychiatry 162:1805–1819, 2005. regard. 37. Tondo L, Baldessarini RJ. Reduced suicide risk during lithium maintenance treatment. J Clin Psychiatry 61(Suppl. 9):97–104, Access the complete reference list and multiple choice questions 2000. (MCQs) online at https://expertconsult.inkling.com 41. Malone RP, Delaney MA, Luebbert JF, et al. A double-blind placebo-controlled study of lithium in hospitalized aggressive children and adolescents with conduct disorder. Arch Gen Psychia- try 57(7):649–654, 2000. KEY REFERENCES 42. Bauer M, Döpfmer S. Lithium augmentation in treatment-resistant 4. Chalecka-Franaszek E, Chuang DM. Lithium activates the serine/ depression: meta-analysis of placebo-controlled studies. J Clin threonine kinaseAkt-1 and suppresses glutamate-induced Psychopharmacol 19:427–434, 1999. inhibition of Akt-1 activity in neurons. Proc Natl Acad Sci USA 45. Prien RF, Kupfer DJ, Mansky PA, et al. Drug therapy in the preven- 96(15):8745–8750, 1999. tion of recurrences in unipolar and bipolar affective disorders. 9. O’Brien WT, Harper AD, Jove F, et al. Glycogen synthase kinase-3 Report of the NIMH Collaborative Study Group comparing beta haploinsufficiency mimics the behavioral and molecular lithium carbonate, imipramine, and a lithium carbonate- effects of lithium. J Neurosci 24(30):6791–6798, 2004. imipramine combination. Arch Gen Psychiatry 41(11):1096–1104, 12. Prien RF, Caey EM Jr, Klett CJ. Comparison of lithium carbonate 1984. and chlorpromazine in the treatment of mania. Report of the 48. Calabrese JR, Shelton MD, Rapport DJ, et al. Lithium combined Veterans Administration and National Institute of Mental Health with neuroleptics in chronic schizophrenic and schizoaffective Collaborative Study Group. Arch Gen Psychiatry 26:146–153, patients. J Clin Psychiatry 42(3):124–128, 1981. 1972. 52. Johnson RE, McFarland BH. Lithium use and discontinuation in 15. Bowden CL, Brugger AM, Swann AC, et al. Efficacy of divalproex a health maintenance organization. Am J Psychiatry 153(8):993– vs lithium and placebo in the treatment of mania. The Depakote 1000, 1996. Mania Study Group. JAMA 271(12):918–924, 1994.

24. Geddes JR, Burgess S, Hawton K, et al. Long-term lithium therapy REFERENCES for bipolar disorder: systematic review and meta-analysis of ran- 47 1. Bauer M, Grof P, Müller-Oerlinghausen B, editors: Lithium in domized controlled trials. Am J Psychiatry 161:217–222, 2004. neuropsychiatry: the comprehensive guide, London, 2006, Taylor & 25. Young AH, Newham JI. Lithium for maintenance therapy in Francis. bipolar disorder. J Psychopharmacol 20:17–22, 2006. 2. Berridge MJ, Downes CP, Hanley MR. Neural and developmental 26. Tohen M, Greil W, Calabrese JR, et al. Olanzapine versus lithium actions of lithium: a unifying hypothesis. Cell 59(3):411–419, in the maintenance treatment of bipolar disorder: a 12-month, 1989. randomized, double-blind, controlled clinical trial. Am J Psychia- 3. Williams RS, Eames M, Ryves WJ, et al. Loss of a prolyloligopep- try 162(7):1281–1290, 2005. tidase confers resistance to lithium by elevation of inositol (1,4,5) 27. Gelenberg AJ, Kane JM, Keller MB, et al. Comparison of standard trisphosphate. EMBO J 18(10):2734–2745, 1999. and low serum levels of lithium for maintenance treatment of 4. Chalecka-Franaszek E, Chuang DM. Lithium activates the serine/ bipolar disorder. N Engl J Med 321:1489–1493, 1989. threonine kinaseAkt-1 and suppresses glutamate-induced inhibi- 28. Perlis RH, Sachs GS, Lafer B, et al. Effect of abrupt change from tion of Akt-1 activity in neurons. Proc Natl Acad Sci USA standard to low serum levels of lithium: a reanalysis of double- 96(15):8745–8750, 1999. blind lithium maintenance data. Am J Psychiatry 159:1155–1159, 5. Klein PS, Melton DA. A molecular mechanism for the effect of 2002. lithium on development. Proc Natl Acad Sci USA 93(16):8455– 29. McElroy SL, Keck PE Jr, Pope HG Jr, et al. Valproate in the treat- 8459, 1996. ment of bipolar disorder: literature review and clinical guide- 6. Stambolic V, Ruel L, Woodgett JR. Lithium inhibits glycogen syn- lines. J Clin Psychopharmacol 12(1 Suppl.):42S–52S, 1992. thase kinase-3 activity and mimics wingless signaling in intact 30. Bowden CL, Calabrese JR, McElroy SL, et al. A randomized, cells. Curr Biol 6(12):1664–1668, 1996. placebo-controlled 12-month trial of divalproex and lithium in 7. Lucas FR, Salinas PC. WNT-7a induces axonal remodeling and treatment of outpatients with bipolar I disorder. Divalproex increases synapsin I levels in cerebellar neurons. Dev Biol Maintenance Study Group. Arch Gen Psychiatry 57:481–489, 2000. 192(1):31–44, 1997. 31. Calabrese JR, Bowden CL, Sachs G, et al. A placebo-controlled 8. Phiel CJ, Klein PS. Molecular targets of lithium action. Annu Rev 18-month trial of lamotrigine and lithium maintenance treat- Pharmacol Toxicol 41:789–813, 2001. ment in recently depressed patients with bipolar I disorder. J Clin 9. O’Brien WT, Harper AD, Jove F, et al. Glycogen synthase kinase-3 Psychiatry 64:1013–1024, 2003. beta haploinsufficiency mimics the behavioral and molecular 32. Baldessarini RJ, Tondo L, Floris G, et al. Effects of rapid cycling effects of lithium. J Neurosci 24(30):6791–6798, 2004. on response to lithium maintenance treatment in 360 bipolar I 10. Hashimoto R, Hough C, Nakazawa T, et al. Lithium protection and II disorder patients. J Affect Disord 61:13–22, 2006. against glutamate excitotoxicity in rat cerebral cortical neurons: 33. Cipriani A, Wilder H, Hawton K, et al. Lithium in the prevention involvement of NMDA receptor inhibition possibly by decreasing of suicidal behaviour and all-cause mortality in patients with NR2B tyrosine phosphorylation. J Neurochem 80(4):589–597, mood disorders: a systematic review of randomized trials. Am J 2002. Psychiatry 162:1805–1819, 2005. 11. Schou M, Juel-Nielsen N, Stromgren E, et al. The treatment of 34. Goodwin FK, Goldstein MA. Optimizing lithium treatment in manic psychoses by the administration of lithium salts. J Neurol bipolar disorder: a review of the literature and clinical recom- Neurosurg Psychiatry 17(4):250–260, 1954. mendations. J Psychiatr Pract 9(5):333–343, 2003. 12. Prien RF, Caey EM Jr, Klett CJ. Comparison of lithium carbonate 35. Walker AM. Confounding by indication. Epidemiology 7:335–336, and chlorpromazine in the treatment of mania. Report of the 1996. Veterans Administration and National Institute of Mental Health 36. Baldessarini RJ, Tondo L, Hennen J. Treating the suicidal patient Collaborative Study Group. Arch Gen Psychiatry 26:146–153, with bipolar disorder. Reducing suicide risk with lithium. Ann NY 1972. Acad Sci 932:24–38, 2001. 13. Poolsup N, Li Wan PA, De Oliveira IR. Systematic overview of 37. Tondo L, Baldessarini RJ. Reduced suicide risk during lithium lithium treatment in acute mania. J Clin Pharm Ther 25:139–156, maintenance treatment. J Clin Psychiatry 61(Suppl. 9):97–104, 2000. 2000. 14. Grunze H. Lithium in the acute treatment of bipolar disorders: a 38. Kowatch RA, Suppes T, Carmody TJ, et al. Effect size of lithium, stock-taking. Eur Arch Psychiatry Clin Neurosci 253(3):115–119, divalproex sodium, and carbamazepine in children and adoles- 2003. cents with bipolar disorder. J Am Acad Child Adolesc Psychiatry 15. Bowden CL, Brugger AM, Swann AC, et al. Efficacy of divalproex 39(6):713–720, 2000. vs lithium and placebo in the treatment of mania. The Depakote 39. Kafantaris V, Coletti DJ, Dicker R, et al. Lithium treatment of acute Mania Study Group. JAMA 271(12):918–924, 1994. mania in adolescents: a placebo-controlled discontinuation 16. Kushner SF, Khan A, Lane R, et al. Topiramate monotherapy in study. J Am Acad Child Adolesc Psychiatry 43(8):984–993, 2004. the management of acute mania: results of four double-blind 40. Geller B, Cooper TB, Sun K, et al. Double-blind and placebo- placebo-controlled trials. Bipolar Disord 8(1):15–27, 2006. controlled study of lithium for adolescent bipolar disorders with 17. Frye MA, Ketter TA, Leverich GS, et al. The increasing use of poly- secondary substance dependency. J Am Acad Child Adolesc Psychia- pharmacy for refractory mood disorders: 22 years of study. J Clin try 37(2):171–178, 1998. Psychiatry 61(1):9–15, 2000. 41. Malone RP, Delaney MA, Luebbert JF, et al. A double-blind 18. Perlis RH, Welge JA, Vornik LA, et al. Atypical antipsychotics in placebo-controlled study of lithium in hospitalized aggressive the treatment of mania: a meta-analysis of randomized, placebo- children and adolescents with conduct disorder. Arch Gen Psychia- controlled trials. J Clin Psychiatry 67(4):509–516, 2006. try 57(7):649–654, 2000. 19. Bauer MS, Mitchner L. What is a “mood stabilizer”? An evidence- 42. Bauer M, Döpfmer S. Lithium augmentation in treatment-resistant based response. Am J Psychiatry 161:3–18, 2004. depression: meta-analysis of placebo-controlled studies. J Clin 20. Nemeroff CB, Evans DL, Gyulai L, et al. Double-blind, placebo- Psychopharmacol 19:427–434, 1999. controlled comparison of imipramine and paroxetine in the treat- 43. Nierenberg A, Papakostas GI, Petersen T, et al. Lithium augmenta- ment of bipolar depression. Am J Psychiatry 158(6):906–912, tion of nortriptyline for subjects resistant to multiple antidepres- 2001. sants. J Clin Psychopharmacol 23(1):92–95, 2003. 21. Prien RF, Caffey EM, Klett CJ. Prophylactic efficacy of lithium 44. Nierenberg AA, Fava M, Trivedi MH, et al. A comparison of carbonate in manic-depressive illness: report of the Veterans lithium and T(3) augmentation following two failed medication Administration and National Institute of Mental Health Collabo- treatments for depression: a STAR*D report. Am J Psychiatry rative Study Group. Arch Gen Psychiatry 28:337–341, 1973. 163(9):1519–1530, 2006. 22. Suppes T, Baldessarini RJ, Faedda GL, et al. Risk of recurrence 45. Prien RF, Kupfer DJ, Mansky PA, et al. Drug therapy in the preven- following discontinuation of lithium treatment in bipolar disor- tion of recurrences in unipolar and bipolar affective disorders. der. Arch Gen Psychiatry 48:1082–1088, 1991. Report of the NIMH Collaborative Study Group comparing 23. Goodwin GM. Recurrence of mania after lithium withdrawal: lithium carbonate, imipramine, and a lithium carbonate- implications for the use of lithium in the treatment of bipolar imipramine combination. Arch Gen Psychiatry 41(11):1096–1104, affective disorder. Br J Psychiatry 164:149–152, 1994. 1984.

46. Terao T, Oga T, Nozaki S, et al. Lithium addition to neuroleptic 50. Collins PJ, Larkin EP, Shubsachs AP. Lithium carbonate in chronic treatment in chronic schizophrenia: a randomized, double-blind, schizophrenia—a brief trial of lithium carbonate added to neu- placebo-controlled, cross-over study. Acta Psychiatr Scand 92(3): roleptics for treatment of resistant schizophrenic patients. Acta 220–224, 1995. Psychiatr Scand 84(2):150–154, 1991. 47. Biederman J, Lerner Y, Belmaker RH. Combination of lithium 51. Dorus W, Ostrow DG, Anton R, et al. Lithium treatment of carbonate and haloperidol in schizo-affective disorder: a control- depressed and nondepressed alcoholics. JAMA 262(12):1646– led study. Arch Gen Psychiatry 36(3):327–333, 1979. 1652, 1989. 48. Calabrese JR, Shelton MD, Rapport DJ, et al. Lithium combined 52. Johnson RE, McFarland BH. Lithium use and discontinuation in with neuroleptics in chronic schizophrenic and schizoaffective a health maintenance organization. Am J Psychiatry 153(8):993– patients. J Clin Psychiatry 42(3):124–128, 1981. 1000, 1996. 49. Lerner Y, Mintzer Y, Schestatzky M. Lithium combined with haloperidol in schizophrenic patients. Br J Psychiatry 153:359–362, 1988.

MULTIPLE CHOICE QUESTIONS of lithium bromide to treat manic or agitated patients, though he later downplayed the importance of lithium. In the early 47 Select the appropriate answer. 1900s a Danish physician, Lange, published a case series reporting the treatment of manic patients with lithium car- Q1 In which year was lithium first approved by the Food bonate. There is little evidence that lithium was studied and Drug Administration (FDA) for the treatment of further, however, until Garrod proposed that lithium urate mania? could be used to treat gout, opening the door to its broader ○ 1949 therapeutic application. ○ 1956 Unfortunately, despite some early study by Mogen Schou and others, lithium’s wider adoption in the United States was hin- ○ 1963 dered by concerns about lithium toxicity. Lithium chloride ○ 1970 had been used as a sodium substitute in the 1940s, until several deaths were reported from lithium toxicity among ○ 1982 hyponatremic patients. Q2 True or False. Hyponatremia will lead to an increase in Lithium was initially perceived as too dangerous for clinical one’s lithium level. application, and it was only in 1970 that lithium was approved by the U.S. Food and Drug Administration (FDA) for the treat- ○ True ment of mania. ○ False Q2 The answer is: True. Q3 Which of the following lithium levels are generally In the kidney, lithium is primarily reabsorbed in the proximal considered to be solidly in the therapeutic range for renal tubules, with some subsequent absorption in the loop treatment of mood disorders? of Henle. Importantly, in contrast to sodium, no significant ○ 0.1 to 0.3 mEq/L absorption occurs in the distal tubules. Therefore, thiazide diuretics, which act distally, will tend to increase lithium levels ○ 0.4 to 0.6 mEq/L by up to 50%, while those that act more proximally generally ○ 0.6 to 0.8 mEq/L have less effect on lithium levels. ○ 1.0 to 1.2 mEq/L More broadly, hydration status can affect lithium levels: individuals who become salt-avid (e.g., because of hypovolemia ○ 1.3 to 1.5 mEq/L or hyponatremia, perhaps in the context of vomiting, diarrhea, or long-distance running) will cause their lithium levels to Q4 Which of the following is/are the LEAST common side increase. effect(s) of lithium treatment? Q3 The answer is: 0.6 to 0.8 mEq/L. ○ Cognitive complaints Lithium carbonate is available in 300 mg capsules or tablets, ○ Gastrointestinal disturbance as well as 300 mg and 450 mg slow-release forms. A liquid, ○ Sinus arrhythmia lithium citrate, is also available, in 300 mg/5 ml form. For most patients, the immediate-release lithium carbonate is ini- ○ Tremor tiated first, with a switch to another form only to maximize ○ Weight gain tolerability if necessary. The narrow therapeutic window for lithium treatment com- Q5 Which of the following has been associated with plicates lithium dosing. Indeed, in its early application lithium lithium use during the first trimester of pregnancy? levels of 0.8 to 1.2 mEq/L were advised, while more recently ○ Ebstein’s anomaly levels of 0.6 to 0.8 mEq/L—and perhaps even lower in some cases—have been advocated. The optimal dose is driven, not ○ Edelstein’s anomaly merely by considerations of efficacy, but by tolerability as well, ○ DaCosta’s syndrome that is, the “best” dose for prevention of recurrence may be too high for some patients to tolerate. ○ Goldberg’s anomaly Typically, the clinician will begin 600 mg once daily at bedtime ○ Goldstein’s anomaly and then increase it to 900 mg at bedtime. After 5 days, a lithium level is checked and the dose is increased or decreased as required to attain a level in the 0.6 to 0.8 mEq/L range. MULTIPLE CHOICE ANSWERS Q4 The answer is: Sinus arrhythmia. Q1 The answer is: 1970. While rare, lithium can cause a depression of firing at the The history of lithium’s use in psychiatry is instructive, both sinoatrial node, contributing to sinus arrhythmias. Lithium in understanding its present role and in tracing the develop- may also cause benign ECG abnormalities, most typically an ment of psychopharmacology in general. The first modern appearance resembling hypokalemia, including widening of description of the application of lithium to treat mania the QRS complex and an increased PR interval. occurred in 1949, by an Australian named John Cade, who observed that lithium had calming effects on animals and Weight gain is common among lithium-treated patients, and then treated a series of 10 agitated manic patients. In fact, while the mechanism is not known, it is not merely a result however, descriptions of lithium treatment date back to at of lithium-induced edema. Some studies suggest that up to least the U.S. Civil War. An 1883 textbook by Union Army half of lithium-treated patients will experience a 5% to 10% Surgeon General William Hammond recommended the use increase in weight.

Cognitive complaints are common among lithium-treated anxiety. It is typically managed by changing the timing of the patients, sometimes characterized as feeling “foggy” or lithium dose (to ensure peak levels during sleep) and when “cloudy.” Adjunctive thyroid hormone is sometimes used in necessary by adding a beta-adrenergic blocker. This may be an attempt to ameliorate these symptoms; however, there is used on either an as-needed or a standing basis, depending no well-established means of treating such complaints, other on patient preference. than decreasing the lithium dose. Q5 The answer is: Ebstein’s anomaly. Gastrointestinal side effects are common among lithium- treated patients, and may include upper (nausea, vomiting, Lithium use during the first trimester of pregnancy has been and dyspepsia) and lower (diarrhea and cramping) gastroin- associated with a potentially life-threatening cardiac abnor- testinal symptoms. Because slow-release formulations are mality, known as Ebstein’s anomaly, a spectrum of changes absorbed lower in the gut, they are more often associated with that typically includes insufficiency of the tricuspid valve, as the latter symptoms, while immediate-release formulations well as hypoplasia of the right ventricle. The risk appears to are more often associated with the former symptoms. In many be about ten-fold greater among children of lithium-treated cases, dividing dosages or dosing with food can minimize or patients compared to those in the general population, though eliminate these symptoms. the absolute risk is still quite low (on the order of 1 in 2,000). This risk must also be balanced against the substantial dangers Tremor is common with lithium treatment, even at therapeu- to the fetus in a mother with uncontrolled bipolar disorder tic levels and particularly following each dose when peak during pregnancy. levels are achieved. This tremor, which resembles a benign physiological tremor, may be exacerbated by caffeine and by