Treatment with Mood Stabilizers

Lithium was used in the mid-19th century for such varied disorders as gout, diabetes, and rheumatism. With the seminal reports of Cade (1) and Schou et al. (2) almost a century later, it emerged as the standard of treatment for bipolar disorder. John Cade, an Australian physician, serendipitously discovered the antimanic properties of this monovalent cation when he injected lithium urate into guinea pigs. Mistaking toxicity for sedation in the animals, he then used it successfully in an open trial with manic patients (3). Mogens Schou, following up on Cade’s report, was the first European to employ lithium in a series of trials using increasing degrees of methodological rigor (4,5,6 and 7). The term normothymic was proposed by Schou to describe the action of lithium against both phases of a bipolar disorder as well as its ability to prevent recurrences of unipolar depressive disorder (4).

In another serendipitous finding, Cade noted that the first patient treated with lithium relapsed when medication was withdrawn (3). From this, he inferred that lithium may also be effective for maintenance treatment. Baastrup, Schou’s colleague, carried out the first definitive study of its prophylactic properties (5,6 and 7). The ability of lithium to decrease the rate of recurrence in both unipolar and bipolar disorders was then confirmed in a series of studies by Hartigan and Baastrup (6,7 and 8).

Unfortunately, just as lithium had been toxic when used as a salt substitute in cardiac patients, it also caused problems when used for mania. Indeed, until its safe use was mastered, Cade is said to have banned lithium in his own hospital, regarding it as too dangerous for humans.

Lithium was reintroduced in the United States by Gershon in the late 1960s, and since then has remained a standard therapy for bipolar disorder (9). A significant proportion of patients, however, do not sufficiently benefit from or cannot tolerate lithium. This recognition led to the reexamination and clarification of alternative monotherapy and copharmacy strategies (10).

A second strategy is the use of anticonvulsants as mood stabilizers (11). This includes agents such as valproate (VPA), lamotrigine (LTG), carbamazepine (CBZ), and perhaps other anticonvulsants (e.g., oxcarbazepine, levetiracetam, zonisamide). It is of interest that some of these treatments also have a history dating back to the 1960s. For example, the first empirical investigation of CBZ was performed by Dehing in 1968 (12). Although studying the behavioral effects of this anticonvulsant in epileptic patients, he noted that it had antiaggressive properties. Interestingly, lithium also has antiaggressive properties, but its antimanic effect was discovered first. The reverse is true for CBZ, which subsequently was used for the treatment of mania by investigators in Japan (13,14). As Schou in Europe and Gershon in the United States are credited with the introduction of lithium, Ballenger and Post (15) similarly deserve credit for introducing CBZ to the United States.

VPA was first used as a mood stabilizer by Lambert et al. (16) in 1966. Their favorable experience in a heterogeneous population of psychiatric patients was followed by several open and controlled trials. The success of this agent spawned a series of trials with other anticonvulsants, some of which have also received U.S. Food and Drug Administration (FDA) approval
(i.e., LTG for maintenance treatment of bipolar disorder and the extended-release formulation of CBZ for acute mania).

TABLE 10-1 MOOD STABILIZERS FOR TREATMENT OF BIPOLAR DISORDER

Class/Trade Name	Generic Name (Indication)	Dosage (Average Range, Orally, Per Day)
Cation
Lithonate, others	Lithium (AM, M)	600-2,700 mg
Anticonvulsant
Depakote	Divalproex (AM)	750-2,500 mg
Lamictal	Lamotrigine (M)	50-400 mg
Equetro	Carbamazepine (AM)	400-1,600 mg
Antipsychotic
Thorazine	Chlorpromazine (AM)	300-900 mg
Zyprexa	Olanzapine (AM, M, ADJ)	10-20 mg
Risperdal	Risperidone (AM, ADJ)	2-6 mg
Seroquel	Quetiapine (AM, AD, ADJ)	300-800 mg
Abilify	Aripiprazole (AM, M)	10-30 mg
Geodon	Ziprasidone (AM, M)	80-160 mg
Saphris	Asenapine (AM)	10-20 mg
Antipsychotic Plus Antidepressant
Symbyax	OLZ/FLU (AD)	6 or 12 mg (OLZ)/25 or 50 mg (FLU)
AM, acute mania; M, maintenance therapy; ADJ, adjunct; AD, acute depression; OLZ, olanzapine; FLU, fluoxetine.

A third strategy is the use of second-generation antipsychotics (SGAs) for mood stabilization (17,18). In support, a series of class I, similarly designed, controlled trials found that olanzapine (OLZ), risperidone (RISP), quetiapine (QTP), aripiprazole (ARIP), ziprasidone (ZPD), and asenapine (ASEN) possess antimanic effects. There are also data indicating that some of these agents benefit bipolar depression (e.g., OLZ plus fluoxetine; QTP monotherapy) and decrease relapse rates in bipolar patients (e.g., OLZ, ARIP, QTP, ZPD). These properties appear to be separate from their antipsychotic effects, indicating that they also qualify as mood stabilizers.

In summary, although lithium revolutionized the treatment of bipolar disorder, its narrow therapeutic index, numerous adverse effects, and relative ineffectiveness in a large proportion of bipolar patients led to an expanding number of alternative approaches, including

Anticonvulsants (e.g., VPA, LTG, CBZ)
SGAs (e.g., OLZ, RISP, QTP)
Therapeutic neuromodulation (e.g., electroconvulsive therapy [ECT], vagus nerve stimulation [VNS], transcranial magnetic stimulation [TMS])
Investigational therapies (19,20,21,22,23,24,25,26 and 27) (see later section)

Table 10-1 lists the drugs presently approved by the FDA for the management of various phases of bipolar disorder, primarily based on the results of industry-sponsored, randomized, controlled trials.

The NIH-sponsored Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) provides an additional perspective on these data. This “pragmatic” clinical study evaluated a large cohort (n = 4,360) of patients (bipolar I, II, NOS, cyclothymic) for up to 5 years (28). The goal was to assess the effectiveness of recommended treatment approaches applied in a uniform and optimal manner. These strategies are based on the controlled trial literature, published guidelines, and clinical experience. Further, the STEP-BD provides information on relapse/recurrence, comorbidities as well as both pharmacological and psychosocial treatments (29). Thus, we will include data from this study in the following sections as appropriate.

Pharmacodynamics

The monoamine hypothesis of mood disorders was developed in the mid-1960s (30,31). Subsequently, a great wealth of data was generated to test this theory (see also Chapter 7). Because the development of psychotropic agents is based on the concept that altering
neurotransmitters such as norepinephrine (NE) or serotonin (5-HT) benefits both depression and mania, it is not surprising that the action of these drugs support the original concept. It is also recognized that receptor changes modulated by alteration in genetic expression are slower to occur, are more sustained, and may be more consistent with the time course of mood improvement. Models to further clarify the underlying pathophysiology of bipolar disorder also incorporate such factors as genetic vulnerability and cyclicity (32,33).

Although the history of psychiatry contains numerous preliminary findings of biological abnormalities attributed to a given diagnostic entity, the great majority are not replicated. The technology to record blood flow, biological activity, and structural and cellular changes through functional magnetic resonance imaging (fMRI), positron emission tomography (PET), single photon emission computerized tomography (SPECT), and magnetic resonance spectroscopy (MRS), as well as molecular and genetic studies, all provide the potential to better “localize” pathophysiological processes.

NEUROTRANSMITTER AND RELATED HYPOTHESES

Although the therapeutic mechanisms of various mood stabilizers are unknown, studies concentrate on

The classic neurotransmitters (e.g., NE, dopamine [DA] 5-HT, acetylcholine [ACh]) implicated in affective disturbances
Cellular processes involving the classic ligand-receptor interaction and beyond (e.g., G-proteins, second messenger systems)
Subsequent intracellular events (e.g., protein kinase C [PKC] activity, neurotropic effects)
The modulating interactive processes among various neurotransmitter systems, neurohormones, and genetic influences
The role of circadian rhythms

Mood stabilizers interact with several neurotransmitters, including the catecholaminergic, indolaminergic, cholinergic, γ-aminobutyric acid, and glutamatergic systems. Still, the exact mechanisms by which mood stabilizers impact the biological substrates subserving mood disorders remain unclear.

Catecholamine System

Lithium decreases the β-adrenergic receptor number, consistent with the noradrenergic downregulation hypothesis. Lithium also blocks DA receptor supersensitivity, consistent with the postulate that mania is associated with an increased sensitivity of catecholamine receptors.

Lithium has antithyroid effects, including blocking the release of thyroid hormones, which potentiate β-noradrenergic receptor sensitivity. This led to the speculation that excessive thyroid activity may contribute to mania in susceptible patients and that the antimanic effect of lithium is, at least in part, due to its antithyroid action (34). In this context, CBZ also impedes thyroid function.

Another approach is the investigation of agents such as the antihypertensive, clonidine, which stimulates α₂-noradrenergic, presynaptic receptors, setting into motion a short-loop, negative-feedback process. This culminates in a shutdown of tyrosine hydroxylase, the ratelimiting enzyme for the production of the catecholamines, slowing their synthesis and release. Further, this drug acts in a rapid fashion and is highly specific for the locus ceruleus, the central nervous system (CNS) location with the richest concentration of NE-containing neurons. As a result, it rapidly and effectively shuts down NE production.

The DA system is also frequently invoked in theories of mood dysregulation. This neurotransmitter’s role in the reward circuitry and the effectiveness of antipsychotics in treating bipolar disorder are two lines of evidence that support various related hypotheses (35).

Indolamine System

Lithium facilitates 5-HT, perhaps by increasing tryptophan uptake, enhancing its release through activity at presynaptic autoreceptors and/or by increasing activity at postsynaptic receptors. Some data, however, bring into question the long-term effect of lithium on 5-HT enhancement when studied in patients as opposed to healthy control subjects (36). Pandey et al. (37) found an increase in the density of platelet 5-HT_2A receptors in bipolar and schizoaffective subjects on lithium. Further study is required to reconcile this phenomenon with the therapeutic mechanism of lithium and determine whether this is a trait or state marker.

The impact on 5-HT receptors by various antipsychotics is also an important consideration in our understanding of their mood-regulating effects.

Cholinergic System

Lithium can increase red blood cell (RBC) levels of choline, but the significance of this finding is unclear (38). This activity is consistent with the cholinergic-adrenergic balance hypothesis of mood disorders, which promulgates a lack of cholinergic relative to noradrenergic activity in the manic state and the reverse in the depressed phase (39). For example, an early experiment that administered a single intravenous dose of physostigmine switched a manic state into a depression for about 1 hour (39). Such findings led to the study of drugs with cholinomimetic effects to treat acute mania. In support of this hypothesis, a pilot study by Burt et al. found the anticholinesterase inhibitor, donepezil, effective in 9 out of 11 treatment-resistant bipolar patients (40). A subsequent placebo-controlled trial, however, did not find this agent helpful as an adjunct in treatmentresistant bipolar disorder (41).

γ-Aminobutyric Acid System

γ-Aminobutyric acid (GABA) is the major CNS inhibitory neurotransmitter and can attenuate catecholaminergic systems. VPA, CBZ, gabapentin, topiramate, zonisamide, and lithium all enhance GABA activity through various mechanisms (42,43).

Glutamate System

Agents that act in part through the glutamatergic (GLU) system may also benefit bipolar disorder. For example, lithium dampens GLU-mediated calcium signaling and regulates calcium levels and transmission in hippocampal neurons (44). LTG may alter mood through blockade of Na⁺ voltage-gated channels that modulate the GLU system (42,45).

Intraneuronal Effects

The interactions of lithium with second messenger systems, particularly the phosphoinositide cycle, may deplete free inositol and alter intracellular calcium mobilization (i.e., inositol depletion hypothesis) (46). Antagonists such as verapamil decrease calcium channel activity, diminishing intracellular Ca²⁺ concentrations. Guanine nucleotide-binding proteins (G-proteins), adenyl cyclase, PKC, and glycogen synthase kinase-3 (GSK-3) are components of signal transduction pathways that are affected by mood stabilizers such as lithium and VPA (47). In particular, the ability of these agents, as well as tamoxifen, to decrease PKC-related activity, represents a common pathway that may explain the mood regulatory effects of these otherwise seemingly disparate agents (23). The effect of lithium on CNS gene expression regulation may also play an important role in long-term mood stability (48,49 and 50).

MEMBRANE AND CATION HYPOTHESIS

Electrolyte disturbances such as disruptions in calcium and sodium activity are the most consistently reported aberrations in mood disorders. In a series of studies, Dubovsky et al. (51) measured intracellular calcium ion concentrations in bipolar manic and depressed patients. They reported decreases in mean concentrations in four bipolar manic and five bipolar depressed patients, in comparison with seven normothymic subjects without personal or first-degree relative histories of psychiatric disorders. Their findings were consistent with a diffuse abnormality in the mechanisms modulating intracellular calcium homeostasis. Further, the presence of this phenomenon in both platelets and lymphocytes lends credence to a disruption in the cell membrane, the G-protein, or other mechanisms involved in the homeostasis of intracellular calcium ion concentrations. This may also support an extension of their findings from peripheral to neuronal tissue.

Calcium antagonists (e.g., verapamil, nimodipine) can also block DA, 5-HT, and endorphin activity; alter sodium activity via a sodiumcalcium counterexchange; and act as anticonvulsants. Any or all of these actions could be involved in their putative antimanic effects (19).

Cellular ionic transport mechanisms are another line of investigation into amine neurotransmitter and neuroendocrine function in affective disturbances. Lithium and the calcium antagonists block the influx of Ca²⁺, as well as alter intracellular calcium mobilization, thus dampening neuroelectrical activity and enhancing stabilization of
neuronal membranes. Meltzer (52) postulated a specific macromolecular complex composed of at least the sodium, potassium, and calcium pumps; the related ion channels; and ankyrin, which may be abnormally constituted in bipolar illness. Further elucidation of this hypothesis might help identify a specific membrane fault in bipolar disorder as well as more specific pharmacotherapies. For example, omega-3 fatty acids could possibly regulate mood stabilization in bipolar patients by inhibiting signal transduction mechanisms in neuronal membranes (21).

NEUROANATOMICAL AND NEUROPHYSIOLOGICAL HYPOTHESES

Before the development of functional brain imaging techniques, neuroanatomical correlates were primarily related to structural lesions of the CNS.

Jeste et al. (53) reviewed the neuroanatomical studies of major mood disorders, noting that several found no difference in abnormalities between schizophrenic and mood disorders. Swayze et al. (54), using magnetic resonance imaging (MRI), found a nonsignificant trend for ventricular enlargement in bipolar men, whereas bipolar women did not differ significantly from normal control subjects. Using MRI, Dupont et al. (55) reported subcortical abnormalities (i.e., white matter hyperintensities) in some bipolar patients. More recently, Moore et al. (56) reported significant increases in prefrontal and subgenual prefrontal gray matter volume in 28 bipolar depressed patients after 4 weeks of lithium treatment.

Initially, one of the more consistent findings was an apparent association between secondary mania and right frontal-temporal or left parietal-occipital lesions. Such data were also consistent with neuropsychological studies pointing to a right frontal lobe disturbance in these syndromes. Bearden et al. (57) speculate that frontal and subcortical hypometabolism may underlie bipolar disorder. Lyoo et al. (58) reported gray matter density decrease in the left prefrontal cortex and right inferior frontal and precentral gyri using voxel-based morphometry. Dager et al. (59) quantified regional brain metabolism in 32 drug-free bipolar subjects using two-dimensional proton echo planar spectroscopy. They found evidence of altered cellular energy metabolism, suggesting a redox shift toward glycolysis. This could represent compromised mitochondrial metabolism that may benefit from the potential neuroprotective actions of lithium. These findings are further supported by a study that found a significant decrease in nuclear messenger RNA coding for mitochondrial proteins in the hippocampus of bipolar subjects versus subjects with schizophrenia (60). These results suggest multiple pathophysiological origins for structural brain abnormalities underlying bipolar disorder. If confirmed and not simply due to epiphenomena such as drug treatment, substance abuse, or associated medical conditions (e.g., hypertension), these findings can serve as an important way to differentiate the neuroanatomical and neurophysiological mechanisms subserving bipolar disorder.

Central electrophysiological measures are also partially supportive, in that evoked response and computer-assisted electroencephalogram (EEG) mapping indicate bipolar-unipolar differences, with right-sided abnormalities more common in bipolar patients. Further, P300 topographical differences and sensorimotor gating deficits are reported in schizophrenia and other psychotic episodes, including acute psychotic mania (61,62).

The efficacy of ECT for mania is intriguing, vis-a-vis the anticonvulsant mood-stabilizing strategy, because ECT has many of the same antiseizure effects as VPA, LTG, and CBZ. Thus, during a course of ECT, the seizure threshold is usually raised, the duration of a given seizure episode decreases, neurometabolic response to a given seizure episode is diminished, and the phenomenon of amygdaloid kindling is attenuated. Kindling occurs in animals that are exposed to repeated subthreshold electrical stimuli and eventually develops spontaneous seizures. This is particularly interesting in that kindling in the mesolimbic structures has been analogized to the course of some bipolar disorders. For example, Post et al. (63) noted similarities between the increasing intensity of response to subthreshold stimulations and eventual spontaneous seizure activity and the natural course of certain bipolar patients whose illness progressively worsens, culminating in increasing vulnerability to more frequent, non-stress-induced episodes (e.g., behavioral sensitization). In this context, Post and Weiss (64) argue for differing pharmacosensitivity as a function of the stage of illness.
By implication, different treatments (e.g., lithium, antipsychotics, anticonvulsants) may interrupt the natural course of the illness at different phases, favorably altering its progression. Ghaemi et al. (65) postulated that interactions among second messengers, gene regulation, and synthesis of long-acting trophic factors in the context of kindling may explain how environmental stress coupled with genetic vulnerability can lead to bipolar disorder over time.

Other neuromodulatory approaches, such as VNS and TMS, may also be therapeutic for various phases of bipolar disorder (see later section and Chapter 8).

BIOLOGICAL RHYTHM HYPOTHESIS

Chronobiological factors are important to consider, given the cyclic pattern of disturbances in bipolar disorders. Goodwin and Jamison (66) noted that the suprachiasmatic nuclei of the hypothalamus serve as the endogenous pacemaker, temporally ordering various CNS functions. Further, they postulate that the organizing function of this “biological clock” may be dysregulated in bipolar disorder. Studies are hampered, however, by a dearth of longitudinal data and the masking of internal and external oscillator-driven rhythms that can alter rhythmic phase or amplitude.

In the context of this hypothesis, one of the most consistent findings is that sleep disturbances often precede or trigger a manic phase (67). Studies on circadian rhythms demonstrate that many aspects of the sleep cycle are phase advanced in mania (i.e., occur earlier than normal). Often, these patterns resemble the freerunning rhythms seen in normal individuals removed from all time cues. In addition, there is a blunting of amplitude and a doubling of the sleep-wake cycle up to 48 hours. Lithium delays the sleep-wake cycle and often slows such freerunning rhythms, which in turn are partly modulated by neurotransmitters such as NE, 5-HT, and ACh. Further, manipulation of the sleep-wake cycle may prevent a manic episode or treat the depressive phase (e.g., sleep deprivation therapy; see also Chapter 7).

Seasonal variation is another chronobiological rhythm manifested by increases in depression and suicide in spring (with smaller peaks in autumn), as opposed to mania, which increases in the summer months. These observations led to preliminary studies on the alteration of both light and temperature as potential therapies for the two seasonal patterns of affective disturbance. The phenomenon of seasonal variation is discussed in “Seasonal Affective Disorder” and “Bright Light Phototherapy” in Chapters 6 and 7.

NEUROENDOCRINE HYPOTHESIS

Dinan et al. (68) found a significant blunting of desipramine-induced growth hormone stimulation in seven drug-free bipolar patients when compared with seven control subjects and suggested that this phenomenon is consistent with a downregulation of α₂-noradrenergic receptors. Linkowski et al. (69) found elevations in nocturnal cortisol levels and an early nadir of this circadian variation. Another study using the dexamethasone suppression test (DST) reported that changes in cortisol activity may be state dependent (70). Watson and Young (71) reviewed evidence for hypothalamic-pituitaryadrenocortical axis dysfunction and concluded that

It is present in many bipolar patients, even when stable
Early life stressors and genetic predisposition are important factors
Corticotropin-releasing hormone hyperdrive, arginine/vasopressin dysregulation, and glucocorticoid receptor dysfunction are all implicated
Agents targeting this axis hold therapeutic promise

Similar phenomena are reported for major depression and might provide an important marker to identify the susceptibility of patients to clinically relevant mood fluctuations.

Lithium has several effects on the endocrine system. For example, it interferes with the synthesis and the release of testosterone, leading to an increase in luteinizing hormone levels. The thyroid system is implicated in neuroendocrine theories of the antimanic effects of lithium. In particular, thyroid hormones potentiate β-NE activity, and the ability of lithium to block their release may subserve its mood-stabilizing properties (i.e., the thyroid-catecholamine receptor hypothesis) (72). Data also implicate latent hypothalamic-pituitary-thyroid (HPT) system hypofunction and the antithyroid effects of lithium in the development of rapid cycling (73).

IMMUNOLOGICAL HYPOTHESIS

This line of investigation is based on evidence for a close interaction between the immune system and the CNS. For example, immunological abnormalities are reported in relation to psychological stress in patients with disorders such as major depression. Kronfol and House (74) studied different immune variables in manic, schizophrenia, and normal control subjects. In general, they found no significant differences for most measures. Results of the mitogen stimulation assays, however, revealed significant reductions in lymphocyte responsivity to the mitogens phytohemagglutinin-P and concanavalin-A in bipolar patients compared with schizophrenia patients and normal control subjects. These authors speculate that this may represent an impairment in cell-mediated immune response because the mitogens in question stimulate mostly T cells. Potential confounds to the study include its small sample size and the effects of ongoing psychotropic drug treatment. Tsai et al. (75) reported that cell-mediated immunity activation in manic patients appeared to occur through a specific state-dependent immune response. A recent review concludes that bipolar disorder and inflammation are linked through shared genetic polymorphisms and gene expression in addition to changing cytokine levels during symptomatic and a symptomatic periods (75a).

GENETIC HYPOTHESIS

Consistent with the dominant mode of transmission, it appears that first-degree relatives of bipolar patients have a 15% to 35% morbid risk for developing an affective disturbance. Concordance rates for mood disorders in twin studies demonstrate a strikingly higher incidence in monozygotic versus dizygotic twins (Table 10-2) (76). Both bipolar and unipolar disorders occur more often in families with a history of this illness. Therefore, these two variants of affective disease appear to breed true. In addition, bipolar patients seem to have a greater genetic loading for mood disorder than their unipolar counterparts (77). Thus, the rate of this illness in relatives of bipolar probands is 4 to 10 times greater than in relatives of unipolar probands. Further, even though a large proportion of bipolar proband relatives develop only unipolar symptoms, unipolar proband relatives develop predominantly unipolar symptoms.

TABLE 10-2 CONCORDANCE (+) AND DISCORDANCE (-) OF MOOD DISORDER IN TWIN PAIRS

	Monozygotic		Dizygotic
8 studies	(+)	(-)	(+)	(-)
Total	103 (62%)	63 (38%)	55 (18%)	258 (82%)
χ² = 98.7; p < 0.0001 (Mantel-Haenszel)
From Davis JM, Noll KM, Sharma R. Differential diagnosis and treatment of mania. In: Swann AC, ed. Mania: New Research
Treatment. Washington, DC: American Psychiatric Press; 1986:1-58, with permission

Other disorders reported to be cotransmitted with bipolar and, to a lesser extent, unipolar disorders include

Schizoaffective disorder
Cyclothymic personality
Hypomania (without depression)

Another approach is the study of RBC/plasma lithium concentration, which expresses the relationship between intracellular and extracellular levels. The lithium-sodium countertransport (an exchange diffusion process) mechanism is located in the cell membrane and determines the relative concentration of these ions (78). Thus, an abnormality in transport function could represent a heritable marker of susceptibility for the development of bipolar disorder (79). Clinical studies of lithium administration find that the RBC/plasma lithium ratio varies from 0.15 to 0.60 (average = 0.30) and remains constant for individuals independent of change in symptoms (80). There is also evidence that bipolar patients have a higher mean ratio than normal controls or those with other psychiatric diagnoses, and that a greater proportion of their first-degree relatives demonstrated this elevation (81).

Linkage analysis studies are particularly useful in resolving issues regarding the myriad clinical presentations, while contributing to an increased understanding of the basis for vulnerability to various mood disorders. Research in this area, however, has not unequivocally established a
specific genetic heritability for bipolar illness. For example, linkage to the genetic markers of color blindness, associated with a region of the X chromosome, is reported in some, but not all, pedigrees (82). The association of bipolar disorder to chromosome 11 suggested to occur in the Amish pedigree did not replicate with a larger sample (83). A corollary to the chromosome 11 story is the nearby location of genes involved in tyrosine hydroxylase production as well as a muscarinic cholinergic receptor gene. Subsequent genetic linkage data identify bipolar susceptibility loci in multiple regions (i.e., polygenic inheritance of the human genome), including 4p16, 12q24, 13q31, 18p11.2, 18q22, 22q11, and Xq26 (84,85,86,87 and 88). Of interest, linkage studies indicate an overlap at loci 8p22, 10p14, 18p11, and 22q11 for susceptibility to both bipolar disorder and schizophrenia, suggesting shared genetic vulnerability and perhaps less distinction between these two disorders than our present diagnostic system indicates (89,90). Further, Faraone et al. (91) have also linked three regions (i.e., 12p, 14q, and 15q) with the age of onset of bipolar disorder.

Another approach is the application of pharmacogenetic techniques to predict responsivity or proclivity for adverse effects to specific therapies. For example, a genome-wide association study (GWAS) identified multiple novel loci that may be associated with response to lithium, including the GR1A2 gene whose expression is regulated by this mood stabilizer (92). Another example involves the identification of relevant genetic polymorphisms. In this context, Mundo et al. (93) conducted a preliminary trial in 27 bipolar I or II subjects with at least one manic or hypomanic, serotonergic antidepressantinduced episode. They were compared with 29 matched bipolar I or II subjects who did not experience mania/hypomania when exposed to a serotonergic antidepressant. The group who experienced a mood switch had an excess of the short allele of the serotonin transporter (5-HTT) gene compared with those who did not. In a third example, Kasuya et al. (94) used the fruit fly (drosophila melanogaster) to identify a lithium-inducible SLC6 amino acid transporter (LIST) that appears to mediate toxicity associated with lithium. If confirmed, these approaches can help clinicians predict medication response and/or sensitivity to adverse effects in bipolar patients.

CONCLUSION

Whether any or all of these factors relate to the efficacy of mood stabilizers is still uncertain, but the final common pathway may well be modulation of neurotransmitter and neuroreceptormediated signal transduction processes that affect neuroplasticity (i.e., how the brain perceives, adapts to, and responds to stimuli) (95). What is not clear is how these drugs alter neuronal function or interactions, ultimately leading to their normothymic effects. A further complication stems from the combined antimanic and antidepressant properties of at least some of these therapies (e.g., ECT, lithium, QTP), suggesting that opposite clinical effects may be mediated by biochemical changes in different directions based on varying states within the CNS. Table 10-3 summarizes various mechanisms of action for existing and putative mood stabilizers. It depicts the multiplicity of effects and overlap among agents, indicating the complexity of their actions as well as providing future targets for drug therapy.

Pharmacokinetics

Table 10-4 lists the pharmacokinetic properties of the more commonly used mood stabilizers, while Chapter 5 discusses the pharmacokinetic properties of SGAs. The relevance of their pharmacokinetic differences is discussed in the section on each agent.

Treatment of Bipolar Disorder

Bipolar disorder poses a diagnostic and therapeutic challenge for even the most skilled clinician (96). Although approximately 3 million individuals in the United States are presently treated for this disorder, it is estimated that many more go undiagnosed, misdiagnosed, or inadequately managed.

Unfortunately, this condition is also characterized by relapses and recurrence. For example, within 1 year of recovery, data indicate that one half experienced symptoms of hypomania, mania, or depression (97). Factors that may place patients at greater risk for relapse or recurrence include

Number of previous episodes
Clinical presentation (e.g., mixed states, bipolar II)
Comorbid personality disorder (e.g., borderline, narcissistic, histrionic)
Substance use disorder or alcohol abuse

TABLE 10-3 MECHANISMS OF ACTION OF EXISTING AND PUTATIVE MOOD STABILIZERS

Anticonvulsants

Antipsychotics

Cation

Lithium

VPA

CBZ

LTG

TOP

CLZ

RISP

OLZ

QTP

ZPD

ARIP

ASEN

Neurotransmitters

√(↓)

√(↑)

√

√^a

√

5-HT

√(↑)

√

GABA

√(↑)

√(?)

√

GLU

√(↓)

√

ACh

√(↑)

√

Signal Transduction

G-protein

√(↓)

Inositol

√(↓)

cGMP

√(↓)

PKC (α, ε)-marks

√(↓)

GSK-3β

√(↓)

AP-1 (fos, jun) DNA binding

√(↑)

Ion Channels

Ca²⁺

√

Na²⁺

√

Cl⁺

√

K⁺

√

Other

Thyroid

√(↓)

Circadian rhythms

√

Carbonic anhydrase inhibition

√

Neurotropic/neuroprotective (e.g., BDNF)

√(?)

VPA, valproate; CBZ, carbamazepine; LTG, lamotrigine; GBN, gabapentin; TOP, topiramate; CLZ, clozapine; RISP, risperidone; OLZ, olanzapine; QTP, quetiapine; ZPD, ziprasidone; ARIP, aripiprazole; NE, norepinephrine; DA, dopamine; GABA, γ-aminobutyric acid; GLU, glutamate; ACh, acetylcholine; PKC, protein kinase C; cGMP, cyclic guanosine monophosphate; PKC, protein kinase C; GKC, glycogen synthase kinase; AP, apoprotein; BDNF, brain derived neurotrophic factor; (↓), decrease; (↑), increase.

^aPartial DA₂ agonist.

TABLE 10-4 PHARMACOKINETIC PROPERTIES OF STANDARD MOOD STABILIZERS

	Mood Stabilizer
Pharmacokinetic Property	Lithium	Valproate	Carbamazepine	Lamotrigine
Time to reach peak concentration	1-12 hours	3-5 hours	1-5 hours	1-4 hours
Metabolism	Renal excretion	Hepatic oxidation and conjugation	Hepatic microsomal oxidation (P450 2D6)	Hepatic glucuronidation; renal excretion
Half-life	20-27 hours	6-16 hours	15-30 hours initial; 10-15 hours maintenance	24 hours
Time to reach steady state	4-7 days	1-3 days	2-4 days initial; 4-5 days after dosage change	3-15 days
Protein binding	0	90%-95%	70%-80%	55%
Therapeutic range	0.8-1.5 mEq/L	45-150 μg/mL (putative)	4-15 μg/mL (putative)	2-20 μg/mL (unknown for bipolar disorder)

Although our focus is on the evidence for various pharmacological and neuromodulation interventions, we emphasize that without a well-conceived psychosocial strategy the ultimate outcome is usually compromised. Thus, there are several major approaches to effectively manage bipolar disorder. They include

Standard mood stabilizers
- Lithium
- Anticonvulsants (e.g., VPA, CBZ, LTG)
SGAs
Combination drug strategies
Therapeutic neuromodulation (e.g., ECT, VNS, TMS)
Psychosocial strategies (e.g., individual, marital, family, group, and cognitive behavioral therapy [CBT]; various rehabilitation programs) (98,99)

Concerning the last point, Miklowitz (100) reports that adjunctive psychotherapies emphasizing medication adherence and early recognition of mood symptoms appear to impact mania more directly, while cognitive and interpersonal coping strategies benefit depression more. Further, Frank et al. (101) report that the interpersonal and social rhythm therapy appear to improve occupational functioning more effectively than psychosocial and supportive approaches.

In this context, the pharmacotherapy of bipolar disorder is continuing to evolve. Studies of the efficacy of various mood stabilizers for bipolar disorder focus on

Treatment of an acute exacerbation of manic, hypomanic, mixed, or depressive epis-odes
Prevention of relapse after an acute episode is controlled
Prevention of future episodes
Attenuation of suicidal behavior

We review the existing literature, emphasizing the class I controlled trials, comparing these agents with either placebo or standard comparator treatments for acute, maintenance, and prophylactic purposes. As noted earlier, we also comment on the findings from the STEP-BD that complement the data from controlled trials.

Management of an Acute Manic Episode

LITHIUM

For almost a half century, lithium has been the standard drug treatment strategy for bipolar disorder. This is in large part because of the quantity and the quality of evidence supporting its role as an effective maintenance and prophylactic treatment, particularly in preventing
mania (102,103). Prophylactic efficacy is a critical consideration given the recurrent nature of this disorder. Thus, clinicians must choose the optimal strategy for acute treatment with the realization that most patients will need to continue drug therapy indefinitely. In addition, there is support for the longer-term beneficial impact of lithium on the suicide rate in bipolar patients (104,105). The authors of these reports note that the lower suicide risk associated with lithium treatment may be due to

Its mood-stabilizing properties
A lower suicide risk per se in patients who remain in treatment
A specific antisuicidal effect

By contrast, in a case-control analysis of 106 bipolar patients, the STEP-BD did not observe a relationship between lithium use and suicide attempts or completions (106). Possible explanations for these discrepant findings include the small number of patients who engaged in suicidal behavior, the paucity of nonwhite subjects, and the use of a longitudinal prospective design in the STEP-BD.

During the depressive phase, patients often receive an antidepressant. This may further complicate their management, however, because antidepressant monotherapy can propel patients into a manic episode, cycle acceleration, a rapid-cycling course, or a more treatmentresistant phase of their illness (107). In this context, lithium can benefit the depressive phase as a monotherapy, as augmentation to antidepressants, or in combination with anticonvulsants or SGAs for more treatmentresistant courses. Further, the strategy of employing lithium alone or in combination with an antidepressant may minimize the risk of mood destabilization.

Lithium Versus Placebo or Other Psychotropics

Lithium Versus Placebo. Schou et al. (2) charted the natural history of several bipolar patients and found that the introduction of lithium-induced remissions, dramatically altering the course of the disorder. Although some patients also received placebo, the data are not presented systematically and therefore are not included in the meta-analysis. Bunney et al. (108) reported on a patient treated with lithium or placebo in a longitudinal ABA design who failed to respond to placebo, improved when switched to lithium, and then relapsed when lithium was discontinued. A subsequent report by Goodwin et al. (109) described two additional cases with an unequivocal response to lithium, four others with a probable response, one with an equivocal response, and three who deteriorated on lithium; however, there was no control group. Although not as definitive as the class I or II study designs, these naturalistic reports strongly support the efficacy of lithium (see “Evaluation of Drug Study Designs” in Chapter 3).

TABLE 10-5 LITHIUM VERSUS PLACEBO FOR ACUTE MANIA

	Responders (%)
Number of Subjects	Lithium (%)	Placebo (%)	Difference (%)	χ²	pValue
107	54	31	23	6.0	0.01

Subsequent trials included 107 patients evaluated in prospective, random-assignment, double-blind designs and found lithium superior to placebo for the treatment of a manic episode (110,111 and 112) (Table 10-5). Maggs (110) conducted a double-blind, random-assignment, placebocontrolled, parallel trial of lithium without concomitant medications. Although he found lithium clearly superior to placebo, he probably underestimated the true drug-placebo difference because he did not include the placebo nonresponding dropouts, nor did he include data that could be used in the meta-analysis. Stokes et al. (111) also conducted a 10-day, double-blind, placebo-controlled, random-assignment trial of lithium using only a modest amount of adjunctive antipsychotics. He crossed over nonresponders every subsequent 10-day period to the opposite arm, for a total of four switches, and found lithium superior to placebo throughout the study. Again, he may have underestimated the true drug-placebo difference, however, because a period of 10 days is often insufficient to achieve full benefit with lithium and some
patients did receive antipsychotics. In 1976, Stokes et al. (112) used the same crossover design, this time assigning patients to high or low doses of lithium. Combining the data from his two studies, he was able to demonstrate a dose (and plasma level)-response relationship to remission.

Bowden et al. (113) conducted the first class I, placebo-controlled, multicenter study of the divalproex sodium (DVPX) formulation of VPA. A lithium arm was also included, significantly increasing the number of lithium-treated patients evaluated under placebo-controlled conditions (i.e., 110 patients: 74 on placebo, 36 on lithium). Marked improvement, defined as at least a 50% reduction in the manic syndrome subscale score derived from the Schedule for Affective Disorders and Schizophrenia-Change (SADS-C), occurred in 49% of the lithium group versus 25% of the placebo group (p < 0.025).

Lithium Versus First-Generation Antipsychotics. First-generation antipsychotics (FGAs) were the treatment of choice for acute mania before the reintroduction of lithium. Clinical experience finds that in some cases, these agents are the only effective therapy, as well as the only practical (i.e., sufficiently longacting) treatment, in nonadherent or pharmacokinetically idiosyncratic patients. Early in its history, lithium was systematically compared with FGAs for the management of acute mania. Most believe that it produces a better qualitative response, but this observation is difficult to substantiate with controlled studies. A patient once analogized this difference to an automobile engine racing out of control, describing the effect of FGAs as similar to applying the brakes, while lithium is similar to adjusting the carburetor (114).

Cookson et al. (115), in a random-assignment design, found that 10 of 12 manic patients responded to pimozide and 11 of 12 patients to chlorpromazine (CPZ). This is consistent with the observation of Post et al. (116) who, after a placebo lead-in period, found the time course of improvement with pimozide (n = 8) to be similar to that with lithium (n = 8) or with a phenothiazine (n = 9). Because pimozide is a more specific D₂ antagonist than CPZ, this outcome provided further support for a beneficial role with D₂ receptor blockade in the treatment of mania.

Rifkin et al. (117) compared three doses of haloperidol (HPDL) for up to 6 weeks as the only treatment in 47 acutely manic patients. They found that 72% of subjects responded, excluding dropouts (most occurring in the first 2 weeks and evenly distributed across all groups). Furthermore, 10, 30, or 80 mg/day of HPDL were equieffective (i.e., survival analysis showed no difference among the three dosing regimens). This is particularly noteworthy and parallels the finding that lower doses of HPDL to treat schizophrenia and schizoaffective patients are often as effective as moderate to high doses (118).

Lithium primarily benefits bipolar disorder and is relatively ineffective for schizophrenia. Given its narrow spectrum of efficacy, the inclusion of patients with schizoaffective disorder and schizophrenia in studies comparing lithium with antipsychotics may have biased the results against it. Also, because antipsychotics usually have a faster onset of action, a design that allows patients to drop out early may underestimate the relative efficacy of lithium. For example, Prien et al. (119) studied 255 Veterans Administration (VA) patients diagnosed as bipolar, manic, or schizoaffective and assigned them to lithium or CPZ. These investigators subdivided patients into highly versus mildly agitated categories based on the Inpatient Multidimensional Psychiatric Scale, which describes highly agitated patients as “exhibiting overactivity, restlessness, and/or accelerated body movement.” Also, patients were categorized by the Brief Psychiatric Rating Scale (BPRS) for baseline levels of excitement, uncooperativeness, grandiosity, mannerisms, tension, and conceptual disorganization. Twenty-two percent of the lithium-treated and 14% of the CPZ-treated patients dropped out, leaving approximately 60 in each of the four groups: highly active, lithium- or CPZ-treated patients and mildly active, lithium- or CPZtreated patients. There were substantially more dropouts in the highly active group receiving lithium, usually due to poor response or uncooperativeness, but the differences between baseline and posttreatment for all four groups were similar. An analysis of covariance found that the outcome for lithium completers did not differ from that for CPZ completers. An end pointoint analysis, however, found CPZ superior to lithium in the highly active group on such measures as conceptual disorganization, psychoticism,
grandiosity, and suspiciousness. There were no significant differences between the two drugs in the mildly active group. When schizoaffective patients and many of the severely disturbed patients who did not receive lithium for a sufficient duration were included, the antipsychotics were found to be superior.

TABLE 10-6 LITHIUM VERSUS CHLORPROMAZINE FOR ACUTE MANIA

	Lithium		First-Generation Antipsychotics
	Responders	Nonresponders	Responders	Nonresponders
Total	64 (89%)	10 (11%)	38 (54%)	33 (46%)
χ² (Mantel-Haenszel) = 13.1; df = 1, p = 0.0003.
Adapted from Janicak PG, Newman RH, Davis JM. Advances in the treatment of mania and related disorders: a reappraisal
Psychiatr Ann. 1992;22(2):94.

Garfinkel et al. (120) compared HPDL, lithium, and their combination for the treatment of mania. Initially, there were seven patients in each group. By day 15, three in the lithium group, two in the HPDL group, and one in the combined drug group dropped out. Although the two HPDL groups improved slightly more than the lithium monotherapy group, considering the small sample and the high dropout rate, we speculate that these patients were highly disturbed and did not receive lithium alone for a sufficient period.

Braden et al. (121) did a study similar to that of Prien et al. (119) in primarily schizophrenic or schizoaffective patients, but the sample also included some affectively ill cases (i.e., 21% met Feighner criteria for mood disorder). Of the 43 patients on lithium, 15 dropped out because they did not improve, worsened, or were unmanageable or confused. By comparison, only one of 35 treated with CPZ dropped out in the first 10 days. CPZ produced better results in the overactive group, whereas the drugs were comparable in the less active group. As with the VA study, the poor results with lithium in the overactive group may be an artifact of insufficient duration due to the high dropout rate or because of the inclusion of patients with a core schizophrenic illness.

Johnstone et al. (122) performed a randomassignment, double-blind trial comparing pimozide, lithium, a combination of these two, and placebo in 120 patients with a variety of psychotic disorders (e.g., psychotic mania, psychotic depression, schizoaffective disorder, schizophreniform disorder, schizophrenia). Patients were characterized as having an elevated or depressed mood or no consistent change in mood. Pimozide had a robust effect on positive symptoms across all groups, whereas lithium had only a modest beneficial effect (p = 0.07) in affectively disordered patients with an elevated mood. Again, heterogeneity in the diagnostic categories tempers interpretation of these results.

Table 10-6 summarizes the results from four of five well-designed, albeit small, trials comparing lithium with CPZ in classic manic patients (123,124,125,126,127 and 128). Each study found lithium superior to CPZ, and a meta-analysis of the combined studies demonstrated this difference was statistically significant.

Lithium Versus Anticonvulsants. Some early trials compared lithium to various anticonvulsant mood stabilizers (e.g., VPA, CBZ, and LTG) (113,129,130). The results are discussed in more detail in the sections on these specific agents. In general, lithium was comparable, but some data support better efficacy with agents such as VPA and CBZ in certain subgroups (e.g., mixed states, rapid cyclers) and better prophylaxis with LTG to prevent depressive relapses. It is not clear, however, whether these agents have the same impact on suicide as lithium.

Lithium Versus Second-Generation Antipsychotics. Some SGA trials for bipolar mania included lithium as a standard comparator and generally found that these agents had similar antimanic effects. These results indicate a moodstabilizing effect for this generation of antipsychotics and will be discussed in more detail in the section on SGAs.

Lithium Plus Other Psychotropics

From the perspective of clinical trial methodology, concurrent medications can complicate the
interpretation of results. Intermediate rescue medications are often required because standard mood stabilizers are relatively slow in their onset of action. Further, if rescue medications are avoided, this usually introduces the confound of dropouts before the experimental drug can be fully effective. When feasible, a reasonable compromise is the use of modest amounts of a benzodiazepine (BZD) when necessary for a limited time (e.g., 7 to 10 days) into the active phase of treatment. This can reduce the number of nonresponding, highly agitated patients who may otherwise drop out of treatment, and in a trial of several weeks, the initial BZD effect should dissipate by the final assessments.

Several studies combined lithium with other treatments such as FGAs, SGAs, BZDs, anticonvulsants, and thyroid supplementation (131). Generally, in partial responders, the addition of these medications was beneficial and well tolerated.

Lithium Plus Antipsychotics. Many acute manic patients present in a very explosive, belligerent, and agitated manner and waiting several days to weeks to gain control of an episode is not feasible. Thus, antipsychotics are often required in the earliest phases of treatment, particularly with moderate-to-severe exacerbations, often associated with psychotic features. In addition, many patients require maintenance antipsychotics to prevent frequent relapses. These agents are usually initiated in conjunction with lithium because of their more rapid impact, then are carefully tapered and discontinued when possible after the full effect of lithium is realized.

Unfortunately, manic patients may be exposed to higher than necessary acute antipsychotic doses, perhaps because of the explosivity often associated with an exacerbation. For example, Baldessarini et al. (132) conducted an epidemiological survey in the Boston area and found that dosing schedules of higher-potency FGAs (e.g., HPDL, fluphenazine) were three to five times greater than the CPZ dose equivalents of lower potency FGAs (e.g., CPZ, thioridazine). This was consistent across a number of different diagnostic categories, including affectively disordered patients. They postulated that the different (and at least perceived as more benign) adverse effect profile of the higherpotency drugs encouraged aggressive increases in dose to control severe, acute psychotic exacerbations.

Addressing this issue in hospitalized acutely manic patients, Janicak et al. (133) conducted a 2-week study of lithium plus random assignment to equivalent doses of CPZ or thiothixene. The dose-equivalent ratio used was 5 mg of thiothixene to 100 mg of CPZ (i.e., 1:20). During the waking hours, patients’ antipsychotic doses were titrated on a 2-hour basis so that response and adverse effects to the prior dose were the determinants for administering or holding the next dose. The aim was to compare the efficacy, adverse effect profiles, and optimal dose required for either antipsychotic. By the end of the 2-week trial, the mean dose in the thiothixene-treated group was 36 mg, and the mean dose of CPZ was 480 mg (i.e., a ratio of 1:13). These amounts fell in the low to mid range of the dose-response curve previously reported. Second, the overall improvement was statistically significant for the entire sample when baseline values were compared with day 14, and response did not differ between the two antipsychotic groups. Finally, because relatively lower doses were used, adverse effect profiles, although typical and in the expected direction (i.e., slightly more extrapyramidal symptoms [EPS] for those patients receiving thiothixene and slightly more sedation and hypotension in those receiving CPZ), allowed for optimal and tolerable dosing regimens. The blood levels of lithium by the end of week 1 (i.e., approximately 1.0 mEq/L) were adequate and almost identical for both groups. No other concomitant medications were used. Thus, relatively low-to-moderate doses of either a high- or low-potency FGA were sufficient to control acute mania with associated psychotic features, and at these doses, adverse effects remained tolerable.

Unfortunately, two retrospective chart reviews at the Connecticut Mental Health Center found that a substantial number of manic patients started on an antipsychotic while hospitalized were still on these agents 6 months after discharge and that chronic FGA treatment was common in their outpatient bipolar group (134,135). These authors concluded that antipsychotic adjuncts to lithium should be reconsidered often and reduced or discontinued whenever possible.

SGAs are increasingly the preferred antipsychotics. Their expanding role as both monotherapy and as augmentation to standard mood
stabilizers for bipolar disorder is discussed in a subsequent section.

Lithium Plus Benzodiazepines. The use of adjunctive antianxiety agents for acutely manic patients may avoid the need for concomitant antipsychotics or at least minimize their total amount (136). Although the literature is generally anecdotal and parallels a similar literature for treating acute psychosis, there are some controlled trials (see “Management of Acute Psychosis” in Chapter 5). The most commonly studied BZDs are lorazepam and clonazepam, due to their rapidity of onset and duration of action. In addition, when lorazepam is administered intramuscularly, there is adequate absorption in contrast to other BZDs. Alprazolam and diazepam have also been studied.

The major benefit of BZDs is to diminish some of the secondary symptoms of an acute exacerbation (e.g., insomnia, agitation, panic, and other general anxiety symptoms) that are not necessarily rapidly and specifically affected by lithium or antipsychotics. With this approach, exposure to antipsychotics is avoided in some situations or kept to a minimum in others, thus avoiding the potential for more serious antipsychotic-related adverse effects. Additionally, given the high comorbidity of acute bipolar mania with alcohol abuse or dependence, concurrent withdrawal symptoms can also be managed with short-term BZD administration. Relative contraindications to the use of BZDs include

Previous paradoxical response to BZDs (i.e., behavioral disinhibition)
Known sensitivity to these agents
Acute narrow angle glaucoma
Pregnancy

Lorazepam. Lenox et al. (137) found lorazepam and HPDL comparable in efficacy when used as adjuncts to lithium in a double-blind study of 20 acutely manic patients. Another report comparing lorazepam with clonazepam found a better outcome with lorazepam, using mean doses of 12 to 13 mg (138).

Clonazepam. Clonazepam is approved for petit mal variant, myoclonic, and akinetic seizures. It also enjoys wide off-label psychiatric applications, including the treatment of acute mania or other agitated conditions as an adjunctive treatment with standard mood stabilizers or antipsychotics.

Typical doses of clonazepam are in the range of 2 to 16 mg/day given on a once- or twice-per-day schedule due to its longer halflife. A major advantage of this anticonvulsant is its relative lack of adverse effects and freedom from laboratory monitoring in comparison with CBZ and VPA. In this context, it is viewed as a behavioral suppressor rather than a true mood stabilizer (139).

Lithium Plus Anticonvulsants. There is limited controlled data to support the combination of two or more mood stabilizers (e.g., lithium plus VPA) in partially but insufficiently responsive patients as well as for more effective prophylactic therapy (140,141 and 142). One question is whether the second or third agent produced the benefit or was it an additive effect from the combination. After a reasonable period of stabilization, we recommend a very slow tapering of the original agent(s) to determine whether the latest mood stabilizer alone is sufficient to maintain the beneficial effects. Further, since there is a risk for additive adverse effects, the starting doses of the add-on agents should be low and slowly increased as necessary.

Lithium Plus Thyroid Supplementation. Some treatment-resistant and rapid-cycling bipolar patients may have thyroid dysfunction (e.g., subclinical hypothyroidism) that improves with the addition of thyroid supplementation. In this context, several case reports involving this population found that high doses of the thyroid hormone, levothyroxine sodium, were clinically beneficial (143,144). In support, Kusalic (145) found that 6 of 10 rapid cyclers had hypothyroidism, based on their thyrotropin-releasing hormone stimulation tests. Further, the average number of mood episodes per year decreased by more than 75% (i.e., from 9.7 to 2.2) after thyroxine was added to the treatment regimen.

Bauer and Whybrow (146) entered 11 rapidcycling, treatment-refractory patients into an open trial of high-dose levothyroxine sodium, which was added to their previously stable medication regimen. Levothyroxine was increased by 0.05 to 0.1 mg/day every 1 to 2 weeks, as tolerated, until symptoms improved or adverse effects prevented further increases. Scores on both the depressive and the manic symptom rating scales decreased significantly compared with baseline. These data indicate that levothyroxine in doses
sufficient to produce supranormal circulating hormone levels may induce remission of both depressive and manic symptoms in an otherwise refractory group. These results were supported by a subsequent report, suggesting that rapidcycling treatment nonresponders may benefit from the addition of levothyroxine (147). Since these are open case reports with small sample sizes, preliminary positive results need confirmation in controlled trials.

In summary, lithium monotherapy of sufficient level and duration is the ideal approach in classic, milder presentations of mania. Adjunctive antipsychotics (especially SGAs) are often required in more severe episodes (e.g., mania with psychosis or schizoaffective disorder) due to their faster onset of effect and broader spectrum of activity (see “Alternative Treatment Strategies” later in this chapter). Combination strategies with other psychotropics are commonly used but less well studied.

Administration of Lithium

Lithium is a monovalent cation which is excreted through the kidneys. It is an alkali metal in group IA and shares many properties with similar elements, such as sodium and potassium. It is rapidly absorbed and reaches peak blood levels in approximately 1 to 3 hours (6 to 8 hours with sustained-release preparations), with absorption completed in approximately 8 hours. Unlike other psychotropics, it is not protein bound and steady state levels are usually achieved after 4 to 6 days on a fixed dose (148). The efficacy of lithium is established within a well-defined therapeutic plasma level range for optimal benefit and minimal adverse or toxic reactions. Lithium preparations include lithium carbonate; sustained-release preparations; and the liquid form, lithium citrate. The sustained-release preparations allow for a more gradual absorption of the drug, leading to blunted peak plasma levels. Because lithium has a slow onset of action, it takes days to weeks, and occasionally longer, to obtain an optimal clinical response. Thus, it is important to avoid premature abandonment in those who are slower to respond.

TABLE 10-7 LITHIUM LABORATORY EVALUATIONS

Before Initiation of Treatment
Renal function testing
	General screening—serum creatinine; BUN; urinalysis
	If further testing required:
		24-hour urine volume
		Creatinine clearance
		Test of renal concentrating ability
Thyroid function
	T₃, T₃RU, and T₄
	TSH
Cardiac
	ECG (if elderly or at risk for heart disease)
General (if indicated)
	CBC and differential
	Serum electrolytes
	General chemistries
During Treatment
Renal function
	Urinalysis; BUN (every 6-12 months)
	Serum creatinine (every 6-12 months when clinically indicated)
	Test of renal concentrating ability (when clinically indicated)
Thyroid function
	TSH (every 6-12 months)
	Repeat full battery with elevated TSH and/or clinical signs/symptoms
BUN, blood urea nitrogen; TSH, thyroid-stimulating hormone; ECG, electrocardiogram; CBC, complete blood count.

As part of the standard prelithium workup, a thorough medical evaluation is required. Table 10-7 lists the various laboratory tests recommended to assess overall physical status, especially renal, thyroid, hematological, and cardiac function, before initiation of treatment (see
also Chapter 1). In particular, the renal and the thyroid systems require a baseline assessment and periodic reevaluation with ongoing lithium therapy.

Because lithium has a half-life of approximately 24 hours and it takes four to five halflives to achieve steady state at a fixed dose, plasma levels should be obtained every 5 days until an adequate therapeutic concentration is achieved or adverse effects preclude further increases. Attempts to develop dose prediction formulae to obtain therapeutic concentrations more rapidly do not enjoy widespread use. Although premature monitoring may lead to higher than necessary dosing, more frequent measuring of levels may be warranted in patients with known sensitivity to lithium or if unexpected adverse reactions occur. Once the initial treatment has begun, we recommend blood levels in the range of 0.8 to 1.2 mEq/L for optimal efficacy. Some patients who do not benefit from these levels may respond at slightly higher concentrations (e.g., 1.3 to 1.5 mEq/L). Conversely, a small proportion of patients who are unable to tolerate levels in the lower end of the usual therapeutic range may acclimate to and benefit from concentrations as low as 0.3 to 0.7 mEq/L. Blood samples for lithium levels should be drawn 10 to 12 hours after the last dose to measure the concentration at its trough. Lithium saliva and RBC concentrations have been promoted as more accessible or more accurate measures, respectively, but do not have general acceptance.

Typical starting doses of lithium are 300 to 900 mg/day. The final dose can vary from a low of 300 mg/day to as high as 3,000 mg/day, with the typical range between 900 and 1,800 mg/day. An individual’s age, renal function, and general physical condition are important associated determining factors for the ideal dose. Initially, dosing schedules were on a three- or four-times-a-day basis, but subsequent data indicated that once- or twice-a-day regimens may enhance compliance and minimize certain adverse effects, without compromising efficacy.

Whenever possible, the ideal is to treat with a single mood stabilizer, such as lithium, because of its specificity for bipolar disorder and to minimize adverse effects. This is particularly true in mild to moderately severe episodes. In addition, if the patient can benefit from monotherapy during the acute episode, this would support its benefit for maintenance and prophylactic purposes. Further, monotherapy improves adherence, diminishes the chance for potentially significant drug adverse effects or interactions, and reduces cost. Unfortunately, most bipolar patients require combination drug strategies to achieve optimal benefit.

The narrow therapeutic index and numerous adverse effects of lithium, however, can limit its usefulness, especially in the context of longterm therapy. Indeed, 75% of patients on this agent experience adverse effects, usually involving the renal, gastrointestinal, thyroid, and/or neurological systems (149). Due to the significant proportion of patients who are nonresponders, insufficiently responsive, or intolerant to its adverse effects, and given the alarmingly high suicide rate for untreated or inadequately treated bipolar patients, the need for other effective therapies is clear.

VALPROATE

Since the early 1990s, a major pharmacological strategy for bipolar disorder involves various anticonvulsants. This era was ushered in with the FDA’s approval of a specific formulation of VPA (i.e., DVPX) for the treatment of acute mania. An extended-release formulation of CBZ is also approved in the United States for this indication.

Reports on the benefit of valproate (VPA) for mood disorders date back to the mid-1960s. Lambert et al. (16) first investigated VPA in a series of clinical trials including a wide variety of
patients. Twelve subsequent open-design studies, representing 297 acutely-ill patients, found an overall moderate to marked response rate of 56%. A number of methodological problems complicate the interpretation of these results, however, including most patients were studied under nonblinded conditions. VPA was often used in combination with other psychotropics, plasma concentrations were usually not monitored, and formal diagnostic criteria derived from standard clinical ratings were typically not used. In the early European experience, the focus was on maintenance therapy of manic-depressive disease, with patients stabilized on VPA or valpromide for up to 10 years using the drug as monotherapy or in conjunction with other psychotropics. Some investigations also employed VPA in acute mania, usually in combination with antipsychotics, and reported benefit, often allowing for substantial reductions in the antipsychotic dose (150).

Preliminary trials indicate that VPA is also useful in the management of other psychiatric conditions, such as

Panic disorder (151)
Posttraumatic stress disorder (152)
Personality disorder with aggression (153)
Dementia with agitation (154)

Studies using adjunctive VPA to treat schizophrenia, however, have generally been negative (155,156 and 157).

The divalproex sodium (DVPX) formulation was the first drug since lithium to be approved by the FDA for the treatment of bipolar disorder. This agent may also benefit some of the more lithium-resistant subtypes described earlier (see also Chapter 9). It does not, however, appear to be as beneficial for the acute depressive phase of this illness. DVPX is contraindicated in patients with significant hepatic or pancreatic disease, during pregnancy, and in those with a known hypersensitivity to this agent. It should also be used with caution in younger women due to the possible increased risk of polycystic ovarian syndrome (PCOS) (158) (see section on “Adverse Effects”).

Valproate Versus Placebo or Other Psychotropics

Valproate Versus Placebo. Although early studies suggested efficacy for acute mania, controlled clinical trials subsequently confirmed these findings.

Two class I studies examined the efficacy of the DVPX formulation in comparison with placebo for acute mania. First, Pope et al. (159) treated 36 previously lithium-resistant (i.e., nonresponsive and/or nontolerant) patients for 3 weeks with DVPX or placebo in a randomassignment, double-blind design. DVPX was significantly more effective than placebo, with 12 of 17 patients responding to this agent versus only 6 of 19 to placebo. An early high plasma level of DVPX (i.e., days 2 to 6) predicted response, whereas rapid cycling, predominant euphoria or dysphoria, family history of mood disorder, increased manic severity, and EEG abnormalities did not (160).

Second, Bowden et al. (113) provided clear evidence for the efficacy of DVPX in acute mania in a multicenter, double-blind, random-assignment, parallel-designed study that compared DVPX with placebo, using lithium as an active control. Marked improvement of manic symptoms occurred in

48% of the DVPX group (n = 69)
49% of the lithium group (n = 36)
25% of the placebo group (n = 74)

Thus, both drug-treated groups experienced significantly greater benefit than that achieved with placebo (p = 0.004 for DVPX; p = 0.025 for lithium). DVPX-treated patients demonstrated significantly greater improvement in elevated mood, decreased need for sleep, excessive activity, and motor hyperactivity compared with placebo-treated patients, whereas lithiumtreated patients showed greater improvement in excessive activity and motor hyperactivity when compared with placebo. Notably, DVPX was similarly effective in manic patients with or without a rapid-cycling course, and response to this agent was equivalent in those who were previously responsive or nonresponsive to lithium. Supplemental chloral hydrate or lorazepam was used in the first 10 days; otherwise, there were no additional rescue medications. The definitive assessment of most patients occurred in the second and the third weeks when no adjunctive medication was used. Finally, only the lithium group had a significantly higher dropout rate due to adverse effects in comparison with those in the placebo group.

TABLE 10-8 DIVALPROEX VERSUS PLACEBO FOR ACUTE MANIA^a

	Percent Responders
Study	Number	Divalproex	Placebo	Difference
Pope et al., 1991	36	71	29	42
Bowden et al., 1994	139	64	32	32
Total	175	65	32	33
^aχ² = 18.3; df = 1; p = 0.00002.

The results of a meta-analysis on these two placebo-controlled trials are shown in Table 10-8. A decrease of 7 or more points on the Mania Rating Scale (MRS) was used to define response. Sixty-five percent of the DVPX patients met this criterion, compared with 32% on placebo (χ² = 18.3; p = 0.00002). Using a continuous method, the probability that DVPX is statistically superior to placebo is 0.0004, with a cumulative effect size of 0.55 (CI = 0.25 to 0.86).

A subsequent trial using a similar design also found the divalproex extended-release formulation superior to placebo in 377 bipolar patients experiencing an acute manic or mixed episode (161).

Other reports also suggest that this agent is helpful in subgroups often resistant to lithium. Calabrese et al. (162) used a prospective longitudinal naturalistic open design to examine 101 VPA-treated rapid-cycling bipolar patients and found that predictors of good antimanic response included decreasing or stable episode frequencies and nonpsychotic episodes. Papatheodorou et al. (163) conducted a preliminary open trial of six bipolar adolescents treated with DVPX and found five had marked improvement and one some improvement. Deltito et al. (164) reported that open treatment with DVPX monotherapy or copharmacy produced marked improvement in psychopathology in a group of 31 adolescents (aged 13 to 18 years) suffering from mixed-presentation bipolar disorder (n = 16), major depression (n = 7), mania (n = 4), or psychosis not otherwise specified (n = 4) (see Chapter 14 also). Other, open trials indicated that this agent was effective and well tolerated in elderly patients, those with concurrent substance or alcohol abuse, and those with organic mood disorders (165,166 and 167).

Valproate Versus Lithium. The previously discussed Bowden et al. (113) study found DVPXDR was similar in efficacy to lithium, which was used as an active comparator. Freeman et al. (129) conducted a 3-week, double-blind, parallel-group comparison of DVPX and lithium for acute mania. Both drugs demonstrated clinically significant efficacy (i.e., 9 of 14 responded to DVPX and 12 of 13 to lithium), and there was no difference in the need for rescue medications (i.e., lorazepam or chloral hydrate) between groups. Response to DVPX was associated with high pretreatment depression scores.

Valproate Versus Second-Generation Antipsychotics. Three class I, controlled trials compared VPA with OLZ. The results are discussed in section “Second-Generation Antipsychotics.”

Valproate Plus Other Psychotropics

Valproate Plus Antipsychotics. MullerOerlinghausen et al. (168) compared FGA monotherapy with VPA plus an FGA in a 3-week prospective, double-blind, placebo-controlled, multicenter study. A total of 136 acutely manic patients received a fixed dose of 20 mg/kg/day of VPA or placebo, in addition to an antipsychotic. They found that the mean dose of antipsychotic (in HPDL equivalents) declined during the study in those subjects who were also receiving VPA. Based on a 50% reduction in the baseline Young Mania Rating Scale (YMRS), 70% on combination therapy versus 46% on antipsychotic alone achieved this level of improvement (p = 0.005). Adverse events did not differ significantly between the two groups. The authors concluded that in comparison with an antipsychotic alone, the combination treatment

Allowed for lower amounts of FGA
Produced a more rapid remission of symptoms
Produced a significantly greater improvement in symptoms

Several subsequent trials considered the combination of various SGAs with VPA, and these are discussed in “Second-Generation Antipsychotics” section.

Administration of Valproate

VPA is available in various formulations, including DVPX (a compound composed of sodium VPA and valproic acid); dipropylacetic acid; and a closely related form, valpromide or dipropylacetamide. Due to its rapid absorption, blood levels peak 1 to 4 hours after oral administration, and the half-life ranges from 6 to 16 hours. It is metabolized primarily through the liver and eliminated in the urine. VPA is highly protein bound and usually does not saturate binding sites at serum levels below 45 to 50 μg/mL. Thus, this level would be the expected minimal threshold for its psychotropic effects.

The most commonly used formulation of VPA is DVPX-DR, which is usually started at 750 to 1,000 mg/day. Doses are titrated every few days to achieve plasma levels in the 50 to 125 μg/mL range or until side effects prohibit further increases (169). Wassef et al. (170) and Davis et al. (171), however, reported that higher doses may be needed, particularly with the extended-release formulation (DVPX-ER) to achieve optimal therapeutic plasma levels. For most patients, an adequate trial typically requires total doses ranging from 1,000 to 2,500 mg/day.

DVPX is also safely administered using a loading dose strategy (172,173). For example, Keck et al. (172) found that 20 mg/kg/day led to therapeutic plasma levels within 5 days. This is important, given evidence that the antimanic activity of DVPX is more likely to occur after achieving adequate serum concentrations. In this study, patients tolerated 20 mg/kg/day in divided doses for 5 days, with rapid onset of antimanic response. With increasing pressure to limit the length of hospital stays, such an accelerated response rate could represent a significant advantage. Preliminary reports with intravenous VPA administration, however, generated conflicting results (174,175).

Case Example

A 34-year-old man was hospitalized on CBZ, CPZ, lithium, and lorazepam for an acute exacerbation of mania. Despite a history of rapid cycling, he responded to an initial trial of lithium, thiothixene, and lorazepam but then relapsed. Increasing the thiothixene (up to 120 mg/day) was unsuccessful and poorly tolerated. The patient then improved on a regimen of loxapine (up to 250 mg/day) and clonazepam (up to 20 mg/day), but doses of each had to be reduced because of intolerable adverse effects (i.e., excessive sedation, drooling, pseudoparkinsonism), and he suffered another relapse. He again received CPZ (up to 900 mg/day) and DVPX (plasma level stabilized at 80 μg/mL), while the clonazepam was tapered slowly and discontinued. He gradually became euthymic and was able to leave the hospital after a 4-month stay, stabilized solely on CPZ (600 mg/day) and DVPX (1,250 mg/day).

CARBAMAZEPINE

The primary psychiatric application of CBZ is for bipolar disorder, based on the initial work of two Japanese groups in the early 1970s (13,14). Interestingly, CBZ has a chemical structure resembling imipramine and was originally synthesized as a possible antidepressant agent. The FDA has approved an extended-release formulation of CBZ for treatment of acute bipolar mania.

Our analyses of the efficacy of CBZ in acute mania and for maintenance therapy find it comparable to lithium, VPA, or SGAs. The spectrum of efficacy of CBZ appears to be similar to that of lithium; however, as noted earlier, it may be more effective for mixed or dysphoric mania, rapid cyclers, and more severe episodes.

Dehing (12) conducted the first controlled study of the behavioral effects of CBZ in epileptic patients. He reported that it made them more active and communicative, less egocentric and stubborn; improved dysphoria, emotional lability, aggressiveness, and outbursts of rage; and had a positive effect on apathy, depression, anxiety, and hypochondriasis. He then studied its effects in a mostly nonepileptic, chronic psychiatric population that suffered from such varied disorders as dementia, psychosis, mental deficiency, and psychopathy but not bipolar disorder. In a double-blind, random-assignment design, he treated most of these patients for 1 month with placebo or CBZ. He continued the investigation after the double-blind phase on patients initially assigned to placebo and added others to the trial, for a total sample of 58. Those receiving CBZ showed a marked improvement in contrast to those on placebo, but two patients developed either a slight or a marked aggravation of their disorder.

Dehing (12) qualitatively identified aggressiveness and outbursts of rage as the symptoms most helped by CBZ. It is of interest that in addition to their mood-stabilizing effects, there is evidence that VPA and lithium also exert an antiaggressive effect (176). This raises the question of whether these agents act at a more fundamental level than a specific disorder. Thus, like anti-inflammatory agents, they may benefit various disorders that share phenomenological and pathophysiological similarities.

Carbamazepine Versus Placebo or Other Psychotropics

Carbamazepine Versus Placebo. Post and Uhde (177) studied nine manic patients using an ABA design. Three patients had a good response to CBZ and relapsed when switched to placebo; one had an equivocal response but relapsed when CBZ was discontinued; one responded to CBZ but failed to relapse on placebo; and three neither demonstrated a clearcut response to CBZ nor relapsed when administered placebo later. To some degree, the placebo lead-in period in the ABA design controls for the placebo effect, but with no true control group, improvement could be due to cycling, spontaneous remission, nonspecific effects of hospitalization, or an unrecognized carryover medication effect.

Goncalves and Stoll (178) studied six patients on CBZ and six on placebo; however, substantial amounts of antipsychotic augmentation were used. Virtually, every placebo patient had some additional HPDL, and two of three also received other supplemental antipsychotics. In the CBZ group, four of six received supplemental HPDL, with none needing other antipsychotics. Despite the greater use of antipsychotics in the placebo group, the CBZ group was statistically superior. Because significant amounts of concomitant antipsychotics were used, it is hard to draw any firm conclusion, but the outcome suggests the need for antipsychotic supplementation with CBZ.

Two class I, large, multicenter, parallel-group, placebo-controlled, randomized trials with beaded extended-release CBZ (CBZ-ERC) are published (179,180). In the first trial, manic or mixed episode bipolar I subjects were assigned to active drug (n = 101) or placebo (n = 103) for 3 weeks. All subjects were hospitalized for at least 7 days. The mean final dose of CBZ-ERC was 756 (±413) mg/day. Based on the primary outcome measure (i.e., YMRS), active treatment was significantly better than placebo (p = 0.032, LOCF analysis) beginning with the second week. At study end, 42% on CBZ-ERC and 22% on placebo met criteria for response (i.e., ≥50% decreased in baseline YMRS). Dizziness, nausea, and somnolence were more frequent with active treatment versus placebo.

The second trial involved 293 subjects studied in a similar design. Sixty percent completed the 3-week trial, with a significantly greater number of placebo subjects discontinuing due to lack of efficacy (p = 0.001). As in the previous trial, CBZ-ERC was significantly better than placebo after 1 week based on mean YMRS improvement scores (p = 0.05). Adverse effects were similar to the first trial. In addition,
CBZ-treated subjects had a significant increase in total cholesterol levels.

The results of these two studies led to an FDA-approved indication for acute bipolar mania.

Carbamazepine Versus Lithium. Lerer et al. (181) compared CBZ with lithium without concomitant medication. Fourteen patients were randomly assigned to lithium and 14 to CBZ. Lithium appeared superior, with 11 patients improving, in contrast to only 4 patients improving with CBZ. In a second study, Small et al. (130) found CBZ and lithium comparable in efficacy in a group of 52 hospitalized, treatment-refractory manic patients. This was a double-blind, randomized design that followed patients during both the acute (i.e., 8 weeks) and the maintenance phase (up to 2 years). Of note, there was a trend favoring the lithium group on the survival analysis (p < 0.14).

Carbamazepine Versus Antipsychotics. Because there is evidence that lithium is superior to FGAs in acute episodes of classic mania, it is relevant to review the data comparing CBZ with antipsychotics. Four studies investigate acute mania (two with random-assignment, doubleblind conditions and no concomitant drugs) (182,183,184 and 185). Although the two class I studies found no statistical difference between CBZ and CPZ, one found a trend favoring CBZ and the other CPZ. When a meta-analysis combined these two studies, the pooled data show the relative outcomes to be virtually equal (i.e., effect size = 0.05; Z = 0.2). We note that the four wellcontrolled studies comparing lithium with antipsychotics found lithium significantly superior.

Carbamazepine Plus Other Psychotropics

Carbamazepine Plus Lithium. Okuma et al. (185) added CBZ to the previous treatment of 107 affective, 54 schizophrenic, and 26 schizoaffective patients. Improvement was reported in 73%, 56%, and 62%, respectively. In an open design, Nolen (186) added CBZ to lithium (and, when necessary, an antipsychotic and/or antidepressant) in a small group of treatment-resistant manic patients, who then showed further improvement. Kramlinger and Post (187) added lithium to CBZ in seven patients with varying degrees of mania, noting that six improved and one worsened. Because there was no control group, we do not know whether the patients would have shown similar improvement, had the CBZ alone been continued for a longer period of time. Indeed, one responder had been on CBZ for 2 weeks, and another for 3 weeks, but the other four responders had been on treatment about 1 month, which is sufficient time for the effects of CBZ to peak. Thus, although the data are suggestive that lithium may augment the effects of CBZ, the absence of a control group and the small sample sizes do not allow for a definitive conclusion.

Carbamazepine Plus Antipsychotics. Klein et al. (188) augmented HPDL with CBZ in a group of newly admitted, highly destructive psychotic patients (affectively disordered or schizophrenic) and found that 19 of 23 CBZaugmented patients improved, in contrast to 11 of 20 of the placebo-augmented patients. Mueller and Stoll (189) and Goncalves and Stoll (178), using random- assignment, double-blind designs in small samples of six to ten patients per group, found that the addition of CBZ to HPDL also produced benefit over the antipsychotic alone. Specifically, Mueller and Stoll (189) found that less supplemental medication was needed when CBZ was added, and Goncalves and Stoll (178) found CBZ superior to placebo supplementation of HPDL.

Lenzi et al. (190) compared patients randomly assigned to lithium or CBZ augmented by CPZ. During the first week, every patient required CPZ; in the second week, 14 of 15 patients required it; and in the third week, 11 of 15 patients in each group required it. The therapeutic result of the CPZ-CBZ combination was equal to the CPZ-lithium combination, and the only difference was that patients on CBZ required less CPZ in the first week. Lusznat et al. (191) also found the CBZ-antipsychotic combination equals to the lithium-antipsychotic combination. In studies in which most, but not all, patients receive two active drugs, the design clouds the effectiveness of the drug alone versus an augmentation strategy. When every patient receives a basic drug at a constant dose that is then supplemented, it is easier to determine whether the second drug is helpful.

Administration of Carbamazepine

CBZ is also indicated for the management of temporal lobe epilepsy and paroxysmal pain disorders. Its anticonvulsant actions are apparently associated with the ability to reduce postsynaptic responses and to block posttetanic potentiation. The initial half-life ranges from 25 to 65 hours, but due to the ability of CBZ to induce its own metabolism (i.e., autoinduction), this may be reduced to 12 to 17 hours after several weeks of treatment.

The routine pretreatment workup includes assessment of baseline hematological and hepatic functions because these two organ systems may be adversely affected. Once baseline medical status is established, typical starting doses are 400 to 600 mg/day, given in divided doses. Of note, Weisler et al. (192) reported that efficacy was comparable when administering CBZ-ERC on a BID or QHS schedule. Increments of 200 mg/day can be given every 3 to 5 days until adverse effects preclude higher dosing or desired clinical response is reached. Less aggressive titration and even dose reduction may be required early in the treatment until the patient develops tolerance to the adverse effects. In terms of adequate blood levels, 4 to 12 μg/mL is considered the accepted therapeutic range for CBZ when used as an anticonvulsant, but the ideal blood level of CBZ as an antimanic is unknown. Preliminary data, however, find a relationship with cerebrospinal fluid levels of the principal epoxide metabolite of CBZ and clinical response. Therefore, it is desirable to titrate the dose based on clinical response and adverse effects rather than rigidly relying on plasma levels. As we discuss later, therapeutic drug monitoring (TDM) of CBZ, especially during the first several weeks of therapy, is crucial due to the phenomenon of autoinduction and the potential for clinically significant drug interactions.

Case Example

A 23-year-old woman with a long history of bipolar disorder resistant to lithium monotherapy and characterized by mixed episodes and rapid cycling was hospitalized in a manic phase. She was on lithium, CBZ, and trifluoperazine, with a therapeutic lithium level but a CBZ level of only 5 μ/mL. The patient underwent a medication washout, during which she deteriorated and was then placed on DVPX. Because of increasing confusion, nausea, and vomiting at therapeutic levels, she was switched back to CBZ plus low-dose trifluoperazine. She had minimal response to CBZ blood levels in the range of 6 to 10 μg/mL, but when levels were titrated to 12 to 14 μg/mL, she demonstrated marked stabilization in mood, with minimal adverse effects. The antipsychotic was discontinued, the patient continued to do well, and was discharged to outpatient follow-up.

Although the sequence is strongly suggestive of a beneficial effect with higher CBZ levels, we cannot rule out the possibility of a spontaneous improvement due to her history of rapid cycling.

ANTIPSYCHOTICS

Antipsychotics represent a third major drug therapeutic strategy for bipolar disorder. Although FGAs (e.g., HPDL) are frequently used to manage various symptoms of bipolar disorder, more recent studies and clinical experience focus on the SGAs
that demonstrate mood-stabilizing properties separate and distinct from their antipsychotic effects (193,194 and 195). The lower incidence of acute EPS (and probably tardive dyskinesia) with SGAs usually favors these agents over FGAs for mood disorders. Weight and metabolic complications, however, remain an important consideration. Studies consider the SGAs from several perspectives as a treatment strategy for bipolar disorder, including

Acute monotherapy
- Mania
- Depression
Acute augmentation therapy
- Mania
- Depression
- Maintenance monotherapy
- Maintenance augmentation therapy
- Reduction in suicidality

In this context, a meta-analysis by Scherk et al. (196) concluded that in the treatment of acute mania, the most compelling argument for their use is in combination with standard mood stabilizers (e.g., lithium, VPA). An important consideration is that most of these trials were industry sponsored and for regulatory purposes involved populations and design requirements that may not readily translate to general practice.

The pharmacokinetic properties of antipsychotics are discussed in Chapter 5.

First-Generation Antipsychotics

CPZ and HPDL are the best studied FGAs for management of acute mania. Indeed, CPZ was the first antipsychotic to receive FDA approval for this indication. More recently, HPDL monotherapy was compared to placebo and various SGAs. It has also been considered as an adjunctive treatment combined with standard mood stabilizers (e.g., lithium). The results of these trials generally support antimanic effects and possibly a more rapid onset of action compared with other agents (195,197).

Second-Generation Antipsychotics

Early reports found adjunctive clozapine benefited previously refractory, mood-disordered patients (e.g., bipolar, schizoaffective) who improved rapidly and significantly with the addition of this agent (198,199). Further, many patients sustained their early gains in psychosocial functioning over a 3- to 5-year period. Subsequently, several trials confirmed the efficacy of other SGAs as monotherapy or augmentation therapy for acute mania, as well as for more severe psychotic and/or refractory mood disorders (17,18) (Table 10-9). Based on this evidence, the FDA has approved OLZ, RISP, QTP, ZPD, ARIP, and ASEN for treatment of acute mania. Further, OLZ, ARIP, QTP, and ZPD are currently approved for maintenance treatment; OLZ, RISP, and QTP as adjuncts to standard mood stabilizers; and QTP monotherapy or OLZ plus fluoxetine for acute bipolar depression. Of interest, ARIP monotherapy did not separate from placebo in two controlled trials for bipolar depression (200).

The role of various SGAs, either as primary or adjunctive treatments for mood disorders, continues to be explored. Theories regarding their mechanism of action include differential effects on DA receptor subtypes (e.g., impact on D4 receptors that exist in high density in the limbic system with clozapine (201); DA partial agonism with ARIP); the greater 5-HT₂ to D₂ antagonism of SGAs in comparison with FGAs; modulation of 5-HT_2A and 5-HT_1A activity (e.g., ARIP); and reuptake inhibition at NE and 5-HT transporters (e.g., ZPD) (202). In this context, Swerdlow (203) notes that subcortical DA system dysfunction contributes to the symptoms of mood disorders, particularly increased or decreased goal-directed behaviors and perceived changes in reinforcement. Still, the author comments that it is not enough to view such pathology in the context of too little or too much subcortical DA. Thus, given the complex motor, cognitive, and affective disturbances in depression, the role of subcortical DA dysfunction and antidepressant action should be viewed within the context of an integrated cortico-striato-pallido-thalamic circuitry.

Regarding the 5-HT system, Keck et al. (204) cite several lines of evidence to support that the thymolytic effects of clozapine (and perhaps other agents with similar properties) are mediated by the antagonism of 5-HT₂ receptors, including

These receptors are implicated in the action of standard antidepressants
Most, but not all, platelet radioligand studies in medication-free patients with major depressive disorder found a significant increase in the number of binding sites of this receptor compared with control subjects
Radioligand binding studies of the postmortem brains of patients with major depressive disorder who did not receive antidepressants showed significant increases in the number of these receptors compared with both patients who had received antidepressants and normal control subjects

TABLE 10-9 SECOND-GENERATION ANTIPSYCHOTICS FOR TREATMENT OF ACUTE MANIA

Type of Study

Author(Year)

Sample

Results

Comments

Olanzapine

Monotherapy

OLZ vs. PLAC

Randomized DB-PC 3 weeks

Tohen et al. (1999)

N = 139

Bipolar mania

OLZ produced significantly greater mean improvement in YMRS and response rates vs. PLAC (49% vs.24%)

OLZ = PLAC for EPS

Starting dose = 10 mg/day; dose range = 5-20 mg/day

Somnolence, dizziness, dry mouth, and weight gain were significantly greater with OLZ

OLZ vs. PLAC

Randomized DB-PC 4 weeks

Tohen et al., (2000)

N = 115

Bipolar

Manic/mixed

OLZ produced significantly greater improvement in YMRS and response rates: (65% vs. 43%)

OLZ = PLAC for EPS

Starting dose = 15 mg/day; dose range = 5-20 mg/day

OLZ caused significantly greater weight gain

OLZ vs. DVPX

Randomized DB

3 week acute

9 weeks extension

Tohen et al. (2002) (2008)

N = 248

Bipolar

Manic/mixed

Based on YMRS, remission rates were significantly greater with OLZ at end point, regardless of the presence of psychosis

Improvement in depressive symptoms based on HDRS with OLZ = DVPX

Rates of relapse higher and time to relapse shorter with DVPX durinq a 9-week extension period

OLZ = 2-20 mg/day

DVPX = 500-2,500 mg/day

OLZ caused significantly more weight gain, dry mouth, and somnolence than DVPX

DVPX caused more nausea than OLZ

OLZ vs. DVPX

Randomized DB 12 weeks

Zajecka et al. (2002)

N = 120

Bipolar mania

OLZ = DVPX in terms of efficacy variables (MRS)

DVPX reported to have a better AE profile, including significantly less weight gain than OLZ

One death due to DKA on OLZ

OLZvs. Li +

Randomized DB 4 weeks

Berk et al. (1999)

N = 30

Bipolar mania

OLZ at least as effective as Li+ for mania

No difference in BPRS, CGI, or MRS

OLZ did not differ from Li+ in emergent EPS

OLZvs. Li +

Randomized DB

4 weeks

Niufan et al. (2008)

N = 140

Bipolar

Manic/mixed

OLZ > Li +

Based on CCG-BP change scores

More AEson OLZvs. Li +

OLZ vs. HPDL

Randomized DB

6-week acute phase

6-week continuation phase

Shi et al. (2002)

N = 453

Bipolar mania

Remission rates: OLZ = HPDL

Quality of life and work status OLZ > HPDL

Flexible dosing: OLZ = 5-20 mg/day (n = 234); HPDL = 3-1 5 mg/day (n = 219)

OLZ vs. RISP Randomized DB 3 weeks

Perlis et al. (2006)

N = 329

Bipolar

Manic/mixed

YMRS change score primary outcome measure

OLZ = RISP

Comparable results on several other mood/functional measures

OLZ vs. lorazepam Randomized DB-PC

Meehan et al. (2001)

N = 201

Bipolar, agitated mania

PANSS-EC primary outcome measure

OLZ (10 mg × 2, then 5 mg prn) Lorazepam (2 mg × 2, then 1 mg prn)

Acute parentera 24 hours

OLZ > lorazepam and PLAC at 2 hours OLZ > PLAC at 24 hours

PLAC (× 2, then OLZ 10 mg prn) No differences among groups for AEs

Augmentation therapy

OLZ plus MS

Randomized DB-PC

Added to MS

2 weeks on MS alone

6 weeks on

combination

Tohen et al (2002)

N = 344

Bipolar

Manic/mixed

Partially nonresponsive

OLZ cotherapy reduced YMRS and HDRS significantly more than MS monotherapy (p < 0.003)

OLZ > PLAC for mania

OLZ > PLAC for response rates (68% vs. 45%)

Patients randomized after 2 weeks of nonresponse to MS monotherapy

OLZ caused more somnolence, dry mouth, weight gain, tremor

OLZ plus DVPX

Randomized DB-PC

Added to MS after

2 weeks

Houston (2009)

N = 200

Bipolar mixed episode

OLZ > PLAC in producing a greater and earlier improvement in manic and depressed symptoms

Increases in weight and fasting glucose levels were greater with OLZ versus PLAC

Risperidone

Monotherapy

RISP vs. PLAC

Randomized DB-PC

3 weeks

Hirschfield et al. (2004)

N = 259

Bipolar mania

YMRS was primary outcome measure

RISP superior to PLAC

Rapid onset of action seen as early as day 3

RISP mean modal dose = 4.1 mg/day

RISP > PLAC for somnolence and EPS

RISP vs. PLAC

Randomized DB-PC

3 weeks

Khanna et al (2005)

N = 290

Bipolar Manic/mixed

YMRS was primary outcome measure

Robust efficacy for RISP over PLAC

Onset of action as early as 1 week

RISP mean modal dose = 5.6 mg/day

EPS more common in RISP vs PLAC group

RISP vs. PLAC

Randomized DB-PLAC /HPDL arms

Smulevich et al (2005)

N = 438

Bipolar mania

YMRS change score was primary outcome measure

RISP superior to PLAC

RISP mean dose = 4.2 mg/day

HPDL mean dose = 8.0 mg/day

HPDL > PLAC on ESRS score

3 weeks

9-week open continuation

at 3 weeks and 9 weeks

RISP = HPDL

RISP vs. Li and HPDL

Randomized DB

4 weeks

Segal et al. (1998)

N = 45

Bipolar mania

All three groups showed similar improvement in total MRS and BPRS scores

No EPS difference between RISP and HPDL

RISP vs. HPDL

6 weeks

Janicak et al. (2001)

N = 62

Schizoaffective bipolar and depressed subtypes

RISP = HPDL in reducing mania/psychosis

RISP > HPDL in reducing depression

EPS: HPDL > RISP

Augmentation therapy

RISP vs. HPDL

Randomized DB-PC Added to MS 3 weeks

Sachs et al. (2002)

N = 156

Bipolar

Significant decrease in YMRS scores with RISP and HPDL (with or without psychosis)

Mean modal dose: RISP = 3.2 mg/day; HPDL = 6.2 mg/day

EPS: HPDL > RISP

RISP plus MS

Randomized DB-PC

3 weeks

Yatham et al (2003)

N = 151

Bipolar mania

YMRS: RISP vs. PLAC trend in favor of RISP

RISP > PLAC when CBZ-treated subjects excluded

CGI: RISP > PLAC

Mean dose = 3.7 mg/day

RISP efficacious in both psychotic and nonpsychotic subjects

AEs similar in both groups, except EPS greater with RISP

Quetiapine

Monotherapy

QTP vs. PLAC

Randomized DB-PC

HPDLarm

12 weeks

Mclntyre et al (2005)

N = 302

Bipolar Manic

QTP superior to PLAC and comparable to HPDL on YMRS change scores at days 21 and 84

HPDL but not QTP separated from PLAC at day 21 for response rates

QTP mean last week dose in responders = 559 mg/day (day 21); 84% of responders received 400-800 mg/day

HPDL mean last week dose in responders = 5.2 mg/day (day 21) QTP = PLAC for AEs

QTP vs. PLAC Randomized

DB-PLAC/Li arms 12 weeks

Bowden et al. (2005)

N = 302

Bipolar Manic

QTP superior to PLAC and comparable to Li on YMRS change scores at days 7, 21.and 84

QTP mean last week dose in responders = 586 mg/day (day 21)

Li target serum concentration = 0.6-1.4 mEq/L

Augmentation therapy

QTP plus DVPX

Randomized DB-PC

6 weeks

DelBello et al. (2002)

N = 30

Bipolar

Manic/mixed

Adolescents

QTP plus DVPX superior to QTP plus PLAC on YMRS score reduction from baseline and response rate

QTP mean dose = 432 mg/day

Manic symptoms improved more rapidly with QTP plus DVPX

Somnolence greater in QTP plus DVPX vs. PLAC

QTP plus MS Randomized DB 3 weeks or 6 weeks

Sachs et al. (2004)

Yatham et al. (2004)

N = 402 Bipolar I mania npatients

QTP plus MS superior to MS plus PLAC on mania score reduction and response rates

QTP mean final dose = 580 mg/day in responders

Somnolence, dry mouth, asthenia, and postural hypotension greater with QTP augmentation

Ziprasidone

Monotherapy

ZPD vs. PLAC

Randomized DB-PC 3 weeks

Keck et al. (2003)

N = 210

Bipolar

Manic/mixed

ZPD > PLAC

Rapid, sustained improvement in SADS-C, PANSS, CGI-S

ZPD mean dose = 130 mg/day

Rapid onset of action

ZPD = PLAC in terms of tolerability

ZPD vs. PLAC

Randomized DB-PC

3 weeks

Potkin et al (2004)

N = 202

Bipolar

Manic/mixed

ZPD > PLAC

ZPD dose range = 80-160 mg/day

Rapid onset of action

Well tolerated

ZPD vs. PLAC

DB-PC/HPDLarms

3 weeks acute

4 weeks continuation

Vieta et al. (2008)

N = 438

Bipolar I mania

ZPD/HPDL > PLAC

ZPD = HPDLat 12 weeks

ZPD (20 mg) vs. ZPD (2 mg)

Randomized DB

Acute parentera

Fixed dose

24 hours

Daniel et al (2004)

N = 79

Bipolar

SA, bipolar type

ZPD (20 mg) > ZPD (2 mg)

Based on: BARS; CGI-S; PANSS-agitation subscale

Rapid onset of action

Well tolerated

No excessive sedation

Aripiprazole

Monotherapy

ARIP vs. PLAC

Randomized DB-PC

npatients 3 weeks

Keck et al. (2003)

N = 262

Bipolar

Manic/mixed

ARIP > PLAC by day 4

Response rate based on a 50% reduction in YMRS at week 3:

ARIP = 40% PLAC = 19%

ARIP starting dose = 30 mg/day (Mean dose = 28 mg/day) Well tolerated

ARIP vs. PLAC

Randomized DB-PC

Inpatient 3 weeks

Sachs et al. (2006)

N = 268

Bipolar

Manic/mixed

ARIP > PLAC by day 4

Response rates based on a 50% reduction in YMRS at week 3: ARIP = 53% PLAC = 32%

ARIP starting dose = 30 mg (Mean dose = 27.7 mg/day) Well tolerated

ARIP vs. PLAC

Randomized DB-PC

Inpatient 3 weeks

Data on file with Otsuka American Pharmaceuticals (2004)

N = 386 Bipolar

Manic/mixed

ARIP = PLAC Response rates:

ARIP (15 mg) = 41% ARIP (30 mg) = 45% PLAC = 38%

Based on a50% reduction in YMRS at week 3

ARIP fixed dose = 1 5 or 30 mg

High PLAC response rate

ARIPvs. HPDL

Randomized DB

hpatient/outpatient 12 weeks

Vieta et al. (2005)

N = 338

Bipolar

Manic/mixed

ARIP = HPDL based on YMRS change scores

ARIP > HPDL

Response rates at week 12 (based on >50% reduction in YMRS)

ARIP = 48%

HPDL = 32%

Starting dose

ARIP = 15 mg/day

(Mean dose = 21.6 mg/day)

HPDL = 10 mg/day

(Mean dose = 11.1 mg/day)

ARIP superior to HPDL in time to discontinuation for all reasons

ARIP vs. PLAC

Randomized DB-PC

HPDL controlled

3 weeks acute

9 weeks maintenance

ARIPvs. HPDL

Young et al (2009)

N = 485

Bipolar

Manic/mixed

ARIP and HPDL > PLAC at 3 weeks

Response rates

ARIP = 47%

PLAC = 38%

HPDL = 50%

ARIP > PLAC after 3 weeks

ARIP = HPDL after (9 weeks)

ARIP vs. PLAC

Randomized

DB-PC and lithium

controlled

ARIPvs. lithium

Keck et al. (2009)

N = 480

Bipolar

Manic/mixed

ARIP and lithium > PLAC at 3 weeks

ARIP > PLAC after 3 weeks

ARIP = lithium over 12 weeks

ARIP dose 15-30 mg/day

ARIP vs. PLAC

Randomized DB

PLAC/Lorazepamvs. ARIP

Acute parentera 2 hours

Zimbroff et al. (2007)

N = 301

Bipolar

Manic/mixed with acute agitation

PANSS-EC change score:

ARIP (9.75 mg)> PLAC

ARIP (1 5 mg) > PLAC

Lorazepam (2 mg) > PLAC

Oversedation less frequent with ARIP 9.75 mg vs. ARIP 15 mg or lorazepam 2 mg

ARIP vs. PLAC plus

MS (Li+; VPA)

DB-PC

6 weeks

Vieta et al. (2008)

N = 384

Bipolar

Not responsive to MS

ARIP > PLAC

Primarily due to combination with VPA

ARIP plus Li+ had more akathisia/tremor vs. ARIP plus VPA

Asenapine

Monotherapy

ASEN vs. PLAC

Randomized DB-PC

Initially inpatients

3 weeks

OLZ for assay

sensitivity

Mclntyre et al. (2009)

N = 488

Bipolar I

Manic/mixed

ASEN > PLA

Based on YMRS change score

Response rates based on a 50% reduction in YMRS at week 3:

ASEN = 43%

PLAC = 34%

ASEN flexibly dosed at 5-10 mg SL BID (mean dose 5 18.2 mg/day)

OLZ (mean dose = 1 5.8 mg/day)

ASEN had low rates of EPS and modest weight/metabolic changes

ASEN vs. PLAC

Randomized DB-PC, Initially inpatients 3 weeks OLZ for assay sensitivity

Mclntyre et al. (2007)

N = 480 (ITT)

Bipolar I

Manic/mixed

ASEN > PLAC

Based on YMRS change score

Response rates based on >50%

reduction in YMRS at week 3:

ASEN = 42%

PLAC = 25%

ASEN flexibly dosed at 5-10 mg SL

BID (mean dose 18.4 mg/day)

OLZ, olanzapine; PLAC, placebo; DB, double-blind study; PC, placebo-controlled study; YMRS, Young Mania Rating Scale; DVPX, divalproex sodium; HDRS, Hamilton Depression Rating Scale; MRS, Mania Rating Scale; AE, adverse event; Li, lithium ion; DKA, diabetic ketoacidosis; BPRS, Brief Psychiatric Rating Scale; CGI, Clinical Global Impression; EPS, extrapyramidal symptoms; HPDL, haloperidol; PANSS, Positive and Negative Syndrome Scale; EC, excited component; prn, as needed; MS, mood stabilizer; RISP, risperidone; ESRS, Extrapyramidal Symptom Rating Scale; CBZ, carbamazepine; QTP, quetiapine; Li, lithium; ZPD, ziprasidone; SADS-C, Schedule for Affective Disorders and Schizophrenia-Change; CGI-S, Clinical Global Impression-Severity; SA, schizoaffective; BARS, Behavioral Activity Rating Scale; ARIP, aripiprazole; ASEN, asenapine

Adapted from Nour H, Baslet G, Janicak PG. The role of second-generation antipsychotics in the treatment of mood disorders. Part I: Acute bipolar mania. Janicak PG, Pavuluri M, eds. Contemporary Psychiatry. 2003;2(6):1-10.

In summary, the authors believe that these lines of evidence support the hypersensitivepostsynaptic, serotonergic receptor theory first proposed by Aprison et al. (205).

Clozapine Plus Other Psychotropics

Open Trials. Owen et al. (206) conducted an open-label, compassionate use clozapine treatment protocol in 37 chronically psychotic patients (25 schizophrenia, 12 schizoaffective). Clozapine produced a highly significant improvement in psychopathology as measured by the BPRS. Interestingly, schizoaffective patients had significantly lower total scores than their schizophrenic counterparts at the final rating.

McElroy et al. (207) surveyed the response of 85 consecutive patients, including 14 bipolar patients with psychotic features, who received clozapine for 6 weeks. The response rates of the schizoaffective patients (both bipolar and depressed subtypes), as well as those with bipolar disorders with psychotic features, were excellent (i.e., almost 90%) and substantially better than the response rate for the pure schizophrenic group (i.e., 46%).

Four clozapine open, monotherapy trials found a substantial mood-stabilizing effect for this agent in bipolar or schizoaffective manic or mixed patients (208,209,210 and 211). Case series also report that this agent is helpful as an adjunct for treatmentresistant, rapid-cycling patients (212,213).

Suppes et al. (198) successfully treated seven treatment-refractory, psychotic, dysphoric manic patients with clozapine in a prospective, randomized open trial. They reported that during a 3- to 5-year follow-up, most sustained substantial gains in psychosocial function and that the six who remained on this agent were not rehospitalized. Subsequently, these same authors reported on three rapid-cycling, nonpsychotic, bipolar patients who benefited from clozapine, leading the authors to opine that this drug may have mood-stabilizing properties separate from its antipsychotic effects (199).

Another report suggests that the combination of clozapine plus VPA may result in greater efficacy and good tolerability in most patients (214). In this context, we would avoid using clozapine plus CBZ due to the possibility of further decreases in white blood cell (WBC) counts as well as clozapine plus a BZD because of potential respiratory compromise (215).

Olanzapine Versus Placebo or Other Psychotropics

Olanzapine Versus Placebo. Tohen et al. (216,217) published the results of two doubleblind, placebo-controlled trials of OLZ for the treatment of acute mania (see Tables 10-9 and 10-10). The results of these trials ushered in the rapid development of SGAs as potential mood stabilizers.

The first study was 3 weeks in duration and included 139 patients hospitalized for an acute bipolar manic or mixed episode (216). The starting dose of OLZ was 10 mg/day, and the mean dose was 14.9 mg/day. Active drug treatment produced a statistically greater mean improvement than placebo on the YMRS scores whether psychosis was present or not. Further, 49% of the OLZ-treated group (n = 70) met the a priori criteria for response (i.e., a ≥50% reduction from baseline YMRS score) versus only 24% of the placebo-treated group (n = 69).

The second study was similar in design. It involved 115 patients, used a higher starting dose of OLZ (i.e., 15 mg/day, mean dose of 16.4 mg/day), used less rescue medication, and had a 4-week treatment duration. The outcome was similar with OLZ and statistically superior to placebo for core manic or mixed symptoms as well as for depression (217,218).

Combining these two studies, there is a total of 251 patients with a 56% response rate to OLZ versus a 33% response rate to placebo (see Table 10-10). Both studies found OLZ superior to placebo and also replicated each other in that the degree of superiority was the same. Sedation and weight gain were significantly greater in the OLZ- versus the placebo-treated group; however, there was no difference in EPS occurrence between the groups.

Olanzapine Versus Divalproex. Tohen et al. (219) conducted a 3-week, randomized, doubleblind trial comparing OLZ (5 to 20 mg/day) to DVPX (500 to 2,500 mg/day) for an acute manic or mixed episode (see Table 10-9). The primary outcome variable was baseline to end point
improvement in the mean total YMRS score. OLZ produced a significantly greater mean improvement compared with DVPX. Response (≥50% reduction in YMRS score) and remission (end point YMRS ≤ 12) rates were

Responders: 54.4% of OLZ-treated versus 42.3% of DVPX-treated patients
Remitters: 47.2% of OLZ-treated versus 34.1% of DVPX-treated patients

TABLE 10-10 SECOND-GENERATION ANTIPSYCHOTICS VERSUS PLACEBO FOR ACUTE MANIA

			Responders (%)
Study		N	SGA	Placebo	Difference (%)
			Olanzapine	Placebo
Tohen et al., 1999		136	49	24	25
Tohen et al., 2000		115	65	43	22
	Total	251	56	33	23
Study		N	Risperidone	Placebo
Hirschfeld et al., 2004		243	43	24	19
Khanna et al., 2005		203	73	36	37
Smulevich et al., 2005		294	48	33	15
	Total	740	54	31	23
Study		N	Quetiapine	Placebo
Brecher et al., 2003		302	61	39	22
Bowden et al., 2005		302	53	27	26
	Total	604	57	33	24
Study		N	Ziprasidone	Placebo
Keck et al., 2003 (2:1 ratio; MRS not YMRS)		210	50	35	15
Potkin et al., 2004 (2:1 ratio; MRS not YMRS)		202	46	29	17
	Total	412	48	32	16
Study		N	Aripiprazole	Placebo
Keck et al., 2003		248	40	19	21
Sachs et al., 2006		268	53	32	21
Keck et al., 2009		320	47	34	13
	Total	836	47	28	19
Study		N	Asenapine	Placebo
McIntyre et al., 2010		283	42	25	17
McIntyre et al., 2009		298	43	34	10
	Total	481	43	30	14

The decrease in Hamilton Depression Rating Scale (HDRS) scores and completion rates for the 3-week study were similar in the two treatment groups. Significantly, more weight gain, dry mouth, increased appetite, and somnolence were reported with OLZ, while more cases of nausea were reported with DVPX.

A subsequent placebo-controlled trial considered the efficacy of OLZ versus DVPX in responders at 3 weeks who then entered into a 9-week, double-blind extension period (220). OLZ was significantly more efficacious at 12 weeks compared with DVPX but produced greater increases in weight as well as glucose, cholesterol, triglyceride, uric acid, and prolactin levels.

Zajecka et al. (221) also compared the efficacy, safety, and tolerability of DVPX to OLZ in a multicenter, randomized, 12-week, doubleblind, parallel group involving 120 bipolar manic subjects (n = 63, DVPX; n = 57, OLZ; see Table 10-9). The mean maximum dose of DVPX was 2,115 mg/day and of OLZ, 15 mg/day. The mean MRS score changes from baseline to day 21 were -14.8 for VPA and 17.2 for OLZ (p = ns). A significantly greater proportion of OLZ-treated subjects experienced somnolence, weight gain, edema, rhinitis, and slurred speech. One death occurred during the study on OLZ due to diabetic ketoacidosis. Thus, this study found no significant difference in efficacy between groups and reported that DVPX was associated with a more favorable AE profile. The different outcome in the second
trial may be due to

A more aggressive loading dose of DVPX with higher plasma levels overall
Lower mean dose of OLZ
A longer duration of treatment
Smaller sample size

Olanzapine Versus Lithium. Berk et al. (222) compared OLZ to lithium in 30 acute manic patients in a randomized 4-week, double-blind, controlled design (see Table 10-9). There were no significant differences between the two groups on any of the primary outcome measures, including

The BPRS (lithium 28.2; OLZ 28.0; p = 0.44)
The Clinical Global Impression-Improvement (CGI-I; lithium 2.75, OLZ 2.36; p = 0.163)
The MRS (lithium 13.2, OLZ 10.2; p = 0.32)

OLZ, however, was superior to lithium on the Clinical Global Impression-Severity (CGI-S) Scale at week 4 (lithium 2.83; OLZ 2.29; p = 0.03).

A second, double-blind, controlled trial conducted in China included 140 bipolar patients with acute manic or mixed episodes (223). Over 4 weeks, OLZ was superior to lithium based on change scores in the CGI-BP that was the primary outcome measure. A significantly greater number of patients, however, experienced at least one adverse event on OLZ versus lithium.

Olanzapine Versus First-Generation Antipsychotics. Shi et al. (224) conducted a treatment trial for acute mania in 453 subjects (see Table 10-9). This was a 6-week, doubleblind, randomized, controlled study comparing OLZ (5 to 20 mg/day; n = 234) to HPDL (3 to 15 mg/day; n = 219), followed by a 6-week continuation phase. Remission rates were similar for OLZ- and HPDL-treated patients at weeks 6 and 12. At week 6, however, significant changes in five dimensions of the Medical Outcomes Study 36-Item Short Form Health Survey favored OLZ over HPDL. At the end of week 12, OLZ also produced significantly greater improvement than HPDL in favorable Health-Related Quality of Life (HRQOL) changes, work activities, impairment, and household activities and on the Streamlined Longitudinal Interview Clinical Evaluation from the Longitudinal Interval Follow-up Evaluation (SLICE/LIFE) activities impairment scores.

Olanzapine Versus Second-Generation Antipsychotics. A 3-week, randomized, doubleblind study compared OLZ (5-20 mg/day; n = 165) to RISP (1-6 mg/day; n = 164) in bipolar I patients experiencing an acute nonpsychotic, manic, or mixed episode (225). There were no significant differences between treatments at the end of the study based on the YMRS mean change score (i.e., primary outcome measure) or several other mood and functional assessments.

Parenteral Olanzapine Versus Parenteral Placebo or Lorazepam. Meehan et al. (226) randomly assigned 201 agitated, manic patients up to three acute parenteral injections of OLZ (10 mg, first two injections; 5 mg, third injection), lorazepam (2 mg, first two injections; 1 mg, third injection), or placebo (placebo, first two injections; OLZ, 10 mg, third injection) within a 24-hour period (see Table 10-9). Twenty-four hours after the first injection, OLZ was superior to placebo in reducing agitation in patients with acute mania. No significant differences among the groups were observed in safety measures, including treatment-emergent EPS, acute dystonia, or QTc interval changes.

Olanzapine Plus Standard Mood Stabilizers

Tohen et al. (227) conducted an 8-week, doubleblind, randomized, placebo-controlled trial of OLZ augmentation of standard mood stabilizers in 344 nonresponding subjects with a manic or mixed episode (see Table 10-9). OLZ produced a statistically superior improvement compared with placebo augmentation (p = 0.003). Sixtyeight percent of OLZ-treated patients and 45% of placebo-treated patients achieved at least a 50% decrease from their baseline YMRS score. OLZ produced improvements in subjects with and without psychotic features. A post hoc analysis of these data also indicated a rapid improvement in suicidality in a subgroup of 58 mixed episode patients (228). Weight gain and sedation occurred more often with OLZ versus placebo.

A second controlled trial assessed the benefit of OLZ plus DVPX versus placebo plus DVPX in the management of 200 bipolar patients experiencing an acute mixed episode (229). In patients insufficiently benefited after 2 weeks of DVPX monotherapy, the addition of adjunctive OLZ yielded a greater and earlier improvement in both manic and depressed symptoms compared with continuation of DVPX monotherapy.
Significant adverse effects with OLZ versus placebo included an increase in weight and fasting blood glucose.

Risperidone Versus Placebo or Other Psychotropics

Risperidone Versus Placebo. Hirschfeld et al. (230) conducted a multicenter trial comparing the efficacy and tolerability of RISP (mean modal dose = 4.1 mg/day) to placebo for acute mania (see Tables 10-9 and 10-10). Improvements in baseline YMRS scores were significantly greater in the RISP (n = 125) versus the placebo (n = 134) group. Significant differences in change scores were seen as early as 3 days as well as weeks l, 2, and 3. Mean improvements in CGI-S scores were also significantly greater on RISP versus placebo. Somnolence and EPS were greater in the RISP versus placebo group. Further, a 9-week, open-label extension trial with RISP found significant and clinically meaningful improvement in overall functioning based on Global Assessment Scale (GAS) scores (231).

Khanna et al. (232) reported the results of a second multicenter, randomized, double-blind, placebo-controlled trial conducted in India (see Tables 10-9 and 10-10). Subjects (n = 290) experiencing an acute manic or mixed episode received a mean modal dose of 5.6 mg/day of RISP. Of the 290 subjects, 146 were assigned to RISP and 144 to placebo. Although RISP produced significantly greater improvements in baseline YMRS scores versus placebo, EPS occurred more often with this agent.

Smulevich et al. (233) conducted a doubleblind, placebo-controlled trial for acute bipolar mania. Subjects were randomly assigned to RISP (4.2 ± 1.7 mg/day; n = 154), HPDL (8.0 ± 3.6 mg/day; n = 144), or placebo (n = 140). At 3 weeks, change scores on the YMRS were significantly greater for RISP versus placebo (p < 0.001), with no difference between RISP and HPDL. Further improvement occurred during the next 9 weeks with active drug treatments. EPS was more frequent with HPDL versus placebo. The authors also reported that initial benefit with RISP was maintained for an additional 9 weeks.

Only gold members can continue reading. Log In or Register to continue

Stay updated, free articles. Join our Telegram channel

Tags: Principles and Practice of Psychopharmacotherapy

Aug 27, 2016 | Posted by drzezo in PHARMACY | Comments Off

Basicmedical Key

Fastest Basicmedical Insight Engine

Treatment with Mood Stabilizers

Like this:

Related

Stay updated, free articles. Join our Telegram channel

Full access? Get Clinical Tree

Basicmedical Key

Fastest Basicmedical Insight Engine

Treatment with Mood Stabilizers

Share this:

Like this:

Related

Related posts:

Stay updated, free articles. Join our Telegram channel

Full access? Get Clinical Tree