Principles

1. 1. Clozapine has the best evidence for violence reduction in patients with schizophrenia independent of its impact on psychotic symptom severity.

2. 2. High plasma level antipsychotic therapy may be considered in violent schizophrenia patients prior to a clozapine trial in the following circumstances:

   1. a. Failure of routine clinical doses of medium/potent D₂ antagonists

   2. b. And, the absence of dose limiting adverse effects, particularly extrapyramidal side effects (EPS) or akathisia

3. 3. High plasma level antipsychotic therapy can be used for violent schizophrenia patients who refuse or have failed a clozapine trial and who did not exhibit dose limiting adverse effects during routine antipsychotic exposure.

4. 4. Antipsychotic treatment should be carefully titrated until one of two well-defined clinical endpoints are reached: clinical response or intolerability.

   1. a. Treatment trials should not be routinely terminated on the basis of daily dose or high plasma levels: certain patients may both require and tolerate high doses and high plasma levels of D₂ antagonists.

   2. b. A small subset of schizophrenia patients exhibits enormous tolerance for D₂ antagonism without EPS or akathisia. While not an absolute rule, consideration can be given for terminating a medication trial for futility with any of the following plasma levels: haloperidol >> 30 ng/mL, fluphenazine >> 4.0 ng/mL, or olanzapine >> 200 ng/mL.

The first principle was derived from open label experienceReference Frogley, Taylor, Dickens and Picchioni ⁴ and data from a randomized, double-blind, 12-week, parallel-group trial in which clozapine demonstrated superiority to olanzapine and haloperidol for the treatment of violent schizophrenia patients, especially in those with low cognitive function, and without statistical separation between the medication arms for symptom reduction.Reference Krakowski, Czobor, Citrome, Bark and Cooper ^{11 ,} Reference Krakowski, Czobor and Nolan ¹² Olanzapine was also superior to haloperidol for some measures of aggression.Reference Krakowski, Czobor, Citrome, Bark and Cooper ¹¹ The second principle emerged from two clinical realities: (1) certain individuals may fail to evince adverse effects for kinetic or adherence reasons, and thus are deprived of an adequate antipsychotic trial;Reference Glick, Balon, Ballon and Rovine ¹³ (2) there is an enormous range in antipsychotic tolerability.Reference Simpson and Kunz-Bartholini ¹⁴ Principle 3 may be viewed as a corollary of the first two principles, and reiterates the RCP position regarding the appropriate use of high-dose antipsychotics in a treatment algorithm.

The use of plasma antipsychotic levels as noted in Principle 4 is critical to documenting the antipsychotic trial and guiding treatment decisions. While older literature often used the term high-dose treatment as synonymous with a high level of antipsychotic exposure, this is clearly not the case and engenders significant confusion. Most antipsychotics undergo significant phase I metabolism through the CYP system, with intestinal absorption (and central nervous system penetration across the blood–brain barrier) influenced by antipsychotic affinity for the efflux transporter PGP.Reference Meyer ⁸ The concurrent use of potent PGP and CYP 3A4 inducers, such as carbamazepine, phenytoin, rifampin, efavirenz, or nevirapine, lower antipsychotic levels significantly, ranging from 16% for asenapine ¹⁵ and 35–37% for ziprasidone or paliperidone (due to their limited phase I metabolism)Reference Meyer ⁸ to 50% for haloperidol, fluphenazine, olanzapine, and risperidone,Reference Meyer ⁸ and a 70–80% reduction for aripiprazole, quetiapine,Reference Meyer ⁸ and lurasidone. ¹⁶ Valproate in some reports has also been associated with reductions in plasma levels of olanzapineReference Haslemo, Olsen, Lunde and Molden ¹⁷ and clozapine,Reference Longo and Salzman ¹⁸ although the mechanism and extent of the effect on other agents is not well defined. The use of any of these inducers, and possibly valproate, will often result in subtherapeutic plasma antipsychotic levels for most agents and inadequate response unless markedly higher antipsychotic doses are employed, or the patient can be switched from potent inducers such as carbamazepine to those with lower (eg, valproate) or no propensity (eg, lithium) to induce antipsychotic metabolism. It should be anticipated at the outset of treatment that the use of potent inducers with most antipsychotics may result in kinetic failures, underscoring the need to obtain antipsychotic plasma levels. Several reviews provide a good resource for correlations between threshold plasma antipsychotic levels and response for haloperidol,Reference Coryell, Miller and Perry ¹⁹ fluphenazine,Reference Midha, Hubbard, Marder, Marshall and Van Putten ²⁰ and olanzapine,Reference Perry, Lund, Sanger and Beasley ²¹ and for plasma levels of risperidone and olanzapine with central D₂ occupancyReference Uchida, Takeuchi and Graff-Guerrero ^{10 ,} Reference Gefvert, Eriksson and Persson ^{22 ,} Reference Mamo, Kapur and Keshavan ²³ (Table 1). These correlations between antipsychotic plasma levels and D₂ occupancy hold for both oral and the comparable long-acting injectable forms.Reference Gefvert, Eriksson and Persson ^{22 ,} Reference Mamo, Kapur and Keshavan ²³ The abundance of clinical data for haloperidol, fluphenazine, and olanzapine that provides therapeutic ranges makes these antipsychotics ideal candidates for treatment decisions guided by therapeutic drug monitoring. Such correlations have generally not been found consistently with risperidone.Reference Seto, Dumontet and Ensom ²⁴ It should be noted that D₂ occupancy data are from normal volunteers or stable schizophrenia patients, and not from violent forensic populations.

Table 1

Therapeutic plasma antipsychotic ranges for selected agents and plasma levels associated with 80% dopamine D₂ receptor occupancy

*LAI = long acting injectable.

**Active moiety = risperidone + 9-OH risperidone.

A clinically occult source of kinetic failure for some schizophrenia patients prescribed antipsychotics metabolized via CYP 2D6 relates to duplication of the genes encoding CYP 2D6. The net result of these extra copies is greater enzyme production and increased 2D6 clinical activity.Reference Kirchheiner, Schmidt and Tzvetkov ⁹ While relatively uncommon in white northern Europeans (3%), the prevalence of these 2D6 ultrarapid metabolizers (UM) is 10–29% in populations from northeastern Africa (eg, Ethiopia) and geographically proximate parts of the Middle East (eg, Saudi Arabia).Reference Kirchheiner, Schmidt and Tzvetkov ⁹ Although genetic testing exists for CYP polymorphisms, it is not routinely employed due to expense or lack of local availability, so this effect is difficult to anticipate in those who are not clear descendants of individuals from areas with high 2D6 UM prevalence. In the absence of genetic testing, one might infer the presence of UM status in those who are on antipsychotics metabolized via 2D6 in whom adherence issues have been ruled out (eg, by use of long-acting injectable antipsychotics), and plasma levels are significantly below what might be expected for the degree of medication exposure.Reference Hiemke, Baumann and Bergemann ²⁵

At a minimum, the use of plasma antipsychotic levels thus helps avoid treatment failure due to known or unrecognized kinetic issues.Reference Salzman, Glick and Keshavan ²⁶ As will be illustrated in Case 1, Principle 4a enunciates the proposition that some patients may require high plasma antipsychotic levels for response —— levels that might be outside of the reported therapeutic reference range. When no clinical evidence exists for adverse effects, this fact must documented before further dose titration is pursued if the clinical endpoint (violence reduction) has not been met. Aside from clozapine, few antipsychotics have robust data on response and tolerability, including haloperidol, fluphenazine, and olanzapine (Table 1).Reference Kelly, Richardson, Yu and Conley ^{27 ,} Reference Meyer ²⁸ While tolerability may decrease significantly for plasma fluphenazine levels > 2.0 ng/mLReference Midha, Hubbard, Marder, Marshall and Van Putten ²⁰ (Figure 1), haloperidol levels > 12 ng/mLReference Van Putten, Marder, Mintz and Poland ²⁹ (Figure 2), or olanzapine levels > 100 ng/mL,Reference Kelly, Richardson, Yu and Conley ²⁷ the absence of EPS/akathisia or other significant adverse effects despite higher plasma levels is an indication that the patient might reasonably be tried at a higher dosage, especially if clozapine is not an option or the patient is a clozapine failure.

Figure 1

Relationship between plasma fluphenazine levels and estimated probability of improvement (gray line) and disabling side-effects (black line).Reference Midha, Hubbard, Marder, Marshall and Van Putten ²⁰

Figure 2

The relationship between plasma haloperidol and the estimated probability of disabling side effects.Reference Midha, Hubbard, Marder, Marshall and Van Putten ²⁰

It should be acknowledged that there are rare circumstances in which schizophrenia patients may exhibit no threshold for neurological adverse effects. The data in Table 2 are from a trifluoperazine titration study in which 10 male patients with schizophrenia were started on trifluoperazine 15 mg/d (equivalent to 6 mg haloperidol) without antiparkinsonian treatment, and advanced by 10 mg per week until they reached a predefined level of neurological adverse effects.Reference Simpson and Kunz-Bartholini ¹⁴ Subject 5 remained in the titration arm for over 36 weeks, achieving a daily trifluoperazine dose of 480 mg, yet had the lowest neurological side effect rating among the subjects in this study despite obvious evidence of significant drug exposure (weight gain). In such individuals, an antipsychotic trial may be terminated for futility (Principle 4b), although no definite guidelines exist. Clinical experience within the DSH system has yielded very few responders at fluphenazine levels significantly greater than 4.0 ng/mL, haloperidol levels above 30 ng/mL, and olanzapine levels above 200 ng/mL.

Table 2

Outcomes in an escalating trifluoperazine dose protocolReference Simpson and Kunz-Bartholini ¹⁴

*Interpretation of psychiatric symptom change: − = no improvement with later worsening; 0 = no improvement; ± = mild improvement with later worsening;+ = mild improvement; ++ = moderate improvement; +++ = marked improvement.

The principles listed above do not cover details such as the rapidity of titration or duration for any dosage, as these parameters are often dictated by clinical acuity. Recent analyses of industry-sponsored schizophrenia trials do indicate that, with respect to symptom reduction, the absence of any treatment effect after 2 weeks predicts nonresponse at week 6, and should prompt reconsideration of the treatment approach.Reference Stauffer, Case and Kollack-Walker ³⁰ In many instances, the response to the lack of improvement will include documentation of plasma levels and consideration of dose increases when tolerability is not yet an issue. In patients who have failed high doses of high-potency first-generation antipsychotics with supporting plasma level data, olanzapine is an option for those who cannot or will not take clozapine, bearing in mind its inferiority to clozapine.Reference Krakowski, Czobor, Citrome, Bark and Cooper ¹¹

Case 1: A High-Dose Responder: Plasma Level Guided High-Dose Antipsychotic Therapy

The patient is a 50-year-old African American female with a diagnosis of schizophrenia, and multiple medical problems including type 2 diabetes mellitus (DM), who was admitted to a forensic psychiatric hospital following a conviction for manslaughter. Due to ongoing assaultiveness and suicidality, she was started on clozapine with significant clinical response, but also the development of poorly controlled DM, necessitating the withdrawal of clozapine therapy. Trials of other atypical antipsychotics at high dosages proved ineffective in managing her psychotic symptoms and violence, so the decision was made to embark on a course of high-dose haloperidol therapy, given her prior tolerance for lower dosages. Over several months, the patient was titrated to a daily dose of 45 mg with a marked reduction in psychotic symptoms, requirements for PRN medications, and acts of aggression toward herself and others. Neither EPS nor akathisia were evident, and she was not maintained on routine antiparkinsonian medication. A new psychiatrist assumed the patient's care and obtained a plasma haloperidol level of 29.5 ng/mL. Alarmed that this greatly exceeded the laboratory reference range (5–12 ng/mL), the haloperidol dose was decreased to 20 mg/d (plasma level 8.1 ng/mL). Over the next month, the patient deteriorated, started to refuse routine medication, and required numerous PRN medications for stability. The haloperidol dose was increased to 30 mg for several months, but the patient was frequently assaultive despite this dose increase and the addition of adjunctive lithium and lamotrigine. Another psychiatrist agreed to assume the patient's care, but the patient continued to intermittently refuse oral haloperidol, primarily for delusional reasons. The patient was convinced to try fluphenazine in lieu of haloperidol, and titrated to 20 mg/d with a trough plasma level of 0.7 ng/mL, but limited clinical improvement. To increase the plasma level and also to limit any confounding issues related to adherence, fluphenazine decanoate was started at 25 mg every 2 weeks, oral fluphenazine was tapered off, and the depot dose increased over ensuing months to 100 mg every 2 weeks. With moderate improvement on this regimen and the absence of EPS and akathisia, the psychiatrist chose to supplement fluphenazine decanoate with oral fluphenazine 15 mg with marked reduction in psychosis and complete resolution of aggressive acts. The plasma fluphenazine level on this combination ranged from 3.0–3.2 ng/mL.

Case 2: High-Dose Failure: The Value of Plasma Levels

The patient is a 29-year-old Latino with a diagnosis of schizophrenia, who was admitted to a forensic psychiatric hospital after being found incompetent to stand trial on numerous charges including assault with a deadly weapon, criminal threats, and arson of an inhabited structure. The court commitment included an involuntary medication order in the case of treatment refusal. The patient was admitted to the hospital from custody on a combination of oral olanzapine 40 mg at bedtime and divalproex 1000 mg twice per day, but extent of adherence was unknown, and he was grossly psychotic in the admission suite, requiring intramuscular medications to control agitation. The medication combination started in jail was continued with good adherence, as documented by trough plasma olanzapine levels of 73 ng/mL and 79 ng/mL and therapeutic serum valproate levels, but on hospital day 12, the patient engaged in an unprovoked assault on a peer with serious injury, and required seclusion and restraints.

Given the severity of the assault, the ongoing significant level of psychotic symptoms, and records from a prior hospitalization that this patient tolerated oral haloperidol doses of 20 mg/d or more, a decision was made to load the patient on haloperidol decanoate while continuing olanzapine and divalproex at prior dosages. The most aggressive depot haloperidol load permissible in DSH facilities is 300 mg weekly for three doses; this regimen is designed to mimic plasma levels achievable with 30 mg/d of oral haloperidol.Reference Meyer ²⁸ This dosing equivalency is an extrapolation of prior loading studies demonstrating that, after 3 weekly injections of haloperidol decanoate 100 mg, mean plasma haloperidol concentrations from the depot were comparable to 10 mg/d of oral haloperidol (7.95 ± 4.94 ng/mL vs 7.79 ± 4.79 ng/mL).Reference Wei, Jann, Lin, Piao-Chien and Chang ³¹ The 3 loading injections of 300 mg were administered in weekly intervals without evidence of adverse effects. Of note, a plasma haloperidol level obtained 1 week after the last loading injection (ie, at T_Max) was 6.4 ng/mL. An assault occurred two days after the third loading injection, but the patient was subsequently free from assaultive behavior and continued to show no evidence of EPS or akathisia. However, his psychotic symptoms appeared to worsen slightly in the week prior to the scheduled maintenance depot dose, so haloperidol decanoate 300 mg was administered 22 days after the last loading injection. A haloperidol level obtained 4 days later was 2.8 ng/mL. The following week, this patient was again involved in a serious unprovoked assault that necessitated seclusion and restraints. A psychopharmacology consultation was obtained to assist with management. Based on the unexpectedly low plasma haloperidol levels obtained despite aggressive loading, it was concluded that this patient's inadequate response to haloperidol was a kinetic failure, but one that might not be easily overcome without the administration of extremely high weekly or biweekly injections of haloperidol decanoate—doses that would be unlikely to be continued outside of the forensic setting due to unfamiliarity with the rationale for a high-dose regimen in this patient. As this patient manifested expected plasma levels from oral olanzapine, a high-dose strategy for this medication was recommended, noting that doses of 60–80 mg/d might be required to achieve plasma levels in the 100–150 ng/ml range, levels associated with 80–85% D₂ occupancy.

Disclosures

Jonathan M. Meyer discloses the following: BMS, speaker, speaker's fee; Genetech, speaker, speaker's fee; Genetech, advisor, honoraria; Otsuka, speaker, speaker's fee.