Antipsychotic Polypharmacy

Ms. X is a 37-year-old Hispanic woman with a history of schizophrenia with significant thought and behavior disorganization and multiple past inpatient admissions for bizarre behavior (e.g., trying to run through a store window, walking into traffic) motivated by command-type auditory hallucinations. Ms. X is not using birth control, has no known medical conditions, does not use drugs, and drinks alcohol only occasionally. When evaluated by a psychiatrist for a second opinion, she is pleasant and appears cooperative, but she is unable to offer a coherent history because of her loose thought process and bizarre, poorly formed delusions. Family reports that although Ms. X is able to take a shower, dress herself, and do some household chores, she has to be redirected to follow a daily routine. She was partially stabilized on 500 mg/day quetiapine and 50 mg/day sertraline without any notable weight gain or obvious metabolic adverse effect. However, she reports severe constipation from the current medication. She has not had an inpatient admission in the past six months. However, she remains disorganized and incoherent at times, with odd delusions. She cannot drive and is unable to conduct any social engagements without the assistance of her parents.

The consulting psychiatrist suggests clozapine and explains the agranulocytosis monitoring requirements and other possible side effects: seizures, orthostatic hypotension, cardiomyopathy, increased risk of cardiovascular events, and—highly relevant to this case—constipation. The patient and family decide not to pursue clozapine because of the perceived burden of blood draw. They are also reluctant to switch antipsychotics because they perceive that the quetiapine successfully reduced Ms. X’s admissions. The patient and her family choose to add paliperidone to quetiapine to address her thought disorder and hallucinations. The patient also looks forward to eventually taking a lower dose of quetiapine, which may limit its only perceived side effect of chronic constipation.

This case potentially illustrates Kapur et al.’s (21) “kiss and run” hypothesis of quetiapine’s psychopharmacological action. Quetiapine is thought to have a low, apparently subtherapeutic, D₂ receptor occupancy 12 hours after administration. However, at four to six hours postadministration, the D₂ receptor occupancy measured with positron emission tomography is high, leading to the hypothesis that quetiapine dissociates from receptors faster than other antipsychotics (21). Adding a low-dose antipsychotic with consistently high D₂ occupancy may be of utility in this case by addressing two domains of psychosis not fully addressed by the initial drug, namely the delusions and disorganized speech and behavior that continue to limit the patient’s function. In addition, paliperidone has minimal affinity for the muscarinic receptors, whose blockade is implicated in the side effects of constipation from antipsychotic medication (22), thus maximizing treatment of Ms. X’s symptoms without exacerbation of the adverse effects she is currently experiencing. Table 2 (19, 22–24) shows the receptor affinity of commonly used antipsychotics. The greater the number of plus signs, the stronger the affinity at that receptor. Antipsychotic combinations may also aim to optimize nondopaminergic (e.g., serotonergic, glutamatergic, and adrenergic) receptor occupancy to alleviate positive and negative symptoms (13). Adding a second antipsychotic can also address the affective domain of psychosis (depression or mania), insomnia, or aggression toward self or others.

Current evidence shows that antipsychotic therapeutic action is more likely to occur after at least 60% occupancy of D₂ receptors is achieved; however, at 80% occupancy, the movement-associated adverse effects of antipsychotics are thought to begin (25). When using multiple antipsychotics, the likelihood of surpassing the therapeutic window is higher and motor-adverse effects may occur, including extrapyramidal symptoms, tardive dyskinesia, and akathisia. Motor side effects of antipsychotic treatment are largely attributed to dopamine blockade, with contributions from the antipsychotic effect of other neurotransmitters, for example, D₂ blockade in the nigrostriatal pathway leading to Parkinsonism; dopamine blockade accompanied by acetylcholine, gamma‑aminobutyric acid, norepinephrine, serotonin, and neuropeptides in akathisia; and dopamine blockade and an increase in acute cholinergic activity in acute dystonia (26–28). Because the motor side effects of antipsychotics are rarely quantified in practice, they cannot be easily identified in the electronic medical record. In the absence of prospective studies addressing the motor side effects of antipsychotic combinations versus monotherapy, researchers have studied treatment with anticholinergic drugs prescribed concomitantly with antipsychotic combinations to generate data on the frequency of motor side effects. Generally, combinations that included first- and second-generation or two second-generation antipsychotics led to increased anticholinergic use, possibly reflecting increased motor side effects (28). It can be argued that anticholinergic medications may be prescribed prophylactically and therefore do not reflect the presence of motor side effects. However, in either case, the addition of another medication with its own side effect profile can be considered to be a negative outcome for the patient (26). Carnahan et al. (26) noted that the low propensity of extrapyramidal side effects found with second-generation antipsychotic monotherapy does not persist when second-generation antipsychotic drugs are combined. Neuroleptic malignant syndrome, which is thought to occur as a result of a marked and sudden reduction in central dopaminergic activity resulting from D₂ receptor blockade within the nigrostriatal, hypothalamic, mesolimbic, and mesocortical pathways, has been found to be associated with antipsychotic combinations, although this association has primarily been noted in case reports (23). Currently, tardive dyskinesia and akathisia have not been clearly associated with combination antipsychotics (28).

Dopamine blockade also can lead to a lack of inhibition of prolactin release. This in turn leads to hyperprolactinemia and in some cases galactorrhea, amenorrhea, sexual dysfunction, osteoporosis, and even breast cancer (24). It appears that adding an antipsychotic drug with high D₂ blockage potential to a drug with low D₂ propensity increases the risk of hyperprolactinemia (25). The apparent exception is the addition of aripiprazole, which has been shown to decrease prolactin levels and is a possible treatment for antipsychotic-induced hyperprolactinemia (24, 29).

Mr. Y is a 56-year-old nonsmoking man with schizoaffective disorder who has had multiple depressed psychotic episodes, each leading to two- to three-week-long hospitalizations for paranoia and hostile behavior. He was discharged from his most recent hospitalization on 30 mg/day olanzapine and 40 mg/day fluoxetine. Despite good medication adherence, he was still depressed and at times had passive suicidal thoughts, paranoia, and auditory hallucinations with derogatory content. Eventually, he was switched from fluoxetine to amitriptyline, and his depressive symptoms resolved. He continued to have auditory hallucinations that interfered with his ability to perform activities of daily living and attend his peer-support program. He repeatedly refused clozapine monotherapy in inpatient and outpatient settings because of concerns about blood draw requirements and frequent trips to the pharmacy, even though his case manager offered to arrange home medication delivery. Eventually, Mr. Y’s outpatient psychiatrist offered to add fluphenazine to his regimen to address the hallucinations after a thorough discussion of possible motor side effects. The patient agreed to take fluphenazine and had remission of the hallucinations with 5 mg/day, while his fasting lipids, serum glucose, waist circumference, weight, and onset of involuntary movements were monitored. The patient became able to work 16 hours/week and consistently kept all his clinic appointments. However, he developed involuntary movements of his arms and trunk that, although manageable, did not resolve with further treatment.

Patients in both case examples 1 and 2 would have been, according to existing guidelines, good candidates for clozapine. Among eligible patients, clozapine is infrequently prescribed (between 15% in the United States and 54% in the United Kingdom) despite guidelines recommending it after nonresponse to two other antipsychotics (see Figure 1) (30). Reluctance to use clozapine perpetuates the low frequency of clozapine prescriptions and thus prescribers’ and trainees’ limited exposure to patients who remit or recover on clozapine (31). Moreover, the mandatory monitoring of white blood and absolute neutrophil counts compounds prescribers’ reluctance to initiate the drug. This reluctance is further increased by the presence of other potentially serious adverse effects: orthostatic hypotension, bradycardia, syncope, seizure, myocarditis and cardiomyopathy, and increased mortality in elderly patients with dementia-related psychosis. Recent evidence has also suggested that clozapine may cause increased mortality risk among patients with pneumonia (32). Despite these concerns, the effectiveness of clozapine among patients with schizophrenia who are treatment resistant or who experience suicidal ideation remains superior (33, 34). In their 2019 meta-analysis on clozapine and long-term mortality risk, Vermuelen et al. (33) also found that continuous clozapine use has the lowest all-cause mortality risk of any antipsychotic.

Clozapine is a highly effective antipsychotic used for patients with treatment-resistant schizophrenia. However, 40%−60% of patients who meet this criterion (13) will have an incomplete response (35) to clozapine monotherapy. Clozapine in combination with a second antipsychotic currently has the best, although limited, evidence in terms of antipsychotic polypharmacy. Clozapine is commonly used in combination with amisulpride, haloperidol, and sulpiride and is increasingly used with aripiprazole as the augmenting antipsychotic to maximize D₂ receptor blockade.

The combination of clozapine and risperidone is possibly the most well studied of the clozapine combination strategies. RCTs comparing clozapine and risperidone with clozapine monotherapy are equivocal (35, 36). Some studies have indicated that combination therapy is more efficacious than clozapine monotherapy for positive and negative symptoms and disorganized thoughts, some have indicated that it is less effective, and some have shown no difference. These mixed findings may indicate that specific populations benefit from the combination therapy, but the criteria to identify those populations have not yet been found.

Sulpiride, amisulpride, and levosulpiride are atypical antipsychotics in the benzamide class that are not approved in the United States, Canada, or Australia but are used in other countries. Clozapine with sulpiride has been reported to improve positive and negative symptoms (37). Sulpiride and risperidone have been found to be equally efficacious in combination with clozapine. In their 2017 review of clozapine combinations for the treatment of schizophrenia, Barber et al. (37) found that amisulpride combined with clozapine is reported to be more efficacious than quetiapine combined with clozapine and noted improvement in the global assessment of function, clinical global impression, and depression, but not psychosis.

The combination of clozapine and aripiprazole shows promise as a way to ameliorate some of the side effects of clozapine. Jeon and Kim (36), in their evaluation of the concerns with antipsychotic polypharmacy and metabolic syndrome (i.e., the state of hyperlipidemia, obesity, hypertension, and glucose intolerance), found two RCTs in which body weight and thus body mass index improved and cholesterol levels decreased, although psychosis did not improve. Jeon and Kim (36) also found open-label trials in which patients showed improved metabolic markers and decreased somnolence. Aripiprazole and clozapine combinations have also been shown to have fewer side effects than clozapine and haloperidol combinations but are similarly efficacious (37). Tiihonen et al. (13) explored the association of antipsychotic polypharmacy versus monotherapy with risk of rehospitalization among persons with schizophrenia and found that the lowest risk was observed with the combination of clozapine and aripiprazole (7%−14% lower than any antipsychotic monotherapy). They also found that patients treated with clozapine and aripiprazole had a better outcome in terms of both psychiatric hospital readmission and all-cause hospitalization than those treated with any other monotherapy or combination of antipsychotics.

A 23-year-old man with schizophrenia, cocaine use disorder, and medication nonadherence was admitted for crisis stabilization five times in 12 months for active psychosis, a suicide attempt by an overdose of ibuprofen, and threats to harm others. At the last admission, he was grossly psychotic and threatened to harm his case manager. The patient did not improve after one week of inpatient care on 30 mg/day olanzapine. The inpatient unit team applied for the patient’s transfer to long-term care, and it took two additional weeks for a bed to be secured for him. His new inpatient treatment team, after reviewing the records and history obtained from the patient and his family, proposed starting treatment with a long-acting injectable antipsychotic, potentially olanzapine, because the patient had eventually showed improvement with an oral formulation. However, the patient adamantly refused to take a monthly injection, stating that he did not like needles. The patient also declined clozapine given the need for a weekly blood draw. After approximately four weeks of hospitalization, the patient became calmer and less paranoid, started attending groups, and frequently played basketball in the recreation area. However, he maintained active hallucinations with derogatory content that bothered him, particularly at night. The patient had a family history of diabetes and gained seven pounds on olanzapine. Given his partial response to high-dose olanzapine, recent weight gain, and family history, the psychiatrist recommended adding 5 mg/day aripiprazole, with the intent of addressing the remaining hallucinations and delusions and diminishing the metabolic side effects. After another two weeks of inpatient care, the patient stabilized on olanzapine (decreased to 20 mg/day) and 5 mg/day aripiprazole and was discharged in the care of an assertive community treatment (ACT) team. Subsequently, the patient attended Narcotics Anonymous and the local psychosocial clubhouse program. He played basketball with friend three times a week and lost five pounds. He inquired about scholarship opportunities at the local community college. Given the patient’s marked improvement, the ACT team psychiatrist and patient alike were reluctant to change the patient's medication, although it involved a combination of antipsychotics.

As reflected in this case example, if weight gain or metabolic disturbance is present but psychotic symptoms persist, clinicians may choose to add another antipsychotic with higher D₂ receptor affinity and address the persisting symptoms while lowering the dose of the antipsychotic with higher liability of metabolic syndrome. Long-term use of antipsychotics among patients with schizophrenia increases the risk for metabolic syndrome (2, 38). The relationship between antipsychotic drug combinations and metabolic syndrome is complex and likely dependent on multifactorial issues such as multiple receptors, failed glucose homeostasis, and lifestyle factors. The mechanisms for the metabolic side effects of atypical antipsychotics are multiple. Coccurello and Moles (39) found that antipsychotics with the greatest weight gain liability share a high affinity for serotonin (5-HT_2A, 5-HT_2C, 5-HT₆, and 5-HT₇), muscarinic (M₁, M₂, M₃, and M₅), histamine (H₁), adrenergic (alpha₁ and alpha₂ but also beta₃), and dopamine (D₁ and D₂-like) receptors. Ijaz et al. (2) found insufficient data for the potential harm of antipsychotic polypharmacy, including metabolic syndrome, noting a potential protective effect of antipsychotic combinations that included aripiprazole for dyslipidemia and glucose metabolism compared with other combinations and monotherapy. The finding of a protective role of aripiprazole calls for further investigation. Data are emerging about the effectiveness of aripiprazole added to decrease the metabolic burden of clozapine (34).

The treatment team in case example 3 also had the option to completely cross-titrate a second antipsychotic (in this case, aripiprazole). However, improvement and stabilization while cross-titrating two antipsychotic drugs may result in the patient being kept on both antipsychotics to prevent illness relapse, which could lead to long-term polypharmacy. Although each of the patients in the case examples presented could ideally benefit from taking only one antipsychotic, the clinical and practical considerations in each case make antipsychotic polypharmacy necessary to stabilize the patients and allow them to progress toward recovery.

Among patients with residual symptoms despite a partial response to a single antipsychotic, distress could be caused by symptoms that belong to other domains of psychosis, requiring treatment with an alternative psychotropic agent. Here, we briefly discuss the augmentation of antipsychotics with various other classes of medications.

In their systematic review and meta-analysis, Helfer et al. (40) noted that depressive symptoms can occur in all phases of schizophrenia—prodromal, acute, and postpsychotic—and lead to poor outcome, increased relapse, and a higher suicide rate. The prevalence of depression in schizophrenia has been reported to be around 40% (41). On their own, depressive symptoms represent a domain of schizophrenia that can be difficult to disentangle clinically from the negative symptoms with which they overlap significantly in symptomatology, genetics, neurobiology, and neurocognition. Thus, among patients with schizophrenia, treating depressive symptoms is imperative. Helfer et al.’s systematic review of 82 trials suggests that antidepressants added to antipsychotics are efficacious in treating negative and depressive symptoms among persons with schizophrenia, albeit with small effect sizes (40). They also found that antidepressants were safe, with no difference between persons receiving antidepressant-antipsychotic combinations and respective control groups in dropout rates, exacerbation of psychosis, or adverse effects. Helfer et al. also found that selective serotonin reuptake inhibitors as a class—and fluvoxamine and citalopram in particular—showed a particularly beneficial effect for depressive and negative symptoms, and mirtazapine has promise particularly in the treatment of negative symptoms (40). An investigation of the use of other adjunctive psychotropics versus antipsychotic polypharmacy found that adjunctive antidepressant use reduced emergency department visits and hospitalization, even among patients with comorbid substance use disorder (42). In addition, the use of other adjunctive psychotropics versus antipsychotic polypharmacy led to a lower risk of diabetes mellitus and reduced mortality (42). Tiihonen et al. (43), in their examination of psychotropic use among persons with schizophrenia, also found that the concurrent use of antidepressants and antipsychotics was associated with decreased all-cause mortality and decreased suicide deaths when compared with concurrent use of benzodiazepines and antipsychotics.

A meta-analysis found that divalproex augmentation of antipsychotics among persons with schizophrenia and schizoaffective disorder significantly decreased psychiatric symptoms compared with treatment with antipsychotic monotherapy (44). However, in a subgroup analysis, the significant benefit on psychiatric symptoms was found to persist only in open trials, not in RCTs, and only in studies that lasted four weeks or less (44). Lithium augmentation of antipsychotic drugs has been evaluated in relatively small studies, yielding low-quality evidence. Although patients who received lithium augmentation had significantly better clinical response than patients treated only with antipsychotics, this benefit was not maintained after participants with schizoaffective disorder were excluded or when studies that were not double-blinded were excluded (45, 46). Stroup et al. (42) reported that there was no evidence for the efficacy of mood stabilizers in the treatment of schizophrenia. Moreover, they found that gabapentin was associated with high mortality and noted that it is associated with dizziness, somnolence, increased suicidal behavior, and suicide. As such, they recommended caution in its use. No other mood stabilizer used with this sample (e.g., lithium, divalproex, carbamazepine, lamotrigine) seemed to be associated with higher mortality.

Reviews by Tiihonen et al. (43) and Stroup et al. (42) found that combinations of antipsychotics and benzodiazepines were associated with higher mortality, postulated to be caused by the use of benzodiazepines at a higher than the daily defined dose. This, in turn, may cause benzodiazepine withdrawal, which could lead to increased mortality. They also postulated that discontinuation of long-term benzodiazepine use may lead to increased anxiety and suicidal behavior. Stroup et al. (42) found that, in addition to their association with the higher rate of mortality in the United States, benzodiazepines were associated with a higher number of emergency department visits and hospitalizations. Stroup et al. (42) found no benefit in the use of adjunct benzodiazepines when compared with the use of adjunctive antipsychotics, antidepressants, or mood stabilizers.

Electroconvulsive therapy (ECT) is one of the oldest and most effective treatments in psychiatry. ECT was originally delivered to patients with schizophrenia and was the most popular treatment for acute psychosis until the introduction of chlorpromazine in 1952 (47). The use of ECT in the Western world later shifted to patients with unipolar and bipolar disorder, although it is still frequently used to treat patients schizophrenia in developing countries (48). ECT can be an effective treatment for acute phases of schizophrenia in patients with catatonia and those with marked positive symptoms or affective symptoms (47, 49). In their recent review of ECT and schizophrenia, Sanghani et al. (47) found growing evidence for the use of ECT as an augmentation strategy for antipsychotics, including evidence that the combination of clozapine and ECT is particularly effective among patients with treatment-resistant schizophrenia or those who do not respond to clozapine (48–50).

A less invasive option is the use of transcranial magnetic stimulation (TMS). A recent meta-analysis supported an effect of low-frequency repetitive TMS (rTMS) applied over the left temporo-parietal region on refractory auditory hallucinations and positive symptoms (51). There was no evidence supporting the use of other rTMS modalities, such as prefrontal or right-side temporo-parietal rTMS, in treatment for auditory hallucinations and positive symptoms. Treatment of negative and cognitive symptoms with this modality did not provide consistent results (51, 52).

Each of the current schizophrenia treatment guidelines mentions the importance of psychosocial intervention in treatment response, particularly providing the patient with enough support beyond direct medication management, as illustrated in Figure 1. Interventions such as ACT, cognitive-behavioral therapy, cognitive remediation, family psychoeducation, illness self-management training, social skills training, supported employment, and art therapy have been shown to significantly assist in improving outcomes for those with severe mental illness. Chien et al. (53) found evidence for the efficacy of psychosocial intervention in relieving psychotic symptoms and improving function. However, studies and clinical implementation have been hindered by limited access to and inconsistent applications of psychosocial interventions.

Given the risk of pharmacokinetic and pharmacodynamic drug interactions with polypharmacy, it would be beneficial to have an understanding of how individual patients metabolize drugs before combination therapy is pursued. This becomes increasingly relevant as patients age and may be prescribed nonpsychotropic medication that affects the liver enzymes involved in drug metabolism and overall treatment response. Personalized medicine appeared to hold significant promise in psychiatry in the early 2000s. Research has been translated into practice primarily by understanding cytochrome (CYP) P450 metabolism and thus the side effects of psychiatric drugs, whereas research on candidate genes with potential in predicting drug response has not yet yielded clinical outcomes. CYP450 genes code for liver enzymes contributing to phase 1 of the metabolism of antipsychotics, among many psychiatric drugs (54). Recently, Arranz et al. (55) demonstrated that in antipsychotic treatment among patients who receive pharmacogenetic testing and demonstrate CYP functional variants, careful dosing adjustments lead to a significant improvement in global (p=0.02), psychic (p=0.05), and other side effects (p=0.01) measured with a validated scale. CYP2D6 metabolizes chlorpromazine, haloperidol, perphenazine, and thioridazine, as well as risperidone, paliperidone, iloperidone, and aripiprazole. CYP1A2 metabolizes chlorpromazine, perphenazine, thioridazine, thiothixene, loxapine, and trifluoperazine, as well as clozapine and olanzapine. CYP3A4 metabolizes haloperidol, pimozide, lurasidone, aripiprazole, ziprasidone, and quetiapine. These enzymes are important in the metabolism of antipsychotic drugs, and it is thus important to consider the possibility of drug interactions when two or more antipsychotics are combined (56).

These enzymes are highly polymorphic, and their polymorphisms are thought to lead to distinct phenotypes, which can in turn influence drug response, particularly adverse effects, although factors such as age, ethnicity, drug interactions, active drug metabolites, and medical comorbidities can also influence testing results (57). Combining medications that are metabolized by the same CYP450 enzyme may be problematic, particularly when the enzyme is nonfunctional. Moreover, the Food and Drug Administration­–approved prescription information for aripiprazole and risperidone contains guidance on adjusting dosing on the basis of CYP2D6 metabolism. Finally, the literature has recommended “CYP2D6 and CYP2C19 genotyping, preferably combined with therapeutic drug monitoring, for the four second-generation antipsychotics that are dependent on CYP2D6 for their metabolism (aripiprazole, brexpiprazole, iloperidone, and risperidone), clozapine, and atomoxetine” (58–60). Clinically, listening carefully to patients’ report of side effects occurring from low doses of an antipsychotic or, conversely, report of individual ineffective trials of large-dose antipsychotics, can offer clues to a possible CYP450 polymorphism (57).