The second-generation antipsychotics (SGAs) are united by a lower propensity for extrapyramidal side effects (EPS) and tardive dyskinesia (TD) at clinically effective doses, and mechanistically by a higher ratio of serotonin 5-HT2A to dopamine (DA) D2 receptor blockade relative to first-generation antipsychotics (FGAs). Within this broad definition, the agents are pharmacologically diverse.
The mechanistic rationale for the atypical designation was formalized in the early 1990s by Meltzer and colleagues, who proposed that the serotonin 5-HT2A to D2 affinity ratio discriminated atypical from typical antipsychotics more reliably than any single receptor binding property.1 The logic was as follows: 5-HT2A receptors on dopaminergic neurons in the nigrostriatal and mesocortical pathways tonically inhibit DA release. Blocking these receptors disinhibits DA release in the striatum, which partially offsets the motor consequences of D2 blockade and raises the effective EPS threshold. Simultaneously, disinhibition of DA release in the prefrontal cortex (PFC) partially counteracts the mesocortical DA deficiency that underlies negative symptoms and cognitive dysfunction. An agent with a high 5-HT2A to D2 affinity ratio can therefore achieve mesolimbic D2 occupancy in the therapeutic range while producing less nigrostriatal blockade and potentially some mesocortical DA restoration relative to a purely selective D2 antagonist.
This framework, while heuristically useful, has important limitations. The serotonin 5-HT2A to D2 ratio does not fully predict EPS risk across agents, because receptor dissociation kinetics also matter: agents that bind D2 with fast-off kinetics (quetiapine, clozapine) achieve high momentary occupancy during plasma concentration peaks but rapidly dissociate, allowing endogenous DA to compete and reducing sustained EPS-producing blockade. This "fast-off" theory, developed by Seeman, offers a complementary explanation for atypicality in agents such as quetiapine and clozapine whose 5-HT2A to D2 ratios do not fully account for their very low EPS burden.2 In practice, EPS risk in the SGA class is best predicted by a combination of D2 affinity, the 5-HT2A to D2 ratio, and receptor dissociation rate.
The SGAs also carry adverse effects largely absent from FGAs, most prominently weight gain, glucose dysregulation, and dyslipidemia constituting a metabolic syndrome profile. These metabolic effects are concentrated in the agents with the highest H1 and 5-HT2C receptor affinity (clozapine and olanzapine) and represent a substantial long-term cardiovascular and metabolic health burden. The narrative of SGAs as categorically superior to FGAs was substantially qualified by the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study, which showed that olanzapine's modest efficacy advantage over other agents came with the highest metabolic adverse effect burden, and that perphenazine performed comparably to quetiapine, risperidone, and ziprasidone on all-cause discontinuation.3 The appropriate clinical conclusion is not that SGAs are inferior to FGAs, but that agent selection should be individualized based on the patient's specific symptom targets, comorbidities, and risk profile rather than on class membership.
Clozapine occupies a singular position in antipsychotic pharmacology. It was the first agent recognized as atypical on clinical grounds, predating the receptor pharmacological framework used to define the class. Its D2 receptor affinity is relatively modest compared with both FGAs and most other SGAs, yet it is the only antipsychotic with demonstrated superior efficacy in treatment-resistant schizophrenia (TRS), defined as failure of two adequate trials of antipsychotics at adequate doses.4 This efficacy cannot be explained by D2 blockade alone and remains incompletely understood, though its uniquely broad multi-receptor profile, including substantial D4, D1, 5-HT2A, 5-HT2C, H1, M1, and alpha-1 adrenergic blockade, is likely contributory. Its fast-off kinetics at D2 are also relevant: clozapine maintains very low sustained D2 occupancy (approximately 40 to 60% at therapeutic doses) while achieving antipsychotic effect, a fact that challenged the prevailing understanding of D2 occupancy thresholds and suggested clozapine's mechanism involves pathways beyond simple D2 blockade.2
Clozapine's evidence base in TRS is anchored by the Kane et al. 1988 trial, which demonstrated a 30% response rate in patients who had failed adequate first-generation antipsychotic (FGA) trials, compared with 4% on chlorpromazine.4 Subsequent studies have confirmed not only superior antipsychotic efficacy in TRS but also a unique anti-suicidal effect: clozapine is the only antipsychotic with a Food and Drug Administration (FDA) indication specifically for reducing suicidal behavior in schizophrenia and schizoaffective disorder, supported by the International Suicide Prevention Trial (InterSePT).5 The mechanism of this anti-suicidal effect is not fully established but may involve serotonergic modulation of impulsivity and hopelessness independent of its antipsychotic effects. Current guidelines recommend that clozapine be initiated after failure of two antipsychotic trials, and that delays beyond this point represent a missed opportunity to address both positive symptoms and suicide risk in the most vulnerable patients.
Clozapine causes agranulocytosis, defined as an absolute neutrophil count (ANC) below 500 cells per microliter, in approximately 0.8 to 1% of patients, with peak risk in the first 3 to 6 months of treatment and a long tail of risk throughout the treatment course.6 Agranulocytosis is idiosyncratic, not dose-dependent, and potentially fatal if not detected early. In the United States, clozapine is available only through the Clozapine Risk Evaluation and Mitigation Strategy (REMS) program, which mandates ANC monitoring before dispensing: weekly for the first 6 months, biweekly for months 6 to 12, and monthly thereafter for patients who remain stable. If ANC falls to 1000 to 1499 cells per microliter (mild neutropenia), monitoring frequency increases and dose interruption is considered. If ANC falls below 1000 cells per microliter (moderate to severe neutropenia), clozapine must be interrupted or discontinued depending on severity, and the patient is placed on a rechallenge restriction list.6
The mechanism of agranulocytosis involves toxic clozapine metabolites (primarily clozapine N-oxide and norclozapine) that damage neutrophil precursors in the bone marrow; an immune-mediated component involving anti-neutrophil antibodies has also been proposed. Benign ethnic neutropenia (BEN), more common in individuals of African, Middle Eastern, and Afro-Caribbean ancestry, results in a lower baseline ANC that is not associated with increased infection risk; REMS guidelines now include BEN-adjusted monitoring thresholds to prevent inappropriate clozapine discontinuation in this population.
Clozapine produces the most pronounced weight gain of any antipsychotic, with mean weight gains of 4 to 5 kg in the first year and continued accumulation thereafter in susceptible patients.7 The mechanism involves combined H1 blockade (appetite stimulation, reduced metabolic rate), 5-HT2C blockade (reduced satiety signaling), and possibly direct effects on insulin sensitivity independent of weight gain. Fasting glucose dysregulation and new-onset type 2 diabetes mellitus occur at higher rates with clozapine than with other antipsychotics, and cases of diabetic ketoacidosis (DKA) have been reported in patients on clozapine without prior diabetes history. Dyslipidemia, particularly hypertriglyceridemia, is also prominent. A comprehensive metabolic monitoring schedule including fasting glucose, lipid panel, blood pressure, and weight assessment at baseline, 12 weeks, and annually is the minimum standard for any patient on clozapine.
Additional clozapine-specific adverse effects include sialorrhea (excessive salivation), which paradoxically occurs despite its M1 anticholinergic activity and is thought to reflect agonist activity at M4 receptors in the submandibular glands; dose-dependent reduction in seizure threshold (risk approaches 5% at doses above 600 mg per day), making clozapine the antipsychotic with the highest seizure risk and requiring prophylactic antiepileptic coverage at high doses; and orthostatic hypotension from alpha-1 blockade, most pronounced during initiation and requiring slow titration starting at 12.5 mg once or twice daily.1 Myocarditis and cardiomyopathy are rare but serious cardiac adverse effects reported predominantly in the first 6 to 8 weeks; baseline troponin and C-reactive protein (CRP) monitoring during the initiation period is recommended in some guidelines, particularly in younger patients.
Olanzapine is a thienobenzodiazepine structurally related to clozapine and shares much of its receptor binding breadth, including significant D2, D4, 5-HT2A, 5-HT2C, H1, M1, and alpha-1 blockade. It lacks clozapine's agranulocytosis risk and does not require REMS monitoring, which made it the dominant antipsychotic in the years following its 1996 approval. Its D2 affinity is substantially higher than clozapine's, placing it in the moderate D2 affinity range with a favorable 5-HT2A to D2 ratio that produces low EPS at standard doses.1 EPS risk is not absent, however; at doses above 20 mg per day, EPS, including akathisia and parkinsonism, become more common, and the EPS threshold is lower than with quetiapine or clozapine.
The clinical evidence base for olanzapine is extensive. In the CATIE trial, olanzapine had the longest time to all-cause discontinuation of any agent studied, suggesting patients experienced meaningful benefit; however, this advantage was offset by the highest rate of metabolic adverse effects in the trial, including the greatest weight gain (mean 0.9 kg per month), highest fasting glucose elevation, and most dyslipidemia of any agent evaluated.3 Olanzapine's FDA-approved indications include schizophrenia, acute and maintenance treatment of manic episodes in bipolar I disorder, and (in combination with fluoxetine as Symbyax) bipolar depression. Intramuscular (IM) olanzapine is available for acute agitation; a critical safety constraint is that IM olanzapine must not be co-administered with IM or intravenous (IV) benzodiazepines within the same session because of cases of severe respiratory depression and death attributed to their combination.
Olanzapine's metabolic consequences require explicit clinical management. Weight gain of 7 to 10 kg or more in the first year is common in susceptible patients and substantially increases long-term cardiovascular risk. The metabolic mechanism involves H1 blockade increasing appetite and reducing energy expenditure, combined with 5-HT2C blockade impairing hypothalamic satiety signaling.7 For patients who develop clinically significant weight gain on olanzapine, pharmacological adjuncts including metformin (strongest evidence base for antipsychotic-induced metabolic syndrome), topiramate, or naltrexone have been studied; metformin at 500 to 1000 mg per day attenuates weight gain and improves insulin sensitivity and is the best-supported pharmacological strategy.8 Switching to a metabolically more favorable agent (aripiprazole, lurasidone, ziprasidone) is the most effective long-term strategy when the antipsychotic choice permits it.
Olanzapine is metabolized primarily by CYP1A2 and to a lesser extent by CYP2D6, with glucuronidation as an additional route. The CYP1A2 pathway makes olanzapine subject to the same smoking-related pharmacokinetic variability described for chlorpromazine in CNS-Antipsy-02, but with greater clinical magnitude. In heavy smokers, CYP1A2 induction reduces olanzapine plasma levels by approximately 40 to 50% relative to non-smokers at the same dose.9 A patient stabilized at 20 mg per day while smoking heavily may be pharmacokinetically equivalent to a non-smoker on 10 to 12 mg per day. When a patient on olanzapine is admitted to an inpatient unit, reduces or ceases smoking, and olanzapine levels rise toward non-smoker pharmacokinetics, the result can be emergence of adverse effects (sedation, orthostatic hypotension, metabolic worsening) at previously well-tolerated doses. The converse, loss of efficacy when a patient resumes heavy smoking after discharge, is also documented. Fluvoxamine, a potent CYP1A2 inhibitor, raises olanzapine plasma levels substantially and requires dose reduction. Carbamazepine, a broad CYP inducer, reduces olanzapine levels and may require dose adjustment upward.
Quetiapine's pharmacological profile is unusual among antipsychotics and explains several features of its clinical behavior that appear paradoxical when viewed through a simple D2 blockade framework. Its D2 receptor affinity is low, and its dissociation from D2 is rapid, resulting in peak D2 occupancy during the first hours after a dose that falls substantially as plasma levels decline.2 Positron emission tomography (PET) studies show D2 occupancy of 58 to 64% at the time of peak plasma concentration, falling below 30% within 12 hours of a standard dose, consistent with its low EPS rate even at doses producing antipsychotic effect. Its affinity for H1 is very high (among the highest of all antipsychotics), for 5-HT2A and alpha-1 is moderate-to-high, and for D2 is low relative to other SGAs. At low doses (25 to 100 mg), H1 and alpha-1 blockade dominate the clinical picture, producing sedation and orthostatic hypotension with minimal antipsychotic effect. Antipsychotic effect requires doses of 400 to 800 mg per day in most patients, at which the transient peak D2 occupancy reaches the therapeutic range despite rapid dissociation.1
This receptor profile generates quetiapine's unusually broad clinical use. FDA-approved indications include schizophrenia (acute and maintenance), bipolar I mania (acute and maintenance), bipolar depression (as quetiapine extended-release), and as adjunctive therapy in major depressive disorder (MDD). The approval for bipolar depression rests on the BOLDER I and II trials, which demonstrated quetiapine's superiority to placebo on depressive symptoms in bipolar patients, making it one of the few agents with Class I evidence for this indication.10 Off-label, quetiapine at low doses (25 to 100 mg) is widely used as a sleep aid and for anxiety in psychiatric populations, a practice driven by its H1-mediated sedation but one that carries the full metabolic and cardiac risk of the agent at any dose.
Quetiapine's metabolic adverse effects are intermediate between olanzapine (highest burden) and the lower-metabolic-impact SGAs. Weight gain is meaningful, averaging 2 to 3 kg in the first year at antipsychotic doses, and is primarily H1-mediated. Glucose dysregulation is less pronounced than with olanzapine or clozapine but more than with ziprasidone, lurasidone, or aripiprazole. QTc prolongation is a class consideration; quetiapine produces modest QTc prolongation that is rarely clinically significant at standard doses but warrants attention in patients with baseline QTc prolongation or concurrent QTc-prolonging medications. Orthostatic hypotension from alpha-1 blockade is particularly prominent during initiation; quetiapine should be started at 25 to 50 mg twice daily and titrated gradually, especially in elderly patients or those on antihypertensive agents. Cataracts have been reported in animal studies at high doses, and annual slit-lamp examinations are included in the prescribing information, though the clinical significance in humans at therapeutic doses remains uncertain.
Quetiapine extended-release (XR) was developed to allow once-daily dosing and to reduce the peaks of H1-mediated sedation associated with the immediate-release formulation. The XR formulation blunts peak plasma concentrations and extends the time-concentration profile, improving tolerability of initiation and supporting adherence. For bipolar depression, once-daily quetiapine XR at 300 mg is the approved dose. For schizophrenia, doses of 400 to 800 mg once daily at bedtime are effective and align the highest plasma concentrations with the overnight period when sedation is less disruptive. The two formulations are not bioequivalent at the same total daily dose and cannot be substituted milligram for milligram without patient monitoring during the transition.
Risperidone is a benzisoxazole with high affinity for both D2 and 5-HT2A receptors and a 5-HT2A to D2 ratio that places it squarely in the atypical range at doses below approximately 6 to 8 mg per day. At doses in this range, EPS rates are low and comparable to other SGAs. Above 8 mg per day, D2 occupancy crosses the EPS threshold and risperidone begins to produce parkinsonism and akathisia at rates approaching those of high-potency FGAs; unlike olanzapine or quetiapine, it lacks the H1 or anticholinergic buffering that would attenuate this EPS emergence.1 This dose-dependent EPS profile makes dose optimization particularly consequential for risperidone, and doses above 6 mg per day offer diminishing antipsychotic returns with increasing EPS cost.
Risperidone is also the SGA most consistently associated with hyperprolactinemia, elevating prolactin levels to a degree comparable to many FGAs. This reflects its relatively selective DA and serotonin receptor pharmacology without meaningful anticholinergic or other receptor activity to modulate tuberoinfundibular D2 blockade. Prolactin elevation on risperidone is sustained, dose-dependent, and particularly relevant in younger patients where amenorrhea, sexual dysfunction, and bone density effects have long-term implications. Risperidone's metabolic profile is intermediate: meaningful weight gain (less than olanzapine or clozapine, more than aripiprazole or ziprasidone) and modest glucose dysregulation.7
Risperidone is metabolized by CYP2D6 to 9-hydroxyrisperidone (paliperidone), which is the principal active metabolite and itself an approved antipsychotic. In CYP2D6 extensive metabolizers, the parent drug and metabolite coexist in plasma in roughly equal proportions; in poor metabolizers, risperidone accumulates and paliperidone is reduced. Because the combined D2 occupancy of risperidone plus paliperidone determines the clinical effect, CYP2D6 phenotype affects the dose-response relationship but less dramatically than for purely single-compound agents. Strong CYP2D6 inhibitors including fluoxetine and paroxetine raise risperidone plasma levels by reducing its conversion to paliperidone; dose reduction of risperidone should be considered when these agents are added.9
Paliperidone (9-hydroxyrisperidone) has a pharmacokinetic profile that distinguishes it from virtually all other antipsychotics: it undergoes minimal hepatic metabolism and is eliminated primarily unchanged in urine through renal excretion (approximately 59% of a dose excreted unchanged).1 This means that paliperidone plasma levels are largely unaffected by CYP enzyme inducers or inhibitors, and that dose adjustment is required in renal impairment rather than hepatic impairment, the reverse of most antipsychotics. Its receptor profile is essentially identical to risperidone (high D2 and 5-HT2A affinity, significant alpha-2 blockade), and its clinical efficacy and adverse effect profile closely mirror risperidone's, including equivalent hyperprolactinemia. The primary clinical advantage of paliperidone over risperidone is pharmacokinetic predictability: because it bypasses CYP2D6, paliperidone's plasma levels are less affected by CYP2D6 polymorphism and drug interactions, making it preferable when CYP2D6 drug interaction burden is a concern.
Risperidone and paliperidone have the most extensively developed long-acting injectable (LAI) program in the SGA class. Risperidone microspheres (Risperdal Consta) are administered every 2 weeks by IM injection, with a unique pharmacokinetic lag of approximately 3 weeks before therapeutic plasma levels are achieved, requiring oral risperidone supplementation during this period. Paliperidone palmitate (Invega Sustenna) is available in monthly (Invega Sustenna) and 3-monthly (Invega Trinza) formulations. The 3-monthly formulation requires prior stabilization on the monthly formulation for at least 4 months before conversion. A 6-monthly subcutaneous paliperidone palmitate formulation (Invega Hafyera) is also approved for patients stable on Invega Trinza. This hierarchy of LAI formulations from biweekly to 6-monthly represents the most fully developed adherence-support system available for any antipsychotic, and trial data consistently show reduced hospitalization rates and improved adherence outcomes with LAI versus oral formulations in patients with documented adherence difficulties.11
The core SGAs covered in this module differ substantially in their pharmacokinetic profiles, and these differences translate directly into clinical management decisions. Understanding the primary metabolic route of each agent is the foundation for anticipating drug interactions and adjusting doses in the context of comedication changes or organ impairment.
Clozapine is metabolized primarily by CYP1A2 (responsible for approximately 70 to 80% of its hepatic clearance) and to a lesser extent by CYP3A4 and CYP2D6. Its oral bioavailability is approximately 50 to 60%, with a plasma half-life of 12 to 16 hours supporting twice-daily dosing in most patients. Volume of distribution is very large (approximately 6 L/kg), reflecting extensive tissue binding. The principal active metabolite is norclozapine (N-desmethylclozapine), which has partial D2 agonist and full M1 agonist activity and may contribute to clozapine's cognitive effects. Plasma protein binding is approximately 97%, primarily to albumin and alpha-1-acid glycoprotein (AGP); acute illness raises AGP levels and reduces the free fraction, meaning that plasma total clozapine levels may overestimate free drug concentration during acute inflammatory states.1
Olanzapine is also primarily CYP1A2-metabolized (with CYP2D6 and direct glucuronidation as secondary routes), oral bioavailability of approximately 60%, and a half-life of 21 to 54 hours supporting once-daily dosing. The wide half-life range reflects inter-individual variability in CYP1A2 activity, largely driven by smoking status. Quetiapine is primarily CYP3A4-metabolized, with an oral bioavailability of approximately 9% (highly variable, substantially food-dependent) and a short half-life of 6 to 7 hours for the immediate-release formulation, necessitating twice-daily dosing. Risperidone is CYP2D6-metabolized to paliperidone, with an oral half-life of 3 hours for the parent drug and 21 hours for paliperidone in extensive metabolizers. Paliperidone itself is renally cleared with minimal hepatic metabolism and a half-life of approximately 23 hours.1,9
CYP1A2 interactions are the most clinically consequential for clozapine and olanzapine. Fluvoxamine, a potent CYP1A2 inhibitor used for obsessive-compulsive disorder (OCD) and sometimes added to clozapine regimens as an augmenting agent, raises clozapine plasma levels by 5 to 10-fold at typical fluvoxamine doses of 50 to 100 mg per day.9 This interaction has been intentionally exploited to allow lower clozapine doses while maintaining therapeutic plasma levels, reducing agranulocytosis risk, and improving tolerability, a practice that requires careful plasma level monitoring. Ciprofloxacin, a moderate CYP1A2 inhibitor, raises clozapine levels by approximately 60% and has been associated with clozapine toxicity in case reports; the course of a fluoroquinolone antibiotic in a patient on clozapine warrants temporary dose reduction and symptom monitoring. Carbamazepine, a broad CYP inducer (CYP1A2, CYP3A4, CYP2D6), reduces clozapine plasma levels substantially and is also independently associated with additive bone marrow suppression risk; the combination of clozapine and carbamazepine is generally contraindicated.9
CYP3A4 interactions are most relevant for quetiapine. Potent CYP3A4 inhibitors including azole antifungals (fluconazole, ketoconazole, itraconazole), clarithromycin, ritonavir, and grapefruit juice raise quetiapine plasma levels substantially; the quetiapine prescribing information recommends dose reduction to one sixth of the standard dose when a potent CYP3A4 inhibitor is co-administered. Potent CYP3A4 inducers including carbamazepine, rifampin, phenytoin, and St. John's Wort reduce quetiapine levels dramatically (up to 90% reduction with carbamazepine), potentially rendering standard doses subtherapeutic; dose increases of five-fold or more may be needed if the combination is unavoidable, with corresponding dose reduction required when the inducer is stopped.9 CYP2D6 inhibitors (fluoxetine, paroxetine, bupropion) inhibit risperidone's conversion to paliperidone, raising risperidone levels; dose reduction by approximately 50% should be considered when adding a strong CYP2D6 inhibitor to established risperidone therapy.
Pharmacodynamic interactions relevant to this module include the additive CNS depression from combining any SGA with benzodiazepines, opioids, or other central nervous system (CNS) depressants, a concern particularly in elderly patients and in the context of acute agitation management. The IM olanzapine and benzodiazepine contraindication is discussed in Section 3. Additive QTc prolongation is a lesser concern for clozapine, olanzapine, and risperidone than for thioridazine, pimozide, or ziprasidone, but it is not zero, particularly in the context of electrolyte abnormalities, and should be assessed whenever a QTc-active co-medication is prescribed.
Stahl SM. Stahl's Essential Psychopharmacology: Neuroscientific Basis and Practical Applications. 4th ed. Cambridge: Cambridge University Press; 2013:129–237.
Seeman P. Atypical antipsychotics: mechanism of action. Can J Psychiatry. 2002;47(1):27–38.
doi:10.1177/070674370204700106Lieberman JA, Stroup TS, McEvoy JP, et al. Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med. 2005;353(12):1209–1223.
doi:10.1056/NEJMoa051688Kane J, Honigfeld G, Singer J, Meltzer H. Clozapine for the treatment-resistant schizophrenic: a double-blind comparison with chlorpromazine. Arch Gen Psychiatry. 1988;45(9):789–796.
doi:10.1001/archpsyc.1988.01800330013001Meltzer HY, Alphs L, Green AI, et al. Clozapine treatment for suicidality in schizophrenia: International Suicide Prevention Trial (InterSePT). Arch Gen Psychiatry. 2003;60(1):82–91.
doi:10.1001/archpsyc.60.1.82Citrome L. Clozapine for schizophrenia: life-threatening or life-saving treatment? Curr Psychiatry. 2009;8(12):56–63. [REMS monitoring thresholds and neutropenia management protocol]
Allison DB, Mentore JL, Heo M, et al. Antipsychotic-induced weight gain: a comprehensive research synthesis. Am J Psychiatry. 1999;156(11):1686–1696.
doi:10.1176/ajp.156.11.1686Fiedorowicz JG, Miller DD, Bishop JR, Calarge CA, Ellingrod VL, Haynes WG. Systematic review and meta-analysis of pharmacological interventions for weight gain from antipsychotics and mood stabilizers. Curr Psychiatry Rev. 2012;8(1):25–36.
doi:10.2174/157340012798994867Spina E, de Leon J. Metabolic drug interactions with newer antipsychotics: a comparative review. Basic Clin Pharmacol Toxicol. 2007;100(1):4–22.
doi:10.1111/j.1742-7843.2007.00017.xCalabrese JR, Keck PE Jr, Macfadden W, et al. A randomized, double-blind, placebo-controlled trial of quetiapine in the treatment of bipolar I or II depression. Am J Psychiatry. 2005;162(7):1351–1360.
doi:10.1176/appi.ajp.162.7.1351Kishimoto T, Robenzadeh A, Leucht C, et al. Long-acting injectable vs oral antipsychotics for relapse prevention in schizophrenia: a meta-analysis of randomized trials. Schizophr Bull. 2014;40(1):192–213.
doi:10.1093/schbul/sbs150