Skip to main content Accessibility help
×
Home
Hostname: page-component-56f9d74cfd-5k9ck Total loading time: 1.375 Render date: 2022-06-26T00:21:38.823Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "useRatesEcommerce": false, "useNewApi": true }

Book contents

2 - The Therapeutic Threshold and the Point of Futility

Published online by Cambridge University Press:  19 October 2021

Jonathan M. Meyer
Affiliation:
University of California, San Diego
Stephen M. Stahl
Affiliation:
University of California, Riverside and San Diego

Summary

Prior to our modern understanding of differences in drug metabolism, it was apparent that a wide range of response to antipsychotic dosages existed, with some schizophrenia patients responding minimally regardless of dose. Whether this variation was due to nonadherence or other biological factors [1, 2], the need to quantify antipsychotic exposure and clinical response spurred interest in development of reliable laboratory assays and improved psychiatric research instruments in the 1960s. The Brief Psychiatric Rating Scale (BPRS) was developed in 1962 to rate major symptom characteristics of acute psychosis, and represented an important advancement over older methods that lacked “efficiency, speed and economy” [3].

Type
Chapter
Information
The Clinical Use of Antipsychotic Plasma Levels
Stahl's Handbooks
, pp. 34 - 59
Publisher: Cambridge University Press
Print publication year: 2021

Principles

  • The bulk of the data correlating plasma antipsychotic levels and psychiatric response relates to the treatment of schizophrenia spectrum disorders, with extremely limited information for other diagnoses. For that reason, the discussions within this handbook about symptomatic improvement pertain only to the management of schizophrenia. Antipsychotic plasma levels can be used for monitoring adherence regardless of the indication (see Chapter 4).

  • The therapeutic threshold represents a value achieved by a synthesis of the available data to arrive at a plasma level that best discriminates true responders from nonresponders. This information comes primarily from clinical studies of oral antipsychotics, supported by imaging data on dopamine receptor occupancy. There are a range of numbers presented in the literature, but there is often a convergence among the more widely studied antipsychotics (e.g. haloperidol, fluphenazine, clozapine, olanzapine, risperidone, etc.).

  • Some patients may respond adequately below the usual response threshold, and this may be more common with long-acting injectable agents. For clinical use, the therapeutic threshold represents an initial target plasma level for patients who are inadequate responders and who do not have dose-limiting adverse effects.

  • The definition of an upper limit is quite variable and confusing. What laboratories report as the upper limit varies dramatically between sources, and may represent a reference range based on doses instead of a therapeutic range. Consensus guidelines may also vary in their definitions, and at times will employ tolerability thresholds or plasma levels based on maximum approved dosages.

  • The point of futility is a term of art utilized within the extensive California Department of State Hospitals to educate clinicians about two important concepts with regard to an upper limit for antipsychotic plasma levels: (1) a small proportion of patients may never exhibit dose-limiting adverse effects and will tolerate further titration; (2) ongoing titration beyond a certain plasma level (the point of futility) is fruitless as < 5% of patients may respond to these higher plasma levels.

Introduction

Prior to our modern understanding of differences in drug metabolism, it was apparent that a wide range of response to antipsychotic dosages existed, with some schizophrenia patients responding minimally regardless of dose. Whether this variation was due to nonadherence or other biological factors [Reference Willcox, Gillan and Hare1, Reference Curry, Davis and Janowsky2], the need to quantify antipsychotic exposure and clinical response spurred interest in development of reliable laboratory assays and improved psychiatric research instruments in the 1960s. The Brief Psychiatric Rating Scale (BPRS) was developed in 1962 to rate major symptom characteristics of acute psychosis, and represented an important advancement over older methods that lacked “efficiency, speed and economy” [Reference Overall and Gorham3]. Papers discussing the measurement and the significance of plasma chlorpromazine levels appeared in 1968 [Reference Curry4, Reference Curry and Marshall5], followed by improved precision of laboratory drug assays, and the advent of newer rating scales for schizophrenia clinical trials (e.g. the Positive and Negative Syndrome Scale [PANSS] by Kay, Fiszbein, and Opler in 1987) [Reference Kay, Fiszbein and Opler6]. The increased availability of antipsychotic plasma level data allowed schizophrenia investigators to explore fundamental questions about the relationships between antipsychotic exposure, clinical response, and tolerability [Reference Sakalis, Curry and Mould7]. By 1981, Dr. Theodore Van Putten, a professor at UCLA and an early proponent of this research, was able to define a therapeutic range for chlorpromazine (CPZ), and described one clinical scenario wherein a plasma level could inform treatment: “It is in the inaccessible patient whose illness is only minimally, or not at all, sensitive to CPZ that a plasma level might be especially useful” [Reference Van Putten, May and Jenden8].

To arrive at therapeutic plasma level ranges, these early investigators used mathematical models for calculating the response threshold, but defining an upper limit relied less on computational analysis and more on agreeing to a point where the plasma level became intolerable for the vast majority of patients (e.g. 90%) [Reference Van Putten, Marder and Wirshing9]. Though overdose with certain first-generation antipsychotics could lead to fatal QT prolongation (e.g. chlorpromazine, pimozide, thioridazine) [Reference Nielsen, Graff and Kanters10], antipsychotics as a class possessed a broader therapeutic index than other commonly used psychotropics in the 1960s–1980s (e.g. tricyclic antidepressants, lithium), so the upper limit definition was not primarily a safety issue, but a tolerability one [Reference Van Putten, Aravagiri and Marder11]. As the initial studies utilized first-generation antipsychotics (FGAs), neurological adverse effects (e.g. parkinsonism, akathisia, dystonia) were the common dose-limiting issue (Figure 2.1) [Reference Van Putten, Marder and Wirshing9]. The development of atypical antipsychotics altered tolerability profiles, but the basic questions persist: whether this upper limit should be defined by a certain intolerability rate (e.g. 50%, 90%, etc.), or by other measures. Where to draw this upper limit remains a healthy source of debate, and one area where the field has not reached consensus [Reference Meyer12Reference Schoretsanitis, Kane and Correll15].

Figure 2.1 Plasma haloperidol levels and the proportion of patients with intolerable adverse effects [Reference Van Putten, Marder and Wirshing9, Reference Midha, Hubbard and Marder16]

(Adapted from: K. K. Midha, J. W. Hubbard, S. Marder, et al. [1994]. Impact of clinical pharmacokinetics on neuroleptic therapy in patients with schizophrenia. J Psychiatry Neurosci, 19, 254–264.)

Over 25 years of research documents the clinical value of directly measuring antipsychotic exposure, yet routine use waned dramatically in the early 1990s despite the publication of Marder, Davis, and Janicak’s edited 1993 book entitled Clinical Use of Neuroleptic Plasma Levels [Reference Marder, Davis and Janicak17]. As noted in the Introduction, practical issues in obtaining rapid results can frustrate attempts to make clinical decisions within time frames achieved when ordering other drug levels (e.g. lithium). The 1993 Marder et al. volume is scholarly but contains discussions about mathematical concepts such as receiver operating characteristic (ROC) analyses that are pivotal to academic debates over response thresholds but have less appeal to a broader clinical audience. Importantly, there has been a paucity of discussion regarding the variable reporting of upper limit thresholds in the literature, and how clinicians should make sense of this literature [Reference Meyer18]. In addition to the aforementioned, perhaps the biggest impediment to widespread use of antipsychotic plasma levels was the development of second-generation antipsychotics (SGAs). To arrive at robust correlations between a drug dose and a plasma level, or between a plasma level and clinical response, requires multiple fixed-dose studies. Unfortunately, this data is less available than with FGAs – clozapine being the notable exception [Reference Remington, Agid and Foussias19]. A 2013 review covering SGAs (other than clozapine) available in the US (risperidone, olanzapine, quetiapine, aripiprazole, ziprasidone, paliperidone, iloperidone, asenapine, lurasidone) found 192 papers, of which only 11 provided data from prospective trials of acute psychosis that correlated levels with efficacy [Reference Lopez and Kane20]. Moreover, of the nine antipsychotics involved in the search, only four had relevant articles, and only two medications were involved in multiple trials. Studies also varied widely in methodology, with only four employing a fixed-dose model. Based on the limited data for SGAs, the authors concluded: “The utility of therapeutic drug monitoring of SGAs (other than clozapine) remains an open question, although limited evidence from fixed-dose studies is encouraging” [Reference Lopez and Kane20].

While the 2013 review might engender a certain degree of nihilism, there have been concerted attempts to refocus clinicians’ attention on the significant value that measurement of antipsychotic levels can bring to clinical decision making, led by the Therapeutic Drug Monitoring (TDM) task force of the German Arbeitsgemeinschaft für Neuropsychopharmakologie und Pharmakopsychiatrie (AGNP). AGNP issued their first guidelines for TDM across all psychiatric medications in 2004, followed by updates in 2011 and 2018, and by a consensus statement in 2020 issued jointly with the American Society for Clinical Psychopharmacology (ASCP) that covered use of antipsychotic levels [Reference Hiemke, Bergemann and Clement13, Reference Schoretsanitis, Kane and Correll15]. In crafting the consensus paper, AGNP and ASCP sought to remind clinicians that levels are obtainable, that there are numerous clinical reasons to check levels, and that antipsychotic response is low below certain threshold levels, whether due to poor adherence or rapid drug metabolism [Reference Schoretsanitis, Kane and Correll15]. The discussion below will assist all clinicians in developing a fuller understanding of how level recommendations are arrived at, and, importantly, how to make sense of the recommendations in this volume and other sources about the use of plasma level information to optimize likelihood of clinical response for schizophrenia patients.

A Evidence Used to Define a Therapeutic Threshold

Astute clinicians will appreciate that therapeutic thresholds for psychotropic medications are consensus conclusions which attempt to provide clarity from a range of values [Reference Grandjean and Aubry21, Reference Machado-Vieira, Otaduy and Zanetti22, Reference Hiemke, Bergemann and Clement13]. For antipsychotics, the degree of certainty is based on the extent and quality of the data [Reference Schoretsanitis, Kane and Correll15], and whether the estimate comes primarily from fixed-dose trials, is inferred from single photon (SPECT) or Positron (PET) Emission Tomography studies of dopamine D2 receptor occupancy, or is calculated from minimum effective doses [Reference Uchida, Takeuchi and Graff-Guerrero23, Reference Takeuchi, MacKenzie and Samaroo24]. Bearing in mind that these are all monotherapy studies and thus cannot be used for antipsychotic combination therapy or for uses other than schizophrenia, estimates of the therapeutic threshold derived from fixed-dose antipsychotic trials provide the most indisputable connection between plasma levels and response, and, unlike imaging, require no assumption that the therapeutic mechanism is primarily correlated with D2 receptor occupancy [Reference Hiemke, Bergemann and Clement13, Reference Schoretsanitis, Kane and Correll15]. Regardless of the data source, the identification of correlation between response likelihood and a plasma level, dose, or imaging result comes from clinical studies that are typically performed in adult patients with an acute exacerbation of schizophrenia. Conclusions can also be gleaned from longer-term maintenance studies by examining the plasma level threshold below which higher relapse rates are experienced [Reference Melkote, Singh and Vermeulen25, Reference Correll, Jain and Meyer26].

For acute psychosis trials, response is based on a certain degree of symptom relief using a standard rating scale administered by trained research personnel (e.g. PANSS, BPRS). Varying response definitions are seen throughout the literature, but a 50% reduction in the primary outcome measure is most commonly employed, although lower values may be acceptable in more treatment-resistant populations [Reference Samara, Leucht and Leeflang27]. Differences in study design, rating scales, and patient populations generate some of the mathematical ‘noise’ that results in differing plasma level values reported for the therapeutic threshold. Clinicians are sometimes frustrated that there is more than one value reported in the literature as various scientists choose to weigh certain types of data in different ways. However, as the body of information increases for a particular antipsychotic (e.g. clozapine, haloperidol), there is a much narrower range of suggested response thresholds (Box 2.1). Having an understanding of how the various data sources translate into a recommended response threshold is helpful for appreciating the currently available recommendations, and for understanding changes that might occur based on newer studies.

Box 2.1 Important Concepts in Using Response Thresholds for Schizophrenia Treatment

  1. a. There is no one ‘right’ answer with respect to therapeutic threshold levels for any antipsychotic, but a range of values presented in the literature, whose degree of certainty is enhanced by the breadth of data. When the sample size is large across clinical studies, there is greater agreement (e.g. a clozapine response threshold of 350 ng/ml) [Reference Meyer and Stahl28]. The AGNP/ASCP consensus paper divides antipsychotics into four groups based on the extent of supporting data to justify therapeutic drug monitoring: Level 1 – strongly recommended; Level 2 – recommended; Level 3 – useful; Level 4 – potentially useful [Reference Schoretsanitis, Kane and Correll15].

  2. b. Throughout this volume, a justification will be provided for why a particular value is selected, but clinicians are free to settle on any evidence-based value for a response threshold within the range reported in the literature. For haloperidol, there are response thresholds of 1 ng/ml, 2 ng/ml, 3 ng/ml, and 5 ng/ml described in various papers [Reference Van Putten, Marder and Mintz29, Reference Van Putten, Marder and Wirshing9, Reference Midha, Hubbard and Marder16, Reference Meyer12, Reference Schoretsanitis, Kane and Correll15]. The crucial issue is to decide on values for those antipsychotics one uses frequently, and modify these as newer literature dictates.

  3. c. Some patients may respond adequately to levels below the response threshold. As discussed in Chapter 5, the therapeutic threshold is a guidepost to assist clinicians in maximizing the chances of converting an inadequate responder who is tolerating a medication into a responder.

  4. d. Having a plasma level above the threshold is no guarantee of adequate response. Where in the therapeutic range a patient will respond, or whether the patient responds at all, relates to pharmacogenetic and other variables that cannot be estimated as of 2021. (See Chapter 5 for an extensive discussion on using plasma antipsychotic levels to maximize the likelihood of response.)

1 Receiver Operating Characteristic (ROC) Curve Analysis of Clinical Data

In World War II, radar engineers were motivated to develop methods that better discriminated between enemy objects and friendly forces on the battlefield. In doing so, these engineers sought to improve the signal detection ability of the receivers used by radar operators, and the plots created were termed receiver operating characteristic (ROC) curves. To use the appropriate statistical terminology, the goal was to maximize the chances of identifying all enemy objects (the true positive rate, or sensitivity, which is plotted on the y-axis), while minimizing the odds of incorrectly labeling a friendly force as an enemy (the false positive rate, calculated as 1-specificity and plotted on the x-axis). These concepts quickly moved from the battlefield to the scientific arena where the applications focused on optimizing diagnostic test results to find a value that might accurately distinguish true responders from nonresponders (false positives). An early ROC analysis for clozapine, published in 1991 from a trial of 29 treatment-resistant adult schizophrenia patients, suggested that a threshold of 350 ng/ml best identified true responders without sacrificing the extent of false positives [Reference Perry, Miller and Arndt30]. A subsequent 1995 study of 45 treatment-resistant schizophrenia patients came to similar conclusions [Reference Kronig, Munne and Szymanski31]. Figure 2.2 presents the graph achieved from the 1995 study with the corresponding percentage of responders and nonresponders plotted for each plasma clozapine level. Treatment response was defined as a 20% decrease in the BPRS score, and either a clinical global impression (CGI) severity rating of mildly ill (score of 3 or below) or a BPRS score of 35 or less.

Figure 2.2 Receiver operating characteristic (ROC) curve from 1995 data on the proportion of clozapine responders and nonresponders for plasma clozapine levels in 50 ng/ml increments plotted from the data in Table 2.1 [Reference Kronig, Munne and Szymanski31]

(Adapted from: M. H. Kronig, R. A. Munne, S. Szymanski, et al. [1995]. Plasma clozapine levels and clinical response for treatment-refractory schizophrenic patients. Am J Psychiatry, 152, 179–182.)

Table 2.1 Responders (sensitivity) and nonresponders (1-specificity) for plasma clozapine levels in 50 ng/ml increments [Reference Kronig, Munne and Szymanski31]

Blood level (ng/ml)% responders (sensitivity)% nonresponders (1-specificity)p
25080.063.60.28
30080.050.00.06
35080.045.5<0.04
40066.745.50.20
45066.740.90.12
50066.736.40.07
55060.031.80.09
60060.022.7<0.03
65046.722.70.12
70046.713.6<0.03

Not surprisingly, the proportion of nonresponders increases with lower plasma levels on the x-axis, with 35% being classified as nonresponders with clozapine levels < 250 ng/ml. Conversely, the proportion of responders increases along with the plasma level; however, one can see by the shape of the curve that, for this data set, plasma clozapine levels > 350 ng/ml, > 600 ng/ml and > 700 ng/ml identify intervals with marked reduction in the proportion of nonresponders (i.e. lower false positive rate). Mathematically, one selects a cutoff that maximizes sensitivity (i.e. the distance from the x-axis is maximized), while false positives (calculated as 1-specificity) are minimized (i.e. distance from the y-axis is minimized) [Reference Kronig, Munne and Szymanski31]. Statistical analysis found that the plasma clozapine level of 350 ng/ml best satisfied those criteria, and this value was chosen as the response threshold. Figure 2.3 represents data from a 2016 study of olanzapine response in 151 Taiwanese adult schizophrenia patients illustrating the cutoff of 22.77 ng/ml that was mathematically calculated to discriminate responders from nonresponders using a PANSS total score ≤ 58 as the criterion for response [Reference Lu, Wu and Chen32].

Figure 2.3 2016 ROC curve of the plasma olanzapine level required to be rated as mildly ill (PANSS score ≤ 58) from a cohort of Taiwanese schizophrenia patients (n = 151) [Reference Lu, Wu and Chen32]

(Adapted from: M. L. Lu, Y. X. Wu, C. H. Chen, et al. [2016]. Application of plasma levels of olanzapine and N-desmethyl-olanzapine to monitor clinical efficacy in patients with schizophrenia. PLoS One, 11, e0148539.)

The differing definitions of treatment response represent one of many reasons why various threshold cutoffs are reported in the literature from ROC analyses. In some instances, investigators who want to increase the precision of an ROC analysis by combining results of several studies will use the original raw PANSS scoring data to look specifically at an outcome measure that can be calculated for all participants (e.g. a PANSS reduction of 50%). When there is a commonly used definition of response, and sufficient data from multiple samples, ROC analysis is considered a gold standard method for arriving at response threshold plasma levels, and this method has been used in papers on antipsychotic response dating back to 1981 [Reference Marder, Davis and Janicak17]. These ROC analyses have also confirmed that plasma antipsychotic threshold levels significantly correlate with clinical response, although conclusions for individual agents are often limited by the paucity of studies reporting the proportion of responders and nonresponders at various plasma levels.

2 Imaging of Dopamine D2 Occupancy Relationships with Plasma Antipsychotic Levels

Based on the common mechanism of action for first-generation antipsychotics, the paradigm for drug development through the late 1980s rested on animal models designed to predict significant dopamine D2 receptor antagonism [Reference Meyer, Brunton, Hilal-Dandan and Knollmann33]. The synthesis of ligands in the mid-1980s for imaging research allowed investigators to use SPECT and later PET scans to quantify D2 receptor occupancy in human subjects, correlate occupancy with response and neurological adverse effects (e.g. parkinsonism, akathisia, dystonia), and infer expected plasma antipsychotic levels for the response threshold [Reference Farde, Hall and Ehrin34]. As with ROC analyses derived from plasma level data, varying definitions of response gave rise to slightly different conclusions using ROC curves or other methods to calculate the threshold D2 occupancy for antipsychotic response. Moreover, there are technical aspects of imaging studies that introduce another layer of complexity (e.g. ligand choice, time since last antipsychotic dose), with PET data providing superior results compared to older SPECT studies. Moreover, PET studies typically have very small sample sizes (under 10 subjects), and often include normal volunteers or younger (nongeriatric) schizophrenia patients, possibly limiting the generalizability of findings [Reference Kirino, Suzuki and Takeuchi35]. Using antipsychotics whose presumed mechanism is D2 antagonism, PET scans obtained at steady state, 12h post-dose, show a range of occupancy thresholds for antipsychotic response from 50% to 65% [Reference Uchida, Takeuchi and Graff-Guerrero23, Reference Arakawa, Takano and Halldin36]. While 60% or 65% are the more commonly cited numbers, lower figures have emerged from studies of first-episode schizophrenia patients, perhaps related to several differences compared to more chronic patients: younger age, limited (or absent) prior antipsychotic exposure, generally higher response rates, and sensitivity to D2 antagonism (Figure 2.4) [Reference de Haan, van Bruggen and Lavalaye37].

Figure 2.4 Adverse effects and response based on fitted D2 occupancy curves for D2 antagonist antipsychotics [Reference Kapur and Seeman38]

(Adapted from: S. Kapur and P. Seeman [2001]. Does fast dissociation from the dopamine d(2) receptor explain the action of atypical antipsychotics? A new hypothesis. Am J Psychiatry, 158, 360–369.)

Imaging of D2 occupancy provided the psychiatric community with in vivo data clearly establishing the link between peripherally obtained plasma antipsychotic levels and effects at a known central nervous system (CNS) receptor target. The consensus conclusion that 60%–65% 12h trough D2 occupancy best predicts response from D2 antagonists is drawn from a large body of data with chronic schizophrenia patients. Unfortunately, estimates of the plasma level thresholds needed to achieve 60%–65% D2 occupancy are often obtained from studies using younger schizophrenia patients or normal volunteers. The use of younger patients or those without schizophrenia may be one factor giving rise to plasma levels required for 60% or 65% D2 occupancy that are different from, and usually much lower than, those derived from ROC analyses of plasma level thresholds in a chronic schizophrenia population. For example, a 2011 review calculated the plasma concentration needed to achieve 60% D2 occupancy for a range of antipsychotics: risperidone + 9-OH risperidone level: 10.5 ng/ml; olanzapine: 13.9 ng/ml; haloperidol: 0.8 ng/ml; and ziprasidone: 70.0 ng/ml [Reference Uchida, Takeuchi and Graff-Guerrero23]. The implied olanzapine level of 13.9 ng/ml derived from this analysis is well below values of 22.77 ng/ml (n = 151) [Reference Perry, Lund and Sanger39] and 23.2 ng/ml (n = 84) [Reference Perry, Lund and Sanger39] obtained by ROC analyses from clinical trials enrolling chronic schizophrenia patients with mean ages of 41.3 and 36.8 years, respectively.

The development of imaging techniques has been a boon to our understanding of D2 antagonism, and helped to define ranges of D2 occupancy best suited to antagonist antipsychotics for most patients with schizophrenia. The difficulty in translating thresholds averaged from numerous participants is the need to treat outliers on the dose and plasma level response curve [Reference Preskorn40, Reference Meyer12]. Unlike medications with very narrow therapeutic indices, there is an extremely broad range of plasma antipsychotic levels which some patients tolerate and where they respond. PET imaging has shown that the risk for neurological adverse effects is seen (on average) when one exceeds 78%–80% D2 occupancy [Reference Kapur, Zipursky and Jones41]; however, clinical studies clearly indicate that a substantial proportion may tolerate significantly higher levels of D2 receptor blockade without experiencing neurological adverse effects. One example is seen in the 1994 pivotal risperidone trial results. The authors noted that 39% of subjects receiving risperidone 16 mg/d, and 47% receiving haloperidol 20 mg/d required antiparkinsonian medications despite exposure to doses associated with > 90% D2 blockade [Reference Marder and Meibach42]. Many schizophrenia patients do not benefit from very high plasma antipsychotic levels, and, in routine usage, this is often discouraged by practice guidelines or by local prescribing regulations that restrict maximum dosages to the licensed dosage range [Reference Hiemke, Bergemann and Clement13, Reference Barnes, Drake and Paton14]. Nonetheless, there are certain outpatients or inpatients in specialized forensic settings who require and tolerate > 80% D2 antagonism for optimal positive symptom control [Reference Meyer12]. Using antipsychotics in this manner for these patients is not inherently unsafe, but must be pursued with careful documentation of plasma levels and other safety monitoring as appropriate, and with a full understanding that the benefit to most schizophrenia patients from high plasma levels that exceed 80% D2 occupancy is limited, even when tolerated [Reference Barnes, Drake and Paton14].

The ‘ideal’ range of 65%–80% D2 occupancy may not apply to certain outliers, and also does not apply to antipsychotics whose primary therapeutic mechanism is not D2 antagonism. Since 2002, three partial agonist antipsychotics have been approved in the US (aripiprazole, brexpiprazole, cariprazine), all of which have intrinsic activity at D2 receptors estimated in the range 18%–24% [Reference Tadori, Forbes and McQuade43, Reference Maeda, Sugino and Akazawa44]. To achieve clinical response in schizophrenia, the partial agonists not only have very high in vitro affinity for the D2 receptor, they generally require a minimum of 80% D2 occupancy. For aripiprazole and cariprazine, the expected D2 occupancy threshold for response is close to 80%, and the upper end of the range is essentially 100% (Figure 2.5) [Reference Mizrahi, Mamo and Rusjan45Reference Girgis, Slifstein and D’Souza47]. Brexpiprazole is also a partial agonist, but appears to have a slightly lower threshold for response of 65% to 80% as seen with D2 antagonists. The recommended brexpiprazole dose range for schizophrenia is 2–4 mg/d, and a PET study of schizophrenia patients (mean age 42 ± 8 years) found that the subtherapeutic dose of 1 mg yielded 65% D2 receptor occupancy, while 4 mg averaged 80% [Reference Girgis, Forbes and Abi-Dargham48].

Figure 2.5 Fitted D2 occupancy curve for aripiprazole in the dosage range of 10–30 mg/d [Reference Mamo, Graff and Mizrahi49]

(Adapted from: D. Mamo, A. Graff, R. Mizrahi, et al. [2007]. Differential effects of aripiprazole on D(2), 5-HT(2), and 5-HT(1A) receptor occupancy in patients with schizophrenia: a triple tracer PET study. Am J Psychiatry, 164, 1411–1417.)

Dopamine D2 occupancy thresholds also do not apply to agents whose mechanism is not highly dependent on D2 antagonism. Clozapine was the first example of an antipsychotic for which trough levels of D2 occupancy were well under 65%, and even at peak plasma levels only transiently exceeded this occupancy threshold [Reference Wiesel, Farde and Nordstrom50]. This same finding is also seen with the weak D2 antagonist quetiapine [Reference Gefvert, Bergstrom and Langstrom51], and the newly approved antipsychotic lumateperone [Reference Vanover, Davis and Zhou52]. Perhaps the most extreme example is the antipsychotic pimavanserin, a medication approved for Parkinson’s disease psychosis with positive data for dementia-related psychosis. Pimavanserin is a highly selective inverse agonist at serotonin 5HT2A (Ki 0.087 nM) and 5HT2C (Ki 0.44 nM) receptors, and lacks affinity for any other monoamine receptor [Reference Meltzer and Roth53, Reference Stahl54]. While pimavanserin is not approved for the treatment of schizophrenia, there are agents in development for schizophrenia which lack any D2 binding: the muscarinic M1 and M4 agonist xanomeline, and the trace amine associated receptor-1 (TAAR1) agonist SEP-363856 [Reference Shekhar, Potter and Lightfoot55Reference Dedic, Jones and Hopkins58]. A range of D2 occupancy is thus associated with antipsychotic response depending on the molecule (Figure 2.6), but for medications whose mechanism is not fully understood (e.g. TAAR1 agonists), or that have limited or no D2 binding (e.g. clozapine, lumateperone, pimavanserin, xanomeline, SEP-363856), PET imaging will not be the primary means to decide on a plasma level cutoff for response. For these agents, one can still create ROC analyses from clinical trials to calculate a plasma level response threshold. ROC analyses are based solely on the proportion of responders/nonresponders at various plasma levels, and are thus agnostic with regards to the presumed mechanism of action.

Figure 2.6 Range of D2 occupancy for antipsychotics

3 Inference from Minimum Effective Dose

Since the fortuitous discovery of chlorpromazine’s antipsychotic property, it has become apparent that the extremely high antipsychotic doses routinely used in the 1950s–1980s are not necessary for most patients [Reference Takeuchi, MacKenzie and Samaroo24]. PET studies find that a haloperidol dose of 2–5 mg/d generates roughly 65%–80% D2 occupancy (Figure 2.4), yet starting doses of 40 mg/d or more were not uncommon in the 1970s [Reference Cohen and Cohen59]. The large NIMH-sponsored Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) schizophrenia study used perphenazine as the first-generation comparator to the second-generation antipsychotics for that study [Reference Lieberman, Stroup and McEvoy60]. In this group of chronic schizophrenia patients (n = 1493), the mean modal dose of perphenazine was 20.8 mg/d, equivalent to only 5.55 mg/d of haloperidol [Reference Meyer, Brunton, Hilal-Dandan and Knollmann33]. Based on analyses of clinical trials outcomes, investigators have sought to find the minimum effective dose (MED) for antipsychotics as a means of expressing relative equivalent doses [Reference Leucht, Samara and Heres61]. The leading author of these studies is Stefan Leucht, a professor of psychiatry at Technische Universität in Munich, and his many papers examine a variety of means to express antipsychotic equivalent dosages, including the World Health Organization defined daily dose (DDD) method, and consensus guidelines [Reference Leucht, Samara and Heres61Reference Rothe, Heres and Leucht63, Reference Takeuchi, MacKenzie and Samaroo24]. MED is clinically useful as it is the lowest dose that significantly differs from placebo in a double-blind randomized trial. MED thus serves as a point of reference for an initial dose, and can be used to guide dosing based on an international consensus study which recommended that an antipsychotic dose for acute schizophrenia should be in the range of 2–3 times the MED [Reference Gardner, Murphy and O’Donnell64]. From MED estimates, one can calculate the corresponding plasma level for any antipsychotic dose from known data on plasma concentration–dose relationships.

MED estimates have a significant amount of clinical value, but the approach is quite different than ROC analyses of responder and nonresponders drawn directly from the plasma level data. Any method which uses dosing cannot account for low rates of adherence with oral antipsychotics, and population variations in drug metabolism due to genetic or environmental exposures (e.g. inhibitors/inducers, smoking, etc.). However, plasma levels for certain antipsychotics are often not available due to the lack of qualified laboratories performing the assay locally, or the fact that the medication is very new (e.g. cariprazine, lumateperone), and there is not as yet widespread demand to develop a commercial laboratory assay. In those instances, the MED is extremely helpful and clinicians can read the publications employing this method to arrive at clinically relevant doses for oral and long-acting injectable antipsychotics [Reference Rothe, Heres and Leucht63, Reference Leucht, Crippa and Siafis65].

B Evidence Used to Define an Upper Limit

Increased systemic antipsychotic exposure not only increases the likelihood of adverse effects, this often comes with little promise of increased efficacy [Reference Barnes, Drake and Paton14]. It is due to the compelling need to balance efficacy and tolerability that the British Association for Psychopharmacology (BAP) 2020 guidelines suggest aiming for doses that are effective in 50% of the population [Reference Barnes, Drake and Paton14]. The BAP guidelines also note that: “There is little evidence to support exceeding British National Formulary (BNF) maximum recommended doses or even maximum effective doses identified in reviews and meta-analyses of trials, and certainly not unless blood concentrations are unexpectedly low on testing.” Yet the BAP also acknowledges that outliers do exist, and comments that for cases where the risk of continuing illness is high, clinicians might consider doses effective in 90% of the population. The question confronting societies and clinicians is where to draw a line based on plasma levels that acknowledges that the spectrum of tolerability and response might necessarily need to incorporate outliers. ROC analyses performed to estimate a response threshold only demand that one come to a definition of response (e.g. 50% symptom reduction) – from this response definition, the ROC result is immediately understandable [Reference Samara, Leucht and Leeflang27]. Depending on how one wishes to weight tolerability or prioritize the inclusion of outliers on the response curve, a myriad of possible cutoffs can be generated to justify an upper limit.

Unfortunately, the numerous suggested upper limits for antipsychotic levels are at times confusing, and often do not make explicit the reasoning regarding trade-offs between decreasing tolerability and incremental response. (What laboratories choose to report as their upper limit is even more problematic, and is discussed extensively in Chapter 3). There are also varying rationales about the weighting of imaging data in this process, and whether the upper limit is based on regulations that preclude exceeding licensed dose maxima. The difficulties encountered by experts in this area reflect these issues. As noted previously, the AGNP/ASCP consensus paper represents the most current well-reasoned source on use of plasma antipsychotic levels, drawing heavily on prior extensive work from AGNP [Reference Hiemke, Bergemann and Clement13, Reference Schoretsanitis, Kane and Correll15]. The suggested upper limit for haloperidol is 10 ng/ml, a level where slightly more than 50% might experience intolerable adverse effects [Reference Van Putten, Marder and Wirshing9], yet the recommended upper limit for fluphenazine, another potent D2 antagonist, is 10 ng/ml, a level that presumably would be intolerable to > 95% of patients if one extrapolates from available data (Figure 2.7) [Reference Midha, Hubbard and Marder16]. That thoughtful investigators struggle with defining plasma level upper limits reflects the myriad of inputs involved, and the limitations of existing clinical data. Clinicians should appreciate the sources of information that go into these decisions, and the evolution of a concept elaborated for treatment of more severely ill inpatients, the point of futility, which provides a clear rationale for a plasma level choice [Reference Meyer12].

Figure 2.7 Plasma fluphenazine levels and proportion of patients with response (dotted line) or intolerable adverse effects (solid line) [Reference Midha, Hubbard and Marder16]

(Adapted from: K. K. Midha, J. W. Hubbard, S. R. Marder, et al. [1994]. Impact of clinical pharmacokinetics on neuroleptic therapy in patients with schizophrenia. J Psychiatry Neurosci, 19, 254–264.)

1 Upper Limits Based on Maximum Approved Dosages for Schizophrenia

The recommendations from BAP and AGNP acknowledge that in certain countries, or in certain practice locations, there are strict limits placed on the highest dose that can be used for a particular antipsychotic derived from what is listed in the approved package insert for schizophrenia treatment [Reference Barnes, Drake and Paton14, Reference Schoretsanitis, Kane and Correll15]. Such decisions represent the compelling interest that national or local health agencies have in limiting unsafe practices, and the choice to enact these policies in a method that applies equally to all antipsychotics, and uses information readily available to every clinician. While the AGNP/ASCP consensus papers describe upper limits estimated from the clinical literature, clinicians must be mindful that certain reviews and laboratory ranges may be based solely on the licensed dosage range for that country, despite abundant subsequent data demonstrating the safety and utility of higher dosages (and higher corresponding plasma levels). This may explain why the upper limits expressed in national or hospital policy documents differ greatly from those espoused in this handbook or the AGNP/ASCP review.

2 Upper Limits Based on PET Imaging of D2 Occupancy

For antipsychotics whose primary mechanism relies on D2 antagonism, exceeding 78%–80% D2 occupancy incurs greater risk for neurological adverse effects, although there are clearly patients who can tolerate > 90% D2 occupancy [Reference Kapur, Remington and Zipursky66, Reference Nord and Farde67]. In addition to the more overt neurological manifestations, > 80% D2 blockade is also associated with decreased subjective well-being for D2 antagonist antipsychotics [Reference de Haan, Lavalaye and Linszen68, Reference de Haan, van Bruggen and Lavalaye37, Reference Lataster, van Os and de Haan69, 70]. As noted above, the partial agonists aripiprazole and cariprazine operate between 80% and 100% D2 occupancy, so this relationship with subjective well-being and D2 occupancy is not found [Reference Mizrahi, Mamo and Rusjan45]. As the risk for parkinsonism, akathisia, and dystonia emerges at 80% D2 blockade, this does not imply that the majority of patients at that threshold are experiencing those adverse effects. An early PET imaging study noted that a daily risperidone dose between 6 and 8 mg crossed the 80% occupancy threshold (Figure 2.4) [Reference Kapur and Seeman38], and a subsequent study by that same group found that 80% D2 blockade corresponded to an active moiety level (risperidone + 9-OH risperidone) of 45 ng/ml, equivalent to 6.43 mg/d of oral risperidone [Reference Remington, Mamo and Labelle71]. However, in the pivotal risperidone trial, only 20% required antiparkinsonian medication at 6 mg/d and 31% at 10 mg/d [Reference Marder and Meibach42]. A decision to use a plasma level that generates 80% D2 blockade as the upper limit would seem overly conservative, but the choice of higher occupancy levels involves the same judgment decision alluded to above: how much D2 antagonism is it acceptable to recommend so that a small subgroup of patients are given access to high levels that might be intolerable for most. There is no universal agreement on this point, but the understanding of D2 occupancy–plasma level relationships for many antipsychotics helps inform the discussion about upper limits. This information does not apply to medications whose antipsychotic response is not dependent on > 60% D2 antagonism (e.g. aripiprazole, cariprazine, clozapine, lumateperone, quetiapine).

3 95% Effective Dose

The BAP recommendations to consider doses effective in 90% of patients for select cases involves both the decision to use the 90% threshold (as opposed to 80%, 85%, etc.), and acknowledgment of the reality that some patients both respond to and tolerate high plasma antipsychotic levels [Reference Barnes, Drake and Paton14]. The classic graphic on fluphenazine response and adverse effects (Figure 2.7) indicates that, at the level where 90% respond, a nearly equal proportion will experience intolerable adverse effects [Reference Midha, Hubbard and Marder16]; however, this is grouped data that may not apply to specific individuals, a finding described in the literature for over 50 years [Reference Simpson and Kunz-Bartholini72]. Using the logic that clinical trials data might provide insight into dosage ranges effective in the majority of chronic schizophrenia patients with acute exacerbation, Stefan Leucht published a meta-analysis in 2020 of 68 placebo-controlled dose-finding studies for oral and long-acting injectable second-generation antipsychotics and for haloperidol [Reference Leucht, Crippa and Siafis65]. Based on changes in BPRS or PANSS total scores, doses that produced 95% of the maximum symptom reduction were identified, some examples of which are: haloperidol 6.3 mg/d, olanzapine 15.2 mg/d, and risperidone 6.3 mg/d. The limits of using dose to define drug exposure have been discussed previously, but this type of analysis is quite useful for examining trials data in a clinically relatable manner. The risperidone result of 6.3 mg maps very closely to the plasma level associated with 80% D2 occupancy, but the olanzapine dose of 15.2 mg falls under that threshold, and the expected plasma olanzapine level from 15 mg/d (30 ng/ml) is well below the upper limit of 80 ng/ml suggested by the AGNP/ASCP review [Reference Kapur and Seeman38, Reference Meyer18, Reference Schoretsanitis, Kane and Correll15].

That the conclusions from this analysis vary from other sources does not represent a methodological flaw, but differences between this type of analysis and a responder analysis. This method does not identify the dose at which 95% of patients might respond (with response defined as 30%–50% symptom reduction), but the dose that, on average, is expected to achieve 95% of the possible symptom reduction seen for that antipsychotic. Providing another source of upper limits is important to the evolving literature in this area, but may be confusing to clinicians whose basic question revolves around the maximum plasma level one can pursue in search of adequate response. This meta-analysis also relies heavily on registrational trial data. Industry-sponsored studies typically screen out individuals with certain medical disorders, substance use, and other psychiatric comorbidities, and thus utilize a subgroup of schizophrenia patients who may not be representative of a more general schizophrenia population. For example, the oral olanzapine calculations are based on two studies [Reference Leucht, Crippa and Siafis65].

C The Point of Futility

The multiple definitions of upper limits can be confusing to clinicians who are not intimately familiar with the information each source used. This can lead to situations where providers conflate maximal levels based on responder analyses with imminent safety issues, leading to erroneous conclusions that a level beyond the upper limit must be dangerous and should be reduced immediately [Reference Meyer18]. This distinction between an upper limit for response and issues of possible safety is embodied in the AGNP history of separating therapeutic reference ranges from laboratory levels. In discussing this issue, AGNP clearly states that for most psychotropics such as antipsychotics “concentrations in blood with an increased risk of toxicity are normally much higher than the upper threshold levels of the therapeutic reference ranges” [Reference Hiemke, Bergemann and Clement13]. In doing so, they conceptualized a level that would mandate laboratory notification of the prescriber – the laboratory alert level. For some psychotropic medications with narrow therapeutic indices (e.g. lithium), these laboratory alert levels are derived from case reports of drug intoxication, but in most instances the alert level was arbitrarily defined as a drug concentration in blood that is two-fold higher than the upper limit of the therapeutic reference range [Reference Hiemke, Bergemann and Clement13]. The underlying method is thus clearly spelt out, and this information is included in the AGNP/ASCP consensus paper [Reference Schoretsanitis, Kane and Correll15].

There are some practical issues for implementation of this dual system of reference ranges and laboratory alert levels:
  1. a Laboratory reports generally only contain one range (see Chapter 3), and there has not been acceptance in many countries of this dual limit system, or from laboratories themselves to incorporate the recommended alert levels.

  2. b While laboratories may never adopt the AGNP system, large hospital systems could potentially make this information available internally. As noted previously, having two different limits (an upper limit for the therapeutic range and a laboratory alert level) has the potential to engender confusion, as noted above; moreover, the concept that reported laboratory plasma level ranges for antipsychotics may not represent an imminent safety issue is difficult for most clinicians to accept, often leading to inappropriate sizable dosage decreases in asymptomatic patients [Reference Meyer18]. The concern is that, when presented with two systems, clinicians will practice based on one set of values (typically the lower value reported as the upper limit of the therapeutic range) since this is most consistent with what current laboratory reports include.

As a means of simplifying clinical decision making around these upper limits, the term point of futility was created within the California Department of State Hospitals (Cal-DSH) system as a term of art [Reference Meyer12]. Cal-DSH has five campuses with over 6500 patients, 90% of whom have schizophrenia spectrum disorders, with care provided by approximately 200 psychiatrists. Given the varying levels of expertise among the psychiatrists, the term point of futility was developed to educate clinicians about two important concepts with regards to an upper limit for antipsychotic plasma levels [Reference Meyer12]:

  1. a While intolerability is a hard endpoint signaling the end of an antipsychotic trial, a small proportion of patients may never exhibit dose-limiting adverse effects and will tolerate further antipsychotic titration, even with high potency first-generation antipsychotics [Reference Simpson and Kunz-Bartholini72]. Although a plasma fluphenazine level of 4.0 ng/ml is intolerable to > 90% of the population [Reference Midha, Hubbard and Marder16], levels ranging from 6.13 to 16.19 ng/ml were documented in early PET studies, and one trial reported a subject with a level of 27.9 ng/ml who was not taking medication for parkinsonism or for akathisia [Reference Miller, Peterson and McLean73, Reference Nyberg, Dencker and Malm74].

  2. b Although intolerability may not limit the drug trial in some instances, ongoing titration beyond a certain plasma level (the point of futility) is fruitless as a minuscule fraction of patients will respond to these higher plasma levels [Reference Midha, Hubbard and Marder16, Reference Meyer12]. For example, the compelling reason why a fluphenazine level of 4.0 ng/ml is viewed as a point of futility is that the probability of response in those patients who can tolerate a plasma level of 4.0 ng/ml is < 5%, as illustrated elegantly in Figure 2.8 [Reference Midha, Hubbard and Marder16, Reference Meyer18].

Figure 2.8 Estimated probability of improvement in the absence of disabling adverse effects based on plasma fluphenazine level [Reference Midha, Hubbard and Marder16]

(Adapted from: K. K. Midha, J. W. Hubbard, S. R. Marder, et al. [1994]. Impact of clinical pharmacokinetics on neuroleptic therapy in patients with schizophrenia. J Psychiatry Neurosci, 19, 254–264.)

The latter point is crucial, as clinicians are sometimes wary of starting clozapine and will instead continue to hope that ‘more is better’ by pursuing seemingly heroic plasma levels of other antipsychotics to achieve schizophrenia response [Reference Cohen75, Reference Gee, Vergunst and Howes76]. The concept of a point of futility also permits the clinician to have a single reference for an upper limit which suffices as both the top of the therapeutic range and an informal alert level demanding that the clinician assess the situation and examine the patient for adverse effects [Reference Meyer18]. While the proportion of responders in the interval between the upper limit of the AGNP/ASCP therapeutic range and the alert level is expected to be small, for the management of more severely ill inpatients, that additional yield may be important. If it is deemed reasonable to have a clozapine laboratory alert level of 1000 ng/ml, placing the top of the therapeutic range at 600 ng/ml might dissuade some psychiatrists from pursuing levels of 601–999 ng/ml that could benefit inadequate responders to lower plasma levels [Reference Remington, Agid and Foussias19].

The point of futility largely overlaps with the laboratory alert level, but not in every instance. For example, there appears to be no safety concern to mandate a laboratory alert until fluphenazine levels reach 15 ng/ml [Reference Hiemke, Bergemann and Clement13], yet in clinical trials there are vanishingly few responders at levels above 4.0 ng/ml, as noted in Figures 2.7 and 2.8 and documented elsewhere, so the value of 4.0 ng/ml has been used as the Cal-DSH point of futility for fluphenazine [Reference Midha, Hubbard and Marder16, Reference Meyer12]. Throughout this handbook, a justification will be provided for each therapeutic threshold and point of futility estimate, along with the AGNP/ASCP laboratory alert level for comparison. These terminologies represent thoughtful approaches by various groups on how to best educate clinicians about plasma antipsychotic levels, and provide the information in an accessible form to optimize schizophrenia treatment. By outlining the basic evidence supporting a choice of threshold level or point of futility, it is hoped that clinicians will be better able to make sense of conflicting ranges from laboratory reports and other sources, to decide the best course of action for their patient.

Summary Points

  1. a The best sources of information to estimate a therapeutic threshold are ROC curves from multiple fixed-dose schizophrenia studies with plasma level data on responders and nonresponders at various levels. Other types of information (e.g. PET imaging of D2 occupancy, inference from minimally effective doses) is supportive, but not as robust as the clinically based data.

  2. b What laboratories report as the upper limit of the therapeutic range varies dramatically between sources. The AGNP/ASCP laboratory alert level is a useful construct, although in most instances the alert level was arbitrarily defined as a plasma level two-fold higher than the upper limit of the therapeutic reference range. Nonetheless, it represents a threshold demanding that the clinician be alerted by the laboratory to assess the situation and examine the patient for adverse effects.

  3. c The point of futility is a term devised to provide clinicians a single upper limit value, and to educate clinicians about two important concepts: (1) a small proportion of patients may never exhibit dose-limiting adverse effects and will tolerate further titration; (2) ongoing titration beyond a certain plasma level (the point of futility) is fruitless as < 5% of patients may respond to these higher plasma levels.

References

Willcox, D. R., Gillan, R., and Hare, E. H. (1965). Do psychiatric out-patients take their drugs? BMJ, 2, 790792.
Curry, S. H., Davis, J. M., Janowsky, D. S., et al. (1970). Factors affecting chlorpromazine plasma levels in psychiatric patients. Arch Gen Psychiatry, 22, 209215.
Overall, J. E. and Gorham, D. R. (1962). The Brief Psychiatric Rating Scale. Psychological Reports, 10, 799812.
Curry, S. H. (1968). The determination and possible significance of plasma levels of chlorpromazine in psychiatric patients. Agressologie, 9, 115121.
Curry, S. H. and Marshall, J. H. (1968). Plasma levels of chlorpromazine and some of its relatively non-polar metabolites in psychiatric patients. Life Sci, 7, 917.
Kay, S. R., Fiszbein, A., and Opler, L. A. (1987). The positive and negative syndrome scale (PANSS) for schizophrenia. Schizophr Bull, 13, 261276.
Sakalis, G., Curry, S. H., Mould, G. P., et al. (1972). Physiologic and clinical effects of chlorpromazine and their relationship to plasma level. Clin Pharmacol Ther, 13, 931946.
Van Putten, T., May, P. R., and Jenden, D. J. (1981). Does a plasma level of chlorpromazine help? Psychol Med, 11, 729734.
Van Putten, T., Marder, S. R., Wirshing, W. C., et al. (1991). Neuroleptic plasma levels. Schizophr Bull, 17, 197216.
Nielsen, J., Graff, C., Kanters, J. K., et al. (2011). Assessing QT prolongation of antipsychotic drugs. CNS Drugs, 25, 473490.
Van Putten, T., Aravagiri, M., Marder, S. R., et al. (1991). Plasma fluphenazine levels and clinical response in newly admitted schizophrenic patients. Psychopharmacol Bull, 27, 9196.
Meyer, J. M. (2014). A rational approach to employing high plasma levels of antipsychotics for violence associated with schizophrenia: case vignettes. CNS Spectr, 19, 432438.
Hiemke, C., Bergemann, N., Clement, H. W., et al. (2018). Consensus guidelines for therapeutic drug monitoring in neuropsychopharmacology: update 2017. Pharmacopsychiatry, 51, 962.
Barnes, T. R., Drake, R., Paton, C., et al. (2020). Evidence-based guidelines for the pharmacological treatment of schizophrenia: Updated recommendations from the British Association for Psychopharmacology. J Psychopharmacol, 34, 378.
Schoretsanitis, G., Kane, J. M., Correll, C. U., et al. (2020). Blood levels to optimize antipsychotic treatment in clinical practice; a joint consensus statement of the American Society of Clinical Psychopharmacology (ASCP) and the Therapeutic Drug Monitoring (TDM) Task Force of the Arbeitsgemeinschaft für Neuropsychopharmakologie und Pharmakopsychiatrie (AGNP). J Clin Psychiatry, 81, https://doi.org/10.4088/JCP.4019cs13169.
Midha, K. K., Hubbard, J. W., Marder, S. R., et al. (1994). Impact of clinical pharmacokinetics on neuroleptic therapy in patients with schizophrenia. J Psychiatry Neurosci, 19, 254264.
Marder, S. R., Davis, J. M., and Janicak, P. G., eds. (1993). Clinical Use of Neuroleptic Plasma Levels. Washington, DC: American Psychiatric Press Inc.
Meyer, J. M. (2020). Monitoring and improving antipsychotic adherence in outpatient forensic diversion programs. CNS Spectr, 25, 136144.
Remington, G., Agid, O., Foussias, G., et al. (2013). Clozapine and therapeutic drug monitoring: Is there sufficient evidence for an upper threshold? Psychopharmacology (Berl), 225, 505518.
Lopez, L. V. and Kane, J. M. (2013). Plasma levels of second-generation antipsychotics and clinical response in acute psychosis: A review of the literature. Schizophr Res, 147, 368374.
Grandjean, E. M. and Aubry, J.-M. (2009). Lithium: Updated human knowledge using an evidence-based approach. Part II: Clinical pharmacology and therapeutic monitoring. CNS Drugs, 23, 331349.
Machado-Vieira, R., Otaduy, M. C., Zanetti, M. V., et al. (2016). A selective association between central and peripheral lithium levels in remitters in bipolar depression: A 3T-(7) Li magnetic resonance spectroscopy study. Acta Psychiatr Scand, 133, 214220.
Uchida, H., Takeuchi, H., Graff-Guerrero, A., et al. (2011). Predicting dopamine D2 receptor occupancy from plasma levels of antipsychotic drugs: A systematic review and pooled analysis. J Clin Psychopharmacol, 31, 318325.
Takeuchi, H., MacKenzie, N. E., Samaroo, D., et al. (2020). Antipsychotic dose in acute schizophrenia: A meta-analysis. Schizophr Bull, 46, 1439–1458.
Melkote, R., Singh, A., Vermeulen, A., et al. (2018). Relationship between antipsychotic blood levels and treatment failure during the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study. Schizophr Res, 201, 324328.
Correll, C. U., Jain, R., Meyer, J. M., et al. (2019). Relationship between the timing of relapse and plasma drug levels following discontinuation of cariprazine treatment in patients with schizophrenia: Indirect comparison with other second-generation antipsychotics after treatment discontinuation. Neuropsychiatr Dis Treat, 15, 25372550.
Samara, M. T., Leucht, C., Leeflang, M. M., et al. (2015). Early improvement as a predictor of later response to antipsychotics in schizophrenia: A diagnostic test review. Am J Psychiatry, 172, 617629.
Meyer, J. M. and Stahl, S. M. (2019). The Clozapine Handbook. Cambridge: Cambridge University Press.
Van Putten, T., Marder, S. R., and Mintz, J. (1985). Plasma haloperidol levels: Clinical response and fancy mathematics. Arch Gen Psychiatry, 42, 835838.
Perry, P. J., Miller, D. D., Arndt, S. V., et al. (1991). Clozapine and norclozapine plasma concentrations and clinical response of treatment-refractory schizophrenic patients. Am J Psychiatry, 148, 231235.
Kronig, M. H., Munne, R. A., Szymanski, S., et al. (1995). Plasma clozapine levels and clinical response for treatment-refractory schizophrenic patients. Am J Psychiatry, 152, 179182.
Lu, M. L., Wu, Y. X., Chen, C. H., et al. (2016). Application of plasma levels of olanzapine and N-desmethyl-olanzapine to monitor clinical efficacy in patients with schizophrenia. PLoS One, 11, e0148539.
Meyer, J. M. (2018). Pharmacotherapy of psychosis and mania. In Brunton, L. L., Hilal-Dandan, R., and Knollmann, B. C., eds., Goodman & Gilman’s The Pharmacological Basis of Therapeutics, 13th edn. Chicago, IL: McGraw-Hill, pp. 279302.
Farde, L., Hall, H., Ehrin, E., et al. (1986). Quantitative analysis of D2 dopamine receptor binding in the living human brain by PET. Science, 231, 258261.
Kirino, S., Suzuki, T., Takeuchi, H., et al. (2017). Representativeness of clinical PET study participants with schizophrenia: A systematic review. J Psychiatr Res, 88, 7279.
Arakawa, R., Takano, A., and Halldin, C. (2020). PET technology for drug development in psychiatry. Neuropsychopharmacol Rep, 40, 114121.
de Haan, L., van Bruggen, M., Lavalaye, J., et al. (2003). Subjective experience and D2 receptor occupancy in patients with recent-onset schizophrenia treated with low-dose olanzapine or haloperidol: a randomized, double-blind study. Am J Psychiatry, 160, 303309.
Kapur, S. and Seeman, P. (2001). Does fast dissociation from the dopamine d(2) receptor explain the action of atypical antipsychotics? A new hypothesis. Am J Psychiatry, 158, 360369.
Perry, P. J., Lund, B. C., Sanger, T., et al. (2001). Olanzapine plasma concentrations and clinical response: Acute phase results of the North American Olanzapine Trial. J Clin Psychopharmacol, 21, 1420.
Preskorn, S. H. (2010). Outliers on the dose–response curve: How to minimize this problem using therapeutic drug monitoring, an underutilized tool in psychiatry. J Psychiatr Pract, 16, 177182.
Kapur, S., Zipursky, R., Jones, C., et al. (2000). Relationship between dopamine D(2) occupancy, clinical response, and side effects: A double-blind PET study of first-episode schizophrenia. Am J Psychiatry, 157, 514520.
Marder, S. R. and Meibach, R. C. (1994). Risperidone in the treatment of schizophrenia. Am J Psychiatry, 151, 825835.
Tadori, Y., Forbes, R. A., McQuade, R. D., et al. (2011). In vitro pharmacology of aripiprazole, its metabolite and experimental dopamine partial agonists at human dopamine D2 and D3 receptors. Eur J Pharmacol, 668, 355365.
Maeda, K., Sugino, H., Akazawa, H., et al. (2014). Brexpiprazole I: In vitro and in vivo characterization of a novel serotonin-dopamine activity modulator. J Pharmacol Exp Ther, 350, 589604.
Mizrahi, R., Mamo, D., Rusjan, P., et al. (2009). The relationship between subjective well-being and dopamine D2 receptors in patients treated with a dopamine partial agonist and full antagonist antipsychotics. Int J Neuropsychopharmacol, 12, 715721.
Sparshatt, A., Taylor, D., Patel, M. X., et al. (2010). A systematic review of aripiprazole – dose, plasma concentration, receptor occupancy, and response: Implications for therapeutic drug monitoring. J Clin Psychiatry, 71, 14471556.
Girgis, R. R., Slifstein, M., D’Souza, D., et al. (2016). Preferential binding to dopamine D3 over D2 receptors by cariprazine in patients with schizophrenia using PET with the D3/D2 receptor ligand [(11)C]-(+)-PHNO. Psychopharmacology (Berl), 233, 35033512.
Girgis, R. R., Forbes, A., Abi-Dargham, A., et al. (2020). A positron emission tomography occupancy study of brexpiprazole at dopamine D2 and D3 and serotonin 5-HT1A and 5-HT2A receptors, and serotonin reuptake transporters in subjects with schizophrenia. Neuropsychopharmacology, 45, 786792.
Mamo, D., Graff, A., Mizrahi, R., et al. (2007). Differential effects of aripiprazole on D(2), 5-HT(2), and 5-HT(1A) receptor occupancy in patients with schizophrenia: A triple tracer PET study. Am J Psychiatry, 164, 14111417.
Wiesel, F. A., Farde, L., Nordstrom, A. L., et al. (1990). Central D1- and D2-receptor occupancy during antipsychotic drug treatment. Prog Neuro-Psychopharmacol Biol Psychiatry, 14, 759767.
Gefvert, O., Bergstrom, M., Langstrom, B., et al. (1998). Time course of central nervous dopamine-D2 and 5-HT2 receptor blockade and plasma drug concentrations after discontinuation of quetiapine (Seroquel) in patients with schizophrenia. Psychopharmacology, 135, 119126.
Vanover, K. E., Davis, R. E., Zhou, Y., et al. (2019). Dopamine D2 receptor occupancy of lumateperone (ITI-007): A Positron Emission Tomography Study in patients with schizophrenia. Neuropsychopharmacology, 44, 598605.
Meltzer, H. Y. and Roth, B. L. (2013). Lorcaserin and pimavanserin: Emerging selectivity of serotonin receptor subtype-targeted drugs. J Clin Invest, 123, 49864991.
Stahl, S. M. (2016). Mechanism of action of pimavanserin in Parkinson’s disease psychosis: Targeting serotonin 5HT2A and 5HT2C receptors. CNS Spectr, 21, 271275.
Shekhar, A., Potter, W. Z., Lightfoot, J., et al. (2008). Selective muscarinic receptor agonist xanomeline as a novel treatment approach for schizophrenia. Am J Psychiatry, 165, 10331039.
Barak, S. and Weiner, I. (2011). The M1/M4 preferring agonist xanomeline reverses amphetamine-, MK801- and scopolamine-induced abnormalities of latent inhibition: Putative efficacy against positive, negative and cognitive symptoms in schizophrenia. Int J Neuropsychopharmacol, 14, 12331246.
Simmler, L. D., Buchy, D., Chaboz, S., et al. (2016). In vitro characterization of psychoactive substances at rat, mouse, and human trace amine-associated receptor 1. J Pharmacol Exp Ther, 357, 134144.
Dedic, N., Jones, P. G., Hopkins, S. C., et al. (2019). SEP-363856, a novel psychotropic agent with a unique, non-D2 receptor mechanism of action. J Pharmacol Exp Ther, 371, 114.
Cohen, W. J. and Cohen, N. H. (1974). Lithium carbonate, haloperidol, and irreversible brain damage. JAMA, 230, 12831287.
Lieberman, J. A., Stroup, T. S., McEvoy, J. P., et al. (2005). Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med, 353, 12091223.
Leucht, S., Samara, M., Heres, S., et al. (2014). Dose equivalents for second-generation antipsychotics: The minimum effective dose method. Schizophr Bull, 40, 314326.
Leucht, S., Samara, M., Heres, S., et al. (2016). Dose equivalents for antipsychotic drugs: The DDD method. Schizophr Bull, 42 Suppl 1, S9094.
Rothe, P. H., Heres, S., and Leucht, S. (2018). Dose equivalents for second generation long-acting injectable antipsychotics: The minimum effective dose method. Schizophr Res, 193, 2328.
Gardner, D. M., Murphy, A. L., O’Donnell, H., et al. (2010). International consensus study of antipsychotic dosing. Am J Psychiatry, 167, 686693.
Leucht, S., Crippa, A., Siafis, S., et al. (2020). Dose-response meta-analysis of antipsychotic drugs for acute schizophrenia. Am J Psychiatry, 177, 342353.
Kapur, S., Remington, G., Zipursky, R. B., et al. (1995). The D2 dopamine receptor occupancy of risperidone and its relationship to extrapyramidal symptoms: A PET study. Life Sci, 57, PL103107.
Nord, M. and Farde, L. (2011). Antipsychotic occupancy of dopamine receptors in schizophrenia. CNS Neurosci Ther, 17, 97103.
de Haan, L., Lavalaye, J., Linszen, D., et al. (2000). Subjective experience and striatal dopamine D(2) receptor occupancy in patients with schizophrenia stabilized by olanzapine or risperidone. Am J Psychiatry, 157, 10191020.
Lataster, J., van Os, J., de Haan, L., et al. (2011). Emotional experience and estimates of D2 receptor occupancy in psychotic patients treated with haloperidol, risperidone, or olanzapine: An experience sampling study. J Clin Psychiatry, 72, 13971404.
Veselinovicć, T., Scharpenberg, M., Heinze, M., et al. (2019). Dopamine D2 receptor occupancy estimated from plasma concentrations of four different antipsychotics and the subjective experience of physical and mental well-being in schizophrenia: Results from the randomized NeSSy Trial. J Clin Psychopharmacol, 39, 550560.
Remington, G., Mamo, D., Labelle, A., et al. (2006). A PET study evaluating dopamine D2 receptor occupancy for long-acting injectable risperidone. Am J Psychiatry, 163, 396401.
Simpson, G. M. and Kunz-Bartholini, E. (1968). Relationship of individual tolerance, behavior and phenothiazine produced extrapyramidal system disturbance. Dis Nerv System, 29, 269274.
Miller, R. S., Peterson, G. M., McLean, S., et al. (1995). Monitoring plasma levels of fluphenazine during chronic therapy with fluphenazine decanoate. J Clin Pharm Ther, 20, 5562.
Nyberg, S., Dencker, S. J., Malm, U., et al. (1998). D(2)- and 5-HT(2) receptor occupancy in high-dose neuroleptic-treated patients. Int J Neuropsychopharmacol, 1, 95101.
Cohen, D. (2014). Prescribers fear as a major side-effect of clozapine. Acta Psychiatr Scand, 130, 154155.
Gee, S., Vergunst, F., Howes, O., et al. (2014). Practitioner attitudes to clozapine initiation. Acta Psychiatr Scand, 130, 1624.
Figure 0

Figure 2.1 Plasma haloperidol levels and the proportion of patients with intolerable adverse effects [9, 16]

(Adapted from: K. K. Midha, J. W. Hubbard, S. Marder, et al. [1994]. Impact of clinical pharmacokinetics on neuroleptic therapy in patients with schizophrenia. J Psychiatry Neurosci, 19, 254–264.)
Figure 1

Figure 2.2 Receiver operating characteristic (ROC) curve from 1995 data on the proportion of clozapine responders and nonresponders for plasma clozapine levels in 50 ng/ml increments plotted from the data in Table 2.1 [31]

(Adapted from: M. H. Kronig, R. A. Munne, S. Szymanski, et al. [1995]. Plasma clozapine levels and clinical response for treatment-refractory schizophrenic patients. Am J Psychiatry, 152, 179–182.)
Figure 2

Table 2.1 Responders (sensitivity) and nonresponders (1-specificity) for plasma clozapine levels in 50 ng/ml increments [31]

Figure 3

Figure 2.3 2016 ROC curve of the plasma olanzapine level required to be rated as mildly ill (PANSS score ≤ 58) from a cohort of Taiwanese schizophrenia patients (n = 151) [32]

(Adapted from: M. L. Lu, Y. X. Wu, C. H. Chen, et al. [2016]. Application of plasma levels of olanzapine and N-desmethyl-olanzapine to monitor clinical efficacy in patients with schizophrenia. PLoS One, 11, e0148539.)
Figure 4

Figure 2.4 Adverse effects and response based on fitted D2 occupancy curves for D2 antagonist antipsychotics [38]

(Adapted from: S. Kapur and P. Seeman [2001]. Does fast dissociation from the dopamine d(2) receptor explain the action of atypical antipsychotics? A new hypothesis. Am J Psychiatry, 158, 360–369.)
Figure 5

Figure 2.5 Fitted D2 occupancy curve for aripiprazole in the dosage range of 10–30 mg/d [49]

(Adapted from: D. Mamo, A. Graff, R. Mizrahi, et al. [2007]. Differential effects of aripiprazole on D(2), 5-HT(2), and 5-HT(1A) receptor occupancy in patients with schizophrenia: a triple tracer PET study. Am J Psychiatry, 164, 1411–1417.)
Figure 6

Figure 2.6 Range of D2 occupancy for antipsychotics

Figure 7

Figure 2.7 Plasma fluphenazine levels and proportion of patients with response (dotted line) or intolerable adverse effects (solid line) [16]

(Adapted from: K. K. Midha, J. W. Hubbard, S. R. Marder, et al. [1994]. Impact of clinical pharmacokinetics on neuroleptic therapy in patients with schizophrenia. J Psychiatry Neurosci, 19, 254–264.)
Figure 8

Figure 2.8 Estimated probability of improvement in the absence of disabling adverse effects based on plasma fluphenazine level [16]

(Adapted from: K. K. Midha, J. W. Hubbard, S. R. Marder, et al. [1994]. Impact of clinical pharmacokinetics on neuroleptic therapy in patients with schizophrenia. J Psychiatry Neurosci, 19, 254–264.)
You have Access

Save book to Kindle

To save this book to your Kindle, first ensure coreplatform@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Available formats
×

Save book to Dropbox

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Dropbox.

Available formats
×

Save book to Google Drive

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Google Drive.

Available formats
×