Future of Depot Neuroleptic Therapy:
`Pharmacokinetic and Pharmacodynamic Approaches
`LARRY ERESHEFSKY, Pharm.D., STEPHEN R. SAKLAD, Pharm.D., MICHAEL W. JANN, Pharm.D.,
`CHESTER M. DAVIS, Ph.D., ANN RICHARDS, Pharm.D., and DONALD R. SEIDEL, M.D.
`
`The future of depot neuroleptic therapy is discussed
`in terms of pharmacokinetic and pharmacodynamic re-
`search opportunities. Analytic methods for neuroleptic
`assays, including chromatographic, radioreceptor, nu-
`clear magnetic resonance, and radioimmunoassay tech-
`niques, are briefly reviewed. Elucidation of depot
`neuroleptic multicompartment kinetics utilizing nonlin-
`ear mixed-effects modeling and the usefulness of these
`data in interpreting plasma levels are discussed. The
`clinical significance of plasma monitoring of depot
`fluphenazine, including the development of dosage con-
`version guidelines, is presented. The relationships be-
`tween haloperidol and its metabolite reduced
`haloperidol (RH) are discussed in terms of dosage form
`and response. Clinical advantages resulting from the
`availability of more depot neuroleptics are discussed.
`(J Clin Psychiatry 45 [5, Sec. 21:50-59, 1984)
`
`One of the most common reasons for the readmission of
`psychiatric patients to the hospital stems from the lack of
`medication compliance. The introduction of injectable de-
`pot neuroleptics in the early 1960s provided clinicians with
`a method of accurately assessing medication compliance in
`the mentally ill patient. Several studies have suggested that
`depot neuroleptics reduce the frequency of relapse or length
`of hospital stay when compared to oral agents." However,
`other investigators have reported that depot neuroleptics did
`not reduce the number of relapses among schizophrenics
`when compared with oral medications.' It is beyond the
`scope of this article to review all the comparative studies;
`the reader is referred to the articles by Kane and Johnson in
`this issue:6
`The major perceived use of depot neuroleptics currently
`is in cases of patient refusal or failure to take oral medica-
`tion as prescribed.' Injectable depot therapy bypasses ab-
`sorption variability, gut wall metabolism, and first pass
`extraction by the liver. This method of drug administration
`is useful with patients who absorb inadequate amounts of
`
`From the Departments of Pharmacology and Psychiatry, University of
`Texas Health Science Center at San Antonio; the College of Pharmacy,
`University of Texas at Austin; and San Antonio State Hospital and The
`Texas Research Institute for the Mental Sciences, Houston, TX.
`The authors acknowledge the contributions of Charles A. Harrington,
`Ph.D., and Jeffrey L Browning.
`Reprint requests to: Larry Ereshefsky, Pharm. D., Department of
`Pharmacology, University of Texas Health Science Center of San Antonio,
`7703 Floyd Curl Drive, San Antonio, TX 78284.
`
`50
`
`drug from oral therapy, resulting in subtherapeutic plasma
`levels.'
`Assessment of the action of drugs in psychiatric disor-
`ders is complicated because drug effects are only one of
`several inputs to the brain. Many environmental factors,
`including drugs, affect the brain's biochemical responses.
`Behavior is the outcome of a complex interplay of genetics,
`prior skills, events stored as memories, and biochemical
`factors. When a drug is added to this complicated set of
`interactions, there may be initially only a minor change in
`behavior. Substantial changes in behavior, however, can oc-
`cur over extended periods of time.
`The longitudinal course of drug response is partially de-
`pendent on individual pharmacodynamic and pharmaco-
`kinetic factors. Pharmacokinetic studies of drugs are
`needed to investigate how bioequivalence, therapeutic effi-
`cacy, and drug distribution characteristics influence the pa-
`tient's response. Failure to consider pharmacokinetics
`explains much of the disparity in results found among clini-
`cal studies. These principles and their applications will play
`an important role in the future use of long-acting neurolep-
`tic drugs.
`
`OVERVIEW OF DEPOT NEUROLEPTICS
`
`The synthesis of long-acting neuroleptics is accom-
`plished by an esterification of fluphenazine or other hy-
`droxyl-containing neuroleptics to a long-chain fatty acid.
`These esters are usually dissolved in sesame seed oil, coco-
`nut oil, or Viscoleo. Medications such as fluspirilene are
`not esterified, but are formulated as an aqueous microcrys-
`talline suspension with slow release properties.' The medi-
`cation, in oil or suspension, when injected into muscle,
`forms a reservoir of slowly released drug. The esterified
`compounds are reportedly taken up into fat stores in various
`body areas, except the brain and spinal cord, within days.'
`The drug is slowly and steadily released from the depot and
`hydrolyzed by plasma esterases to the parent compound,'
`although preliminary reports suggested that fluphenazine
`decanoate was released "erratically, in spurts," from the
`injection site.'
`At the present time, fluphenazine enanthate and deca-
`noate are the only long-acting depot neuroleptics available
`in the United States. An increase in the number of depot
`neuroleptics available will lead to potentially improved pa-
`tient care for the following reasons:
`1. Easier conversion from oral to depot medication without
`switching drug class.
`
`Mylan v. Janssen (IPR2020-00440) Ex. 1020 p. 001
`
`

`

`J CLIN PSYCHIATRY 45:5 (Sec. 2) MAY 1984
`
`PHARMACOKINETICS AND PHARMACODYNAMICS
`
`TABLE 1. Comparison of Depot Neuroleptics
`
`Neuroleptics
`Fluphenazine
`Fluphenazine
`Haloperidol
`Flupenthixol
`Clopenthixol
`Perphenazine
`Pipotiazine
`
`Fluspirilene
`
`Fatty Acid Chain
`Decanoate
`Enanthate
`Decanoate
`Decanoate or palmitate
`Decanoate
`Enanthate
`Palmitate
`Undecylenate
`None
`
`Vehicle
`Sesame oil
`Sesame oil
`Sesame oil
`Viscoleo
`Viscoleo
`Sesame oil
`Sesame oil
`Sesame oil
`Aqueous
`suspension
`
`Dosage and Interval
`12.5-100 mg, 1-4 weeks
`12.5-100 mg, 1-2 weeks
`20-400 mg, 4 weeks
`10-50 mg, 2-4 weeks
`50-600 mg, 1-4 weeks
`50-200 mg, 1-4 weeks
`25-600 mg, 4 weeks
`2-200 mg, 1-4 weeks
`2-30 mg, 1-4 weeks
`
`Plasma
`Levels
`0.5-3.0
`NA
`3.0-10.0
`0.5-2.0
`0.5-9.0
`2.0-10.0
`NA
`NA
`NA
`
`Single-Dose Kinetics
`Time to Peak
`lialfrUfe
`(daysl
`(days-
`0.3-1.5
`6-9
`2
`3.5-4.0
`3-9
`21
`1 1 -17
`17
`4-7
`19
`2-3
`3.5-4.5
`NA
`NA
`NA
`NA
`NA
`NA
`
`2. Flexible selection of depot agent to minimize side ef-
`fects.
`3. Increased selections for alternative therapies of the re-
`fractory patient.
`4. Decreased costs of depot therapy secondary to competi-
`tion in the marketplace.
`5. More widespread use of depot therapy, leading to in-
`creased clinical experience and to decreased morbidity
`and health care costs.
`Table 1 summarizes the studies examining the pharma-
`cokinetics of depot neuroleptics available worldwide. m-16
`Most of the clinical studies with the various depot neurolep-
`tics have compared their side effect profile to that of
`fluphenazine decanoate or simply reported the various sed-
`ative, autonomic, and extrapyramidal symptoms (EPS) oc-
`curring in the study population. Fluphenazine enanthate
`was reported to cause increased extrapyramidal and motor
`side effects compared to fluphenazine decanoate during a 7-
`month double-blind study with 50 schizophrenic outpa-
`tients." Haloperidol decanoate was reported to cause
`sedation in 4 of 38 patients.18 In a larger study of haloperi-
`dol decanoate involving 239 chronic schizophrenic pa-
`tients, 27% had autonomic side effects or EPS.19
`Fluspirilene and clopenthixol decanoate were compared to
`fluphenazine decanoate, and their sedative effects ("som-
`nolence," "drowsiness") were found to be similar.2°'21 In a
`report on 10 schizophrenics, 5 displayed akathisia and
`pseudoparkinsonism after administration of perphenazine
`enanthate. 22 Extrapyramidal symptoms with pipotiazine
`palmitate and undecylenate were reported to range between
`10% and 43% among the patients in the various published
`studies!'
`In several studies, all of the extrapyramidal side effects
`were adequately treated with antiparkinsonian agents.
`There were no reports of abnormal effects on liver, renal,
`or other organ systems, as measured by routine laboratory
`analysis. The most serious problem that can occur with a
`long-acting neuroleptic is the neuroleptic malignant syn-
`drome (NMS). Because of the apparently long half-life of
`the injected drug, it can be difficult to effectively treat pa-
`tients in whom this syndrome develops. Of the 12 deaths
`reported from NMS secondary to neuroleptics, 6 were in
`
`patients treated with fluphenazine decanoate." Finally, it
`has been reported that the relative occurrence of tardive
`dyskinesia is not specifically related to any particular
`neuroleptic agent.25 Careful epidemiologic studies to accu-
`rately assess the effects of drug product selection and
`plasma levels on the incidence of tardive dyskinesia have
`not been performed to date. Additional carefully controlled
`studies that consider pharmacokinetic and pharmacody-
`namic variables need to be performed as new depot agents
`become available.
`
`NEUROLEPTIC ASSAY
`
`A prerequisite to the study of the pharmacokinetics of
`neuroleptic agents is the ability to measure the clinically
`relevant subnanomolar quantities of drug in biological flu-
`ids. In general, the best correlation between drug at the site
`of action within the central nervous system and a peripheral
`measure is the free (unbound) level of drug within the
`plasma. Free drug levels are more useful than total drug
`levels because they measure the drug that is available to
`diffuse across the blood brain barrier to the site(s) of
`action.' Some studies suggest that models of the amount of
`drug within the central nervous system have produced bet-
`ter correlations when based on red blood cell levels.27-29
`Several low-affinity binding sites in plasma are located
`on albumin. Cholesterol and alpha-1-acid glycoprotein are
`high-affmity binding sites!'-'' Alpha-1-acid glycoprotein
`levels in plasma are important because of the dramatic in-
`crease in the concentration of this protein found in patients
`during stressful events!' Perhaps higher total levels of
`drug, with free levels relatively unchanged, are required
`during an acute stressful event to maintain a constant
`amount of drug at the site of action.
`Pharmacokinetic descriptions of neuroleptic drugs in-
`volve the use of multicompartment models. From the
`plasma, usually termed the central compartment, unbound
`drug may be actively or passively transported into
`"deeper" compartments (e.g., skeletal muscle, brain tis-
`sue, or fat). Until these deeper compartments equilibrate
`with the free drug in the central compartment, the medica-
`tion will not achieve steady-state concentrations at the site
`
`51
`
`Mylan v. Janssen (IPR2020-00440) Ex. 1020 p. 002
`
`

`

`J CLIN PSYCHIATRY 45:5 (Sec. 2) MAY 1984
`
`ERESHEFSKY ET AL.
`
`of action. Furthermore, when the administration of the drug
`is discontinued, these deep compartments will leach the ac-
`tive drug out through the central compartment and back to
`the site of action:4 Therefore, effects secondary to oral or
`depot neuroleptic therapy can persist for weeks and months
`after the discontinuation of routine therapy.
`As a general rule, three drug plasma levels must be ob-
`tained in a standard kinetic analysis to define each compart-
`ment. For a neuroleptic drug that has at least three
`significant compartments, nine levels drawn at appropriate
`times would be required to define one individual's pharma-
`cokinetic parameters.34 Analysis of the kinetics of depot
`neuroleptics is complicated, since prolonged studies are re-
`quired. One method of obtaining more information in a
`shorter time with fewer levels is to use a nonlinear mixed
`effects model (NONMEM) for analysis of the data. NON-
`MEM requires knowledge of how much drug was given,
`and the times of administration and plasma level sam-
`pling."'"
`Assays of the neuroleptics are usually performed either
`by physical detection of the drug by high performance liq-
`uid chromatography (HPLC),37'" gas liquid chromatogra-
`phy (GLC):94° high performance thin layer chromatogra-
`phy (HPTLC)," or by radioimmunoassay (RIA).38.42'43 An in
`vitro measure of neuroleptic activity can also be approx-
`imated by determining the total amount of dopamine-
`blocking activity by its competition with a radioactively la-
`belled ligand bound to dopamine receptors (radioreceptor
`assay, RRA):64446 Alternately, biological responses in vivo
`have been measured by assaying prolactin.""
`In addition to a reliable, sensitive, and specific assay
`method, blood collection procedures must be stringently
`controlled. Because these drugs are present in such low
`concentrations, small changes in the collection procedure
`may result in a dramatic change in the assay result. One
`example of a poorly controlled collection procedure is to
`draw a sample into a vacuum collection tube capped with a
`plastic stopper. These stoppers, however, contain a plastici-
`ser that causes redistribution of the drug into erythrocytes,
`which will be spun out of the sample when centrifuged. A
`second example is photo-oxidation of the drug by exposure
`to ultraviolet light prior to assay. Both examples result in
`lower drug levels measured than actually exist in the sam-
`ple.
`Once these sampling procedures have been perfected,
`reproducible separation of the drug from contaminants must
`be performed with a high extraction efficiency, monitored
`by the use of an internal standard. A clinically usable assay
`should have a coefficient of variation (both intra- and in-
`terassay) of less than the minimum change in plasma level
`that might yield a clinical effect (e.g., 10%-20%). Many
`methods of detection, such as electron capture, scintilla-
`tion, and nitrogen or fluorescent detection, have been em-
`ployed; the small quantity of drug present in clinical
`samples makes this a critical procedure. Our laboratory
`currently employs an HPTLC assay for fluphenazine sepa-
`
`52
`
`ration, followed by paraffin oil treatment of the chromatog-
`raphy plates to permit increased resonance within the
`molecule, then ultraviolet development and fluorescent
`scanning with computerized quantification." We have
`found this to be an extremely reliable and efficient method
`of detecting subnanomolar quantities of fluphenazine." An
`advantage of HPTLC over GLC or HPLC is that many
`samples may simultaneously be run on a single plate, pro-
`viding excellent speed and efficiency. An additional advan-
`tage is that HPTLC is a closed system and the plates may be
`stored for future reanalysis; in contrast, GLC and HPLC
`are open systems, in which the sample is lost following
`chromatographic separation and detection.'
`RIAs potentially offer the most efficient detection of
`drug, but there are still problems with specificity. The ap-
`propriate antibody must be produced to measure only the
`drug to be examined, and not the metabolites or other con-
`taminants in the sample. To do this, one must have a reli-
`able, sensitive, and specific assay (such as HPTLC or
`HPLC) to standardize the RIA.38'42'43
`The neuroleptic RRA, although a substantial improve-
`ment over many of the quantitative assays in terms of mea-
`surement of total biological effects, has numerous
`limitations. These include low sensitivity, nonspecific pro-
`tein binding limiting the sample aliquot, and the use of cau-
`date tissue instead of limbic tissue, the presumed site of
`neuroleptic action. Maximum sensitivity achieved with the
`RRA usually is > 1 ng/ml for neuroleptic agents. In the
`case of fluphenazine, because of its high affinity binding to
`dopamine receptors, a sensitivity of 0.5 ng/ml is achievable
`(personal communication, Charles L. Bowden, January 10,
`1983). This approximates the sensitivity of most quantita-
`tive analytic processes. Another approach, the measure-
`ment of prolactin increase as a function of the plasma
`concentration of a neuroleptic, has not provided consistent
`correlations with clinical effect.'" Further development of
`analytic techniques should allow us to clarify the role of
`neuroleptic metabolites and drug response:I'52
`Another area in which dramatic progress in assay tech-
`nology is being made is in the use of quantitative nuclear
`magnetic resonance (QNMR). QNMR is a technique using
`a strong magnetic field and a perpendicular radio frequency
`field to cause the fluorine nuclei in a neuroleptic to absorb
`energy quantifiably. QNMR may permit us to measure the
`drug at the site of action. This might be done by a device
`that looks similar to a computerized axial tomography
`(CAT) scanner. Refinement of this technique may permit
`quantification and detection of the molecular interaction at
`the site of action in the brain, e.g., protein, tissue, or recep-
`tor binding. Improvements in QNMR resolution might
`yield the capability to produce sectional images of the
`drug's distribution and binding within the body (personal
`communication from David Young, December 5, 1983)."
`Because of the specialization and sophistication of the
`personnel and equipment needed to perform these types of
`quantitative analysis, we envision regional analytic centers
`
`Mylan v. Janssen (IPR2020-00440) Ex. 1020 p. 003
`
`

`

`J CLIN PSYCHIATRY 45:5 (Sec. 2) MAY 1984
`
`PHARMACOKINETICS AND PHARMACODYNAMICS
`
`FIGURE 2. Idealized plasma level versus time curve for a single
`oral dose of a neuroleptic, illustrating absorption phase (a), dis-
`tribution into equilibrating compartments (b), metabolic elimi-
`nation (c), and leakage back into plasma from equilibrating
`compartments (d).
`
`d
`
`20
`
`25
`
`b
`
`a
`
`Log Plasma Concentration
`
`0
`
`5
`
`10
`15
`Time (Hours)
`
`rate of elimination. This "flip-flop" model is useful for the
`understanding and application of depot neuroleptic pharma-
`cokinetics to patient care. It also must be reemphasized that
`these highly lipophilic compounds do not possess simple
`one-compartment pharmacokinetic characteristics. Drugs
`that are highly ionized or hydrophilic usually follow a sin-
`gle-compartment kinetic model, where the kinetics can be
`characterized over the entire concentration time curve as a
`biexponential function (absorption and elimination). For
`depot neuroleptics and other lipophilic drugs, multicom-
`partment kinetic models are necessary. The measured decay
`phase half-life of neuroleptic agents is longer during studies
`performed at steady state (once plateau concentrations have
`been reached) than during single-dose studies. An early il-
`lustration of multicompartment effects was the description
`of chlorpromazine urinary metabolite detection after cessa-
`tion of chronic therapy for up to 6 months!'
`
`FIGURE 3. Detail of a patient's plasma level versus time curve
`following the injection of a single 25 mg dose of fluphenazine
`decanoate.
`
`1.5
`oo
`• 1.4
`• 1.3
`0
`
`o .9
`C.)
`•
`.8
`E
`.7
`g .6
`0
`ar
`.—c .4
`
`a)
`.c
`Q
`
`=8days
`
`10 20 30 40 50 60 70
`Time (Hours)
`
`I
`I
`I
`i
`r
`90 100 110 120 130
`
`53
`
`FIGURE 1. Pharmacokinetic variables that affect the amount of
`active drug able to bind to the site of action.
`
`Receptor
`Binding
`
`Cerebrospinal
`Fluid
`
`Nonspecific
`Binding
`
`Blood Brain1Barrier
`
`Esterified Drug
`In Plasma
`
`Esterase
`
`Unbound
`Drug
`In Plasma
`
`Metabolism
`
`Plasma Protein
`Binding
`
`Esterified Drug
`In Depot Oil
`
`Rena and
`Biliary
`Elimination
`
`Multicompartment
`Tissue Binding
`
`as a cost-effective solution to the growing demand for psy-
`chotropic drug assay. These centers should utilize methods
`that have a large sample capability and rapid turnaround
`time.
`
`PHARMACOKINETIC AND
`PHARMACODYNAMIC APPLICATIONS
`
`Pharmacokinetics is the study of drug absorption, distri-
`bution throughout the body, metabolism, and eventual elim-
`ination. 54 Pharmacodynamics is the study of the
`biochemical and physiological effects of drugs and, most
`importantly, their mechanisms of action. A clearer under-
`standing of the unique advantages and disadvantages of de-
`pot neuroleptics versus oral therapy requires further study
`of the interplay between drug concentration at the site of
`action over time. Drug effects on receptors, such as kin-
`dling phenomena, denervation hypersensitivity, and down
`regulation, also need further study.
`Depot neuroleptic pharmacokinetic variables that inter-
`relate with brain effects are illustrated in Figure 1," which
`shows that once the drug is absorbed from its injection site,
`redistribution and extensive binding to tissue and protein
`occur. As stated, the fraction of the free drug in plasma is
`the only portion that can equilibrate with cerebral spinal
`fluid and brain tissue. Therefore, it is essential in any study
`where only total plasma levels are measured that the protein
`binding be constant over the entire concentration range for
`the drug. Nonlinear protein binding, as with valproic acid,
`decreases the usefulness of total plasma levels.' Figures 2
`and 3 depict the relevant pharmacokinetic parameters ob-
`tained from the plasma level versus time curves for both
`oral and depot neuroleptic therapy. As Figure 3 shows, for
`the depot neuroleptic, absorption from the injection site
`may be the rate-limiting step. The typical "decay phase" of
`the depot curve provides an estimate of the rate of absorp-
`tion of other deep compartment rate constants and not the
`
`Mylan v. Janssen (IPR2020-00440) Ex. 1020 p. 004
`
`

`

`J CLIN PSYCHIATRY 45:5 (Sec. 2) MAY 1984
`
`ERESHEFSKY ET AL.
`
`The valid study of the relationships between plasma lev-
`els and clinical effects depends on the methods used for
`analysis of the drug, the pharmacokinetic derivation of the
`individual parameters, clinical assessments, statistical anal-
`ysis, and the nature of the psychiatric illness. The response
`that is the best measure of neuroleptic effectiveness is rela-
`tive improvement, rather than absolute severity of symp-
`toms. For this reason, it is recommended that correlations
`be based on the difference between the patient's baseline
`severity of illness and severity at the time of the plasma
`level determination. This is especially important in evaluat-
`ing drug effects in a population in which large numbers of
`minimal and nonresponders are anticipated. If a patient's
`psychotic symptoms were considered extremely severe on
`admission and have improved to a "moderately ill" rating,
`then a definite clinical benefit has resulted from the treat-
`ment. On the other hand, if a patient was moderately ill on
`admission and remains moderately ill, no clinical benefit
`has resulted. Both patients are considered moderately ill,
`but one has responded to treatment, while the other has
`not.58 Acutely psychotic patients, although more difficult to
`enroll in studies of plasma level-response relationships, are
`the best candidates in whom to obtain clinical correlations.
`Chronically ill patients with multiple dependencies and per-
`sonality characteristics related to institutionalization are
`less suitable for clinical response studies. A response study
`should use a fixed-dose design, in which patients are ran-
`domly assigned to various dosages that result in a uniform
`distribution from subtherapeutic to supratherapeutic levels.
`In the past, most of the plasma level response studies were
`technically flawed by the use of a variable dose titration
`scheme. However, the ethical difficulties in using acutely
`psychotic patients as subjects in this type of fixed-dose de-
`sign are apparent.
`A parallel strategy for the elucidation of clinical re-
`sponse, adverse effects, and plasma level relationships is
`the use of large numbers of patients in naturalistic environ-
`ments, rather than a few patients in a clinical research cen-
`ter. Many of the flaws and biases present in small
`populations (e.g., inadequate statistical power and lack of a
`uniform distribution of patients who have drug levels above
`and below the therapeutic range) are easily avoided by us-
`ing large populations of patients. The problems inherent in
`the analysis of data obtained from a naturalistic environ-
`ment are overcome by the use of a NONMEM technique to
`develop response and kinetic relationships. We are using
`NONMEM-assisted analysis and subnanomolar quantifica-
`tion of psychotropic drugs in a naturalistic environment to
`develop models of response. The application of these tech-
`niques is illustrated below by data on kinetics for fluphena-
`zine decanoate and metabolite relationships for haloperidol.
`
`METABOLITE RELATIONSHIPS
`FOR HALOPERIDOL
`
`The development of our RIA for haloperidol and the
`
`54
`
`reduced metabolite (hydroxyhaloperidol) shows the benefit
`of collaboration between members of our interdisciplinary
`research group." Our initial RIA procedure did not dis-
`criminate between the parent and metabolite moiety. Clini-
`cal response studies indicated that two compounds were
`being detected on the basis of the pharmacokinetic profile.
`Subsequent review of our analytic techniques revealed that
`our antibody, prepared from the oximoconjugate of halo-
`peridol, was co-specific for both haloperidol and reduced
`haloperidol. We then developed a procedure utilizing selec-
`tive succinylation of the hydroxyl group, followed by silica
`gel chromatographic separation and reassay."
`Our studies with oral haloperidol indicate that reduced
`haloperidol concentrations range between zero and twice
`the parent level.' When patients are given intravenous hal-
`operidol in a cardiac intensive care unit, no detectable re-
`duced haloperidol is found. We speculate that the use of
`haloperidol decanoate may similarly result in a different
`parent-to-metabolite relationship. The dopaminergic antag-
`onist activity of reduced haloperidol is 20% and 25% of
`that of haloperidol, as measured by apomorphine-induced
`stereotypy in rats and in calf caudate RRA, respectively. No
`reduced haloperidol was observed after intraperitoneal in-
`jection of haloperidol in these rats. In addition, the effects
`on weight gain, appetite, and spontaneous motor activity
`were different in rats administered these two compounds
`intraperitoneally." The differing biological activities of re-
`duced haloperidol and the parent compound may partially
`explain our initial findings in schizophrenic patients who
`were poor responders to oral haloperidol therapy."
`The NONMEM analysis of our data on 41 schizo-
`phrenic patients is now in progress. This retrospective data
`set includes 66 plasma levels and clinical global impres-
`sions rating scales.52 Pearson product-moment correlation
`analysis yielded the following: dose vs. the summed plasma
`concentrations of haloperidol + reduced haloperidol, r =
`0.69 (t = 5.42, df = 64, p < .01); dose vs. plasma con-
`centration of haloperidol, r = 0.76 (t = 5.99, p < .01);
`and dose vs. plasma concentration of reduced haloperidol, r
`= 0.57 (t = 4.60, p < .01). A log-linear correlation analy-
`sis showed weaker relationships, with the exception of dose
`vs. plasma concentration of reduced haloperidol, r = 0.71
`(t = 5.59, p < .01). Two groups of patients were identified
`based on our earlier work": those with reduced-haloperi-
`dol/haloperidol ratios > 1 (N = 17, Group 1) and reduced-
`haloperidol/haloperidol ratios < 1 (N = 24, Group 2).
`Age, sex, and diagnosis were not found to differ signifi-
`cantly between the two groups. Significant differences
`identified between Groups 1 and 2, respectively, were as
`follows: haloperidol dosage = 44.5 ± 22.9 vs. 35.4 ±
`17.6 mg/day (grouped Student's t-test, (t = 2.20, df = 64,
`p < .05); reduced haloperidol plasma concentration =
`40.3 + 24.9 vs. 10.6 ± 14.3 ng/ml (t = 2.75, p < .01);
`severity of illness = 4.50 ± 1.18 vs. 3.68 ± 1.11, t =
`2.75, p < .01); and therapeutic effect = 3.36 ± 0.95 vs.
`2.81 ± 1.08 t = 2.00, p < .05). Haloperidol plasma con-
`
`Mylan v. Janssen (IPR2020-00440) Ex. 1020 p. 005
`
`

`

`J CLIN PSYCHIATRY 45:5 (Sec. 2) MAY 1984
`
`PHARMACOKINETICS AND PHARMACODYNAMICS
`
`TABLE 2. Demographics and Descriptive Information for Population Studied
`Fluphenazine Dosage Form
`
`Variable
`Sex
`Male
`Female
`Age
`Range
`Mean ± SD
`Weight (kg)
`Range
`Mean ± SD
`Diagnosis
`
`Decanoate (N=39)
`
`Hydrochloride (N=22)
`
`29
`10
`
`20-64
`36.2±13.2
`
`15
`7
`
`18-59
`33.0+11.5
`
`45-114
`71.0±16.2
`Chronic undifferentiated schizophrenia, 35.9%;
`paranoid schizophrenia, 30.8%; schizoaffective, 15.4%;
`schizophreniform, 7.7%; disorganized schizophrenia,
`5.1%; bipolar manic, 2.6%; OBS with psychosis, 2.6%
`
`50-108
`76.6+14.6
`Paranoid schizophrenia, 45.4%; schizoaffective,
`22.7%; chronic undifferentiated schizophrenia, 18.2%;
`schizophreniform, 9.1%; bipolar manic, 4.5%
`
`same rater, although numerous raters were used throughout
`the 2-year period of the review.
`2. The CGI severity of illness and side effects subscales
`were rated on a scale of 1 to 7, where 1 indicated the item
`was not present, and 7 indicated an extremely severe pre-
`sentation. The change in severity of illness score (delta se-
`verity) was used to assess therapeutic effect of the drug.
`3. Fluphenazine plasma levels for depot therapy were
`obtained either 7 or 14 days after injection, usually preced-
`ing the next dosage administration.
`4. The fluphenazine plasma samples were analyzed by
`HPTLC. Our methodology included careful screening for
`compliance, elimination of data points in patients who re-
`ceived a fluphenazine hydrochloride dose within 48 hours
`of plasma level samplings, controlling for patients with
`known enzyme-inducing or inhibiting substances, and the
`use of an acutely psychotic population.
`Table 2 gives demographic data for the study popula-
`tion, and Table 3 provides fluphenazine dosage information
`and Table 4 plasma level distributions. The 39 patients on
`fluphenazine decanoate were analyzed for the change in
`their severity of illness and side effects from baseline rat-
`ings to final evaluation, at least 4 weeks later.
`Linear Pearson product-moment correlation analysis of
`
`centrations were not significantly different, 24.4 ± 13.3
`vs. 18.2 ± 14.3 ng/ml (t = 1.70, p = .10). A log-linear
`Pearson product-moment correlation analysis yielded the
`best fit for haloperidol plasma concentration vs. reduced
`haloperidol plasma concentration, r = 0.80 (t = 6.21, df =
`64, p < .01). The log-linear least-squares regression line
`was as follows: In reduced haloperidol plasma concentra-
`tion = 0.073 x haloperidol plasma concentration + 0.81.
`Graphical analysis of the data indicates a nonlinear relation-
`ship for reduced haloperidol vs. haloperidol plasma con-
`centration at haloperidol plasma concentrations > 30
`ng/ml.
`Increasing plasma levels of haloperidol result in a
`greater than linear increase in reduced haloperidol plasma
`levels. If reduced haloperidol is a less potent dopamine an-
`tagonist, then the therapeutic window61'62 observed when
`measuring only haloperidol plasma levels might be partially
`explained by the production of reduced haloperidol in the
`intestinal tract. Further studies are currently in progress to
`elucidate this model.
`
`CLINICAL SIGNIFICANCE OF PLASMA LEVEL
`MONITORING OF FLUPHENAZINE DECANOATE
`
`Our pilot data on the use of fluphenazine plasma level
`monitoring in 39 patients receiving fluphenazine decanoate
`and 22 patients on oral fluphenazine therapy in a naturalis-
`tic setting were obtained retrospectively."'64 An additional 3
`patients had plasma samples obtained serially following an
`injection of fluphenazine decanoate. We have been using
`fluphenazine plasma levels and behavioral assessments in
`the San Antonio State Hospital/University of Texas, Col-
`lege of Pharmacy Clinical Psychopharmacology Program
`since 1980. An illustration of the type of information result-
`ing from the analysis of data obtained from our routine
`plasma level monitoring program is presented here.
`1. Before each drug plasma level was obtained, the
`patient was interviewed by a clinical psychopharmacologist
`who completed an assay request form that included the
`Clinical Global Impressions (CGI) rating scale, diagnosis,
`dosage, and demographic data. Each patient always had the
`
`TABLE 3. Descriptive Information for Fluphenazine Decanoate
`and Oral Fluphenazine
`Item
`Starting dose
`
`i
`
`Oral
`23.2+22.9
`mg/day
`
`Decanoate
`29.7+14.7
`mg/injection'
`
`Final dose
`41.0±20.7
`Mean
`27.3 ± 26.9
`Range
`12.5-100
`0-80
`'Starting injection interval = 7 days; injection interval after dis-
`charge = 14-28 days.
`
`TABLE 4. Relationship Between Steady-State Fluphenazine
`Decanoate Plasma Levels and Dosage Range
`Depot Dose (mg/week)
`Plasma Level (ng/mI) Patient N
`25 mg
`0.76
`32
`37.5 mg
`1.11