Gerhard Gründer, MD

One of the most challengingaspects of psychiatric pharmacotherapyfaced by clinicians isto administer a drug dosage that is mostclinically effective yet causes minimalside effects. This is not a trivial problem.Owing to marked individual differencesin the pharmacologic profiles ofeach psychotropic drug, there can beenormous variations in plasma concentrations.

 Since the therapeutic range isrelatively narrow for the mood stabilizerslithium, carbamazepine, andvalproic acid, their plasma concentrationsare routinely monitored. As aclassic example of routine therapeuticdrug monitoring (TDM), plasma lithiumconcentrations must be carefully monitoredto avoid dangerous and potentiallylife-threatening toxicity.

TDM of other widely used psychotropicdrugs, such as antipsychoticsand antidepressants, is much lesscommon. Recently, however, positronemission tomography (PET) and singlephoton emission computed tomography(SPECT) have been used extensivelyto characterize the relationships betweenoccupancy of target moleculesin the brain (neurotransmitter receptorsand transporters), plasma concentrationsof the respective drug, and clinicalefficacy and side effects. Indeed,PET occupancy measures have approachedthe status of surrogate markersfor probable drug effectiveness andhave proved critical in moving potentialpsychotropic agents from the preclinicalto the clinical stage of development.

It is now widely accepted that determiningthe appropriate occupancy levelfor various classes of drugs is vital tostreamlining drug development. Forsome drug classes, these findings are ofconsiderable clinical importance. In thisarticle, I will provide a brief review ofthe latest data on antipsychotics, antidepressants,and other psychotropic drugsin relation to brain occupancy and plasma levels. It will be shown that PET has validatedthe concept that plasma concentrationsdirectly reflect concentrations ofpsychotropic drugs at brain target molecules;thus, the simple determination ofa drug plasma concentration helps toachieve optimal brain concentrations inthe respective therapeutic window.

With regard to the relationship betweendrug kinetics in plasma and brain and theclinical effects and side effects, antipsychoticsare among the most studied drugs.Farde and colleagues

 demonstrated thatclinically effective doses of typicalneuroleptics occupy between 65% and90% of D

-like dopamine receptors. Thesuggestion of a therapeutic windowbetween 60% and 80% striatal D

 receptoroccupancy for sufficient treatmentresponse and a ceiling of approximately80% occupancy for extrapyramidal sideeffects (EPS) was later confirmed byFarde and colleagues and a number ofother groups (

Although antipsychotics are characterizedby marked pharmacologicheterogeneity, the general rule that theincidence of EPS increases in a dosedependentmanner also applies for mostof the second-generation antipsychotics,including olanzapine and risperidone.They lose their atypical properties whenstriatal D

 receptor occupancy passes athreshold of approximately 80%. Theserotonin type 2 (5-HT

) antagonism thatcharacterizes most of the second-generationantipsychotics seems to protectfrom EPS only at moderate doses; thispreservation is lost at striatal D

 receptoroccupancies above the 80% threshold.Thus, when the doses of these drugsare raised above a certain threshold(approximately 6 mg/d for risperidone,30 mg/d for olanzapine, and 160 mg/dfor ziprasidone), striatal D

 occupancyincreases to levels that are associated witha higher incidence of EPS.

There are some exceptions to thisgeneral rule. A number of second-generationantipsychotics do not induce EPS,even when their doses (or plasma concentrations)are raised to unusually highlevels. In those cases, the upper thresholdof the therapeutic window is definedby other side effects, rather than EPS.

This pharmacologically heterogeneousgroup of antipsychotics can bedifferentiated by the distinct characteristicsof these drugs in PET studies.Clozapine and quetiapine occupy a maximumof 60% to 70% of striatal D

-likedopamine receptors even at extremelyhigh doses or plasma levels (

receptors is no longer detectable 24 hoursafter the last administration of quetiapine.

 These particular characteristics ofclozapine and quetiapine are most likelydue to their low affinity rather than theirrapid dissociation from the D

 receptor.Thus, the tolerability and safety of thesedrugs is limited by anticholinergic, antihistaminic,and α-adrenolytic sideeffects, rather than EPS.

On the other hand, administration ofaripiprazole leads to complete saturationof striatal D

 receptors at clinicaldoses, although the incidence of EPS isvery low (

 Again, tolerabilityis not restricted by the occurrence ofEPS but rather by other side effects, suchas psychomotor activation. The characterization of aripiprazole as a partialdopamine agonist explains this uniquefeature. Nevertheless, there are severalreported cases of EPS caused by aripiprazolewhen used in combination withserotonergic antidepressants, suggestingthat aripiprazole may exert strongerdopamine antagonistic properties undercertain real world conditions.

Consequently, the therapeutic windowfor treatment with most antipsychoticdrugs is defined at the low endby a minimum D

 receptor occupancy,which has to be achieved, and at thehigh end by the EPS threshold. Withthe exceptions of aripiprazole, clozapine,and quetiapine, this rule applies toboth first- and second-generationantipsychotics.

When the plasma concentrationrange reaches the therapeutic occupancywindow (established by means of PET),we can test whether this range is the mostuseful in large patient populations. Thishas been demonstrated for variousantipsychotic drugs, and a well-documentedand clinically important exampleis risperidone. PET studies show thatrisperidone at 6 mg/d, which is the highestdosage administered under usual clinicalconditions, enables a significantproportion of patients to have striatal D

-like occupancies sufficient to cause EPS,while a daily oral dose of 3 to 4 mgproduces a striatal D

-like occupancy inthe range of 70% to 80% in most patients.

In a sample of 9 patients, Kapur andassociates

 documented mean receptoroccupancies of 66% at 2 mg/d, 73% at4 mg/d, and 79% at 6 mg/d of risperidone.Three patients, those with thehighest receptor occupancies, exhibitedmild EPS in that study. Striatal D2-likeoccupancy was 25%, 40%, and 48% in3 patients who were given 25 mg ofrisperidone long-acting injectable(RLAI)

; 50 mg RLAI led to occupanciesof 59%, 71%, and 83%; and 75 mgRLAI to 62% and 72%. The correspondingactive-moiety concentrations(active moiety: risperidone and its activemetabolite hydroxyrisperidone) were5.2 to 7.4 ng/mL, 15.0 to 37.0 ng/mL,and 20.9 to 22.5 ng/mL, respectively.

Thus, the available data suggest thatthe 80% threshold that is associated witha higher probability of EPS is reachedwith approximately 40 ng/mL of bothrisperidone and RLAI.A plasma concentrationof approximately 15 ng/mL ofboth formulations is associated with60% striatal D

 occupancy. It can beconcluded from this study that the therapeuticwindow for risperidone (activemoiety) is approximately 15 to 40 ng/mLand that determination of plasma concentrationsis warranted in those patientswho do not respond sufficiently (ie,very low plasma levels) or who sufferfrom EPS (ie, very high plasma levels).

The knowledge of an individualplasma level can provide meaningfulinformation on psychotropic drug concentrationsat brain target molecules.Baumann and associates

 provide a thoroughlisting of recommended targetplasma concentration ranges for psychoactivedrugs and levels of recommendationfor routine monitoring. However,the therapeutic ranges are based onclinical studies in most cases and lackconfirmation by imaging studies.

More recently, PET (and SPECT) technologyhas been applied to the characterizationof antidepressant drugs.However, since radioligands for thenoradrenaline transporter are still essentiallylacking, the selective serotoninreuptake inhibitors (SSRIs) have beencharacterized with regard to their bindingto the serotonin transporter (5-HTT). The quantification of binding ofpsychotropic drugs to serotonin receptorssuch as 5-HT2 and 5-HT1A is stillsomewhat experimental.

 have demonstratedthat clinical doses of variousSSRIs occupy more than 80% of the 5-HTT in depressed patients. For patientstreated with 20 mg/d of paroxetine, themean 5-HTT occupancy was 83%, andfor those treated with 20 mg/d of citalo-pram, it was 77%. The 5-HTT occupancyincreased in a nonlinear manner; withparoxetine, 5-HTT occupancy reacheda plateau of approximately 85% at serumparoxetine levels above 28 g/mL(

, page 61]). Fluvoxamine has been shownto lead to 80% 5-HTT occupancy at50 mg/d or a plasma concentration ofapproximately 30 ng/mL.

Assuming that the basic therapeuticprinciple of treatment with SSRIs is ablockade of 5-HTT, a substantial increasein serum level above the 30 ng/mLthreshold (for both paroxetine andfluvoxamine) might lead to a higher incidenceof side effects and higher drugcosts without increasing clinical effectiveness.Furthermore, saturated 5-HTTscombined with high doses of antidepressantscan explain the fact that, ingeneral, relationships between plasmaconcentrations of SSRIs and clinicaltreatment response to relatively high oraldoses of these drugs cannot be detected. It might also explain the relatively flatdose-response relationship of SSRIs inpatients with depression. Increasing theSSRI dose in these nonrespondingpatients does not substantially increaseremission rates (however, the situationmight be different with other diseasessuch as obsessive-compulsive disorder).

It must be noted, however, that individualsdiffer in their absorption andmetabolism of SSRIs. Some patientsmay require unusually high SSRI dosesin order to reach the blood level requiredfor optimal 5-HTT binding. Thus, thenotion of a flat response curve is onlyvalid in the aggregate; an individualpatient may not fit that model. On theother hand, a rapid metabolizer in theP450 CYP2D6 system might presentwith very low paroxetine (or any otherdrug that is metabolized via CYP2D6)plasma concentrations at clinical doses.These patients usually do not respond to treatment because their corresponding5-HTT occupancy is also too low.Interestingly, the threshold values determinedwith PET are markedly lowerthan those of Bauman and associates.

These discrepancies underline the needfor further research with larger clinicalsamples and in combination with nuclearimaging technology.

Information on the relationship betweenplasma and brain pharmacokinetics andtheir relationship to clinical parametersis almost completely lacking for mostof the other psychotropic agents. Forlithium, carbamazepine, and valproicacid, therapeutic plasma levels havebeen determined, but their effectivebrain targets are virtually unknown.Benzodiazepines and nonbenzodiazepinehypnotics, such as zolpidem as receptor agonists, occupy only a smallpercentage of their target receptors atclinical doses, which makes their measurementwith PET or SPECT technologyunreliable. Quantification ofcholinesterase inhibitor effects in thehuman brain has not reached clinicalimportance yet.

Modern nuclear imaging technology hasimproved our understanding of the relationshipsbetween the binding of psychotropicdrugs to their molecular targets inthe human brain and to their concentrationsin plasma and from there to clinicalvariables such as efficacy and sideeffects. This is especially true for the classof antipsychotics, and to some extent, forantidepressants. Knowledge of some ofthese essential relationships helps todecrease the number of patients who arenonresponders, reduce the incidence ofside effects, avoid serious toxicity, anddecrease the total cost of drug therapy.

Dr Gründer is a professor of psychiatry in theDepartment of Psychiatry and Psychotherapy atRWTH Aachen University in Aachen, Germany.

Dr Gründer reports that he has served as aconsultant for Bristol-Myers Squibb, Janssen,Otsuka, Pfizer, and Astra Zeneca. He hasserved on the speakers bureau of Bristol-Myers Squibb, Otsuka, Pfizer, Astra Zeneca,Eli Lilly, and Wyeth. He has received grantsupport from Bristol Myers Squibb, Pfizer,and Sanofi Synthelabo.

 Hiemke C, Dragicevic A, GrÃ¼nder G, et al.Therapeutic monitoring of new antipsychotic drugs.Ther Drug Monit. 2004;26:156-160.


 Farde L, Wiesel FA, Halldin C, Sedvall G. CentralD2-dopamine receptor occupancy in schizophrenicpatients treated with antipsychotic drugs. Arch GenPsychiatry. 1988;45:71-76.


 Gründer G, Landvogt C, Vernaleken I, et al. Thestriatal and extrastriatal D2/D3 receptor-bindingprofile of clozapine in patients with schizophrenia.Neuropsychopharmacology. Oct 12, 2005 [E pubahead of print].


 Kapur S, Seeman P. Does fast dissociation fromthe dopamine d(2) receptor explain the action of atypicalantipsychotics? A new hypothesis. Am JPsychiatry. 2001;158:360-369.


 Gründer G, Carlsson A, Wong DF. Mechanism ofnew antipsychotic medications: occupancy is not justantagonism. Arch Gen Psychiatry. 2003;60:974-977.


 Cohen ST, Rulf D, Pies R. Extrapyramidal sideeffects associated with aripiprazole coprescriptionin 2 patients. J Clin Psychiatry. 2005;66:135-136.


 Kapur S, Remington G, Zipursky RB, et al. The D2dopamine receptor occupancy of risperidone and itsrelationship to extrapyramidal symptoms: a PETstudy. Life Sci. 1995;57:PL103-PL107.


 Gefvert O, Eriksson B, Persson P, et al. Pharmacokineticsand D2 receptor occupancy of long-actinginjectable risperidone (Risperdal Consta) in patientswith schizophrenia. Int J Neuropsychopharmacol.2005;8:27-36.


 Baumann P, Hiemke C, Ulrich S, et al; Arbeitsgemeinschaftfur neuropsychopharmakologie und pharmakopsychiatrie.The AGNP-TDM expert groupconsensus guidelines: therapeutic drug monitoring inpsychiatry. Pharmacopsychiatry. 2004;37:243-265.


 Meyer JH, Wilson AA, Ginovart N, et al. Occupancyof serotonin transporters by paroxetine and citalopramduring treatment of depression: a [(11)C]DASB PETimaging study. Am J Psychiatry. 2001;158:1843-1849.


 Suhara T, Takano A, Sudo Y, et al. High levels of serotonin transporter occupancy with low-doseclomipramine in comparative occupancy study withfluvoxamine using positron emission tomography.Arch Gen Psychiatry. 2003;60:386-391.

Articles

Administering drug dosages that are clinically effective while causing minimal side effects is a challenge for physicians. The latest data on antipsychotics, antidepressants, and other psychotropic drugs in relation to brain occupancy and plasma levels are reviewed here.

Gerhard Gründer, MD

Articles

Psychotropic Drugs: Brain and Plasma Pharmacokinetics and the Therapeutic Window