Comparison of the Peak-to-trough Fluctuation in Plasma Concentration of Long-acting Injectable Antipsychotics and Their Oral Equivalents

Abstract

Background: Small peak-to-trough drug levels have been suggested to be related to improved tolerability. The aim of this study is to review the steady-state, peak-to-trough, plasma-concentration fluctuation of long-acting injectable antipsychotics and equivalent oral formulations.

Methods: A review of published literature and clinical study reports identified references that reported, depicted, or permitted derivation of the steady-state, peak-to-trough, plasma-concentration fluctuation of antipsychotics (the ratio of maximum concentration to minimum concentration following administration according to the recommended dosing interval) over the dosing interval. Suitable references were identified for haloperidol decanoate, olanzapine pamoate, paliperidone palmitate, risperidone long-acting injectable, and zuclopenthixol decanoate and their oral equivalents except zuclopenthixol. The single-dose time to maximum plasma concentration (T_max) and half-life (t_1/2) were also identified.

Results: The steady-state, peak-to-trough, plasma-concentration ratios of oral antipsychotics varied from 1.47 (paliperidone extended-release, once daily) to 3.30 (active-moiety risperidone, once daily). Among long-acting injectable antipsychotics, the ratios varied from 1.56 (paliperidone palmitate, once monthly) to approximately 4 (olanzapine pamoate, once every four weeks). Among drugs with similar dosing intervals, longer T_max and/or t_1/2 generally correlated with less peak-to-trough fluctuation.

Conclusion: Peak-to-trough fluctuations in plasma concentrations vary widely and may be affected by differences in dosing, pharmacokinetic sampling, subjects’ phenotypes, concomitant medications, comorbid diseases, and formulation. These fluctuations may affect clinical response and tolerability. Along with other patient-specific and drug-specific factors, these fluctuations warrant consideration when selecting an antipsychotic and antipsychotic formulation. Further study is needed with more robust and generalizable peak-to-trough fluctuation data.

Introduction

Dopamine type 2 (D₂)–receptor occupancy has been shown to correlate with antipsychotic therapeutic response.¹ A linear relationship is usually observed between antipsychotic dose and plasma concentration, and an asymptotic nonlinear (i.e., plateau-shaped) relationship exists between plasma concentration and D₂-receptor occupancy.^1–3 Most studies that explore this issue suggest that D₂-receptor occupancy or plasma concentration correlates with tolerability, and some studies suggest a possible correlation with efficacy.^1,2,4–8 For example, the antipsychotic effect of risperidone and haloperidol is reported to be evident in a range of 60 to 65 percent D₂-receptor occupancy, with occupancy above 80 percent reported to increase the risk for developing extrapyramidal symptoms (EPSs)² and other adverse events (AEs). More specifically, recent brain imaging studies suggest that the clinical efficacy of antipsychotic drugs was correlated to dopamine-receptor occupancy in the temporal cortex and striatum, whereas extrapyramidal symptoms were primarily related to striatal dopamine-receptor occupancy.⁹ An increase in plasma concentration peak-to-trough ratio may result in increased fluctuations in D₂-receptor occupancy, although D₂-receptor occupancy changes may lag behind plasma concentration changes due to delays in drug distribution between plasma and the central nervous system.⁴

If plasma concentration correlates with downstream pharmacologic effect on D₂-receptor occupancy, antipsychotic formulations with smaller plasma-concentration fluctuations should cause smaller fluctuations in D₂-receptor occupancy. This smaller fluctuation in occupancy may, in turn, lead to an improved pharmacodynamic profile. Formulations with a shorter half-life (t_1/2) will have greater peak-to-trough fluctuations,⁵ given an equivalent dosing interval. When dose selection is optimized, the advantages of small plasma-concentration fluctuations often seen in controlled-release formulations include reduced peak concentrations (potentially decreasing the incidence of AEs), increased trough concentrations (potentially decreasing the chance of subtherapeutic drug concentrations), and the potential to improve adherence because of a possibly improved risk-benefit profile.^10,11 As early as 1973, Johnson¹² reported that reducing the dosing interval of fluphenazine decanoate (thereby reducing peak-to-trough fluctuation) could reduce AEs without diminishing efficacy. We hypothesized that this finding related to serum drug concentrations remaining below a theoretic threshold concentration may be associated with poor tolerability.¹² Therapeutic drug monitoring groups have recommended a dosing interval that results in a peak-to-trough fluctuation of 2 or less for antipsychotics, although the proposed therapeutic ranges for specific medications vary in width.^13,14

Long-acting injectable (LAI) antipsychotics have been developed with a wide variety of formulation technologies (e.g., covalent linkage of the active molecule to a fatty acid by esterification for paliperidone palmitate, haloperidol decanoate, and zuclopenthixol decanoate; ionic crystalline salt for olanzapine pamoate; microsphere technology for risperidone long-acting injection), suggesting the possibility of widely varied pharmacokinetic profiles.^14–18

The aim of this analysis is to review the steady-state, peak-to-trough, plasma-concentration fluctuation and the pharmacokinetic stability of LAI antipsychotics and equivalent oral formulations and to review the clinical effects of peak-to-trough fluctuation. We present the steady-state, peak-to-trough fluctuation of antipsychotics over the dosing interval, including two pharmacokinetic parameters: time to maximum plasma concentration (T_max) and t_1/2. We present the peak-to-trough fluctuation over the dosing interval because this value likely affects the efficacy, safety, and tolerability of a drug. We present T_max and t_1/2 because these two parameters, along with dosing interval, substantially underlie peak-to-trough fluctuation.

Methods

Study selection. In an effort to identify and use all relevant published sources that could be used to validly determine steady-state, peak-to-trough, plasma-concentration fluctuations, the following literature search strategy was used: first, the references had to explicitly state the peak-to-trough fluctuation in the text or a figure or provide a figure depicting the plasma concentration over time that could be used to derive the mean steady-state, peak-to-trough fluctuation. Second, oral antipsychotics were only included if a suitable reference was identified for the corresponding LAI formulation. In addition, articles were not included if they only showed animal data or were case reports or only showed plasma concentrations before or after steady-state during titration or discontinuation. The first literature search of PubMed from 1960 through April 2011 consisted of the following terms: (clopenthixol or flupenthixol or fluphenazine or haloperidol or olanzapine or paliperidone or risperidone or zuclopenthixol) and (peak-to-trough or fluctuation index or [C_max and C_min] or steady state and pharmacokinetic). If suitable references for a given formulation were not identified in the first literature search, a second literature search of PubMed from 1960 through April 2011 was conducted for the remaining molecules and “pharmacokinetics.” One of the authors (J.J.S.) reviewed both literature searches. The reference lists from the identified articles were further examined to identify additional pertinent references that may not have been identified from the original search; unpublished data on Janssen Scientific Affairs, LLC products were also evaluated. In the event two or more appropriate references were identified, the peak-to-trough fluctuation was averaged across the studies.

Calculations. Whenever possible, peak-to-trough fluctuations were calculated as the ratio of the steady-state, mean maximum plasma concentration (C_max) to the steady-state, mean minimum plasma concentration (C_min) after administration according to the recommended dosing interval. Reported plasma concentrations represent the pharmacologically active moiety, when applicable. Peak-to-trough calculations for each medication are presented in Table 1.^14,19–25 The key studies that yielded the C_max and C_min values for each antipsychotic are listed in Table 2.^14,19–24

TABLE 1.

Peak-to-trough methods of calculation and data source

DRUG	METHOD OF CALCULATION AND DATA SOURCE
LAI medications
Haloperidol decanoate	Derived from the reported dose-normalized Cmax and Cmin19
Olanzapine pamoate	Reported in the literature20
Paliperidone palmitate	Predicted median Cmax and median Cmin in a PK model21
Risperidone LAI	PK model22
Zuclopenthixol decanoate	Reported in the literature14
Oral medications
Haloperidol	Calculated from Cmax and Cmin in the literature19
Olanzapine	Derived from graph depicting steady-state PK23
Paliperidone ER	Calculated from the steady-state mean Cmax and Cmin in a PK study24,25
Risperidone	Calculated from the average peak-to-trough fluctuation of the active moiety derived from a PK study and predicted from a PK model22,24

LAI: long-acting injectable; C_max: maximum plasma concentration; C_min: minimum plasma concentration; ER: extended release; PK: pharmacokinetics.

TABLE 2.

Key clinical studies used to obtain pharmacokinetic data

STUDY	DRUG(S)	NUMBER OF PATIENTS	DURATION OF STUDY
Nayak, 198719	Haloperidol decanoate, haloperidol	30	≥6 months
Taylor, 200920	Olanzapine pamoate	NS	NS
Samtani, 200921	Paliperidone palmitate	NA	NA
Mannaert, 200522	Risperidone LAI, oral risperidone	26	84 days
Poulsen, 199414	Zuclopenthixol decanoate	58	Analysis was during long-term treatment
Callaghan, 199923	Oral olanzapine	NS	NS
Berwaerts, 201024	Paliperidone ER, oral risperidone	38 for each drug	6 days

LAI: long-acting injection; ER: extended release; NA: not applicable (modeling study); NS: not stated (review article)

Results

An initial literature search resulted in identification of 147 publications for haloperidol (oral and LAI), paliperidone extended-release (ER), risperidone (oral and LAI), and zuclopenthixol (LAI). A second, broader literature search for the remaining agents and pharmacokinetics resulted in identification of 460 additional publications. From the second literature search, the authors identified suitable references for olanzapine and olanzapine pamoate. The authors combined these results with internal results for paliperidone palmitate.^16–31 Most publications were not suitable for use in the calculation of peak-to-trough ratios because they did not meet the selection criteria outlined in the methods.

For most agents, only a single reference met the criteria. For zuclopenthixol decanoate, although a few different references reported the peak-to-trough fluctuation, the authors report the results from the most recent article because previous articles did not sample plasma concentrations at peak concentration. For oral risperidone, two articles met the search criteria and were appropriate for inclusion (one pharmacokinetic model and one pharmacokinetic study with human subject data);^{22 24} the average peak-to-trough fluctuation was derived from these two publications. No relevant values were identified for human, steady-state, peak-to-trough fluctuations in plasma concentrations for clopenthixol decanoate, flupenthixol decanoate, fluphenazine decanoate, fluphenazine enanthate, or oral zuclopenthixol.

The steady-state, peak-to-trough, plasma-concentration ratios of oral antipsychotics over the dosing interval varied from 1.47 (paliperidone extended release [ER], once daily) to 3.30 (active-moiety risperidone, once daily) (Figure 1).^14,19–25 Among LAI antipsychotics, the steady-state ratios varied from 1.56 (paliperidone palmitate, once monthly) to approximately 4.00 (olanzapine pamoate, once every four weeks) (Figure 1).^14,19–25

Figure 1.

Steady-state, peak-to-trough fluctuation in plasma concentration over the recommended dosing interval.

Peak-to-trough fluctuations were calculated as the ratio of mean C_max to mean C_min at the steady-state following administration according to the recommended dosing interval. Source data.^14,19–25

*Dose and injection site affect the pharmacokinetics (including peak-to-trough fluctuation) of paliperidone palmitate. Here the steady-state fluctuation of paliperidone palmitate 117mg was administered in the deltoid muscle. Administration of higher doses and/or gluteal administration may result in less steady-state, peak-to-trough fluctuation. Source data.^27,40,41

Selected pharmacokinetic parameters are described in Table 3.^{14,16–19,26–31} Because t_1/2 represents the time for plasma drug concentration to decrease by half, among drugs administered with identical dosing intervals, a formulation with a longer t_1/2 will result in a smaller peak-to-trough fluctuation. Similarly, assuming an identical dosing interval, a longer time to maximum plasma concentration (T_max) also may result in a smaller peak-to-trough fluctuation because the elimination phase would constitute a smaller proportion of the dosing interval (Figure 1 and Table 2).^14,16–31

TABLE 3.

Selected pharmacokinetic parameters

DRUG	Tmax	T1/2
LAI medications, days
Haloperidol decanoate18	6	21
Olanzapine pamoate16,26	4	30
Paliperidone palmitate27	13	37*
Risperidone LAI†,17	35*	4.5*
Zuclopenthixol decanoate14	3	7.4
Oral medications, hours
Haloperidol19,28	4.9	25.6*
Olanzapine29	6	30
Paliperidone ER31	24	23
Risperidone†,30	1.3	19.5

LAI: long-acting injectable; ER: extended-release; T_1/2: terminal half-life; T_max: time to maximum plasma concentration

^*Where a range was reported, the midpoint of the range is presented here.

^†Risperidone: data reported for the active moiety (risperidone + 9-OH-risperidone).

Discussion

Our analyses show that the steady-state, plasma-concentration fluctuations of antipsychotics vary and are also related to how they have been formulated. Among the oral and long-acting antipsychotics evaluated here, steady-state, peak-to-trough plasma concentrations varied least with paliperidone ER and paliperidone palmitate. Among the antipsychotic formulations with a similar dosing interval that were evaluated, such as risperidone and paliperidone, a longer T_max and/or t_1/2 generally corresponded with less peak-to-trough fluctuation.

Over the entire dosing interval, some antipsychotics (e.g., haloperidol and paliperidone) had similar peak-to-trough fluctuations for both the LAI formulation and the oral formulation. For others (e.g., risperidone), the LAI form had less peak-to-trough fluctuation than the oral form (Figure 1).^14,19–25

Bai et al³² hypothesized that improved adherence with risperidone LAI compared with oral antipsychotics may have been due to stable serum concentrations and a subsequent significant reduction in AEs.³² The majority of the literature supports a significant correlation between plasma concentration and/or large fluctuations in peak-to-trough plasma concentration with increased AEs,^2,7,8,33,34 but some studies do not demonstrate this relationship.^35,36

Long-acting or ER antipsychotics have a narrow peak-to-trough fluctuation and have been reported to be better tolerated than drugs with immediate release (i.e., with a larger peak-to-trough fluctuation). For example, comparisons of paliperidone ER and risperidone immediate-release (IR) may provide the best information on the impact of peak-to-trough fluctuation on tolerability because those two formulations use molecules with similar receptor binding profiles but widely different peak-to-trough fluctuations. In 2009, the National Institute for Health and Clinical Excellence compared the tolerability profile of oral atypical antipsychotics and reported a lower odds ratio of developing acute extrapyramidal symptoms with paliperidone ER (0.35) compared with risperidone (0.55) over one year.³⁷ Moreover, a single-dose pharmacokinetic study of intravenous (IV), IR, and ER formulations of paliperidone reported a reduced rate of somnolence with the ER formulation compared with the IR and IV formulations (30% vs. 55–60%, respectively).³⁸ Finally, a randomized, six-week, prospective, blinded-initiation study evaluated medication satisfaction as a primary outcome measure in a schizophrenia trial.³⁹ Participants with suboptimal response to oral risperidone reported improved medication satisfaction after initiation of paliperidone ER at the two-week time point compared with participants who continued to receive oral risperidone.³⁹ Improved patient satisfaction may be linked to the lower peak-to-trough fluctuations of paliperidone ER compared with oral risperidone.

This analysis has several limitations. Steady-state, peak-to-trough fluctuations reported in, or derived from, the literature can vary for many reasons, including differences in study design (e.g., duration, dosing interval evaluated, dose selected, dose stability, frequency of pharmacokinetic sampling). Subjects' phenotypes can be important, particularly for medications with complex metabolism via multiple enzymes. In addition, concomitant medications, substance abuse, and comorbid diseases (e.g., renal failure) may affect plasma concentrations and associated pharmacokinetic parameters. Because the rate of absorption varies depending on whether an antipsychotic is administered into the gluteal or deltoid muscle, or, more significantly, whether it is injected into fatty tissue rather than muscle, injection sites may influence peak-to-trough fluctuation. For example, dose and injection sites affect the pharmacokinetics (including peak-to-trough fluctuation) of paliperidone palmitate, and administration of higher doses and/or gluteal administration may result in less steady-state, peak-to-trough fluctuation.^27,40,41

Not all available antipsychotic drugs were included in this study; for some drugs (clopenthixol decanoate, flupenthixol decanoate, fluphenazine decanoate, fluphenazine enanthate, or oral zuclopenthixol), no adequate studies were available or data were limited. Further, the total number of subjects in the key studies included here was less than 200; therefore, it may not be valid to extrapolate the conclusions of this study to all patients taking antipsychotic drugs.

The literature generally supports a correlation between plasma concentrations and tolerability; however, exact relationships for specific drugs have not been established. Furthermore, substantial interpatient variability in peak-to-trough fluctuation exists. This variability may also be dose related. Nevertheless, although the clinical significance of peak-to-trough profiles identified for antipsychotic treatments is not fully predictive of individual response, these fluctuations warrant consideration, along with other patient- and drug-specific factors, when selecting an antipsychotic and its formulation. As peak-to-trough fluctuations may be affected by differences in dosing, pharmacokinetic sampling, subjects' phenotypes, concomitant medications, comorbid diseases, and injection site, further study is needed to more clearly demonstrate relationships between changes in efficacy and safety parameters and peak-to-trough fluctuations in plasma concentrations of antipsychotic drugs.

Acknowledgments

The authors acknowledge Matthew Grzywacz, PhD, and Sheena Hunt, PhD, for providing writing and editorial assistance; and Mahesh Samtani, PhD, principal scientist, Johnson & Johnson Pharmaceutical Research and Development, for technical assistance.