Pill Counts to Track Nonadherence in Schizophrenia Patients

Although the cause may be lack of insight, cognitive dysfunction, or other aspects of the disorder (e.g. substance use and homelessness), antipsychotic nonadherence is exceedingly common among schizophrenia patients.Reference Dufort and Zipursky ¹⁰ Whether the adherence metric is pill count, medication possession ratio, or plasma antipsychotic levels, an estimated 50% of schizophrenia patients are nonadherent with antipsychotic therapy.Reference Dufort and Zipursky ¹⁰ Even for schizophrenia patients knowingly enrolled in studies employing specially designed Medication Event Monitoring System electronic chips (MEMSCap) to record each pill bottle opening,Reference Brain, Sameby and Allerby ¹¹ adherence rates (by pill count) remain low. One MEMSCap study (n = 52) found that only 48% of schizophrenia patients took 80% of their doses over 4 weeks, with 17.3% of patients taking ≤ 20% of their antipsychoticReference Remington, Teo and Mann ¹² (Figure 1). Employing 70% as the adherence threshold, another MEMSCap study noted a nonadherence rate of 57% among a group of chronic schizophrenia spectrum patients (n = 61).Reference Byerly, Thompson and Carmody ¹³ The latter study also illustrates the concept that patients and clinicians are poor judges of the extent of nonadherence: the patient self-report of nonadherence was 5%, and the physician estimate of nonadherence was only 7%.Reference Byerly, Thompson and Carmody ¹³ These MEMSCaps studies illustrate several important concepts: pill count is an excellent measure of adherence for patients on oral medication; plasma antipsychotic levels are not as well correlated with adherence as are pill counts; pill counts and plasma levels are superior to clinician estimates of adherence.Reference Brain, Sameby and Allerby ^{11 ,} Reference Velligan, Wang and Diamond ¹⁴ As will be discussed later, the additional value provided by plasma antipsychotic levels rests in allowing one to differentiate nonresponders due to inadequate drug levels from those who are treatment resistant.Reference McCutcheon, Beck and Bloomfield ¹⁵

Figure 1.

Four-week MEMSCap data on 52 outpatients with schizophrenia.Reference Remington, Teo and Mann ¹²

For patients who respond to an antipsychotic that is available as an LAI, conversion to the depot formulation is appropriate and can be guided by manufacturer recommendations, and a recent review paper is dedicated to this topic.Reference Meyer ¹⁶ Unfortunately, most antipsychotics and the commonly used first-line mood stabilizers lithium and divalproex are available only in oral forms. The use of pill counts can rapidly identify poor oral medication adherence and permit interventions to avert the psychiatric and forensic consequences. Diversion programs typically have measures to assure that patients have a supply of oral psychiatric medication, so refill records are of less importance than in other outpatient programs. However, the fact that a patient has refilled a medication is not a guarantee that the medication is ingested, a fact borne out by numerous studies.Reference Velligan, Lam and Glahn ^{17 ,} Reference Xu, Gong and Gloyd ¹⁸ Pill counts are thus a critical tool for ensuring oral medication adherence. As diversion patients typically meet with a staff member every week, weekly pill counts are an appropriate interval, with the exact day of the week varying (if possible) to avert pill dumping by higher-functioning patients when the assessment day is preassigned. As diversion enrollees transition to more independent living situations and also demonstrate the ability to self-monitor, the frequency of pill counts can decrease.

Use of Long-Acting Injectable Antipsychotics

Despite the best efforts of diversion program staff, many patients will struggle with oral medication adherence, thus necessitating conversion to an LAI formulation. Moreover, diversion enrollees often have a historical pattern of oral medication nonadherence when not in supervised treatment, so instituting an LAI prior to exiting the diversion program is appropriate to forestall symptomatic relapse and recidivism.Reference Case, Steadman and Dupuis ² That schizophrenia patients with criminal justice contact have high rates of treatment nonadherence served as the basis for a pragmatically informed randomized trial of usual oral medications versus an LAI (paliperidone palmitate) in adult schizophrenia patients who had been in custody within the prior 2 years.Reference Alphs, Mao and Rodriguez ¹⁹ The primary endpoint was time to treatment failure over 15 months defined by one of many outcomes: arrest or incarceration; psychiatric hospitalization; increased psychiatric services to prevent hospitalization; treatment discontinuation or supplementation due to inadequate efficacy, safety, or tolerability; completed suicide. Among the 444 subjects included in the final analysis, paliperidone palmitate was associated with a significant delay in time to first treatment failure compared to oral antipsychotics (hazard ratio 1.43; 95% CI, 1.09–1.88; p = 0.011). Treatment failure rates in the LAI and oral cohorts over 15 months were 39.8% and 53.7%, respectively, with arrest and psychiatric hospitalization the most common reasons for treatment failure (21.2% vs. 29.4% and 8.0% vs. 11.9%, respectively).Reference Alphs, Benson and Cheshire-Kinney ²⁰ The superiority of an LAI versus comparable oral antipsychotics seen in that trial parallels conclusions from meta-analyses of other randomized and controlled study designs.Reference Kishimoto, Hagi and Nitta ²¹

Table 1 provides basic kinetic details on the LAI antipsychotics available in the USA as of January 2019. Although there are 10 separate agents listed, these fall into 4 broad groups: the first-generation antipsychotics (FGAs) (haloperidol or fluphenazine decanoate), risperidone or its metabolite paliperidone (Consta®, Perseris®, Invega Sustenna®, Invega Trinza®), aripiprazole (Abilify Maintena®, Aristada®, Aristada Initio®), and olanzapine (Replrevv®). Detailed information on the use of these LAI antipsychotics and unique safety issues with olanzapine pamoate (Relprevv®) can be found in recent reviews.Reference Meyer ^{16 ,} Reference Meyer ²⁷ While some patients will refuse an LAI, and others might not respond to or tolerate the available LAI options, successful conversion to an LAI vastly simplifies concerns about antipsychotic nonadherence. The failure of a patient to be seen at the prearranged time for the injection immediately alerts clinical staff that something is amiss. Unlike oral medications, the long half-lives of LAI preparations also provide a cushion of time to locate the patient before antipsychotic plasma levels become subtherapeutic.Reference Kishimoto, Hagi and Nitta ²¹

TABLE 1.

Kinetic properties of depot antipsychotics available in the United StatesReference Meyer ^{22 – 24}

Abbreviation: IM = intramuscular.

* Only for those on paliperidone palmitate monthly for 4 months. Cannot be converted from oral medication.

** See US FDA bulletin: (http://www.fda.gov/Drugs/DrugSafety/ucm356971.htm).

*** 21 days oral overlap unless starting with aripiprazole lauroxil nanocrystal + single 30 mg oral dose.

**** Aripiprazole lauroxil nanocrystal (AL_NCD) is only used for initiation of treatment with aripiprazole lauroxil, or for resumption of treatment. It is always administered together with the clinician determined dose of aripiprazole lauroxil, although the latter can be given up to 10 days after the aripiprazole lauroxil nanocrystal injection.

The choice of an LAI is dependent on patterns of response and tolerability during prior antipsychotic exposure. Use of FGA LAIs is necessary for patients who require higher levels of dopamine D₂ antagonism than that provided by second-generation antipsychotics (SGAs); however, for those who are sensitive to D₂ antagonism, risperidone/paliperidone and aripiprazole are better choices. Risperidone and paliperidone have a risk for hyperprolactinemia which exceeds that for any other antipsychotic, including FGAs and aripiprazole, but risperidone and paliperidone also have lower risk for parkinsonism, akathisia, and dystonia than an FGA.Reference Leucht, Cipriani and Spineli ²⁸ Aripiprazole and the FGAs are the LAIs with the lowest weight-gain risk and therefore might be preferable for patients at higher risk for weight gain due to younger age or nonwhite race/ethnicity.Reference Meyer, Rosenblatt and Kim ²⁹

Monitoring Plasma Antipsychotic Levels to Track Adherence and Optimize Antipsychotic Response

In certain circumstances there might be a suspicion of pill dumping by patients on oral antipsychotics, so plasma antipsychotic levels can be used to monitor adherence.Reference Brain, Sameby and Allerby ¹¹ As noted later, fluctuations of more than 30% in trough plasma levels are usually indicative of adherence issues, assuming that levels are obtained at approximately the same time after the last dose and that no new medications have been introduced to influence antipsychotic kinetics.Reference Velligan, Wang and Diamond ¹⁴ The other important use of plasma antipsychotic levels is to guide dosing adjustments, especially in nonresponders. Even when adherence is assured on the basis of regular pill counts or use of an LAI formulation, there is extensive interindividual variation in antipsychotic metabolism, and a wide range of antipsychotic levels at which patients respond.Reference Meyer ¹⁶ When adherence is removed from the equation, inadequate response can only be due to one of two issues: subtherapeutic plasma antipsychotic levels (due to underdosing or kinetic interactions) or treatment resistance.Reference McCutcheon, Beck and D’Ambrosio ³⁰ Before deciding that an adherent patient who failed two or more antipsychotics is treatment resistant, subtherapeutic antipsychotic levels must be ruled out. In a British study of 99 outpatients deemed to have treatment-resistant schizophrenia and in need of a clozapine trial, 35% had subtherapeutic antipsychotic levels, 34% of which (n=12) were undetectable.Reference McCutcheon, Beck and D’Ambrosio ³⁰

Basic principles for obtaining plasma antipsychotic levels

Table 2 outlines the important issues in obtaining plasma antipsychotic levels. For oral medication in particular, consistency in the time since last dose is critical to interpreting changes in levels over time. Table 3 provides estimated trough plasma levels for antipsychotics and commonly used LAI formulations. A single data point can be difficult to interpret, so when levels are unexpectedly high or low based on a given dose a repeat level must be ordered. How one reacts to a high plasma level will depend on the clinical scenario as described in Table 2. Embedded in the discussion of response to high plasma antipsychotic levels is the reality that there is no agreed-upon maximum level for antipsychotics, a fact reflected in the significant disparity between laboratory reference ranges (Table 4). The following section on managing antipsychotic nonresponse provides a nuanced discussion of the concepts that should underlie the use of plasma antipsychotic levels to pursue efficacy in nonresponders.

TABLE 2.

How to obtain and respond to plasma antipsychotic levelsReference Meyer and Stahl ³¹

TABLE 3.

Predicted steady-state trough plasma levels by dose for select antipsychoticsReference Meyer ^{16 ,} Reference Linnet and Wiborg ³⁴

Abbreviations: C/D = concentration/dose; PM = poor metabolizers.

TABLE 4.

Variations in laboratory plasma antipsychotic reference ranges

* Sites accessed January 15, 2019:ARUP Laboratories: https://www.aruplab.comLabCorp: https://www.labcorp.com/test-menu/searchMayo Clinic Laboratories: https://www.mayomedicallaboratories.comQuest: http://www.questdiagnostics.com/testcenter/TestDetail.action(866-697-8378; for Cloz: 800-421-4449 x5440; for Flu: 866-522-6649)

** ARUP reports that the toxic haloperidol level is > 42.0 ng/mlLabCorp reports a potentially toxic haloperidol level is > 50.0 ng/mlMayo provides the following text: “A therapeutic window exists for haloperidol; patients who respond at serum concentrations between 5 and 16 ng/mL show no additional improvement at concentrations >16 to 20 ng/mL. Some patients may respond at concentrations < 5 ng/mL, and others may require concentrations significantly > 20 ng/mL before an adequate response is attained.”

*** ARUP reports that the toxic clozapine level is ≥ 1500 ng/mlQuest reports a level range only for norclozapine: 25–400 ng/ml. Patient reports contain the following referring to the plasma clozapine level: “The therapeutic response begins to appear at 100 mcg/L. Refractory schizophrenia appears to require a therapeutic concentration of at least 350 mcg/L (trough, at steady state). Toxic range: > 1000 mcg/L.” [NB: mcg/L = ng/mL].

Use of antipsychotic plasma levels to manage nonresponse

With respect to oral antipsychotic treatment, meta-analyses of numerous clinical trials have shown that minimal response to a given dose after 2 weeks portends low likelihood of response to that dose at week 6.Reference Samara, Leucht and Leeflang ⁵¹ For patients who fail to respond meaningfully to a given oral antipsychotic dose after 2 weeks, further dose escalation is indicated in the absence of limiting adverse effects. The time frame for dosing adjustments is markedly different for LAI antipsychotics and depends on the half-life of the agent. In general, antipsychotic titration should be terminated only for one of three reasons: (a) intolerable adverse effects, (b) marked improvement, and (c) plasma levels that indicate low likelihood of further improvement (i.e. levels are at the point of futility).Reference Meyer ³² Despite the known dose correlations (Table 3), even adherent patients or those on LAI antipsychotics may have surprisingly low plasma levels due to genetic variations that make the individual an ultrarapid metabolizer for specific cytochrome P450 enzymes.Reference Meyer ³² Determination of the antipsychotic level is the only means to distinguish those who are kinetic failures (as manifested by subtherapeutic plasma levels) and those who are treatment resistant and in need of clozapine.Reference McCutcheon, Beck and D’Ambrosio ³⁰

While there is limited agreement on the high end of the therapeutic range, most laboratories present a narrow range of values that represent the therapeutic threshold for antipsychotic therapy. Table 5 presents working definitions of therapeutic threshold, tolerability threshold, and point of futility to help guide treatment decisions, and evidence-based values derived from the literature are noted in Table 6.Reference Meyer ³² In inadequate responders without adverse effects, the plasma antipsychotic level is important in deciding whether the minimum response threshold has been met. Having a value at or slightly above the response threshold is no guarantee of response as the patient may need a plasma level at the high end of the range, but it does serve as a useful initial target for nonresponders with levels below the threshold.Reference Midha, Hubbard and Marder ⁵² As the titration proceeds to higher plasma levels in search of greater efficacy, most patients will not be able to tolerate levels above the tolerability threshold; however, for those who do, there is little clinical value in exceeding the point of futility beyond which response rates are close to zero.Reference Meyer ³² As laboratories present a wide range of values for the upper end of their reference range (or in some instances no value at all), the point of futility should serve as an evidence-based marker that the clinician can aim for, assuming tolerability, but once reached serves as the defining point to terminate that antipsychotic trial.

TABLE 5.

Concepts for use of plasma antipsychotic levels

TABLE 6.

Response, tolerability and futility thresholds for select antipsychoticsReference Meyer ³²