Is monitoring of plasma antipsychotic levels useful?

Although there has been success in defining the minimum therapeu­tic response threshold for certain antipsychotics—for example, clozap­ine (350 to 450 ng/mL), haloperidol (3 to 5 ng/mL), and fluphenazine (0.8 ng/mL)—one aspect of antipsychotic plasma levels not widely discussed is their value as a marker of adherence.

Many schizophrenia patients achieve an optimal response to agents for which there is no depot formulation. For them, maintenance of symptom con­trol depends wholly on oral medication adherence.¹ Regrettably, nonadherence with oral antipsychotic treatment is prevalent among patients with schizo­phrenia; yet, in routine clinical practice, the extent of nonadherence rarely is measured.

Studies have been able to quantify oral medication nonadherence using moni­toring devices, such as the Medication Event Monitoring System (MEMS) that electronically records opening of the medication container and strongly cor­relates with pill count. Although patients knew they were participating in a trial using MEMS technology, only 48% were able to take their medication at least 80% of the time in a 4-week study,² and only 43% met the 70% adherence threshold in a 6-month trial.³

Clinicians, patients: Unreliable indicators of adherence
Neither clinician rating nor patient self-reporting is a reliable predictor of adherence with an oral medication regimen. In the 6-month adherence trial,³ clinicians estimated that 95% of their patients met the 70% adherence threshold (the actual percentage was 43%); in a 12-week study, clinician rat­ings correlated weakly with adherence (r = 0.32; P = .001), but patient self-reporting showed no significant correla­tion (r = 0.18; P = .08) with pill count.⁴

Clinicians underestimate not only the extent of nonadherence but also the impact that even a brief period of modest nonadherence has on the risk of relapse. In an 18-month prospec­tive study of patients who recently had been given a diagnosis of schizo­phrenia, and in whom clinician and patient reports were supplemented with a pill count every 1 to 2 weeks and plasma antipsychotic levels every 4 weeks, any period of at least mild nonadherence was significantly pre­dictive of symptom exacerbation or relapse (hazard ratio [HR], 3.4; 95% CI, 1.4–8.4; P < .002).¹ Moreover, 50% to 75% adherence for ≥ 2 consecutive weeks increased the HR to 5.8, and moderate nonadherence (<50% for 2.0 to 3.9 weeks) to an HR of 28.5.

There might be a better method already available
Given the poor correlation between a cli­nician’s judgment and a patient’s actual pill-taking, it is clear that more accurate methods of tracking adherence must be devised. Because MEMS technology is not widely available, and because pill counts require a home visit or a coop­erative patient who brings medications to office visits, plasma antipsychotic monitoring potentially is an appealing method of tracking adherence.

Correlation between the plasma antipsychotic level and relapse is not consistently seen in the literature,⁴ but plasma levels obtained during periods of clinical stability offer the opportunity to define, for the individual patient, a range of drug exposure that is associated with clinical response. The ideal plasma level baseline is obtained at steady state during a presumed period of observed adherence, such as during a hospital stay. Although patients can be nonadherent in the hospital, this setting offers the best proxy for an acute clinical response to a given plasma level. The alternative is to obtain several plasma levels during a period of outpatient clinical stability.

Clinicians must be mindful that changes in the plasma antipsychotic level after hospital discharge might not reflect poor adherence; environmental factors (eg, exposure to cytochrome P450 or P-glycoprotein inducers) can have a sig­nificant impact on results. Resumption of smoking is a classic example, and rou­tinely is associated with a 50% reduction in plasma clozapine levels.5

There also is expected variability in plasma antipsychotic levels based on (1) the timing of prior doses with regard to trough levels, and (2) the effects of an occasional missed dose. Nevertheless, in a sample of adherent clozapine-treated patients, investigators found that 98% of patients had a coefficient of variability (CV) of 30% for sequential plasma con­centrations (mean CV, 14%).⁶

Clinicians should inform patients that the plasma antipsychotic level is a tool for helping track treatment engagement before relapse—the same way metabolic monitoring helps track abnormalities that can be associated with future cardiovascular events. (Clinicians also must be charitable with their patients when discussing a significant drop in the plasma antipsychotic level [eg, >30%], acknowledging that many patients often miss doses.)

Using the patient’s input about specific difficulties with a medication regimen, clinicians should strive to find ways to improve oral medication adherence. In many cases, the clinician can assist through medication simplifi­cation, consolidation of multiple daily doses, provision of pill boxes, and dis­cussions about long-acting injectable (LAI) antipsychotics.