Do Long-Acting Injectable Antipsychotics Improve Adherence and Related Outcomes?

The high incidence of inconsistent medication-taking among patients with chronic illnesses is well established, but nonadherence is especially problematic in schizophrenia. Total and partial, overt and covert nonadherence is a serious public health concern because poor adherence is a major reason for relapse and rehospitalization, with all of the attendant negative consequences. Logic suggests that adherence would be improved by using medication that patients do not have to take every day.

We have conducted several reviews of the evidence as to whether long-acting injections (LAIs) of antipsychotic medications lead to better outcomes than oral antipsychotics (OAs). Evidence from randomized controlled trials (RCTs) and mirror-image studies has been examined in meta-analyses, with over 5,000 patients in each meta-analysis. Studies comparing LAIs and OAs provide conflicting evidence, but the results differ by trial design. In general, RCTs tend to show equivalence for the two treatments, whereas mirror-image studies show consistent advantages for LAIs. These diverging results highlight strengths and weaknesses of each design.

Reasons for the seeming contradiction between results from RCTs and those from mirror-image studies seem to lie in the selection, assessment, and treatment of patients. Because the goal of treatment with LAIs is the improvement of adherence over that with OAs, potential problems arise for the generalizability of results from RCTs. Although the traditional RCT is usually considered the “gold standard” design for trials, it might not be the most informative methodology to compare the effectiveness of OAs and LAIs due to selective recruitment of more adherent patients who are able and willing to consent to a double-blind, double-dummy design study. Moreover, in RCTs, patients are seen frequently, are often handed the medication, and know that their adherence is being monitored.

The mirror-image study, on the other hand, utilizes each patient as his or her own control by comparing the frequency of a specific outcome (eg, hospitalization or duration of inpatient stay) during an equal period of time before and after the implementation of the new treatment. A potential problem with this design is that changes that are unrelated to the study might take place over time (such as hospitalization rates declining due to the availability of fewer beds or more rigid application of “medical necessity” to reduce rates of potentially unnecessary hospitalization). Additionally, expectation bias or greater surveillance in the prospective mirror–phase may limit usefulness of this design. However, results from retrospective mirror-image studies that are highly generalizable are fully consistent with the remaining mirror-image data, confirming the significant advantages of LAI treatment in patients selected for clinical eligibility for LAI treatment in usual care settings.

The best approach to determining comparative effectiveness of LAIs versus OAs may be a pragmatic RCT , a large, simple trial with broad inclusion criteria and few exclusion criteria in order to minimize selection bias. Such a trial can mirror “real world” naturalistic treatment without elements of traditional RCTs like research assessments, more frequent follow-up visits, appointment reminders, and assistance obtaining oral medication. Although such studies have rarely been conducted, they are sorely needed. Another opportunity for showing the adherence-enhancing and relapse-reducing effects of LAIs compared to OAs may be a focus on first-episode and early-phase patients with schizophrenia, who tend to benefit substantially from antipsychotic treatment but who also frequently stop taking antipsychotics. For example, a recent 1-year RCT in first-episode schizophrenia patients demonstrated significant and impressive advantages of risperidone LAI over oral risperidone, translating to a number needed to treat of 4 for prevention of psychotic exacerbation or relapse. “Excellent” adherence was found in 95% of patients receiving risperidone LAI compared to 33% taking oral risperidone.

Financial disclosure:Dr Correll is a consultant for AbbVie, Acadia, Actavis, Alkermes, Genentech, Gerson Lehrman, IntraCellular Therapies, Janssen/J&J, Lundbeck, MedAvante, Otsuka, Pfizer, ProPhase, Reviva, Roche, Sunovion, Supernus, and Takeda; has received grant/research support from Bristol-Myers Squibb, Otsuka, and Takeda; has received honoraria from MedScape; is a member of the advisory boards for AbbVie, Acadia, Actavis, Alkermes, Genentech, IntraCellular Therapies, Lundbeck, Otsuka, Reviva, Roche, Sunovion, and Supernus; and has provided expert testimony for Janssen.

Dr Kishimoto is a consultant for Dainippon Sumitomo, Novartis, and Otsuka; has received grant/research support from Pfizer, Takeda, Tanabe-Mitsubishi, Dainippon-Sumitomo, Otsuka, and Mochida; and has received honoraria from Banyu, Eli Lilly, Dainippon Sumitomo, Janssen, Novartis, Otsuka, and Pfizer.

Dr Kane is a consultant for Alkermes, Eli Lilly, Forum, Forest, Genentech, Lundbeck, Intracellular Therapies, Janssen, Johnson & Johnson, Otsuka, Reviva, Roche, and Teva; has received honoraria for lectures from Janssen, Genentech, Lundbeck, and Otsuka; and is a stock shareholder of MedAvante and Vanguard.