Comparison of Injectable and Oral Antipsychotics in Relapse Rates in a Pragmatic 30-Month Schizophrenia Relapse Prevention Study

It is well recognized that relapse in schizophrenia is detrimental with respect to both the long-term functioning of individual patients and the attendant burden on caregivers and the mental health care delivery system (1,2). Both psychological and pharmacological approaches have been investigated—with variable success, although with the best results obtained when delivered in tandem—to mitigate the risk of relapse over the course of illness (3).

The advent of second-generation antipsychotic medications, initially in oral formulations and subsequently in long-acting injectable formulations, raised expectations that perhaps a more stable, long-term illness trajectory might be attainable. Also, it has been speculated that new long-acting injectable formulations might provide a superior treatment option in combating the effect of medication nonadherence as a major driver of relapse in schizophrenia. Largely, these suppositions have not been borne out in recent meta-analyses of already published literature or in contemporary relapse prevention trials (4–8). Counter to hypothesis, in a large U.S. Department of Veterans Affairs (VA) study that compared one-year relapse rates between oral medications and injectable risperidone, no significant difference was detected (9). The PROACTIVE trial (Preventing Relapse Oral Antipsychotics Compared to Injectables Evaluating Efficacy [10]) is another recent relapse prevention study of similar methodology, longer duration, and a more pragmatic design than the earlier VA study by Rosenheck and colleagues (9). In this 30-month, clinical trial comparing long-acting injectable risperidone with oral second-generation antipsychotic medication, the overall rate of relapse was not significantly different between treatment groups (42% relapsed with injectable risperidone, and 32% relapsed with oral medications). Moreover, there was no significant difference between groups in time to first relapse (10).

Pragmatic trials and analyses of practice can bridge the gap between more rigorous, double-blind clinical trials and clinical practice (8,11,12). Because PROACTIVE was a pragmatic trial, patients could continue in randomized treatment and follow-up after they reached the primary outcome of first relapse. In this study, we compared second and third relapses between treatment groups to assess whether experience of relapse had a differential impact on subsequent relapse among those receiving injectable or oral medication.

Methods

The study design, methodology, and results for first relapse in the PROACTIVE study are described in detail elsewhere (10). In brief, PROACTIVE was a 30-month relapse prevention study at eight U.S. academic sites that enrolled 305 patients with schizophrenia or schizoaffective disorder randomly assigned to receive either long-acting injectable risperidone or oral second-generation antipsychotic medication. With use of two-way videoconferencing technology, centralized raters without knowledge of treatment group assignment assessed patients and then followed up biweekly at each site. The study was pragmatic in design in that it included both efficacy (randomization and raters blinded to study condition) and effectiveness (realistic relapse criteria; allowed flexibility with dosing of medications) features to enhance broader applicability of study findings (8,10). Relapse criteria were similar to those used in an earlier seminal study (9), and relapse was independently determined by a relapse monitoring board whose members were blind to treatment assignment.

Similar to the primary report on PROACTIVE (10), this study was an intent-to-treat analysis that compared second and third relapses between treatment groups using Kaplan-Meier survival analysis.

Results

Regarding first relapse, 109 patients (36%) experienced one relapse—61 of 146 (42%) in the long-acting medication group and 48 of 150 (32%) in the oral medication group—and there was no significant difference between groups in time to first relapse. Overall, there were few subsequent relapses in the PROACTIVE cohort (Figure 1). A total of 32 patients (11%) experienced two relapses; 16 of 147 (11%) in the injectable medication group and 16 of 151 (11%) in the oral medications group. There was no significant difference between groups in time to second relapse. A total of 13 patients (4%) experienced three relapses; five of 147 (3%) in the injectable medication group and eight of 151 (5%) in the oral medication group. There was no significant difference between groups in time to third relapse. There also was no statistically significant difference between the two treatment groups in terms of patients who discontinued treatment (for whatever reason) before study end (53% [N=74] of the injectable medication group and 53% [N=78] of the oral medication group).

Discussion and Conclusions

This analysis revealed similar second and third relapse rates between patients receiving either long-acting or oral second-generation antipsychotic medications. Although the primary analysis of first relapse in the PROACTIVE study yielded similar results between long-acting and oral treatment groups (10), we speculated that any between-groups differences might become more apparent among patients who have subsequent relapses. Indeed, several earlier relapse prevention studies showed greater benefit of the long-acting formulation over oral medications when a more severely ill population was considered (4). However, our analyses failed to support this contention.

Consistent with the pragmatic design (8,10), this study showed that there was more latitude in the oral group to change oral medications at either the clinician’s discretion or the patient’s request. This flexibility was not available for the patients receiving the long-acting injectable medication. That said, the proportion of patients experiencing subsequent relapses and the time to relapse were strikingly similar between the two treatment groups. Several factors may account for this finding. First—and likely foremost—the biweekly contact with patients in both treatment groups by experienced staff may have mitigated the extent of later relapses overall in this cohort as well as “washed out” any potential differential impact on relapse across the two treatment groups (8). Second, several of the most recent long-acting injectable comparisons—of either old and new long-acting injectables or long-acting injectable versus oral medications—have not shown differences (4–6,8,9). This pattern mirrors prior studies, as well as the initial PROACTIVE results. In contrast, a recent comparative study of long-acting injectable and oral risperidone used with first-episode schizophrenia patients showed a real benefit of the injectable formulation in reducing relapses (5% relapse rate in injectable medication group versus 33% relapse rate in oral risperidone group) (13). Third, the PROACTIVE study cohort largely comprised persistently ill, although not severely ill, patients, such that there were few relapses over time. Comparison with other relapse prevention studies of one-year duration or longer is consistent with that assumption (9). That said, the low rate of subsequent relapse among patients in the PROACTIVE study precludes the detection of any differential treatment effect (even if it were present).

As previously described, all patients in both the injectable and oral treatment arms were seen in the clinic every two weeks (at which time staff met with them individually to discuss the illness in a confidential and comfortable atmosphere), and missed appointments received immediate attention to determine the reason and to reschedule. The contribution of this level of care to the generally low rates of relapse provides indirect evidence that frequent outpatient contact with patients with schizophrenia may itself be a valid relapse prevention strategy. Finally, relatives can sometimes be concerned that their loved ones’ participation in research might compromise care and outcomes that might otherwise be attainable through regular services. Although this study clearly lacked direct comparative details, the low relapse rate sustained over 30 months of care within a pragmatic treatment trial design is reassuring that high-quality care is provided in research settings as well as in routine practice. Similar models of pragmatic research are being adopted to enhance generalizability of findings to clinical practice and to aid in the recruitment of patient samples that are more broadly representative of community mental health practices (3,14). The extent to which each informs the other offers the potential for future gains in the delivery of care for people with schizophrenia.