Long acting risperidone in Australian patients with chronic schizophrenia: 24-month data from the e-STAR database

Patients with schizophrenia or schizoaffective disorder (diagnoses on the basis of DSM-IV criteria [8]) who started treatment with RLAI at an in- or out-patient setting during the course of their routine clinical management were eligible to participate. Patients with co-morbid psychiatric disorders were eligible to participate. All participants provided written, informed consent prior to any data collection, either personally, or via their appointed legal guardian.

Participants were prescribed RLAI according to the Australian local label [9]. For risperidone naïve patients, it was recommended that a test dose of immediate release oral formulation was given to establish tolerability [9]. The recommended starting dose for all patients was 25 mg given intramuscularly every two weeks [9]. Sufficient antipsychotic therapy coverage was recommended during the three-week lag period following the first RLAI injection [9]. For patients switched from a long-acting depot medication, RLAI was initiated at the next scheduled depot injection time, with sufficient oral coverage required for 3-weeks. Upward dose adjustments were recommended to not be made more frequently than every four weeks [9].

Data collection in the study has been described previously [7]. Briefly, however, three data collection periods were specified: retrospective, baseline and prospective. The retrospective data were collected from patient medical records from the treating hospital and/or community mental health clinic for a period of 12-months prior to initiation of RLAI. Baseline data were collected at the time of initiation of RLAI. Prospective data were collected every 3-months from baseline for a period of 24-months.

Data collected during the retrospective period included treatment and hospitalisation history. At baseline (defined as the initiation of RLAI) the following information was collected: patient demographics, illness characteristics (using the Clinical Global Impression of illness Severity [CGI-S [10]] and Global Assessment of Functioning [GAF [8]]) including dose of RLAI, the reason for RLAI initiation, other psychiatric medications, and community treatment orders (CTO). Prospectively, we collected information on CGI-S, GAF, psychiatric hospitalisations, concomitant therapies, clinical deterioration (defined as one of the following: partial or full hospitalisation for exacerbation of psychotic symptoms; need to increase level of care combined with an increase in CGI-S of at least 2 points; the emergence of suicidal or homicidal ideation; deliberate self harm, and violent behaviours), treatment adherence and discontinuation, compulsory treatment orders, and adverse events.

Data were entered directly into the web-based e-STAR tool. Data presented here are based upon intent-to-treat (ITT) and per protocol (PP) analyses of Australian patients enrolled in this 24-month study. The ITT population were those subjects who received at least one injection of RLAI. The PP population were those subjects who received RLAI at baseline and completed the study (defined as having at least one recorded piece of information 720 days (24-months) post baseline). The PP population did not necessarily remain on RLAI for the 24-month period. We did not use a last observation carried forward procedure. Therefore, comparisons between the PP and ITT populations can only be made at baseline. Safety data are presented for the ITT and PP populations of Australian patients.

A Kaplan-Meier procedure was used to estimate the percentage of patients who discontinued from RLAI in the first 24-months following initiation. The first reason for RLAI discontinuation was reported with descriptive statistics.

Data are presented as mean (standard deviation) unless otherwise noted. To adjust for multiple testing, statistical significance was assigned where P < 0.001.

A total of 784 patients were available for inclusion in the ITT analysis. The PP population consists of 359 patients. The demographic, baseline clinical characteristics and previous hospitalisations of ITT and PP patients are shown in Table 1. For those patients in the ITT population with "other" diagnoses, the diagnoses included schizophreniform (n = 13); delusional disorder (n = 6); bipolar affective disorder (n = 4); psychosis (n = 4); "likely" schizophrenia (n = 1); and schizoaffective with organic personality disorder (n = 1). The ITT and PP were similar with respect to baseline demographics, although those who completed the study (PP population) had a longer duration of illness compared to the whole (ITT) population.

The main reasons patients were initiated on RLAI were due to adherence (51.8%), unacceptable tolerability/adverse events (24.7%), and insufficient response (14.3%). In the 3-months prior to initiation of RLAI, most patients who were on antipsychotic medications received oral atypical antipsychotics either as monotherapy (27.0%) or in combination (13.4%); or were administered a conventional depot medication either as monotherapy (23.1%), or in combination (11.2%).

The majority of patients in the ITT population were initiated on 25 mg RLAI (90.8%). The mean dose at baseline was 26.4 ± 5.0 mg with a median of 25 mg. At 24 months, the mean dose in patients still receiving RLAI was 43.4 ± 15.7 mg with a median of 37.5 mg. Slightly more than a third of patients had no dose adjustments (38.0%); others subsequently went on to have 1 (22.3%) or 2 (23.6%) dose adjustments. Three or more dose adjustments were required in only 16.1% of patients. The proportion of patients taking each dose of RLAI over the 24-months is presented in Figure 1.

Similar to the ITT population, the majority of patients in the PP population were initiated on 25 mg RLAI (89.4%). The mean dose at baseline was 26.6 ± 5.2 mg, which increased to 43.6 ± 15.7 mg at 24 months. A greater proportion of patients required 2 dose adjustments (28.4%), than none (22.3%), 1 (25.3%), or 3 or more (24.0%).

For PP patients with 24 months of follow up, there was a decrease in the use of other psychiatric medications over the course of the study (70.3% at baseline vs. 64.3% at 24-months, p = 0.048). There was a reduction in benzodiazepine use (42.3% at baseline vs. 29.7% at 24-months, p < 0.001), a non-significant increase in the use of antidepressants (16.9% at baseline vs. 18% at 24-months, p = 0.71), and a reduction in concurrent anticholinergics (22% at baseline vs. 16% at 24-months, p = 0.026). There was no marked change in the use of mood stabilisers or somatic medications. As the percentage of patients at baseline using other psychiatric medications was similar between the ITT (71.0%) and PP (70.3%) populations, the PP population is considered a representative subpopulation of the ITT.

A significant reduction in mean CGI-S score occurred over the 24-month period (from 4.48 ± 1.04 at baseline to 3.59 ± 1.09 at 24-months; p < 0.001; Figure 2). The majority of this reduction was seen in the first 3-months and sustained over the 24-month period. This pattern was reflected in the GAF scores, which improved significantly over the 24-month period (from 42.3 ± 14.5 at baseline to 58.7 ± 15.3 at 24-months; p < 0.001). Again, the majority of change occurred within the first 3-months and was sustained over the full 24-month period (Figure 3). Similarly, the ITT results were: 4.52 ± 1.04 at baseline to 3.56 ± 1.10 at 24-months. And, were also reflected in the improvement in GAF scores over 24-months: 42.9 ± 14.5 at baseline to 59 ± 15.4 at 24-months.

At baseline, 51.1% of patients were being treated on an in-patient basis. During the 12-months post baseline (using discharge date as the baseline for inpatients), 48.5% of patients were fully hospitalised at least once, as opposed to 77% in the 12-months prior to baseline (p < 0.001). The percentage of patients fully hospitalised at least once post baseline increased from 48.5% at 12-months to 58.1% at 24-months. However, if the initiation date of RLAI is used as a baseline for all patients, 70.2% of patients were fully hospitalised at least once in the first 12-months (including the initial episode). And, by 24-months, the percentage of patients fully hospitalised increased slightly to 73.5%. Only patients with data at the specified time point were included in the analyses (i.e., completers only).

There was a significant reduction in the proportion of patient rehospitalisations from baseline to 24-months (ITT, 54.2% at baseline vs. 12.7% at 24-months; PP, 47.9% at baseline vs. 12.7% at 24-months, p < 0.001 McNemar's test). There was also a significant reduction in the proportion of patients who were suicidal or exhibited homicidal ideation (ITT: 7.9% at baseline vs. 2.3% at 24-months; PP, 7.6% at baseline vs. 2.3% at 24-months, p = 0.001). Although improved at 24-months compared with baseline, violent behaviour (p = 0.84) and self-injury (p = 0.18) did not show a significant change over the same time period.

The Kaplan-Meier estimate for the percentage of ITT patients who discontinued RLAI before 24-months was 66.7% (95% CI: 63.2% to 70.2%; Figure 4). For the PP population, the Kaplan-Meier estimate was 46.0% (95% CI 41.0% to 51.3%). For the 472 patients in the ITT analysis who discontinued RLAI before 24-months the mean time to discontinuation was 239 ± 180 days. The main reasons cited for discontinuation were insufficient response to treatment (25.8%) and lost to follow-up (23.7%). Excluding patients with lost to follow-up given as the reason for RLAI discontinuation (i.e. patients with no 24-month follow-up); the rate decreases to 46%. The most frequently prescribed antipsychotic at the time of discontinuation of RLAI was oral risperidone (21.6%) either alone (15.3%) or in combination with other antipsychotics (6.3%). Other agents frequently prescribed (either alone or in combination) were olanzapine (9.5%), clozapine (9.1%) and zuclopenthixol (8.1%).

There were 2273 adverse events reported by 411 (52.4%) patients in the ITT population during the prospective 24-month period of the study. Relationship to RLAI was not reported. The most common events (occurring in > 5% patients) were sedation (10.3%), increased weight (8.7%), headache (7.1%), and tremor (6.5%). Sixty-seven (8.6%) patients reported 153 serious adverse events. The most common (occurring in > 1% of patients) were suicidal ideation (1.4%), death (1.3%), and hospitalisation (1.0%).

Of the 359 patients in the PP population, 194 (54.0%) received RLAI for at least 24-months. A total of 628 events were reported by 109 (56.2%) patients who were still receiving RLAI at study end. The most common events (occurring in > 5% patients) were increased weight (12.4%), sedation and headache (8.2%), tremor (7.7%), akathisia (7.7%), insomnia (5.2%), dyspnoea (5.2%) and vomiting (5.2%). Nineteen (9.8%) patients who received RLAI at study end reported 44 serious adverse events. The most common (occurring in ≥ 1% of patients) were hospitalisation (1.5%), suicidal ideation (1.5%) pneumonia (1.0%), intentional overdose (1.0%), overdose (1.0%), aggression (1.0%) and suicide attempt (1.0%).

The following ethics committees provided ethical approval for the study: Alfred Hospital, The Alfred Ethics Committee (48/04); Box Hill Hospital, Eastern Health Research & Ethics Committee (E31/0304); Dandenong Hospital, Southern Health-Human Research Ethics Committee (03139D); Fremantle Hospital, Human Research Ethics Committee-Fremantle Hospital and Health Service (03/324); Glenside Campus, Royal Adelaide Hospital-Research Ethics Committee (030904a); Gold Coast Hospital, Gold Coast Health Service District-Queensland Health (200401); Graylands Hospital, Human Research Ethics Committee, Area Mental Health Services, North Metropolitan Health Service (1.15.13 (03/15)); Hornsby Ku-ring-gai Hospital, Northern Sydney Health-Human Research Ethics Committee (0311-220 M); Royal Brisbane Hospital, Royal Brisbane & Women's Hospital & Health Service District Office of the Human Research Ethics Committee (2003/157); Royal Hobart Hospital, Department of Health and Human Services-Tasmania, (H0007646); Royal Melbourne Hospital, Melbourne Health Behavioural and Psychiatric Research & Ethics Committee (E/04/004); St Vincent's Hospital, St Vincent's Health, Research and Grants Unit (HREC-D 139/03); The Prince Charles Hospital, Human Research Ethics Committee, The Prince Charles Hospital Health Service District (EC2366); The Queen Elizabeth Hospital, Central Northern Adelaide Health Service-Ethics of Human Research Committee (2003116); and Westmead Hospital, Western Sydney Area Health Service-Human Research Ethics Committee (HREC2003/12/3.2(1712)).