Effectiveness of aripiprazole once-monthly in schizophrenia patients pretreated with oral aripiprazole: a 6-month, real-life non-interventional study

This multicenter, prospective, non-interventional study was designed according to the German Medicinal Products Act and approved by the Freiburg ethics commission international (Approval number: 014/1336). Diagnosis, treatment and monitoring of patients were conducted in a naturalistic, usual care treatment setting. Prospective collection of the data was chosen to collect high quality data, and a multicenter approach was chosen to ensure an adequate sample size. Patients were only recruited for the study after the treating psychiatrist had chosen AOM treatment for these patients based on clinical grounds. Since this was a naturalistic sample, combinations with other psychiatric medications, including oral antipsychotics were based on clinician’s choice. Participation in the study had no influence on treatment choice.

Seventy-five centers participating in this study were located throughout Germany. The observation period for each patient was about 6 months. Data were collected from July 2014 to March 2016, when the last patient of the target sample had reached 6 months. Data were collected at seven time points (T0-T6), each about 4 weeks apart (− 2/+ 5 days). BPRS data were collected at T0, T3 (month 3) and T6 (month 6), and data for all other endpoints were collected at each monthly time point (except CGI-I, no baseline assessment possible; Fig. 1).

Patients in the study cohort were ≥ 18 years old and diagnosed with schizophrenia (F20.X) according to ICD-10. Starting from the time of inclusion in the study, patients were treated with AOM on an outpatient basis according to prescribing information and clinical need. The treatment choice was independent from the study. All patients gave written informed consent.

Exclusion criteria for the study were contraindications for AOM, being a member or being related to a member of the study staff, pregnancy, planning a pregnancy, breastfeeding, or expected reluctance to follow the prespecified monitoring plan (as assessed by the treating psychiatrist).

The BPRS is a clinician-rated scale [23] which is well established in clinical routine, and therefore it was used as the primary endpoint in this study. Eighteen items were assessed that are grouped into 5 domains: anxiety/depression, anergia, thought disorders, activation and hostility/mistrust. Each of the respective items was rated on a seven-point scale ranging from 1 (not present) to 7 (extremely severe).

The Clinical Global Impression - Severity scale (CGI-S) was used to report the current severity of the patient’s mental illness on a seven-point scale ranging from 1 (normal, not at all ill) to 7 (extremely ill) [24]. This rating was complemented by the CGI-I (Improvement) scale, in which improvement of the mental illness from the beginning of the study is rated on a seven-point scale (1, very much improved to 7, very much worse).

Any patient reported adverse events were documented during the study and rated by the clinicians as treatment related adverse events (TRAE) or unrelated to treatment. Patients were asked about adverse events, comorbidities and comedications at every visit. Adverse events were coded according to MedDRA 19.0.

Other rating scales used in the same study will be discussed elsewhere.

The primary outcome, change in BPRS total score from baseline to endpoint was analyzed using the Wilcoxon Signed-Rank test for paired samples, and the Wilcoxon’s rank-sum test for independent samples. Changes in marginal distributions in contingency tables of categorical outcomes were analyzed using Bhapkar’s test and proportions within one group of patients with the binomial test. Fisher’s exact test was used to compare proportions between groups of patients. Missing values were imputed using the Last Observation Carried Forward (LOCF) method if there was a value for T0 and at least one post-baseline time point. There was one prespecified subgroup analysis for the primary outcome, i.e., the comparison of patients aged ≤35 years old vs patients aged > 35 years old, as the younger age group had performed significantly better than the older age group in a prior AOM efficacy study [25]. Subgroup comparisons by age group were covaried using a linear regression analysis with backward selection of effects for variables that differed significantly between the two age groups at baseline at p < 0.05 (see Table 1). All data were processed using SAS™ software (SAS Institute, Cary, NC/USA), with all tests being two-sided and alpha = 0.05, without correction for multiple testing for secondary outcomes.

Two hundred seventy eight patients were reported by the treating clinicians to be eligible for the study. After screening, 243 patients were included in the study. Patient baseline demographics are presented in Table 1. Altogether, 204 patients (84.3%) completed all scheduled visits, and 23 patients (9.5%) came for at least the first and last visits.

Diagnoses included paranoid schizophrenia (ICD-10: F20.0) in 202 patients (83.5%), non-differentiated schizophrenia in 22 patients (9.1%; ICD-10: F20.3) and others for the remaining 18 patients (7.4%). All patients had been previously treated with oral aripiprazole for a mean duration of 9.7 months (SD 22.3 months), with 33.3% of patients having received oral aripiprazole for < 1 month, 39.9% for 1–6 months, and 26.8% for more > 6 months. Apart from aripiprazole treatment, 141 patients (58.3%) had been treated with additional medication, and the most common substances were risperidone, quetiapine, and olanzapine (35, 34 and 28 patients, respectively, being 14.5%, 14,1, and 11.6% of the patient population).

At study start, 20 patients (8.5%) received 5 mg oral aripiprazole, 183 (77.5%) received 10–20 mg, and 33 (14.0%) received a higher oral aripiprazole dose. During oral aripiprazole treatment, most patients had been symptomatically stable, as assessed by the treating psychiatrist (without use of a dedicated rating scale), for a mean duration of 5.9 months (standard deviation 18.2), with 91 patients (39.2%) being stable for < 1 month and 28 (12.1%) being not stable at all. At the start of the study 81 patients (33.5%) received further medication apart from aripiprazole (Table 2).

Reasons for the decision to switch from treatment with oral aripiprazole to AOM were most often (48.4%) easier adherence to treatment, followed by good/better tolerability (16.9%), patient’s request (13.6%), good/better efficacy (12.0%) and easier use (9.1%) than oral treatment. Switching was a subjective decision of treating clinicians to further improve outcome in a naturalistic sample of patients. At study start, 79.3% of patients received 400 mg of injectable aripiprazole, 17.4% received 300 mg, 2.9% received 200 mg and one patient (0.4%) received 160 mg. Most patients (132, 54.6%) were treated with 400 mg of injectable aripiprazole at every time point T0-T6, and 12 patients (5.0%) received 300 mg at all time points. Of the remaining patients, 30.6% were treated with different AOM doses, and 9.9% discontinued treatment. Reasons for discontinuation of AOM included patient’s request (9, 3.7%), lack of effectiveness (7, 2.9%), adverse drug reactions (6, 2.5%), and patients not arriving at agreed appointments, having moved to a new house, or undergoing inpatient drug addiction therapy (1 case each, 0.4% each; one patient gave two reasons for discontinuation). AOM adherence was monitored based on injection visits taking place, yielding that 84% of the patients were fully adherent, 8% were adherent to receiving > 70% of injections, and that 8% had lower adherence rates. After the 6-month observational period, 200 patients (89.3%) decided to continue AOM treatment.

At baseline, the CGI-S category with the most patients (41.7%) was “markedly ill”. The mean global BPRS value was 54.1 (±15.6, standard deviation, SD) (Fig. 2), and 94.5% presented with a GCI-S value of ≥4 (moderately severe) (Fig. 5).

The psychopathological status of the patients improved significantly during treatment, as assessed via the BPRS (p < 0.0001, Fig. 2). One hundred eighty-seven patients (82.0%) improved during the study, 12 (5.3%) remained the same, and 29 (12.7%) had worse global BPRS scores at T6, seven of which worsened by more than 10 points. Altogether, 131 patients (57.5%) had a reduction of their global BPRS value by at least 20%. In total, the difference in global BPRS between T6 and T0 was − 13.8 (SD: 16.0; 95% CI: [− 15.9; − 11.7]; p < 0.001). The largest improvements were found in the scores anxiety/depression (T6-T0: -0.98; SD: 1.21; 95% CI: [− 1.14; − 0.83]) and activation (T6-T0: -0.82; SD: 1.15; 95% CI: [− 0.97; − 0.67]) (Fig. 3; score reduction must be multiplied by number of respective items to obtain total score reduction).

For patients ≤35 years, the difference between global BPRS values at T6 and T0 was − 17.2 (SD: 17.6; 95% CI: [− 21.1; − 13.4]). For patients > 35 years, the difference was − 11.9 (SD: 14.6; 95% CI: [− 14.3; − 9.5]) (Fig. 4). This difference was not significant (p = 0.0746, Wilcoxon two-sample test). The differences in subscores are presented in Table 3.

Severity of illness as the secondary study outcome was measured using the CGI-S score. During the study, the proportion of patients with high scores became smaller, while the proportion with low scores increased (Fig. 5). At baseline, 146 patients (62.1%) had a severity score of 5 (markedly ill) or worse. At T6, only 75 (31.9%) remained with a score of 5 or 6. By contrast, at baseline only 12 patients (5.1%) scored as “mildly ill” (3), whereas at T6, 87 patients (37.0%) had a score of 3 or better. These improvements were found to be highly significant (p < 0.001) using Bhapkar’s test.

According to the CGI-S scores, 222 patients (94.5%) were stable (scores improved or remained the same) during the study. Altogether, 57 patients (24.3%) improved by two points or more (Fig. 6). Thirteen patients (5.4%) deteriorated, with a mean score of 1.3 points on the scale (four patients deteriorated by two points, another nine patients by one point each).

These improvements were also reflected by CGI-I scores: At T6, 35 (15.2%) of patients were rated as very much improved (score of 1) compared to baseline, and 96 (42.0%) were rated as much improved (score of 2).

In patients > 35 years, 18.8% (n = 28) improved by 2 points or more on the CGI-S scale, whereas in younger patients (≤35 years), this was true for 33.7% (n = 29) of patients. Only 8.8% (n = 13) of older patients had a CGI-I score of 1 (very much improved) at T6, whereas in younger patients, this was the case for 26.2% (n = 22). The difference between age groups regarding CGI-I scores at T6 was significant (p = 0.0074, Fisher’s exact test).

During the study, a total of 153 adverse events (AE) were reported by patients (Table 4) and coded according to MedDRA 19.0. One hundred thirty-three of these AEs were rated by the clinician as probably or possibly treatment-related (TRAE). Only one type of event, “medication taken at an inappropriate time”, applied to more than 5% of patients (24.8%). In most of these cases, oral aripiprazole had been discontinued earlier than recommended (< 2 weeks of the recommended concomitant treatment after first injection). Altogether, 52 TRAEs (39.1%) improved during the treatment period. Extrapyramidal symptoms and weight disturbances that were reported as treatment-related were rare (Table 4). No TRAEs were recorded related to sexual dysfunction, such as hyperprolactinemia. No patients died during the study.