Patient population and treatment
In total, 1076 patients (rollover, n = 706; de novo, n = 370) provided informed consent for study participation, of whom 1002 entered the open-label treatment phase (rollover, n = 706; de novo, n = 296). Patient disposition is shown in and the characteristics of patients included in the safety analyses are shown in . In total, 323 patients (32.2%) completed 52 weeks of open-label treatment; completion rates were similar between the rollover and de novo groups (). Overall, the most common reasons for withdrawal from the open-label treatment phase were adverse events (23.0%), lack of efficacy/relapse (13.5%), and withdrawal of consent (12.5%). Time to discontinuation of aripiprazole due to any reason during open-label treatment is shown in .
Enrollment, randomization, and disposition of patients.
Open-label baselinea demographic characteristics (safety sample)
Time to discontinuation of aripiprazole due to any reason during open-label treatment, safety sample.*
Treatment and dosing
The distribution of antidepressant therapy at study endpoint (n = 984) was consistent with the distribution at open-label study baseline and was as follows: escitalopram, n = 275 (27.7%); venlafaxine extended-release, n = 249 (25.1%); sertraline, n = 171 (17.2%); fluoxetine, n = 143 (14.4%); paroxetine controlled-release, n = 61 (6.1%); paroxetine, n = 29 (2. 9%); bupropion extended-release, n = 35 (3.5%); bupropion sustained-release, n = 11 (1.1%); duloxetine, n = 7 (0.7%); and mirtazapine, n = 3 (0.3%).
At endpoint (n = 987), the mean dose of aripiprazole was 10.1 mg/day for the total population. During the last four-weekly dosing interval during the open-label phase (open-label treatment weeks 48–52, n = 320), the distribution of adjunctive aripiprazole dosing was as follows: 2 mg/day, 10.9%; 5 mg/day, 25.6%; 10 mg/day, 28.8%; 15 mg/day, 20.3%; 20 mg/day, 7.2%; and >20 mg/day, 7.2%.
The most commonly used (>5% of patients) concomitant central nervous system medications during open-label treatment were other analgesics and antipyretics (57.0%), anticholinergics (10.6%), opioids (9.2%), hypnotics and sedatives (7.5%), and anxiolytics (7.2%). Overall, 15.2% of patients received concomitant medication for the potential treatment of extrapyramidal symptoms. These included propranolol (5.3%), amantadine (0.1%), benztropine (10.6%), and trihexyphenidyl (0.1%).
During long-term treatment, 931 (93.7%) patients experienced at least one adverse event. Treatment-emergent adverse events that occurred at an incidence ≥10% are shown in . The most common (>15% of the total population) adverse events with long-term adjunctive aripiprazole treatment were akathisia (26.2%), fatigue (18.0%), and weight increase (17.1%). The majority (75.2%) of treatment-emergent adverse events were mild or moderate in nature.
Treatment-emergent adverse events (≥10% of patients, safety sample)
Overall, 226 (22.7%) patients in the safety sample discontinued study treatment due to adverse events; the rate of discontinuation was 23.7% for aripiprazole rollover patients and 22.4% for the placebo rollover/de novo patients. The most common adverse events leading to discontinuation (>1% of total population) were weight increase (3.3%), akathisia (3.3%), somnolence (2.0%), anxiety (1.7%), fatigue (1.7%), and sedation (1.1%); no other adverse events resulted in discontinuation of more than 1% of patients.
The incidence of serious adverse events was 4.0%; five serious adverse events occurred in two placebo rollover/de novo patients during long-term treatment (suicidal ideation, depression, chest pain, myocardial infarction, and intentional overdose); cellulitis, cholecystitis, and pneumonia were each also experienced by two patients (one placebo rollover/de novo patient and one aripiprazole rollover patient). There were no reports of neuroleptic malignant syndrome, completed suicide, or death due to other causes in this study.
The rates of reported extrapyramidal symptom-related adverse events were as follows: dystonic events, 4.0%; parkinsonian events, 10.6%; akathisia events (ie, akathisia, psychomotor hyperactivity), 26.5%; dyskinetic events, 1.8%; and residual events, 2.2%. Akathisia occurred in 26.2% of this open-label population (placebo rollover/de novo patients, 24.0%; aripiprazole rollover patients, 31.8%) and the majority of cases had their onset within the first six weeks of treatment. Akathisia led to discontinuation in 3.3% of study patients. There were small mean changes from baseline in the AIMS, SAS, and BARS scores; mean change from baseline to week 52 (last observation carried forward) was 0.06 for the AIMS total score (mean baseline 0.07), 0.17 points for the SAS (mean baseline 10.32), and 0.11 for the BARS (mean baseline 0.15).
Four spontaneous reports of possible tardive dyskinesia were submitted during the study (0.4%); all the patients involved were receiving aripiprazole adjunctive to escitalopram. Interventions to manage tardive dyskinesia included dose reduction (n = 2) and drug discontinuation (n = 2). In all four cases, highest AIMS total scores were ≤4, and the symptoms completely resolved within 45 days of discontinuing adjunctive aripiprazole treatment.
At week 26/32 (n = 491), the mean change in body weight from baseline (observed case) was 3.6 kg (placebo rollover/de novo patients, 3.4 kg; aripiprazole rollover patients, 4.3 kg). For the patients who completed 52/58 weeks of open-label adjunctive treatment (n = 303), the mean change in body weight over time showed that most weight gain occurred in the first 26–32 weeks on treatment. At week 26/32, the mean change in body weight was 4.0 kg, and at week 52/58 was 4.4 kg (placebo rollover/de novo patients, 3.9 kg at week 26, and 4.3 kg at week 52; aripiprazole rollover patients, 4.4 kg at week 32 and 4.9 kg at week 58). Mean change in body weight for aripiprazole rollover patients completing the study at week 6 was 1.8 kg. For aripiprazole rollover patients who had clinically significant weight gain (≥7%) at week 6, the mean change in body weight from baseline (observed case) was 8.7 kg at week 32 and 8.5 kg at week 58 (both n = 7). In aripiprazole rollover patients without clinically significant weight gain at week 6, mean change in body weight from baseline (observed case) was 4.0 kg (n = 106) and 4.5 kg (n = 64) at weeks 32 and 58, respectively.
Clinically significant weight gain (≥7%) occurred in 28.0% of patients (based on a last observation carried forward analysis) and 36.6% of subjects who completed the study (observed case analysis).
shows the mean change in fasting metabolic parameters from adjunctive aripiprazole baseline, by treatment period, for patients who continued to receive treatment (observed case analysis). National Cholesterol Education Program (NCEP)-defined cut-offs (see ) show that, on average, LDL, HDL, and glucose levels remained within normal limits. Baseline levels of cholesterol were higher, but tended to decrease over the 52-week exposure to aripiprazole. Baseline triglyceride levels were also above the normal 150 mg/dL criterion and remained above baseline throughout the course of aripiprazole treatment. Overall, there were no clinically important findings in the mean changes from baseline in fasting cholesterol, HDL, LDL, triglycerides, or glucose. For adjunctive aripiprazole exposure >46 weeks, the median change (range) in fasting metabolic parameters from adjunctive aripiprazole baseline were as follows: fasting cholesterol (n = 264), −4.0 (−154.0 to 90.0) mg/dL; fasting HDL (n = 264), −5.0 (−85.0 to 24.0) mg/dL; fasting LDL (n = 264), 1.0 (−117.0 to 94.0) mg/dL; fasting triglycerides (n = 264), 8.0 (−385.0 to 354.0) mg/dL, and fasting glucose (n = 262), 2.5 (−128.0 to 151.0) mg/dL.
Time-course of fasting metabolic mean changes from baseline (observed case analysis).
The incidence of treatment-emergent potentially clinically relevant abnormalities in fasting total cholesterol levels (≥240 mg/dL) was 11.9% (n = 33/277) with adjunctive aripiprazole and 13.6% (n = 38/280) with adjunctive placebo during placebo-controlled trials and was 19.4% (n = 135/697) for a pooled population of patients treated with aripiprazole in both the short-term studies and this long-term trial. The incidence of treatment-emergent potentially clinically relevant abnormalities in fasting glucose levels (≥126 mg/dL) was 2.2% (n = 8/358) with adjunctive aripiprazole and 2.8% (n = 10/362) with adjunctive placebo during placebo-controlled trials, and was 4.9% (n = 44/906) for the pooled total population from short-term and long-term studies.
Vital signs and laboratory findings
The incidence of potentially clinically relevant vital sign and electrocardiographic abnormalities, as well as serum chemistry, hematology, and electrolyte measurements, was low. The exception to this was the incidence of potentially clinically relevant prolactin elevations at at least one blood draw (13.7%). For a number of these patients, there were relevant prolactin elevations at study baseline and aripiprazole treatment was associated with decreased prolactin levels. At week 52 (last observation carried forward), patients showed a mean decrease of 0.5 ng/dL in serum prolactin levels from baseline (12.9 ng/dL).
Mean CGI-S scores over the course of treatment showed sustained improvement in clinical symptoms, regardless of treatment during the double-blind study (). At baseline, 33 patients (10.4%) had a CGI-S score of 1 or 2, indicating “normal” or “borderline ill” at adjunctive aripiprazole baseline, whereas at the end of open-label treatment (week 52/58) 221 patients (69.7%) had a CGI-S score of 1 or 2.
Mean CGI-S scores by treatment week.