The mean age of the cohort was 46.7 years (SD 16.0) at the start of follow-up, they had spent a mean of 8.8 years (SD 9.0) since their first schizophrenia diagnosis, and 50.3% (N = 31 104) were men. Oral olanzapine was the most commonly used antipsychotic (16 131 users), followed by risperidone (N = 13 083), clozapine (N = 11 828), and quetiapine (N = 10 838). Of specific dose categories of specific antipsychotic monotherapies, the most commonly used antipsychotic was oral olanzapine ≥1.6 DDDs/day, followed by risperidone <0.6 DDDs/day (Table 1). Distribution of used doses varied between specific antipsychotics. All dose categories of clozapine were used relatively frequently. Oral olanzapine was more commonly used with ≥1.6 DDDs/day dose (N = 11 351 users vs N = 7365 users of standard dose) whereas quetiapine and risperidone were more commonly used <0.6 DDDs/day dose than as standard dose (Table 1).
Compared to nonuse of antipsychotics within the same individual, 13 out of 15 antipsychotic monotherapies showed a U-shaped or J-shaped dose-response curve, showing the lowest aHR for relapse for standard dose of 0.9–1.1 DDDs/day (nine specific antipsychotics out of 13, namely oral levomepromazine, haloperidol oral and LAI, oral zuclopenthixol, oral clozapine, oral olanzapine, oral quetiapine, risperidone LAI and oral aripiprazole), or 0.6–<0.9 DDDs/day (four out of 13 specific antipsychotics, namely perphenazine LAI, chlorprothixene, zuclopenthixol LAI, oral risperidone) compared with nonuse of antipsychotics (Figure 1). The exceptions in dose categories with the lowest HRs were oral perphenazine (aHR = 0.72, 95% CI = 0.68–0.76, <0.6 DDDs/day) and olanzapine LAI (aHR = 0.17, 95% CI = 0.11–0.25, 1.4–<1.6 DDDs/day), the latter of which had the lowest aHR for any antipsychotic (54% lower than for second best, nonolanzapine LAI drug, which was risperidone LAI with a dose of 0.9–1.1 DDDs/day, HR = 0.37, 95% CI = 0.33–0.41), P < .0001 for difference. In addition to the dose of 1.4–1.6 DDD/day, all doses of olanzapine LAI above 0.9 DDDs/day were associated with excellent outcomes (aHRs below 0.25). Oral risperidone dose 0.6–0.9 DDDs/day was associated with 21% lower risk of hospitalization compared with standard dose 0.9–1.1 DDD/day (aHR = 0.54, 95% CI = 0.51–0.58 vs aHR = 0.68, 95% CI = 0.62–0.74), P < .001 for difference.
Adjusted Hazard Ratios for 15 most widely used antipsychotics in monotherapy, reference: nonuse of antipsychotics. Note the different scale for perphenazine oral.
High doses (>1.6 DDDs/day, corresponding to >48 mg/day) of oral perphenazine (aHR=3.35, 95% CI = 2.37–4.73), risperidone LAI (>60 mg/2 weeks; aHR 1.41, 95% CI = 1.25–1.59), and oral risperidone (>8 mg/day; aHR = 1.34, 95% CI = 1.22–1.47) were associated with substantially higher risk of rehospitalization due to schizophrenia than nonuse of antipsychotic (Table 1). Concerning oral perphenazine, even standard dose (0.9–<1.1 DDDs/day) was associated with higher risk of relapse (aHR = 1.32, 95% CI = 1.09–1.61), than nonuse of antipsychotics. Sensitivity analyses censoring the first 30 days of all dose categories showed similarly shaped dose-response curves (see Supplementary Table 2).
In head-to-head comparison (including only those patients having used standard dose) using standard dose of each antipsychotic as a reference, similar results were observed (Supplementary Table 3). All antipsychotics with dose ≥1.6 DDDs/day were associated with higher risk of relapse than standard dose of the same drug. For most common oral antipsychotics, low dose (<0.6 DDDs/day) of clozapine, olanzapine, and quetiapine were associated with higher risk of relapse whereas no difference was found for low dose risperidone compared with standard dose of the same drug. A statistically significant superior outcome was seen for the lowest doses of perphenazine (37%–46% lower risk of rehospitalization compared with standard dose), and 0.6–0.9 DDD/day dose for risperidone (20% lower risk). Also, head-to-head comparison with the most commonly used antipsychotic and dose, oral olanzapine >1.6 DDD (ie, including only those patients who had used high-dose oral olanzapine), showed very similar results as the primary analysis (Supplementary Table 4).
The superior outcome for olanzapine LAI 1.4–<1.6 DDDs/day was confirmed in the secondary between-individual analyses (aHR = 0.37, 95% CI = 0.21–0.66, being the lowest aHR for rehospitalization) when compared with the most common antipsychotic/dose category combination oral olanzapine >1.6 DDDs/day (Supplementary Table 5).
Sensitivity analysis stratified on the basis of age, schizoaffective diagnosis, and first-episode status were well in line with the primary analysis (see Supplementary Tables 6, 7, and 8). In the first-episode cohort, the results on high-dose perhenazine (aHR = 6.12, 95% CI = 1.17–32.03), risperidone (aHR = 2.00, 95% CI = 1.46–2.73), and olanzapine LAI (aHR = 0.05, 95% CI = 0.01–0.16) were even more extreme than in the primary analysis.
Schizophr Bull. 2022;48(4):774-784. © 2022 Oxford University Press