Effectiveness of amisulpride augmentation versus other antipsychotics in clozapine-treated adolescents: a real-world retrospective study Authors: Binay Kayan Ocakoglu, Ayşegül Tonyalı, Uğur Karabağ, Zeynep Gültekin Taş, Damla Kasap Üstündağ, Gül Karaçetin Published: 2026 ABSTRACT BackgroundIn this study, we compared the clinical response to clozapine-amisulpride augmentation versus clozapine augmentation with other antipsychotics among children and adolescents with schizophrenia spectrum disorders in clinical practice. Crucially, the comparative efficacy of specific augmentation strategies in this vulnerable, early-onset population remains a largely unanswered question in current literature.MethodsSeventy-six inpatients who were treated with clozapine between 2013 and 2023 years at a tertiary care psychiatric university teaching hospital in Turkey were included in this retrospective study. Patients were divided into clozapine+amisulpride (n = 22) and clozapine+others (n = 54) groups according to their treatment regimen. Clinical response was assessed by comparing baseline and discharge Positive and Negative Syndrome Scale (PANSS) scores and response rates defined as ≥20% and ≥30% improvement. The effect of group assignment on clinical response was examined by multiple linear regression analysis while controlling for clozapine dose, baseline PANSS scores, and adjunctive antipsychotic load standardized via the Defined Daily Dose (DDD) method.ResultsMean age of the patients was 15.9 ± 2.3 years. Baseline PANSS total scores were similar between groups (p = 0.977). Mean changes in PANSS total scores (−25.1±34.7 vs. −25.5±32.1; p = 0.968) and subscale scores were not significantly different between groups at discharge. Response rates according to 30% and 20% thresholds were similar between groups (38.1% vs. 51.1%; p = 0.324, p = 0.954). In the multiple linear regression analysis, clozapine dose (p = 0.026) and baseline PANSS score (p < 0.001) were significant predictors of change in PANSS total score. In the negative subscale model, higher clozapine doses predicted greater symptom reduction (p = 0.028), whereas higher standardized doses of adjunctive antipsychotics predicted poorer improvement (p = 0.031).ConclusionsClozapine augmentation with amisulpride was not superior to augmentation with other antipsychotics in our sample. Treatment response appears to be driven more by clozapine dose and baseline severity. In addition, high standardized doses of adjunctive agents may counterproductively limit negative symptom recovery, perhaps due to secondary negative symptoms caused by excessive D2 receptor blockade.Clinical trial numberNot applicable.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12888-026-08548-2. Background Schizophrenia spectrum disorders (SSD) are severe psychiatric conditions associated with substantial functional disability in youth [1]. Although clozapine is the the gold standard for refractory schizophrenia, many patients still have residual symptoms despite adequate monotherapy. This condition, called clozapine-resistant schizophrenia [2–4], often needs augmentation with a second antipsychotic agent to address this partial response. A large body of evidence has focused on the use of amisulpride, a second-generation antipsychotic agent, among a host of potential augmentation strategies [3, 5–9]. Because clozapine has modest D2 receptor blockade, augmentation with a potent D2 antagonist such as amisulpride theoretically leads to a more robust antipsychotic response [10, 11]. In addition, low-dose amisulpride has an action on primary negative symptoms which is a major unmet need in the treatment of schizophrenia [12, 13]. Despite this strong theoretical basis, there is currently a significant lack of clinical evidence supporting the use of amisulpride as an augmentation strategy in adult populations. Several randomized controlled trials and open-label trials have shown significant benefit when compared with placebo [11, 14–16], while other large-scale, high-quality studies such as the AMICUS trial found no significant benefit [6, 7]. Recent meta-analyses have also highlighted this ambiguity. The conclusions of these analyses suggest that, as yet, no antipsychotic augmentation strategy, including amisulpride, is clearly superior [9, 17]. In fact, a recent review by Grover et al. highlights the negative evidence on antipsychotic polypharmacy and suggests that non-pharmacological approaches, especially electroconvulsive therapy (ECT), may be the most effective intervention for clozapine-resistant schizophrenia [18]. Importantly, there is a lack of evidence regarding clozapine augmentation in pediatric and adolescent populations. Thus, clinicians are required to extrapolate from adult studies, despite the fact that there are significant age-related differences in pharmacodynamics and susceptibility to side effects during this vulnerable developmental window. Moreover, most existing clinical trials (predominantly in adults) have involved direct comparison of amisulpride to a placebo [5–7, 11]. This design fails to capture the most relevant clinical question, which is whether augmentation with amisulpride is superior to augmentation with other commonly-used antipsychotics in this challenging-to-treat patient population. In light of this striking deficiency in the pediatric literature, the present study aimed to compare the real-world clinical effectiveness of clozapine-amisulpride augmentation versus clozapine augmentation with other antipsychotics. The research sought to assess the potential of amisulpride to produce superior clinical outcomes, as measured by PANSS score reductions and clinical response rates, in a tertiary inpatient cohort of children and adolescents with schizophrenia spectrum disorders. Materials and methods Data analysis All statistical analyses were performed in Python (version 3.x) using scipy.stats, scikit-learn and pandas packages. Continuous variables are presented as mean ± standard deviation (SD). Categorical variables are presented as frequencies and percentages. p value of 0.05 was used as the significance threshold. Mann–Whitney U tests were used to compare the continuous variables (PANSS scores, change scores and medication doses) between the clozapine + other antipsychotic and amisulpride groups. Chi-squared tests were used to compare the categorical outcomes, that is, the proportions of patients achieving ≥ 20% and ≥30% improvement in PANSS total scores between the treatment groups. Pearson correlation coefficients were calculated to examine the relationship between medication dose and treatment outcomes (PANSS change scores and percentage improvement) within the clozapine group. Multiple linear regression was used to examine percent change in PANSS scores while controlling for the same covariates. Cohen’s d was calculated for continuous variables to assess the clinical significance of differences between groups. Standardised doses were calculated using z-score transformation to enable comparison across different medications. Complete case analysis was performed. Patients with missing baseline, discharge or dose data were excluded from the relevant analyses. Study sample and baseline characteristics The study sample of this retrospective chart review consists of the medical records of the patients admitted to the acute inpatient child and adolescent psychiatry service of Bakırköy Prof. Dr. Mazhar Osman Mental Health and Neurological Diseases Training and Research Hospital, a tertiary care psychiatric hospital in Turkey. The study was approved by the Institutional Review Board (IRB) of Bakırköy Prof. Dr. Mazhar Osman Research and Training Hospital for Psychiatry, Neurology and Neurosurgery (Protocol No: 2022/160). All data were anonymized to ensure patient confidentiality. We screened 114 inpatients with clozapine between January 2013 and December 2023 and 76 adolescents met the inclusion criteria and were included in the final analysis. Inclusion and exclusion criteria Inclusion criteria were as follows:Admission to the inpatient unit and a diagnosis of a Schizophrenia Spectrum Disorder (SSD) (schizophrenia, schizophreniform disorder, or schizoaffective disorder) according to DSM-5 criteria.Age between 12-18 years.Treatment with clozapine augmentation (clozapine + 2nd antipsychotic) in response to partial response/resistance to clozapine monotherapy.Complete Positive and Negative Syndrome Scale (PANSS) scores at both admission (baseline) and discharge. Admission to the inpatient unit and a diagnosis of a Schizophrenia Spectrum Disorder (SSD) (schizophrenia, schizophreniform disorder, or schizoaffective disorder) according to DSM-5 criteria. Admission to the inpatient unit and a diagnosis of a Schizophrenia Spectrum Disorder (SSD) (schizophrenia, schizophreniform disorder, or schizoaffective disorder) according to DSM-5 criteria. Age between 12-18 years. Age between 12-18 years. Treatment with clozapine augmentation (clozapine + 2nd antipsychotic) in response to partial response/resistance to clozapine monotherapy. Treatment with clozapine augmentation (clozapine + 2nd antipsychotic) in response to partial response/resistance to clozapine monotherapy. Complete Positive and Negative Syndrome Scale (PANSS) scores at both admission (baseline) and discharge. Complete Positive and Negative Syndrome Scale (PANSS) scores at both admission (baseline) and discharge. Exclusion criteria:Diagnosis of intellectual disability or a primary organic brain disorder.Psychotic symptoms primarily due to substance use.Treatment with clozapine monotherapy (no adjunctive AP agent).Incomplete medical records or missing dosage information. Diagnosis of intellectual disability or a primary organic brain disorder. Diagnosis of intellectual disability or a primary organic brain disorder. Psychotic symptoms primarily due to substance use. Psychotic symptoms primarily due to substance use. Treatment with clozapine monotherapy (no adjunctive AP agent). Treatment with clozapine monotherapy (no adjunctive AP agent). Incomplete medical records or missing dosage information. Incomplete medical records or missing dosage information. The final sample (n = 76) was categorized into two groups according to the augmentation strategy: Clozapine + Amisulpride group (n = 22), Clozapine + Other AP group (n = 54) Diagnostic process and treatment protocol Diagnoses were made by trained child and adolescent psychiatrists based on semi-structured clinical interviews and long-term observation during hospitalization according to DSM-5. Augmentation was defined as adding a second antipsychotic agent to an existing clozapine regimen. To exclude real augmentation strategies from cross titration during medication switch, the following cut-off was set: all patients included received the adjunctive antipsychotic for a minimum concomitant treatment period of 2 weeks (considered to be a minimum period necessary to judge initial tolerability and clinical response in this retrospective cohort). In Türkiye, clozapine is approved for the treatment of treatment-resistant schizophrenia (TRS) and for patients with schizophrenia who develop severe, untreatable neurologic adverse reactions (e.g., extrapyramidal side effects) to other antipsychotics [19]. In line with these guidelines, all patients in our cohort were treated with clozapine initially for TRS. Standardizing antipsychotic dosages Because the adjunctive antipsychotics used in the ‘Clozapine + Other AP’ group may have different potencies, daily doses of adjunctive antipsychotics were converted into chlorpromazine (CPZ) equivalents. We used the Defined Daily Dose (DDD) method as suggested by Leucht et al. [20] to overcome methodological limitations of older conversion factors, in particular concerning second-generation antipsychotics. The estimated DDD-based CPZ equivalent for the adjunctive medications was then used as a continuous covariate in the regression models to correctly control the effect of different antipsychotic loads on treatment outcomes. Clinical assessments (PANSS) Symptom severity and treatment response were measured with the Positive and Negative Syndrome Scale (PANSS). PANSS is a standardized, 30-item rating scale that measures three domains [21, 22]:Positive Scale (7 items): rates hallucinations, delusions, and disorganized thinking.Negative Scale (7 items): rates blunted affect, emotional withdrawal, and poor rapport.General Psychopathology Scale (16 items): rates anxiety, depression, and other general symptoms. Positive Scale (7 items): rates hallucinations, delusions, and disorganized thinking. Positive Scale (7 items): rates hallucinations, delusions, and disorganized thinking. Negative Scale (7 items): rates blunted affect, emotional withdrawal, and poor rapport. Negative Scale (7 items): rates blunted affect, emotional withdrawal, and poor rapport. General Psychopathology Scale (16 items): rates anxiety, depression, and other general symptoms. General Psychopathology Scale (16 items): rates anxiety, depression, and other general symptoms. Ratings were done by trained child and adolescent psychiatry residents and child and adolescent psychiatry specialists as a part of the routine clinical protocol at admission (baseline) and discharge. The Turkish version of the PANSS has shown high validity and reliability [23]. Results Seventy-six patients were eligible for the final analyses. The sample was divided into two groups for comparison of the clozapine + amisulpride (n = 22) vs. clozapine + other AP (n = 54) treatment. The median length of inpatient hospitalization, which was taken as the observational treatment window for each participant, was 75 days (IQR: 39–103 days) for the entire sample. The length was comparable between the clozapine+amisulpride group (107 days, range: 23–177) and the clozapine+other AP group (83 days, range: 16–380) (p = 0.155). Table 1 displays the baseline demographic and clinical characteristics of the sample. The treatment groups were well-matched in terms of mean age (16.2 ± 1.3 vs. 15.8 ± 2.6 years; p = 0.521), education, and clinical diagnosis, with the amisulpride group having a significantly higher proportion of male participants (90.9% vs. 64.8%; p = 0.043). Cannabis use (27.3% vs 7.4%) and other substance use (27.3% vs 11.1%) was more common in the amisulpride augmentation group. The distribution of diagnoses of schizoaffective disorder, schizophrenia and schizophreniform disorder were not statistically different between the clozapine + amisulpride and clozapine + other AP treated patients. In each treatment group the most common diagnosis was schizophrenia. First-episode psychosis and catatonia were similar between the treatment groups.Table 1Demographic and clinical Characteristics of the study groupsCharacteristicClozapine + Amisulpride Group (n = 22)Clozapine + Other AP Group (n = 54)Total (n = 76)p-valueAge (years), Mean ± SD16.2 ± 1.315.8 ± 2.615.9 ± 2.30.521Education (years), Mean ± SD8.6 ± 1.58.3 ± 1.88.4 ± 1.70.552Gender, n (%)0.043 Male20 (90.9%)35 (64.8%)55 (72.4%) Female2 (9.1%)19 (35.2%)21 (27.6%)Educational Status, n (%)0.633 Primary School2 (9.1%)2 (3.7%)4 (5.3%) High School8 (36.4%)14 (25.9%)22 (28.9%) Open High School1 (4.5%)4 (7.4%)5 (6.6%) Dropout11 (50.0%)34 (63.0%)45 (59.2%)Cannabis Use, n (%)6 (27.3%)4 (7.4%)10 (13.2%)0.051Other Substance Use, n (%)6 (27.3%)6 (11.1%)12 (15.8%)0.610Diagnosis0.699 Schizophrenia17 (77.3%)41 (85.4%)58 (82.8%) Schizphreniform disorder3 (13.6%)4 (8.3%)7 (10.0%) Schizoaffective disorder2 (9.1%)3 (6.2%)5 (7.14%)First episode psychosis6 (37.5%)10 (62.5%)16 (100%)0.545Catatonia2 (33.3%)4 (66.6%)6 (100%)1.000Duration of Untreated Psychosis (months), median2.5 (0–24)4.0 (0–20)4 (0–24)0.522Number of Prior Antipsychotic Trials, Median (Min-Max)3 (2–5)3 (1–5)3 (1–5)0.359Data are presented as mean ± standard deviation (SD), median (min-max), or as count (n, %). Group comparisons were made using the Mann-Whitney U test for continuous variables and the Chi-square test for categorical variables. A p-value < 0.05 was considered statistically significant. Abbreviation: AP, Antipsychotic Table 1Demographic and clinical Characteristics of the study groupsCharacteristicClozapine + Amisulpride Group (n = 22)Clozapine + Other AP Group (n = 54)Total (n = 76)p-valueAge (years), Mean ± SD16.2 ± 1.315.8 ± 2.615.9 ± 2.30.521Education (years), Mean ± SD8.6 ± 1.58.3 ± 1.88.4 ± 1.70.552Gender, n (%)0.043 Male20 (90.9%)35 (64.8%)55 (72.4%) Female2 (9.1%)19 (35.2%)21 (27.6%)Educational Status, n (%)0.633 Primary School2 (9.1%)2 (3.7%)4 (5.3%) High School8 (36.4%)14 (25.9%)22 (28.9%) Open High School1 (4.5%)4 (7.4%)5 (6.6%) Dropout11 (50.0%)34 (63.0%)45 (59.2%)Cannabis Use, n (%)6 (27.3%)4 (7.4%)10 (13.2%)0.051Other Substance Use, n (%)6 (27.3%)6 (11.1%)12 (15.8%)0.610Diagnosis0.699 Schizophrenia17 (77.3%)41 (85.4%)58 (82.8%) Schizphreniform disorder3 (13.6%)4 (8.3%)7 (10.0%) Schizoaffective disorder2 (9.1%)3 (6.2%)5 (7.14%)First episode psychosis6 (37.5%)10 (62.5%)16 (100%)0.545Catatonia2 (33.3%)4 (66.6%)6 (100%)1.000Duration of Untreated Psychosis (months), median2.5 (0–24)4.0 (0–20)4 (0–24)0.522Number of Prior Antipsychotic Trials, Median (Min-Max)3 (2–5)3 (1–5)3 (1–5)0.359Data are presented as mean ± standard deviation (SD), median (min-max), or as count (n, %). Group comparisons were made using the Mann-Whitney U test for continuous variables and the Chi-square test for categorical variables. A p-value < 0.05 was considered statistically significant. Abbreviation: AP, Antipsychotic Demographic and clinical Characteristics of the study groups Data are presented as mean ± standard deviation (SD), median (min-max), or as count (n, %). Group comparisons were made using the Mann-Whitney U test for continuous variables and the Chi-square test for categorical variables. A p-value < 0.05 was considered statistically significant. Abbreviation: AP, Antipsychotic Regarding the cohort’s clinical history, no statistically significant differences were observed between the clozapine+amisulpride and clozapine+other AP groups in median duration of untreated psychosis (DUP) (2.5 vs. 4.0 months; p = 0.522) or median number of previous antipsychotic trials prior to clozapine initiation (3 trials in both groups; p = 0.359) (Table 1). For clozapine initiation, 13.2% of the patients (n = 10) were on clozapine therapy prior to their current admission. The majority of the sample (86.8%, n = 66) started clozapine during their index hospitalization (n = 2 on the day of admission, and n = 64 during the subsequent inpatient stay). The median duration of treatment with clozapine was 51.5 days (range 27–142) in the clozapine + amisulpride group and 57.0 days (range 14–312) in the clozapine + other AP group. The Mann-Whitney U test confirmed that there was no significant difference in the treatment duration between the cohorts (p = 0.954). Clinical outcomes Comparisons of clinical outcomes indicated that there were no statistically significant differences in baseline symptom severity or clinical improvement at discharge between treatment groups (Table 2). Baseline PANSS total scores were very similar between the amisulpride and other AP groups (86.2 ± 29.3 vs. 86.4 ± 26.6, respectively; p = 0.977; Cohen’s d = 0.007). The lack of a significant between-group difference at baseline also applied to the Positive, Negative and General Psychopathology subscales.Clozapine + amisulpride augmentation was not significantly different from augmentation with other atypical antipsychotics with respect to the improvement from baseline to discharge, as assessed by change in PANSS total score (−25.1 ± 34.7 vs. −25.5 ± 32.1; p = 0.968; ‘Cohen’s d = 0.012). Effect sizes for the change scores on PANSS subscales were similarly negligible (Positive: d = 0.01, Negative: d = 0.10, General: d = 0.09).Table 2Comparison of clinical outcomes and treatment doses between augmentation groupsMeasureClozapine + Amisulpride (n = 22)Clozapine + Other AP (n = 54)Statistical Testp-value/ Cohen’s dPANSS Total Score Baseline86.2 ± 29.386.4 ± 26.6−0.0290.977 Discharge61.0 ± 20.060.9 ± 25.50.0250.980 Change from Baseline−25.1 ± 34.7−25.5 ± 32.1−0.040.968/ 0.01PANSS Positive Baseline20.3 ± 8.919.6 ± 7.60.2910.773 Discharge12.8 ± 4.812.1 ± 5.60.5460.587 Change from Baseline−7.4 ± 11.2−7.5 ± 8.7−0.0280.977/ 0.01PANSS Negative Baseline25.7 ± 9.925.9 ± 10.7−0.0750.940 Discharge17.9 ± 8.319.2 ± 10.2−0.5800.564 Change from Baseline−7.8 ± 9.2−6.7 ± 11.70.4410.661/ 0.10PANSS General Baseline40.2 ± 13.840.8 ± 13.4−0.1850.854 Discharge30.3 ± 10.129.6 ± 12.50.2720.787 Change from Baseline−9.9 ± 16.2−11.3 ± 15.4−0.3440.733/ 0.09Medication Doses (mg/day) Clozapine Dose360.0 ± 117.0342.0 ± 122.00.7290.562/ 0.15 Amisulpride Dose565.2 ± 210.8N/A-- Other AP Dose (CPZ-eq.)N/A502 ± 1247a--Values are presented as Mean ± Standard Deviation (SD) for continuous variables or as percentage (%) for categorical variables. Negative change scores indicate clinical improvement. A p-value less than 0.05 was considered statistically significant (*). aCalculated using the Defined Daily Dose (DDD) method by Leucht et al. [20]. Abbreviations:AP, Antipsychotic; PANSS, Positive and Negative Syndrome Scale; SD, Standard Deviation. Cohen’s d values < 0.2 indicate negligible effect size Table 2Comparison of clinical outcomes and treatment doses between augmentation groupsMeasureClozapine + Amisulpride (n = 22)Clozapine + Other AP (n = 54)Statistical Testp-value/ Cohen’s dPANSS Total Score Baseline86.2 ± 29.386.4 ± 26.6−0.0290.977 Discharge61.0 ± 20.060.9 ± 25.50.0250.980 Change from Baseline−25.1 ± 34.7−25.5 ± 32.1−0.040.968/ 0.01PANSS Positive Baseline20.3 ± 8.919.6 ± 7.60.2910.773 Discharge12.8 ± 4.812.1 ± 5.60.5460.587 Change from Baseline−7.4 ± 11.2−7.5 ± 8.7−0.0280.977/ 0.01PANSS Negative Baseline25.7 ± 9.925.9 ± 10.7−0.0750.940 Discharge17.9 ± 8.319.2 ± 10.2−0.5800.564 Change from Baseline−7.8 ± 9.2−6.7 ± 11.70.4410.661/ 0.10PANSS General Baseline40.2 ± 13.840.8 ± 13.4−0.1850.854 Discharge30.3 ± 10.129.6 ± 12.50.2720.787 Change from Baseline−9.9 ± 16.2−11.3 ± 15.4−0.3440.733/ 0.09Medication Doses (mg/day) Clozapine Dose360.0 ± 117.0342.0 ± 122.00.7290.562/ 0.15 Amisulpride Dose565.2 ± 210.8N/A-- Other AP Dose (CPZ-eq.)N/A502 ± 1247a--Values are presented as Mean ± Standard Deviation (SD) for continuous variables or as percentage (%) for categorical variables. Negative change scores indicate clinical improvement. A p-value less than 0.05 was considered statistically significant (*). aCalculated using the Defined Daily Dose (DDD) method by Leucht et al. [20]. Abbreviations:AP, Antipsychotic; PANSS, Positive and Negative Syndrome Scale; SD, Standard Deviation. Cohen’s d values < 0.2 indicate negligible effect size Comparison of clinical outcomes and treatment doses between augmentation groups Values are presented as Mean ± Standard Deviation (SD) for continuous variables or as percentage (%) for categorical variables. Negative change scores indicate clinical improvement. A p-value less than 0.05 was considered statistically significant (*). aCalculated using the Defined Daily Dose (DDD) method by Leucht et al. [20]. Abbreviations:AP, Antipsychotic; PANSS, Positive and Negative Syndrome Scale; SD, Standard Deviation. Cohen’s d values < 0.2 indicate negligible effect size Response rate (proportion of patients reaching a predefined clinical response threshold) was also assessed for the proportion of patients with a ≥ 20% or ≥30% reduction in PANSS scores, and was compared between the two treatment groups. Detailed comparisons (number of responders, response rates, odds ratios) are reported in Table 3. Figure 1 shows the clinical response rates at the ≥30% improvement in PANSS threshold. Response rates for the positive, negative and general psychopathology subscales were comparable in the clozapine + amisulpride and clozapine + other atypical AP groups. The response rate on the total PANSS score was 38.1% in the amisulpride group and 51.1% in the other atypical AP group, a non-statistically significant difference (p = 0.324). The response rate with a 20% threshold was essentially identical in the clozapine + amisulpride group (66.7%) and clozapine + other AP group (66.0%).Fig. 1Comparison of clinical response rates (≥20 and 30% improvement) by treatment group, this chart displays the percentage of patients who achieved a clinical response, defined as a ≥ 20 and 30% improvement on the PANSS total and subscale scores, for both the clozapine+amisulpride and clozapine+other AP groups. While response rates were broadly similar across the subscales, the most notable numerical difference was observed in the PANSS total score Comparison of clinical response rates (≥20 and 30% improvement) by treatment group, this chart displays the percentage of patients who achieved a clinical response, defined as a ≥ 20 and 30% improvement on the PANSS total and subscale scores, for both the clozapine+amisulpride and clozapine+other AP groups. While response rates were broadly similar across the subscales, the most notable numerical difference was observed in the PANSS total score Table 3Comparison of Clinical Response Rates Between Treatment Groups at 20 and 30% Improvement ThresholdsPANSS DomainImprovement ThresholdAmisulpride GroupOther AP GroupOdds Ratio (95% CI)p-valueTotal Score≥20%66.7%66.0%0.97 (0.31 - 2.84)0.954≥30%38.1%51.1%1.70 (0.60 - 5.01)0.324Positive Symptoms≥20%52.4%66.0%1.76 (0.61 - 5.08)0.290≥30%52.4%53.2%1.03 (0.36 - 2.91)0.951Negative Symptoms≥20%71.4%61.7%0.64 (0.20 - 1.91)0.440≥30%38.1%42.6%1.20 (0.42 - 3.56)0.730General Psychopathology≥20%61.9%59.6%0.91 (0.31 - 2.58)0.856≥30%47.6%46.8%0.97 (0.34 - 2.74)0.951Data are presented as Responders %. Odds Ratio (OR) is for the clozapine+other AP group compared to the amisulpride+clozapine group (reference). An OR > 1 suggests an advantage for the “clozapine+other AP “strategy; an OR < 1 suggests an advantage for the amisulpride+clozapine group strategy. Abbreviations:AP, Antipsychotic; CI, Confidence Interval; PANSS, Positive and Negative Syndrome Scale Table 3Comparison of Clinical Response Rates Between Treatment Groups at 20 and 30% Improvement ThresholdsPANSS DomainImprovement ThresholdAmisulpride GroupOther AP GroupOdds Ratio (95% CI)p-valueTotal Score≥20%66.7%66.0%0.97 (0.31 - 2.84)0.954≥30%38.1%51.1%1.70 (0.60 - 5.01)0.324Positive Symptoms≥20%52.4%66.0%1.76 (0.61 - 5.08)0.290≥30%52.4%53.2%1.03 (0.36 - 2.91)0.951Negative Symptoms≥20%71.4%61.7%0.64 (0.20 - 1.91)0.440≥30%38.1%42.6%1.20 (0.42 - 3.56)0.730General Psychopathology≥20%61.9%59.6%0.91 (0.31 - 2.58)0.856≥30%47.6%46.8%0.97 (0.34 - 2.74)0.951Data are presented as Responders %. Odds Ratio (OR) is for the clozapine+other AP group compared to the amisulpride+clozapine group (reference). An OR > 1 suggests an advantage for the “clozapine+other AP “strategy; an OR < 1 suggests an advantage for the amisulpride+clozapine group strategy. Abbreviations:AP, Antipsychotic; CI, Confidence Interval; PANSS, Positive and Negative Syndrome Scale Comparison of Clinical Response Rates Between Treatment Groups at 20 and 30% Improvement Thresholds Data are presented as Responders %. Odds Ratio (OR) is for the clozapine+other AP group compared to the amisulpride+clozapine group (reference). An OR > 1 suggests an advantage for the “clozapine+other AP “strategy; an OR < 1 suggests an advantage for the amisulpride+clozapine group strategy. Abbreviations:AP, Antipsychotic; CI, Confidence Interval; PANSS, Positive and Negative Syndrome Scale Across all of the subscale analyses (positive, negative and general), the response rates at both the 20 and 30% thresholds differed only modestly between the two groups. None of these comparisons reached statistical significance. Overall, these results support the conclusion that augmentation with amisulpride and other antipsychotics results in comparable clinical response rates in all domains. Medication doses and other APs The mean daily clozapine dose was not significantly different between groups (360.0 ± 117.0 mg vs. 342.0 ± 122.0 mg; p = 0.562; Cohen’s d = 0.15)). The mean dose of the augmenting agent was 565.2 ± 210.8 mg/day in the amisulpride group. The mean chlorpromazine (CPZ) equivalent dose of the augmenting agents in the other AP group was 502 ± 1247 mg/day. Table 2 shows a summary of all the clinical and medication dose comparisons. In the clozapine plus other antipsychotics group (n = 54) adjunctive agents prescribed were: risperidone (n = 22, 43.1%), quetiapine (n = 24, 47.1%), aripiprazole (n = 12, 23.5%), olanzapine (n = 1, 2%) haloperidol (n = 6, 11.8%), and chlorpromazine (n = 1, 2%). Treatment-emergent adverse effects during clozapin For safety and tolerability, adverse effects that were specifically monitored and documented as related to clozapine during the combined treatment period were extracted from the clinical records. The most frequent adverse effects of clozapine in the cohort (N = 76) included sedation (n = 23, 32.9%), hypersalivation (n = 30, 42.9%), EPS (n = 50, 71.4%), fasciculation (n = 27, 38.6%) and eosinophilia (n = 15, 21.4%). Supplementary Table S1 lists a comprehensive summary of all adverse effects reported for clozapine, including cardiovascular and haematological parameters. Predictors of treatment response A multiple linear regression analysis was conducted in order to ascertain the factors that predict treatment response, with the change in PANSS total score designated as the dependent variable. The comprehensive results of this analysis are presented in Table 4.Table 4Multiple linear regression model predicting change in PANSS scoresOutcome (Dependent Variable)Predictor VariableBSEt-statisticp-valuePANSS Total Score(Constant)33.316.032.080.041Group (Other AP vs. Amisulpride)−2,254.43−0.500.61Clozapine Dose−0.0440.019−2.270.026*Other AP Dose (DDD equivalent)0.0020.00181.240.21Baseline Total Score−0.380.11−3.33<0.001*PANSS Positive Subscale(Constant)7.134.151.710.091Group (Other AP vs. Amisulpride)−0.0731.56−0.040.962Clozapine Dose−0,0070,007−1,0450,300Other AP Dose (DDD equivalent)0,0000,001−0,1250,901Baseline Positive Score−0,4510,091−4,961<0.001*PANSS Negative Subscale(Constant)−0,1424,744−0,0300,976Group (Other AP vs. Amisulpride)−0,3811,561−0,2440,808Clozapine Dose−0,0160,007−2,2590,028*Other AP Dose (DDD equivalent)0,0010,0012,2100,031*Baseline Negative Score−0,0560,091−0,6130,542PANSS General Subscale(Constant)−6,1095,747−1,0630,292Group (Other AP vs. Amisulpride)−2,9892,561−1,1670,248Clozapine Dose0,0020,0120,1880,851Other AP Dose (DDD equivalent)0,0010,0011,1410,259Baseline General Score−0,0420,025−1,7110,092Note: Table displays the results of four separate multiple linear regression models. A p-value < 0.05 was considered statistically significant (*) •PANSS Total Score Model: F(4, 59) = 5.27, p = 0.004; R2 = 0.226, Adjusted R2 = 0.173 •PANSS Positive Subscale Model: F(4, 58) = 6.85, p < 0.001; R2 = 0.321, Adjusted R2 = 0.274 •PANSS Negative Subscale Model: F(4, 58) = 3.52, p = 0.044; R2 = 0.152, Adjusted R2 = 0.118 •PANSS General Subscale Model: F(4, 71) = 4.41, p = 0.003; R2 = 0.199, Adjusted R2 = 0.0931 The Group variable represents the Clozapine + Other AP group compared to the reference group (Clozapine + Amisulpride). Abbreviations:B, Unstandardized Coefficient; SE, Standard Error; AP, Antipsychotic; PANSS, Positive and Negative Syndrome Scale Table 4Multiple linear regression model predicting change in PANSS scoresOutcome (Dependent Variable)Predictor VariableBSEt-statisticp-valuePANSS Total Score(Constant)33.316.032.080.041Group (Other AP vs. Amisulpride)−2,254.43−0.500.61Clozapine Dose−0.0440.019−2.270.026*Other AP Dose (DDD equivalent)0.0020.00181.240.21Baseline Total Score−0.380.11−3.33<0.001*PANSS Positive Subscale(Constant)7.134.151.710.091Group (Other AP vs. Amisulpride)−0.0731.56−0.040.962Clozapine Dose−0,0070,007−1,0450,300Other AP Dose (DDD equivalent)0,0000,001−0,1250,901Baseline Positive Score−0,4510,091−4,961<0.001*PANSS Negative Subscale(Constant)−0,1424,744−0,0300,976Group (Other AP vs. Amisulpride)−0,3811,561−0,2440,808Clozapine Dose−0,0160,007−2,2590,028*Other AP Dose (DDD equivalent)0,0010,0012,2100,031*Baseline Negative Score−0,0560,091−0,6130,542PANSS General Subscale(Constant)−6,1095,747−1,0630,292Group (Other AP vs. Amisulpride)−2,9892,561−1,1670,248Clozapine Dose0,0020,0120,1880,851Other AP Dose (DDD equivalent)0,0010,0011,1410,259Baseline General Score−0,0420,025−1,7110,092Note: Table displays the results of four separate multiple linear regression models. A p-value < 0.05 was considered statistically significant (*) •PANSS Total Score Model: F(4, 59) = 5.27, p = 0.004; R2 = 0.226, Adjusted R2 = 0.173 •PANSS Positive Subscale Model: F(4, 58) = 6.85, p < 0.001; R2 = 0.321, Adjusted R2 = 0.274 •PANSS Negative Subscale Model: F(4, 58) = 3.52, p = 0.044; R2 = 0.152, Adjusted R2 = 0.118 •PANSS General Subscale Model: F(4, 71) = 4.41, p = 0.003; R2 = 0.199, Adjusted R2 = 0.0931 The Group variable represents the Clozapine + Other AP group compared to the reference group (Clozapine + Amisulpride). Abbreviations:B, Unstandardized Coefficient; SE, Standard Error; AP, Antipsychotic; PANSS, Positive and Negative Syndrome Scale Multiple linear regression model predicting change in PANSS scores Note: Table displays the results of four separate multiple linear regression models. A p-value < 0.05 was considered statistically significant (*) •PANSS Total Score Model: F(4, 59) = 5.27, p = 0.004; R2 = 0.226, Adjusted R2 = 0.173 •PANSS Positive Subscale Model: F(4, 58) = 6.85, p < 0.001; R2 = 0.321, Adjusted R2 = 0.274 •PANSS Negative Subscale Model: F(4, 58) = 3.52, p = 0.044; R2 = 0.152, Adjusted R2 = 0.118 •PANSS General Subscale Model: F(4, 71) = 4.41, p = 0.003; R2 = 0.199, Adjusted R2 = 0.093 1 The Group variable represents the Clozapine + Other AP group compared to the reference group (Clozapine + Amisulpride). Abbreviations:B, Unstandardized Coefficient; SE, Standard Error; AP, Antipsychotic; PANSS, Positive and Negative Syndrome Scale Predicting treatment response factors Before interpreting the regression models, the multicollinearity assumptions were examined. The multicollinearity diagnostics showed that there was no multicollinearity that could pose a problem. All predictor variables had Variance Inflation Factor (VIF) values well below the recommended threshold of 5 in both the total and subscale models. There was no indication of problematic multicollinearity in the baseline and discharge scores. The overall regression model was significant (F(4, 59) = 5.27, p = 0.004) and explained 22.6% of the total variance of PANSS score change (R2 = 0.226). Significant predictors were baseline PANSS score (B = −0.38, p < 0.001) and clozapine dose (B = −0.044, p = 0.026). The analysis was primarily designed to ascertain if augmentation with amisulpride was related to a different outcome. After adjusting for clozapine dose, adjunctive antipsychotic dose (in DDD equivalents), and baseline severity, the model indicated that treatment group was not a significant predictor of PANSS score change (B = −2.25, p = 0.610). This finding implies that amisulpride augmentation does not produce a statistically significant difference in the level of symptom reduction relative to other atypical antipsychotics in this model. Subscales analyses To more specifically test treatment effects on different symptom domains, we performed separate multiple linear regression analyses with change in the Positive, Negative and General Psychopathology subscales of the PANSS as dependent variables. Full regression results are shown in Table 3. All three subscale models were significant (p < 0.05). The model for the Positive subscale accounted for 32.1% of the variance in change in this subscale (R2 = 0.321) and the model for the General Psychopathology subscale accounted for 19.9% of the variance (R2 = 0.199). The model for the Negative subscale accounted for 15.2% of variance in the change of the respective subscale (R2 = 0.152). Predictor patterns differed by symptom domain. The only significant predictor in the Positive subscale model was the baseline positive score (B = −0.451, p < 0.001). In the Negative subscale model the baseline score was not a significant predictor (p = 0.542) but clozapine dose (B = −0.016, p = 0.028) and the standardized adjunctive AP dose (B = 0.001, p = 0.031) were significant predictors of symptom change. General Psychopathology subscale baseline score was marginally significant (B = −0.042, p = 0.092). The treatment group (amisulpride vs other antipsychotics) remained non-significant in predicting symptom change in the positive (p = 0.962), negative (p = 0.808) and general psychopathology (p = 0.248) domains. Discussion This study had an aim to compare the outcomes of children and adolescents with SSDs, augmented with amisulpride or other APs in combination with clozapine. The results of the study do not confirm a statistically significant difference in clinical outcomes in patients whose treatment was augmented with amisulpride in comparison with those augmented with other antipsychotics. The groups demonstrated comparable age ranges and educational levels together with equivalent clinical statuses yet the amisulpride group included a larger percentage of male participants. While our data set cannot provide firm conclusions as to the exact reasons for this discrepancy, one possible explanation in naturalistic clinical settings relates to expected side-effect profiles. The well-known propensity of amisulpride to induce hyperprolactinemia and related side effects (eg, amenorrhea and galactorrhea) [24–26] may make clinicians reluctant to prescribe it to young women [27]. Cannabis and other substance use rates were also higher in the amisulpride augmentation group, although this did not reach statistical significance. Diagnosis and rate of first-episode psychosis/catatonia were similar between the groups. The mean clozapine dosages remained consistent across groups which eliminates drug titration as a cause for varied treatment effects. Recent literature points to particular patterns and persistent challenges in the management of early onset schizophrenia spectrum (EOSS) disorders. Clozapine remains the most effective intervention for treatment resistant EOSS with higher efficacy than agents such as olanzapine or haloperidol with a few but highly evidenced pediatric randomized controlled trials [28]. In addition, recent long-term data indicate that clozapine is well tolerated in youth, with a striking 38% lower relapse rate over 10-year follow-up compared to other antipsychotics, without significant differences in adverse events [29]. It is effective but regulatory frameworks differ. Clozapine is approved by the European Medicines Agency (EMA) for adolescents 16 and over, but is off-label for under 18s under the US FDA, leading to common off-label use [30]. Lower doses to achieve lower blood concentrations are often effective in clinical practice for paediatric patients. However the best approach when clozapine monotherapy is insufficient (switching, dose increase, within approved limits, augmentation) is unclear. If augmentation is tried, current literature generally recommends using a second-generation antipsychotic, specifically aripiprazole, as an augmentation strategy [28, 30]. However, our real-world data show a highly heterogeneous prescribing pattern with frequent use of drugs such as quetiapine (47.1%) and risperidone (43.1%) in addition to aripiprazole (23.5%). This discrepancy underscores the clinical challenges in treating severe clozapine-resistant early-onset schizophrenia spectrum disorders. Aripiprazole augmentation is still the most evidence-based recommendation, but the common off-label use of other agents in our cohort reflects the pragmatic and symptom-tailored choices clinicians are forced to make. For example, off-label use of broad-receptor agents may be driven by the urgent need to treat specific residual clusters such as severe agitation, treatment-emergent insomnia, or acute behavioral dysregulation. Symptom-tailored off-label strategies are pragmatic approaches, frequently necessary in highly refractory pediatric naturalistic settings where strict evidence-based guidelines are often limited [30]. No significant differences in the reduction of the Positive and Negative Syndrome Scale (PANSS) total scores from baseline to discharge were found in the Clozapine + Amisulpride group versus the Clozapine + Other AP group. Clozapine combined with amisulpride or with another AP demonstrated a similar degree of treatment response, including the total score and all three subscales of the PANSS (positive, negative, and general psychopathology symptoms). A similar degree of baseline symptom severity was observed in both treatment groups. In addition, no significant differences in clinical response rates, including patients with ≥20% and ≥30% improvement in the total PANSS score, were found. A numerical trend favouring the other AP group was observed at the 30% improvement threshold for the total score; however, this difference did not reach statistical significance. Similar results were found for all subscales of the PANSS, with no statistically significant differences in any of the response rate comparisons. Several randomised controlled trials (RCTs) and systematic reviews have compared the use of amisulpride for augmentation in patients with schizophrenia who are partially or completely unresponsive to clozapine. The results of most high-quality RCTs in adults have shown only a modest or statistically nonsignificant improvement in clinical outcomes with an increased risk of side effects [6, 7, 9, 11]. Amisulpride is an SGA with well-documented efficacy in the treatment of both positive and negative symptoms of schizophrenia [31]. Unlike most other atypical antipsychotics, amisulpride has been studied in patients with predominant negative symptoms in randomised controlled trials. Placebo-controlled data consistently show that amisulpride is effective in the treatment of negative symptoms in the licensed dose range of 50–300 mg/day [32–34]. As with all antipsychotic treatment, however, response is usually heterogeneous. In view of the aforementioned data, the absence of any substantial improvement in negative symptoms in our study constitutes a critical finding that necessitates further discussion. The mean amisulpride dose administered in the present study was 565.2 ± 210.8 mg/day, which is clearly in excess of the approved range for negative symptoms and falls within the high-dose range intended for treating positive symptoms. It can be hypothesised that the significant dopamine blockade triggered by the combination of high-dose amisulpride and clozapine may manifest as secondary negative symptoms, thereby obscuring any potential therapeutic benefits for primary negative symptoms. This provides a compelling pharmacological explanation for the absence of observed outcomes in our findings [11]. The debates that continue around therapeutic options for clozapine resistant schizophrenia have also led to an investigation of the limits of the traditional dopaminergic therapy. The logic of this strategy has been based on adding a drug with high D2 receptor antagonist properties, for example amisulpride, to a pharmaceutical antipsychotic to potentiate the drug’s D2 blockade. Clozapine is seen as weak in its D2 receptor blockade [6, 7]. However, it has been shown that the neurobiology of clozapine resistant schizophrenia is not as simple as a dopaminergic failure and that non-dopaminergic systems such as the glutamate pathway are much more responsible for the resistance [9]. The study’s findings indicate that augmenting amisulpride does not yield superior outcomes in comparison with other active antipsychotics, which is in alignment with the principles of this alternative perspective. This viewpoint is supported by substantial evidence from recent large-scale studies. For instance, AMICUS established that there was no statistically significant benefit to augmenting with amisulpride in comparison with a placebo [7]. Conversely, a thorough network meta-analysis revealed that no antipsychotic augmentation, including amisulpride, exhibited superior efficacy in comparison to a placebo in the management of CRS [9]. Our results, in line with previous studies, indicate that treatment strategies for this challenging population, which is resistant to clozapine, must consider targets other than those related to dopamine. A randomised, double-blind, placebo-controlled trial conducted by Zhu et al. [11] discovered that augmenting clozapine-resistant schizophrenia patients with amisulpride is significantly more efficacious than augmenting them with a placebo. In this post, it is clear that the patients in the amisulpride group experienced significantly greater reductions in PANSS total, positive, and general psychopathology subscale scores at both weeks 6 and 12 in comparison to the placebo group [11]. Our finding that augmentation with amisulpride is no more efficacious than augmentation with other active antipsychotics is in contrast to the results of Zhu et al. [11] who showed it to be effective in comparison to a placebo in adult patients. In this commentary, it is clear that this apparent discrepancy can be explained by several key differences in the design of the studies (active comparator vs placebo-controlled) and the populations studied (children and adolescents vs adults). It is more difficult to show superiority to an active treatment than it is to a placebo. The investigation conducted by our team alongside Zhu et al.‘s research provided unequivocal evidence that treatment failed to improve negative symptoms. In addition, Zhu et al. demonstrated a potential positive which was not evaluated in our study, an improvement in cognition (especially language). This clearly delineates a promising area for future paediatric research. We then investigated predictors of treatment response using a multiple linear regression analysis. This indicates that the type of augmentation agent (amisulpride vs. other antipsychotics) was not a significant predictor of change in PANSS total score or any of its subscales. Analysis revealed that baseline PANSS scores and clozapine dose were significant predictors of overall treatment outcome. This underscores the significance of severity of illness at treatment initiation and degree of primary antipsychotic in this patient population over choice of augmentation agent. In exploring individual symptom domains, baseline severity emerged as a robust predictor of change in the positive subscale. But a distinctive and clinically important pattern emerged for negative symptoms. Baseline severity was not a significant predictor in the negative subscale model, with higher clozapine doses predicting greater reduction in negative symptoms and higher improvement. Clozapine has been well established as effective in reducing negative symptoms in children and adolescents. Comparative studies and reviews show it to be superior to other antipsychotics, especially in treatment-resistant early-onset schizophrenia [28, 35, 36]. Recent pediatric retrospective data have shown significant improvements in SANS (The Scale for the Assessment of Negative Symptoms) scores in schizophrenia spectrum groups [37], and significant improvements across PANSS subscales in inpatient cohorts [38]. However, the evidence available to date defining the dose-response relationship for negative symptoms is weak. Pediatric studies typically report overall dosages; however, most do not conduct dose-response analyses or stratify negative symptom outcomes by dose levels [37, 39, 40]. Thus, our finding that higher clozapine dose is a significant predictor of negative symptom reduction offers valuable and novel findings to address this critical gap in the pediatric literature. Conversely, the contrary finding that greater standardized doses of adjunctive antipsychotics were associated with worse negative symptom improvement is very consistent with the clinical phenomenon of secondary negative symptoms from excessive D2 receptor blockade. Direct pediatric randomized data correlating high D2 blockade to negative symptom worsening are relatively limited, but mechanistic adult data and pediatric tolerability profiles strongly support this hypothesis. Neuroimaging and clinical trial data suggest that high D2 receptor occupancy, particularly at or above the extrapyramidal symptom (EPS) threshold (~77–78% in young adults, and sharply increasing above 75–85%), greatly diminishes tolerability and limits therapeutic benefit [41, 42]. In fact, high dose D2 antagonists have been associated with more severe negative symptoms than partial agonists, even in remitted patients with first episode psychosis [43]. It is noteworthy that children and adolescents have a higher risk of antipsychotic-induced EPS compared with adults [44–47]. Meta-analytic evidence in pediatric populations shows a 2.24-fold increase in EPS risk with atypical antipsychotics, with a Number Needed to Harm (NNH) of 7, with high-affinity D2 blockers (e.g., risperidone or ziprasidone) bearing the greatest burden [48–50]. However, antipsychotics with rapid D2 dissociation or lower D2 affinity, such as clozapine and quetiapine, carry a much lower burden of EPS [44, 51, 52]. Mesocortical and other cerebral circuits may also be strongly D2 antagonized, leading to a state of “blunted affect and behavior” characterized by apathy and avolition [52] in addition to nigrostriatal pathways. This is further worsened by sedation and somnolence, which represent highly frequent side effects in pediatric cohorts, particularly with medications such as olanzapine, ziprasidone, or high-dose adjunctive therapies [44, 47, 53–55]. Sedation and EPS are strong independent detractors of functional quality of life in adolescent schizophrenia [56]. Consequently, high doses of adjunctive D2-blocking agents may produce overlapping extrapyramidal stiffness, cognitive blunting and sedation, which clinically resemble, or actively exacerbate primary negative features. The increased vulnerability of early onset populations to these adverse effects and secondary negative symptoms is fundamentally due to the ongoing maturation of the adolescent brain. In this vulnerable period of development, dopamine and other neurotransmitter receptor systems are still actively developing and critical communication pathways between the cortico-striatal systems have not yet been fully established [57]. Adolescents are biologically more sensitive to dopamine receptor blockade, so the addition of a potent D2 antagonist to clozapine can be particularly problematic. This profound neurodevelopmental immaturity affects both the clinical course and vulnerability to adverse events, making youth much more sensitive to the consequences of aggressive D2 blockade than adults [51]. The following limitations of the current study are noted. First, as this is a retrospective analysis of clinical practice, it was not conducted under the rigorous conditions of an experimental study.The PANSS ratings were not part of an inter-rater reliability exercise and possible rater-related differences in administration are thus not excluded. Second, the length of time of treatment with clozapine and of exposure to adjunctive antipsychotics was not matched between individuals. Third, while the DDD method was used to strictly standardize adjunctive antipsychotic doses and control for varying potencies, this approach cannot fully account for the different pharmacodynamic receptor-binding profiles of the different agents pooled in the ‘Other AP’ comparator group. Consequently, the efficacy of each individual adjunctive antipsychotic against amisulpride could not be statistically analyzed separately. Fourth, despite the relatively good sample size (n = 76) for a rare population like that of clozapine-treated adolescents, we may have had an inadequate statistical power to detect moderate but clinically significant differences between the two active strategies. Fifth, some clinically relevant baseline imbalances between groups may have been acting as potential confounders. In particular, the Clozapine+Amisulpride group included a significantly higher proportion of male patients (90.9% vs. 64.8%) and a trend towards a higher frequency of cannabis users (27.3% vs. 7.4%) compared to the comparison group. Male gender and comorbid substance use are both known to be associated with more severe and treatment-resistant illness in the early-onset schizophrenia. Although our regression models took into account the potential effect of baseline severity, one cannot completely rule out the potential impact of these demographic and clinical imbalances on treatment outcomes. Another limitation of our retrospective design is that we cannot determine the exact order of prior clozapine augmentation trials. In pediatric populations long lines of sequential augmentations are relatively rare since the duration of illness is shorter. The lack of this specific data limits a more detailed analysis of the patients’ previous resistance profiles. In addition, the retrospective chart-review design did not allow us to access the precise clinical rationale for drug choices, nor could we systematically extract data on specific treatment-emergent adverse effects (eg, hyperprolactinemia or metabolic syndrome). Hence, the expected or experienced side-effect profiles could not be assessed quantitatively as a factor influencing the choice of treatment. Conclusions To conclude, this observational, retrospective cohort study conducted in the child and adolescent psychiatric inpatient setting could not find support for the clinical superiority of clozapine augmentation with amisulpride over augmentation with other antipsychotics. Baseline illness severity and the target clozapine dose seemed to be more important determinants of response. This study’s results underline the need for adequately powered randomised controlled trials to better guide the augmentation of clozapine in this vulnerable adolescent population. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary material 1 Supplementary material 1