Early Antipsychotic Nonresponse as a Predictor of Nonresponse and Nonremission in Adolescents With Psychosis Treated With Aripiprazole or Quetiapine: Results From the TEA Trial.

OBJECTIVE: To evaluate 1) whether early nonresponse to antipsychotics predicts nonresponse and nonremission, 2) patient and illness characteristics as outcome predictors, and 3) response prediction of 30-item Positive and Negative Syndrome Scale (PANSS-30) compared with 6-item PANSS (PANSS-6) and Clinical Global Impressions-Improvement Scale (CGI-I) in youths with first-episode psychosis. METHOD: Post hoc analysis from a 12-week, double-blinded, randomized trial of aripiprazole vs extended-release quetiapine in adolescents (age 12-17 years) with first-episode psychosis was performed. Early nonresponse (week 2 or week 4) was defined as <20% symptom reduction (PANSS-30) (or <20% symptom reduction [PANSS-6] or CGI-I score 4-7 [less than "minimally improved"]). Nonresponse (week 12) was defined as <50% symptom reduction (PANSS-30). Nonremission (week 12) was defined as a score of >3 on 8 selected PANSS-items. Positive/negative predictive values (PPV/NPV) and receiver operating characteristics, binary logistic regression models, and PPV/NPV using PANSS-6 and CGI-I were analyzed. RESULTS: Of 113 randomized patients, 84 were included in post hoc analysis (mean [SD] age = 15.7 [1.3] years; 28.6% male). The 12-week symptom decrease was 31.9% [27.9%], most pronounced within the first 2 weeks (61.1% of total PANSS reduction). Response (27.4%) and remission (22.6%) rates were low. Results indicated that early nonresponse reliably predicted 12-week nonresponse (PPV: week 2, 82.2%; week 4, 90.0%) and nonremission (PPV: week 2, 80.0%; week 4, 90.0%); early nonresponse at week 4 was a statistically significant baseline predictor for 12-week nonresponse; and PANSS-6 had similar predictive significance as PANSS-30. However, outcomes were heterogeneous using CGI-I. CONCLUSION: In youths with first-episode psychosis showing early nonresponse to aripiprazole or extended-release quetiapine, switching antipsychotic drug should be considered. PANSS-6 is a feasible and clinically relevant alternative to PANSS-30 to predict 12-week nonresponse/nonremission. CLINICAL TRIAL REGISTRATION INFORMATION: Tolerance and Effect of Antipsychotics in Children and Adolescents With Psychosis; https://www. CLINICALTRIALS: gov/; NCT01119014.

Pretreatment Cardiometabolic Status in Youth With Early-Onset Psychosis: Baseline Results From the TEA Trial.

OBJECTIVE: To describe pretreatment cardiometabolic constitution in children and adolescents with first-episode psychosis (FEP). METHODS: Baseline cardiometabolic assessment was performed in youths aged 12-17 years with FEP entering the Tolerability and Efficacy of Antipsychotics (TEA) trial and matched healthy controls. Patients were included between June 10, 2010, and January 29, 2014. ICD-10 was used as the diagnostic classification system. Cardiometabolic risk markers were compared between patients versus controls and antipsychotic-naive versus antipsychotic-exposed patients. RESULTS: Comparing 113 youths with FEP (age ± SD = 15.74 ± 1.36 years, males = 30.1%, schizophrenia-spectrum disorders = 92.9%, antipsychotic-naive: n = 57) to 60 controls, patients had higher waist circumference (WC) z scores (1.13 ± 1.65 vs 0.42 ± 1.27, P = .018), cholesterol (4.10 ± 0.71 vs 3.79 ± 0.49 mmol/L, P = .014), low-density lipoproteins (2.37 ± 0.56 vs 2.13 ± 0.51, P = .012), and non-high-density lipoproteins (2.58 ± 1.60 vs 2.52 ± 0.52, P = .018). More patients than controls (42.9% vs 20.3%, P = .019) and antipsychotic-naive than antipsychotic-exposed (51.9% vs 34.0%, P = .023) had a WC > 90th percentile. Hypercholesterolemia (34.0% vs 12.5%, P = .015) was more frequent in patients, while decreased high-density lipoprotein cholesterol was more frequent in controls (32.5% vs 19.0%, P = .032). Family history of type 2 diabetes mellitus was associated with increased body mass index (BMI) z score (P < .001), WC z score (P = .001), insulin (P = .038), and homeostatic model assessment of insulin resistance (HOMA-IR; P = .025). Dyslipidemia was associated with significantly increased insulin (P = .041), HOMA-IR (P = .032), and low-density lipoprotein cholesterol (P = .041). Previous antipsychotic exposure was not associated with increased cardiometabolic risk. Early age at onset predicted increased BMI and WC z scores, while diagnosis of schizophrenia and higher Clinical Global Impression-Severity score were associated with increased blood lipids. CONCLUSIONS: Youths with FEP had significantly greater WC and lipid abnormalities than matched controls, regardless of antipsychotic exposure. In youths with FEP, elevated metabolic risk predates antipsychotic exposure. TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT01119014; European Clinical Trials Database (EudraCT): 2009-016715-38​​​.