Treatment of acute schizophrenia with paliperidone ER: Predictors for treatment response and benzodiazepine use

- Baseline characteristics predict paliperidone ER dosing.
- Higher baseline disease severity predicts concomitant benzodiazepine use.
- Early clinical response predicts overall clinical response.
- Female gender, baseline CGI-S, or early high response predicts high response.

The Paliperidone ER Treatment in Acute Intervention (PERTAIN) study was designed to explore treatment response, tolerability, and safety of flexible doses of paliperidone ER in patients with schizophrenia admitted for an acute exacerbation. This paper addresses a secondary analysis of PERTAIN data designed to explore predictors for treatment response, flexible dosing, and concomitant benzodiazepine use. This prospective, multicenter, phase 3b, open-label, single-arm, 6-week study used flexible doses of paliperidone ER (3 to 12 mg once daily) to treat patients hospitalized for an acute exacerbation of schizophrenia, reflecting more closely daily clinical practice. Predictive models were evaluated for paliperidone ER flexible dosing, treatment response, and concomitant treatment with benzodiazepines as distinct independent variables. For the analysis of explanatory variables, a stepwise logistic regression was used, taking into account patient age, gender, body mass index, diagnosis and duration of schizophrenia, number of prior hospitalizations, psychotic symptoms (PANSS), disease severity (CGI-S), and patient functioning (PSP) at baseline. Early response (defined as response within 2 weeks of treatment initiation) was also used as a predictor. Clinical response (defined as ≥ 30% decrease in PANSS total score and ≥ 1 point decrease in CGI-S from baseline to endpoint) was predicted by early clinical response (p < 0.001) and there was a trend for the diagnosis of paranoid schizophrenia vs. other types of schizophrenia to predict clinical response (p = 0.0525). High response (defined as ≥ 50% decrease in PANSS total score and ≥ 2 points decrease in CGI-S from baseline to endpoint) was predicted by early high response, higher baseline CGI-S, or female gender. More severely ill patients with a higher baseline CGI-S were twice likely to be treated concomitantly with a benzodiazepine.