Large effect of baseline treatment with long acting antipsychotic drugs on randomized treatment outcomes [Abstract]

Background: In the CUtLASS 1 Trial (Jones et al., 2006) patients with an inadequate clinical response or intolerance were randomised to either a first generation antipsychotic (FGA) drug or a (non-clozapine) second generation antipsychotic (SGA) with assessments at baseline, 12, 26 and 52 weeks following randomisation. The primary outcome was quality of life (QOL) measured using the QLS, with secondary outcome measures including symptoms (PANSS), depression (CDSS), overall functioning (GAF), drug attitude (DAI) and adherence (Kemp). Non-neurological side effects (ANNSERS) and neurological side effects (Simpson-Angus, AIMS, Barnes) were also assessed. Would outcome during the course of the trial be affected by the delivery route of the antipsychotic drug prescribed at trial entry? Methods: Forty per cent (N=90) of the 227 patients entering the CUtLASS 1 Trial were being treated with a depot FGA antipsychotic prior to randomisation. Results: Fitting multi-level mixed-effects models using Stata 11 and including demographic variables and baseline attitudes to medication as predictors showed that: QLS was significantly reduced (-5.7 points; CI -10.1, -1.4) at final visit in those receiving depot before randomisation. There was no significant difference in this effect between those who were randomised to first or second generation antipsychotics during the trial. The same pattern of results held for PANSS total score and GAF. Modelling centre as a separate level had little effect on coefficients of baseline covariates. Baseline DAI score indicating adherent attitudes predicted better outcome on all three measures (p