RESUMO
BACKGROUND: Superiority of long acting injectable antipsychotics (LAI) over oral antipsychotics remains controversial and dependent on study design and inclusion criteria. Meta-analysis of 21 RCTs demonstrated no difference in their effectiveness, but meta-analysis of 25 mirror-image studies did. None of these included paliperidone palmitate (PP). METHODS: We challenged efficiency of PP in a multicentric mirror-image study. Primary outcome was total hospitalization days. Mirror periods were 365days either side of the first injection in model-1, and either side of index admission in model-2. Inclusion criteria were: 18 to 65years, schizophrenia spectrum disorder, ≥3 injections received, and oral antipsychotic prescriptions before PP trial. Exclusion criteria were: prior clozapine or LAI trial. Cost-effectiveness was calculated from a public payer's perspective. RESULTS: 114 patients were recruited (77% males, mean 37years, mean disease duration 10years). Oral antipsychotics adherence was 43%. Mean PP treatment lasted 297days (adherence 81%). Mean annual hospitalization days weren't significantly different in model-1 (45.8days vs 38.5days, p=0.058), but were significantly lower in model-2, (14.4days vs 24.2days, p=0.003). 1.9 admissions per patient-year fell to 0.64 on PP (p<0.0001). PP was approximately cost-neutral: differences were -$326 and $1788 for model-1 and model-2. DISCUSSION: PP as a first LAI improved adherence, decreased hospital visits and duration was cost neutral. Drawbacks are the retrospective design and lack of comparator and safety data. Strengths are naturalistic design and adherence calculation. A subset of patients responds well to LAI, leading to meaningful reductions in hospital services requirements.