RESUMO
OBJECTIVE: The present review evaluates systematically the published quantitative studies of correlates of PA in patients with schizophrenia. METHOD: EMBASE, PsycINFO, PubMed, ISI Web of Science, CINAHL and PEDro were searched from their inception to 1 July 2011 combining the medical subject heading 'schizophrenia' with 'physical activity' or 'physical inactivity' or 'exercise' or 'health education' or 'health behaviour' or 'health promotion'. RESULTS: Out of 68 potentially eligible studies, 25 papers (n = 25,013) evaluating 36 correlates were included. Correlates consistently associated with lower PA participation are the presence of negative symptoms and cardio-metabolic comorbidity. Also, side-effects of antipsychotic medication, lack of knowledge on cardiovascular disease risk factors, no belief in the health benefits, a lower self-efficacy, other unhealthy lifestyle habits and social isolation correlated with lower PA participation. The quality of the PA measurement was not related to the proportion of significant associations (χ(2) = 3.8, P = 0.07). Current gaps in literature that need to be examined more in detail are the role of environmental and policy-level factors on PA participation in patients with schizophrenia. CONCLUSION: All correlates should be confirmed in prospective studies, and interventions to improve the modifiable variables should be developed and evaluated.