Cortical gray matter deficits have been found in patients with schizophrenia, with evidence of progression over time. The aim of this study was to determine the extent of progressive cortical gray matter volume changes over time in schizophrenia, their site and time of occurrence, and the role of potential moderators of brain changes. English language articles published between 1 January 1983 and 31 March 2012 in the MEDLINE and EMBASE databases were searched. Longitudinal magnetic resonance imaging studies comparing changes in cortical gray matter volume over time between patients with schizophrenia and healthy controls were included. Hedges g was calculated for each study. Analyses were performed using fixed- and random-effects models. A subgroup analysis was run to explore the pattern of brain changes in patients with first-episode schizophrenia. A meta-regression statistic was adopted to investigate the role of potential moderators of the effect sizes (ESs). A total of 19 studies, analyzing 813 patients with schizophrenia and 718 healthy controls, were included. Over time, patients with schizophrenia showed a significantly higher volume loss of total cortical gray matter, left superior temporal gyrus (STG), left anterior STG, left Heschl gyrus, left planum temporale and posterior STG bilaterally. Meta-analysis of first-episode schizophrenic patients showed a more significant pattern of progressive loss of whole cerebral gray matter volume involving the frontal, temporal and parietal lobes, and left Heschl gyrus compared with healthy controls. Clinical, pharmacologic and neuroradiological variables were found to be significant moderators of brain volume changes in patients with schizophrenia. The meta-analysis demonstrates that progressive cortical gray matter changes in schizophrenia occur with regional and temporal specificity. The underlying pathological process appears to be especially active in the first stages of the disease, affects the left hemisphere and the superior temporal structures more and is at least partly moderated by the type of pharmacological treatment received.