Abstract Parvalbumin interneurons are fast-spiking GABAergic neurons that provide inhibitory control of cortical and subcortical circuits and are thought to be a key locus of the pathophysiology underlying schizophrenia. In view of the contradictory results regarding the nature of parvalbumin post-mortem findings in schizophrenia, we conducted a quantitative meta-analysis of the data on parvalbumin cell density and parvalbumin mRNA levels in pre-frontal regions in the brains of patients with schizophrenia (n = 274) compared with healthy controls (n = 275). The results suggest that parvalbumin interneurons are reduced in density in the frontal cortex of patients with schizophrenia (Hedges’ g = − 0.27; p = 0.03) and there is a non-significant reduction in parvalbumin mRNA levels (g = − 0.44; p = 0.12). However, certain methodological issues need to be considered in interpreting such results and are discussed in more detail. A meta-regression was conducted for post-mortem interval and year of publication as covariates which were both non-significant, except in the mRNA meta-analysis where post-mortem interval was found to be significant. Overall our findings provide tentative support for the hypothesis that the GABAergic system is deficient in schizophrenia and that parvalbumin-containing interneurons offer a potential target for treatment. However, further well-controlled studies that examine multiple regions and layers are warranted to determine whether parvalbumin alterations are region or layer specific and to test the robustness of the findings further.