Background and purpose The debate over the fact that experimental drugs proposed for the treatment of stroke fail in the translation to the clinical situation, has attracted considerable attention in the literature. In this context, we present a retrospective pooled analysis of a large dataset from pre-clinical studies, in order to examine the effects of early versus late administration of intravenous recombinant tissue type plasminogen activator (rt-PA). Methods We collected data from 26 individual studies from 9 international centers (13 researchers, 716 animals) that compared rt-PA to controls, in a unique mouse model of thromboembolic stroke induced by an in situ injection of thrombin into the middle cerebral artery. Studies were classified into early (<3h) versus late (≥3h) drug administration. Final infarct volumes, assessed by histology or MRI, were compared in each study and the absolute differences were pooled in a random-effect meta-analysis. The influence of time of administration was tested. Results When compared to saline controls, early rt-PA administration was associated with a significant benefit (absolute difference = −6.63 mm3; 95%CI, −9.08 to −4.17; I2=76%) whereas late rt-PA treatment showed a deleterious effect (+5.06 mm3; 95%CI, +2.78 to +7.34; I2=42%, Pint<0.00001). Results remained unchanged following subgroup analyses. Conclusion Our results provide the basis needed for the design of future pre-clinical studies on recanalization therapies using this model of thromboembolic stroke in mice. The power analysis reveals that a multi-center trial would require 123 animals per group instead of 40 for a single center trial.