Introduction: Cell-derived extracellular vesicles ([EVs], i.e., exosomes and microparticles) have been reported to mediate the cardioprotective effects of stem cells. However, a head-to-head comparison of their effects with those of the cells from which they derive has not been reported. Hypothesis: The effects of the EV content of human embryonic stem cell (hESC)-derived cardiac progenitors may be equivalent to those of the parent cells. Methods: Two series of immunodeficient (nude) mice underwent permanent coronary artery ligation. Three weeks later, those with an echocardiographically-determined LV ejection fraction ≤ 50% were randomly allocated to receive transcutaneous echo-guided peri-infarct injections of alpha-MEM media (controls), hESC-derived SSEA-1+ cardiac progenitors (500000 cells in 30 μL), or total EV secreted by those 500000 cells in 48 hours in an equivalent volume. The second series also included sham-operated animals for which the needle was inserted into the myocardium in three places without injection. EVs were collected from the cell-secreted medium by ultracentrifugation and characterized by flow cytometry, and nanosight NTS. Outcomes were assessed after 6 weeks by echocardiography, taking LV end-systolic volume (ESV) as the primary end point. Hearts were processed for the assessment of fibrosis and angiogenesis. Total RNA was extracted from carefully selected animals for gene expression analysis by Affymetrix chip. Only animals with similar VTS at baseline were used for this analysis. All data were collected and analyzed blindly. Results: Cardiac progenitors were successfully generated by exposure of the pluripotent ESC to bone morphogenetic protein-2, purified by anti-CD15 immunomagnetic sorting and then cultured on vitronectin for 48 hours, after which cells or EVs derived from the same cell batch were injected. After 6 weeks in the first series, LVESV [m±SEM] in controls did not significantly differ from the pre-injection value (difference: -2.64 ± 1.54 µL, p=0.12, n= 12). Conversely, in the cell-treated group, LVESV significantly decreased by -4.20 ± 0.96 µL (p=0.0007, n=16). A similar decrease was seen in the EV-treated group: -5.73 ± 1.21 µL (p=0.0003, n=15). Similar patterns were seen for LV end-diastolic volumes: controls (-2.46 ± 1.19 µL, p=NS), cells (-4.48 ± 1.47 µL, p=0.009), EV (-4.29 ± 1.31 µL, p=0.005). For the second series, the VTD of cell- and EV-treated animals remained stable (cell: -1.85 ± 3.27 µL, p=NS, n= 4; EV: +0.05 ± 1.53 µL, p=NS, n = 4), whereas sham- and control-treated animals deteriorated significantly (sham: +2.20 ± 0.35 µL, p = 0.003, n = 5; control: +3.27 ± 0.87 µL, p = 0.03, n = 4). The trend held for VTS, though differences were not significant. Analysis of gene expression data revealed interesting pathways that were more highly expressed in cell- and EV-treated animals than in controls and sham-operated animals. Conclusions: These data support the hypothesis that EVs may be critical mediators of the paracrine effects of cell therapy, and for the sake of streamlining translational processes, deserve to be considered as potential alternatives to stem cell transplantation. Gene expression levels were compared between mouse hearts.