The prognostic implications of myocardial computed tomography perfusion (CTP) analyses are unknown. In this sub-study to the CATCH-trial we evaluate the ability of adenosine stress CTP findings to predict mid-term major adverse cardiac events (MACE). In 240 patients with acute-onset chest pain, yet normal electrocardiograms and troponins, a clinically blinded adenosine stress CTP scan was performed in addition to conventional diagnostic evaluation. A reversible perfusion defect (PD) was found in 38 patients (16 %) and during a median follow-up of 19 months (range 12-22 months) 25 patients (10 %) suffered a MACE (cardiac death, non-fatal myocardial infarction and revascularizations). Accuracy for the prediction of MACE expressed as the area under curve (AUC) on receiver-operating characteristic curves was 0.88 (0.83-0.92) for visual assessment of a PD and 0.80 (0.73-0.85) for stress TPR (transmural perfusion ratio). After adjustment for the pretest probability of obstructive coronary artery disease, both detection of a PD and stress TPR were significantly associated with MACE with an adjusted hazard ratio of 39 (95 % confidence interval 11-134), p < 0.0001, for visual interpretation and 0.99 (0.98-0.99) for stress TPR, p < 0.0001. Patients with a PD volume covering >10 % of the LV myocardium had a worse prognosis compared to patients with a PD covering <10 % of the LV myocardium, p = 0.0002. The optimal cut-off value of the myocardial PD extent to predict MACE was 5.3 % of the left ventricle [sensitivity 84 % (64-96), specificity 95 % (91-97)]. Myocardial CT perfusion parameters predict mid-term clinical outcome in patients with recent acute-onset chest pain.