Background: Although it is known that elderly patients have high cardiac event rates, there are no separate guidelines for risk stratification, due to insufficient data.Presently, exercise tolerance testing alone remains the test of choice in this group.We evaluated the published data of exercise ECG, nuclear myocardial perfusion imaging (MPI) and stress echo in elderly with regards to prediction of cardiac events.Method: We searched PubMed and Cochrane databases from Jan '90 to Sep '08 and reviewed articles obtained for the search terms: 'elderly', 'old', 'aged', and 'risk stratification', 'myocardial perfusion imaging', 'stress Echo, 'stress electrocardiogram'.We included prospective cohort studies of subjects who underwent stress ECG, MPI or Echo, for known or suspected CAD, and provided data on cardiac outcomes (death or MI) with at least 1 year of followup.Studies with N , 100 patients, mean age ,65 or early revascularization were excluded.A weighted t test was used for cardiac events, and fixed effects models were used for summary odds ratios and relative risks.Results: A total of 7 papers on exercise ECG (N -4595, mean age 71.4), 7 papers on stress echo (N -7366, m.age 72) and 9 papers on stress MPI (N 2 2472, m. age 74.8) met the criteria.The summary odds ratio for stress MPI (RR 4.97, 95% CI 3.83-6.46)was superior [p , 0.05] to that of stress Echo (RR 3.00, 95% CI 2.44-3.68)and stress ECG (RR 1.84, 95% CI 1.63-2.09).The pooled annualized event rates were 1.2% for a negative MPI and 8.9% for positive MPI.Conclusion: Stress Myocardial Perfusion Imaging can provides excellent risk stratification in elderly patients and is superior to exercise testing alone, or echocardiography.These findings suggest stress MPI may be the preferred method of evaluation in the elderly.