Objective.-To assess the feasibility of carrying out a second-opinion trial for patients urged to undergo coronary angiography and to assess the long-range out-come of such patients denied that procedure, and the criteria evolved for reaching such a conclusion. Design.-A case series of patients referred for a second opinion as to the need for coronary angiography. Patients were followed up by questionnaire, telephone call, and center visits. Setting.-Cardiovascular referral center and teaching hospital in Boston, Mass. Patients.-One hundred seventy-one patients with coronary artery disease (144 men, average age 60 years; range, 36 to 88 years). Three patients became unavailable for follow-up during a mean of 46.5 months. Outcome Measures.-Concordant-discordant outcome as to the second opinion, cardiac events, invasive interventions, quality of life questionnaire, and level of symptoms. Results.-One hundred thirty-four (80%) of the 168 patients were judged not to require angiography; it was recommended in six. In 28 (16%) recommendation was deferred pending further studies. At a mean follow-up of 46.5 months among the 168 patients, there were seven cardiac deaths (annualized cardiac mortality of 1.1%); 19 patients experienced a new myocardial infarction (2.7% annualized rate), while 27 patients (4.3%) were judged to have developed unstable angina. Twenty-six patients (1 5.4%) ultimately underwent either coronary bypass or angioplasty. Conclusions.-In a large fraction of medically stable patients with coronary disease who are urged to undergo coronary angiography, the procedure can be safely deferred. While there may be a limitation in terms of generalizing this experience to all patients with coronary disease, we reasonably conclude that an estimated 50% of coronary angiography currently being undertaken in the United States is unnecessary, or at least could be postponed.