Deaths and delays in queues for coronary surgery in Canada have been highlighted by American interest groups opposed to "socialized medicine". Since 1991 all nine cardiac surgery centres in Ontario register and follow patients after acceptance for surgery. We examined the experience of 8517 consecutive patients leaving the registry from October 1991 to July 1993. Individual acuity scores were determined based on symptoms, angiographic findings, left ventricular function, and, where available, non-invasive tests of ischaemic jeopardy. Planned surgery was declined or deferred for 3.2% of registrants. While in the queue, 31 (0.4%) patients died and three had surgery indefinitely deferred after a non-fatal myocardial infarction. Among 8213 patients receiving surgery, the median wait was 17 days (inter-quartile range [IQR]: 4, 51), ranging from one day (IQR 0:4) for patients needing very urgent surgery (acuity score 2-3) to 42 days (IQR: 18, 77) for those rated low priority (acuity score 6-7). In a multivariate analysis, the most important determinant of waiting time was symptom status (p < 0.001), followed by anatomy (p < 0.001). Age did not alter waiting time; depending on statistical methods, female sex was either not significant or independently associated with approximately 11% relative delay (p = 0.001). Whether controlling for significant clinical factors or the multifactorial acuity scores, waiting times clearly varied (p < 0.001) among hospitals. We conclude that, during 1991-93, patients queuing for coronary surgery in Ontario rarely suffered critical events or extreme delays, and individual variation in waiting times primarily reflected clinical acuity. Nonetheless, symptoms provoked by very modest exertion were commonplace in the queue, and waiting times did vary inequitably among hospitals.