Our objective was to determine the outcome of onsite CPR in elderly patients receiving institutional care who had access to 24-hour on-site full ACLS capability. We used a retrospective chart and CPR protocol review of all patients who received CPR during the study period; these were patients or residents in a multi-level geriatric long-term care facility that provides various degrees of facility-based and ambulatory care. Immediate, short-term, and long-term survivorship were determined and correlated where possible with the category of patient seen (short-term care versus long-term care) and the relationship between the witnessing of the arrest and survival. Of the 41 patients who underwent CPR, there were only four survivors of 60 days or greater. Of these, three subjects were short-term, not long-term care patients of the Centre; they returned to their previous level of function. One fully dependent bed-ridden patient returned to that level of care for a survival period of 100 days. Of the four long-term survivors, three of the arrests were witnessed, and one was indeterminate. No unwitnessed arrests resulted in long-term survival. Our experience suggests that CPR in the elderly long-term care patient is unlikely to be successful even when it is available on-site. Unwitnessed arrests in this population were universally fatal. This information has helped Baycrest Centre for Geriatric Care develop protocols, guidelines, and policies for CPR that are suitable for our population.