The likely mortality benefit from pre-optimisation of high risk surgical patients depends on the baseline risk of death. The baseline risk estimated from the control groups of studies of pre-optimisation may not reflect the risk in other institutions. We determined the 28-day mortality of 'high risk' adult surgical patients in our institution by performing a retrospective cohort study of 554 patients undergoing major or ultra-major general or vascular surgery lasting more than two hours. Patients were identified as being high risk based on the presence of risk factors used in previous studies of pre-optimisation. The 28-day mortality (95% CI) was 0% for patients with no risk factors and 2.3 (0.8 to 3.7)% for patients with at least one risk factor. These data indicate that the risk of death amongst high risk surgical patients may vary considerably between institutions.