Ninety-seven patients aged 88 +/- 4 years (range, 80-97 years) (study group), and 74 aged 75 +/- 3 years (range, 70-79 years) (control group), were prospectively studied to investigate whether basic medical variables can predict in-hospital mortality in very old patients undergoing hip surgery because of femoral fracture. Mortality was 16.5% and 6.7% in the study and control groups, respectively (p = 0.054). In the study group, mortality was significantly correlated with age (p < 0.01), venous disorders (p < 0.05), malnutrition (p < 0.0001), duration of surgery (p < 0.006), and postoperative noninfectious complications (p < 0.005). In the control group, age was the only significant correlate of mortality (p < 0.005). After exclusion of surgery-related variables, the logistic regression analysis confirmed the predictive role of venous disorders (odds ratio = 2.04, confidence limits = 1.09-3.79) and malnutrition (odds ratio = 6.01, confidence limits = 1.85-19.47) but not of age in the study group. However, the goodness-of-fit test showed that the statistical model did not fit the data adequately. We conclude that in-hospital mortality after hip surgery in the very old cannot be predicted on the basis of underlying medical conditions alone.