PURPOSE OF THE STUDY To assess the effects of medical history, fracture type, method of treatment and complications on the risk of death in elderly patients treated for proximal femoral fracture. MATERIAL AND METHODS The group comprised of 269 patients (219 women and 50 men) older than 70 years who underwent surgery for proximal femoral fracture at the Level I Traumacentre between January 2003 and June 2005. The follow-up ranged from 12 to 38 months. In a prospective consecutive manner the following was recorded: age, gender, serious internal diseases, pre-injury level of mobility, place of living before injury, fracture type (AO classification), time between injury and surgery, anaesthesia, surgical technique, complications and death. Statistical significance at the 95% level was ascertained using null-hypothesis tests for qualitative and quantitative variables by means of multivariate analysis. RESULTS The average age of the patients was 81 years (range, 70 to 99 years). The higher the age, the shorter the time of survival (p=0.015) as each additional year reduced survival by 4.6%. The women to men ratio was 4.2:1. The male gender had a significantly shorter time of survival (p=0.007). Two and more serious internal diseases in the patient's medical history, as compared with a single one, also made survival significantly shorter (p<0.001). When, before injury, a patient moved without walking support, survival was significantly longer than in a patient using a walking cane/crutch (p=0.022) or two canes/crutches or a walker (p<0.001), or in a bedridden patient (p=0.014). The fact that, before injury, a patient was living in a pensioner's home had no effect on a shortened time of survival (p=0.136). Similarly, the fracture type (31A or 31B and subgroups) was not significant for the length of survival (p=0.903). The interval between injury and surgery was not a risk factor for survival (p=0.269). No effect of the type of anaesthesia on survival was found (p=0.450). Neither the surgical technique nor the type of implant was significant for survival time. When general internal complications occurred, they affected the length of survival significantly (p<0.001). Also, pressure sores developing in the post-operative period significantly shortened the time of survival (p=0.037). Early complications did not result in shorter survival (p=0.867), but late deep infection in ive patients significantly shortened their lives (p=0.008). Failed osteosynthesis and consequent revision surgery shortened survival time only when it occurred during the primary treatment (p=0.003); after the primary hospitalization was terminated, it had no effect on the length of survival (p=0.398). DISCUSSION The study focused attention only to elderly patients with proximal femoral fracture treated surgically. The authors suggest that, by excluding patients treated conservatively, the group became more homogeneous and a more exact assessment of each factor's effect on death risk was possible. The investigated factors were recorded only during the first year following injury; however, death was recorded by the end of the study. Using the assessment of cumulative survival, this allowed for a more accurate statistical evaluation of the effect of each factor on the risk of death. CONCLUSIONS In patients over 70 years, a significantly shorter time of survival following the surgical treatment of proximal femoral fracture was found to be related to high age, male gender, multiple morbidity in the patient's medical history, poor patient's mobility before injury, general complications, development of pressure sores post-operatively, failed osteosynthesis requiring revision surgery and deep infection of the affected hip. No relation to significantly shorter survival was found for the following factors: living in a pensioner's home before injury, fracture type, time between injury and surgery, type of anaesthesia and operative technique. Key words: hip fracture, surgical treatment, mortality, risk factors.