Between 1978 and 1994, 130 deaths after hip fractures were investigated at the Institute of Forensic Medicine of the Free University of Berlin. In this period more than 3500 such cases occurred in Berlin. Eight persons were found dead at home, 17 had conservative treatment or died before the planned operation, and 105 had an operation. Hip replacement was carried out in 49, nailing in 17, and other kinds of osteosynthesis (especially using screws) were used in 39 cases. In the surgically treated group, death occurred during the operation in 6 cases (all hip replacements), on the day of operation or the first postoperative day in 7 cases, and within the first week in 21 cases. In the conservative group 10 cases died during the first week after trauma. The main, clearly detectable causes of death were pneumonia (30) and thromboembolism of the lungs (22). Fat embolism was responsible for death in at least 6 cases (3 intra- and 3 postoperative); in 2 additional intraoperative and 7 early postoperative deaths, a potentially lethal fat or bone marrow embolism occurred together with other deadly factors. A fat embolism was present in 9 cases within the first week, diminishing with survival time. Hip replacement was clearly dominant for the genesis of pulmonary fat embolism: in none of the early post-endoprosthetic deaths was there little or no fat embolism in contrast to the other types of operation. Bone marrow embolization can induce the formation of multiple small, red intravenous thrombi-possibly a factor explaining the lethal outcome in only a few cases with intraoperative fat embolism. This form of "thromboembolism" must be differentiated from typical lung thromboembolism, originating from a preceding venous thrombosis.