The difficult task of assessing perimortem and postmortem fractures on the skeleton : a blind test on 210 fractures of known origin

this study aims at highlighting the difficulties and pitfalls a forensic anthropologist has to face when diagnosing bone fractures as perimortem and postmortem. Impact statement: any morphological and macroscopic evaluation on bone fractures has to be handled with care since the evaluation of a bone fracture as being peri or postmortem may be difficult, treacherous and at times observer-dependent and thus influenced by knowledge, intuition and training of the observer. If postmortem fractures can be detected more easily, on the other hand perimortem fractures can be wrongly identified, especially when spongy bone is involved and when time has taken its toll on the remains. In the field of forensic anthropology, the difference between perimortem and postmortem fractures is one of the most difficult challenges. Indicators of perimortality have been suggested (for example “green” aspect or colouring of the fractured margins) but most forensic anthropologist know how many times it may be impossible to reach a decision. Few studies have focused on the actual error anthropologists can run into when classifying bone fractures as post or perimortem. How many times will a postmortem or taphonomical fracture be mistaken for a perimortem fracture and viceversa? Which bones are the trickiest? The present study aims at verifying the error behind such a diagnosis by blind testing two experienced anthropologists (both trained as anthropologists, observer A with a 7 year working experience in the field and observer B with a 3 year working experience in the field) on 210 fractures of known origin. Four skeletons were selected from a skeletal series of 250 individuals who died in 1991 and whose skeletons were exhumed in 2001, unclaimed and thus available for scientific research according to Italian Mortuary Police Regulations. Of these four skeletons three had died in traffic accidents (case 1: pedestrian run over by a tram; case 2: pedestrian run over by a car; case 3: pedestrian run over by a truck, thus all presenting blunt force trauma) and one of natural causes. Autopsy reports were available for all with detailed descriptions of soft tissue and bone lesions. It was evident upon an initial examination of the skeletons (before the test) that several postmortem fractures were due to taphonomical events related to burial and exhumation; furthermore actual perimortem lesions (registered upon autopsy) had been evidently ruined by soil and other taphonomical variables. For all cases the number and site of bone fractures detected at autopsy in 1991 were recorded as well as those known to have been certainly caused by postmortem events (because not present at autopsy). The total number of fractures was 210. Then the two observers (forensic anthropologists) were asked to blindly score all lesions on the four skeletons as perimortem, postmortem or uncertain. Results were then evaluated by comparing the scores to the real perimortem or postmortem nature of the fracture. The results of the osteological analyses show the highest success rate for both observers in the correct identification of postmortem fractures, with percentages for correct identification between 75% and 100%. On the other hand, with perimortem fractures, the correct classification falls to a mean of 45%, with very minimum differences between observers (43.5% and 46%). In other words only 45% of all perimortem fractures were identified as such by the two observers. In 16.5% of perimortem fractures and in 7% of postmortem fractures the origin was defined as non assessable (uncertain) by the observers, and thus considered “dubious”. When observing the bones involved, the highest number of mistakes were noticed when the observer had to evaluate spongy bone (mostly ribs or bones of the pelvis). Much fewer errors were performed when evaluating fractures on long bones and skull. Globally, the present study clearly shows the difficulty and dangers of this crucial task. The correct identification of peri- and postmortem lesions is fundamental, but macroscopic and morphological criteria are still unsatisfactory, and sometimes misleading. The present results therefore should serve as a cautionary note concerning interpretation of peri and postmortem fractures as well as an invitation to search for novel methods of analysis (such as histology, immunochemistry, electronic microscopy), in order to find new tools towards a more comprehensive solution of this problem.