Résumé -Administrés par voie intraveineuse, les bisphosphonates trouvent de nombreuses indications en oncologie et en hématologie (prise en charge des myélomes multiples, prévention des complications osseuses de certaines tumeurs malignes avancées, traitement des hypercalcémies malignes). Depuis 2003, de nombreuses publications ont rapporté un effet indésirable grave imputable aux bisphosphonates : l'ostéonécrose des maxillaires (ONM). Une revue de la littérature portant sur 10 ans de recherche montre que les données épidémiologiques relatives à cette pathologie restent contradictoires. L'incidence des ONM couramment acceptée (0,8 à 12 %) semble sous-estimée en raison des limites méthodologiques des études disponibles (durée de suivi des patients limitée, absence d'examen bucco-dentaire systématique, inhomogénéité des échantillons, etc.). L'absence de consensus autour des facteurs de risque de survenue des ONM résulte de ces mêmes limites méthodologiques. Ce manque de données hypothèque la compréhension des ONM et donc leur prévention et leur prise en charge thérapeutique. Un recueil systématisé et exhaustif des cas de ONM et de leur contexte de survenue (pathologie primaire, données relatives au traitement par bisphosphonates, facteurs de risque, etc.) devrait permettre de pallier ce problème.Abstract -Osteonecrosis of the jaw in patients treated with intravenous bisphosphonates: incidence and risk factors. Intravenous bisphosphonates are effective in the treatment and management of cancer-related conditions including hypercalcemia of malignancy, skeletal-related events associated with bone metastases in the context of solid tumors such as breast cancer, prostate cancer and lung cancer, and management of lytic lesions in the setting of multiple myeloma. In 2003, oral and maxillofacial surgeons first recognized and reported cases of non-healing exposed bone in the maxillofacial region in patients treated with intravenous bisphosphonates. This adverse effect was called bisphosphonate-related osteonecrosis of the jaw (BOJ). A review of literature based on ten years of research highlights a lack of reliable epidemiological data. BOJ incidence commonly accepted (0.8 to 12%) seems to be underestimated. The main raisons for conflicting results concerning BOJ incidence and risk factors are the weakness of study design, the underpowered study with the use of mixed samples (patients with different pathologies and treatments) during short follow-up. To increase level of evidence well designed studies are needed with a longer follow-up of patients, a systematic dental exam by a trained dentist, the set up of homogeneous samples, etc. The lack of reliable data prevents the scientific community from understanding BOJ and so from setting up effective prevention and treatments. A systematic and exhaustive data collection of cases of BOJ and their context of occurrence (primary disease, information concerning the treatment with bisphosphonates, risk factors, etc.) should overcome these problems.