Drugs in the bisphosphonate class are the most commonly prescribed therapeutic agents for the treatment of osteoporosis. Prospective, randomized, placebo-controlled clinical trials have demonstrated efficacy in reducing fracture risk, with favourable safety profiles, in women with postmenopausal osteoporosis. However, long-term use in clinical practice has been associated with reports of undesirable events not previously recognized. These have included gastrointestinal intolerance, osteonecrosis of the jaw, atypical femur fractures, oesophageal cancer, atrial fibrillation and chronic musculoskeletal pain. Physicians must be alert to newly recognized safety concerns, understand the level of evidence supporting them and be able to effectively communicate the balance of expected benefit and potential risk to patients. Usually, post-marketing adverse events are first presented as case reports or observational studies with variable levels of supporting evidence for plausibility, pathophysiology and causality. Widespread coverage in the news media, which can be alarming to patients and their physicians, may not present a balanced view of the proven benefits, the uncertain risks of therapy and the relative magnitude of these events. There may be confusion about the risks associated with bisphosphonate use for the treatment of osteoporosis versus treatment of other conditions, such as cancer, which typically involves a very different patient population and different doses or frequency of drug administration. Often reports of possible adverse events do not provide information on the number of patients exposed to the drug in proportion to the reported adverse event, or do not describe the incidence of the adverse event in a comparator population not exposed to the drug. Gastrointestinal intolerance with oral bisphosphonates, and hypocalcaemia, acute phase reactions and renal toxicity with intravenous bisphosphonates are characterized by biological plausibility and demonstration of causality. Safety concerns with uncertain biological plausibility and unproven causality include osteonecrosis of the jaw, atypical femur fractures, oesophageal cancer and ocular inflammation. Suspected concerns that are unlikely to be causally related include atrial fibrillation and hepatotoxicity. When making the decision to use a bisphosphonate for the treatment of osteoporosis, the balance between benefit and potential risks according to clinical circumstances of each patient should be considered.