The aim of this systematic review was to summarize the evidence about the risks of spinal manipulation. Articles were located through searching three electronic databases (MEDLINE, EMBASE, Cochrane Library), contacting experts (n =9), scanning reference lists of relevant articles, and searching departmental files. Reports in any language containing data relating to risks associated with spinal manipulation were included, irrespective of the profession of the therapist. Where available, systematic reviews were used as the basis of this article. All papers were evaluated independently by the authors. Data from prospective studies suggest that minor, transient adverse events occur in approximately half of all patients receiving spinal manipulation. The most common serious adverse events are vertebrobasilar accidents, disk herniation, and cauda equina syndrome. Estimates of the incidence of serious complications range from 1 per 2 million manipulations to 1 per 400,000. Given the popularity of spinal manipulation, its safety requires rigorous investigation.